The Liver

Dedication

This book is dedicated to Win Arias, whose enthusiasm, insight, tools for building bridges between basic and clinical hepatolo-
and scientific rigor have served as an inspiration to several gen- gists as they elucidate together the mysteries of liver function in
erations of investigators, providing the essential foundation and health and disease.
The Liver
Biology and Pathobiology
Sixth Edition

Edited by
Irwin M. Arias MD
Emeritus Senior Scientist, National Institutes of Health, Bethesda, MD, USA
Emeritus Professor of Physiology, Tufts University School of Medicine, Boston, MA, USA
Emeritus Professor of Medicine, Albert Einstein College of Medicine, Bronx, NY, USA

Harvey J. Alter MD, MACP

Distinguished NIH Scientist Emeritus, Department of Transfusion Medicine,
National Institutes of Health, Bethesda, MD, USA

James L. Boyer MD
Ensign Professor of Medicine, Department of Internal Medicine and Liver Center, Yale University School of Medicine,
New Haven, CT, USA

David E. Cohen MD, PhD

Vincent Astor Distinguished Professor of Medicine
Chief, Division of Gastroenterology and Hepatology, Joan & Sanford I. Weill Department of Medicine,
Weill Cornell Medical College
Co‐Director, Center for Advanced Digestive Care,
New York‐Presbyterian Hospital and Weill Cornell Medical Center, New York, NY, USA

David A. Shafritz MD
Professor of Medicine, Cell Biology & Pathology, Albert Einstein College of Medicine
Associate Director, Marion Bessin Liver Research Center, Bronx, NY, USA

Snorri S. Thorgeirsson MD, PhD

Senior Scientist, Laboratory of Human Carcinogenesis, Center for Cancer Research, National Cancer Institute,
National Institutes of Health, Bethesda, MD, USA

Allan W. Wolkoff MD
The Herman Lopata Chair in Liver Disease Research
Professor of Medicine and Anatomy and Structural Biology
Associate Chair of Medicine for Research
Chief, Division of Hepatology
Director, Marion Bessin Liver Research Center
Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, NY, USA
 This edition first published 2020
© 2020 John Wiley & Sons Ltd

Edition History
John Wiley & Sons Ltd. (5e, 2009)

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is
available at http://www.wiley.com/go/permissions.

The right of Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz, Snorri S. Thorgeirsson, and Allan W. Wolkoff to be
identified as the authors of the editorial material in this work has been asserted in accordance with law.

Registered Offices
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial Office
9600 Garsington Road, Oxford, OX4 2DQ, UK

For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book
may not be available in other formats.

Limit of Liability/Disclaimer of Warranty

The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be
relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In view of ongoing research,
equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and
devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device
for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors
have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents
of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular
purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an
organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and
authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the
understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your
situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or
disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other
commercial damages, including but not limited to special, incidental, consequential, or other damages.

Library of Congress Cataloging‐in‐Publication Data

Names: Arias, Irwin M., editor.
Title: The liver : biology and pathobiology / edited by Irwin M. Arias [and six others].
Other titles: Liver (Arias)
Description: Sixth edition. | Hoboken, NJ : Wiley-Blackwell, 2020. | Includes bibliographical references and index.
Identifiers: LCCN 2019024961 (print) | LCCN 2019024962 (ebook) | ISBN 9781119436829 (hardback) |
  ISBN 9781119436836 (epub) | ISBN 9781119436843 (adobe pdf)
Subjects: MESH: Liver–physiopathology
Classification: LCC QP185 (print) | LCC QP185 (ebook) | NLM WI 702 | DDC 612.3/52–dc23
LC record available at https://lccn.loc.gov/2019024961
LC ebook record available at https://lccn.loc.gov/2019024962

Cover images: main image and background image courtesy of Sandor Paku; top left image courtesy of Scott Friedman;
top middle images courtesy of Sandor Paku; top right image courtesy of Peter Kim
Cover design by Wiley

Set in 9.5/11.5pt Times LT Std by SPi Global, Pondicherry, India

10 9 8 7 6 5 4 3 2 1

ffirs.indd 4 11/30/2019 9:33:25 AM

 Contents

List of Contributors x

Preface xx

Acknowledgments xxi

PART ONE:  INTRODUCTION 1

  1 Organizational Principles of the Liver 3

Peter Nagy, Snorri S. Thorgeirsson, and Joe W. Grisham

  2 Embryonic Development of the Liver 14

Kenneth S. Zaret, Roque Bort, and Stephen A. Duncan

PART TWO:  THE CELLS 23

SECTION A:  CELL BIOLOGY OF THE LIVER 25

  3 Cytoskeletal Motors: Structure and Function in Hepatocytes 27

Mukesh Kumar, Arnab Gupta, and Roop Mallik

  4 Hepatocyte Surface Polarity 36

Anne Müsch and Irwin M. Arias

  5 Primary Cilia 50
Carolyn M. Ott

  6 Endocytosis in Liver Function and Pathology 62

Micah B. Schott, Barbara Schroeder, and Mark A. McNiven

  7 The Hepatocellular Secretory Pathway 75

Catherine L. Jackson and Mark A. McNiven

  8 Mitochondrial Function, Dynamics, and Quality Control 86

Marc Liesa, Ilan Benador, Nathanael Miller, and Orian S. Shirihai

  9 Nuclear Pore Complex 94

Michelle A. Veronin and Joseph S. Glavy

10 Protein Maturation and Processing at the Endoplasmic Reticulum 108

Ramanujan S. Hegde
vi Contents

11 Protein Degradation and the Lysosomal System 122

Susmita Kaushik and Ana Maria Cuervo

12 Peroxisome Assembly, Degradation, and Disease 137

Rong Hua and Peter K. Kim

13 Organelle–Organelle Contacts: Origins and Functions 151

Uri Manor

14 Gap and Tight Junctions in Liver: Structure, Function, and Pathology 160

John W. Murray and David C. Spray

15 Ribosome Biogenesis and its Role in Cell Growth and Proliferation in the Liver 174
Katherine I. Farley‐Barnes and Susan J. Baserga

16 miRNAs and Hepatocellular Carcinoma 183

Yusuke Yamamoto, Isaku Kohama, and Takahiro Ochiya

17 Hepatocyte Apoptosis: Mechanisms and Relevance in Liver Diseases 195

Harmeet Malhi and Gregory J. Gores

SECTION B:  THE HEPATOCYTE 207

18 Copper Metabolism and the Liver 209

Cynthia Abou Zeid, Ling Yi, and Stephen G. Kaler

19 The Central Role of the Liver in Iron Storage and Regulation of Systemic Iron Homeostasis 215
Tracey A. Rouault, Victor R. Gordeuk, and Gregory J. Anderson

20 Disorders of Bilirubin Metabolism 229

Namita Roy Chowdhury, Yanfeng Li, and Jayanta Roy Chowdhury

21 Hepatic Lipid Droplets in Liver Function and Disease 245

Douglas G. Mashek, Wenqi Cui, Linshan Shang, and Charles P. Najt

22 Lipoprotein Metabolism and Cholesterol Balance 255

Mariana Acuña‐Aravena and David E. Cohen

SECTION C:  TRANSPORTERS, BILE ACIDS, AND CHOLESTASIS 269

23 Bile Acid Metabolism in Health and Disease: An Update 271

Tiangang Li and John Y.L. Chiang

24 TGR5 (GPBAR1) in the Liver 286

Verena Keitel, Christoph G.W. Gertzen, Lina Spomer, Holger Gohlke, and Dieter Häussinger

25 Bile Acids as Signaling Molecules 299

Thierry Claudel and Michael Trauner

26 Hepatic Adenosine Triphosphate‐Binding Cassette Transport Proteins and Their Role in Physiology 313
Peter L.M. Jansen

27 Basolateral Plasma Membrane Organic Anion Transporters 327

M. Sawkat Anwer and Allan W. Wolkoff

28 Hepatic Nuclear Receptors 337

Raymond E. Soccio

29 Molecular Cholestasis 351

Paul Gissen and Richard J. Thompson

30 Pathophysiologic Basis for Alternative Therapies for Cholestasis 364

Claudia D. Fuchs, Emina Halilbasic, and Michael Trauner
 Contents vii

31 Adaptive Regulation of Hepatocyte Transporters in Cholestasis 378

James L. Boyer

SECTION D:  NON‐HEPATOCYTE CELLS 391

32 Cholangiocyte Biology and Pathobiology 393

Massimiliano Cadamuro, Romina Fiorotto, and Mario Strazzabosco

33 Polycystic Liver Diseases: Genetics, Mechanisms, and Therapies 408

Tatyana Masyuk, Anatoliy Masyuk, and Nicholas LaRusso

34 The Liver Sinusoidal Endothelial Cell: Basic Biology and Pathobiology 422

Karen K. Sørensen and Bård Smedsrød

35 Fenestrations in the Liver Sinusoidal Endothelial Cell 435

Victoria C. Cogger, Nicholas J. Hunt, and David G. Le Couteur

36 Stellate Cells and Fibrosis 444

Youngmin A. Lee and Scott L. Friedman

PART THREE:  FUNCTIONS OF THE LIVER 455

SECTION A:  METABOLIC FUNCTIONS 457

37 Non‐alcoholic Fatty Liver Disease and Insulin Resistance 459

Max C. Petersen, Varman T. Samuel, Kitt Falk Petersen, and Gerald I. Shulman

38 AMPK: Central Regulator of Glucose and Lipid Metabolism and Target of Type 2 Diabetes Therapeutics 472
Daniel Garcia, Maria M. Mihaylova, and Reuben J. Shaw

39 Insulin‐Mediated PI3K and AKT Signaling 485

Hyokjoon Kwon and Jeffrey E. Pessin

40 Ca2+ Signaling in the Liver 496

Mateus T. Guerra, M. Fatima Leite, and Michael H. Nathanson

41 Clinical Genomics of NAFLD 509

Frank Lammert

SECTION B:  LIVER GROWTH AND REGENERATION 521

42 Stem Cell‐Fueled Maturational Lineages in Hepatic and Pancreatic Organogenesis 523

Wencheng Zhang, Amanda Allen, Eliane Wauthier, Xianwen Yi, Homayoun Hani, Praveen Sethupathy,
David Gerber, Vincenzo Cardinale, Guido Carpino, Juan Dominguez‐Bendala, Giacomo Lanzoni,
Domenico Alvaro, Eugenio Gaudio, and Lola Reid

43 Developmental Morphogens and Adult Liver Repair 539

Mariana Verdelho Machado and Anna Mae Diehl

44 Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 550
David A. Shafritz and Markus Grompe

45 Liver Regeneration 566

George K. Michalopoulos

46 β‐Catenin Signaling 585

Satdarshan P.S. Monga

47 Polyploidy in Liver Function, Mitochondrial Metabolism, and Cancer 603

Evan R. Delgado, Elizabeth C. Stahl, Nairita Roy, Patrick D. Wilkinson, and Andrew W. Duncan
viii Contents

PART FOUR:  PATHOBIOLOGY OF LIVER DISEASE 615

48 Hepatic Encephalopathy 617

Roger F. Butterworth

49 The Kidney in Liver Disease 630

Moshe Levi, Shogo Takahashi, Xiaoxin X. Wang, and Marilyn E. Levi

50 α1‐Antitrypsin Deficiency 645

David A. Rudnick and David H. Perlmutter

51 Pathophysiology of Portal Hypertension 659

Yasuko Iwakiri and Roberto J. Groszmann

52 Non‐alcoholic Fatty Liver Disease: Mechanisms and Treatment 670

Yaron Rotman and Devika Kapuria

53 Alcoholic Liver Disease 682

Bin Gao, Xiaogang Xiang, Lorenzo Leggio, and George F. Koob

54 Drug‐Induced Liver Injury 701

Lily Dara and Neil Kaplowitz

55 Oxidative Stress and Inflammation in the Liver 714

John J. Lemasters and Hartmut Jaeschke

56 The Role of Bile Acid‐Mediated Inflammation in Cholestatic Liver Injury 728

Shi‐Ying Cai, Man Li, and James L. Boyer

57 Toll‐like Receptors in Liver Disease 737

So Yeon Kim and Ekihiro Seki

PART FIVE:  LIVER CANCER 747

58 Experimental Models of Liver Cancer: Genomic Assessment of Experimental Models 749

Sun Young Yim, Jae‐Jun Shim, Bo Hwa Sohn, and Ju‐Seog Lee

59 Epidemiology of Hepatocellular Carcinoma 758

Hashem B. El‐Serag

60 Mutations and Genomic Alterations in Liver Cancer 773

Jessica Zucman‐Rossi and Jean‐Charles Nault

61 Treatment of Liver Cancer 782

Tim F. Greten

PART SIX:  HEPATITIS 793

62 Molecular Biology of Hepatitis Viruses 795

Christoph Seeger, William S. Mason, and Michael M.C. Lai

63 Immune Mechanisms of Viral Clearance and Disease Pathogenesis During Viral Hepatitis 821

Carlo Ferrari, Valeria Barili, Stefania Varchetta, and Mario U. Mondelli

64 Clinical Implications of the Molecular Biology of Hepatitis B Virus 851

Timothy M. Block, Ju‐Tao Guo, and W. Thomas London

65 Viral Escape Mechanisms in Hepatitis C and the Clinical Consequences of Persistent Infection 868
Marc G. Ghany, Christopher M. Walker, and Patrizia Farci

66 Tracking Hepatitis C Virus Interactions with the Hepatic Lipid Metabolism: A Hitchhiker’s

Guide to Solve Remaining Translational Research Challenges in Hepatitis C 889
Gabrielle Vieyres and Thomas Pietschmann
 Contents ix

67 Nucleoside Antiviral Agents for HCV: What’s Left to Do? 906

Franck Amblard, Seema Mengshetti, Junxing Shi, Sijia Tao, Leda Bassit, and Raymond F. Schinazi

68 Hepatitis E Virus: An Emerging Zoonotic Virus Causing Acute and Chronic Liver Disease 915
Xiang‐Jin Meng

69 Biological Principles and Clinical Issues Underlying Liver Transplantation for Viral‐Induced
End‐Stage Liver Disease in the Era of Highly Effective Direct‐Acting Antiviral Agents 926
Michael S. Kriss, James R. Burton, Jr., and Hugo R. Rosen

70 Time for the Elimination of Hepatitis C Virus as a Global Health Threat 935

John W. Ward, Alan R. Hinman, and Harvey J. Alter

PART SEVEN:  HORIZONS 953

71 Genome Editing by Targeted Nucleases and the CRISPR/Cas Revolution 955

Shawn M. Burgess

72 Imaging Cellular Proteins and Structures: Smaller, Brighter, and Faster 965

Aubrey V. Weigel and Erik Lee Snapp

73 Liver‐Directed Gene Therapy 979

Patrik Asp, Chandan Guha, Namita Roy Chowdhury, and Jayanta Roy Chowdhury

74 Telomeres and Telomerase in Liver Generation and Cirrhosis 992

Sonja C. Schätzlein and K. Lenhard Rudolph

75 Toxins and Biliary Atresia 1000

Michael Pack and Rebecca G. Wells

76 The Dual Role of ABC Transporters in Drug Metabolism and Resistance to Chemotherapy 1007
Jean‐Pierre Gillet, Marielle Boonen, Michel Jadot, and Michael M. Gottesman

77 Stem Cell‐Derived Liver Cells: From Model System to Therapy 1015

Helmuth Gehart and Hans Clevers

78 Extracellular Vesicles and Exosomes: Biology and Pathobiology 1022

Gyongyi Szabo and Fatemeh Momen‐Heravi

79 Integrated Technologies for Liver Tissue Engineering 1028

Tiffany N. Vo, Amanda X. Chen, Quinton B. Smith, Arnav Chhabra and Sangeeta N. Bhatia

80 Pluripotent Stem Cells and Reprogramming: Promise for Therapy 1036

James A. Heslop and Stephen A. Duncan

81 Chromatin Regulation and Transcription Factor Cooperation in Liver Cells 1043

Ido Goldstein

82 Drug Interactions in the Liver 1050

Guruprasad P. Aithal and Gerd A. Kullak‐Ublick

83 Metabolic Regulation of Hepatic Growth 1058

Wolfram Goessling

84 The Gut Microbiome and Liver Disease 1062

Lexing Yu, Jasmohan S. Bajaj, and Robert F. Schwabe

85 Lineage Tracing: Efficient Tools to Determine the Fate of Hepatic Cells in Health and Disease 1069
Frédéric Lemaigre

86 The Hepatocyte as a Household for Plasmodium Parasites 1075

Vanessa Zuzarte‐Luis and Maria M. Mota

Index 1081
 List of Contributors

Cynthia Abou Zeid M. Sawkat Anwer

Section on Translational Neuroscience, Molecular Medicine Tufts Clinical and Translational Science Institute
Branch Tufts University Cummings School of Veterinary Medicine
Intramural Research Program, National Institutes of Health, Department of Biomedical Sciences
Bethesda, MD, USA North Grafton, MA, USA
Mariana Acuña‐Aravena Irwin M. Arias
Division of Gastroenterology and Hepatology, Joan & National Institutes of Health
Sanford I. Weill Department of Medicine Bethesda, MD, USA
Weill Cornell Medical College
New York, NY, USA Patrik Asp
Marion Bessin Liver Research Center
Guruprasad P. Aithal Department of Surgery, Albert Einstein College of Medicine
Nottingham Digestive Diseases Centre, School of Medicine, Bronx, NY, USA
University of Nottingham, Nottingham, UK;
NIHR Nottingham Biomedical Research Centre, Nottingham Jasmohan S. Bajaj
University Hospitals NHS Trust and the University of Virginia Commonwealth University and McGuire
Nottingham, Nottingham, UK VA Medical Center
Richmond, VA, USA
Amanda Allen
Department of Cell Biology and Physiology Valeria Barili
University of North Carolina School of Medicine Unit of Infectious Diseases and Hepatology
Chapel Hill, NC, USA Department of Medicine and Surgery, University of Parma
Parma, Italy
Harvey J. Alter
Department of Transfusion Medicine, Clinical Center Susan J. Baserga
National Institutes of Health Department of Molecular Biophysics & Biochemistry
Bethesda, MD, USA Department of Genetics
Department of Therapeutic Radiology, Yale University School
Domenico Alvaro
of Medicine
Department of Medico‐Surgical Sciences and Biotechnologies,
New Haven, CT, USA
Sapienza University of Rome, Latina, Italy
Department of Medicine and Medical Specialties Leda Bassit
Sapienza University of Rome Laboratory of Biochemical Pharmacology, Emory University
Rome, Italy School of Medicine
Atlanta, GA, USA
Franck Amblard
Laboratory of Biochemical Pharmacology, Emory University Ilan Benador
School of Medicine Department of Medicine, Division of Endocrinology and
Atlanta, GA, USA Department of Molecular and Medical Pharmacology
David Geffen School of Medicine at UCLA
Gregory J. Anderson
Los Angeles, CA, USA
QIMR Berghofer Medical Research Institute
Brisbane, Queensland, Australia
 List of Contributors xi

Sangeeta N. Bhatia Guido Carpino

Institute for Medical Engineering and Science, Massachusetts Department of Movement, Human and Health Sciences,
Institute of Technology, Cambridge, MA; Division of Health Sciences
Harvard‐MIT Department of Health Sciences and Technology, University of Rome “Foro Italico”;
Institute for Medical Engineering and Science, Massachusetts Department of Anatomical, Histological, Forensic Medicine
Institute of Technology, Boston, MA; and Orthopedics Sciences
Department of Electrical Engineering and Computer Science, Sapienza University of Rome
Massachusetts Institute of Technology, Cambridge, MA; Rome, Italy
David H. Koch Institute for Integrative Cancer Research,
Amanda X. Chen
Massachusetts Institute of Technology, Cambridge, MA;
Department of Biological Engineering, Massachusetts
Howard Hughes Medical Institute, Chevy Chase, MD, USA
Institute of Technology
Timothy M. Block Cambridge, MA, USA
Baruch S. Blumberg Institute
Arnav Chhabra
Doylestown, PA, USA
Harvard‐MIT Department of Health Sciences and Technology,
Marielle Boonen Institute for Medical Engineering and Science, Massachusetts
Laboratory of Intracellular Trafficking Biology, URPhyM, Institute of Technology
NARILIS Boston, MA, USA
Faculty of Medicine, University of Namur
John Y.L. Chiang
Namur, Belgium
Department of Integrative Medical Sciences
Roque Bort Northeast Ohio Medical University
Instituto de Investigación Sanitaria La Fe (IIS La Fe) Rootstown, OH, USA
Unidad de Hepatología experimental
València, Spain Thierry Claudel
Hans Popper Laboratory of Molecular Hepatology, Division
James L. Boyer of Gastroenterology and Hepatology Department of Internal
Department of Internal Medicine and Liver Center, Medicine III
Yale University School of Medicine Medical University of Vienna
New Haven, CT, USA Vienna, Austria
Shawn M. Burgess Hans Clevers
National Human Genome Research Institute Oncode Institute, Hubrecht Institute, Royal Netherlands
National Institutes of Health Academy of Arts and Sciences (KNAW) and University
Bethesda, MD, USA Medical Centre (UMC) Utrecht;
Princess Máxima Centre for Paediatric Oncology
James R. Burton, Jr.
Utrecht, The Netherlands
University of Colorado School of Medicine
Aurora, CO, USA Victoria C. Cogger
Roger F. Butterworth Centre for Education and Research on Ageing
Department of Medicine University of Sydney and Concord RG Hospital
University of Montreal Sydney, NSW, Australia
Montreal, QC, Canada David E. Cohen
Massimiliano Cadamuro Division of Gastroenterology and Hepatology,
Department of Molecular Medicine Joan & Sanford I. Weill Department of Medicine
University of Padua, Padova, Italy; Weill Cornell Medical College
International Center for Digestive Health (ICDH) New York, NY, USA
University of Milan‐Bicocca, Monza, Italy
Ana Maria Cuervo
Shi‐Ying Cai Department of Developmental and Molecular Biology
The Liver Center, Yale University School of Medicine Institute for Aging Research, Albert Einstein College of
New Haven, CT, USA Medicine
Bronx, NY, USA
Vincenzo Cardinale
Department of Medico‐Surgical Sciences and Wenqi Cui
Biotechnologies Department of Biochemistry, Molecular Biology and Biophysics
Sapienza University of Rome University of Minnesota
Latina, Italy Minneapolis, MN, USA
xii List of Contributors

Lily Dara Liver Center and Section of Digestive Diseases, Department

Research Center for Liver Disease, Department of Medicine, of Internal Medicine, Section of Digestive Diseases, Yale
Division of Gastrointestinal and Liver Diseases University School of Medicine
Keck School of Medicine, University of Southern California New Haven, CT, USA
Los Angeles, CA, USA
Scott L. Friedman
Evan R. Delgado Division of Liver Diseases
Department of Pathology Icahn School of Medicine at Mount Sinai
McGowan Institute for Regenerative Medicine, Pittsburgh New York, NY, USA
Liver Research Center
Claudia D. Fuchs
University of Pittsburgh
Hans Popper Laboratory of Molecular Hepatology, Division
Pittsburgh, PA, USA
of Gastroenterology and Hepatology Department of Internal
Anna Mae Diehl Medicine III
School of Medicine, Duke University Medical University of Vienna
Durham, NC, USA Vienna, Austria
Juan Dominguez‐Bendala Bin Gao
Department of Medicine and Medical Specialties Laboratory of Liver Diseases, National Institute on Alcohol
Sapienza University of Rome Abuse and Alcoholism
Rome, Italy; National Institutes of Health
Diabetes Research Institute, Miller School of Medicine Bethesda, MD, USA
University of Miami
Daniel Garcia
Miami, FL, USA
Molecular and Cell Biology Laboratory
Andrew W. Duncan The Salk Institute for Biological Studies
Department of Pathology, McGowan Institute for Regenerative La Jolla, CA, USA
Medicine, Pittsburgh Liver Research Center, University of
Eugenio Gaudio
Pittsburgh
Department of Anatomical, Histological, Forensic Medicine
Pittsburgh, PA, USA
and Orthopedics Sciences
Stephen A. Duncan Sapienza University of Rome
Department of Regenerative Medicine and Cell Biology Rome, Italy
Medical University of South Carolina
Helmuth Gehart
Charleston, SC, USA
Oncode Institute, Hubrecht Institute, Royal Netherlands
Hashem B. El‐Serag Academy of Arts and Sciences (KNAW) and University
Department of Medicine Medical Centre (UMC) Utrecht
Baylor College of Medicine and Michael E. DeBakey Veterans Utrecht, The Netherlands
Affairs Medical Center
David Gerber
Houston, TX, USA
Department of Surgery
Patrizia Farci University of North Carolina School of Medicine
Hepatic Pathogenesis Section, Laboratory of Infectious Chapel Hill, NC, USA
Diseases, National Institute of Allergy and Infectious Diseases,
Christoph G.W. Gertzen
National Institutes of Health
Institute of Pharmaceutical and Medicinal Chemistry,
Bethesda, MD, USA
Heinrich‐Heine University Düsseldorf, Düsseldorf, Germany
Katherine I. Farley‐Barnes
Marc G. Ghany
Department of Molecular Biophysics & Biochemistry
Liver Diseases Branch, National Institute of Diabetes and
Yale University School of Medicine
Digestive and Kidney Diseases
New Haven, CT, USA
National Institutes of Health
Carlo Ferrari Bethesda, MD, USA
Unit of Infectious Diseases and Hepatology
Jean‐Pierre Gillet
Department of Medicine and Surgery, University of Parma
Laboratory of Molecular Cancer Biology, URPhyM, NARILIS
Parma, Italy
Faculty of Medicine, University of Namur
Romina Fiorotto Namur, Belgium
International Center for Digestive Health (ICDH), University
Paul Gissen
of Milan‐Bicocca
UCL Great Ormond Street Institute of Child Health and Great
Monza, Italy;
Ormond Street Hospital for Children, London, UK
 List of Contributors xiii

Joseph S. Glavy Chandan Guha

University of Texas at Tyler Marion Bessin Liver Research Center, Albert Einstein College
Fisch College of Pharmacy of Medicine, Bronx, NY;
Tyler, TX, USA Departments of Radiation Oncology and Pathology, Albert
Einstein College of Medicine
Wolfram Goessling
Bronx, NY, USA
Division of Gastroenterology
Massachusetts General Hospital Ju‐Tao Guo
Harvard‐MIT Division of Health Sciences and Technology Baruch S. Blumberg Institute
Harvard Medical School Doylestown, PA, USA
Boston, MA, USA
Arnab Gupta
Holger Gohlke Department of Biological Sciences, Indian Institute of Science
Institute of Pharmaceutical and Medicinal Chemistry, Education and Research Kolkata, Mohanpur, India
Heinrich‐Heine University Düsseldorf, Düsseldorf, Germany
Emina Halilbasic
Ido Goldstein Hans Popper Laboratory of Molecular Hepatology, Division
Institute of Biochemistry, Food Science and Nutrition of Gastroenterology and Hepatology Department of Internal
The Robert H. Smith Faculty of Agriculture, Food and Medicine III
Environment Medical University of Vienna
The Hebrew University of Jerusalem Vienna, Austria
Rehovot, Israel
Homayoun Hani
Victor R. Gordeuk Department of Cell Biology and Physiology
University of Illinois at Chicago University of North Carolina School of Medicine
Chicago, IL, USA Chapel Hill, NC, USA
Gregory J. Gores
Dieter Häussinger
College of Medicine, Division of Gastroenterology
Clinic for Gastroenterology, Hepatology and Infectious
and Hepatology
Diseases, University Hospital Düsseldorf Medical Faculty at
Mayo Clinic
Heinrich‐Heine‐University
Rochester, MN, USA
Düsseldorf, Germany
Michael M. Gottesman
Ramanujan S. Hegde
Laboratory of Cell Biology, Center for Cancer Research
MRC Laboratory of Molecular Biology
National Cancer Institute, National Institutes of Health
Cambridge, UK
Bethesda, MD, USA
Tim F. Greten James A. Heslop
Thoracic and Gastrointestinal Malignancies Branch Department of Regenerative Medicine and Cell Biology
Center for Cancer Research, National Cancer Institute Medical University of South Carolina
Bethesda, MD, USA Charleston, SC, USA

Joe W. Grisham Alan R. Hinman

Department of Pathology and Laboratory Medicine Task Force for Global Health
University of North Carolina Decatur, GA, USA
Chapel Hill, NC, USA Rong Hua
Markus Grompe Cell Biology Program, Hospital for Sick Children
Papé Pediatric Research Institute Toronto, ON, Canada
Oregon Health Sciences University
Nicholas J. Hunt
Portland, OR, USA
Centre for Education and Research on Ageing
Roberto J. Groszman University of Sydney and Concord RG Hospital
Section of Digestive Diseases, Yale School of Medicine Sydney, NSW, Australia
New Haven, CT, USA
Yasuko Iwakiri
Mateus T. Guerra Section of Digestive Diseases
Departments of Medicine and Cell Biology Yale University School of Medicine
Yale University School of Medicine New Haven, CT, USA
New Haven, CT, USA
xiv List of Contributors

Catherine L. Jackson Isaku Kohama

Institut Jacques Monod, UMR7592 CNRS Université Paris‐ Division of Molecular and Cellular Medicine, National Cancer
Diderot, Sorbonne Paris Cité Center Research Institute
Paris, France Tsukiji, Tokyo, Japan
Michel Jadot George F. Koob
Laboratory of Physiological Chemistry, URPhyM, NARILIS, National Institute on Alcohol Abuse and Alcoholism and
Faculty of Medicine National Institute on Drug Abuse
University of Namur, Belgium National Institutes of Health
Bethesda, MD, USA
Hartmut Jaeschke
Department of Pharmacology, Toxicology and Therapeutics Michael S. Kriss
University of Kansas Medical Center Division of Gastroenterology & Hepatology
Kansas City, KS, USA University of Colorado School of Medicine
Peter L.M. Jansen Aurora, CO, USA
Department of Hepatology and Gastroenterology Gerd A. Kullak‐Ublick
Academic Medical Center University Hospital Zurich and University of Zurich
Amsterdam, The Netherlands; Zurich, Zurich, Switzerland;
LiSyM research network Mechanistic Safety, Chief Medical Office and Patient Safety,
University of Freiburg Novartis Global Drug Development
Freiburg, Germany Basel, Switzerland
Stephen G. Kaler
Mukesh Kumar
Section on Translational Neuroscience, Molecular
Department of Biological Sciences, Tata Institute of
Medicine Branch
Fundamental Research, Navy Nagar, Colaba, Mumbai, India
Intramural Research Program, National Institutes of Health
Bethesda, MD, USA Hyokjoon Kwon
Neil Kaplowitz Department of Medicine, Division of Endocrinology,
Research Center for Liver Disease, Department of Medicine, Metabolism and Nutrition
Division of Gastrointestinal and Liver Diseases Rutgers‐Robert Wood Johnson Medical School
Keck School of Medicine, University of Southern California New Brunswick, NJ, USA
Los Angeles, CA, USA Michael M.C. Lai
Devika Kapuria China Medical University and China Medical
Liver Energy and Metabolism Section, Liver Diseases Branch, University Hospital
National Institute of Diabetes and Digestive and Kidney Taichung, Taiwan
Diseases
Frank Lammert
National Institutes of Health
Department of Medicine II
Bethesda, MD, USA
Saarland University Medical Center
Susmita Kaushik Homburg, Germany
Department of Developmental and Molecular Biology
Institute for Aging Research, Albert Einstein College of Giacomo Lanzoni
Medicine Diabetes Research Institute, Miller School of Medicine
Bronx, NY, USA University of Miami
Miami, FL, USA
Verena Keitel
Clinic for Gastroenterology, Hepatology and Infectious Nicholas LaRusso
Diseases, University Hospital Düsseldorf Medical Faculty at Division of Gastroenterology and Hepatology
Heinrich‐Heine‐University Mayo Clinic College of Medicine
Düsseldorf, Germany Rochester, MN, USA

Peter K. Kim David G. Le Couteur

Cell Biology Program, The Hospital for Sick Children; Centre for Education and Research on Ageing
Department of Biochemistry, University of Toronto University of Sydney and Concord RG Hospital
Toronto, ON, Canada Sydney, NSW, Australia

So Yeon Kim Ju‐Seog Lee

Division of Digestive and Liver Diseases, Department of Department of Systems Biology
Medicine The University of Texas M.D. Anderson Cancer Center
Cedars‐Sinai Medical Center Houston, TX, USA
Los Angeles, CA, USA
 List of Contributors xv

Youngmin A. Lee Mariana Verdelho Machado

Department of Surgery, Vanderbilt University Medical Center Gastroenterology and Hepatology Department
Nashville, TN, USA Hospital de Santa Maria
CHLN, Lisbon;
Lorenzo Leggio Faculty of Medicine
Section on Clinical Psychoneuroendocrinology and Lisbon University
Neuropsychopharmacology, National Institute on Alcohol Lisbon, Portugal
Abuse and Alcoholism and National Institute on Drug Abuse
National Institutes of Health Harmeet Malhi
Bethesda, MD, USA College of Medicine, Division of Gastroenterology and Hepatology
Mayo Clinic
M. Fatima Leite Rochester, MN, USA
Department of Physiology and Biophysics
UFMG Roop Mallik
Belo Horizonte, Brazil Department of Biological Sciences, Tata Institute of
Fundamental Research, Navy Nagar, Colaba, Mumbai, India
Frédéric Lemaigre
Université catholique de Louvain, de Duve Institute Uri Manor
Brussels, Belgium Waitt Advanced Biophotonics Center, Salk Institute for
Biological Studies
John J. Lemasters La Jolla, CA, USA
Departments of Drug Discovery and Biomedical Sciences and
Biochemistry and Molecular Biology Douglas G. Mashek
Medical University of South Carolina Department of Biochemistry, Molecular Biology and
Charleston, SC, USA Biophysics and Department of Medicine, Division of Diabetes,
Endocrinology and Metabolism
Marilyn E. Levi University of Minnesota
Department of Medicine, Division of Infectious Diseases, Minneapolis, MN, USA
University of Colorado
Aurora, CO, USA William S. Mason
Fox Chase Cancer Center
Moshe Levi Philadelphia, PA, USA
Department of Biochemistry and Molecular and Cellular
Biology Anatoliy Masyuk
Georgetown University Division of Gastroenterology and Hepatology
Washington, DC, USA Mayo Clinic College of Medicine
Rochester, MN, USA
W. Thomas London (deceased)
Formerly, Fox Chase Cancer Center Tatyana Masyuk
Philadelphia, PA, USA Division of Gastroenterology and Hepatology
Mayo Clinic College of Medicine
Man Li Rochester, MN, USA
The Liver Center, Yale University School of Medicine
New Haven, CT, USA Mark A. McNiven
Department of Biochemistry and Molecular Biology, Division
Tiangang Li of Gastroenterology and Hepatology, Mayo Clinic
Department of Pharmacology, Toxicology and Therapeutics Rochester, MN, USA
University of Kansas Medical Center
Kansas City, KS, USA Xiang‐Jin Meng
Department of Biomedical Sciences and Pathobiology
Yanfeng Li Virginia Polytechnic Institute and State University
Departments of Medicine and Genetics and Marion Bessin Blacksburg, VA, USA
Liver Research Center
Albert Einstein College of Medicine, Bronx, NY, USA Seema Mengshetti
Laboratory of Biochemical Pharmacology, Emory University
Marc Liesa School of Medicine
Department of Medicine, Division of Endocrinology and Atlanta, GA, USA
Department of Molecular and Medical Pharmacology
David Geffen School of Medicine at UCLA George K. Michalopoulos
Los Angeles, CA, USA Department of Pathology
University of Pittsburgh School of Medicine
Pittsburgh, PA, USA
xvi List of Contributors
Maria M. Mihaylova
Molecular and Cell Biology Laboratory
The Salk Institute for Biological Studies
La Jolla, CA, USA
Nathanael Miller
Department of Medicine, Division of Endocrinology and
Department of Molecular and Medical Pharmacology
David Geffen School of Medicine at UCLA
Los Angeles, CA, USA
Fatemeh Momen‐Heravi
Department of Medicine
University of Massachusetts Medical School
Worcester, MA, USA
Mario U. Mondelli
Division of Infectious Diseases and Immunology, Fondazione
IRCCS Policlinico S.Matteo, Pavia;
Department of Internal Medicine and Therapeutics
University of Pavia, Pavia, Italy
Satdarshan P.S. Monga
Pittsburgh Liver Research Center
University of Pittsburgh, School of Medicine and UPMC
Pittsburgh, PA, USA
Maria M. Mota
Instituto de Medicina Molecular João Lobo Antunes,
Faculty of Medicine, University of Lisbon, Lisbon,
Portugal
Anne Müsch
Department of Developmental & Molecular Biology
Albert Einstein College of Medicine
Bronx, NY, USA
John W. Murray
Marion Bessin Liver Research Center
Albert Einstein College of Medicine;
Department of Anatomy and Structural Biology, Albert
Einstein College of Medicine
Bronx, NY, USA
Peter Nagy
First Department of Pathology and Experimental Cancer
Research, Semmelweis University, Budapest, Hungary
Charles P. Najt
Department of Biochemistry, Molecular Biology and
Biophysics
University of Minnesota
Minneapolis, MN, USA
Michael H. Nathanson
Departments of Medicine and Cell Biology
Yale University School of Medicine
New Haven, CT, USA
Jean‐Charles Nault
Centre de Recherche des Cordeliers, Sorbonne Université,
Inserm, Université de Paris, Functional Genomics of Solid
Tumors Laboratory, Paris, France
Liver Unit, Hôpital Jean Verdier, Hôpitaux Universitaires
Paris‐Seine‐Saint‐Denis, Assistance‐Publique Hôpitaux de
Paris, Bondy, France
Takahiro Ochiya
Division of Molecular and Cellular Medicine, National Cancer
Center Research Institute
Tsukiji, Tokyo
Institute of Medical Science, Tokyo Medical University
Shinjuku, Tokyo, Japan
Janelia Research Campus, Ashburn, VA, USA
Carolyn M. Ott
Janelia Research Campus
Ashburn, VA, USA
Michael Pack
Departments of Medicine (GI Division) and Cell and
Developmental Biology
Perelman School of Medicine, University of Pennsylvania
Philadelphia, PA, USA
David H. Perlmutter
Department of Pediatrics, Division of Gastroenterology,
Hepatology, and Nutrition and Department of
Developmental Biology, Washington University School
of Medicine in St. Louis, St. Louis Children’s Hospital,
St. Louis, MO, USA
Jeffrey E. Pessin
Albert Einstein‐Mount Sinai Diabetes Research Center and
the Fleischer Institute for Diabetes and Metabolism, Albert
Einstein College of Medicine, Bronx, NY;
Department of Medicine and Department of Molecular
Pharmacology, Albert Einstein College of Medicine,
Bronx, NY, USA
Kitt Falk Petersen
Department of Internal Medicine, Section of Endocrinology
Yale University School of Medicine
New Haven, CT, USA
Max C. Petersen
Department of Internal Medicine, Section of Endocrinology,
and Department of Molecular and Cellular Physiology
Yale University School of Medicine
New Haven, CT, USA
Thomas Pietschmann
Institute of Experimental Virology, TWINCORE,
Centre for Experimental and Clinical Infection Research,
Hannover;
German Centre for Infection Research (DZIF)
Braunschweig, Germany
Lola Reid
Department of Cell Biology and Physiology
Program in Molecular Biology and Biotechnology
University of North Carolina School of Medicine
Chapel Hill, NC, USA
Hugo R. Rosen
University of Southern California Keck School of Medicine
Los Angeles, CA, USA
 List of Contributors xvii

Yaron Rotman Barbara Schroeder

Liver Energy and Metabolism Section, Liver Diseases Branch, Department of Biochemistry and Molecular Biology, Division
National Institute of Diabetes and Digestive and Kidney of Gastroenterology and Hepatology, Mayo Clinic
Diseases Rochester, MN, USA
National Institutes of Health
Bethesda, MA, USA Robert F. Schwabe
Columbia University
Tracey A. Rouault New York, NY, USA
Eunice Kennedy Shriver National Institute of Child Health and
Human Development Christoph Seeger
National Institutes of Health Fox Chase Cancer Center
Bethesda, MD, USA Philadelphia, PA, USA

Nairita Roy Ekihiro Seki

Department of Pathology Division of Digestive and Liver Diseases, Department of
McGowan Institute for Regenerative Medicine, Pittsburgh Medicine, Cedars‐Sinai Medical Center, Los Angeles, CA;
Liver Research Center Department of Biomedical Sciences, Cedars‐Sinai
University of Pittsburgh Medical Center,
Pittsburgh, PA, USA Los Angeles, CA, USA
Praveen Sethupathy
Jayanta Roy Chowdhury
Department of Genetics
Departments of Medicine and Genetics and Marion Bessin
University of North Carolina School of Medicine
Liver Research Center
Chapel Hill, NC;
Albert Einstein College of Medicine
Department of Biomedical Sciences
Bronx, NY, USA
Cornell University
Namita Roy Chowdhury Ithaca, NY, USA
Departments of Medicine and Genetics and Marion Bessin
David A. Shafritz
Liver Research Center
Marion Bessin Liver Research Center
Albert Einstein College of Medicine
Albert Einstein College of Medicine
Bronx, NY, USA
Bronx, NY, USA
David A. Rudnick
Linshan Shang
Department of Pediatrics, Division of Gastroenterology,
Department of Biochemistry, Molecular Biology and Biophysics
Hepatology, and Nutrition and Department of Developmental
University of Minnesota
Biology, Washington University School of Medicine in St.
Minneapolis, MN, USA
Louis, St. Louis Children’s Hospital St. Louis, MO, USA
Reuben J. Shaw
K. Lenhard Rudolph
Molecular and Cell Biology Laboratory
Research Group on Stem Cell Aging, Leibniz Institute on
The Salk Institute for Biological Studies
Aging – Fritz Lipmann Institute (FLI)
La Jolla, CA, USA
Jena, Germany
Junxing Shi
Varman T. Samuel Laboratory of Biochemical Pharmacology, Emory University
Department of Internal Medicine, Section of Endocrinology School of Medicine
Yale University School of Medicine Atlanta, GA, USA
New Haven;
Veterans Affairs Medical Center Jae‐Jun Shim
West Haven, CT, USA Department of Internal Medicine, Kyung Hee University
College of Medicine
Sonja C. Schätzlein Seoul, Korea
Spark@FLI, Leibniz Institute on Aging – Fritz Lipmann
Institute (FLI) Orian S. Shirihai
Jena, Germany Department of Medicine, Division of Endocrinology and
Department of Molecular and Medical Pharmacology
Raymond F. Schinazi David Geffen School of Medicine at UCLA
Laboratory of Biochemical Pharmacology, Emory University Los Angeles, CA, USA
School of Medicine
Atlanta, GA, USA Gerald I. Shulman
Department of Internal Medicine, Section of Endocrinology
Micah B. Schott Department of Molecular and Cellular Physiology and Howard
Department of Biochemistry and Molecular Biology, Division Hughes Medical Institute
of Gastroenterology and Hepatology, Mayo Clinic Yale University School of Medicine
Rochester, MN, USA New Haven, CT, USA
xviii List of Contributors

Bård Smedsrød Gyongyi Szabo

Vascular Biology Research Group, Department of Beth Israel Deaconess Medical Center
Medical Biology Harvard Medical School
UiT The Arctic University of Norway Boston, MA, USA
Tromsø, Norway
Shogo Takahashi
Quinton B. Smith Department of Biochemistry and Molecular and Cellular Biology
Institute for Medical Engineering and Science, Massachusetts Georgetown University
Institute of Technology, Cambridge, MA, USA Washington, DC, USA
Erik Lee Snapp Sijia Tao
Janelia Research Campus of the Howard Hughes Medical Laboratory of Biochemical Pharmacology, Emory University
Institute School of Medicine
Ashburn, VA, USA Atlanta, GA, USA
Raymond E. Soccio Richard J. Thompson
University of Pennsylvania Institute of Liver Studies, King’s College London
Perelman School of Medicine, Department of Medicine London, UK
Division of Endocrinology, Diabetes, and Metabolism
Institute for Diabetes, Obesity, and Metabolism Snorri S. Thorgeirsson
Philadelphia Crescenz VA Medical Center Laboratory of Human Carcinogenesis, Center for Cancer
Philadelphia, PA, USA Research, National Cancer Institute, NIH, Bethesda, MD, USA
Bo Hwa Sohn Michael Trauner
Department of Systems Biology Hans Popper Laboratory of Molecular Hepatology, Division
The University of Texas M.D. Anderson Cancer Center of Gastroenterology and Hepatology Department of Internal
Houston, TX, USA Medicine III
Medical University of Vienna
Karen K. Sørensen
Vienna, Austria
Vascular Biology Research Group, Department of Medical
Biology Stefania Varchetta
UiT The Arctic University of Norway Division of Infectious Diseases and Immunology, Fondazione
Tromsø, Norway IRCCS Policlinico S.Matteo, Pavia; Department of Internal
Medicine and Therapeutics
Lina Spomer
University of Pavia, Italy
Clinic for Gastroenterology, Hepatology and Infectious
Diseases, University Hospital Düsseldorf Medical Faculty at Michelle A. Veronin
Heinrich‐Heine‐University University of Texas at Tyler
Düsseldorf, Germany Fisch College of Pharmacy
David C. Spray Tyler, TX, USA
Marion Bessin Liver Research Center, Albert Einstein Gabrielle Vieyres
College of Medicine Institute of Experimental Virology, TWINCORE, Centre for
Department of Neuroscience, Albert Einstein College of Experimental and Clinical Infection Research
Medicine Hannover, Germany
Department of Medicine, Albert Einstein College of Medicine
Bronx, NY, USA Tiffany N. Vo
Institute for Medical Engineering and Science, Massachusetts
Elizabeth C. Stahl Institute of Technology
Department of Pathology Cambridge, MA, USA
McGowan Institute for Regenerative Medicine, Pittsburgh
Liver Research Center Christopher M. Walker
University of Pittsburgh Center for Vaccines and Immunity, The Research Institute at
Pittsburgh, PA, USA Nationwide Children’s Hospital, and College of Medicine, The
Ohio State University
Mario Strazzabosco
Columbus, OH, USA
International Center for Digestive Health (ICDH), University
of Milan‐Bicocca, Monza, Italy; Xiaoxin X. Wang
Liver Center and Section of Digestive Diseases, Department Department of Biochemistry and Molecular and Cellular Biology
of Internal Medicine, Section of Digestive Diseases, Yale Georgetown University
University School of Medicine, New Haven, CT, USA Washington, DC, USA
 List of Contributors xix

John W. Ward Ling Yi

Task Force for Global Health, Decatur, GA; Section on Translational Neuroscience, Molecular
Centers for Disease Control and Prevention Medicine Branch
Atlanta, GA, USA Intramural Research Program, National Institutes of Health,
Bethesda, MD, USA
Eliane Wauthier
Department of Cell Biology and Physiology Xianwen Yi
University of North Carolina School of Medicine Department of Surgery
Chapel Hill, NC, USA University of North Carolina School of Medicine
Chapel Hill, NC, USA
Aubrey V. Weigel
Janelia Research Campus of the Howard Hughes Medical Institute Sun Young Yim
Ashburn, VA, USA Department of Internal Medicine, Korea University
College of Medicine
Rebecca G. Wells
Seoul, Korea
Departments of Medicine (GI Division) and Pathology and
Laboratory Medicine, Perelman School of Medicine, and Lexing Yu
Bioengineering, School of Engineering and Applied Sciences Second Military Medical University
University of Pennsylvania Shanghai, China
Philadelphia, PA, USA
Kenneth S. Zaret
Patrick D. Wilkinson Institute for Regenerative Medicine, Department of Cell and
Department of Pathology Developmental Biology
McGowan Institute for Regenerative Medicine, Pittsburgh Perelman School of Medicine, University of Pennsylvania
Liver Research Center Philadelphia, PA, USA
University of Pittsburgh
Wencheng Zhang
Pittsburgh, PA, USA
Department of Cell Biology and Physiology
Allan W. Wolkoff University of North Carolina School of Medicine
Division of Gastroenterology and Liver Diseases Chapel Hill, NC, USA
Marion Bessin Liver Research Center
Jessica Zucman‐Rossi
Albert Einstein College of Medicine and Montefiore
Centre de Recherche des Cordeliers, Sorbonne Université,
Medical Center
Inserm, Université de Paris, Functional Genomics of Solid
Bronx, NY, USA
Tumors Laboratory, Paris, France;
Xiaogang Xiang Hôpital Européen Georges Pompidou, Assistance Publique‐
Laboratory of Liver Diseases, National Institute on Alcohol Hôpitaux de Paris
Abuse and Alcoholism Paris, France
National Institutes of Health
Vanessa Zuzarte‐Luis
Bethesda, MD, USA;
Instituto de Medicina Molecular João Lobo Antunes
Department of Infectious Diseases, Ruijin Hospital,
Faculty of Medicine, University of Lisbon
School of Medicine
Lisbon, Portugal
Shanghai Jiao Tong University
Shanghai, China
Yusuke Yamamoto
Division of Molecular and Cellular Medicine, National Cancer
Center Research Institute
Tsukiji, Tokyo, Japan
 Preface
The pace of discoveries in basic biomedical sciences and engi-
neering and their application to diagnosis and treatment of liver
disease continues to exceed greatly the expectations expressed
in the Preface to the previous editions published in 1982, 1988,
1994, 1999, and 2009. Concomitantly, the challenge addressed
by this book has not changed since first appearing in the Preface
to the first edition over 30 years ago:
The amazing advances in fundamental biology that have
occurred within the past two decades have brought hepatol-
ogy and other disciplines into new, uncharted and exciting
waters. The dynamic changes in biology will profoundly influ-
ence our ability to diagnose, treat and prevent liver disease.
How can a student of the liver and its diseases maintain a link
to these exciting advances? Most physicians lack the time to
take post‐graduate courses in basic biology; most basic
researchers lack an understanding of liver physiology and
disease. This book strives to bridge the ever‐increasing gap
between the advances in basic biology and their application
to liver structure, function and disease.
Molecular biology was not the only great wave in contempo-
rary science, nor is it surely the last. Remarkable advances in
genetics and various omics are increasingly linked with dynamic
super‐resolution light microscopy, which permits the study of
cellular, molecular, and organ‐based physiology at nano‐levels.
The expanding worlds of RNA structure and function, CRISPR‐
type gene editing, and chromatin biology coupled with single‐
cell and single‐molecule genomic analyses are facilitating
discoveries with great importance in organ physiology and
medicine, including personalized diagnosis and treatment.
­ David A. Shafritz
Unexpected discoveries are certain to emerge from the ongoing
bridge‐building between chemical and physical structural
analysis, engineered drug design, signaling networks, immune
mechanisms and tolerance, the brain, and metabolic/digestive
functions. Discoveries in these disciplines have already facili-
tated diagnosis, treatment, and improved clinical outcome of
many liver diseases. Much more is undoubtedly yet to come.
This sixth edition contains new chapters that present major
progress that has been achieved in research laboratories and
clinics around the world. All other chapters have been com-
pletely revised and updated. Following the death of our col-
league Nelson Fausto, Snorri Thorgeirsson became an Associate
Editor. Previous editions included a section called “Horizons,”
devoted to extraordinary advances in areas of potentially major
importance to the liver. Virtually all of these fields have rapidly
expanded and become topics for later chapters. Sixteen new
“Horizons” chapters are presented in this edition. One may
safely predict that their impact on the field of hepatology will be
considerable. As stated in the Preface to previous editions:
The amazing advance in science proceeds at an ever‐increasing
pace. The implications for students of liver disease are
­considerable. The authors and editors will have achieved our
goals if the reader finds within this volume glimpses into the cur-
rent state and future direction of our discipline and p­ erspectives
that lead to better understanding of liver ­function and disease.
Irwin M. Arias
Harvey J. Alter
James L. Boyer
David E. Cohen
Snorri S. Thorgeirsson
Allan W. Wolkoff
 Acknowledgments

We thank the distinguished authors for their expertise, of Wiley Blackwell and also to the freelance project manager
­enthusiastic participation, and patience in responding to edi- Gillian Whitley.
torial suggestions. Appreciation is also extended to the staff
PART ONE:
INTRODUCTION
 Organizational Principles
1 of the Liver
Peter Nagy1, Snorri S. Thorgeirsson2, and Joe W. Grisham3
1
First Department of Pathology and Experimental Cancer Research, Semmelweis University, Budapest, Hungary
2
Laboratory of Human Carcinogenesis, Center for Cancer Research, National Cancer Institute, NIH, Bethesda,
MD, USA
3
Department of Pathology and Laboratory Medicine, University of North Carolina, Chapel Hill, NC, USA

PRINCIPLES OF LIVER STRUCTURE
AND FUNCTION
The liver is the largest organ of the mammalian body and has a
highly versatile and complex function. Its specialized role is
shown by the fact that, despite intense efforts, the activity of the
liver cannot be replaced by artificial equipment. The liver par­
ticipates in the maintenance of the organism’s homeostasis as
an  active, bidirectional biofilter. It is classed as bidirectional
because it filters the portal blood that transports nutritional and
toxic compounds from the environment through the gastrointes­
tinal tract and also filters the systemic blood (the body’s own
products, e.g. bilirubin), providing the only channel of the body,
the biliary system, through which non‐water‐soluble substances
can be removed. It is classed as an active filter because it rapidly
metabolizes most nutritional compounds and neutralizes and
prepares for removal toxic exogenous (xenobiotics) and endog­
enous (worn out) materials. Because of these major functions
the liver is constantly exposed to intense microbiological and
antigenic stimuli which require function of the innate and adap­
tive immune systems. These diversified functions are executed
by a structurally complex, multicellular tissue with a unique
angioarchitecture, and by the combined and integrated activities
of the participants.
There are only two unique cell types in the liver – hepatocytes
and biliary cells (or cholangiocytes). The hepatocytes are “the
most valuable” parenchymal cells of the hepatic tissue. They do
not constitute a homogeneous cell population. They are highly
polarized cells (i.e. molecular specializations of the various
­surface membranes, including receptors, pumps, transport chan­
nels and carrier proteins) and their functions and to a certain
extent morphology depend on their location in the parenchyma.
This polarization makes the hepatocytes the logical center of the
liver. In addition, they perform the most complex metabolic
tasks of the mammalian organism.
The cholangiocytes form the channels that constitute the
­biliary system, which drains the parenchyma and guarantees the
permanent flow of the bile, a highly toxic solution. Cholangiocytes
also modify the composition of the bile and, in case of adverse
conditions, can participate in repair mechanisms. These liver
cells could not carry out their specific functions, of course,
­without the support of several “communal” cell types, which
are  highly adapted to the special function and architecture of
the liver. The endothelial cells of the parenchyma have a unique
fenestrated structure and various different ­subpopulations can
be  distinguished. There are several subpopulations of hepatic
myofibroblasts as well. In addition to their mechanical functions
the myofibroblasts can store special substances (e.g. vitamin
A in stellate cells) and are a major source of growth factors and
cytokines. The Kupffer cells are the resident macrophages in the
liver. In addition to filtering the blood, they perform their tradi­
tional immunregulatory function. The presence of almost all
subtypes of lymphocytes and dendritic cells makes the liver the
largest organ of the immune system. The mesothelial cells of the
Glisson capsule are, beside their mechanical function, an impor­
tant source of lymph production and can contribute to the gen­
eration of other hepatic cell types. The features of the hepatic
extracellular matrix are unique. The components of the basement
membrane are present around the sinusoids in an “unstructured”
fashion, and cannot be detected by electron microscope, yet they
can perform certain functions.
Another fundamental feature of liver organization is its
unique vascular pattern. Two afferent vessels supply blood to
the liver: the portal vein and the hepatic artery. The blood of
the  portal vein, having already “drained” the stomach, gut,
­pancreas, and spleen, is reduced in oxygen and pressure, and is

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
4 THE LIVER:  FUNCTIONAL ANATOMY OF THE LIVER

enriched in nutrients and toxic materials absorbed from the hepatocytes runs in a centrifugal direction in the bile canalicules
alimentary tract and in viscerally generated hormones and
­ formed by neighboring hepatocytes and is collected by the inter­
growth factors. The arterial blood of the hepatic artery has sys­ lobular bile ducts of the portal triads. There is thus a countercur­
temic levels of oxygen, pressure, and composition. The major rent between the flow of the blood and bile at lobular level.
function of the hepatic artery is to supply the peribiliary vascu­
lar plexus, the portal tract interstitium, the hepatic capsule, and
the vasa vasorum of major vessels. In some species, the hepatic
artery forms anastomosis with the branches of the portal vein, FUNCTIONAL ANATOMY OF THE LIVER
but even then this blood also ends up in the sinusoids. The blood
of the liver is collected by one efferent draining system, the Macroanatomy
hepatic or “central” veins, which reach the systemic circulation
The liver is a continuous sponge‐like parenchymal mass
via the inferior vena cava. The sinusoids form a very special
­penetrated by tunnels (lacunae) that contain the interdigitating
vascular system, which is interposed between the afferent and
networks of afferent and efferent vessels [1]. The adult human
efferent vessels. The large number and capacity of the sinusoids
liver weighs from 1300 to 1700 g, depending on sex and body
and the special arrangement of the supplying vessels provide a
size. It is relatively small compared to other species (2% of the
large volume of blood at a high flow rate via the large vessels
body weight) – in rat and mouse the liver is 4–5% of the body.
with high compliance and capacity. At the same time the
In most mammalian species the liver is multilobed, the indi­
­sinusoids are perfused with blood at low pressure and flow rate.
vidual lobes reflecting the distribution of the major branches of
These arrangements (i.e. low flow, specifically fenestrated
afferent and efferent blood vessels. In contrast, the human liver
(­perforated) endothelial cells, and the lack of the structured
parenchyma is fused into a continuous parenchymal mass with
basement membrane) provide an especially efficient communi­
two major lobes, right and left, delineated only by being supplied
cation between the blood and hepatocytes. This is well illus­
and drained by separate first‐ and second‐order branches of the
trated by the pathological condition of liver cirrhosis, when the
portal and hepatic veins. Right and left lobes are topographically
changes in hemodynamic condition (i.e. the “capillarization” of
separated by the remnants of the embryonic umbilical vein (the
the sinusoids) disrupts this communication, resulting in severe
falciform ligament), but this landmark does not locate the true
dysfunction of the liver.
anatomic division. Anatomically, the medial segment of the left
Bile acids and their enterohepatic circulation are another good
lobe is located to the right side of the falciform ligament, ­centered
example of the cumulating functions. The bile acids are synthe­
on the anterior branches of the left portal vein. Interdigitation of
tized in the hepatocytes by a complex biochemical process that
first‐ and second‐order branches of the portal and hepatic veins
requires 16 different enzymes, which are further modified by the
produces eight macrovascular parenchymal segments centered
gut microbiota. The primary physiological function of the bile
on large portal veins and separated by large hepatic veins [2].
acids is to convert lipid bilayers into micelles. This makes pos­
Hemodynamic watersheds or fissures separating afferent and
sible the excretion of important waste products from the blood.
efferent macrovascular segments permit the surgical resection of
The bile acids also emulsify elements of the food in the gut and
individual or adjacent segments.
aid their ­absorption. In addition, bile acids act as signaling mol­
Liver transplantation and surgery has reached such a com­
ecules, synchronizing the cooperation of the liver and gut.
plexity, however, that the traditional eight‐segment scheme is no
The different types of cells and vessels mentioned above can
longer sufficient. Detailed histological and imaging investiga­
operate only if they are organized in a well “designed” structure.
tions have revealed that the number of second‐order branches
The most widely studied and analyzed morphological and func­
given off by the left and right portal veins is much higher, and
tional unit or module of the liver is the hepatic lobule. The popu­
the mean of their number is 20, leading to the “1–2–20” concept
larity of this structure for studies can be partly explained by the
of portal venous segmentation [3]. The recognition of the water­
fact that lobules are outlined nicely in some species (pig, camel,
shed septa between the variable actual segments is helped by
bear) by connective tissue septa, and can therefore be easily
intraoperative imaging techniques in real operative situations.
­recognized on the two‐dimensional histological sections com­
monly used in structural studies. The idealized lobule has a
polygonal (usually hexagonal) shape. The terminal branch of
the hepatic vein (central vein) is in the center of the lobule while
Microanatomy
the corners are occupied by the “portal triads.” The components Normal liver function requires the unique arrangement of basic
of the triad are the interlobular bile ducts and the terminal components of hepatic tissue: portal vein, hepatic artery, bile
branches of the portal vein and hepatic artery. The blood carried duct, hepatic vein, and hepatocytes. These form in two‐­
by these afferent vessels is distributed by the inlet venules and dimensional sheets the above‐mentioned hepatic (classical of
arteries along the virtual “vascular septa.” This vascular frame is Kiernan’s) lobules. Profiles of portal tracts and hepatic veins of
filled up columns (or sheets in three‐dimensional space) of the various sizes are a prominent feature of liver histology [4–6].
hepatocytes constructed as “plates” arranged in a radial fashion. Smaller branches of the afferent and efferent vessels (together
The hepatic plates are separated by the similarly distributed with their stromal components) predominate in tissue sections
sinusoids. The blood runs in a centripetal direction from the taken from peripheral, subcapsular locations, whereas tissue
­vascular septa to the central vein. The vascular septa secure sections taken from more proximal areas nearer to the hilum
the mixing of the portal venous and arterial blood and the more‐ contain larger vascular structures  [6]. These vascular/stromal
or‐less equal supply to the sinuses. The bile produced by the elements are contained in tunnels (lacunae) that penetrate the
 1:  Organizational Principles of the Liver 5
Figure 1.1  Schematic drawing of the organization of blood vessels
(arteries, red; portal veins, purple; central veins, blue; bile ducts, green;
lumen of the biliary channels, including bile canaliculi, yellow) in two
adjacent lobules of human liver. One sixth of a lobule is visible on the
right and one third of a lobule is visible on the left. A terminal portal
venule, arteriole, and bile ductule (canal of Hering) are present in the
vascular septum between the lobules. The arteriole is connected directly
to the sinusoids or enters the inlet venule. Bile is drained over the whole
surface of the lobule. Arterioles and bile ductules are not present in the
vascular septum in rodents. The bile is drained through canals of Hering
connected to hepatocytes of the limiting plate. The arterioles anasto­
mose with the portal system at higher level as well. Courtesy of Sandor
Paku, Semmelweis University, Budapest, Hungary.
parenchymal mass [4]. The hepatocytes arranged in plates fill in
the space between the portal tracts and hepatic veins (Figure 1.1).
The hepatic plates form brick‐like walls (muralia) of hepato­
cytes one cell (one brick) thick. The first hepatocytes of the
hepatic plates form a virtual barrier between the periportal con­
nective tissue and the liver parenchyma called a limiting plate.
The blood vessels and their investments of connective tissue
provide the soft, spongy liver with its major structural support,
or “skeleton.” Larger afferent vessels, portal veins, and hepatic
arteries are contained together with bile ducts in connective
­tissue – the portal tracts – which are continuous with the mesen­
chymal components of the liver’s mesothelium‐covered surface
capsule (Glisson’s capsule). Portal tracts also contain lymphatic
vessels, nerves, and varying populations of other types of cells,
such as macrophages, immunocytes, myofibroblasts, and pos­
sibly hematopoietic stem cells (see [7] and references therein).
The collagenous investment of the efferent vessels is less robust
and lacks large numbers of adventitious cells.
The hepatic artery is distributed to the tissues of portal tracts,
the liver capsule, and the walls of large vessels [4–6]. In portal
tracts arterial branches form a capillary network (the peribiliary
plexus) arborized around bile ducts [8, 9]. Efferent twigs from
the peribiliary plexus empty into adjacent portal veins in rat and
mouse but not in human and hamster [10]. The portal vein sup­
plies blood to the parenchymal mass through the so‐called inlet
venules [9, 11].
In histological sections of mammalian liver, afferent and
efferent vessels interdigitate regularly in an approximate ratio of
5–6 portal tracts for each profile of a hepatic vein, to form a pat­
tern of cross‐sections of portal tracts and hepatic veins separated
by parenchyma [5, 6]. Most of the cross‐sectioned portal tracts
contain preterminal hepatic venules. These vessels represent the
seventh‐ to tenth‐order branches from the hilar portal vein in
large mammals, such as humans. These small portal tracts and
hepatic (central) veins penetrate the parenchyma in nearly paral­
lel orientations about 0.5–1.0 mm apart. The portal inlet venules
are very short vessels with no smooth muscle in their walls. They
branch from preterminal and terminal venules at points on the
circumference of the lobules at about 120 radial degrees (­triradial
branching) and penetrate the parenchyma together with terminal
arteriolar branches approximately perpendicular to and midway
between two adjacent terminal hepatic venules [5, 6]. During
their course through the parenchyma portal inlet venules break
up completely into sinusoids, which are oriented more or less
perpendicularly to the veins. Because they are hardly larger than
sinusoids, the inlet venules are not conspicuous in humans and
other mammals that lack a definite connective sheath around
them. However, in adult swine their course through the paren­
chyma is clearly marked by connective tissue.
Capillary‐size sinusoids occupy the smallest and most
­numerous tunnels (lacunae) in the parenchymal mass [4]. Unlike
capillaries elsewhere, liver sinusoids are composed of endothe­
lial cells that are penetrated by holes (fenestrae) and lack a
structured basal membrane [12], features that allow free egress
of the fluid components and solutes of the perfusing blood.
For example, tagged albumin has access to a space in the liver
that is about 48% larger than the sinusoidal volume, in contrast
to other tissues in which capillary space and albumin space are
nearly the same [13]. In favorably oriented histological sections,
more or less parallel, longitudinal profiles of sinusoids alternate
with hepatic plates  [14]. A narrow cleft, called the space of
Disse, separates sinusoids from hepatocytes located in adjacent
hepatic plates [12, 15]. At their proximal (portal venous) ends,
sinusoids are narrow and somewhat tortuous, whereas their mid­
dle and distal (hepatic venous) portions are larger and straighter
[9, 16, 17]. Sinusoids and hepatic plates are disposed radially
around the draining hepatic veins and extend directly to the
­supplying inlet venules [17].
Three‐dimensional reconstruction of the interlobular zone
revealed the existence of a small vessel in this plane, the ­vascular
septum, that serves as a starting pool for intralobular sinusoids.
This is a hemodynamic barrier, a “watershed” between the
two  neighboring lobules. This “interlobular” septum contains
­connective tissue matrix in pig, camel, bear, etc. and outlines the
lobules nicely; it also exists in a rudimentary form in human
liver [18].
The bile  –  the excretion product of the hepatocytes  –  is
­collected and transported in bile canaliculi, which are formed by
the apical sides of two adjacent hepatocytes in the hepatic plate.
The network of canaliculi is drained into the interlobular bile
ducts through interface structures called canals of Hering. These
are intermediary structures constructed partly by hepatocytes or
cholangiocytes (Figure  1.2). Since these structures are the
­primary candidates to harbor the hepatic stem cell compartment,
they are the subject of intensive investigations [19]. The distribu­
tion of canals of Hering shows variation among different spe­
cies. They are characterized by a distinct (EMA−/CD56+/CD133+)
immunophenotype in humans, leave the periportal space and
spread into the parenchyma along the rudimentary interlobular
septa, and thus do not enter into the hepatic lobule  [18].
6 THE LIVER:  FUNCTIONAL ANATOMY OF THE LIVER
Figure 1.2  Normal human liver stained for CD10. The hepatocytes
form trabeculae, 1–2 cells thick, radiating from a central vein. CD10 is
expressed on the canalicular domain, indicating the polarization of the
cells. Courtesy of Sandor Paku, Semmelweis University, Budapest,
Hungary.
Figure 1.3  Normal rat liver stained for pankeratin (green) and laminin
(red), the nuclei are labeled by TOTO (blue). There is a cross‐section of
a canal of Hering beside a small portal vein. The cytoplasm of the chol­
angiocytes is strongly positive for keratin and these cells are outlined by
laminin (basement membrane) but basement membrane is absent at the
pole where the ductile is connected to an adjacent hepatocyte. Courtesy
of Sandor Paku, Semmelweis University, Budapest, Hungary.
The interlobular bile ducts are lined by a single layer of cuboidal
cholangiocytes (Figure  1.3). They anastomose and unite larger
septal and hilar branches. The connective tissue around the larg­
est biliary branches contain peribiliary glands which also secrete
into the biliary tract (Figure 1.4).
Teutsch and coworkers [20, 21] analyzed serial sections of rat
and human livers to reconstruct the three‐dimensional structure
of hepatic tissue. Although there were differences between the
two species, the basic arrangement was similar. The reconstruc­
tion revealed primary “modules,” which constructed a more
complex “secondary” module. The integration was based on a
common drainage by branches of the hepatic veins and supply­
ing portal veins, and the modules were covered by continuous
vascular septa. The primary modules correspond to the two‐
dimensional hepatic lobules. Quite a substantial variation in the
shape and size of the modules was found, which provides
­morphogenetic plasticity to construct the whole organ. This
Figure 1.4  Normal human liver stained for pankeratin (green) and kera­
tin 7 (red). The dark lane in the center represents an interlobular vascular
septum. The double positive (yellow) biliary ductules have several con­
nections with the limiting hepatocytes but do not enter into the lobules.
Courtesy of Sandor Paku, Semmelweis University, Budapest, Hungary.
modular arrangement can improve the interpretation of lesions,
especially in pathologically altered livers, but it is certainly not
easy to transform the two‐dimensional observations into three‐
dimensional space.
Functional unit of the liver
The concept of the primary functional unit of the liver has been
the subject of debate for more than 350 years since its descrip­
tion by Wepfler in 1664 [22]. The first and most widely accepted
traditional unit of the liver is “Kiernan’s lobule” [23], as
described earlier. This is the efferent microvascular segment,
being the smallest unit of parenchyma that is drained of blood by
a single efferent (terminal hepatic or central) vein. It is quite
easy to identify it, especially in species where they are outlined
by connective tissue. The major criticism of the concept is that
the terminal afferent vessels through the vascular septa contrib­
ute to the blood supply of adjacent lobules, and therefore the
lobule cannot be a “basic functional unit.” Rappaport defined the
basic unit as the compartment of the hepatic parenchyma sup­
plied with blood by a single terminal portal vein and called this
unit the “liver acinus” [24]. Now we know that this unit is also
supplied by a single terminal branch of the hepatic artery. The
simple acinus is a parenchymal mass around a portal tract and it
is drained by more hepatic venules. The acinus is subdivided
into three zones, based on the distance from the portal vein.
The distribution of these areas fits to the functional zonality of
the hepatic parenchyma. Pathological lesions (e.g. steatosis or
necrosis) also often follow this zonal pattern, which made this
unit attractive. However, this zonality did not correspond per­
fectly to the distribution of enzyme activities and the hepatic
modules described by Teutsch [20, 21] were also not compatible
with the concept of the acinus.
Matsumoto and his colleagues [6] investigated the angioarchi­
tecture of the human liver on thousands of serial sections, dis­
tinguishing conducting and parenchymal portions of the portal
venous tree. A cone‐shaped parenchymal portion (primary lob­
ule) was defined which was supplied by a terminal portal venule.
 1:  Organizational Principles of the Liver 7
The circulatory network of this unit was named the hepatic
microcirculatory subunit (HMS). Ekataksin and Wake [25]
showed that the afferent microvascular segment is also the small­
est unit of parenchyma drained of bile by terminal bile ducts
[25], demonstrating that this hemodynamic segment is also the
smallest excretory unit of parenchyma. The compartment of the
hepatic parenchyma associated with an HMS contains all of
the elementary structures of the hepatic tissue and may represent
the elementary functional and morphological unit of the liver. It
was named cholehepaton. The cholehepaton is also the smallest
bile–blood unit and conforms with the principle of countercur­
rent flow. The cholehepaton has no anatomical or structural
­borders, and cannot be recognized on histological sections. It is
defined by its function and mostly corresponds to Matsumoto’s
primary lobules. Six of these Matsumoto’s primary lobules con­
stitute the secondary lobule, which is almost identical with the
classical Kiernan’s lobule. This long and complicated detour
returned us to the classical lobule, which is widely used today to
analyze reactions of the hepatic tissue.
However, it is worth remembering that there are other types
of hepatic functional units and some workers contend that the
hepatic tissue is an indivisible continuum that has no definable
units [26].
Liver hemodynamics
The hepatic vasculature is characterized by high capacity, high
compliance, and low resistance [27]. Blood vessels comprise
about 22% of the liver’s mass/volume [13] and the liver contains
about 12% of the body’s total blood volume under physiological
conditions [27], a sizeable fraction of which can be expelled by
contraction of larger vessels by sympathetic nerve stimulation. In
other words, the liver is a blood reservoir. The pressure of portal
venous blood is reduced as the major afferent vessels dichotomize
through the parenchyma, from about 130 mmH2O in the extrahe­
patic portal vein to about 60 mmH2O in the preterminal portal
veins of the exteriorized liver of the anesthetized rat, amounting to
about 60% of the total transhepatic pressure gradient [13]. A
­similar portal pressure gradient has been found in humans [27].
Blood flow through the liver amounts to about 1500–2000 mL
min−1 in adult humans, about 25% of the resting cardiac output
[27]. About 25% of the total liver blood flow is derived from the
hepatic artery at prevailing arterial pressure and oxygenation. The
portal venous blood (about 75% of total liver blood flow) arrives at
the liver partially depleted of oxygen and at a reduced pressure as
a consequence of having already perfused the splanchnic viscera.
In aggregate, sinusoids comprise about 60% of the liver’s
vascular volume, or about 13% of the total liver mass/volume
[13]. A significant decrease in blood pressure occurs in sinu­
soids (about 40% of the transhepatic pressure gradient), the
pressure declining to about 25 mmH2O in terminal hepatic veins
of exteriorized liver of anesthetized rats [28]; the pressure gradi­
ent in the short sinusoids is especially steep. Blood pressure in
the inferior vena cava approximates that in the terminal hepatic
vein [27]. Consequently, although flow of blood through sinu­
soids faces little resistance, it is slow and somewhat intermittent
and is assisted by negative pressure produced by respiratory
expiration [27]. Possible mechanisms of regulating blood flow
within sinusoids are controversial. The sinusoidal circulation is
regulated by a sphincter in the terminal arterioles [29]. In addi­
tion, hepatic stellate cells are the pericytes of the sinusoids and
are the principal regulators of sinusoidal blood flow. They have
receptors for molecular mediators regulating contraction and
are closely associated with nerve endings, indicating neurogenic
influence [30]. Sinusoids appear to have limited contractile abil­
ity, possibly produced by contraction of encircling stellate cells
(pericytes) [29, 31]. Studies of the exteriorized liver of rodents
suggest that sinusoidal flow may be regulated at both inlet and
outlet levels [32], although other similar studies have not
detected sphincters at either point [28]. However, sinusoidal
flow is strongly affected by post‐sinusoidal resistance [27].
THE CELLS OF THE LIVER
Hepatocytes
Hepatocytes are commonly denoted as “parenchymal cells” and
the other cells of the liver tissue matrix as “non‐parenchymal
cells.” This convention is somewhat artificial since hepatocytes
alone are not competent to perform all ­ essential hepatic
­functions, and several types of cells in the liver tissue matrix
function as an integrated community to carry out conjointly the
multiplicity of hepatic functions. Functional integration of this
cellular community is accomplished by several communication
mechanisms, including signaling networks involving numerous
cytokines and chemokines, and by direct transfer of small mol­
ecules through gap junctions [33]. Hepatocytes, responsible for
most of the synthetic and many of the metabolic functions of the
liver (see Chapters 23 and 24), are large polygonal cells (averag­
ing about 25–30 μm in cross‐section and 5000–6000 μm3 in
­volume [34]). They are the most numerous cells in the liver
parenchyma; the adult human liver probably contains about 1011
hepatocytes, representing about 60% of all cells in the paren­
chymal matrix and comprising about 80% of its mass/volume
[15]. Hepatocytes are shaped as complex rhomboids with
­several distinct surfaces [34]. They are polarized by molecular
specializations of their various surface membranes in the forms
of receptors, pumps, transport c­hannels, and carrier proteins
(see Chapters 5 and 6) that comprise three functionally distinct
domains (see Chapter  6): the basolateral domain, the lateral
domain, and the canalicular domain.
• Basolateral domain: The basolateral (or sinusoidal) domain
constitutes about 35% of the total hepatocyte surface facing the
sinusoids. The area of this surface is greatly amplified by the
folding of the plasma membrane to form innumerable micro­
villi that extend into the space of Disse [34]. This membrane is
equipped for extraction of a great variety of molecules from the
blood and for the simultaneous secretion into the blood of other
molecules that have been modified or synthesized by hepato­
cytes. The cell matrix adhesion molecules are also present on
this side of the hepatocyte. Although there is no structured
basement membrane between the hepatocytes and sinusoidal
endothelial cells, the lack of integrin‐linked kinase 1 (Ilk1)
results in disruption of the normal structure  [35], supporting
the necessity of hepatocyte–matrix interaction. About 50% of
the total hepatocyte surface faces adjacent hepatocytes [34].
8 THE LIVER:  THE CELLS OF THE LIVER
• Lateral domain: The plasma membranes of these intercellular
surfaces are mostly flat lateral domain, which is the least com­
plex. This surface contains intercellular adhesion complexes
(tight junctions, intermediate junctions, and desmosomes)
that pin together the adjacent hepatocytes, form a permeabil­
ity barrier between the peri‐sinusoidal space of Disse and
bile canaliculi, and include gap junctions that allow commu­
nication between adjacent hepatocytes by transfer of small
molecules.
• Canalicular domain: Infoldings of the lateral surfaces create
bile canaliculi, which comprise about 13% of the total hepato­
cyte surface. This is termed the canalicular domain [34]. It is
also amplified by microvilli and modified for bile excretion.
The canalicular surface is confined by strong junctional com­
plexes. Bile canaliculi form a belt‐like extracellular space
(about 1 μm in diameter) that is continuous along the lengths
of the hepatic plates, connecting at the portal ends with bile
ducts. The molecular mechanisms responsible for the polarity
of the canalicular domain are well characterized. Hepatocyte
nuclear factor 4 (Hnf4α) is the master regulator of morpho­
logical and functional hepatocyte differentiation [36], but sev­
eral other factors, such as liver kinase B1 (Lkb1) [37],
vacuolar sorting protein 33b (Vps33b) [38], and claudin‐15
[39], are also required for the polarization.
As befits their numerous metabolic functions, hepatocytes
contain a complex array of mitochondria (~1700 per cell on
average), peroxisomes (~370 per cell), lysosomes (~250 per
cell), Golgi complexes (~50 per cell), aggregates of rough and
smooth endoplasmic reticulum (~15% of cell volume), and
numerous microtubules/microfilaments [34].
Polyploidization is another unique feature of the hepatocytes.
This is the result of defective cytokinesis, which seems to be
regulated by the insulin–Akt signaling pathway [40]. The extent
of ploidy increases with age. The presence of aneuploid hepato­
cytes in the normal liver is also well established [41]. The exact
function of this process is still unknown but it is thought to help
in adaptation to chronic injury.
Hepatic plates and adjacent sinusoids form associations that
are structurally similar in all parts of the liver. Various liver cells
show numerical, structural, and functional heterogeneities
related to their location along the afferent–efferent axis of
hepatic plates and sinusoids. Among the structural differences
are ploidy variations in hepatocytes; in adult mammals hepato­
cytes located at the portal ends of hepatic plates are diploid,
whereas cells of higher ploidy are located further downstream
[42]. Gap junctions containing connexin 26 are more numerous
on portal hepatocytes, whereas junctions containing connexin
32 are distributed on hepatocytes in all parts of hepatic
plates [43]. These variations in structure and cellular composi­
tion are associated with functional differences among hepato­
cytes located at different points along the afferent–efferent axis
of plates and sinusoids. Rappaport divided the portal‐hepatic
(afferent–efferent) lengths of hepatic plates into three arbitrary
zones (termed I, II, and III). Hepatocytes located in these zones
differ in their functional capabilities and susceptibilities to path­
ological damage [24], and pathways performing opposing
­functions follow an inverse distribution along the portocentral
axis. This is strikingly exemplified by regional differences in
carbohydrate metabolism (gluconeogenesis and glycogen stor­
age by periportal hepatocytes; glycolysis by perihepatic vein
­hepatocytes). The ammonia and fatty acid metabolisms are also
zonally distributed. Zone 1 hepatocytes are engaged in urea pro­
duction and β‐oxidation of fatty acids, while zone 3 hepatocytes
remove nitrogen by glutamine synthetase and perform lipogen­
esis. The metabolism of xenobiotics is more prominent in the
pericentral hepatocytes. Recent research has shown that many
liver functions are dispersed heterogeneously, with dispersed
functions often acting as integrated parts of coordinate meta­
bolic systems [44].
Zonation of liver functions was thought to be related to
­sinusoidal hemodynamics, which produced gradients in blood‐
borne substances available to hepatocytes and other cells of the
parenchymal matrix [44]. Recent evidence gained mostly from
experiments with genetically engineered mice proved that the
Wnt/β‐catenin pathway is the master regulator of hepatocytic
zonation, but the interaction with lymphoid enhancer factor 4
(Lef4) and HNF4α is also critical [45]. Hepatocytes and other
cells located at afferent and efferent ends of hepatic plates are
subjected to different microenvironmental conditions. Certain
molecules are largely extracted by the first hepatocytes that
encounter the perfusing blood, lowering their concentration
downstream. For example, oxygen levels in the blood at affer­
ent and efferent ends of sinusoids differ greatly because oxygen
is efficiently extracted by hepatocytes located at the afferent
ends of hepatic plates, exposing downstream hepatocytes to
relatively hypoxic conditions; the oxygen gradient alone can
explain much of the heterogeneity of hepatocyte function
related to position in plates [46]. Other molecules modified or
produced by upstream hepatocytes are excreted into the sinu­
soidal blood and may be removed by hepatocytes located fur­
ther downstream. The complex interplay of metabolite
concentration in the perfusing blood, coupled with extraction,
modification, secretion, re‐extraction, and further modification,
influence the metabolic events that occur in individual cells and
define ­unequal parenchymal territories that produce zonal vari­
ations in different physiological capabilities and pathological
susceptibilities [44]. Disruption of metabolic zonation has
­
severe consequences.
Physiological turnover of hepatocytes occurs slowly. They
have a lifespan of about 400 days in an adult steady‐state hepat­
ocyte population, about 0.025% of which typically are cycling
[32] and the remainder rest in G0 phase. Although the hepato­
cytes are ready to re‐enter the cell cycle upon injury, this capac­
ity decreases with age. There are, however nonresolved opposing
opinions about the replacement of lost hepatocytes during
homeostatic conditions. The contribution of hematopoietic cells
to liver maintenance was a very popular idea 20 years ago.
Repopulation of bone marrow‐derived cells (macrophages,
myofibroblasts) in the liver is evident in recipients of liver trans­
plants, in which these types of liver cells are replaced with cells
of the host genotype [47]. In contrast, hepatocytes are not gener­
ated from bone marrow cells in significant numbers under either
circumstance [47].
However, the ancient and very attractive theory of “streaming
liver” keeps returning. It was originally proposed by Zajicek and
colleagues [48] that periportal hepatocytes have enhanced repli­
cative potential and the progeny of these cells spread under
 1:  Organizational Principles of the Liver 9
homeostatic condition in a portal–central direction. Although few
studies applying lineage‐tracing methodology have supported
this notion, most of the studies have ruled out this option. Most
recently, Axin2+ hepatocytes [49] abutting hepatic veins as well
as hybrid periportal hepatocytes [50] have been reported to have
selective growth advantage. Based on these results a bidirectional
streaming hypothesis, “somehow like ocean water entering into
the delta of the Amazon River at high tide” has been proposed
[51]. This issue will no doubt be resolved in the near future.
Cholangiocytes
Cholangiocytes, or biliary epithelial cells, comprise much less
than 1% of the total number of cells in the liver parenchyma, since
most are located in bile ducts in portal tracts [52]. Only the small­
est bile ducts penetrate the parenchymal mass in the company of
terminal portal veins, where they connect with bile canaliculi in
hepatic plates. The points of connection of ducts with hepatic
plates are defined by tubular structures called the canals of
Hering, which are composed of both cholangiocytes and hepato­
cytes [53]. They are also the location of liver epithelial stem cells
that can differentiate into both hepatocytes and cholangiocytes
[54, 55]. Larger bile ducts contain cholangiocytes that rest on a
basal membrane and vary in number and size in proportion to
duct size [56]. The cholangiocytes are also polarized cells, with
an apical (luminal) and a basolateral domain. Their luminal sur­
face membranes are expanded by microvilli and a single primary
cilium, which is a sensor for mechanical, osmolar, and chemical
stress [57]. Although they contain fewer mitochondria and sparser
endoplasmic reticulum than do hepatocytes, cholangiocytes in
intrahepatic bile ducts, together with the network of capillaries
that surrounds them (the peribiliary plexus), form a metabolic
unit that modifies the composition of canalicular bile [58].
The biliary tree is not just a draining pipe and 70–90% of the
bile volume is produced by the cholangiocytes. They change the
composition of the bile by secretion and absorption. The main
secretory product is bicarbonate, while they absorb bile acids,
glucose, and glutamate. The cholangiocytes also play an impor­
tant immunoregulatory role. They are in the first line of defense
against microbial components of the biliary tract, xenobiotics,
and foreign antigens. The cholangiocytes tackle these chal­
lenges by maintaining immunotolerance. They are equipped
with pathogen recognition receptors (PRR), all members of the
toll‐like receptors (TLR1–10), as well as related signaling mol­
ecules. Cholangiocytes produce antibacterial products (e.g.
defensins, lactoferrin, lysozyme, and transport IgA) into the
lumen and express MHC class I and II molecules and antigen
presenting cells [59]. It is not surprising, therefore, that the bil­
iary tree is often affected by immunological disorders, such as
primary sclerosing cholangitis, primary biliary cholangitis, and
graft‐versus‐host disease. The cholangiocytes are slow turnover
cells, but under special condition they can also participate in the
regeneration of liver parenchyma (see later).
Endothelial cells
Endothelial cells of sinusoids comprise about 3% of the paren­
chymal mass/volume [15] and probably number about 3 × 1010
in an adult human liver. Liver sinusoidal endothelial cells are
highly specialized endothelial cells with peculiar morphology
and function. They are extremely thin in normal liver, 150–170
nm across, and this attenuated cytoplasm is fenestrated. The
diameter of these “perforations” ranges from 50 to 200 nm, and
they are clustered together in groups to form “sieve plates” [60].
The fenestrae are dynamic structures, and their actual diameter
is influenced by blood pressure, composition of extracellular
matrix, hormones, etc. The porosity of the sinusoidal endothe­
lial cells is polarized, and the fenestrae are more numerous in
the centrilobular zone. The maintenance of the fenestration
requires paracrine and autorcrine signals, which are mostly pro­
vided by hepatocytes and stellate cells [61]. Surprisingly, these
endothelial cells have mostly anaerobic metabolism, providing
lactate to the adjacent hepatocytes.
High‐resolution in vivo microscopy has revealed swelling
and contracting of these cells as a response to vasoactive sub­
stances, suggesting that they participate in the regulation of
blood flow [62]. A unique functional feature of the sinusoidal
endothelial cells is their high endocytic capacity. This process is
mediated by all sorts of scavenger receptors, providing the main
pathway for clearance of weakened molecules from circulation
[63, 64] (see Chapters 26–28). Endothelial sinusoidal cells also
express several types of PRR (e.g. mannose receptor several
TLRs) and are able to produce inflammatory cytokines (e.g.
TNF and IL‐6) as well as playing a significant role in the innate
immunity. In spite of intensive studies, their role in adaptive
immune reactions is still controversial [52].
The lifespan of sinusoidal endothelial cells is not known;
they divide rarely and progenitor cells seem to be important in
their maintenance. Two populations of progenitor cells can be
distinguished [65]: the liver‐resident population is thought to be
responsible for normal cell turnover, while bone marrow‐derived
cells help to replenish the sinusoidal endothelial cells when
necessary.
Hepatic immune cells
The human liver is exposed to 1.5 L of blood every minute and
a massive load of harmless dietary and commensal antigens, to
which it must remain tolerant. The predominantly tolerogenic
role of the hepatic immune system is well known [52], but the
liver is also exposed to a variety of viruses, bacteria, parasites,
and metastatic tumor cells, so needs mechanisms to override
immune tolerance. In addition, the liver’s native immune sys­
tem has a major regulatory role in the repair of the liver after
cell injury and loss. The liver‐centered immune system is
largely segregated from the rest of the body’s immune system
[66, 67]. The human liver is estimated to contain about 1010
lymphocytes of different phenotypes, located along sinusoids
and in portal tracts [66]. It includes a major fraction of the
body’s innate (native) immune capacity, as well as a small com­
ponent of its acquired (adaptive) immune capacity [66, 67]. The
major components of the hepatic immune system are innate
lymphocytes, which include a variety of T cells and non‐T cells
that are able to respond rapidly to conserved ligands. These
cells do not express T cell receptor (TCR) antigens. This group
of immune cells includes NK cells, CD56+ T cells, natural killer
T (NKT) cells, γ/δ T cells, and mucosal associated invariant
T  (MAIT) cells. The liver contains multiple types of antigen
10 THE LIVER:  ONTOGENESIS OF THE LIVER
presenting cells such as hepatic myeloid dendritic cells (DCs),
plasmacytoid DCs, CD11+ DCs, and NK1.1+ cytotoxic DCs
[68]. In addition to the professional lymphocytes and DCs, sev­
eral populations of hepatic resident cells (e.g. Kupffer cells,
sinusoidal endothelial cells, stellate cells, and cholangiocytes)
are also important and active players of the liver‐centered
immune system.
Macrophages
Macrophages are myeloid cells that are widely distributed
throughout the tissues of mammalian organisms. The liver har­
bors 80% of all macrophages of the body. In addition, it is also
patrolled by blood monocytes [69]. Hepatic macrophages can
be divided into two classes based on their origin [70]: resident
macrophages and bone marrow‐derived macrophages. Resident
macrophages of the liver, traditionally termed Kupffer cells, are
established during embryonic development from the yolk sac
and persist independent of blood monocytes. These cells self‐
renew during homeostatic conditions. Bone marrow‐derived
blood‐borne monocytes give rise to monocyte‐derived hepatic
macrophages that are more characteristic of liver injury. Kupffer
cells are not optimally suited to migration, so hepatic injuries
massively recruit blood monocytes to the liver. These resemble
Kupffer cells phenotypically, but they remain functionally
­different. Macrophages are highly versatile cells that play a sub­
stantial role in liver homeostasis, promoting and resolving
inflammatory processes and fibrosis [69].
Liver macrophages are avidly phagocytic through C3 and
Fc  receptors, clearing the sinusoidal blood of relatively large
particulate materials, including bacteria and weakened cells
(worn‐out erythrocytes, dead or damaged hepatocytes, etc.) [63,
64]. Together with sinusoidal endothelial cells they form the
organism’s major system for removing worn‐out cells and pro­
teins from perfusing blood. Activated macrophages produce
many chemokines and cytokines that have a fundamental role in
the implementation of the liver’s acute‐phase reaction, coordi­
nating the responses of all the parenchymal cells to injury [67].
Myofibroblasts
Myofibroblasts are not present in the normal liver, but several
cell types which are present physiologically can transform or be
activated or transdifferentiated into the phenotype, which is the
major source of extracellular matrix components and plays a
basic role in pathological processes of the liver [71].
Hepatic stellate cells (HSCs) [72] are liver‐resident mesen­
chymal cells which play an important role in liver physiology
and pathology. They are found in the space of Disse, between
the sinusoidal endothelial cells and hepatocytes. This special
location makes them able to respond to numerous kinds of
injury. In addition, by encircling the sinusoids they can func­
tion as pericytes and are thought to be the most important
regulator of sinusoidal diameter and blood flow. They com­
prise about 1.5% of the parenchymal volume/mass [15] and
are multifunctional (see Chapters 28 and 29). The embryonic
origin of stellate cells is still uncertain. Most likely they origi­
nate from the septum transversum‐derived mesothelial cells,
but other options are still open. In the healthy liver they are the
largest reservoir of vitamin A, hence their former name “­fat‐
storing cells.” When the liver is injured the stellate cells trans­
differentiate into myofibroblasts and are the major producer of
extracellular matrix (ECM). HSCs are important sources of
cytokines and growth factors. This way they have impact on
the proliferation, ­differentiation, and morphogenesis of the
other hepatic cell types during liver development and regen­
eration [72].
Portal fibroblasts are spindle‐shaped mesenchymal cells in
the periportal connective tissue [73]. They are distinct from
HSCs in both distribution and phenotype. They do not store
vitamin A but express elastin and Thy‐1. Portal fibroblasts take
part in physiological ECM turnover and can contribute to the
myofibroblast population in cholestatic liver injury.
Bone marrow‐derived mesenchymal stem cells can also dif­
ferentiate into myofibroblasts. The contribution of bone marrow
cells is well established in the fibrotic processes of kidney and
lung but it seems to be quite limited in the liver. Epithelial–­
mesenchymal transition (EMT) is another recently described
mechanism. Both hepatocytes and cholangiocytes can undergo
such phenotypic change in tissue culture, but lineage‐tracing
experiments in mice provide evidence against the role of EMT
in myofibroblast ­generation in hepatic tissue in vivo [74, 75].
ONTOGENESIS OF THE LIVER
The evolutionary steps that resulted in the emergence of the
mammalian liver with its multiple types of functioning cells are
obscure. To the extent that the ontogenesis of the liver mirrors
its phylogenesis, the embryonic development of the mammalian
liver suggests the way in which the aggregation of multiple
types of cell into the hepatic parenchyma may have evolved
(for details see Chapter 2).
The development sequences of the liver in fish, birds, and
mammals are similar [74–76], but endothelial cells do not
appear to direct the emergence of endodermal cells from the gut
during development of the liver in zebrafish [76]. Furthermore,
endothelial cells with scavenger activity are in the gills and kid­
neys of cartilaginous and bony fish, and not in the liver as in all
terrestrial animals [77]. The location of scavenger endothelial
cells in the liver reflects a late step in the evolution of the mam­
malian liver. Nevertheless, the general pattern of expression of
transcription factors and genes involved in liver development is
conserved in all these species [78], suggesting a common tran­
scriptional strategy for assembling the liver. Information on
when this strategy first emerged awaits further genetic analysis
of gut appendages in chordate ancestors of vertebrates.
Liver parenchymal repair
Three distinct processes have evolved to generate new hepato­
cytes needed to meet both increased physiological functional
demands and to replace hepatocytes lost to trauma and/or toxicity.
These processes comprise either a temporary reactivation of cell
cycle transit in fully differentiated, mitotically quiescent hepato­
cytes, or the generation of entirely new hepatocyte lineages from
adult liver stem cells (see Chapters 36 and 38).
 1:  Organizational Principles of the Liver 11
The most direct and rapid parenchymal augmentation/
replacement processes involve the induction of hepatocyte rep­
lication in the absence of a preceding increase in hepatocyte
death, often associated with increased hepatic functional
demand due to physiological need [78, 79]. Hyperplasia of
hepatocytes by this mechanism enlarges the parenchymal mass
and increases hepatocyte functional capacity. This process is
regulated by the binding of ligands to hepatocyte nuclear recep­
tors, nearly 50 of which have been identified [79, 80]. Nuclear
receptors are transcription factors that, when bound to ligands,
directly upregulate the combination of genes required to drive
hepatocytes through the cell cycle [79, 80]. Several ligands for
nuclear receptors (termed “primary hepatocyte mitogens”),
including adrenal corticoids, bile acids, sex steroids, thyroid
hormone, peroxisome proliferators, and 9‐cis,cis‐retinoic acid,
directly stimulate the proliferation of hepatocytes and increase
liver mass after binding to nuclear receptors [80]. Although it
would appear that endothelial cells would be needed to support
the additional hepatocytes, no documentation of coordinate
endothelial cell proliferation has been presented; it is, however,
possible that new endothelial cells could be derived from the
bone marrow.
Next in process complexity and speed of response is the
replacement of the diverse liver parenchyma by the sequential
proliferation of all of the component cells (hepatocytes, chol­
angiocytes, endothelial cells, macrophages, stellate cells, and
immunocytes), and the merging of the new cells into a tissue
that closely resembles the functional units of the undamaged
liver [78]. This process, which can replace up to 70% of the
parenchymal mass in mammals, is often called “liver regenera­
tion,” although this is a misnomer since in mammals the part of
the liver removed surgically does not “regenerate” in the way
that body parts in certain lower animal species do. Instead, the
liver, after resection, is enlarged by the expansion of remaining
units (lobes), in a biological process defined as “compensatory
hyperplasia.” In contrast to liver repair in mammals, liver repair
after partial hepatectomy in fish most intensively involves cells
at the resection margin [81, 82] and may culminate in the
regrowth (regeneration) of the resected tissue [81].
The cell proliferation phase of this reparative process in
mammals has been subjected to intensive kinetic and regulatory
analyses (see Chapter  45 and references therein). After tissue
loss, residual hepatocytes are activated to proliferate within few
hours. Hepatocyte proliferation begins at the portal ends of
plates  [83], and successive waves of hepatocyte proliferation
ultimately involve virtually all residual hepatocytes [83].
Hepatocyte replacement is followed sequentially by prolifera­
tion of sinusoidal endothelial cells and macrophages [83, 84],
and the other cells of the parenchymal matrix. To the extent that
it has been elucidated (see Chapter 45), regulation of hepatocyte
proliferation is regulated by a complex mixture of cytokines and
growth factors [85]. Most of the regulatory molecules are pro­
duced by various liver cells or are released from storage sites
within the liver [85], and many are components of the acute‐
phase reaction [86] and other elements of the liver’s native
immune system [87–89]. The less completely analyzed remod­
eling phase primarily involves endothelial cells and likely the
other cells of the liver parenchyma. For example, proliferating
hepatocytes initially form focal multicellular clumps [90, 91],
which are cleaved into one‐hepatocyte‐wide plates by signaling
from and separation by endothelial and stellate cells [90, 91].
Eventually, the remaining lobes increase exclusively by the
enlargement of preexistent hepatic lobules, contrary to the
­physiological liver growth in young animals, when new lobules
are formed [92].
Although known regulatory mechanisms drive the reparative
process, the mechanism that “triggers” the onset of repair after
loss of liver tissue is still obscure. Since the liver vasculature
must accept the entire portal blood volume, it has long been
suspected that the trigger may be the massive increase in portal
blood flow per unit of residual mass that follows loss of liver
tissue [93]. Increased portal blood flow and pressure cause shear
stress in sinusoids [94], which produces a burst of nitric oxide
and prostaglandin production by sinusoidal endothelial cells,
possibly providing the molecular trigger [95, 96]. Alternatively
(or in concert), early activation of the nuclear receptor mecha­
nism of hepatocyte proliferation may function as a trigger [96],
and it is possible that multiple alterations in the physiological
status of the liver remaining after tissue loss may converge to
produce a “mass action” trigger.
If the hepatocytes are compromised, there are alternative
mechanisms of liver regeneration. Enlargement or hypertrophy
of hepatocytes can compensate for the lost parenchyma [97], but
this response usually provides just a transient solution.
There has been much debate about the participation of stem or
progenitor cells in liver regeneration. A peculiar cell population,
named after the shape of their nuclei as “oval cells,” were
observed in hepatocarcinogenesis experiments in rodents. Similar
cells have been described in several other species and their emer­
gence has been named “ductular reaction.” There is convincing
evidence that these cells can replace the lost liver parenchyma
[98] and behave as the amplification compartment of hepatic
stem cells. Several candidates for hepatic stem cells are known,
but most data indicate that the terminal segment of the biliary
system, the canals of Hering, harbor the adult hepatic stem cells.
Lineage‐tracing experiments in rats [54, 55, 57, 99, 100] and
zebrafish [101] demonstrated the regeneration of hepatocytes
from biliary stem cells. The application of the cre/lox lineage
tracing in mice did not support this prevailing model, because
no biliary cell‐derived hepatocytes were observed, although the
hepatocytic origins of biliary cells and cholangiocarcinomas
were demonstrated [102]. However, eventually hepatic progeni­
tor cells of biliary origin with liver repopulation capacity were
shown in mice following complete blockage of hepatocyte pro­
liferation [103]. At present there seem to be general agreement
[104] that both hepatocytes and cholangiocytes (or their sub­
populations) are able to behave as stem cells, and under specific
conditions are capable of regenerating both epithelial cell
­compartments of the liver. The capacity of these highly differen­
tiated cells is also referred to as “plasticity” [105] but this is
mostly a debate about terminology – how we should refer to a
peculiar biological reaction. Initial observations indicate that
these “back‐up” stem cells, which support regenerative pro­
cesses in rat and human, are organized along the branches of the
portal vein [106, 107] and are regulated by elements of hepatic
immunomodulation centered on the acute‐phase reaction [98,
108], similar to the organization of liver architecture during
embryonic development.
12 THE LIVER:  REFERENCES
Although the subject of intensive scrutiny recently, there is
no substantial evidence that hematopoietic stem cells or mesen­
chymal components of the liver are a significant source for the
generation of hepatocytes or biliary epithelial cells in either
humans or experimental animals [47]. This situation contrasts
with the replenishment from hematopoietic sources of other
cells of the liver parenchyma [47].
REFERENCES
1. Elias, H. A re‐examination of the mammalian liver, I. Parenchymal
­architecture. Am J Anat, 1949;84:311–33.
2. Couinand, C. Liver anatomy: portal (and suprahepatic) or biliary segmentation.
Dig Surg, 1999;16:459–67.
3. Majno, P., Mentha, G., Toso, C. et al. Anatomy of the liver: an outline with
three levels of complexity: a further step towards tailored territorial liver
resections. J Hepatol, 2013;60:654–62.
4. Elias, H. A re‐examination of the mammalian liver, II. The hepatic lobule
and  its relation to the vascular and biliary systems. Am J Anat, 1949;84:
379–456.
5. Rappaport, A.M., Borowy, Z.J., Lougheed, W.M. et al. Subdivision of hex­
agonal liver lobules into a structural and functional unit. Role in hepatic
physiology and pathology. Anat Rec, 1954;119:11–34.
6. Matsumoto, T., Komori, R., Magara, M. et al. A study of the normal structure
of the human liver, with special reference to its angioarchitecture. Jikeikai
Med J, 1979;26:1–40.
7. Kotton, D.W., Fabian, A.J., and Mulligan, R.C. A novel cell population in
adult liver with potent hematopoietic reconstitution activity. Blood,
2005;106:1574–80.
8. Mitra, S.K. The terminal distribution of the hepatic artery with special refer­
ence to arterio‐portal anastomosis. J Anat, 1966;100:651–63.
9. Kardon, R.H. and Kessel, R.G. Three‐dimensional organization of the
hepatic microcirculation in the rodent liver as observed in the scanning elec­
tron microscopy of corrosion casts. Gastroenterology, 1980;79:72–81.
10. Yamamoto, K., Sherman, I., Phillips M.J. et al. Three‐dimensional observations
of the hepatic arterial terminations in rat, hamster and human liver by scanning
electron microscopy of microvascular casts. Hepatology, 1985;5:452–6.
11. Watanabe, Y., Püschel, G.P., Gardemann, A. and Jungermann, K. Presinusoidal
and proximal intrasinusoidal confluence of hepatic artery and portal vein in
rat liver: functional evidence by orthograde and retrograde bivascular per­
fusion. Hepatology, 1994;19:1198–207.
12. Wisse, E., DeZanger, R.B., Charels, K. et al. The liver sieve: considerations
concerning the structure and function of endothelial fenestrae, the sinusoidal
wall and the space of Disse. Hepatology, 1985;5:683–92.
13. Goresky, C.A. A linear method for determining the liver sinusoidal and
extravascular volumes. J Physiol, 1963;204:626–40.
14. Grisham, J.W., Nopanitaya, W., Compagno, J., and Nagel, A.E.H. Scanning
electron microscopy of normal rat liver: the surface structure of its cells and
tissue components. Am J Anat, 1975;144:295–322.
15. Blouin, A., Bolender, R.P., and Weibel, E.R. Distribution of organelles and
membranes between hepatocytes and nonhepatocytes in the rat liver paren­
chyma. A stereological study. J Cell Biol, 1977;72:441–55.
16. Miller, D.L., Zanolli, C.S., and Gumucio, J.J. Quantitative morphology of
the sinusoids of the hepatic acinus. Gastroenterology, 1979;76:965–69.
17. Grisham, J.W. and Nopanitaya, W. Scanning electron microscopy of casts of
hepatic microvessels, in Hepatic Circulation in Health and Disease (ed. W.
Laut), Raven Press, New York, 1981, pp. 87–107.
18. Dezső, K., Paku, S., Papp, V. et al. Architectural and immunohistochemical
characterization of biliary ductules in normal human liver. Stem Cells Dev,
2009;18:1417–22.
19. Theise, N.D., Saxena, R., Portmann, B.C. et al. The canals of Hering and
hepatic stem cells in humans. Hepatology 1999;30:1425–33.
20. Teutsch, H.F., Schuerfeld, D., and Groezinger, E. Three‐dimensional
­reconstruction of parenchymal units in the liver of the rat. Hepatology,
1999;29:494–505.
21. Teutsch, H.F. The modular microarchitecture of human liver. Hepatology,
2005;42:317–25.
22. Bloch, E.H. The termination of hepatic arterioles and functional unit of the
liver as determined by microscopy of the living organ. Ann NY Acad Sci,
1970;170:78–87
23. Kiernan, F. The anatomy and physiology of the liver. Trans R Soc Lond,
1833;123:711–70.
24. Rappaport, A.M. The microcirculatory hepatic unit. Microvasc Res,
1973;6:212–28.
25. Ekataksin, W. and Wake, K. Liver units in three dimensions: I. Organization
of argyrophil connective tissue skeleton in porcine liver with particular refer­
ence to the “compound hepatic lobule.” Am J Anat, 1991;191(2):113–53.
26. Elias, H. and Sherrick, J.C. Morphology of the Liver. Academic Press,
New York, 1969.
27. Lautt, W.W. and Greenway, C.V. Conceptual view of the hepatic vascular
bed. Hepatology, 1987;7:952–63.
28. Nakata, K., Leong, G.F., and Brauer, R.W. Direct measurement of blood
pressure in the minute vessels of the liver. Am J Physiol, 1960;199:1181–8.
29. McKuskey, R.S. Morphological mechanisms for regulating blood flow
through hepatic sinusoids. Liver, 2000;20:3–7.
30. Hellerbrand, C. Hepatic stellate cells  –  the pericytes in the liver. Pflugers
Arch Eur J Physiol, 2013;465:775–8.
31. Reynaert, H., Thompson, M.G., Thomas, T., and Geerts, A. Hepatic stellate
cells: role in microcirculation and pathophysiology of portal hypertension.
Gut, 2002;50:571–81.
32. Grisham, J.W. A morphologic study of deoxyribonucleic acid synthesis and
cell proliferation in regenerating rat liver; autoradiography with thymidine
H3. Cancer Res, 1962;22:842–49.
33. Kmeic, Z. Cooperation of liver cells in health and disease. Adv Anat Embryol
Cell Biol, 2001;161:1–151.
34. Weibel, E.R., Stäubli, W., Gnägi, H.R., and Hess, F.A. Correlated
­morphometric and biochemical studies on the liver cell. 1. Morphometric
model, stereologic methods, and normal morphometric data for rat liver.
J Cell Biol, 1969;42:68–90.
35. Gkretsi, V., Apte, U., Mars W.M. et al. Liver specific ablation of integrin‐
linked kinase in mice results in abnormal histology, enhanced cell prolifera­
tion, and hepatomegaly. Hepatology 2008;48:1932–41.
36. Parviz, F., Matullo, C., Garrison, W.D. et  al. Hepatocyte nuclear factor
4 alpha controls the development of a hepatic epithelium and liver morpho­
genesis. Nat Genet, 2003;34:292–6.
37. Woods, A., Heslegrave, A.J., Muckett, P.J. et al. LKB1 is required for hepatic
bile acid transport and canalicular membrane integrity in mice. Biochem J,
2011;434:49–60.
38. Cullinane, A.R., Straatman‐Iwanowska, A., Zaucker, A. et  al. Mutations in
VIPAR cause an arthrogryposis, renal dysfunction and cholestasis syndrome
phenotype with defects in epithelial polarization. Nat Genet, 2010;42:303–12.
39. Cheung, I.D., Bagnat, M., Ma, T.P. et al. Regulation of intrahepatic bile duct
morphogenesis by Claudin 15‐like b. Dev Biol, 2012;361, 68–78.
40. Celton‐Morizur, S., Merlen, G., Conton, D. et al. The insulin/Akt pathway
controls a specific cell division program that leads to generation of binucle­
ated tetraploid liver cells in rodents. J Clin Invest, 2009;122:3307–15.
41. Duncan, A.W., Hanlon Newell, A.E., Sith, L. et  al. Frequent aneuploidy
among normal human hepatocytes. Gastroenterology, 2012;142:25–8.
42. Gupta, S. Hepatic polyploidy and liver growth control. Semin Cancer Biol,
2000;10:161–71.
43. Traub, O., Look, J., Dermietzel, R. et  al. Comparative characterization of
21‐kD and 26‐kD junction proteins in murine liver and cultured hepatocytes.
J Cell Biol, 1989;108:1039–51.
44. Jungermann, K. and Katz, N. Functional specialization of different hepato­
cyte populations. Physiol Rev, 1989;69:708–64.
45. Gougelet, A., Torre, C., Veber, P. et al. T cell factor 4 and β‐catenin chromatin
occupancies pattern zonal liver metabolism. Hepatology, 2014;59:2344–57.
46. Jungermann, K. and Kietzmann, T. Oxygen: modulator of metabolic zona­
tion and disease of the liver. Hepatology, 2000;31:255–60.
47. Thorgeirsson, S.S. and Grisham, J.W. Hematopoietic cells as hepatocyte
stem cells: a critical review of the evidence. Hepatology, 2006;43:2–8.
48. Zajicek, G., Oren, R., and Weinreb, M. Jr. The streaming liver. Liver,
1985;5:293–300.
49. Wang, B., Zhao, L., Fish, M. et al. Self‐renewing diploid Axin 2+ cells fuel
homeostatic renewal of the liver. Nature, 2015;524:180–5.
50. Font‐Burgada, J., Shalapour, S., Ramaswamy, S. et  al. Hybrid periportal
hepatocytes regenerate the injured liver without giving rise to cancer. Cell,
2015;162:766–79.
 1:  Organizational Principles of the Liver 13
51. Allison, M.R. Hepatocytes come out of left field. Hepatology, 2016;63:
1041.–2.
52. Jennem, C.N. and Kubes, P. Immune surveillance by the liver. Nat Immunol,
2013;14:996–1006.
53. Saxena, R. and Theise, N. Canals of Hering: recent insights and current
knowledge. Semin Liver Dis, 2004;24:43–8.
54. Thorgeirsson, S.S. and Grisham, J.W. Liver stem cells, in Stem Cells (ed.
C.S. Potten), Academic Press, London, 1997, pp. 233–85.
55. Kuwahara, R., Kofman, A.V., Landis, C.S. et al. The hepatic stem cell niche:
identification fy label‐retaining assay. Hepatology, 2008;47:1994–2002.
56. Benedetti, A., Bassotti, C., Rapno, K. et al. A morphometric study of the epi­
thelium lining the rat intrahepatic biliary tree. J Hepatol, 1996;24:335–42.
57. Bing, Q., Masyuk, T.V., Muff, M.A. et al. Isolation and characterization of
cholangiocyte primary cilia. Am J Physiol Gastrointest Liver Physiol,
2006;291:G500–9.
58. Alpini, G., Ulrich, C., Roberts, S. et al. Molecular and functional heteroge­
neity of cholangiocytes from rat liver after bile duct ligation. Am J Physiol
Gastrointest Liver Physiol, 1997;272:G289–97.
59. Zhang, H. How biliary tree maintains immune tolerance? BBA Mol Basis
Dis, 2018;1864:1367–73.
60. Braet, F. and Wisse, E. Structural and functional aspects of liver sinusoidal
endothelial cell fenestrae: a review. Comp Hepatol, 2002;1:1–10.
61. Sorensen, K.K., Simon‐Santamaria, J., McCuskey, R.S. et al. Liver sinusoi­
dal endothelial cells. Compr Physiol, 2001;5:1751–74.
62. McCuskey, R.S. and Reilly, F.D. Hepatic microvasculature: dynamic struc­
ture and its regulation. Semin Liver Dis, 1993;13:1–12.
63. Smedsrød, B., DeBleser, P.J., Braet, F. et al. Cell biology of endothelial and
Kupffer cells. Gut, 1994;35:1509–16
64. Smedsrød, B. Clearance function of scavenger endothelial cells. Comparative
Hepatol, 2004;3(Suppl 1):S22.
65. Wang, L., Wang, X., Wang, L. et  al. Hepatic vascular endothelial growth
factor regulates recruitment of rat liver sinusoidal endothelial cell progenitor
cells. Gastroenterology, 2012;143:1555–63.
66. Doherty, D.G. and O’Farrelly, C. Innate and adaptive lymphoid cells in the
human liver. Immunol Rev, 2000;174:5–20.
67. Parker, G.A. and Picut, C.A. Liver immunobiology. Toxicol Pathol,
2005;33:52–62.
68. Doherty, D.G. Immunity, tolerance and autoimmunity in the liver: a compre­
hensive review. J Autoimmun, 2016;66:60–75.
69. Ju, C. and Tacke, F. Hepatic macrophages in homeostasis and liver disease:
from pathogenesis to novel therapeutic strategies. Cell Mol Immunol,
2016;13:316–27.
70. Gomez, P.E., Klapproth, K., Schulz, C. et al. Tissue‐resident macrophages
originate from yolk‐sac derived erythro‐myeloid progenitors. Nature,
2015;518:547–51.
71. Lee, U.E. and Friedman, S.L. Mechanisms of hepatic fibrogenesis. Best
Pract Res Clin Gastroenterol, 2011;25:195–206.
72. Yin, C., Evason, K. J., Asashina, K. et al. Hepatic stellate cells in liver devel­
opment, regeneration and cancer. J Clin Invest, 2013;123:1902–10.
73. Wells, R.G. The portal fibroblast  –  not just a poor man’s stellate cell.
Gastroenterology, 2014;147:41–7.
74. Elias, H. Origin and early development of the liver in various vertebrates.
Acta Hepatol, 1955;3:1–567.
75. Le Douarin, N.M. An experimental analysis of liver development. Med Biol,
1975;53:427–55.
76. Field, H.A., Ober, E.A., Roeser, T., and Stainier, D.Y.R. Formation of the
digestive system in zebrafish. I. Liver morphogenesis. Dev Biol, 2003;253:
279.–90.
77. Seternes, T., Sørensen, K., and Smedsrød, B. Scavenger endothelial cells of
vertebrates: a nonperipheral leukocyte system for high capacity elimination
of waste macromolecules. Proc Natl Acad Sci U S A, 2002;99:7594–97.
78. Columbano, A. and Ledda‐Columbano, G.M. Mitogenesis by ligands of
nuclear receptors: an attractive model for the study of molecular mechanisms
implicated in liver growth. Cell Death Differ, 2003;10:S19–21.
79. Zaret, K.S. Regulatory phases of early liver development: paradigms of
organogenesis. Nat Rev Genet, 2007;3:499–512.
80. Nagy, L. and Schwabe, J.W. Mechanism of the nuclear receptor molecular
switch. Trends Biochem Sci, 2004;29:317–24.
81. Zhao, R. and Duncan, S.A. Embryonic development of the liver.
Hepatology, 2005;41:956–67.
82. Zaret, K.S. Regulatory phases of early liver development: paradigms of
organogenesis. Nat Rev Genet, 2007;3:499–512.
83. Sadler, K.L., Krahn, K.N., Gaur, N.A., and Ukomadu, C. Liver growth in
the embryo and during liver regeneration in zebrafish requires the cell cycle
regulator, uhrf1. Proc Natl Acad Sci U S A, 2007;104:1570–5.
84. Okihiro, M.S. and Hinton, D.E. Partial hepatectomy and bile duct ligation
in rainbow trout (Oncorhynchus mykiss): histologic, immunohistochemical
and enzyme histochemical characterization of hepatic regeneration and
hyperplasia. Toxicol Pathol, 2000;28:342–335.
85. Michalopoulos, G.K. Hepatostat: liver regeneration and normal tissue
maintenance. Hepatology 2017;65:1384–92.
86. Fulop, A.K., Pocsik, E., Brozik, E. et al. Hepatic regeneration induces tran­
sient acute phase reaction: systemic elevation of acute phase reactants and
soluble cytokine receptors. Cell Biol Int, 2001;25:585–92.
87. Fausto, N. Involvement of innate immune system in liver regeneration and
injury. J Hepatol, 2006;45,347–9.
88. Dong, Z., Wei, H., Sun, R., and Tian, Z. The role of innate cells in liver
injury and regeneration. Cell Mol Immunol, 2007;4:241–52.
89. Preziasi, M.E. and Monga, S.P. Update on the mechanisms of liver regen­
eration. Semin Liver Dis, 2017;37:141–51.
90. Martinez‐Hernandez, A. and Amenta, P. The extracellular matrix in hepatic
regeneration. FASEB J, 1995;9:1401–10.
91. Wack, K.E., Ross, M.A., Zegara, V. et al. Sinusoidal ultrastructure during
revascularization of regenerating liver. Hepatology, 2001;33:363–78.
92. Papp, V., Dezso, K., Laszlo, V. et al. Architectural changes during regenera­
tive and ontogenic growth in the rat. Liver Transplant, 2009;18:177–83.
93. Sato, Y., Kayama, S., Tsukada, K., and Hatakeyama, K. Acute portal hyper­
tension reflecting shear stress as a trigger to liver regeneration following
partial hepatectomy. Surg Today (Jpn J Surg), 1997;27:518–26.
94. Schoen, J.M., Wang, H.H., Minuk, G.Y., and Lautt, W.W. Shear stress‐
induced nitric oxide release triggers the liver regeneration cascade. Nitric
Oxide, 2001;5:453–64.
95. Schoen‐Smith, J.M. and Lautt, W.W. The role of prostaglandins in trigger­
ing the liver regeneration cascade. Nitric Oxide, 2005;23:111–17.
96. Huang, W., Ma, K., Zhang, J. et al. Nuclear receptor‐dependent bile acid sign­
aling is required for normal liver regeneration. Science, 2006;312:233–6.
97. Matot, I. and Nochmansson, N. Impaired liver regeneration after hepatec­
tomy and bleeding is associated with shift from hepatocyte proliferation to
hypertrophy. FASEB J, 2017;31:5283–95.
98. Thorgeirsson, S.S., Factor, V.M., and Grisham, J.W. Early activation and
expansion of hepatic stem cells, in Handbook of Stem Cells, Introduction to
Adult and Fetal Stem Cells, vol. 3 (eds. R. Lanza, H. Blau, D. Melton et al.),
Academic Press, San Diego, 2004, pp. 497–512.
99. Evarts, R.P., Nagy, P., Marsden, E., and Thorgeirsson, S.S. A precursor‐­
product relationship exists between oval cells and hepatocytes in rat liver.
Carcinogenesis, 1987;8:1737–40.
100. Paku, S, Schnur, J., Nagy, P et  al. Origin and structural evolution of the
early proliferating oval cells in rat liver. Am J Pathol, 2001;158:1313–23.
101. Choi, T.Y., Ninov, N., Stainer, D.Y. et al. Extensive conversion of hepatic
biliary epithelial cells to hepatocytes after near total loss of hepatocytes in
zebra fish. Gastroenterology, 2014;146:776–88.
102. Schaub, J.R., Malato, Y., Gormond, C. et al. Evidence against a stem cell
origin of new hepatocytes in a common mouse model of chronic liver
injury. Cell Rep, 2014;8:933–939.
103. Lu, W.Y., Bird, T.G., Boulter, L. et al. Hepatic progenitor cells of biliary
origin with liver repopulation capacity. Nature Cell Biol, 2015;17:971–83.
104. Michalopoulos, G.K. and Khan, Z. Liver stem cells: experimental findings
and implications for human disease. Gastroenterology, 2015;149:876–82.
105. Kopp, J.L., Grompe, M., and Sander, M. Stem cells versus plasticity in liver
and pancreas regeneration. Nature Cell Biol, 2016;18:238–44.
106. Dezso, K., Papp, V., Bugyik, E. et al. Structural analysis of oval‐cell‐medi­
ated liver regeneration in rat. Hepatology, 2012;56:1457–67.
107. Dezso, K., Rokusz, A., Bugyik, E. et al. Human liver regeneration in advanced
cirrhosis is organized by the portal tree. J Hepatol, 2017;66:778–86.
108. Cast, A.E., Walter, T.J., and Huppert S.S. Vascular patterning sets the stage
for macro and micro hepatic architecture. Dev Dynam, 2015;244:497–506.
 Embryonic Development
2 of the Liver
Kenneth S. Zaret1, Roque Bort2, and Stephen A. Duncan3
1
Institute for Regenerative Medicine, Department of Cell and Developmental Biology, Perelman School of
Medicine, University of Pennsylvania, Smilow Center for Translational Research, Philadelphia, PA, USA
2
Instituto de Investigación Sanitaria La Fe (IIS La Fe), Unidad de Hepatología experimental, València, Spain
3
Department of Regenerative Medicine and Cell Biology, Medical University of South Carolina, Charleston, SC, USA

INTRODUCTION ACQUISITION OF HEPATIC

The liver is one of the first organs to develop in the embryo and
it rapidly becomes one of the largest organs in the fetus. The
most essential function of the mammalian fetal liver is to pro-
vide a site for hematopoiesis. The early dependence of the fetus
on its own blood cell supply makes embryonic liver growth and
viability a sensitive phenotypic indicator for gene inactivation
studies. From an experimental perspective, the large size and
small number of cell types in the developing liver make it easy
to study, and new insights have emerged from recent work on
genetically modified mice, explants of embryonic tissues, pluri-
potent stem cell differentiation, and other vertebrate models
such as zebrafish and frogs. Much is being learned about how
gene function is orchestrated to control tissue morphogenesis,
helping to establish liver development as a paradigm for the
genesis of other gut‐derived tissues. Understanding the mecha-
nisms that govern liver development should also provide insight
into future therapies for liver diseases. Examples of such appli-
cations include activating stem cells in the adult liver, replenish-
ing diseased livers with cells from pluripotent stem cells,
transdifferentiating cells from different organs, and reconstitut-
ing proper liver morphology and function.
This review will describe the early stages of liver develop-
ment from the initial specification to hematopoietic cell inva-
sion, which covers the competence of progenitor cells to become
hepatocytes, the formation of the liver bud, and the early mor-
phogenesis and differentiation of the liver. The review will also
present the key effectors of hepatocyte maturation and discuss
the control of liver size and regeneration in the embryo and in
the adult. The reader is referred to other reviews for summaries
of the mid‐ and late‐fetal stages of liver development [1–9].
COMPETENCE WITHIN
THE ENDODERM
The liver, lung, pancreas, thyroid, and gastrointestinal tract are
derived from the anterior‐ventral definitive endoderm, which is
one of the three germ layers that arise during gastrulation. Initially,
the endoderm is an epithelial sheet that lines the ventral surface of
the embryo. Infolding of the sheet at the anterior and posterior of the
embryo generates the foregut and hindgut (Figure  2.1). When
these morphogenetic movements reach the middle of the embryo, the
gut tube closes off. During gut tube formation, different tissues
are specified along the anterior–posterior and dorsal–ventral axes
of the embryo, with the liver arising from the prospective ventral
endoderm domain of the foregut (Figure  2.1). Are there “pre‐
patterns” or local domains of endoderm that are competent to
differentiate into the liver? LeDouarin demonstrated, using heter-
otypic grafts of quail embryo segments into recipient chick
embryos, that only the prospective anterior‐ventral domain of
endoderm had the capacity to develop into the liver [10, 11].
High‐resolution lineage tracing in mouse embryos has revealed
that distinct domains within the endoderm contribute to the devel-
oping liver [12]. The majority of cells derive from two bilateral
patches within the lateral endoderm, and a smaller population
derives from a small clutch of cells at the ventral midline of the
anterior endoderm. Morphogenic movements ensure that these
populations converge during the formation of the liver bud.
More recent studies with mouse embryos and more sensitive
assays of gene expression revealed that while the prospective
dorsal endoderm (Figure 2.1) does not normally activate liver
genes or become liver, in a tissue explant assay the dorsal endo-
derm cells could initiate liver gene expression when isolated

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
 2:  Embryonic Development of the Liver 15
cardiac head
mesoderm
(FGF source)
septum
transversum
mesenchyme fore- somites
(BMP source) gut hind-
gut
ventral-anterior
endoderm
(suppression of Wnt
allows hepatic dorsal-posterior endoderm
potential) (Wnt signaling suppresses
hepatic potential)
Figure 2.1  Parasagittal diagram of a mammalian embryo at the time
of hepatic specification. Relevant tissues and signaling molecules are
indicated. The time of development corresponds to about 8.25 days’
gestation in the mouse and about three weeks in the human.
from its adjacent mesodermal tissues [13]. The molecular
mechanism underlying these transplantation studies has been
deciphered using Xenopus (frog) embryos, and similar mecha-
nisms likely operate in mammalian embryos. It was found that
due to the anterior expression of Wnt inhibitors such as Dkk,
sFrp‐1 or sFrp‐5, the function of the Wnt/β‐catenin signaling
pathway is repressed upon gastrulation in the anterior endo-
derm [14], but it is active posteriorly, where the Wnt inhibitors
are absent. Wnt downregulation in the anterior endoderm was
shown to be crucial for liver and pancreas specification; in fact,
posterior endoderm that was forced to activate GSK‐3β, a Wnt
inhibitor, possessed hepatogenic competence [15]. Wnt signal-
ing induces the transcription factor Vent in the endoderm,
which in turn represses the homeobox gene Hex that is required
in the endoderm for liver and pancreas development [16–21].
In summary, Wnt repression in the anterior endoderm allows
liver and pancreas specification, whereas active Wnt signaling
in the posterior endoderm suppresses those fates. These find-
ings provide a molecular basis to help explain the patterning of
the endoderm.
Another approach to understanding how the endoderm gains
the competence to develop into liver is to identify the regulatory
transcription factors that directly enable the process. Regulatory
factors that are known to be important for liver differentiation
are expressed in prehepatic endoderm, prior to hepatic specifi-
cation. Such factors could operate by helping to open up chro-
matin structure for genes that need to be transcribed during liver
differentiation [22]. Transcription factors that are expressed in
the prehepatic domain of ventral foregut endoderm, and later in
the liver, include FOXA1, FOXA2 [23–26], GATA4, and
GATA6 [27–30]. The roles of these transcription factors can be
discerned by understanding how the factors function in a chro-
matin context. DNA binding by both FOXA and GATA factors
is required for the activity of a transcriptional enhancer of alb1,
the gene encoding serum albumin [31, 32]; alb1 is one of the
earliest genes to be activated in hepatic development [13, 33].
An analysis of the DNA binding sites for FOXA2 and GATA
[34] showed that both sites are occupied on the alb1 enhancer in
the endoderm (Figure  2.2), prior to alb1 transcriptional
­activation or hepatic commitment [13, 31]. Once hepatic speci-
fication has occurred, adjacent binding sites for a variety of
endoderm
alb1–
–10 kb X
GATA FoxA
liver bud
alb1+
C/EBP FoxA eY FoxA NF1
Figure 2.2  Transcriptional competence factors in the endoderm.
DNA binding sites for GATA and FOXA transcription factors are occu-
pied on the alb1 gene prior to alb1 expression or hepatic commitment.
During hepatic specification, other transcription factors bind DNA at
adjacent sites and the albumin gene becomes active.
other transcription factors become occupied at the enhancer and
alb1 becomes active (Figure 2.2).
Genetic experiments have confirmed the essential role of
FOXA and GATA factors in hepatogenesis. A transgenic mouse
with a conditional ablation of both FOXA1 and FOXA2 in the
foregut endoderm completely lacked the induction of hepatic
markers of specification, such as alb1, afp, or ttr [35]. Similarly,
while the liver is specified in GATA6−/− or GATA4−/− mouse
embryos (though subsequent development is blocked), a double
knockout of both genes in zebrafish revealed a complete failure in
liver induction [36]. The role of GATA factors in controlling endo-
dermal fate appears to be conserved in humans. Several groups
have shown that depletion of GATA6 affects the differentiation of
endoderm from human induced pluripotent stem cells (iPSCs)
[37–39]. These findings provide genetic evidence supporting the
concept that FOXA and GATA factors serve as “pioneer factors”
in the endoderm, being among the first to bind a target gene in
development and endow competence for the gene to be activated,
under the influence of cell‐type inductive signals.
FROM DEFINITIVE TO HEPATIC
ENDODERM
Once the hepatogenic competence in the definitive endoderm is
acquired, what causes the liver to be specified from the ventral
foregut endoderm? The mesoderm secretes patterning signals
that instruct the differentiation of the underlying endoderm.
Several lines of research conducted in chicken and mouse dem-
onstrated that cardiac mesoderm helps instruct the underlying
ventral foregut endoderm to become hepatic endoderm
(Figure  2.1). In vitro culture of mouse embryonic explants
together with the use of zebrafish and Xenopus embryos have
identified some of the relevant molecular signals.
There are over 20 genes for fibroblast growth factors (FGFs)
and 4 genes for FGF receptors [40, 41]. Each of the FGF
receptors has different binding specificities for FGFs, and
­
the existence of multiple spliced isoforms of the receptors pro-
vides further complexity to the receptor–ligand relationships.
16 THE LIVER:  FROM HEPATIC ENDODERM TO LIVER BUD
Cardiogenic mesoderm expresses FGF1, FGF2, FGF8, and
FGF10 in the period prior to hepatic induction in the endoderm.
In an embryonic tissue explant system, purified FGF1 or FGF2
can efficiently activate liver gene expression in ventral foregut
explants where cardiogenic mesoderm has been removed, and
an FGF antagonist can inhibit liver gene induction in foregut
explants where cardiogenic mesoderm has been retained [42].
FGF binding induces tyrosine phosphorylation activity by the
cytoplasmic domain of the receptors, resulting in activation of
mitogen‐activated protein kinase (MAPK) signaling pathways
within cells [41, 43]. A combination of in vivo‐genetic, whole‐
embryo culture and tissue explant approaches revealed that a
transient activation of the MAPK pathway by FGF signaling in
foregut endoderm explants is necessary for the initiation and
stabilization of the hepatic program [44]. Although the PI3K/
AKT pathway is activated in the endoderm shortly after the
MAPK pathway, PI3K/AKT pathway activation appears not to
be downstream of FGF signaling and the pathways do not cross-
regulate in the hepatic endoderm [44]. Ultimately, the action of
FGF signaling must activate gene expression programs that
drive hepatic fate. The differentiation of human iPSCs to a
hepatic fate has allowed researchers to identify the immediate
targets of FGF signaling [45]. Genes that are directly activated
during the endoderm‐to‐hepatic transition include several tran-
scription factors, growth factors, and signaling molecules.
Interestingly, one of the direct targets of FGF encodes NKD1,
which is a repressor of WNT signaling. When NKD1 was
depleted, differentiation of the endoderm to a hepatic fate was
inhibited [45]. These findings imply that FGF drives hepatic fate
in part by transiently suppressing canonical WNT signaling.
BMP4 is a member of the TGFβ superfamily. BMP2, BMP4,
BMP5, and BMP7 are highly expressed in the septum transver-
sum mesenchyme (STM), the latter consisting of loose mesen-
chyme cells that surround the cardiac and ventral endoderm
domains [46–50]. The BMP receptors BMPRIA, BMPRII, and
ActRIIA are expressed in the endoderm [51, 52]. The mouse
foregut explant system described above was used to reveal that
BMP signaling from STM cells is also crucial for hepatic gene
induction in the endoderm [53]. Thus, hepatic induction requires
positive signals (FGF and BMP) from two different cell sources
(cardiac mesoderm and STM), indicating the importance of
combinatorial signaling. The general relevance of FGFs and
BMPs for hepatic induction has been highlighted by genetic
experiments in zebrafish [54].
The zebrafish studies of hepatic induction, which cleverly
traced the fate of individual endoderm cells [54], confirmed an
initial observation made with populations of endoderm cells iso-
lated from mouse embryos. That is, the foregut endoderm makes
a cell fate choice for the liver or pancreas, under the influence of
FGF and BMP signaling, as no FGF and BMP allow pancreas
induction, whereas elevated levels of both signaling molecular
induce a liver fate [55]. Subsequent studies revealed that the
signaling network is dynamic, with intracellular FGF and BMP
responses changing in the endoderm tissue as it moves past the
cardiac mesoderm and STM signaling centers [56]. During this
period, the hepatogenic BMP signal in endoderm is transmitted
to target genes via the transcription factor SMAD4, which in
turn recruits the p300 coactivator [57]. P300 is a histone acetyl-
transferase and its activity modulates liver gene induction in the
endoderm and thus the number of cells that become specified to
liver, instead of pancreas [57].
The role of Wnt signaling during hepatic induction appears
dynamic. The FGF‐BMP induction of the liver occurs when
Wnt signaling is being suppressed in the foregut. But shortly
after the induction of the hepatic program in the endoderm, Wnt
signaling appears to be required for further outgrowth of the
endoderm into a liver bud [15]. In zebrafish, expression of
Wnt2b in the lateral plate mesoderm, acting through the β‐catenin
canonical pathway, appears essential for liver specification in
the endoderm and bud induction [58]. Differences in the devel-
opment of the liver in fish versus amniotes may explain an
­earlier positive role for Wnt signaling in hepatic induction [5].
Like most organs, the liver is asymmetrically positioned with
the developing embryo. Patients with heterotaxy syndrome
(isomerism), in which organ asymmetry is disrupted, commonly
have defects associated with liver function including biliary
atresia [59]. The molecular basis for the asymmetric nature of
the liver has only recently been explored and our understanding
remains rudimentary. However, studies in zebrafish have found
that communication between EphrinB1 and EphB3 are crucial
to coordinate the movement of the hepatic endoderm with the
adjacent lateral plate mesoderm [60]. EphrinB1 controls hepa-
toblast migration while EphB3 acts in the mesoderm to repel the
hepatoblasts. By working together these proteins control the
directional migration of the hepatoblasts that is required to cor-
rectly position the liver.
FROM HEPATIC ENDODERM
TO LIVER BUD
The progress from hepatic endoderm to liver bud has been
divided into three morphogenetic stages [17]: stage I, the forma-
tion of a thickened, columnar hepatic epithelium; stage II, the
formation of a pseudostratified epithelium (Figure  2.3c); and
stage III, laminin breakdown and hepatic cell migration from
the epithelium into the STM. Hepatoblast migration is accom-
panied by major remodeling of the extracellular matrix sur-
rounding the hepatic cells [61]. During this period, the entire
ventral foregut domain extends toward the midgut, bringing the
liver region with it (Figure 2.3a,b). The mass of cells emerging
from the endodermal epithelium and concentrating in the
­septum transversum is referred to as the liver bud, and the cells
within the liver bud are referred to as hepatoblasts. Hepatoblasts
will later differentiate into hepatocytes and cholangiocytes
(­biliary cells).
At stage I of liver bud formation in mammals, endothelial
cells, not yet assembled into a vasculature, are adjacent to the
hepatoblast epithelium [62]. Genetic depletion of endothelial
cells at this stage led to the discovery that endothelial cells are
an essential stimulus to further liver bud development: the
absence of endothelial cells halted liver bud development at
stage II, with hepatoblasts remaining within the limits of the
basal epithelial membrane. A similar organogenic stimulatory
role has been attributed to endothelial cells in the developing
pancreas [63]. In the pancreas, VEGF‐A secreted by beta cells
attracts endothelial cells to the developing islet [64]. Endothelial
 2:  Embryonic Development of the Liver 17

(a) (b)
neural
tube

dorsal
aorta
gut tube

pericardio/
peritoneal
canal
sinus
venosus
liver bud

(c)

hepatic endoderm
cells
(pseudo-stratified
epithelium)

endothelial cells and

septum transversum
mesenchyme cells

Figure 2.3  The beginning of liver development. (a) Mouse embryo, 9 days gestation. Tail is removed; arrow indicates plane of section in
(b). (b) 100× magnification of transverse section. Note the thickening of the endodermal epithelium of the gut tube at the region of the liver bud.
Boxed area is magnified to 400× in (c). Arrows depict the liver bud and other landmarks of the embryo. Photographs courtesy of J. Rossi.

cells, in turn, form a basal membrane rich in laminins, collagen
IV, and fibronectin required for the correct function of beta
cells. The nature of the signals originating from the endothelial
cells that trigger morphogenesis and cell differentiation in the
emerging liver bud is not known.
As already stated, the formation of the liver bud requires the
degradation of the basal membrane to allow migration of hepat-
ocytes into the STM. One of the genes controlling this step is
Prox1 [61]. The Prox1−/− mouse embryonic hepatic endoderm is
unable to degrade the basal epithelial membrane, resulting in
the clustering of hepatoblasts into a smaller liver. It is not known
whether this activity is caused by a slow degradation or produc-
tion of excess basal membrane components by the embryonic
hepatoblasts. Expression of Prox1 is regulated in part by the
transcription factor Tbx3. As in Prox1−/− embryos, outgrowth of
the liver bud is severely inhibited in Tbx3−/− embryos; however,
it has been suggested that Tbx3 acts to control hepatic progeni-
tor cell fate and proliferation through regulating expression of
multiple transcription factors and cell cycle regulators, rather
than acting solely through Prox1 [65, 66].
As noted above, the homeodomain factor Hex is crucial for
liver development [20] and plays an essential role in transforma-
tion from stage I to stage II [17]. Hex seems to control different
aspects of liver bud formation, including the adequate prolifera-
tion of hepatic endoderm cells, the formation of a pseudostrati-
fied epithelium, and the stable maintenance of the hepatic cell
type [16–18]. How Hex exerts its role is not known, but it has
been suggested that it represses sonic hedgehog signaling within
the ventral foregut endoderm, contributing to the exclusion of
the intestinal fate [17].
BEYOND THE LIVER BUD:
INTRAHEPATIC CHOLANGIOCYTE
DIFFERENTIATION
The complex signals that cause the emergence of the liver bud
are followed closely by distinct signals required to grow
the  bud into the liver organ. The hepatoblasts differentiate
into hepatocytes and intrahepatic cholangiocytes at about
­embryonic day of gestation 13.5 (E13.5) in the mouse and
7  weeks in humans [67–71]. Substantial advances in our
understanding of the mechanisms that control the formation
of the intrahepatic ducts have been provided through the
study of conditional knockout mouse embryos and the use of
Cre‐mediated lineage tracing [72]. The intrahepatic bile ducts
consist of cholangiocytes (biliary epithelial cells). They
derive from hepatoblasts that surround branches of the portal
vein to form a structure called the ductal plate. Portal vein
mesenchyme secretes high levels of TGFβ which drives the
differentiation of cholangiocytes and suppresses hepatocyte
fate. The action of TGFβ is strictly regulated and only a single
layer of hepatoblasts that surround the portal vein differenti-
ate to cholangiocytes [73]. The molecular effectors involved
in the fate decisions involve the ONECUT transcription
18 THE LIVER:  ROLE OF MESENCHYMAL CELLS IN HEPATOCYTE DIFFERENTIATION
Pre-natal liver development
BMP4
FGF2
FGF8
Definitive Hepatic
endoderm (low endoderm
FoxA1/2 Wnt) Hex
Gata4/6 Prox-1
Figure 2.4  Schema of steps of liver development and relevant signals and transcription factors that help mediate the steps. See text for details.
factors HNF6 (OC1) and OC2. Double homozygous OC1/2
mutant embryos present a disturbance of the TGFβ gradient
across the axis of the portal vein to the hepatic parenchyma.
As a result of the increase in TGFβ signaling, hepatoblasts
lying distal to the portal veins express both hepatocyte and
cholangiocyte markers [73] (Figure  2.4). Control of TGFβ
signaling occurs at multiple levels, including regulation of the
concentration of TGFβ type II receptors (TBRII) on the cell
surface of periportal hepatoblasts [74]. The Hippo‐YAP sign-
aling pathway has also been implicated in controlling the
TGFβ axis by directly promoting expression of TGFβ2 and
inhibiting expression of key hepatocyte transcription factors,
such as HNF4 [75].
Together with TGFβ, Notch signaling is also required for
­normal biliary tract morphogenesis. Alagille syndrome, a devel-
opmental disorder characterized by a paucity of intrahepatic
bile ducts (IHBD), is caused predominantly by mutations in the
Jagged1 (JAG1) gene, which encodes a ligand for Notch family
receptors [76]. Studies using notch loss‐of‐function mice and
zebrafish models have indicated that Notch regulates bile duct
abundance rather than cell fate specification [77–79]. However,
overexpression of the Notch intracellular domain, which drives
expression of Notch target genes, increases the expression of
cholangiocyte transcription factors and represses hepatocyte
transcription factors [80]. These findings imply that Notch
could directly promote cholangiocyte differentiation. The
mechanism through which Notch controls cholangiocyte forma-
tion is complicated by the fact that Notch interaction with its
ligands Jagged 1 and Delta can have both cell‐extrinsic and
cell‐intrinsic effects [81].
Although during development cholangiocytes derive from
the hepatoblasts within the ductal plate, it has recently been
shown that hepatocytes retain the capacity to fully transdiffer-
entiate into cholangiocytes in the livers of adult mice that lack
an intrahepatic biliary system [82]. Unlike the differentiation of
cholangiocytes from hepatoblasts, transdifferentiation of adult
hepatocytes occurred independently of Notch and was driven
by TGFβ signaling. These findings suggest new strategies for
the potential treatment of Alagille syndrome and cholestatic
liver disease.
On the other hand, a recent report suggests that the cells of
the extrahepatobiliary system (gallbladder, the hepatic, cystic
and common ducts), share a common origin with the ventral
pancreas in a group of cells that are slightly caudal to the liver
intrahepatic Post-natal
bile duct
cholangiocytes
HNF-6
Notch
HNF-1b
Bipotential Embryonic Adult
hepatoblasts hepatocytes hepatocytes
Hex HNF-4
OC1(HNF-6) C/EBPs
OC2
bud, or within the caudal portion of the liver bud [82A]. Cell‐
tracing experiments will definitively assess this issue.
ROLE OF MESENCHYMAL CELLS
IN HEPATOCYTE DIFFERENTIATION
Hepatocyte differentiation and liver morphogenesis is
dependent upon the cellular microenvironment. During early
stages of liver bud development, as discussed above, the
microenvironment is provided by endothelial and mesenchy-
mal cells in the STM and, after E10.5 in the mouse, when
liver becomes a hematopoietic organ, hematopoietic stem
cells (HSCs) also contribute. The relevance of these mesen-
chymal cell types is highlighted by genetic loss‐of‐function
studies in mice, where genes expressed in mesenchymal cells
surrounding the nascent liver but not in hepatoblasts are
found to be essential for correct liver formation. One of these
cases, Lhx2, is a LIM‐homeobox gene expressed in the STM
and mesenchymal components in the adult liver (presumably
stellate cells). Lhx2−/− livers display a disrupted cellular
organization with increased deposition of ECM and an altered
gene expression pattern of the early hepatocytes [83]. The
mesenchymal cells are also required for the proliferation of
the fetal hepatocytes. When Gata4 was deleted specifically in
the liver mesenchymal cells, embryos exhibited severe defects
in liver growth, hepatocyte proliferation and survival, and
fetal hematopoiesis [84].
Besides the early role of individual endothelial cells in pro-
moting liver bud growth, blood vessels develop de novo within
the liver bud (Figure 2.3b), forming a capillary bed that becomes
interspersed within the expanding hepatoblast population [85].
These transitions establish the liver’s sinusoidal architecture,
which is critical for organ function and sets the stage for the fetal
liver to support hematopoiesis. Hematopoietic cells migrate to
the early liver first from the yolk sac [86] and later from the
aorta–gonad–mesonephros region [87]. Proper intermediate fila-
ment expression is critical for blood cell homing, as erythrocytes
accumulate excessively in the fetal liver of keratin 8 mutants
[88]. Embryos deficient in the heavy metal‐responsive transcrip-
tion factor MTF‐1 exhibit reduced cytokeratin expression as well
as enlarged sinusoids, and dissociated epithelial cells, but not
anemia [89]. The primary defect in MTF‐1‐deficient embryos
 2:  Embryonic Development of the Liver 19
appears to be in a failure to control metal homeostasis and the
oxidation–reduction state in hepatocytes at mid‐gestation.
Both hematopoietic and endothelial cells provide differentiat-
ing signals to the hepatoblasts, because heterogeneity in hepatic
gene expression has been found to be related to the vascular
architecture of the fetal liver [90]. More specifically, mouse
gene inactivation studies have shown that oncostatin M signal-
ing from hematopoietic cells to nascent hepatocytes is critical
for liver growth [91]. Impaired hematopoietic cell proliferation
in c‐myb mutant embryos [92] or impaired erythrocytic cell pro-
liferation in retinoblastoma gene (Rb) mutant embryos [93, 94]
also result in impaired liver growth. The failure of hematopoi-
etic cells to migrate to the liver in β1 integrin‐null embryos is
therefore also thought to contribute to the liver developmental
defect in the β1 mutants [95, 96].
As the fetal liver matures and grows, a capsule forms around
it from mesothelial cells. Although the role of the mesothelium
during liver development has not been appreciated, a growing
body of evidence supports a role in promoting fetal hepatocyte
proliferation. N‐myc gene expression in the liver capsule and
jumonji gene expression in the hepatic stromal cells help pro-
mote growth during mid‐gestation [97–99]. The mesothelial
capsule is a rich source of growth factors that may act on the
fetal hepatocytes [100]. Moreover, deletion of the zinc finger
transcription factor Wt1, which is highly expressed in the liver
capsule during development, inhibits proliferation of fetal hepat-
ocytes, diminishes liver growth, and affects formation of the
liver lobules [100, 101]. Although many of the genes that affect
hepatoblast proliferation are probably most important during the
initial transition from the liver bud to the organ stage, inactiva-
tion of these genes usually manifests itself as a hematopoietic
defect well after the organ is formed. In these cases, a liver
­capsule develops and hematopoietic cells migrate to the liver,
but the paucity of hepatoblasts leads to a defect in the hemat-
opoietic environment and, consequently, embryonic lethality.
Furthermore, mutations of certain liver regulatory factors yield a
fetal liver growth defect due to apoptosis of the hepatoblasts.
These proteins include c‐jun [102, 103], IKK2 [104], RelA
[105], and XBP‐1 [106].
EMBRYOLOGIC CONTROL OF LIVER
REGENERATION
The liver is among the few internal organs that can rapidly
regenerate after removal of tissue in the adult. Recent studies
indicate that the regenerative capacity of the liver is attained as
early as the hepatoblast stage in embryos. Tissue complementa-
tion experiments, where Hex−/− mouse embryonic stem cells
were injected into Hex+/+ blastocysts, discovered that at the
onset of liver morphogenesis (E9.0), wild‐type hepatic progeni-
tor cells can increase their proliferation rate to compensate for
the failure of Hex−/− cells in the liver bud to survive [17]. In an
independent study, when two‐thirds of liver progenitors are
genetically ablated between E9.5 and E13.5, the remaining cells
are still able to engage in a compensatory growth, compensating
to generate a normal‐sized fetal liver within 4 days [107].
The  pancreas, also an endoderm‐derived organ, is unable to
grow in this fashion, resulting in major losses of pancreatic
mass after partial ablation of its embryonic progenitors [107]. It
seems that liver, but not the pancreas, has a full regenerating
potential at nearly all developmental stages and while the
growth limit of the liver is set by the needed organ size, in the
case of the pancreas it is set by the cell division number. It is
noteworthy that even when the pancreas does not recover its
original size after 50% pancreatectomy, it does recover nearly
100% of its beta cell mass after four weeks [108], suggesting
different regenerating capabilities between the endocrine and
exocrine compartment. Another interesting aspect is the differ-
ence between human and animals, since humans, unlike rodents,
do not have a significant increase in beta cell proliferation after
partial pancreatectomy [109]. Identifying the genetic programs
that cause the marked difference in hepatic and pancreatic
regeneration at the embryonic stage, when the cells are other-
wise quite immature, could be a convenient way to reveal the
basis for adult liver regeneration.
HEPATOCYTE DIFFERENTIATION
Hepatocyte differentiation spans from liver specification in ven-
tral foregut endoderm until the postnatal maturation of hepato-
cytes (Figure  2.4). Downstream of signaling molecules that
induce liver differentiation are the transcription factors that
execute the liver program, including HNF1, HNF4, HNF6,
FOXA, and C/EBP [4–6]. These liver‐specific genes activate, in
a cross‐regulatory fashion, each other’s promoters. By estab-
lishing both positive and negative feedback loops, the transcrip-
tion factors generate stable gene regulatory networks that ensure
expression of genes that are critical for liver function [72, 110].
In mice, HNF4 acts as a central regulator of hepatocyte gene
expression, but is dispensable for early formation of the liver
[111]. However, when its role is examined during the differen-
tiation of human iPSCs, the requirement for HNF4 in control-
ling the onset of hepatocyte gene expression is even more strict
[112]. Whether the strict requirement for HNF4α in human cells
reflects differences between species or between the model sys-
tems remains to be established. Reverse transcription polymer-
ase chain reaction (RT‐PCR) analyses in the mouse revealed
that the expression of the primary liver‐enriched transcription
factors was unchanged in HNF4−/− livers at E12, except for PXR
and HNF1α [111]. However, a large number of tissue‐specific
genes involved in the maturation of hepatoblasts to hepatocytes
failed to be induced in HNF4−/− embryos. The extensive effects
were explained by a genome‐wide chromatin analysis, whereby
the location of HNF4 protein bound to promoter sequences of
nearly 10 000 genes was assessed simultaneously, using micro-
array technology [113]. Such analysis was also performed for
HNF6 and HNF1α [113]. HNF1α in human hepatocytes was
bound to 1.6% of the genes on the array; HNF6 bound 1.7% of
the genes; and HNF4α bound to a striking 12% of the genes
on the array. The genes bound by HNF4α corresponded to ~42%
of the genes bound by RNA polymerase II. That is, nearly half of
the active genes tested in the liver are bound by HNF4α, and
similar results were obtained by analyzing HNF4α binding to
the entire genome [114]. Subsequent studies found that HNF4
20 THE LIVER:  REFERENCES
regulates genes involved in cell junction assembly and adhesion
in the developing liver [115], consequently promoting epithelial
maturation of the liver parenchyma [116].
The apparently limited role of relevant transcription factors
such as HNF1α, in liver development, based on gene inactiva-
tion studies in animals, contrasts sharply with ectopic and over-
expression studies in cultured hepatic cell lines, which suggested
that HNF1α is critical for the expression of a wide variety of
liver‐specific genes [117]. Such distinctions indicate how
strongly the mechanisms of gene regulation are influenced by
whole‐animal physiology, beyond simple notions of gene redun-
dancy (e.g. [118, 119]). Ultimately, it is crucial to determine
how inductive signaling pathways and epigenetic modifications
converge on regulatory transcription factor genes to coordi-
nately promote early liver differentiation and morphogenesis.
FUTURE AND PERSPECTIVES
This review has highlighted ways that our understanding of
embryonic liver development has expanded greatly in the past
25 years or so. The regulatory molecules that control liver
development are now being used successfully to program
hepatic cells from embryonic stem cells and other sources,
including the combination of various cell types to make orga-
noids [120–122]. Given the large number of proteins that were
discovered in studies of adult livers and yet give embryonic
liver phenotypes when deleted in mice, there is high confidence
that many of these proteins’ functions in the adult are a reca-
pitulation of activities in the embryo. Thus, continued investiga-
tion of embryonic liver development is certain to be instructive
about the function, regeneration, and repair of the adult liver,
and seems likely to provide new sources of cells and molecules
to combat liver disease [123]. Further refinements in the ability
to inactivate genes in the early liver and better methods of
embryo tissue culture are needed to advance the analysis of
liver development. Such advances may come from adaptations
of methodology for other endoderm‐derived organ systems
[124]. In addition, relevant new genes will emerge from studies
of model organisms where genetic screens can be coupled with
knowledge of genomic sequences and interrelationships of pro-
tein function. New models have also emerged that allow high‐
resolution insight into the molecular basis of cell differentiation
[125]. The differentiation of human embryonic stem cells and
induced pluripotent stem cells toward a hepatic fate appears to
closely mimic known developmental processes [126]. Having
access to a tissue culture model of liver cell differentiation can
complement the study of traditional animal models. The differ-
entiation of iPSCs in culture is relatively synchronous, which
allows researchers to address the molecular changes that rapidly
and dynamically occur in response to growth factor signaling
[45, 127]. Moreover, tissue culture models allow researchers to
apply high‐­throughput methods and genome‐wide assays to
identify novel pathways and mechanisms that control cell fate
[127–129]. Considering that our mechanistic understanding of
liver development has emerged so recently, the prospects are
bright for much deeper knowledge and new applications for
liver therapies in the future.
ACKNOWLEDGMENTS
Thanks to Melanie Song for help in preparing the manuscript.
R.B. is supported by a grant from the Spanish Ministry of
Science (SAF‐51991R). K.Z.’s research on liver development is
supported by a grant from the NIH (GM36477).
REFERENCES
1. Du Bois, A.M. The embryonic liver, in The Liver, Vol. 1 (ed. C.H. Rouiller),
Academic Press, New York, 1963, pp. 1–39.
2. Elias, H. Origin and early development of the liver in various vertebrates.
Acta Hepat, 1955;3:1–57.
3. Severn, C.B. A morphological study of the development of the human liver.
Am J Anat, 1968;133:85–108.
4. Zaret, K.S. Regulatory phases of early liver development: paradigms of
organogenesis. Nat Rev Genet, 2002;3(7):499–512.
5. Zaret, K.S. Genetic programming of liver and pancreas progenitors: lessons
for stem‐cell differentiation. Nat Rev Genet, 2008;9(5):329–40.
6. Zhao, R. and Duncan, S.A. Embryonic development of the liver. Hepatology,
2005;41(5):956–67.
7. Si‐Tayeb, K., Lemaigre, F.P., and Duncan, S.A. Organogenesis and develop-
ment of the liver. Dev Cell, 2010;18(2):175–89.
8. Ober, E.A. and Lemaigre, F.P. Development of the liver: Insights into organ
and tissue morphogenesis. J Hepatol, 2018;68(5):1049–62.
9. Gordillo, M., Evans, T., and Gouon‐Evans, V. Orchestrating liver develop-
ment. Development, 2015;142(12):2094–108.
10. Fukuda‐Taira, S. Hepatic induction in the avian embryo: specificity of
reactive endoderm and inductive mesoderm. J Embryol Exp Morphol,
­
1981;63:111–25.
11. Le Douarin, N.M. et al. Origin of hemopoietic stem cells in embryonic bursa
of Fabricius and bone marrow studied through interspecific chimeras. Proc
Natl Acad Sci U S A, 1975;72(7):2701–5.
12. Tremblay, K.D. and Zaret, K.S. Distinct populations of endoderm cells
­converge to generate the embryonic liver bud and ventral foregut tissues.
Dev Biol, 2005;280:87–99.
13. Gualdi, R. et  al. Hepatic specification of the gut endoderm in vitro: cell
­signaling and transcriptional control. Genes Dev, 1996;10:1670–82.
14. Schohl, A. and Fagotto, F. A role for maternal beta‐catenin in early ­mesoderm
induction in Xenopus. Embo J, 2003;22(13):3303–13.
15. McLin, V.A., Rankin, S.A., and Zorn, A.M. Repression of Wnt/beta‐catenin
signaling in the anterior endoderm is essential for liver and pancreas
­development. Development, 2007;134(12):2207–17.
16. Bort, R. et al. Hex homeobox gene‐dependent tissue positioning is required for
organogenesis of the ventral pancreas. Development, 2004;131(4):797–806.
17. Bort, R. et al. Hex homeobox gene controls the transition of the endoderm
to  a pseudostratified, cell emergent epithelium for liver bud development.
Dev Biol, 2006;290(1):44–56.
18. Hunter, M.P. et al. The homeobox gene Hhex is essential for proper hepatoblast
differentiation and bile duct morphogenesis. Dev Biol, 2007;308(2):355–67.
19. Keng, V.W. et al. Homeobox gene Hex is essential for onset of mouse embry-
onic liver development and differentiation of the monocyte lineage. Biochem
Biophys Res Commun, 2000;276(3):1155–61.
20. Martinez‐Barbera, J.P. et al. The homeobox gene hex is required in definitive
endodermal tissues for normal forebrain, liver and thyroid formation.
Development, 2000;127(11):2433–45.
21. Wallace, K.N. et al. Zebrafish hhex regulates liver development and diges-
tive organ chirality. Genesis, 2001;30(3):141–3.
22. Zaret, K., Developmental competence of the gut endoderm: genetic potentia-
tion by GATA and HNF3/fork head proteins. Dev Biol, 1999;209:1–10.
23. Ang, S.L. et al. The formation and maintenance of the definitive endoderm
lineage in the mouse: involvement of HNF3/forkhead proteins. Development,
1993;119(4):1301–15.
24. Monaghan, A.P. et al. Postimplantation expression patterns indicate a role
for the mouse forkhead/HNF‐3α, β, and γ genes in determination of the
definitive endoderm, chordamesoderm and neuroectoderm. Development,
1993;119:567–78.
 2:  Embryonic Development of the Liver 21
25. Ruiz i Altaba, A. et  al. Sequential expression of HNF‐3α and HNF‐3β by
embryonic organizing centers: the dorsal lip/node, notochord, and floor
plate. Mech Dev, 1993;44:91–108.
26. Sasaki, H. and Hogan, B.L.M. Differential expression of multiple fork head
related genes during gastrulation and pattern formation in the mouse embryo.
Development, 1993;118:47–59.
27. Arceci, R. et  al. Mouse GATA‐4: a retinoic acid‐inducible GATA‐binding
transcription factor expressed in endodermally derived tissues and heart. Mol
Cell Biol, 1993;13:2235–46.
28. Gao, X. et al. Distinct functions are implicated for the GATA‐4, ‐5, and ‐6
transcription factors in the regulation of intestine epithelia cell differentia-
tion. Mol Cell Biol, 1998;18:2901–11.
29. Laverriere, A.C. et al. GATA‐4/5/6, a subfamily of three transcription factors
transcribed in developing heart and gut. J Biol Chem, 1994;269:23177–84.
30. Suzuki, E. et al. The human GATA‐6 gene: structure, chromosomal location,
and regulation of expression by tissue‐specific and mitogen‐responsive sig-
nals. Genomics, 1996;38:283–90.
31. Bossard, P. and Zaret, K.S. GATA transcription factors as potentiators of gut
endoderm differentiation. Development, 1998;125:4909–17.
32. Liu, J.K., DiPersio, C.M., and Zaret, K.S. Extracellular signals that regulate
liver transcription factors during hepatic differentiation in vitro. Mol Cell
Biol, 1991;11(2):773–84.
33. Cascio, S. and Zaret, K.S. Hepatocyte differentiation initiates during endo-
dermal‐mesenchymal interactions prior to liver formation. Development,
1991;113:217–25.
34. Mueller, P.R. and Wold, B. In vivo footprinting of a muscle specific enhancer
by ligation mediated PCR. Science, 1989;246:780–6.
35. Lee, C.S. et  al. The initiation of liver development is dependent on Foxa
transcription factors. Nature, 2005;435(7044):944–7.
36. Holtzinger, A. and Evans, T. Gata4 regulates the formation of multiple
organs. Development, 2005;132(17):4005–14.
37. Fisher, J.B. et al. GATA6 is essential for endoderm formation from human
pluripotent stem cells. Biol Open, 2017;6(7):1084–95.
38. Tiyaboonchai, A. et al. GATA6 plays an important role in the induction of
human definitive endoderm, development of the pancreas, and functionality
of pancreatic beta cells. Stem Cell Rep, 2017;8(3):589–604.
39. Shi, Z.D. et al. Genome editing in hPSCs reveals GATA6 haploinsufficiency
and a genetic interaction with GATA4 in human pancreatic development.
Cell Stem Cell, 2017;20(5):675–688.e6.
40. McKeehan, W.L., Wang, F., and Kan, M. The heparan sulfate‐fibroblast
growth factor family: diversity of structure and function. Prog Nucl Acid Res
Mol Biol, 1998;59:135–76.
41. Szebenyi, G. and Fallon, J.F. Fibroblast growth factors as multifunctional
signaling factors. Int Rev Cytol, 1999;185:45–106.
42. Jung, J. et al. Initiation of mammalian liver development from endoderm by
fibroblasts growth factors. Science, 1999;284:1998–2003.
43. Wang, J.‐K., Gao, G., and Goldfarb, M. Fibroblast growth factor receptors have
different signaling and mitogenic potentials. Mol Cell Biol, 1994;14:181–8.
44. Calmont, A. et  al. An FGF response pathway that mediates hepatic gene
induction in embryonic endoderm cells. Dev Cell, 2006;11(3):339–48.
45. Twaroski, K. et al. FGF2 mediates hepatic progenitor cell formation during
human pluripotent stem cell differentiation by inducing the WNT antagonist
NKD1. Genes Dev, 2015;29(23):2463–74.
46. Dudley, A.M., Rougeulle, C., and Winston, F. The Spt components of SAGA
facilitate TBP binding to a promoter at a post‐activator‐binding step in vivo.
Genes Dev, 1999;13(22):2940–5.
47. Lyons, K.M., Pelton, R.W., and Hogan, B.L. Patterns of expression of
murine Vgr‐1 and BMP‐2a RNA suggest that transforming growth factor‐
beta‐like genes coordinately regulate aspects of embryonic development.
Genes Dev, 1989;3(11):1657–68.
48. Solloway, M.J. and Robertson, E.J. Early embryonic lethality in Bmp5;Bmp7
double mutant mice suggests functional redundancy within the 60A sub-
group. Development, 1999;126(8):1753–68.
49. Winnier, G. et al. Bone morphongenetic protein‐4 is required for mesoderm
formation and patterning in the mouse. Genes Dev, 1995;9:2105–16.
50. Zhang, H. and Bradley, A. Mice deficient for BMP2 are nonviable and
have  defects in amnion/chorion and cardiac development. Development,
1996;122(10):2977–86.
51. Mishina, Y. et al. Bmpr encodes a type I bone morphogenetic protein recep-
tor that is essential for gastrulation during mouse embryogenesis. Genes
Dev, 1995;9(24):3027–37.
52. Roelen, B.A. et al. Differential expression of BMP receptors in early mouse
development. Int J Dev Biol, 1997;41(4):541–9.
53. Rossi, J.M. et  al. Distinct mesodermal signals, including BMP’s from the
septum transversum mesenchyme, are required in combination for hepa-
togenesis from the endoderm. Genes Dev, 2001;15:1998–2009.
54. Shin, D. et al. Bmp and Fgf signaling are essential for liver specification in
zebrafish. Development, 2007;134(11):2041–50.
55. Deutsch, G. et al. A bipotential precursor population for pancreas and liver
within the embryonic endoderm. Development, 2001;128:871–81.
56. Wandzioch, E. and Zaret, K.S. Dynamic signaling network for the specifica-
tion of embryonic pancreas and liver progenitors. Science, 2009. 324(5935),
1707–10.
57. Xu, C.R. et al. Chromatin “prepattern” and histone modifiers in a fate choice
for liver and pancreas. Science, 2011;332(6032):963–6.
58. Ober, E.A. et  al. Mesodermal Wnt2b signalling positively regulates liver
specification. Nature, 2006;442(7103):688–91.
59. Lakshminarayanan, B. and Davenport, M. Biliary atresia: A comprehensive
review. J Autoimmun, 2016;73:1–9.
60. Cayuso, J. et  al. EphrinB1/EphB3b coordinate bidirectional epithelial‐­
mesenchymal interactions controlling liver morphogenesis and laterality.
Dev Cell, 2016;39(3):316–28.
61. Sosa‐Pineda, B., Wigle, J.T., and Oliver, G. Hepatocyte migration during
liver development requires Prox1. Nat Genet, 2000;25(3):254–5.
62. Matsumoto, K. et  al. Liver organogenesis promoted by endothelial cells
prior to vascular function. Science, 2001;294:559–63.
63. Lammert, E., Cleaver, O., and Melton, D. Induction of pancreatic differentia-
tion by signals from blood vessels. Science, 2001;294(5542):564–7.
64. Nikolova, G. et  al. The vascular basement membrane: a niche for insulin
gene expression and beta cell proliferation. Dev Cell, 2006;10(3):397–405.
65. Suzuki, A. et  al. Tbx3 controls the fate of hepatic progenitor cells in liver
­development by suppressing p19ARF expression. Development, 2008;135(9):
1589–95.
66. Ludtke, T.H. et  al. Tbx3 promotes liver bud expansion during mouse
­development by suppression of cholangiocyte differentiation. Hepatology,
2009;49(3):969–78.
67. Germain, L., Blouin, M.J., and Marceau, N. Biliary epithelial and hepato-
cytic cell lineage relationships in embryonic rat liver as determined by the
differential expression of cytokeratins, α‐fetoprotein, albumin, and cell
­surface‐exposed components. Cancer Res, 1988;48(17):4909–18.
68. Shiojiri, N. Enzymo‐ and immunocytochemical analyses of the differentia-
tion of liver cells in the prenatal mouse. J Embryol Exp Morphol,
1981;62:139–52.
69. Shiojiri, N. Analysis of differentiation of hepatocytes and bile duct cells in
developing mouse liver by albumin immunofluorescence. Dev Growth
Differ, 1984;26:555–61.
70. Shiojiri, N. Development and differentiation of bile ducts in the mammalian
liver. Microsc Res Technol, 1997;39(4):328–35.
71. Antoniou, A. et al. Intrahepatic bile ducts develop according to a new mode
of tubulogenesis regulated by the transcription factor SOX9.
Gastroenterology, 2009;136(7):2325–33.
72. Gerard, C., Tys, J., and Lemaigre, F.P. Gene regulatory networks in differen-
tiation and direct reprogramming of hepatic cells. Semin Cell Dev Biol,
2017;66:43–50.
73. Clotman, F. et al. Control of liver cell fate decision by a gradient of TGF beta
signaling modulated by Onecut transcription factors. Genes Dev, 2005;19(16):
1849–54.
74. Takayama, K. et al. CCAAT/enhancer binding protein‐mediated regulation
of TGFbeta receptor 2 expression determines the hepatoblast fate decision.
Development, 2014;141(1):91–100.
75. Lee, D.H. et al. LATS‐YAP/TAZ controls lineage specification by regulating
TGFbeta signaling and Hnf4alpha expression during liver development. Nat
Commun, 2016;7:11961.
76. Oda, T. et  al. Mutations in the human Jagged1 gene are responsible for
Alagille syndrome. Nat Genet, 1997;16(3):235–42.
77. Kodama, Y. et al. The role of notch signaling in the development of intrahe-
patic bile ducts. Gastroenterology, 2004;127(6):1775–86.
78. Lorent, K. et  al. Inhibition of Jagged‐mediated Notch signaling disrupts
zebrafish biliary development and generates multi‐organ defects compatible
with an Alagille syndrome phenocopy. Development, 2004;131(22):5753–66.
79. Lozier, J., McCright, B., and Gridley, T. Notch signaling regulates bile duct
morphogenesis in mice. PLoS ONE, 2008;3(3):e1851.
22 THE LIVER:  REFERENCES
80. Tanimizu, N. and Miyajima, A. Notch signaling controls hepatoblast
­differentiation by altering the expression of liver‐enriched transcription
­factors. J Cell Sci, 2004;117(15):3165–74.
81. Kaylan, K.B. et al. Combinatorial microenvironmental regulation of liver
progenitor differentiation by Notch ligands, TGFbeta, and extracellular
matrix. Sci Rep, 2016;6:23490.
82. Schaub, J.R. et al. De novo formation of the biliary system by TGFbeta‐
mediated hepatocyte transdifferentiation. Nature, 2018;557(7704):247–51.
82. A. Spence, J.R. et al. Sox17 regulates organ lineage segregation of ventral
foregut progenitor cells. Dev Cell, 2009;17(1):62–74.
83. Wandzioch, E. et al. Lhx2−/− mice develop liver fibrosis. Proc Natl Acad
Sci U S A, 2004;101(47):16549–54.
84. Delgado, I. et  al. GATA4 loss in the septum transversum mesenchyme
­promotes liver fibrosis in mice. Hepatology, 2014;59(6):2358–70.
85. Medlock, E.S. and Haar, J.L. The liver hemopoietic environment: I.
Developing hepatocytes and their role in fetal hemopoiesis. Anat Rec,
1983;207(1):31–41.
86. Johnson, G.R. and Moore, M.A. Role of stem cell migration in initiation of
mouse foetal liver haemopoiesis. Nature, 1975;258:726–8.
87. Medvinsky, A. and Dzierzak, E. Definitive hematopoiesis is autonomously
initiated by the AGM region. Cell, 1996;86(6):897–906.
88. Baribault, H. et  al. Mid‐gestational lethality in mice lacking keratin 8.
Genes Dev, 1993;7:1191–202.
89. Günes, C. et al. Embryonic lethality and liver degeneration in mice lacking the
metal‐responsive transcriptional activator MTF‐1. EMBO J, 1998;17:2846–54.
90. Gaasbeek Janzen, J.W. et al. Gene expression in derivatives of embryonic
foregut during prenatal development of the rat. J Histochem Cytochem,
1988;36(10):1223–30.
91. Kamiya, A. et al. Fetal liver development requires a paracrine action of oncos-
tatin M through the gp130 signal transducer. Embo J, 1999;18(8):2127–36.
92. Mucenski, M.L. et  al. A functional c‐myb gene is required for normal
murine fetal hepatic hematopoiesis. Cell, 1991;65(4):677–89.
93. Jacks, T. et  al. Effects of an Rb mutation in the mouse. Nature,
1992;359:295–300.
94. Lee, E.Y. et al. Mice deficient for Rb are nonviable and show defects in
neurogenesis and haematopoiesis. Nature, 1992;359(6393):288–94.
95. Hirsch, E. et al. Impaired migration but not differentiation of haematopoietic
stem cells in the absence of beta1 integrins. Nature, 1996;380(6570):171–5.
96. Houssaint, E., Differentiation of the mouse hepatic primordium. I. An
­analysis of tissue interactions in hepatocyte differentiation. Cell Differ,
1980;9:269–79.
97. Giroux, S. and Charron, J. Defective development of the embryonic liver in
N‐myc‐deficient mice. Dev Biol, 1998;195:16–28.
98. Motoyama, J. et  al. Organogenesis of the liver, thymus and spleen is
affected in jumonji mutant mice. Mech Dev, 1997;66:27–37.
99. Sawai, S. et al. Defects of embryonic organogenesis resulting from targeted
disruption of the N‐myc gene in the mouse. Development, 1993;117:1445–55.
100. Onitsuka, I., Tanaka, M., and Miyajima, A. Characterization and functional
analyses of hepatic mesothelial cells in mouse liver development.
Gastroenterology, 2010;138(4):1525–35, 1535.e1–6.
101. Ijpenberg, A. et al. Wt1 and retinoic acid signaling are essential for stellate
cell development and liver morphogenesis. Dev Biol, 2007;312(1):157–70.
102. Eferl, R. et al. Functions of c‐Jun in liver and heart development. J Cell
Biol, 1999;145(5):1049–61.
103. Hilberg, F. et al. c‐Jun is essential for normal mouse development and hepa-
togenesis. Nature, 1993;365:179–81.
104. Li, Q. et  al. Severe liver degeneration in mice lacking the 1κB kinase 2
gene. Science, 1999;284:321–5.
105. Beg, A.A. et al. Embryonic lethality and liver degeneration in mice lacking
the RelA component of NF‐KB. Nature, 1995;376:167–70.
106. Reimold, A.M. et al. An essential role in liver development for transcription
factor XBP‐1. Genes Dev, 2000;14(2):152–7.
107. Stanger, B.Z., Tanaka, A.J., and Melton, D.A. Organ size is limited by the
number of embryonic progenitor cells in the pancreas but not the liver.
Nature, 2007;445(7130):886–91.
108. Lee, C.S. et al. Regeneration of pancreatic islets after partial pancreatec-
tomy in mice does not involve the reactivation of neurogenin‐3. Diabetes,
2006;55(2):269–72.
109. Menge, B.A. et  al. Partial pancreatectomy in adult humans does not
­provoke beta‐cell regeneration. Diabetes, 2008;57(1):142–9.
110. Kyrmizi, I. et  al. Plasticity and expanding complexity of the hepatic
transcription factor network during liver development. Genes Dev,
­
2006;20(16):2293–305.
111. Li, J., Ning, G., and Duncan, S.A. Mammalian hepatocyte differentiation
requires the transcription factor HNF‐4alpha. Genes Dev, 2000;14(4):
464–74.
112. DeLaForest, A. et  al. HNF4A is essential for specification of hepatic
progenitors from human pluripotent stem cells. Development,
­
2011;138(19):4143–53.
113. Odom, D.T. et al. Control of pancreas and liver gene expression by HNF
transcription factors. Science, 2004;303(5662):1378–81.
114. Hoffman, B.G. et  al. Locus co‐occupancy, nucleosome positioning, and
H3K4me1 regulate the functionality of FOXA2‐, HNF4A‐, and PDX1‐
bound loci in islets and liver. Genome Res, 2010;20(8):1037–51.
115. Battle, M.A. et al. Hepatocyte nuclear factor 4alpha orchestrates expression
of cell adhesion proteins during the epithelial transformation of the devel-
oping liver. Proc Natl Acad Sci U S A, 2006;103(22):8419–24.
116. Parviz, F. et al. Hepatocyte nuclear factor 4alpha controls the development
of a hepatic epithelium and liver morphogenesis. Nat Genet, 2003;34(3):
292–6.
117. Tronche, F. et al. Hepatocyte nuclear factor 1(HNF1) and liver gene expres-
sion, in Liver Gene Expression (eds. F. Tronche and M. Yaniv), R.G. Landes
Company, 1994, pp. 155–82.
118. Barbacci, E. et al. Variant hepatocyte nuclear factor 1 is required for vis-
ceral endoderm specification. Development, 1999;126(21):4795–805.
119. Cereghini, S. et al. Expression patterns of vHNF1 and HNF1 homeopro-
teins in early postimplantation embryos suggest distinct and sequential
developmental roles. Development, 1992;116:783–97.
120. Gouon‐Evans, V. et  al. BMP‐4 is required for hepatic specification of
mouse embryonic stem cell‐derived definitive endoderm. Nat Biotechnol,
2006;24(11):1402–11.
121. Loh, K.M. et  al. Efficient endoderm induction from human pluripotent
stem cells by logically directing signals controlling lineage bifurcations.
Cell Stem Cell, 2014;14(2):237–52.
122. Takebe, T. et  al. Vascularized and functional human liver from an iPSC‐
derived organ bud transplant. Nature, 2013;499(7459):481–4.
123. Goldman, O. and Gouon‐Evans, V. Human pluripotent stem cells: myths and
future realities for liver cell therapy. Cell Stem Cell, 2016;18(6):703–6.
124. Wells, J.M. and Melton, D.A. Vertebrate endoderm development. Annu Rev
Cell Dev Biol, 1999;15:393–410.
125. Vallier, L. Heps with pep: direct reprogramming into human hepatocytes.
Cell Stem Cell, 2014;14(3):267–9.
126. Yiangou, L. et  al. Human pluripotent stem cell‐derived endoderm for
modeling development and clinical applications. Cell Stem Cell,
­
2018;22(4):485–99.
127. Bertero, A. et al. Activin/nodal signaling and NANOG orchestrate human
embryonic stem cell fate decisions by controlling the H3K4me3 chromatin
mark. Genes Dev, 2015;29(7):702–17.
128. Jing, R., Duncan, C.B., and Duncan, S.A. A small‐molecule screen reveals
that HSP90beta promotes the conversion of induced pluripotent stem cell‐
derived endoderm to a hepatic fate and regulates HNF4A turnover.
Development, 2017;144(10):1764–74.
129. Camp, J.G. et al. Multilineage communication regulates human liver bud
development from pluripotency. Nature, 2017;546(7659):533–8.
PART TWO:
THE CELLS
SECTION A:
CELL BIOLOGY
OF THE LIVER
 Cytoskeletal Motors: Structure
3 and Function in Hepatocytes
Mukesh Kumar1, Arnab Gupta2, and Roop Mallik1
1
Department of Biological Sciences, Tata Institute of Fundamental Research, Navy Nagar, Colaba, Mumbai, India
2
Department of Biological Sciences, Indian Institute of Science Education and Research Kolkata, Mohanpur, India
INTRODUCTION
The hepatocyte is the major epithelial cell of the liver, making
up to 70–80% of liver mass. Hepatocytes produce and secrete
bile, extract specific molecules from blood, and secrete oth­
ers into blood flow. This “sieving” function makes the liver a
key homeostatic organ and requires intracellular transport of
a vast variety of lipids/proteins towards multiple distinctly
polarized surfaces in hepatocytes. In order to do this, hepatocytes
must partition as well as connect two different environments –
the blood and the bile. To serve this function in the liver,
hepatocytes are arranged in cords. An individual cell is
polygonal and faces at least two blood sinusoids (the basal
domain). A branched network of grooves between adjacent
cells forms the bile canaliculus, representing the apical
domain. Such a polygonal shape means that hepatocytes do
not have a single basolateral‐to‐apical axis, and the transcy­
totic pathways between basolateral and apical membranes are
more complex than other single apical–basolateral axis polar­
ized epithelial cells, such as enterocytes, renal epithelial
cells, etc. [1]. Most newly synthesized membrane and secre­
tory proteins exiting the trans‐Golgi network (TGN) are
directed via the basolateral membrane, with less TGN‐exit
traffic directed towards the apical side (Figure  3.1a).
Similarly, endocytosed cargos including plasma membrane
(PM) fragments internalized at the basolateral (blood) side
are transported to lysosomes, the apical PM, or bile via tran­
scytotic pathways. Since these multiple and distinct sorting
steps of the cargos in the biosynthetic and endocytic path­
ways rely on microtubules, microfilaments (actin), and their
respective motor proteins, the cytoskeletal system functions
as the primary coordinator between the apical and basolateral
regions in hepatocytes [2].
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
HEPATOCYTE POLARITY
AND THE ACTIN NETWORK
Although hepatocyte polarity is well‐described at an anatomical
level, little is known about the molecules and the mechanisms
that initiate, establish, and maintain this polarization. During
mouse development, a liver bud composed of nonpolarized hepa­
toblasts is detected as early as embryonic day 9.5. Studies on
mouse liver development by Feracci et al. have demonstrated that
initiation of hepatocyte polarization occurs very early in develop­
ment (day 15). From embryonic day 17 onwards, hepatoblasts
aggregate together resembling acini. The acini exhibit a simple
polar phenotype, with their apical surfaces facing a central lumen [3].
Gradually, into the postnatal stages, the simple polarity changes
to polygonal hepatic polarity, a phenomenon adeptly emulated in
WIF‐B cells. This entire process of ­hepatocyte as well as WIF‐B
polarization is heavily dependent on the spatiotemporal location
and function of the cytoskeletal ­proteins. The anatomy of micro­
filament in the polarized hepatocyte was studied in great detail by
Ishii et  al. [4]. WIF‐B cells maintain PM proteins that are
restricted to either the apical (bile  canalicular) or basolateral
domain, with tight junctions demarcating the apical–basolateral
boundaries. Phalloidin staining reveals a thin but highly intense
ringed network of actin f­ ilaments around the bile canaliculus/api­
cal membrane (BC), along with a sparse network close to the
basolateral membrane (Figure 3.1a). Densely concentrated actin
filaments were identified around the bile canaliculi in the form of
microvillous core filaments and pericanalicular web filaments.
The microvillous core filaments exhibited growth at their apical
ends. In contrast, filaments of the pericanalicular web, running
parallel to the cell surface, showed no fixed polarities.
Interestingly, adjacent ­ filament pairs often showed opposite
polarities, an alignment necessary for filament sliding. A group
28 THE LIVER:  HEPATOCYTE POLARITY AND THE ACTIN NETWORK

(a) (b)
Bile
canaliculus
Cargo binding domain Myosin V

SAC

?
Golgi
Coiled coil domain
?

? Light chain binding domain

+ ? Nucleus

Basola + Motor domain

te
endos ral
ome
MyoVb with ATP7B as cargo
Cortical actin

Figure 3.1  Actin‐dependent trafficking in hepatocytes and the myosin‐V motor. (a) Most of the apically targeted proteins (copper transporting
ATPase, ATP7B in this case) travels to the destination via basolateral endosomes. The microtubule‐based motor proteins that drive the cargo from
the TGN to the basolateral endosomes and subsequently to the subapical compartment (SAC) has not been determined (shown as interrogative
clause). Once the cargo‐carrying vesicles reach the SAC, myosin Vb drives its recycling between the SAC and the bile canaliculus/apical mem­
brane. The myosin Vb remains anchored to the actin microfilaments that form a dense network beneath the apical membrane. (b) The general
structure of myosin V that is conserved in all the myosin V proteins (MyoVa, Vb, and Vc).

of sporadic actin f­ ilaments with no uniform polarity acts as a link
between the canalicular membrane and coated vesicles [4]. Such
spatial organization of actin filaments is implicated in mainte­
nance of microvilli length, contraction of the canalicular walls
facilitating bile secretion, and transport of coated vesicles in the
apical membrane (bile canalicular membrane). Treatment of
hepatocytes with cytochalasin D, which targets actin, completely
inhibited self‐assembly of rat hepatocytes into spheroids. Thus,
hepatocytes require an intact actin network to self‐assemble effi­
ciently into functional tissue‐like structures [5].
Of all the actin‐related motors, most work has been done on
myosin V (Figure  3.1b). Myosin Vb is most abundant in the
liver as well as the WIF‐B cells [6]. It is an unconventional myo­
sin regulating vesicle trafficking by binding Rab11a [7, 8].
Myosin Vb inhibition induces failure to attain complete polarity
as well as trafficking defects of various apically targeted cargos,
marked by accumulation of metabolites leading to a diseased
state. Understanding human genetic diseases linked to muta­
tions in myosin Vb and its effector proteins has been greatly
helpful in comprehending liver polarity and polarized protein
trafficking. Mutations in myosin Vb were found to cause micro­
villus inclusion disease (MVID), which is characterized by a
lack of apical microvilli and intracellular structures containing
microvilli in the intestine, leading to a severe form of congenital
diarrhea [9]. Years before detailed mechanistic understanding of
MVID phenotypes, Wakabayashi et  al., while investigating
mechanisms of apical targeting, serendipitously discovered that
knocking down of Rab11 before its polarization prevented cana­
licular formation in WIF‐B9 cells. Polarization could also be
arrested by overexpression of the Rab11a‐GDP locked form or
the myosin Vb tail dominant negative (DN) construct. In WIF‐
B9 cells, which lack bile canaliculi, apical ABC (ATP‐binding
cassette) transporters colocalize with transcytotic membrane
proteins in Rab11a‐containing endosomes and did not distribute
to the PM [10]. The study demonstrated that polarization of
hepatocytes requires recruitment of Rab11a and myosin Vb to
intracellular membranes that contain apical ABC transporters,
permitting their targeting to the PM.
Stable MYO5 knockdown (MY05B‐KD) in CaCo‐2 cells
induced loss of microvilli, alterations in tight junction proteins,
and disruption of polarized trafficking [9, 11, 12]. Due to short
life expectancy, liver phenotypes usually go undetected in
MVID patients [13]. However, MVID patients developed a
cholestatic liver disease similar to progressive familial intrahe­
patic cholestasis (PFIC) and benign recurrent cholestasis (BRC)
[14]. PFIC and BRC have been linked to mutations in ABCB11
(encoding bile salt export pump, BSEP) and ATP8B1 (encoding
FIC1 protein) genes, respectively [15, 16]. Interestingly, in
neither of those two genes were disease‐causing mutations
­
detected in any of these patients. Immunohistochemistry stud­
ies ­suggested mislocalization of BSEP in cytoplasmic vesicular
staining as compared to controls, where the protein primarily
localized at the apical membrane. This phenotype is a conse­
quence of impairment of the MY05B/RAB11A apical recycling
endosome pathway in hepatocytes [14]. Utilizing polarized
WIF‐B cells, it was demonstrated that the copper‐exporting
ATPase ATP7B (also known as the Wilson disease protein) is a
major cargo of myosin Vb in the liver. Mutations in ATP7B
imparts a nontrafficking phenotype to the protein, causing a
­disorder of liver copper accumulation termed Wilson disease [6].
Interestingly, a similar copper accumulation phenotype was
demonstrated in polarized WIF‐B cells when myosin Vb was
acutely knocked down using a DN mutant. Detailed micro­
scopic studies revealed that in the absence of functional myosin
Vb, ATP7B‐containing vesicles accumulated beneath the
­apical membrane and failed to fuse with it. This study further
 3:  Cytoskeletal Motors: Structure and Function in Hepatocytes 29

delineated the physical and functional localizations of micro­
tubule and cortical actin with respect to apical trafficking.
Apically targeted cargos exit the TGN and traffic basolaterally
before getting redirected apically in vesicles utilizing the micro­
tubular network [17]. The vesicles are then transferred to the
myosin Vb motor anchored at the cortical actin that is located
beneath the apical membrane (Figure  3.1a). This transfer of
­cargos occurs in specialized endosomal compartments called
subapical compartments (SACs) lying in close proximity to the
microtubule organizing center (MTOC). Studies on the role of
myosin Vb in trafficking of apically targeted proteins reveals a
fascinating spatiotemporal correlation with the polarization
­status of WIF‐B cells. In completely polarized WIF‐B cells,
myosin Vb is primarily localized as a ring‐like pattern around
the F‐actin ring at the bile canaliculus (Figure 3.1a). However,
in pre‐polarized cells, myosin Vb, as well as the myosin Vb tail
DN protein, localize to a tight juxtanuclear intracellular site
described as the “apical compartment” [18]. This apical compart­
ment is an intracellular membranous structure clustered near
the minus ends of microtubules that are labeled with γ‐tubulin,
where apically destined cargos accumulate selectively.
MICROTUBULES: HISTORY, STRUCTURE
AND DYNAMICS
Microtubules (MTs) are tracks for transport of intracellular cargo
by microtubule‐dependent motors such as dynein and kinesin
(Figure  3.2a). MTs were observed as rod‐like structures in
­spermatozoid of Sphagnum and ciliated epithelia [19, 20], and
were hypothesized to power the beating movements of cilia.
Improvement in structural preservation techniques [21] showed
that MTs contain multiple (for example 13) subunits arranged in a
circular fashion [22]. Ledbetter and Porter named the structure
“microtubule,” because the protofilaments were organized around
a hollow center and therefore appeared tubular. Taylor and cow­
orkers characterized a single [3H]‐colchicine binding protein with
a sedimentation coefficient of 6S and an Mr of 110 000–120 000
[23]. Later, it was shown that the denaturation of the 6S protein
produced α and β monomeric subunits (Mr = 55 000). These two
very similar subunits were found to exist within MTs in a 1 : 1
molar ratio, forming αβ heterodimers [24]. The term “tubulin”
was coined by Mohri in 1968 [25]. The ability of tubulin purified
from brain homogenate to polymerize in vitro in the presence of
GTP, Mg2+, and EGTA greatly improved our understanding of MT
assembly and dynamics [26]. A third type of tubulin, γ‐tubulin,
was discovered and was found to be associated with the minus end
of MTs near the centrosome. Although γ‐tubulin is not a major
component of MTs, it plays an i­ mportant role in the assembly and
expression of MT arrays in eukaryotic cells [27].
MTs undergo rapid stochastic growth and shrinkage
(Figure 3.2b), known as “dynamic instability” [28, 29]. This phe­
nomenon depends on various intrinsic and extrinsic factors,
including the concentration of tubulin. The tubulin subunits asso­
ciate/dissociate at different rates at the plus and minus ends of
protofilaments. The critical concentration is the minimum con­
centration of tubulin dimers above which there is a net growth of
tubulin polymer. Nucleotides at the N‐site are not available for
exchange, unlike nucleotides at the E‐site which can exchange
with the solution [30]. Preferential polymerization of tubulin
dimers at the β‐tubulin end defines the MT polarity. The rapidly
polymerizing β‐end is called the plus end whereas the slow‐
growing end with exposed α‐tubulin is called the minus end.
GTP hydrolysis within the polymer is slower than GTP addition
at the plus end. Resultant accumulation of a “GTP cap” at the
plus end favors dimer addition [28, 30]. As the tubulin dimer
concentration decreases below a critical value, dimer addition at
the plus end is reduced in comparison to GTP hydrolysis. This
leads to a rapid shrinkage of MTs and is termed catastrophe.
Catastrophe can be rescued by the addition of GTP‐tubulin [28].
The competition between polymerization and catastrophe can be
regulated by MT‐associated proteins and MT‐severing proteins
inside cells [31, 32]. Although dynamic, the MTs still collec­
tively maintain a stiff intracellular skeleton to impart shape to a
cell and support intracellular cargo transport.

(a) (c) (d)

Light chain
+ IC
Cargo Tail
domain
LIC

Linker Heavy
Depolymerization chain
Stalk
phase
Polymerization domain
phase Head/motor
ule
tub

domain
cro

β
Mi

– MT binding Motor
domain domain
α Dynein
(b) Kinesin

Figure 3.2  Microtubules and associated motor proteins. (a) Microtubules normally extend from the minus end of the microtubule (anchored at the
microtubule organizing center (MTOC)) to the plus end (fast‐growing). (b) Microtubules consisting of αβ‐tubulin dimers (shown) remain in a
dynamic mode, alternating between polymerization and depolymerization phases. The molecular structures of (c) dynein (retrograde motor) and (d)
conventional kinesin (anterograde motor) are shown. IC, intermediate chain; LIC, light intermediate chain.
30 THE LIVER:  MICROTUBULE ORGANIZATION, POLARIZATION, AND CARGO TRAFFICKING IN HEPATOCYTES
MICROTUBULE MOTORS
MTs also serve as tracks against which molecular motors can
generate mechanical force (Figure 3.2a). Motors utilize ATP as
fuel to generate force and movement for intracellular transport,
cell division etc. Gibbons and Rowe were the first to report MT
motor‐based ATPase activity when they purified the dynein
ATPase from Tetrahymena cilia [33]. The term dynein was
derived from the Greek word dyne, meaning “force,” as this
motor was thought to be responsible for beating of cilia.
Discovery of dynein in the flagella of sea urchin spermatozoa
confirmed that dynein was responsible for ciliary and flagellar
beating [34]. Although the MT motor proteins were suspected to
power mitosis and organelle transport, molecular characteriza­
tion was difficult due to their low cytosolic concentration. The
use of extracts from squid giant axon paved the way for studying
intracellular transport in vitro [35, 36]. This was followed by the
biochemical purification of two organelle‐associated cytoplas­
mic motors, namely dynein and kinesin, that usually transport
cargo in opposite directions along MTs [37, 38].
Cytoplasmic dynein
Dynein (Figure  3.2c) is a multi‐subunit protein complex of
~2000 kDa consisting of two heavy chains (DHC; each ~530
kDa), a variable number of light chains (LCs) and intermedi­
ate chains (ICs). The ICs bind to DHC and multiple LCs bind
to ICs at separate sites [39]. DHC is a single polypeptide and
an unusual member of the hexameric AAA+ superfamily
(ATPase associated with various cellular activities). The two
motor domains of DHC interact with MTs. Each motor
domain contains six AAA+ sub‐domains, two of which
(AAA1 and AAA3) can bind and hydrolyze ATP [40, 41].
Force is generated by a shift in the linker domain that joins the
dynein tail with AAA1 [42]. Intermediate chains are thought
to bind dynein to its many cargos, such as vesicles, Golgi
body, lipid droplets, kinetochores, and mRNA [43]. Recent
studies suggest that different adaptor proteins also mediate
the cargo binding and motor activities of dynein. Jun N‐termi­
nal kinase interacting proteins (JIP), Rab7‐interacting lysoso­
mal proteins RILP/p150Glued and sorting nexins are well‐known
adaptor proteins for dynein [39, 42, 44, 45]. Intriguingly,
although the dynein motor generates a smaller force than
kinesin, it can work in large teams by virtue of an in‐built gear
mechanism and the ability to catch‐bond to MTs against high
opposing load [46, 47]. Taking these findings further, choles­
terol‐enriched lipid rafts on an endosome/phagosome mem­
brane were shown to serve as a platform where many dynein
motors can cluster together. This geometric clustering allowed
a large team of dyneins to simultaneously contact a single
MT, and generate a collective force that rapidly transports the
phagosome towards degradative lysosomes [48, 49]. Because
cytoplasmic dynein transports cargos towards the minus end
of MTs, it is known as a retrograde motor [50]. The orienta­
tion of MTs differs in epithelial cells of different origins.
Therefore, the dynein‐driven motion may localize cargos
towards the apical or the basolateral membrane, depending on
cell type. In hepatocytes, dynein likely drives cargos towards
the perinuclear region [51, 52].
Kinesin
Since the discovery of conventional kinesin in 1985 by Brady and
Vale [36, 53], a large number of ATPases have been categorized
into the kinesin superfamily [53, 54]. Kinesins are classified into
14 classes (kinesins 1–14), all these motors contain an ATP‐bind­
ing domain with significant homology [55]. Kinesins are protein
complexes of approximately 380 kDa, consisting of two heavy
chains (each 120 kDa) and two light chains (each 64 kDa) [56]. As
shown in Figure 3.2d, the N‐terminal globular head domains of
the heavy chains are the “motor domains” that bind to the MT [57]
and also hydrolyze ATP to provide energy for movement [58].
The processive motion of conventional kinesins is explained by
their “hand‐over‐hand” walking [59]. The flexible stalk domains
help kinesin to dimerize, and the neck linker acts as a lever to
facilitate steps [60]. The C‐terminal tail domain, in association
with the light chain, forms the cargo‐binding domain of kinesin
(Figure  3.2d). The kinesin tail domain forms complexes with
adaptor proteins such as Miro/Milton, syntabulin, DENN/MADD,
etc. [55]. These combinations achieve a high level of specificity in
cargo recognition. Spatiotemporal regulation of cargo delivery is
achieved by the phosphorylation of kinesin, Rab GTPase activity
and Ca2+ signaling [55]. Most kinesins have a motor domain at the
N‐terminal region (known as N‐kinesins) that drive cargos
towards the plus end of the microtubule (anterograde motors).
Others kinesins (known as C‐kinesins, e.g. kinesin‐14) have the
motor domain at the C‐terminal, while a few kinesins have a cata­
lytic domain in the middle (M‐kinesin, e.g. kinesin‐13). C‐kinesin
drives minus end‐directed motility of cargos, while M‐kinesin
depolymerizes MTs [55, 61]. Similar to dynein, the net direction
of kinesin‐driven cargo movement depends on the MT orientation
within a particular cell type.
MICROTUBULE ORGANIZATION,
POLARIZATION, AND CARGO
TRAFFICKING IN HEPATOCYTES
Like other epithelial cells, hepatocytes must exchange macro­
molecules between the external environment and their interior
milieu. To achieve this, hepatocytes are polarized with distinct
apical (or bile canalicular) and basolateral (or sinusoidal)
domains segregated by tight junctions. The lateral surfaces of
hepatocytes form cell–cell contacts while the basal surface inter­
acts with the extracellular matrix. Unlike typical epithelial cells,
hepatocytes have a multipolar organization. Each hepatocyte
shares a boundary with multiple narrow lumina, bile canaliculi,
and basal domains of neighboring cells that face the endothelial
lining. A hepatocyte sandwiched in this manner has to sustain
two countercurrent flow systems: the synthesis/secretion of bile
and uptake/secretion of blood components at the apical and
basolateral surfaces, respectively [1]. The two surface domains
exhibit distinct protein and lipid compositions [62], and suste­
nance of this complex polarity likely requires vesicular transport
along MTs and actin inside hepatocytes.
Novikoff et al. imaged the MT organization in cultured ­primary
rat hepatocytes by immunofluorescence labeling of tubulin [63].
A “starburst” pattern, commonly identified as radially organized
 3:  Cytoskeletal Motors: Structure and Function in Hepatocytes 31
Bile
canaliculi
Sinusoid
Nucleus
Figure 3.3  Microtubule organization in hepatocytes. Microtubules extend from pericanalicular region towards the sinusoidal region. Lipid drop­
lets (LDs) are driven by conventional kinesin towards peripherally located smooth ER (sER) near the plus ends of microtubules (shown). Endocytic
vesicles are driven dominantly by dynein and exocytic vesicles by kinesin motors (motors are not shown).
MTs (e.g. in melanocytes), was observed with one end of MTs
focused at a centrally located centrosome, and the other ends
emanating out like an umbrella to line the cortical region of the
hepatocyte. In agreement with this geometry, the microtubule
organizing center (MTOC) in polarized hepatocytes is localized
adjacent to a perinuclear region and MTs extend from canalicu­
lar to sinusoidal region [64]. Cytoplasmic dynein is associated
with, and may drive ligand‐containing endosomes towards the
central centrosomal region of hepatocytes [51, 65]. A simplified
view of the microtubule orientation in relation to the apical and
basolateral membrane in hepatocytes is shown in Figure 3.3. As
a consequence of this geometry, MT‐based motion towards the
apical PM in hepatocytes may require a plus end‐directed motor
(e.g. kinesin‐1). The MT orientation and directionality of motors
is also relevant to lipid trafficking. For example, the smooth
endoplasmic reticulum (sER) (where lipids are packaged for
secretion) appears located towards the ­periphery of hepatocytes
and may be supplied with lipids via kinesin‐driven transport of
lipid droplets [66].
The liver is an organ of immense metabolic importance. It is
involved in the production of bile acid, cholesterol, plasma pro­
tein, assembly/secretion of triglyceride‐laden very low‐density
lipoprotein particles (VLDL), elimination of toxic substances,
and processing of hormones and cytokines [1]. All these pro­
cesses involve directed intracellular transport between distinct
subcellular compartments and membranes of a hepatocyte. The
ER–Golgi network is the central pathway in protein sorting and
endosomal targeting. Nascent proteins arising in the ER are
transported to the trans‐Golgi network (TGN) via the Golgi cis­
ternae. The proteins sequestered at TGN are actively sorted and
packaged into vesicles to be further transported to the apical
membrane or to endosomal compartments [67]. MT motors are
essential for the generation and transport of TGN‐derived
­transport carriers [68, 69]. Disruption of MTs results in the mis­
sorting of apical proteins [70]. A function‐blocking antibody
against kinesin also inhibits exit of an apical reporter protein,
NTRp75, from the TGN and leads to accumulation of small
Golgi
sER
LDs
Kinesin-1
Microtubule
Cortical actin
vesicles around the Golgi [71]. Exposure of hepatocytes to col­
chicine reduces endosome–lysosome fusion, suggesting that
dynein transports endocytic vesicles to the apical surface [72].
Fluorescently labeled early endocytic vesicles isolated from rat
liver are driven along MTs by kinesin‐1 and KIFC2 motors [73].
In contrast to the early endocytic vesicles, dynein and KIF3A
drive motion of late endosomes which do not exhibit significant
fission [74].
MICROTUBULE MOTORS
IN TRIGLYCERIDE SECRETION
The liver can be thought of as the “lipid manager” of the body.
Most of this lipid is stored away in intracellular organelles
called lipid droplets (LDs). LDs are loaded with triglyceride
(TG) molecules that consist of fatty acid chains esterified to
glycerol. Our view that LDs are coalesced fat sitting idle in cells
has changed drastically. Proteomic and imaging studies hinted
that LDs interact with mitochondria, peroxisomes, and ER
depending on the cell/tissue types and metabolic requirements
[75–77]. This interaction of LDs with a variety of cellular com­
partments requires long‐distance transport of the LDs within
cells. Live imaging reveals that LDs move in a directed manner
in many cell types ranging from algae to hepatocytes [66, 78].
Cytoplasmic dynein and kinesin are physically associated with
LDs to drive intracellular motion of LDs [66, 79].
The TG in LDs can be utilized to generate energy in h­ epatocytes,
or can be secreted in the form of VLDL. The bulk of TG (~70%)
in VLDL is derived from cytosolic LDs [80]. The earliest report
suggesting that MTs are involved in lipoprotein secretion from the
liver was published in 1973 [81]. Furthermore, Reaven et al. dem­
onstrated that depolymerization of microtubules drastically
decreased VLDL lipidation in hepatocytes [82]. More recently, we
found that LDs purified from rat liver show rapid plus end‐directed
motion along MTs [83]. We also showed that kinesin‐1‐driven LD
32 THE LIVER:  CONCLUSIONS
transport supplies triglyceride to the smooth ER in hepatocytes for
assembling VLDL, and that knockdown of kinesin‐1 in McA‐
RH7777 cells inhibits the characteristic peripheral localization of
LDs [66, 84]. Intriguingly, kinesin‐1 is recruited to LDs in a com­
plex with ADP ribosylation factor 1 (ARF1), a small GTPase and
a key regulator of lipolysis [66]. Because ARF1 induces ­membrane
curvature, it is possible that the additional force from kinesin
causes vesiculation from ARF1‐rich regions on the LD membrane.
ARF1 and kinesin were activated on LDs in an insulin‐dependent
manner, making this pathway responsive to the metabolic state of
an animal. In the fed state kinesin‐1 was more active on LDs,
transporting them to the peripherally localized sER in hepato­
cytes, where sER‐resident lipases could catabolize the TG to gen­
erate diacylglycerol [85]. Because it has two fatty acid chains,
diacylglycerol may equilibrate across the sER membrane, where­
after it is reconverted into TG in the sER lumen and supplied to
synthesize mature VLDL particles [86].
The catabolism of LD‐TG for VLDL lipidation is extremely
efficient [80] and may require a synergy between kinesin, ARF1,
and specific LD/sER‐associated lipids in the fed state. In con­
trast, in the fasted state, lowered insulin‐signaling reduces kine­
sin‐1 on LDs to limit sER–LD contact and tempers TG supply
for VLDL lipidation [66]. Tuning kinesin‐dependent transport
of LDs in hepatocytes may therefore enable the liver to protec­
tively sequester massive amounts of TG after fasting (fasting‐
induced steatosis), thus preventing lipotoxic effects of TG on
peripheral tissues. The cellular mechanisms that control TG
secretion from the liver, and therefore systemic TG homeostasis
across daily metabolic cycles, can help understand TG imbal­
ance leading to lipodystrophies such as fatty liver and diabetes
[87]. Indeed, we have seen significant changes in lipid/protein
trafficking to LDs across feeding–fasting cycles [66, 88]. It is
surprising that relatively little work has been done employing
animal models and the simple feeding–fasting response. We
also hope that disease‐relevant mutations in animal models can
answer how motor proteins transport LDs within hepatocytes to
control trafficking in response to metabolic states.
MT‐dependent transport is also essential for hepatitis C virus
(HCV) replication in host hepatocytes. HCV replicates its
genome at the ER‐derived membranous web, and assembles new
infectious particles using ER resident structural proteins [89].
Newly synthesized viral proteins (e.g. HCV‐core) must transfer
from ER to LDs for HCV to propagate. This transfer and HCV
assembly was diminished upon kinesin‐1 knockdown [66].
HCV‐core induces LD redistribution in an MT‐ and dynein‐
dependent manner [90]. HCV also cleaves the Rab‐interacting
lysosomal protein to redirect Rab7‐containing vesicles from
dynein‐ to kinesin‐dependent transport, and in this manner pro­
motes virion secretion [91]. MT‐dependent transport, therefore,
impacts multiple aspects of the viral lifecycle in hepatocytes, and
is a rich area for further scientific investigations.
MICROTUBULE MODIFICATIONS
AND LIVER PATHOLOGY
MTs and their associated functions are tuned by the post‐trans­
lational modification (PTM) of tubulin. PTMs and various
microtubule‐interacting proteins contribute to the heterogeneity
of MTs within cells. The most common and evolutionarily con­
served modification on tubulin is acetylation. The ε‐amino
group of lysine‐40 (K40) residue of α‐tubulin is predominantly
acetylated with αTAT1 (α‐tubulin acetyltransferase‐1) [92, 93]
and reversed by deacetylases: Sirt2 (sirtuin type 2) [94] or his­
tone deacetylase (HDAC6) [95]. The acetylation‐induced weak
lateral interactions of protofilaments increase MT flexibility and
make MTs resistant to mild cold, nocodazole, and colchicine
exposure [96, 97]. Besides acetylation, the K40 site is also sus­
ceptible to trimethylation [98]. In mitotic cells, α‐tubulin of the
central spindle is trimethylated at K40, and enriched at the plus
ends of MTs at metaphase. Despite numerous reports on the
phosphorylation of α‐ and β‐tubulin at serine, threonine, and
tyrosine residues, the physiological significance of these modi­
fications remains to be understood. The unstructured C‐terminal
tail (CTT) domains of α‐ and β‐tubulin are frequent targets of
tyrosination, Δ2 modification (penultimate glutamate is also
removed from detyrosinated tubulin) and detyrosination.
α‐Tubulin undergoes tyrosination–detyrosination cycles via the
activity of tubulin tyrosinase ligase and carboxypeptidase. An
in vitro study suggests that PTM of tubulin regulates the MT
depolymerization rate. Moreover, modifications in CTT of tubu­
lin are recognized by molecular motors that also govern the
motor velocity and its processivity [99]. In vivo, detyrosinated
MTs persist for hours, but tyrosinated MTs turn over in minutes
[100]. The KIF2C kinesin that depolymerizes MTs has prefer­
ential activity for tyrosinated MTs both in vivo and in vitro. The
binding of KIF5 kinesin to MTs is also regulated by multiple
PTMs. The presence of tyrosinated tubulin also favors interac­
tion of the MTs with plus end‐tracking proteins (+TIPs) that
have conserved CAP‐Gly domains [101]. Some tubulin modifi­
cation, such as glutamylation, succination, and O‐Glc‐N‐acylation
have been recently identified, but their detailed physiological
impact is as yet unexplored.
Alcoholic liver may develop as a result of abnormal PTMs in
hepatocytes. Acetylation of MTs in hepatocytes increases tubu­
lin stability [102]. Treatment of hepatoma‐derived WIF‐B cells
with a drug that specifically depolymerizes dynamic MTs
impairs MT‐dependent trafficking along three distinct cellular
pathways [103, 104]. Tubulin is hyperacetylated in ethanol‐fed
rats, which increases MT stability and might also alter intracel­
lular motility [105]. Acetylated MTs are required for adipogen­
esis, and this acetylation is controlled by AMPK‐mediated
phosphorylation of αTAT1 [106]. AMPK, a master regulator of
lipid homeostasis, also promotes relocation of LDs and mito­
chondria on detyrosinated MTs in VERO cells of nonhepatic
origin to promote fatty acid oxidation [107]. Despite its poten­
tial physiological consequences, the effect of alcohol‐induced
acetylation on LD dynamics in hepatocytes is not known.
CONCLUSIONS
Motor‐dependent motion along actin and MTs sustains the
intracellular organization of many cell types, and is particularly
relevant to polarized epithelial cells such as hepatocytes.
In  hepatocytes, the MTs are organized in a radial fashion
­originating from the centrosome and spreading out towards the
peripheral sinusoidal regions [108, 109]. This arrangement of
 3:  Cytoskeletal Motors: Structure and Function in Hepatocytes 33
MTs, along with the directionality of cytoplasmic dynein and
kinesin motors, defines endo/exocytosis and protein trafficking
in hepatocytes [63]. As an example, kinesin‐dependent lipid
droplet transport maintains lipid homeostasis across metabolic
transitions. Viruses (such as HCV) may also require MT‐
dependent transport for assembly in hepatocytes and secretion
to the blood plasma in complex with VLDL particles [90, 110].
PTM of tubulin affects the dynamics and organization of MTs,
and also motor‐driven transport on MTs [99]. PTMs regulate
intracellular cargo transport and organelle–organelle contacts,
and aberrant PTMs might hinder substrate delivery leading to
physiological anomalies such as fatty liver disease. In conclu­
sion, the MT cytoskeleton in hepatocytes provides a rich and
very disease‐relevant context for further investigations. Of par­
ticular interest in this context is the recent finding that MT
motors control lipid flux from the liver across metabolic cycles
[66]. How cargos are handed over between the actin and MT
systems in hepatocytes is also poorly understood and warrants
investigation.
REFERENCES
1. Treyer, A. and Müsch, A. Hepatocyte polarity. Compr Physiol,
2013;3(1):243–87.
2. Musch, A. Microtubule organization and function in epithelial cells. Traffic,
2004;5:1–9.
3. Feracci, H., Connolly, T.P., Margolis, R.N., and Hubbard, A.L. The estab­
lishment of hepatocyte cell surface polarity during fetal liver development.
Dev Biol, 1987;123(1):73–84.
4. Ishii, M., Washioka, H., Tonosaki, A., and Toyota, T. Regional orientation of
actin filaments in the pericanalicular cytoplasm of rat hepatocytes.
Gastroenterology, 1991;101(6):1663–72.
5. Tzanakakis, E.S., Hansen, L.K., and Hu, W.S. The role of actin filaments and
microtubules in hepatocyte spheroid self‐assembly. Cell Motil Cytoskeleton,
2001;48(3):175–89.
6. Gupta, A., Schell, M.J., Bhattacharjee, A., Lutsenko, S., and Hubbard, A.L.
Myosin Vb mediates Cu+ export in polarized hepatocytes. J Cell Sci,
2016;129(6):1179–89.
7. Hales, C.M., Griner, R., Hobdy‐Henderson, K.C. et  al. Identification and
Characterization of a Family of Rab11‐interacting Proteins. J Biol Chem,
2001;276(42):39067–75.
8. Lapierre, L.A., Kumar, R., Hales, C.M. et al. Myosin vb is associated with
plasma membrane recycling systems. Mol Biol Cell, 2001;12(6):1843–57.
9. Müller, T., Hess, M.W., Schiefermeier, N. et  al. MY05B mutations cause
microvillus inclusion disease and disrupt epithelial cell polarity. Nat Genet,
2008;40(10):1163–5.
10. Wakabayashi, Y., Dutt, P., Lippincott‐Schwartz, J., and Arias, I.M. Rab11a
and myosin Vb are required for bile canalicular formation in WIF‐B9 cells.
Proc Natl Acad Sci U S A, 2005;102(42):15087–92.
11. Ruemmele, F.M., Müller, T., Schiefermeier, N. et  al. Loss‐of‐function of
MY05B is the main cause of microvillus inclusion disease: 15 Novel muta­
tions and a CaCo‐2 RNAi cell model. Hum Mutat, 2010;31(5):544–51.
12. Knowles, B.C., Roland, J.T., Krishnan, M. et al. Myosin Vb uncoupling from
RAB8A and RAB11A elicits microvillus inclusion disease. J Clin Invest,
2014;124(7):2947–62.
13. Kravtsov, D., Mashukova, A., Forteza, R. et al. Myosin 5b loss of function
leads to defects in polarized signaling: implication for microvillus inclusion
disease pathogenesis and treatment. Am J Physiol Liver Physiol,
2014;307(10):G992–1001.
14. Girard, M., Lacaille, F., Verkarre, V. et al. MY05B and bile salt export pump
contribute to cholestatic liver disorder in microvillous inclusion disease.
Hepatology, 2014;60(1):301–10.
15. Strautnieks, S.S., Bull, L.N., Knisely, A.S. et al. A gene encoding a liver‐­
specific ABC transporter is mutated in progressive familial intrahepatic
­cholestasis. Nat Genet, 1998;20(3):233–8.
16. Klomp, L.W.J., Vargas, J.C., Van Mil, S.W.C. et al. Characterization of muta­
tions in ATP8B1 associated with hereditary cholestasis. Hepatology,
2004;40(1):27–38.
17. Nyasae, L.K., Schell, M.J., and Hubbard, A.L. Copper directs ATP7B to
the  apical domain of hepatic cells via basolateral endosomes. Traffic,
2014;15(12):1344–65.
18. Tuma, P.L., Nyasae, L.K., and Hubbard, A.L. Nonpolarized cells selectively
sort apical proteins from cell surface to a novel compartment, but lack apical
retention mechanisms. Mol Biol Cell, 2002;13(10):3400–15.
19. Manton, I. and Clarke, B. An electron microscope study of the spermatozoid
of sphagnum. J Exp Bot, 1952;3(3):265–75.
20. Fawcett, D.W. and Porter, K.R. A study of the fine structure of ciliated
­epithelia. J Morphol, 1954;94(2):221–81.
21. Sabitini, D.D., Bensch, K., and Barrnett, R.J. Cytochemistry and electron
microscopy. The preservation of cellular ultrastructure and enzymatic
­activity by aldehyde fixation. J Cell Biol, 1963;17:19–58.
22. Ledbetter, M.C. and Porter, K.R. Morphology of microtubules of plant cell.
Science, 1964;144(3620):872–4.
23. Borisy, G.G. and Taylor, E.W. The mechanism of action of colchicine.
Binding of colchicine‐3H to cellular protein. J Cell Biol, 1967;34:525–33.
24. Bryan, J. and Wilson, L. Are cytoplasmic microtubules heteropolymers?
Proc Natl Acad Sci U S A, 1971;68(8):1762–6.
25. Mohri, H. Amino‐acid composition of “tubulin” constituting microtubules of
sperm flagella. Nature, 1968;217:1053–4.
26. Borisy, G.G., Olmsted, J.B., and Klugman. R.A. In vitro aggregation of cyto­
plasmic microtubule subunits. Proc Natl Acad Sci U S A, 1972;69(10):2890–4.
27. Oakley, B.R. γ‐Tubulin: the microtubule organizer? Trends Cell Biol,
1992;2(1):1–5.
28. Mitchison, T. and Kirschner, M. Dynamic instability of microtubule growth.
Nature, 1984;312(5991):237–42.
29. Matov, A., Applegate, K., Kumar, P. et al. Analysis of microtubule dynamic
instability using a plus‐end growth marker. Nat Methods, 2010;7(9):761–8.
30. Downing, K.H. and Nogales, E. Tubulin and microtubule structure. Curr
Opin Cell Biol, 1998;10:16–22.
31. McNally, F.J. and Vale, R.D. Identification of katanin, an ATPase that severs
and disassembles stable microtubules. Cell, 1993;75(3):419–29.
32. Dehmelt, L. and Halpain, S. The MAP2/Tau family of microtubule‐­
associated proteins. Genome Biol, 2005;6(1):204.
33. Gibbons, I.R. and Rowe, A.J. Dynein: a protein with adenosine ­triphosphatase
activity from cilia. Science, 1965;149(3682):424–6.
34. Gibbons, B.H. and Gibbons, I.R. Flagellar movement and adenosine triphos­
phatase activity in sea urchin sperm extracted with triton X‐100. J Cell Biol,
1972;54(1):75–97.
35. Allen, R.D., Metuzals, J., Tasaki, I., Brady, S.T., and Gilbert, S.P. Fast axonal
transport in squid giant axon. Science, 1982;218(4577):1127–9.
36. Brady, S.T., Lasek, R.J., and Allen, R.D. Fast axonal transport in extruded
axoplasm from squid giant axon. Science, 1982;218(4577):1129–31.
37. Schroer, T.A., Schnapp, B.J., Reese, T.S., and Sheetz, M.P. The role of kine­
sin and other soluble factors in organelle movement along microtubules.
J Cell Biol, 1988;107(5):1785–92.
38. Schnapp, B.J. and Reese, T.S. Dynein is the motor for retrograde axonal
transport of organelles. Proc Natl Acad Sci U S A, 1989;86(5):1548–52.
39. Allan, V.J. Cytoplasmic dynein. Biochem Soc Trans, 2011;39(5):1169–78.
40. Gibbons, I.R., Lee‐Eiford, A., Mocz, G. et al. Photosensitized cleavage of
dynein heavy chains. Cleavage at the “V1 site” by irradiation at 365 nm in
the presence of ATP and vanadate. J Biol Chem, 1987;262(6):2780–6.
41. Reck‐Peterson, S.L. and Vale, R.D. Molecular dissection of the roles of
nucleotide binding and hydrolysis in dynein’s AAA domains in Saccharomyces
cerevisiae. Proc Natl Acad Sci U S A, 2004;101(6):1491–5.
42. Kardon, J.R. and Vale, R.D. Regulators of the cytoplasmic dynein motor. Nat
Rev Mol Cell Biol, 2009;10:854–65.
43. Vallee, R.B., Williams, J.C., Varma, D., and Barnhart, L.E. Dynein: an
ancient motor protein involved in multiple modes of transport. J Neurobiol,
2004;58:189–200.
44. Johansson, M., Rocha, N., Zwart, W. et al. Activation of endosomal dynein
motors by stepwise assembly of Rab7‐RILP‐p150Glued, ORP1L, and the
receptor βIII spectrin. J Cell Biol, 2007;176(4):459–71.
45. Fu, M.M. and Holzbaur, E.L. Integrated regulation of motor‐driven organelle
transport by scaffolding proteins. Trends Cell Biol, 2014;24:564–74.
46. Mallik, R., Carter, B.C., Lex, S.A., King, S.J., and Gross, S.P. Cytoplasmic
dynein functions as a gear in response to load. Nature, 2004;427(6975):
649–52.
34 THE LIVER:  REFERENCES
47. Rai, A.K., Rai, A., Ramaiya, A.J., Jha, R., and Mallik, R. Molecular adapta­
tions allow dynein to generate large collective forces inside cells. Cell,
2013;152(1–2):172–82.
48. Anishkin, A. and Kung, C. Stiffened lipid platforms at molecular force foci.
Proc Natl Acad Sci U S A, 2013;110(13):4886–92.
49. Rai, A., Pathak, D., Thakur, S. et al. Dynein clusters into lipid microdomains
on phagosomes to drive rapid transport toward lysosomes. Cell,
2016;164(4):722–34.
50. Hirokawa, N. Kinesin and dynein superfamily proteins and the mechanism
of organelle transport. Science, 1998;279:519–26.
51. Oda, H., Stockert, R.J., Collins, C. et  al. Interaction of the microtubule
cytoskeleton with endocytic vesicles and cytoplasmic dynein in cultured rat
hepatocytes. J Biol Chem, 1995;270(25):15242–9.
52. Wang, Y. Cytoplasmic dynein participates in apically targeted stimulated
secretory traffic in primary rabbit lacrimal acinar epithelial cells. J Cell Sci,
2003;116(10):2051–65.
53. Vale, R.D., Reese, T.S., and Sheetz, M.P. Identification of a novel force‐
generating protein, kinesin, involved in microtubule‐based motility. Cell,
1985;42(1):39–50.
54. Brady, S.T., Lasek, R.J., and Allen, R.D. Fast axonal transport in extruded
axoplasm from squid giant axon. Science 1982;218(4577):1129–31.
55. Hirokawa, N., Noda, Y., Tanaka, Y., and Niwa, S. Kinesin superfamily
motor  proteins and intracellular transport. Nat Rev Mol Cell Biol,
2009;10:682–96.
56. Kuznetsov, S.A., Vaisberg, E.A., Shanina, N.A. et al. The quaternary struc­
ture of bovine brain kinesin. EMBO J, 1988;7(2):353–6.
57. Bloom, G.S., Wagner, M.C., Pfister, K.K., and Brady, S.T. Native structure
and physical properties of bovine brain kinesin and identification of the ATP‐
binding subunit polypeptide. Biochemistry, 1988;27(9):3409–16.
58. Asenjo, A.B., Krohn, N., and Sosa H. Configuration of the two kinesin motor
domains during ATP hydrolysis. Nat Struct Biol, 2003;10(10):836–42.
59. Yildiz, A., Tomishige, M., Vale, R.D., and Selvin, P.R. Kinesin walks hand‐
over‐hand. Science, 2004;303(5658):676 LP‐678.
60. Yun, M., Bronner, C.E., Park, C.G. et al. Rotation of the stalk/neck and one
head in a new crystal structure of the kinesin motor protein, Ncd. EMBO J,
2003;22(20):5382–9.
61. Noda, Y., Sato‐Yoshitake, R., Kondo, S., Nangaku, M., and Hirokawa, N.
KIF2 is a new microtubule‐based anterograde motor that transports membra­
nous organelles distinct from those carried by kinesin heavy chain or
KIF3A/B. J Cell Biol, 1995;129(1):157–67.
62. Rosario, J., Sutherland, E., Zaccaro, L., and Simon, F.R. Ethinylestradiol
administration selectively alters liver sinusoidal membrane lipid fluidity and
protein composition. Biochemistry, 1988;27(11):3939–46.
63. Novikoff, P.M., Cammer, M., Tao, L. et al. Three‐dimensional organization
of rat hepatocyte cytoskeleton: relation to the asialoglycoprotein endocytosis
pathway. J Cell Sci, 1996;109(Pt 1):21–32.
64. Murray, J.W. and Wolkoff, A.W. Roles of the cytoskeleton and motor pro­
teins in endocytic sorting. Adv Drug Deliv Rev, 2003;55:1385–403.
65. Goltz, J.S., Wolkoff, A.W., Novikoff, P.M., Stockert, R.J., and Satir P. A role
for microtubules in sorting endocytic vesicles in rat hepatocytes. Proc Natl
Acad Sci U S A, 1992;89(15):7026–30.
66. Rai, P., Kumar, M., Sharma, G. et al. Kinesin‐dependent mechanism for con­
trolling triglyceride secretion from the liver. Proc Natl Acad Sci U S A,
2017;114(49):12958–63.
67. Allan, V.J., Thompson, H.M., and McNiven, M.A. Motoring around the
Golgi. Nat Cell Biol, 2002;4:E236–42.
68. Van der Sluijs, P., Bennett, M.K., Antony, C., Simons, K., and Kreis, T.E.
Binding of exocytic vesicles from MDCK cells to microtubules in vitro. J
Cell Sci, 1990;95(Pt 4):545–53.
69. Fath, K.R., Trimbur, G.M., and Burgess, D.R. Molecular motors and a spec­
trin matrix associate with Golgi membranes in vitro. J Cell Biol, 1997;139(5):
1169–81.
70. Kreitzer, G., Schmoranzer, J., Low, S.H. et al. Three‐dimensional analysis of
post‐Golgi carrier exocytosis in epithelial cells. Nat Cell Biol, 2003;5:126–36.
71. Kreitzer, G., Marmorstein, A., Okamoto, P., Vallee, R., and Rodriguez‐
Boulan, E. Kinesin and dynamin are required for post‐Golgi transport of a
plasma‐membrane protein. Nat Cell Biol, 2000;2(2):125–7.
72. Berg, T., Kindberg, G.M., Ford, T., and Blomhoff, R. Intracellular transport
of asialoglycoproteins in rat hepatocytes. Evidence for two subpopulations
of lysosomes. Exp Cell Res, 1985;161(2):285–96.
73. Bananis, E., Murray, J.W., Stockert, R.J., Satir, P., and Wolkoff, A.W.
Regulation of early endocytic vesicle motility and fission in a reconstituted
system. J Cell Sci, 2003;116(Pt 13):2749–61.
74. Bananis, E., Nath, S., Gordon, K. et al. Microtubule‐dependent movement of
late endocytic vesicles in vitro: requirements for dynein and kinesin. Mol
Biol Cell, 2004;15(8):3688–97.
75. Brasaemle, D.L., Dolios, G., Shapiro, L., and Wang, R. Proteomic analysis
of proteins associated with lipid droplets of basal and lipolytically stimulated
3T3‐L1 adipocytes. J Biol Chem, 2004;279(45):46835–42.
76. Yang, L., Ding, Y., Chen, Y. et al. The proteomics of lipid droplets: structure,
dynamics, and functions of the organelle conserved from bacteria to humans.
J Lipid Res, 2012;53(7):1245–53.
77. Gao, Q. and Goodman, J.M. The lipid droplet – a well‐connected organelle.
Front Cell Dev Biol, 2015;3.
78. Guimaraes, S.C., Schuster, M., Bielska, E. et al. Peroxisomes, lipid droplets,
and endoplasmic reticulum “hitchhike” on motile early endosomes. J Cell
Biol, 2015;211(5):945–54.
79. Boström, P., Rutberg, M., Ericsson, J. et al. Cytosolic lipid droplets increase
in size by microtubule‐dependent complex formation. Arterioscler Thromb
Vasc Biol, 2005;25(9):1945–51.
80. Wiggins, D. and Gibbons, G.F. The lipolysis/esterification cycle of hepatic tria­
cylglycerol. Its role in the secretion of very‐low‐density lipoprotein and its
response to hormones and sulphonylureas. Biochem J, 1992;284(Pt 2):457–62.
81. Orci, L., Marchand, Y.L.E., Singh, A., Assimacopoulos‐Jennnet, F., Rouiller,
C., and Jeanrenaud, B. Role of microtubules in lipoprotein secretion by the
liver. Nature, 1973;244:30.
82. Reaven, E.P. and Reaven, G.M. Evidence that microtubules play a permis­
sive role in hepatocyte very low density lipoprotein secretion. J Cell Biol,
1980;84(1):28–39.
83. Barak, P., Rai, A., Rai, P., and Mallik, R. Quantitative optical trapping on
single organelles in cell extract. Nat Methods, 2013;10(1):68–70.
84. Schulze, R.J. and McNiven, M.A. Fasting inhibits the recruitment of kine­
sin‐1 to lipid droplets and stalls hepatic triglyceride secretion. Hepatology,
2019;69(1):444–6.
85. Gilham, D., Alam, M., Gao, W., Vance, D.E., and Lehner, R. Triacylglycerol
hydrolase is localized to the endoplasmic reticulum by an unusual retrieval
sequence where it participates in VLDL assembly without utilizing VLDL
lipids as substrates. Mol Biol Cell, 2005;16(2):984–96.
86. Lankester, D.L., Brown A.M., and Zammit, V.A. Use of cytosolic triacylg­
lycerol hydrolysis products and of exogenous fatty acid for the synthesis
of  triacylglycerol secreted by cultured rat hepatocytes. J Lipid Res,
1998;39(9):1889–95.
87. Anderwald, C., Bernroider, E., Krssak, M. et al. Effects of insulin treatment
in type 2 diabetic patients on intracellular lipid content in liver and skeletal
muscle. Diabetes, 2002;51(10):3025–32.
88. Sadh, K., Rai, P., and Mallik, R. Feeding‐fasting dependent recruitment of
membrane microdomain proteins to lipid droplets purified from the liver.
PLoS One, 2017;12(8).
89. Scheel, T.K.H. and Rice, C.M. Understanding the hepatitis C virus life cycle
paves the way for highly effective therapies. Nat Med, 2013;19:837–49.
90. Boulant, S., Douglas, M.W., Moody, L. et al. Hepatitis C virus core protein
induces lipid droplet redistribution in a microtubule‐ and dynein‐dependent
manner. Traffic, 2008;9(8):1268–82.
91. Wozniak, A.L., Long, A., Jones‐Jamtgaard, K.N., and Weinman, S.A.
Hepatitis C virus promotes virion secretion through cleavage of the Rab7
adaptor protein RILP. Proc Natl Acad Sci U S A, 2016;201607277.
92. LeDizet, M. and Piperno, G. Identification of an acetylation site of
Chlamydomonas alpha‐tubulin. Proc Natl Acad Sci U S A, 1987;84(16):5720–4.
93. Shida, T., Cueva, J.G., Xu, Z., Goodman, M.B., and Nachury, M.V. The
major alpha‐tubulin K40 acetyltransferase TAT1 promotes rapid ciliogenesis
and efficient mechanosensation. Proc Natl Acad Sci U S A, 2010;107(50):
21517–22.
94. North, B.J., Marshall, B.L., Borra, M.T., Denu, J.M., and Verdin, E. The
human Sir2 ortholog, SIRT2, is an NAD+‐dependent tubulin deacetylase.
Mol Cell, 2003;11(2):437–44.
95. Hubbert, C., Guardiola, A., Shao, R. et  al. HDAC6 is a microtubule‐­
associated deacetylase. Nature, 2002;417(6887):455–8.
96. Xu, Z., Schaedel, L., Portran, D. et  al. Microtubules acquire resistance
from  mechanical breakage through intralumenal acetylation. Science,
2017;356(6335):328–32.
 3:  Cytoskeletal Motors: Structure and Function in Hepatocytes 35
97. LeDizet, M. and Piperno G. Cytoplasmic microtubules containing acety­
lated  ??‐tubulin in Chlamydomonas reinhardtii: spatial arrangement and
properties. J Cell Biol, 1986;103(1):13–22.
98. Park, I.Y., Powell, R.T., Tripathi, D.N. et al. Dual chromatin and cytoskel­
etal remodeling by SETD2. Cell, 2016;166(4):950–62.
99. Sirajuddin, M., Rice, L.M., and Vale, R.D. Regulation of microtubule
motors by tubulin isotypes and post‐translational modifications. Nat Cell
Biol, 2014;16:335.
100. Kreitzer, G., Liao, G., and Gundersen, G.G. Detyrosination of tubulin
regulates the interaction of intermediate filaments with microtubules in
vivo via a kinesin‐dependent mechanism. Mol Biol Cell, 1999;10(April):
1105–18.
101. Peris, L., Thery, M., Fauré, J. et al. Tubulin tyrosination is a major factor
affecting the recruitment of CAP‐Gly proteins at microtubule plus ends. J
Cell Biol, 2006;174(6):839–49.
102. Groebner, J.L. and Tuma PL. The altered hepatic tubulin code in alcoholic
liver disease. Biomolecules, 2015;5:2140–59.
103. Poüs, C., Chabin, K., Drechou, A. et al. Functional specialization of stable
and dynamic microtubules in protein traffic in WIF‐B cells. J Cell Biol,
1998;142(1):153–65.
104. Phung‐Koskas, T., Pilon, A., Poüs, C. et  al. STAT5B‐mediated growth
­hormone signaling is organized by highly dynamic microtubules in hepatic
cells. J Biol Chem, 2005;280(2):1123–31.
105. Kannarkat, G.T., Tuma, D.J., and Tuma, P.L. Microtubules are more stable
and more highly acetylated in ethanol‐treated hepatic cells. J Hepatol,
2006;44(5):963–70.
106. Mackeh, R., Lorin, S., Ratier, A. et  al. Reactive oxygen species, amp‐
Activated protein kinase, and the transcription cofactor p300 regulate α‐
Tubulin acetyltransferase‐1 (αtat‐1/mec‐17)‐dependent microtubule
hyperacetylation during cell stress. J Biol Chem, 2014;289(17):11816–28.
107. Herms, A., Bosch, M., Reddy, B.J.N. et al. AMPK activation promotes lipid
droplet dispersion on detyrosinated microtubules to increase mitochondrial
fatty acid oxidation. Nat Commun, 2015;6:7176.
108. Ihrke, G., Neufeld, E.B., Meads T. et al. WIF‐B cells: An in vitro model for
studies of hepatocyte polarity. J Cell Biol, 1993;123(6 II):1761–75.
109. Meads, T. and Schroer, T.A. Polarity and nucleation of microtubules in
polarized epithelial cells. Cell Motil Cytoskeleton, 1995;32(4):273–88.
110. Coller, K.E., Heaton, N.S., Berger, K.L. et al. Molecular determinants
and dynamics of hepatitis C virus secretion. PLOS Pathog, 2012;8(1):
e1002466.
 Hepatocyte Surface Polarity
4 Anne Müsch1 and Irwin M. Arias2
1
2
Department of Developmental & Molecular Biology, Albert Einstein College of Medicine, Bronx, NY, USA
National Institutes of Health, Bethesda, MD, USA
INTRODUCTION
With exception of erythrocytes and possibly a few others, all
mammalian cells are polarized. Epithelial cell polarity, as pre-
sent in the liver, gastrointestinal tract, and kidney, developed
during the Precambrian Period and is essential for the formation
of multicellular organs and species. Selective absorption and
secretion requires cellular polarization. Hepatocyte polarization
is neglected, unique, complex, and, as shown in recent studies,
critical in many inheritable and acquired liver diseases.
Surprisingly, only one current text in hepatology or pathol-
ogy [1] mentions hepatocellular polarity. Neither the authors
nor a large number of hepatologists or pathologists recall seeing
a pathology report which comments on the polarization status of
hepatocytes! A possible explanation is that hepatology devel-
oped in the late nineteenth century, when efforts were always
made to associate pathology with clinical signs, symptoms, and
outcome. The major tool in this effort was hematoxylin and
eosin (H&E)‐stained sections of liver. The H&E stain does not
reveal the small bile canaliculus, which is the apical polarized
domain of the hepatocyte. The bile canaliculus is functionally
sealed by tight junctions and, with its microvilli, constitutes
~13% of total hepatocyte plasma membrane [1]. Discovery of
the bile canaliculus did not occur until transmission and scan-
ning electron microscopy were introduced in the 1940s. Despite
the availability of numerous bile canalicular proteins and asso-
ciated antibodies, such as 5′ nucleotidase, cCAM 105, ABC B1,
ABC B11, and others, they are rarely used in routine studies of
liver histopathology. Probably because polarity is not described
by pathologists, it is also neglected by clinicians.
This chapter will consider mechanisms responsible for hepa-
tocellular polarity. Canalicular network formation and hepato-
cyte polarity require coordinated expression of several key
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
elements, including extracellular matrix (ECM), adherens and
tight junctions, intracellular protein trafficking machinery,
cytoskeleton, and energy production. Inheritable and acquired
defects in these elements that result in depolarization or failure
to polarize are presented in Chapter 29.
THE UNIQUE HEPATOCYTE POLARITY
PHENOTYPE
The liver presents a remarkable example of how cell shape
serves function. The two liver epithelial cell types – hepatocytes
and bile duct cells  –  adopt radically different polarity pheno-
types that serve their distinct physiological roles. Bile duct cells,
which form simple conduits for bile, organize like other tubule‐
forming epithelial cells with a single luminal domain perpen-
dicular to cell–cell adhesion domains and opposite a basal
surface; they are monopolar. Hepatocytes, which mediate exten-
sive bidirectional molecular exchange with the blood, maximize
contact with blood vessels by establishing a second basal sur-
face in place of the biliary cell’s apical surface. The hepatocyte
apical domains assemble instead at cell–cell contact sites,
interrupting cell–cell adhesion domains (Figure  4.1). Each
­
hepatocyte forms two or more lumina with several of its neigh-
bors, yielding the branched luminal network known as bile
canaliculi. Hepatocytes are therefore multipolar. While the
­
monopolar organization of bile duct cells is similar to that of
most epithelia, the hepatocyte polarity phenotype is unique.
Immunohistochemistry, electron microscopy, and gene
expression analysis collectively indicate that hepatocytes form
functional cell–cell adherens and tight junctions, and that they
express and localize the constituents of evolutionary conserved
 4:  Hepatocyte Surface Polarity 37

(a)
Basal
Hepatocytes
Lateral
Luminal/apical

Bile ducts

(b) (A) (B)

HA 321
HA4

10

(C) (D) Sinusoidal

front
BC

Golgi
complex
BC
High
concentrations
of endo/exocytotic BC
organelles BC

Figure 4.1  (a) The two polarity phenotypes in the liver – hepatocytes and bile duct epithelia – have three distinct membrane domains: basal
domains facing the extracellular matrix and underlying blood vessels; lateral domains engaged in cell–cell adhesion; and luminal domains, which
face the outside world (they are situated at the cell apex in monopolar cells and hence also called apical domains). Hepatocytes have multiple lumi-
nal surfaces that interrupt cell–cell adhesion domains and most feature two basal surfaces facing the space of Disse on either side. Bile duct epithe-
lial cells each have a single luminal and basal domain flanked by lateral surfaces. Adapted from Müsch, Curr Opin Cell Biol, 2018;54:18–23 with
permission of Elsevier. (b) The architecture of the liver is unique. A: A scanning electron micrograph of a portion of a liver lobule is shown. A
continuous network of bile canaliculi runs along the exposed cell surfaces of the liver plate. B: The two distinct PM domains are visualized by
immunofluorescence detection of the basolateral PM protein, HA321/BEN and the apical PM protein, HA4/cell‐CAM105/ectoATPase. An electron
micrograph of a hepatocyte (C) and a corresponding schematic drawing (D) highlight the “active zones” in vesicular trafficking. The major sorting
organelles (TGN and endosomes) and transport vesicles are concentrated in small “clear” zones that are probably the most active in vesicle traffic.
These zones are located between the Golgi and the apical PM and near the basolateral PM (the shaded regions in D). Reproduced from Braiterman
and Hubbard, The Liver, 5th edn, Fig. 6.1.
38 THE LIVER:  EXPERIMENTAL SYSTEMS FOR THE STUDY OF EPITHELIAL POLARIZATION IN THE LIVER
polarity complexes, which establish epithelial surface identities,
in a similar manner to that of monopolar epithelial cells. These
observations make it likely that the distinct polarity phenotypes
are caused by different regulation of the same key polarization
mechanisms. Identification of regulatory mechanisms that
define the two polarity phenotypes is only beginning to emerge.
For a comprehensive listing of polarity‐associated proteins,
their functions, and expression in hepatocytes, see Braiterman
and Hubbard, The Liver, 5th edn, Table 6.2.
ESTABLISHMENT OF CONNECTED BILE
DUCT AND BILE CANALICULAR
NETWORKS DURING LIVER
DEVELOPMENT
The two liver epithelial cells originate from common precur-
sors, called hepatoblasts. Hepatoblasts delaminate from a mon-
olayered epithelial tube, the foregut, proliferate and invade the
surrounding mesenchyme [2, 3] (see Chapter  2 for more
details). Hepatoblasts around the portal vein first form a mon-
olayered epithelium, the ductal plate  [4]. Two prominent
features distinguish this from the adjacent hepatoblasts: a
­ chronous polarization of monopolar epithelial cells in culture,
laminin‐positive basement membrane and strong E‐cadherin
labeling at cell–cell contact sites. Parts of the ductal plate
develop into bile ducts, while the remnants have been proposed
to become hepatic stem cells [5, 6]. Bile ducts develop when
the ductal plate cells signal adjacent hepatoblasts to adopt a
similar biliary fate. The two cell layers establish a basal lamina
encircling them, and then generate a lumen between them.
Hepatoblasts outside the vicinity of portal veins develop into
hepatocytes. Hepatocytes acquire polarity (i.e. form bile cana-
liculi) only in late gestation, and the bile canalicular network
continues to elongate postnatally [7, 8]. The hepatocyte polari-
zation spurt coincides with the onset of bile acid synthesis
[8–10]. Remarkably, the major bile acid taurocholate stimu-
lates hepatocyte polarization in culture, suggesting that tauro-
cholate serves as hepatocyte polarization trigger [11].
For bile excreted into hepatocyte bile canaliculi to reach the
bile duct, the two epithelial tissues must link their respective
luminal networks into a contiguous entity. This is accomplished
when major bile ducts branch into smaller ductules, which con-
nect to the bile canaliculi. Techniques that combined thick sec-
tion immunofluorescence on cleared tissue with carbon ink
injections into the common bile duct revealed that branching of
smaller ductules from the larger ducts happened in parallel with
the spurt in bile canaliculi formation and with the first appear-
ance of bile in the intestine at E18 in the mouse [12].
Pharmacological inhibition of the hepatocyte bile acid trans-
porter Mrp2, which pumps bile acids into the bile canaliculi,
prevented bile duct branching. Bile acids that reach bile ducts
from these newly formed bile canaliculi might thus constitute
signals for ductal branching into narrower ductules, which then
connect to the bile canaliculi. Taurocholate signaling for bile
canaliculi formation involves activation of the kinase liver
kinase B1 (LKB1) and its effector AMP‐regulated kinase
(AMPK) [11], a signaling cascade that has also been linked to
branching morphogenesis in the lung [13]. Future work is
expected to determine whether a taurocholate AMPK signaling
cascades also governs bile duct branching.
EXPERIMENTAL SYSTEMS FOR THE
STUDY OF EPITHELIAL POLARIZATION
IN THE LIVER
Due to lack of a readily available stable culture system for pri-
mary hepatocytes, much of what has been learned about hepato-
cyte polarization comes from polarized hepatocytic cell lines, in
particular from Hep G2 and WIF‐B cells, which are derived
from cancers [14–17]. Both form spherical lumina between
neighboring cells but do not develop a canalicular network. The
recently established Can10 line, which, like WIF‐B, is derived
from a rat hepatoma, is the first hepatocytic cell line in which
individual lumina connect to form canaliculi; however, it has not
yet been extensively characterized [18]. HepG2 and WIF‐B
cells recapitulate some critical hallmarks of the hepatocytic
polarity phenotype observed in primary cultures or in vivo, such
as the gradual, nonsynchronous formation of luminal surfaces at
cell–cell contact sites, which contrast with the rapid and syn-
the asymmetric cell divisions subsequently discussed, and the
indirect targeting of apical single membrane‐spanning and GPI‐
anchored proteins. HepG2 polarization is sensitive to oncostatin
M, a cytokine critical for hepatocyte differentiation in vivo [19];
however, neither cell line responds to taurocholic acid, the pre-
sumptive in vivo trigger of bile canaliculi formation. This is a
serious limitation to the study of signaling mechanisms leading
to hepatocyte lumen formation. Also, the mechanism and regu-
lation of ABC transporter trafficking in WIF‐B and HepG2 cells
differ in several respects from that observed in hepatocytes [20].
Primary hepatocytes, isolated by liver perfusion with colla-
genase, re‐establish polarity and the dense canalicular network
found in intact liver when cultured in collagen sandwiches.
These cultures not only repolarize in a predictable manner, but
they also maintain function and gene expression profiles of
mature hepatocytes for up to two weeks [21–24]. This culture
system, although less amenable to experimental manipulation
than cell lines, provides information directly related to hepato-
cellular processes as confirmed by subsequent gene knockout
studies subsequently described [25].
In addition to models for hepatocyte polarity, a few cell lines
derived from normal biliary cells have also been established
[26, 27]. These lines polarize with monopolar organization in
2D culture on permeable filter substrates where they acquire
high trans‐epithelial resistance, a measure of epithelial barrier
function [28]. The Normal Rat Cholangiocyte line (NRC) devel-
oped by LaRusso’s group [27] develops a primary cilium, which
is a sign of biliary differentiation [29], and recapitulates the in
vivo regulation and the cell surface appearance of biliary water
and ion channels [30–32].
In development and during recovery from injury, hepatocytes
and biliary cells evolve from common progenitors. These
include stem cells and transit‐amplifying cells, called hepato-
blasts [33]. The liver stem cell reservoir is believed to reside at
 4:  Hepatocyte Surface Polarity 39
the canal of Hering (the junction between bile ductules and
hepatocytes) and in peribiliary glands and the gallbladder and is
present throughout life [5, 34]. Hepatoblasts, on the other hand,
are abundant during liver development but in the adult only
detectable in response to liver injury. Both types of progenitors
have been isolated and expanded in culture in attempts to dif-
ferentiate them into the two epithelial lineages [7, 35–37]. Few
of these studies have convincingly demonstrated hepatocytic or
biliary polarity. In those instances where proper polarization
was achieved, the progenitors gave rise to only one but not both
lineages [38, 39]. Thus, in vitro recapitulation of the branching
of the two liver epithelial cell types still eludes the field. A pat-
ented human cancer cell line with hepatoblast characteristics,
HepaRG [40], can be polarized with either hepatocyte polarity
when cultured in spheroids [41], or as monolayered hollow
cysts (i.e. with bile duct polarity) in 3D matrigel culture [42].
Because of its human origin and relatively high degree of func-
tional differentiation when cultured in spheroids, these cells
have become popular for toxicology studies, although they have
yet to be exploited for polarity studies.
Two technical advances have opened the door to elucidating
polarity mechanisms even in the intact liver. The first pertains to
our ability to express or ablate proteins in hepatocytes by tail
vein injection of adenoviruses [43, 44] or of even naked DNA
via hydrodynamic injections [45, 46]. This “poor man’s genet-
ics” approach allows acute in vivo manipulation of hepatocytes
with relative ease when compared to the generation of knockout
and transgenic animals. The second advance is in intravital
imaging, which now allows live cell imaging of cellular pro-
cesses in liver lobes [47]. Such analysis has recently shown that
hepatocytes lacking the kinase LKB1 (discussed in more detail
below) have leaky tight junctions [48].
These novel techniques add to the traditional biochemical,
histochemical, and electron microscopic studies of the liver and
complement dynamic live cell imaging performed in cell lines.
POLARIZATION MECHANISMS
Polarity complexes
Cell surface polarity is established when signals generated by
local cues or by stochastic fluctuation become amplified through
feedback mechanisms to yield a robust segregation of distinct
membrane domains. Work in invertebrates combined with theo-
retical modeling identified a set of core signaling mechanisms
that could generate cell‐autonomous polarity. They include sev-
eral conserved epithelial signaling complexes operating through
a combination of positive feedback and mutual antagonistic
signaling to stake out distinct surface domains (reviewed in
[49–51] and Figure 4.2). The transmembrane protein Crumbs,
in conjunction with a Cdc42–aPKC–Par6–Par3 signaling net-
work generates apical polarity in part by clustering and stabiliz-
ing Crumbs; a Dlg–Lgl–Scribble network promotes basolateral
surface identity by preventing Crumbs accumulation and by
promoting the accumulation of lateral membrane proteins,
including E‐cadherin. The third complex centered around PALs/
Patj and the aPKC–Par3 proteins operates at the intersection
Crb
ls P
Pa Par6 Cdc42
Patj aPKC
TJ TJ
P P P Par3
Par1 Lgl
Crb
Cdc42 Lgl
T
Par6
Scrib
Dlg
P
Par3 Par1
T
Figure 4.2  Polarity complexes organize two distinct membrane
domains (apical and basolateral) through self‐recruitment, positive
feedback, and mutual inhibition. The transmembrane protein Crumbs
(Crb) engages in lateral homotypic interactions via its extracellular
domain and becomes phosphorylated by aPKC at its cytoplasmic
domain. Both events stabilize the protein, which determines the apical
domain. Crb interacts with the Pals/Patj complex, which links the apical
domain to tight junctions. aPKC, which is activated by scaffolding with
active Cdc42 and Par6, also directs its substrate Par3 to tight junctions
and prevents basolateral surface determinants Lgl and Par1 from attach-
ing to the apical domain. Conversely, Lgl, which genetically interacts
with Scribble and Dlg, promotes Crb endocytosis from the basolateral
domain and inhibits the Cdc42/Par6 complex to operate at the lateral
membrane. Par1‐mediated phosphorylation of Par3 removes Par3 from
the lateral domain.
between the apical and lateral complexes and in mammalian
cells is critical for the formation of tight junctions. Available
evidence suggests that the Par, Crumbs, and Scribble complexes
are present in both monopolar epithelia and in hepatocytes.
Thus, in the adult liver, immunohistochemical analysis has
determined that aPKC iota and zeta and Par3 are colocalized
with tight junction markers, specifically ZO‐1, occludin, and
claudin‐3 at the boundary between bile canaliculi and sinusoidal
membranes, suggesting that they function in apical junctional
complexes as in other epithelial cells [52]. The basolateral
polarity determinants Scribble and Lgl‐2 have been localized to
the sinusoidal domains in the polarized hepatocytic WIF‐B cells
and Scribble has been shown to mediate similar protein–protein
interactions in the hepatocytic cell line WIF‐B and in monopo-
lar (kidney‐derived) MDCK cells [53].
Cell matrix adhesion
In the context of the tissue, these “polarity complexes” are
likely positioned by external cues, which include diffusible sig-
nals and signals originating from cell–ECM and cell–cell adhe-
sion. Thus, during bile duct formation signals from the portal
mesenchyme mediate the deposition of a basement membrane
between mesenchymal and ductal cells [54, 55]. The basement
40 THE LIVER:  POLARIZATION MECHANISMS
membrane is a two‐dimensional sheet of matrix molecules
secreted by epithelial cells and the underlying connective tis-
sue. It is organized into sheets primarily by laminin and fibrous
collagen IV networks that are connected with each other via
nidogen and the proteoglycan perlecan. It captures growth fac-
tors and cytokines and activates cell surface receptors, mostly
integrins. ECM–cell signaling triggered by the basement mem-
brane in the ductal cells cues apical surface formation on the
membrane domain opposite the basement membrane.
Consistent with a role for laminin in bile duct formation, 3D
polarization of hepatoblast‐derived biliary cells into a hollow
monolayered cyst in vitro required culture in laminin‐rich
matrices and was dependent on the laminin receptor β1‐integrin
[38, 56]. Remarkably, hepatocytes differ from other epithelial
cells in that they do not assemble a basal lamina [57]. While
collagen IV and laminin expression is stimulated during hepa-
toblast differentiation along the biliary lineage, expression of
these ECM proteins is not stimulated during hepatocyte differ-
entiation [58]. Mature hepatocytes completely lack expression
of the critical basal lamina constituents laminin and nidogen
[59]. Whether this unique epithelial feature contributes to the
unique hepatocyte polarity phenotype has not been directly
tested in vivo but several pieces of circumstantial evidence sup-
port such a scenario: (1) Fibrosis, which results in extensive
matrix deposition between hepatocytes and endothelial cells
results in loss of hepatocyte polarity [60, 61]; (2) lack of ECM
deposition proved critical for hepatocytic polarity in an experi-
mental model in which monopolar and hepatocytic polarity can
be switched. This model relies on the kidney‐derived cell line
MDCK, which exhibits monopolar organization but switches to
hepatocytic organization when induced to overexpress the
kinase Par1b [62]. Hepatocytic polarization upon Par1b over-
expression coincided with reduced ECM deposition and focal
adhesion; importantly, this hepatocytic polarity phenotype was
reversed when the cells were plated on a collagen IV matrix
[63]. Further investigation revealed reduced RhoA activity
downstream of the lack of ECM signaling as the critical polar-
ity cue: RhoA depletion was sufficient to induce hepatocytic
polarity in MDCK cells, while pharmacological Rho activation
in the hepatocytic cell line WIF‐B promoted their monopolar
organization [64, 65]. These findings point to RhoA signaling
downstream of cell adhesion signaling as a putative key regula-
tor of the polarity phenotypes.
Cell–cell junctions
In polarized epithelia the contacting membranes form three
types of intercellular junctions: tight junctions, anchoring junc-
tions, and gap junctions. Tight junctions provide a barrier for
paracellar flow of macromolecules and solutes, thus enforcing
their vectorial transport across epithelial cells; they also restrict
the diffusion of membrane proteins between the apical and
basolateral domains, thereby maintaining surface polarity.
Anchoring junctions, which include adherens junctions and
­desmosomes, couple cytoskeletal elements to the plasma mem-
brane, providing mechanical integrity and allowing mechanical
coupling of cells. Gap junctions mediate cell–cell communica-
tion by permitting the passage of small molecular weight solutes
(up to 1 kDa) directly between neighboring cells. Intercellular
junctions are not just organizers of cell structure but serve as
signaling platforms regulating gene expression, differentiation,
proliferation, and morphogenesis. It is this signaling role that
makes them well suited to translate an external cue (i.e. cell–cell
contact) into a polarization sequence.
Consistent with this idea, hepatocyte luminal domains form at
sites of cell–cell contact. Similarly, in the experimental model of
a single nonpolarized MDCK cell, which first establishes polar-
ity after cell division yields two daughters, the new cell–cell con-
tact site serves as patch for the establishment of a lumen between
the two cells (Figure 4.3). Lumen formation in this experimental
system ensues when endocytosed apical proteins, which consti-
tutively cycle between endosomes and the entire (nonpolarized)
plasma membrane in single cells, are specifically targeted to and
accumulate at the new cell–cell contact site. In this manner, api-
cal proteins are progressively cleared from the plasma membrane
outside the cell–cell contact region and are transported to the
nascent luminal domain [66, 67]. This constitutes a pathway
called basolateral‐to‐apical transcytosis. Once MDCK cells have
established their first luminal surface, however, new cell–cell
contacts generated after additional cell divisions no longer
­support de novo lumen formation, resulting in the characteristic
monopolar organization with a single luminal domain.
Hepatocytes differ from monopolar cells in that each cell–cell
contact may trigger a new lumen. We hypothesize that two prin-
cipal mechanisms are responsible for this difference: the nature
of cell–cell adhesion signaling and apical ­protein trafficking.
(a) Bas Bas Lat Bas
Ap
Bas Ap
(b)
Monopolar Multipolar
Figure 4.3  Model for lumen formation in monopolar (ductal cells)
versus multipolar (hepatocytes) epithelial cells. (a) Both monopolar and
hepatocytic cells initiate de novo lumen formation at cell–cell adhesion
domains. In monopolar MDCK cell doublets embedded in a 3D matrix
the cell–cell contact domain serves as targeting patch for recycling api-
cal membrane cargo, which prior to cell–cell contact formation recycles
in a nonpolarized manner. When the apical targeting patch is sealed
from the surrounding lateral membrane by tight junctions, water trans-
port and the resulting turgor can inflate a lumen. This mechanism is also
hypothesized for hepatocytes, which similarly initiate lumina at cell–
cell contact sites. (b) Monopolar and multipolar epithelia differ in the
formation of secondary luminal domains: Hepatocytes can form lumina
at each of their cell–cell contact sites. Monopolar epithelia, by contrast,
do not form additional lumina at new cell–cell contact sites. Instead,
cell divisions enlarge the existing lumen.
 4:  Hepatocyte Surface Polarity 41
As determined in MDCK cells, basolateral‐to‐apical transcy-
tosis is the mechanism by which apical domains form de novo
when new cell–cell junctions become a target site for apical‐
directed cargo. Basolateral‐to‐apical transcytosis is also the
mechanism by which newly synthesized single membrane‐
spanning and glycosylphosphatidylinositol (GPI)‐anchored api-
cal proteins are targeted to the luminal domain in hepatocytes
[68]. By contrast, MDCK cells and other monopolar epithelia
differ from hepatocytes in that they predominantly target newly
synthesized apical proteins directly to the apical surface [69].
Taken together, these two observations suggest that transcytotic
apical targeting in hepatocytes ensures that there is an abundant
supply of apical proteins to create new apical lumina at each
new hepatocyte contact domain. In contrast, direct apical target-
ing to an existing apical domain in monopolar epithelia, com-
bined with the inherent low rate of endocytic recycling at the
apical surface [70, 71] means apical cargo is unavailable for
endocytic targeting to new cell–cell contact sites. This could
prevent formation of new apical domains at cell–cell junctions
in monopolar epithelia such as bile ducts.
Cell–cell adhesion is initiated by nectins, a family of four
IgG‐like, Ca2+‐independent adhesion molecules, which activate
GTP exchange factors for the small GTPases Cdc42, Rac1, and
Rap1. Signaling by these GTPases then promotes establish-
ment of Ca2+‐dependent adhesion by cadherins, predominantly
E‐cadherin, which in mature adherence junctions clusters in a
zonula adherens adjacent to tight junctions and is linked to a
circumferential actin belt [72, 73]. Hepatocytes also express N‐
cadherin [74, 75], which belongs to the same class as E‐cad-
herin but in other epithelial cells replaces E‐cadherin during
epithelial‐to‐mesenchymal transitions (EMT) [76]. How N‐cadherin‐
expressing hepatocytes avoid the fate of EMT is not known.
Junctional adhesion molecules (JAMs), of which JAM‐A is the
best studied, are, together with claudins and occludin, the adhe-
sion molecules of tight junctions. They initially mingle with
nectin and E‐cadherin at cell–cell contacts in immature junc-
tions prior to their segregation into separate adherens and tight
junctions [77, 78]. Of these TJ proteins, JAM‐A is absolutely
essential for lumen formation in both hepatocytic (WIF‐B) [79]
and monopolar (MDCK) cell lines [80], whereas claudin com-
position appears to contribute to the decision to polarize with
monopolar or heptocytic polarity. Claudins are the multi‐­
membrane‐spanning proteins whose extracellular loops are
thought to establish the barrier for paracellular flow. The cell
type‐specific combination in which its 23 members are
expressed determines the “tightness” of the permeability barrier
as well as its ion selectivity. Curiously, siRNA‐mediated deple-
tion of claudin‐2 from WIF‐B cells [81] and of claudin‐3 from
the related Can10 cells [82] resulted in a switch from hepato-
cytic to monopolar organization, suggesting specific signaling
roles for claudins beyond regulating paracellular flux.
E‐cadherin is considered a key trigger of epithelial polarization
in monopolar cells [83]. This was established in so‐called
“Ca‐switch assays,” which exploit the Ca‐dependence of E‐cadherin
adhesion. In Ca‐free medium monopolar epithelial cells such as
MDCK cells do not polarize even when plated at confluence
(i.e. in the presence of only nectin‐mediated cell–cell adhesion).
Re‐addition of Ca triggers rapid, synchronized tight junction
establishment and lumen formation [84, 85]. It was unexpected,
therefore, that a hepatocytic cell line (HepG2) which failed to
target E‐cadherin and β‐catenin to the cell surface still established
functional tight junctions and bile canaliculi‐like luminal domains
(albeit with delayed kinetics), suggesting that E‐cadherin is not
absolutely essential for the establishment of hepatic polarity [86].
Conversely, increasing E‐cadherin levels in this cell line led to the
formation of a horizontal tight junction belt characteristic of
monopolar cells [87]. In MDCK cells, substitution of E‐cadherin
for an adhesion‐deficient mutant promoted lumen establishment
at cell–cell contact sites as in hepatocytes, at least transiently [88].
Collectively, these data suggest a model by which E‐cadherin‐
mediated adhesion signaling promotes monopolar organization
and antagonizes hepatocytic polarity. Consistent with this model,
immunohistochemistry of liver tissue sections shows significantly
higher E‐cadherin staining of bile ducts when compared with
adjacent hepatocytes [4]. This might be due to lower E‐cadherin
mRNA levels in hepatocytes, as suggested by recent RNA
sequencing data [89].
Even if lumen formation in hepatocyte does not require
E‐cadherin adhesion, it likely depends on other cell–cell adhe-
sion molecules. This can be concluded from cell culture studies
in which mechanical cell compaction [90] or spheroid forma-
tion [91], both conditions that maximize cell–cell contacts,
­significantly stimulated lumen formation.
The taurocholate signaling mechanism
Analysis of taurocholate signaling, the putative hepatocyte
polarization trigger, in primary cultures established that tauro-
cholate increased hepatocyte cAMP levels, which led to activa-
tion of EPAC, a GEF for small GTPases of the Rap family. Rap
signaling in turn stimulated activity of AMP‐regulated kinase
(AMPK) via its main activating kinase in the liver, LKB1 [11]
(see Chapter 38 for further details).
LKB1 was discovered as a tumor suppressor mutated in the
human Peutz–Jegher syndrome, which is characterized by
hamartomas and polyps in the gastrointestinal tract and by an
elevated cancer risk. LKB1 activity depends on association with
a pseudokinase and a scaffolding protein [92]. In some tissues,
various growth factors activate LKB1, suggesting that hepato-
cyte polarity and other functions may be regulated by hormones
and/or growth factors. Notably, in single intestinal cells LKB1
activation induces cell‐autonomous cell surface polarization,
hinting that LKB1 might initiate signaling by the self‐organiz-
ing polarity complexes mentioned above [93]. In addition, in
MDCK cells AMPK was required for tight junction formation
downstream of cell–cell adhesion [94, 95]. Tight junctional pro-
teins appear to be the main AMPK targets in its role as a polarity
kinase. They include the substrate Gα‐interacting vesicle‐asso-
ciated protein, (aka Girdin), which when phosphorylated by
AMPK re‐enforces tight junctions under energy stress [96], and
cingulin [97], the suggested AMPK substrate responsible for
leaky tight junctions observed by intravital imaging in LKB1
knockout livers in vivo [48]. In addition to tight junction defects,
LKB1 knockout livers also presented with defective targeting of
bile acid transporters to the luminal surface, which were cor-
rected by cAMP activation [25]. As can be expected from the
polarity defects, LKB1 conditional liver knockout mice suffered
cholestasis and liver injury [98].
42 THE LIVER:  POLARIZED PROTEIN TRAFFICKING
POLARIZATION REQUIRES ENERGY
Hepatocyte polarization and all of its components are energy‐
dependent and have been increasingly related to two protein
kinase, LKB1 and AMPK (see Chapter 38 for further details).
AMPK and LKB1 contribute to hepatocyte polarization by con-
trolling ATP synthesis and energy metabolism. AMPK, a serine
threonine kinase containing a catalytic α subunit and regulatory
β and γ subunits, controls energy metabolism within cells by
sensing the cellular AMP to ATP ratio. Activation of AMPK by
phosphorylation of the α subunit Thr172 decreases energy con-
sumption and increases energy production during cellular stress,
such as hypoxia, glucose deprivation, and ischemia, and has an
important role in hepatic metabolism through effects on glu-
cose, lipid and protein homeostasis and mitochondrial biogene-
sis. Long‐term effects involve regulation of the glycolytic and
lipogenic pathways.
In collagen sandwich cultured hepatocytes, the stress of iso-
lation resulted in depolarization, ATP depletion, and mitochon-
drial fragmentation [99]. Mitochondrial fusion occurred within
2 days associated with increased ATP synthesis from oxidative
phosphorylation and canalicular network formation. Subsequent
AMPK activation upregulated glucose uptake, glycolysis, and a
further increase in ATP. These studies reveal that, after stress,
hepatocytes preferentially restore polarity even at low ATP lev-
els, suggesting that polarity is a prime requirement for cellular
activity [100].
HEPATOCYTE POLARITY IS LINKED TO
THE EXPRESSION OF DIFFERENTIATION
MARKERS
When hepatocytes are cultured on rigid 2D matrices they rap-
idly lose expression of differentiation markers such as albumin
or various transporters. The degree of de‐differentiation is pro-
portional to the ECM concentration regardless of its nature.
Increasing matrix concentrations also cause increasing cell
spreading which is incompatible with hepatocyte polarization
[101–103]. Conversely, hepatocytes cultured on soft matrices,
particularly in sandwich configuration, maintain expression of
differentiation markers; they also acquire polarity [21, 22].
This correlation between increased expression of markers for
hepatocyte function and better polarization was also observed
when hepatocytes or hepatic cell lines such as HepaRG were
cultured in the absence of exogenous matrix but with extensive
cell–cell contact in spheroids [91]. Its crux is cytoskeletal
­tension, which activates a mechano‐transduction cascade in
which integrin signaling leads to FAK→RhoA/RhoK→ERK1/2
activation, which promotes proliferation and induces an EMT
[104, 105]. Importantly, activation of RhoA/RhoK in this path-
way inhibits expression of the hepatocytic master transcription
factor HNF4a, which controls a hepatocyte‐specific transcrip-
tional network that includes both functional proteins and polar-
ity determinants [106]. Forced HNF4a expression on rigid
matrix remedied loss of several, although not all, functional
markers. As previously discussed, low RhoA activity is also a
prerequisite for heptocytic polarization in cell culture models.
Thus, a RhoA–HNF4a axis might link hepatocyte polarity to
functionality.
POLARIZED PROTEIN TRAFFICKING
The vectorial activities of epithelia require that each transporter/
channel has a clearly defined apical–basal polarity, which
results from sorting events in the biosynthetic and recycling
routes at the trans‐Golgi network (TGN), common recycling
endosomes (CRE), and apical recycling endosomes (ARE)
[107] (Figures 4.4 and 4.5). Biosynthetic protein trafficking itin-
eraries in hepatocytes have been established by combining an in
vivo pulse‐chase protocol with cell fractionation [108]. This
protocol, which exploits the observation that 35S‐methionine,
when injected into the tail vein first reaches the liver and is
mostly incorporated into newly synthesized hepatocyte pro-
teins, is in fact the only in vivo approach to protein trafficking
for any mammalian epithelium. It reveals that hepatocytes target
polytopic membrane proteins, such as ABC transporters, from
the TGN directly to the bile canalicular domain, some of which
first accumulate at a Rab11a‐positive ARE pool, from which
they then cycle to the canalicular membrane [109]. In contrast,
all single membrane‐spanning and GPI‐anchored bile canalicu-
lar membrane proteins are targeted from the TGN to the basolat-
eral plasma membrane, from where they transcytose in
endosomes through the cell to the canalicular membrane [108,
110, 111]. This is different from all monopolar epithelia studied
to date, which utilize the basolateral‐to‐apical transcytotic route
for apical targeting to a lesser extent than hepatocytes. The most
poignant difference is that unlike monopolar epithelia, hepato-
cytes lack a pathway that targets soluble proteins from the TGN
to the bile canalicular domain via vesicular transport. All solu-
ble cargo that passes through the secretory pathway is (predomi-
nantly) directed into the space of Disse [112, 113].
The mechanisms that segregate apical and basolateral pro-
teins and mediate their surface domain‐specific targeting from
the TGN have been largely elucidated in the epithelial model
cell line MDCK. Work over three decades by many groups
using apical and basolateral model proteins yielded the follow-
ing model of polarized protein delivery in the direct biosynthetic
pathway (Figure 4.4): Sorting machinery at the TGN recognizes
apical and basolateral sorting signals and packages apical and
basolateral proteins into separate TGN‐derived transport carri-
ers [107]. Sorting signals in basolateral proteins have affinities
for clathrin adaptors, which mediate their clathrin‐dependent
basolateral surface targeting; in the absence of clathrin basolat-
eral proteins still reach the plasma membrane, but they are mis‐
targeted to the apical domain [114]. Based on their interaction
with different adaptors, basolateral cargoes take at least two dif-
ferent routes out of the TGN: If not captured by TGN‐localized
clathrin adaptors such as AP1A, they are targeted to transfer-
rin‐positive recycling endosomes where basolateral proteins
interact with the endosomal clathrin adaptor AP1B for basolat-
eral surface delivery. AP1A‐interacting cargo, by contrast, is
delivered to the basolateral surface without traversing recycling
compartments. While some basolateral proteins are known to
 4:  Hepatocyte Surface Polarity 43

AP pathways
(a) (b)
BL pathways
AP

MyoVb
Rab11a
AEE
ARE

AP1B (i)
MyoVb
CRE Rab11a

BC
BC ARE
TGN
AP1A
AP4
TGN
(ii)
trans-

TGN

BL BL

Figure 4.4  Targeting pathways from the TGN to the apical/canalicular and basolateral surfaces in MDCK and hepatocytes. (a) Known pathways
in MDCK cells. Apical (AP) pathways (green arrows): depending on their presence in detergent‐insoluble microdomains, raft‐dependent and raft‐
independent apical routes have been distinguished. Several membrane and secretory proteins have been shown to traverse the apical recycling
endosome (ARE) (in a MyoVb and Rab11‐dependent manner) or the apical early endosome (AEE) before reaching the apical domain. Basolateral
(BL) pathways (blue arrows): basolateral protein exiting from the TGN is clathrin mediated. Multiple pathways have been distinguished. Some
basolateral proteins reach the basolateral domain without intermediate (dependent on TGN‐localized clathrin adaptors AP1A and AP4), others
traverse the common recycling endosome (CRE), from where their basolateral targeting is mediated by clathrin adaptor AP1B. (b) Known and
hypothetical pathways in hepatic cells. Two classes of apical proteins have been distinguished: polytopic membrane proteins, such as bile acid
transporters, travel from the TGN to the apical domain directly (at least some via a Rab11 compartment); GPI‐anchored and single membrane‐span-
ning membrane proteins first reach the basolateral domain from the TGN. It is not resolved whether (i) apical and basolateral proteins are sorted
into distinct TGN‐derived transport carriers that both reach the basolateral domain, or (ii) whether apical and basolateral proteins travel in common
carriers to the basolateral domain. Since hepatocytes lack AP1B, it is unlikely that hepatocytes target their biosynthetic cargo via the CRE. Adapted
from Treyer and Müsch, Compr Physiol, 2013;3:243–87 with permission of John Wiley & Sons.

enter both pathways, others are predominantly targeted via one
or the other pathway [115, 116]. Apical sorting signals, on the
other hand, are pleomorphic, constituted by N‐glycans, GPI
anchors, and specialized transmembrane protein domains with
affinity for lipid rafts. Lipid rafts are formed by sphingolipids
and cholesterol, which cluster apical cargo and also establish
membrane domain boundaries that segregate apical and basolat-
eral cargo domains [117]. Apical proteins are targeted from the
TGN in microtubule (MT)‐dependent tubular carriers, indepen-
dently of clathrin [69]. Apical and basolateral transport carriers
also differ in their fission and fusion machineries. Thus, while
basolateral carrier scission depends on the activities of protein
kinase D and on CtBP1‐S/BARS‐induced activation of
lysophosphatidic acid acyltransferase δ [118–120], apical carri-
ers require the scission factor dynamin 2, which functions in
conjunction with branched actin filaments that assemble around
the apical carriers at the scission site [121, 122]. Basolateral
fusion is mediated by SNARE pairings containing VAMP3 and
syntaxin 4, while apical carriers fuse via tetanus‐resistant v‐
SNAREs (VAMP7,8) that pair with syntaxin 3 [44, 123–127].
Hepatocytes can be expected to utilize some of the same tar-
geting mechanisms, yet there also have to be differences to
account for heptocyte TGN‐to‐basolateral targeting of apical
proteins that in MDCK cells are targeted to the apical surface
directly. So far, only one contributing mechanism has been
clearly identified: the lack in hepatocytes of myelin and lym-
phocyte 1 (MAL1), a proteolipid tetraspanning raft‐associated
membrane protein, which in MDCK cells is critical for TGN‐
derived apical transport of proteins that depend on association
with lipid rafts for their polarized targeting [128, 129].
Expression of MAL1 in the hepatocytic line WIF‐B (which
­normally lack it) promoted vectorial delivery of GPI‐apical pro-
teins and single‐pass transmembrane proteins from the TGN to
the apical surface [130]. It was also suggested (for the HepG2
cell line) that the transcytotic mechanism is regulated by the
levels of surface E‐cadherin [86], although no mechanism for
such regulation has been proposed.
Hepatocytes also lack the basolateral cargo adaptor AP1B. In
all other epithelia that naturally fail to express AP1B, AP1B‐
dependent basolateral proteins, such as transferrin receptor,
localize at the apical domain [116]. This is because their endo-
somal sorting in both the exocytic and the recycling pathways
are compromised. However, such polarity reversal is not
observed in hepatocytes or hepatocytic cell lines. This implies
44 THE LIVER:  CYTOSKELETAL MICROFILAMENT AND MICROTUBULAR SYSTEMS IN TRAFFICKING
BL
LYS
AEE
MyoVb
Rab11a
BC
ARE SAC
? maintenance of polarity.
LYS BEE BEE
BL proteins AP proteins
Figure 4.5  Targeting pathways from the apical and basolateral plasma
membranes in hepatocytes. Pathways for apical (AP) proteins (green
arrows): Upon endocytosis from the canalicular domain proteins are
sorted in the apical early endosome (AEE) into either the late endoso-
mal/lysosomal pathway or for recycling to the apical recycling endo-
some (ARE). Apical proteins that have reached the basolateral surface
from the TGN are internalized via clathrin‐coated vesicles (non‐raft) or
in a clathrin‐independent, flotillin‐dependent manner (raft) and reach
the subapical endosome (SAC) via basolateral early endosomes (BEEs).
From the SAC they are targeted to the apical surface via the ARE.
Basolateral (BL) resident proteins (blue arrows) undergo fast recycling
from the BEE directly or via recycling endosomes. Whether they min-
gle with apical‐directed cargo in a common recycling endosome (CRE)
as described for MDCK cells is still debated. While the SAC has been
described by some as the equivalent of the MDCK CRE, others have
suggested that BL and AP proteins are segregated before apical‐directed
cargo reaches the SAC. Similar to the AEE, the BEE also sorts proteins
for degradation to the lysosome (LYS). Adapted from Treyer and
Müsch, Compr Physiol, 2013;3:243–87 with permission of John
Wiley & Sons.
that hepatocytes organize their endosomal recycling itineraries
differently from MDCK cells (Figure 4.5). In the MDCK model,
endocytosed cargo can either undergo fast recycling to their
membrane of origin from early endosomes or, alternatively,
recycle from a CRE, where cargo from both surfaces mix and is
sorted into distinct apical‐ or basolateral‐directed recycling car-
riers [131–133]. The CRE is also where basolateral‐to‐apical
transyctotic cargo is segregated from basolateral recycling cargo
[132, 134]. Apical carriers exiting the CRE either mature into or
fuse with the AREs beneath the apical surface, while basolateral
cargo returns to the surface directly [135, 136]. Studies in the
hepatocytic cell lines WIF‐B and HepG2 as well as electron
microscopy 3D reconstruction of the hepatocyte endosomal sys-
tem have indeed suggested differences in endosome organiza-
tion compared to MDCK cells. In particular, a subapical
compartment (SAC) has been described, which according to dif-
ferent schools either mediates apical and basolateral cargo sort-
ing equivalent to the MDCK CRE or represents a unique
compartment of the basolateral‐to‐apical transcytotic pathway
[16, 111, 137–139]. Regardless of the model and cell type, all
epithelial cells accumulate subapical membranes that appear to
be bottlenecks for protein delivery to the apical surface. They
are enriched in Rab GTPases, including Rab11a and its effector,
the actin‐associated molecular motor myosin Vb, as well as the
Rab11 adaptor proteins Fip1 and Fip2 [140–143]. Inhibition of
Rab11a or myosin Vb indeed prevented polarization of WIF‐B
cells and primary hepatocytes [144] and, when introduced into
polarized cells, prompted depolarization and internalization of
apical proteins [145]. These observations highlight the impor-
tance of apical membrane traffic for the establishment and
The Rab11‐positive endosomes also serve as “holding cells”
for hepatocyte ABC transporters, the majority of which are
maintained intracellularly rather than at the bile canalicular
membrane [25, 146]. This intracellular pool is mobilized by bile
acids circulating in the enterohepatic circuit, and by postprandi-
ally secreted peptide hormones, which increase cAMP produc-
tion in hepatocytes to cope with the increased demand for bile
acid secretion. Taurocholate and cAMP activate distinct signal-
ing pathways to mobilize ABC transporters to the canalicular
domain. Regulation of these two responses differs. Increase in
cAMP concentration results in PKA‐mediated stimulation of
phosphoinositide 3‐kinase but not taurocholate‐stimulated incor-
poration of ABCB11 into the canalicular membrane [147–150].
Bulk endocytosis from the apical surface of polarized epithe-
lia occurs at much lower rate than endocytosis from the basolat-
eral domain [70]. This is likely due to the extensive direct and
indirect linkage of apical membrane proteins to actin filaments,
which makes it hard for the endocytic machinery to deform the
membrane. On the other hand, a branched actin network at the
base of a nascent endocytic vesicle promotes clathrin‐mediated
endocytosis when linked to the scission factor dynamin via the
actin and dynamin‐binding protein cortactin [151, 152]. Apical
bile acid transporters BSEP, MDR1, and MDR2 contain in their
cytoplasmic domains a binding site for HCLS1‐associated pro-
tein X‐1 (HAX‐1) [153], which also binds cortactin and hence
could recruit these ABC transporters into productive endocytic
pits. Consistent with this hypothesis, HAX‐1 depletion increased
BSEP accumulation at the apical domain [153]. Together, their
facilitated endocytosis and regulated re‐insertion from an endo-
cytic pool allows for the highly dynamic behavior of ABC trans-
porters at the bile canalicular domain.
CYTOSKELETAL MICROFILAMENT
AND MICROTUBULAR SYSTEMS IN
TRAFFICKING(SEE CHAPTER 3)
Proper endosomal trafficking and recycling of proteins to all
plasma membrane domains requires an intact actin and microtu-
bular cytoskeletal system. In particular, plus‐end dynamic
microtubules marked by CLIP170 and EB1 mediate trafficking
of secreted and canalicular proteins. Newly synthesized
ABCB11, the canalicular bile acid transporter, and other canali-
cular ABC transporters traffic in post Golgi vesicles from the
TGN along microtubules. However, dynamic microtubules do
not attach to the canalicular membrane and their cargo
endosomes are transferred to the pericanalicular actin system
 4:  Hepatocyte Surface Polarity 45
(Figure 4.5). The complex mechanism for cargo transfer is not
known; however, microtubules become associated with actin
through a pericanalicular actin‐binding complex containing IQ
Gap, APC, Hax‐1, and cortactin proteins. Live cell imaging
studies reveal that selective plasma membrane localization of
transporter proteins is predominantly due to the localization of
specific docking proteins. In polarized WIF‐B cells, ABCB11
and ABCB1 traffic along microtubules throughout the cell but
only attach to specific sites on the canalicular membrane. The
docking site has been proposed to be syntaxin 3 that facilitates
fusion of protein‐sorting vesicles with the inner leaflet of the
canalicular membrane from which they diffuse until reaching
the tight junction which limits presence to the canalicular
domain. The actin‐binding protein radixin links some cargo
molecules, such as ABCC2 and ABCB11, to the pericanalicular
actin system. Radixin knockout mice manifest impaired ABCC2
and ABCB11 localization to the canalicular domain which
becomes progressively devoid of microvilli, resulting in hepato-
cyte injury.
Formin controls the assembly and disassembly of short actin
filaments which are involved in endosomal transport. In the
HepG2 hepatoma cell line INF2, CDC42 and the transmem-
brane protein MAL2 are required for transcytotic trafficking of
canalicular membrane proteins.
HEPATOCYTE POLARITY AND CELL
DIVISION
Although mature polarized hepatocytes are quiescent, they can,
under conditions of severe injury, re‐enter the cell cycle and pro-
liferate [154]. Hepatocytes also actively divide during postnatal
development while their bile canaliclar network matures. How
proliferating epithelial cells orient their mitotic spindle and
the cleavage furrow, which always forms perpendicular to the
­spindle pole axis, is of critical importance for both tissue and
cellular organization [155]. Monopolar epithelial cells orient
their metaphase spindle parallel to the basement membrane
[156, 157]. This ensures that the cleavage furrow bisects their
luminal domain, yielding two identical daughters that both
remain in the epithelial plane, thereby ensuring that the epithe-
lium remains monolayered. Critical to this outcome are cortical
cues positioned at equal distance from the basal surface at oppo-
site lateral domains. The cues consist of an evolutionarily con-
served protein complex in which the α subunit of a trimeric
G‐protein provides the cortical anchor and the minus end‐
directed microtubule motor dynein binds the plus ends of astral
microtubules. In metaphase these complexes on opposite mem-
brane domains capture one of each set of astral spindle microtu-
bules, thereby aligning the metaphase spindle parallel to the
basal domain [158, 159]. Unlike monopolar cells, hepatocytes
rarely bisect their luminal domain in cytokinesis [160]. This
preserves canalicular lumen organization and prevents the gen-
eration of acini. Remarkably, molecular analysis in the WIF‐B
and HepG2 culture models suggested that the different cell divi-
sion outcomes in monopolar and hepatocytic cells occur even
though both cell types use similar spindle orientation mecha-
nism, that is, both place their cortical spindle captures cues
(a) Monopolar Hepatocytic
(b)
Figure 4.6  Metaphase spindle orientation (a) and division outcome
(b) in cultured monopolar and heptocytic cells. Monopolar epithelia
orient their metaphase spindle parallel to their basal surface when
astral microtubules bind to cortical cues (green) below the apical sur-
face (red) (a). The cleavage furrow, which is established perpendicular
to the spindle pole axis, bisects the apical surface and yields two iden-
tical daughters, each attached to the substratum (b). In hepatocytes the
spindle attaches adjacent to two different lumina. The cleavage furrow
therefore does not bisect the luminal domains. In cultured WIF‐B and
HepG2 cells, a stereotypic spindle tilt with respect to the substrate con-
tacting basal domain causes one daughter cells to be removed from the
substratum. Whether this is relevant in vivo remains to be determined.
Adapted from [64].
adjacent to their luminal surfaces [64, 161] (Figure 4.6). Lumen
geometry and the presence of at least two luminal domains per
hepatocyte likely cause each astral microtubule set to attach
next to a different lumen, thereby preventing bisection of either
lumen. By contrast, in monopolar cells, the two astral microtu-
bule attachment sites flank the same (single) luminal domain,
resulting in its bisection. In an additional difference to monopo-
lar cells, hepatocytic cell lines can accommodate the metaphase
spindle only in a quasi‐diagonal position, resulting in a stereo-
typic spindle tilt with respect to their basal domains, which
yields a division out of the monolayer and results in the bilayer-
ing observed in HepG2 and WIF‐B cell cultures [65]. Live cell
imaging of mammalian liver tissue [48, 162] should make it
feasible to test the relevance of this division mode in vivo.
LIVER DISEASE AND POLARITY
As detailed in Chapter 29, mutations in MYO5B encoding myo-
sin Vb and Rab11b, Rab25, and Rab8, cause microvillus inclu-
sion disease (MVID), in which malabsorption results from the
absence of the intestinal brush border. Other patients manifest
cholestasis and progressive liver disease. Mouse Rab8 condi-
tional knockouts mimic MVID. Mutations affecting syntaxin 3,
an apical membrane SNARE (family of membrane proteins that
ensures fusion between opposing membranes), suggesting that
myosin 5b, Rab8, and syntaxin 3 may be involved in the same
46 THE LIVER: REFERENCES
trafficking pathway. Discovery of loss‐of‐function mutations in
genes encoding recycling endosome‐associated proteins such as
myosin 5b in MVID, VPS33B and VIPAR in arthrogryposis,
renal dysfunction, and cholestasis syndrome (ARC), also sup-
ports the importance of the RE in establishment and mainte-
nance of hepatocyte polarity.
REFERENCES
1. Arias, I.M. et al. The Liver: Biology and Pathobiology. 5th edn (eds. I.M.
Arias et al.), Wiley‐Blackwell, Chichester, 2009.
2. Si‐Tayeb, K., Lemaigre, F.P., and Duncan, S.A. Organogenesis and develop-
ment of the liver. Dev Cell, 2010;18(2):175–89.
3. Gordillo, M., Evans, T., and Gouon‐Evans, V. Orchestrating liver develop-
ment. Development, 2015;142(12):2094–108.
4. Antoniou, A. et al. Intrahepatic bile ducts develop according to a new mode
of tubulogenesis regulated by the transcription factor SOX9.
Gastroenterology, 2009;136(7):2325–33.
5. Zhang, L. et  al. The stem cell niche of human livers: symmetry between
development and regeneration. Hepatology, 2008;48(5):1598–607.
6. Carpentier, R. et al. Embryonic ductal plate cells give rise to cholangiocytes,
periportal hepatocytes, and adult liver progenitor cells. Gastroenterology,
2011;141(4):1432–8, 1438.e1–4.
7. Wauthier, E. et al. Hepatic stem cells and hepatoblasts: identification, isola-
tion, and ex vivo maintenance. Methods Cell Biol, 2008;86:137–225.
8. Gallin, W.J. Development and maintenance of bile canaliculi in vitro and in
vivo. Microsc Res Tech, 1997;39(5):406–12.
9. Kanamura, S., Kanai, K., and Watanabe, J. Fine structure and function of hepat-
ocytes during development. J Electron Microsc Tech, 1990;14(2):92–105.
10. Little, J.M. et al. Taurocholate pool size and distribution in the fetal rat. J
Clin Invest, 1979;63(5):1042–9.
11. Fu, D. et al. Bile acid stimulates hepatocyte polarization through a cAMP‐
Epac‐MEK‐LKB1‐AMPK pathway. Proc Natl Acad Sci U S A,
2011;108(4):1403–8.
12. Tanimizu, N. et al. Intrahepatic bile ducts are developed through formation
of homogeneous continuous luminal network and its dynamic rearrangement
in mice. Hepatology, 2016;64(1):175–88.
13. Lo, B. et  al. Lkb1 regulates organogenesis and early oncogenesis
along  AMPK‐dependent and ‐independent pathways. J Cell Biol,
2012;199(7):1117–30.
14. Darlington, G.J., Kelly, J.H., and Buffone, G.J. Growth and hepatospecific
gene expression of human hepatoma cells in a defined medium. In Vitro Cell
Dev Biol, 1987;23(5):349–54.
15. Decaens, C. et al. Establishment of hepatic cell polarity in the rat hepatoma‐
human fibroblast hybrid WIF‐B9. A biphasic phenomenon going from a sim-
ple epithelial polarized phenotype to an hepatic polarized one. J Cell Sci,
1996;109(Pt 6):1623–35.
16. Ihrke, G. et  al. Apical plasma membrane proteins and endolyn‐78 travel
through a subapical compartment in polarized WIF‐B hepatocytes. J Cell
Biol, 1998;141(1):115–33.
17. Kelly, J.H. and Darlington, G.J. Modulation of the liver specific phenotype
in the human hepatoblastoma line Hep G2. In Vitro Cell Dev Biol,
1989;25(2):217–22.
18. Peng, X. et al. How to induce non‐polarized cells of hepatic origin to express
typical hepatocyte polarity: generation of new highly polarized cell models
with developed and functional bile canaliculi. Cell Tissue Res,
2006;323(2):233–43.
19. van der Wouden, J.M., van IJzendoorn, S.C., and Hoekstra, D. Oncostatin M
regulates membrane traffic and stimulates bile canalicular membrane bio-
genesis in HepG2 cells. EMBO J, 2002;21(23):6409–18.
20. Wakabayashi, Y., Lippincott‐Schwartz, J., and Arias, I.M. Intracellular traf-
ficking of bile salt export pump (ABCB11) in polarized hepatic cells: consti-
tutive cycling between the canalicular membrane and rab11‐positive
endosomes. Mol Biol Cell, 2004;15(7):3485–96.
21. Dunn, J.C., Tompkins, R.G., and Yarmush, M.L. Long‐term in vitro function
of adult hepatocytes in a collagen sandwich configuration. Biotechnol Prog,
1991;7(3):237–45.
22. Dunn, J.C. et  al. Hepatocyte function and extracellular matrix geometry:
long‐term culture in a sandwich configuration. FASEB J, 1989;3(2):174–7.
23. LeCluyse, E.L., Audus, K.L., and Hochman, J.H. Formation of extensive
canalicular networks by rat hepatocytes cultured in collagen‐sandwich con-
figuration. Am J Physiol, 1994;266(6 Pt 1):C1764–74.
24. Michalopoulos, G.K. et al. Comparative analysis of mitogenic and morpho-
genic effects of HGF and EGF on rat and human hepatocytes maintained in
collagen gels. J Cell Physiol, 1993;156(3):443–52.
25. Homolya, L. et al. LKB1/AMPK and PKA control ABCB11 trafficking and
polarization in hepatocytes. PLoS One, 2014;9(3):e91921.
26. Grubman, S.A. et al. Regulation of intracellular pH by immortalized human
intrahepatic biliary epithelial cell lines. Am J Physiol, 1994;266(6 Pt
1):G1060–70.
27. Vroman, B. and LaRusso, N.F. Development and characterization of polar-
ized primary cultures of rat intrahepatic bile duct epithelial cells. Lab Invest,
1996;74(1):303–13.
28. Salter, K.D. et al. Modified culture conditions enhance expression of differ-
entiated phenotypic properties of normal rat cholangiocytes. Lab Invest,
2000;80(11):1775–8.
29. Gradilone, S.A. et al. Cholangiocyte cilia express TRPV4 and detect changes
in luminal tonicity inducing bicarbonate secretion. Proc Natl Acad Sci U S A,
2007;104(48):19138–43.
30. Lazaridis, K.N. et  al. Kinetic and molecular identification of sodium‐
dependent glucose transporter in normal rat cholangiocytes. Am J Physiol,
1997;272(5 Pt 1):G1168–74.
31. Spirli, C. et al. Functional polarity of Na+/H+ and Cl−/HCO3− exchangers in a
rat cholangiocyte cell line. Am J Physiol, 1998;275(6 Pt 1):G1236–45.
32. Zsembery, A. et al. Purinergic regulation of acid/base transport in human and
rat biliary epithelial cell lines. Hepatology, 1998;28(4):914–20.
33. Miyajima, A., Tanaka, M., and Itoh, T. Stem/progenitor cells in liver devel-
opment, homeostasis, regeneration, and reprogramming. Cell Stem Cell,
2014;14(5):561–74.
34. Reid, L.M. Stem/progenitor cells and reprogramming (plasticity) mecha-
nisms in liver, biliary tree, and pancreas. Hepatology, 2016;64(1):4–7.
35. Tanimizu, N. et al. Long‐term culture of hepatic progenitors derived from
mouse Dlk+ hepatoblasts. J Cell Sci, 2004;117(Pt 26):6425–34.
36. Wang, Y. et al. Paracrine signals from mesenchymal cell populations govern
the expansion and differentiation of human hepatic stem cells to adult liver
fates. Hepatology, 2010;52(4):1443–54.
37. Huch, M. et al. Long‐term culture of genome‐stable bipotent stem cells from
adult human liver. Cell, 2015;160(1–2):299–312.
38. Tanimizu, N., Miyajima, A., and Mostov, K.E. Liver progenitor cells develop
cholangiocyte‐type epithelial polarity in three‐dimensional culture. Mol Biol
Cell, 2007;18(4):1472–9.
39. Tanimizu, N. et al. Hepatic biliary epithelial cells acquire epithelial integrity
but lose plasticity to differentiate into hepatocytes in vitro during develop-
ment. J Cell Sci, 2013;126(Pt 22):5239–46.
40. Guillouzo, A. et al. The human hepatoma HepaRG cells: a highly differenti-
ated model for studies of liver metabolism and toxicity of xenobiotics. Chem
Biol Interact, 2007;168(1):66–73.
41. Gunness, P. et al. 3D organotypic cultures of human HepaRG cells: a tool for
in vitro toxicity studies. Toxicol Sci, 2013;133(1):67–78.
42. Dianat, N. et  al. Generation of functional cholangiocyte‐like cells from
human pluripotent stem cells and HepaRG cells. Hepatology,
2014;60(2):700–14.
43. Nakatani, T. et al. Assessment of efficiency and safety of adenovirus medi-
ated gene transfer into normal and damaged murine livers. Gut,
2000;47(4):563–70.
44. Shayakhmetov, D.M. et al. Analysis of adenovirus sequestration in the liver,
transduction of hepatic cells, and innate toxicity after injection of fiber‐mod-
ified vectors. J Virol, 2004;78(10):5368–81.
45. Zhang, G., Budker, V., and Wolff, J.A. High levels of foreign gene expression
in hepatocytes after tail vein injections of naked plasmid DNA. Hum Gene
Ther, 1999;10(10):1735–7.
46. Eggenhofer, E. et al. High volume naked DNA tail‐vein injection restores
liver function in Fah‐knock out mice. J Gastroenterol Hepatol,
2010;25(5):1002–8.
47. Marques, P.E. et al. Imaging liver biology in vivo using conventional confo-
cal microscopy. Nat Protoc, 2015;10(2):258–68.
48. Porat‐Shliom, N. et al. Liver kinase B1 regulates hepatocellular tight junc-
tion distribution and function in vivo. Hepatology, 2016;64(4):1317–29.
 4:  Hepatocyte Surface Polarity 47
49. Tepass, U. The apical polarity protein network in Drosophila epithelial cells:
regulation of polarity, junctions, morphogenesis, cell growth, and survival.
Annu Rev Cell Dev Biol, 2012;28:655–85.
50. Thompson, B.J. Cell polarity: models and mechanisms from yeast, worms
and flies. Development, 2013;140(1):13–21.
51. Rodriguez‐Boulan, E. and Macara, I.G. Organization and execution of the
epithelial polarity programme. Nat Rev Mol Cell Biol, 2014;15(4):225–42.
52. Takaki, Y. et al. Dynamic changes in protein components of the tight junction
during liver regeneration. Cell Tissue Res, 2001;305(3):399–409.
53. Kallay, L.M. et  al. Scribble associates with two polarity proteins, Lg12
and  Vang12, via distinct molecular domains. J Cell Biochem,
2006;99(2):647–64.
54. Terada, T. and Nakanuma, Y. Expression of tenascin, type IV collagen and
laminin during human intrahepatic bile duct development and in intrahepatic
cholangiocarcinoma. Histopathology, 1994;25(2):143–50.
55. Shiojiri, N. and Sugiyama, Y. Immunolocalization of extracellular matrix
components and integrins during mouse liver development. Hepatology,
2004;40(2):346–55.
56. Tanimizu, N. et  al. alpha1‐ and alpha5‐containing laminins regulate the
development of bile ducts via beta1 integrin signals. J Biol Chem,
2012;287(34):28586–97.
57. Schaffner, F. and Poper, H. Capillarization of hepatic sinusoids in man.
Gastroenterology, 1963;44:239–42.
58. Yang, L. et al. A single‐cell transcriptomic analysis reveals precise pathways
and regulatory mechanisms underlying hepatoblast differentiation.
Hepatology, 2017;66(5):1387–401.
59. Martinez‐Hernandez, A. and Amenta, P.S. The hepatic extracellular matrix.
I. Components and distribution in normal liver. Virchows Arch A Pathol Anat
Histopathol, 1993;423(1):1–11.
60. Bataller, R. and Brenner, D.A. Liver fibrosis. J Clin Invest,
2005;115(2):209–18.
61. Wells, R.G. The role of matrix stiffness in regulating cell behavior.
Hepatology, 2008;47(4):1394–400.
62. Cohen, D. et al. Mammalian PAR‐1 determines epithelial lumen polarity by
organizing the microtubule cytoskeleton. J Cell Biol, 2004;164(5):717–27.
63. Cohen, D. et al. The serine/threonine kinase Par1b regulates epithelial lumen
polarity via IRSp53‐mediated cell‐ECM signaling. J Cell Biol,
2011;192(3):525–40.
64. Lazaro‐Dieguez, F. et al. Par1b links lumen polarity with LGN‐NuMA posi-
tioning for distinct epithelial cell division phenotypes. J Cell Biol,
2013;203(2):251–64.
65. Lazaro‐Dieguez, F. and Musch, A. Cell‐cell adhesion accounts for the differ-
ent orientation of columnar and hepatocyte cell division. J Cell Biol,
2017;216(11):3847–59.
66. Bryant, D.M. et al. A molecular network for de novo generation of the apical
surface and lumen. Nat Cell Biol, 2010;12(11):1035–45.
67. Galvez‐Santisteban, M. et al. Synaptotagmin‐like proteins control the forma-
tion of a single apical membrane domain in epithelial cells. Nat Cell Biol,
2012;14(8):838–49.
68. Tuma, P. and Hubbard, A.L. Transcytosis: crossing cellular barriers. Physiol
Rev, 2003;83(3):871–932.
69. Weisz, O.A. and Rodriguez‐Boulan, E. Apical trafficking in epithelial cells:
signals, clusters and motors. J Cell Sci, 2009;122(Pt 23):4253–66.
70. Boulant, S. et al. Actin dynamics counteract membrane tension during clath-
rin‐mediated endocytosis. Nat Cell Biol, 2011;13(9):1124–31.
71. Szalinski, C.M. et al. PIP5KIbeta selectively modulates apical endocytosis
in polarized renal epithelial cells. PLoS One, 2013;8(1):e53790.
72. Sato, T. et al. Regulation of the assembly and adhesion activity of E‐cadherin
by nectin and afadin for the formation of adherens junctions in Madin‐Darby
canine kidney cells. J Biol Chem, 2006;281(8):5288–99.
73. Takai, Y. et  al. Nectins and nectin‐like molecules: roles in cell adhesion,
migration, and polarization. Cancer Sci, 2003;94(8):655–67.
74. Tsuchiya, B. et al. Differential expression of N‐cadherin and E‐cadherin in
normal human tissues. Arch Histol Cytol, 2006;69(2):135–45.
75. Straub, B.K. et al. E‐N‐cadherin heterodimers define novel adherens junc-
tions connecting endoderm‐derived cells. J Cell Biol, 2011;195(5):873–87.
76. Wheelock, M.J. et  al. Cadherin switching. J Cell Sci, 2008;121(Pt
6):727–35.
77. Rajasekaran, A.K. et al. Catenins and zonula occludens‐1 form a complex
during early stages in the assembly of tight junctions. J Cell Biol,
1996;132(3):451–63.
78. Asakura, T. et al. Similar and differential behaviour between the nectin‐afa-
din‐ponsin and cadherin‐catenin systems during the formation and disrup-
tion of the polarized junctional alignment in epithelial cells. Genes Cells,
1999;4(10):573–81.
79. Braiterman, L.T. et al. JAM‐A is both essential and inhibitory to develop-
ment of hepatic polarity in WIF‐B cells. Am J Physiol Gastrointest Liver
Physiol, 2008;294(2):G576–88.
80. Rehder, D. et al. Junctional adhesion molecule‐a participates in the forma-
tion of apico‐basal polarity through different domains. Exp Cell Res,
2006;312(17):3389–403.
81. Son, S. et al. Knockdown of tight junction protein claudin‐2 prevents bile
canalicular formation in WIF‐B9 cells. Histochem Cell Biol,
2009;131(3):411–24.
82. Grosse, B. et al. Build them up and break them down: tight junctions of cell
lines expressing typical hepatocyte polarity with a varied repertoire of clau-
dins. Tissue Barriers, 2013;1(4):e25210.
83. Takeichi, M. Cadherin cell adhesion receptors as a morphogenetic regula-
tor. Science, 1991;251(5000):1451–5.
84. Gonzalez‐Mariscal, L., Chavez de Ramirez, B., and Cereijido, M. Tight
junction formation in cultured epithelial cells (MDCK). J Membr Biol,
1985;86(2):113–25.
85. Vega‐Salas, D.E., Salas, P.J. and Rodriguez‐Boulan, E. Exocytosis of vacu-
olar apical compartment (VAC): a cell‐cell contact controlled mechanism
for the establishment of the apical plasma membrane domain in epithelial
cells. J Cell Biol, 1988;107(5):1717–28.
86. Theard, D. et al. Cell polarity development and protein trafficking in hepat-
ocytes lacking E‐cadherin/beta‐catenin‐based adherens junctions. Mol Biol
Cell, 2007;18(6):2313–21.
87. Konopka, G. et al. Junctional adhesion molecule‐A is critical for the forma-
tion of pseudocanaliculi and modulates E‐cadherin expression in hepatic
cells. J Biol Chem, 2007;282(38):28137–48.
88. Cohen, D., Tian, Y., and Musch, A. Par1b promotes hepatic‐type lumen
polarity in Madin Darby canine kidney cells via myosin II‐ and E‐cadherin‐
dependent signaling. Mol Biol Cell, 2007;18(6):2203–15.
89. Su, X. et al. Single‐cell RNA‐Seq analysis reveals dynamic trajectories dur-
ing mouse liver development. BMC Genomics, 2017;18(1):946.
90. Wang, Y. et al. Mechanical compaction directly modulates the dynamics of
bile canaliculi formation. Integr Biol (Camb), 2013;5(2):390–401.
91. Godoy, P. et al. Recent advances in 2D and 3D in vitro systems using pri-
mary hepatocytes, alternative hepatocyte sources and non‐parenchymal
liver cells and their use in investigating mechanisms of hepatotoxicity, cell
signaling and ADME. Arch Toxicol, 2013;87(8):1315–530.
92. Alessi, D.R., Sakamoto, K., and Bayascas, J.R. LKB1‐dependent signaling
pathways. Annu Rev Biochem, 2006;75:137–63.
93. Baas, A.F. et al. Complete polarization of single intestinal epithelial cells
upon activation of LKB1 by STRAD. Cell, 2004;116(3):457–66.
94. Zhang, L. et al. AMP‐activated protein kinase regulates the assembly of epi-
thelial tight junctions. Proc Natl Acad Sci U S A, 2006;103(46):17272–7.
95. Zheng, B. and Cantley, L.C. Regulation of epithelial tight junction assem-
bly and disassembly by AMP‐activated protein kinase. Proc Natl Acad Sci
U S A, 2007;104(3):819–22.
96. Aznar, N. et al. AMP‐activated protein kinase fortifies epithelial tight junc-
tions during energetic stress via its effector GIV/Girdin. Elife, 2016;5.
97. Yano, T. et  al. The association of microtubules with tight junctions is
promoted by cingulin phosphorylation by AMPK. J Cell Biol, 2013;
­
203(4):605–14.
98. Woods, A. et al. LKB1 is required for hepatic bile acid transport and cana-
licular membrane integrity in mice. Biochem J, 2011;434(1):49–60.
99. Fu, D. et al. Coordinated elevation of mitochondrial oxidative phosphoryla-
tion and autophagy help drive hepatocyte polarization. Proc Natl Acad Sci
U S A, 2013;110(18):7288–93.
100. Moghe, P.V. et  al. Culture matrix configuration and composition in the
maintenance of hepatocyte polarity and function. Biomaterials,
1996;17(3):373–85.
101. Mooney, D. et al. Switching from differentiation to growth in hepatocytes:
control by extracellular matrix. J Cell Physiol, 1992;151(3):497–505.
102. Ezzell, R.M. et al. Effect of collagen gel configuration on the cytoskeleton
in cultured rat hepatocytes. Exp Cell Res, 1993;208(2):442–52.
103. Berthiaume, F. et al. Effect of extracellular matrix topology on cell struc-
ture, function, and physiological responsiveness: hepatocytes cultured in a
sandwich configuration. FASEB J, 1996;10(13):1471–84.
48 THE LIVER: REFERENCES
104. Fassett, J., Tobolt, D., and Hansen, L.K. Type I collagen structure regulates
cell morphology and EGF signaling in primary rat hepatocytes through
cAMP‐dependent protein kinase A. Mol Biol Cell, 2006;17(1):345–56.
105. Godoy, P. et al. Extracellular matrix modulates sensitivity of hepatocytes to
fibroblastoid dedifferentiation and transforming growth factor beta‐induced
apoptosis. Hepatology, 2009;49(6):2031–43.
106. Desai, S.S. et al. Physiological ranges of matrix rigidity modulate primary
mouse hepatocyte function in part through hepatocyte nuclear factor 4
alpha. Hepatology, 2016;64(1):261–75.
107. Rodriguez‐Boulan, E., Kreitzer, G., and Musch, A. Organization of vesicu-
lar trafficking in epithelia. Nat Rev Mol Cell Biol, 2005;6(3):233–47.
108. Bartles, J.R. et al. Biogenesis of the rat hepatocyte plasma membrane in
vivo: comparison of the pathways taken by apical and basolateral proteins
using subcellular fractionation. J Cell Biol, 1987;105(3):1241–51.
109. Kipp, H. and Arias, I.M. Newly synthesized canalicular ABC transporters
are directly targeted from the Golgi to the hepatocyte apical domain in rat
liver. J Biol Chem, 2000;275(21):15917–25.
110. Schell, M.J. et al. 5′nucleotidase is sorted to the apical domain of hepato-
cytes via an indirect route. J Cell Biol, 1992;119(5):1173–82.
111. Barr, V.A., Scott, L.J., and Hubbard, A.L. Immunoadsorption of hepatic
vesicles carrying newly synthesized dipeptidyl peptidase IV and polymeric
IgA receptor. J Biol Chem, 1995;270(46):27834–44.
112. Saucan, L. and Palade, G.E. Differential colchicine effects on the transport
of membrane and secretory proteins in rat hepatocytes in vivo: bipolar
secretion of albumin. Hepatology, 1992;15(4):714–21.
113. Bastaki, M. et al. Absence of direct delivery for single transmembrane api-
cal proteins or their “Secretory” forms in polarized hepatic cells. Mol Biol
Cell, 2002;13(1):225–37.
114. Deborde, S. et al. Clathrin is a key regulator of basolateral polarity. Nature,
2008;452(7188):719–23.
115. Gravotta, D. et al. The clathrin adaptor AP‐1A mediates basolateral polar-
ity. Dev Cell, 2012;22(4):811–23.
116. Folsch, H. Role of the epithelial cell‐specific clathrin adaptor complex
AP‐1B in cell polarity. Cell Logist, 2015;5(2):e1074331.
117. Schuck, S. and Simons, K. Polarized sorting in epithelial cells: raft cluster-
ing and the biogenesis of the apical membrane. J Cell Sci, 2004;117(Pt
25):5955–64.
118. Yeaman, C. et al. Protein kinase D regulates basolateral membrane protein
exit from trans‐Golgi network. Nat Cell Biol, 2004;6(2):106–12.
119. Valente, C., Luini, A., and Corda, D. Components of the CtBP1/BARS‐
dependent fission machinery. Histochem Cell Biol, 2013;140(4):407–21.
120. Pagliuso, A. et  al. Golgi membrane fission requires the CtBP1‐S/BARS‐
induced activation of lysophosphatidic acid acyltransferase delta. Nat
Commun, 2016;7:12148.
121. Kreitzer, G. et al. Kinesin and dynamin are required for post‐Golgi trans-
port of a plasma‐membrane protein. Nat Cell Biol, 2000;2(2):125–7.
122.. Salvarezza, S.B. et  al. LIM kinase 1 and cofilin regulate actin filament
population required for dynamin‐dependent apical carrier fission from the
trans‐Golgi network. Mol Biol Cell, 2009;20(1):438–51.
123. Breuza, L., Fransen, J., and Le Bivic, A. Transport and function of syntaxin
3 in human epithelial intestinal cells. Am J Physiol Cell Physiol,
2000;279(4):C1239–48.
124. Ikonen, E. et  al. Different requirements for NSF, SNAP, and Rab pro-
teins  in apical and basolateral transport in MDCK cells. Cell,
1995;81(4):571–80.
125. Low, S.H. et al. Differential localization of syntaxin isoforms in polarized
Madin‐Darby canine kidney cells. Mol Biol Cell, 1996;7(12):2007–18.
126. Pocard, T. et  al. Distinct v‐SNAREs regulate direct and indirect apical
delivery in polarized epithelial cells. J Cell Sci, 2007;120(Pt 18):3309–20.
127. Reales, E. et al. Basolateral sorting of syntaxin 4 is dependent on its N‐ter-
minal domain and the AP1B clathrin adaptor, and required for the epithelial
cell polarity. PLoS One, 2011;6(6):e21181.
128. Puertollano, R. et al. The MAL proteolipid is necessary for normal apical
transport and accurate sorting of the influenza virus hemagglutinin in
Madin‐Darby canine kidney cells. J Cell Biol, 1999;145(1):141–51.
129. Cheong, K.H. et al. VIP17/MAL, a lipid raft‐associated protein, is involved
in apical transport in MDCK cells. Proc Natl Acad Sci U S A,
1999;96(11):6241–8.
130. Ramnarayanan, S.P. et al. Exogenous MAL reroutes selected hepatic apical
proteins into the direct pathway in WIF‐B cells. Mol Biol Cell,
2007;18(7):2707–15.
131. Sheff, D.R. et  al. The receptor recycling pathway contains two distinct
populations of early endosomes with different sorting functions. J Cell
Biol, 1999;145(1):123–39.
132. Wang, E. et al. Apical and basolateral endocytic pathways of MDCK cells
meet in acidic common endosomes distinct from a nearly‐neutral apical
recycling endosome. Traffic, 2000;1(6):480–93.
133. Maxfield, F.R. and McGraw, T.E. Endocytic recycling. Nat Rev Mol Cell
Biol, 2004;5(2):121–32.
134. Odorizzi, G. et al. Apical and basolateral endosomes of MDCK cells are
interconnected and contain a polarized sorting mechanism. J Cell Biol,
1996;135(1):139–52.
135. Apodaca, G., Katz, L.A., and Mostov, K.E. Receptor‐mediated transcytosis
of IgA in MDCK cells is via apical recycling endosomes. J Cell Biol,
1994;125(1):67–86.
136. Barroso, M. and Sztul, E.S. Basolateral to apical transcytosis in polarized
cells is indirect and involves BFA and trimeric G protein sensitive passage
through the apical endosome. J Cell Biol, 1994;124(1–2):83–100.
137. Rahner, C., Stieger, B., and Landmann, L. Apical endocytosis in rat hepato-
cytes In situ involves clathrin, traverses a subapical compartment, and leads
to lysosomes. Gastroenterology, 2000;119(6):1692–707.
138. Marbet, P. et  al. Quantitative microscopy reveals 3D organization and
kinetics of endocytosis in rat hepatocytes. Microsc Res Tech,
2006;69(9):693–707.
139. Hoekstra, D., Tyteca, D., and van IJzendoorn, S.C. The subapical compart-
ment: a traffic center in membrane polarity development. J Cell Sci,
2004;117(Pt 11):2183–92.
140. Casanova, J.E. et al. Association of Rab25 and Rab11a with the apical recy-
cling system of polarized Madin‐Darby canine kidney cells. Mol Biol Cell,
1999;10(1):47–61.
141. Hales, C.M., Vaerman, J.P., and Goldenring, J.R. Rab11 family interacting
protein 2 associates with Myosin Vb and regulates plasma membrane recy-
cling. J Biol Chem, 2002;277(52):50415–21.
142. Jing, J. and Prekeris, R. Polarized endocytic transport: the roles of Rab11
and Rab11‐FIPs in regulating cell polarity. Histol Histopathol,
2009;24(9):1171–80.
143. Lapierre, L.A. et al. Phosphorylation of Rab11‐FIP2 regulates polarity in
MDCK cells. Mol Biol Cell, 2012;23(12):2302–18.
144. Fu, D. et al. Regulation of bile canalicular network formation and mainte-
nance by AMP‐activated protein kinase and LKB1. J Cell Sci, 2010;123(Pt
19):3294–302.
145. Wakabayashi, Y. et al. Rab11a and myosin Vb are required for bile canali-
cular formation in WIF‐B9 cells. Proc Natl Acad Sci U S A,
2005;102(42):15087–92.
146. Wakabayashi, Y., Kipp, H., and Arias, I.M. Transporters on demand: intra-
cellular reservoirs and cycling of bile canalicular ABC transporters. J Biol
Chem, 2006;281(38):27669–73.
147. Gatmaitan, Z.C., Nies, A.T., and Arias, I.M. Regulation and translocation of
ATP‐dependent apical membrane proteins in rat liver. Am J Physiol,
1997;272(5 Pt 1):G1041–9.
148. Misra, S. et  al. The role of phosphoinositide 3‐kinase in taurocholate‐
induced trafficking of ATP‐dependent canalicular transporters in rat liver. J
Biol Chem, 1998;273(41):26638–44.
149. Misra, S. et  al. Phosphoinositide 3‐kinase lipid products regulate ATP‐
dependent transport by sister of P‐glycoprotein and multidrug resistance
associated protein 2 in bile canalicular membrane vesicles. Proc Natl Acad
Sci U S A, 1999;96(10):5814–19.
150. Misra, S., Varticovski, L., and Arias, I.M. Mechanisms by which cAMP
increases bile acid secretion in rat liver and canalicular membrane vesicles.
Am J Physiol Gastrointest Liver Physiol, 2003;285(2):G316–24.
151. Cao, H. et  al. Cortactin is a component of clathrin‐coated pits and
participates in receptor‐mediated endocytosis. Mol Cell Biol,
­
2003;23(6):2162–70.
152. Smythe, E. and Ayscough, K.R. Actin regulation in endocytosis. J Cell Sci,
2006;119(Pt 22):4589–98.
153. Ortiz, D.F. et al. Identification of HAX‐1 as a protein that binds bile salt
export protein and regulates its abundance in the apical membrane of
Madin‐Darby canine kidney cells. J Biol Chem, 2004;279(31):32761–70.
154. Miyaoka, Y. et al. Hypertrophy and unconventional cell division of hepato-
cytes underlie liver regeneration. Curr Biol, 2012;22(13):1166–75.
155. Ragkousi, K. and Gibson, M.C. Cell division and the maintenance of epi-
thelial order. J Cell Biol, 2014;207(2):181–8.
 4:  Hepatocyte Surface Polarity 49
156. Reinsch, S. and Karsenti, E. Orientation of spindle axis and distribution of
plasma membrane proteins during cell division in polarized MDCKII cells.
J Cell Biol, 1994;126(6):1509–26.
157. Tuncay, H. and Ebnet, K. Cell adhesion molecule control of planar spindle
orientation. Cell Mol Life Sci, 2016;73(6):1195–207.
158. McNally, F.J. Mechanisms of spindle positioning. J Cell Biol,
2013;200(2):131–40.
159. Kotak, S. and Gonczy, P. Mechanisms of spindle positioning: cortical force
generators in the limelight. Curr Opin Cell Biol, 2013;25(6):741–8.
160. Bartles, J.R. and Hubbard, A.L. Preservation of hepatocyte plasma mem-
brane domains during cell division in situ in regenerating rat liver. Dev Biol,
1986;118(1):286–95.
161. Slim, C.L. et  al. Par1b induces asymmetric inheritance of plasma mem-
brane domains via LGN‐dependent mitotic spindle orientation in proliferat-
ing hepatocytes. PLoS Biol, 2013;11(12):e1001739.
162. Meyer, K. et  al. A predictive 3D multi‐scale model of biliary fluid
­dynamics in the liver lobule. Cell Syst, 2017;4(3):277–90.
 Primary Cilia
5 Carolyn M. Ott
Janelia Research Campus, Ashburn, VA, USA
INTRODUCTION
A multicellular organism has a need to coordinate multiple pro-
cesses in space and time. During development, multiple cell lin-
eages migrate and differentiate simultaneously. Individual organ
function, though physically isolated, must be coordinated with
the activity of other organs. Similarly, each specialized cell
coordinates with neighboring cells. Coordination is achieved as
individual cells respond to messages from other cells and the
environment, and transmit signals about their own state and
needs. Amid the cacophony of potential messages, individual
cells must discern which messages are relevant. Cilia project
away from the cell surface and function as a tunable sensing
organelle. Because cilia are continuous with both the cellular
cytoplasm and plasma membrane, specialized barriers restrict
entry and exit so that cells can form and adjust the composition
of cilia to sense and respond to signals.
The term “cilia” has been generalized to encompass all
types of ciliated structures, including flagella, and motile and
non‐motile cilia. Cilia and flagella, present throughout the
eukaryotic lineage, are distinct in structure and composition
from prokaryotic flagella. In eukaryotes, microtubules pro-
vide both structural support and a transportation highway
within cilia. Unicellular eukaryotes use cilia to process
responses to stimuli from the environment, mate, and feed. In
mice and other mammals, genetic ablation of primary cilia
terminates development. In adults, ciliary loss leads to dis-
ease [1]. Primary cilia are typically solitary projections from
the cell surface and they lack the structural components that
facilitate active beating in motile cilia and flagella. However,
primary cilia are not stationary, but can move with the
extracellular environment or pivot due to forces from the
intracellular actin cytoskeleton [2].
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
Primary cilia are present on a subset of cells in the liver;
cholangocytes have primary cilia, while hepatocytes do not. The
absence of cilia may be essential to specialization. Outside the
hematopoietic lineage, all progenitor cells have primary cilia, so
cilia‐less, differentiated cells have lost the ability to respond to
messages in the same way, and thus may be insulated. Like
hepatocytes, adipocytes and myocytes lack cilia, but are derived
from ciliated progenitors. The liver is a source of signals that
can be perceived by primary cilia elsewhere in the body. For
example, insulin‐like growth factor‐1 (IGF‐1) is produced by
the liver. The IGF‐1 receptor is expressed in diverse tissues and
can localize to cilia. IGF‐1 signaling through cilia can initiate
cilia resorption as cells enter the mitotic cycle [3].
Cells have many modes of perceiving the extracellular envi-
ronment. What advantages do cilia provide for signal perception
and transmission? Location, structure, isolation, and size make
the primary cilium a unique environment.
Location
Like a probe or an antenna, primary cilia extend away from the
cell and are immersed in the extracellular environment. The
structure is positioned to detect and report the conditions out-
side the cell. Cells can specify which part of the environment to
monitor using the cilium. For example, polarized cells typically
position cilia at the apical surface.
Structure
Two structural features of primary cilia are essential to consider.
First, because the cilium is a long narrow structure, it has a very
high surface area‐to‐volume ratio. The plasma membrane pro-
vides a platform for detecting the extracellular environment.
 5:  Primary Cilia 51
Second, the polarized microtubules that provide structural sup-
port also act as highways for traffic along the length of the cil-
ium. Transport along these microtubules can be regulated and
facilitates signaling. Tubulin in the microtubules themselves
may also have direct effects on proteins in the membrane [4].
Isolation
Although small proteins may be able to access primary cilia,
entry of larger proteins is restricted and targeting sequences or
interacting partners are thought to be essential. Regulation of
both the cytoplasmic and plasma membrane content of the pri-
mary cilium tunes the cilium to detect specific ligands. In addi-
tion, downstream signaling molecules within the cilium are
protected from modification by enzymes in the cytoplasm.
Although ions can pass between the cilium and the cytoplasm,
channels localized in the ciliary plasma membrane create a
unique ionic environment inside cilia that may influence protein
activity [5].
Size
The volume of a 5 μm cilium is approximately 0.35 μm3, which
is less than 0.05% of the total eukaryotic cell volume [6]. As a
result, within the cilium the concentration of a single molecule
is much larger than it would be in the cytoplasm. In addition,
because changes in cilia length change cilia volume, concentra-
tions within the cilium could also be regulated by processes that
affect cilia growth.
CILIUM COMPOSITION
AND ORGANIZATION
In the cytoplasm, most microtubules are hollow tubes created by
assembly of α and β tubulin monomers assembled into 13
protofilaments. Cilia microtubules are doublets composed
of a 13‐protofilament tubule (designated the A tubule) and a
10‐protofilament B tubule that closes onto the outside of
three A tubule protofilaments (Figure 5.1). Ciliary microtubules
receive several post‐translational modifications: acetylation,
detyrosination, glutamylation, and glycylation (Figure 5.2a) [7].
The nine doublet microtubules are direct extensions of centriole
microtubules. As a consequence, primary cilia are always
anchored in the cytoplasm. Collectively, cilia microtubes are
referred to as an axoneme. Although polarized, axoneme micro-
tubules do not treadmill like cytoplasmic microtubules because
there is no depolymerization at the minus end within the centro-
some. Axoneme length is mediated by addition or removal of
tubulin at the plus end at the cilium tip. Unlike flagella and
motile cilia that have internal structures to maintain the radial
symmetry of the microtubules along the entire length of the
cilium, primary cilia microtubules can rearrange toward the dis-
tal end. Primary cilia can also narrow as individual doublets ter-
minate (Figure 5.1) [8].
As protein synthesis does not occur in cilia, growth requires
tubulin monomers to be imported into the cilium and travel to
the tip. The intraflagellar transport (IFT) complex facilitates
directed movement along axonemal microtubules (reviewed in
[9]). IFT was first discovered as movement in the flagella of the
unicellular alga Chlamydomonas reinhardtii [10]. The IFT
complexes form trains (approximately 200 nm long in
Chlamydomonas flagella) that bind to microtubule motor pro-
teins, cargo, such as tubulin, and cargo adaptors that bind solu-
ble and membrane proteins. IFT complexes are large multiprotein
complexes about the size of the ribosomal small subunit (16S)
that can be biochemically separated into two subcomplexes,
IFT‐A and IFT‐B [11]. The IFT‐B subcomplexes associate with
kinesin‐2 to facilitate anterograde transport along the B tubules
of the axoneme doublets. The retrograde motor, dynein, is car-
ried to the plus end of ciliary microtubules through association
with the IFT‐A subcomplex. At the tip, kinesin‐2 dissociates
from the IFT complex, retrograde trains form, and IFT‐A‐asso-
ciated dynein drives retrograde movement along the A tubules
of the axoneme doublets (Figure 5.1) [12]. Two kinesin‐2 family
members, Kif3 and Kif17, process along ciliary microtubules:
Kif3 is required for ciliogenesis while Kif17 appears to partici-
pate in more specialized trafficking. The specific tubulin iso-
types in cilia, as well as the multiple tubulin modifications are
thought to affect IFT transport.
Membrane surrounds cilia microtubules as they extend away
from the centriole (also called a basal body). While some cen-
trosomes are at the cell surface, many are recessed and have a
cililary pocket: a specialized membrane domain with active
exocytosis and endocytosis [13]. The base of the cilium, near
the centriole, serves as the access point for both cytoplasmic and
membrane components. The composition of a cilium defines the
sensitivity and responsiveness of the organelle. While the ciliary
membrane is continuous with the plasma membrane of the cell,
access to both the ciliary membrane and cytoplasm is restricted
by a boundary created by a structure located inside the cilium,
just above the centrosome, called the transition zone. Mutations
in transition zone proteins or other proteins responsible for cre-
ating and regulating cilia composition are responsible for many
cilia‐related diseases, termed ciliopathies. Protein entry through
the transition zone is gated and although small molecules can
pass through the transition zone, several active processes main-
tain the unique environment of the cilium.
Most ciliopathies are pleiotropic diseases that affect several
overlapping organ systems including liver, kidney, retina, heart,
and bones [1]. Several ciliopathies, including nephronophthisis
(NPHP), Meckel syndrome (MKS), and Joubert syndrome, are
caused by mutation of transition zone proteins [14]. While often
considered a single defined structure, transition zone organiza-
tion and length can vary between cell types [15]. Tissue‐specific
differences in transition zones might contribute to the differ-
ences in symptoms upon loss or mutation of different transition
zone components.
Genetic, biochemical, and proteomic strategies identified
many components of the transition zone [16–18]. By electron
microscopy the transition zone is visible as “Y”‐shaped bridges
with two branches attached to the ciliary membrane and a single
anchor at each microtubule doublet (Figure 5.1). Subdiffraction
imaging strategies have been employed to reconcile the parts list
with the visible structure, and to begin to attribute functions to
individual components. Three key transition zone protein
complexes were named in part after the diseases caused by
52 THE LIVER:  CILIUM COMPOSITION AND ORGANIZATION

Tubulin kin
dimer es
in

(+)

IFT-A
IFT-B

Microtubule dynein
doublet
IFT-B IFT-A

kinesin

Primary cilium

dynein

Y-link
Intraflagellar transport
(IFT)

Transition zone
Ciliary pocket

Basal body
(Mother centriole)

Distal appendages

200 nm

Figure 5.1  Primary cilia architecture and intraflagellar transport. The cross‐sections show the structural changes as the microtubules extend from
the centrosome (described from bottom to top). The drawings are based on EM images in [8]. At the centrosome there are nine microtubule triplets.
Distal appendage proteins appear like a pinwheel emanating from the centriole microtubules. At the transition zone, Y‐links are visible. The doublet
microtubules of the ciliary axoneme are extensions of two of the triplet centriolar microtubules. Toward the distal end, the axoneme doublets can
rearrange and terminate. The doublets are composed of α and β tubulin dimers that assemble a closed cylinder (A tubule) and a second cylinder that
closes on the outside of the first (B tubule). On the right is a representation of intraflagellar transport (IFT). The IFT complex is a large multi‐subunit
complex that can be biochemically separated into the IFT‐A and IFT‐B subcomplexes. The IFT‐A subcomplex interacts with dynein and facilitates
retrograde transport along the A tubule. For anterograde transport toward the plus end of the microtubules at the cilium tip, the IFT‐B subcomplex
associates with kinesin‐2. IFT complexes associate with cargo in both directions. The retrograde motor, dynein, is an anterograde cargo, but the
kinesin motor dissociates at the tip and diffuses back to the base. Because the time for kinesin‐2 to return to the base is length‐dependent, the con-
centration of kinesin‐2 has been proposed as a length regulator. Adapted from Chien 2017 and Hendel 2018 [109, 110].

mutation of the complex. MKS and NPHP complex proteins
create the Y‐links while Cep290 localizes adjacent to the Y‐
links, proximal to the centrosome [19, 20]. How stable are the
Y‐link structures? Mating experiments in Chlamydomonas
revealed NPHP4 is static, but CEP290 can exchange between
transition zones of different flagella [21, 22]. Researchers found
no fluorescent recovery within 30 min after photobleaching
green fluorescent protein‐tagged MKS complex proteins indi-
cating that the proteins do not exchange with the unbleached
pool [23]. These results suggests that the MKS and NPHP com-
plexes form stable structures, while the pool of CEP290 may be
more dynamic.
The structure of the transition zone cannot yet explain its
function as a barrier capable of selective transport. It has
been shown, however, that depletion of transition zone pro-
teins that cause disease permit typically excluded proteins
into the cilium [22]. In addition to restricting entry, the tran-
sition zone can also limit exit and may function as a staging
platform because several dynamic cilia proteins have been
found to accumulate above this region [24, 25]. To pass
through the barrier, many proteins utilize escorts such as
nuclear import effectors, the BBSome complex, and tubby
family proteins (the latter two are discussed in the section
later on ciliary signaling).
 5:  Primary Cilia 53

(a) (c)
CEP164 FBF1 Merge

200 nm
(d)
FBF1 and SCLT1 CEP164 and SCLT1

5 m
200 nm
anti-acetylated tubulin anti-detyrosinated tubulin Hoecst
(e)
(b) DAM
Y-links

Ciliary
pocket
DABs

xy 5 m
CEP164 FBF1 SCLT1 CEP89 CEP83
xz
cilia-enriched GPCR centrosome CP110 T TBK2 IFT88 EHD1 Arl13B

Figure 5.2  Conventional and super‐resolution views of primary cilia. (a) Microtubules within cilia are post‐translationally modified. These cilia
on Madin–Darby canine kidney (MDCK) cells have been stained with antibodies that recognize acetylated and detyrosinated tubulin and visualized
using confocal microscopy. The image is a maximum intensity projection of a z‐stack. The inset images show that both acetylated and detyrosinated
tubulin are found along the entire length, however the distributions of the two are not identical. (b) Cilia‐enriched GPCR localization. Live imaging
of NIH3T3 fibroblast cell shows the concentration of the GPCR adjacent to the centrosome in cells expressing the GPCR, melanin concentrating
hormone receptor 1 (tagged with the fluorescent protein tdTomato), and the centrosome localization sequence of pericentrin (PACT; tagged with
tag‐blue fluorescent protein). The upper image is an xy projection; the narrow, lower panel is an xz projection of the same cell. (c–e) Super‐resolu-
tion imaging brings together protein localization and structure to reveal organization at the base of the cilium. (c) Two‐color axial view dSTORM
images of distal appendage proteins, FBF1 and CEP164. This approach was used to calculate the relative angular positions of different distal
appendage proteins (represented by symbols in the far right image). From [92]. (d) Two‐color lateral view dSTORM images of FBF1, SCLT1, and
CEP164 (three representative images of each). The relative height of FBF and SCLT1 are designated by the arrowheads on the far left. From [92].
(e) Three‐dimensional computational model created by combining information about the relative axial and lateral positions of distal appendage
proteins and several ciliary proteins. DAB, dorsal appendage blade; DAM, dorsal appendage matrix. From Yang 2018 https://www.nature.com/
articles/s41467‐018‐04469‐1#rightslink. Licensed under CCBY 4.0 [92].

The mechanism remains to be elucidated, but several lines of
evidence suggest that entry and exit of ciliary matrix proteins is
related to gating at the nuclear pore. Small soluble proteins can
freely enter the cilium, but, like the nuclear pore that excludes
proteins above ~50 kDa, large proteins require active transport
into the primary cilium [26]. Several fluorescently labeled
nucleoporin (NUP) proteins localize to the base of the primary
cilium and disrupting their function alters the transport of pro-
teins into and out of cilia [27, 28]. The NUPs found in cilia are
not the membrane or inner, nuclear basket proteins, but rather
NUPs that contain phenylalanine–glycine repeats that are
thought to create the permeability barrier. To pass through the
nuclear pore, cargo with import or export signals bind to trans-
port adaptor proteins, such as importin β. Adaptor binding and
release on the opposite side of the nuclear envelope is facilitated
by a gradient of RanGTP inside the nucleus and RanGDP in the
cytoplasm. Like the nucleus, the cilium is also rich in RanGTP
and the nuclear import adaptor importin β binds to ciliary locali-
zation sequences and can facilitate delivery of soluble proteins
through the boundary at the base of the cilium [29]. Because
primary cilia lack structures that resemble nuclear pores, it is
not yet clear where NUP proteins assemble or how they create a
permeability barrier.
Although the full definition and diversity of ciliary lipids is
not known, several lines of evidence indicate that, like the pro-
tein composition, the lipid environment of cilia is unique and
actively regulated. Early studies revealed that cilia membranes
are generally cholesterol rich, however, the transition zone may
have less cholesterol [30, 31]. Cilia have condensed membranes
and ciliogenesis requires phosphatidylinositol 4‐phosphate
adaptor protein‐2 (FAPP2) [32]. One of the causes of the
ciliopathy Joubert syndrome is mutation of INPP5E, a ciliary
54 THE LIVER:  CILIA‐MEDIATED SIGNALING
phosphoinositide 5‐phosphatase that localizes along the length
of the cilium, where it facilitates the conversion of phosphoi-
nositide 4,5‐bisphosphate (PI(4,5)P2) to phosphatidylinositol
4‐phosphate (PI(4)P) [33–36]. While adjacent membranes are
rich in PI(4,5)P2, INPP5E localization to cilia creates a PI(4)P‐
rich membrane. Targeting of INPP5E and other farnesylated
proteins to cilia is thought to be mediated by release of cargo
from the prenyl‐binding protein, phosphodiesterase 6 delta
subunit, mediated by Arl3 at the cilium [37]. Disrupting the
transition zone permits PI(4,5)P2 phospholipids to enter ­primary
cilia [38].
As mentioned earlier, the volume of the cilium is a fraction of
the volume of the rest of the cell and it is membrane bound
except at the base. Although ions can pass into and out of the
base of the cilium, transport of ions through ciliary membrane
channels helps maintain the unique ionic environment inside
cilia. The resting membrane potential of cilia on cultured human
retina pigment epithelial cells is ~30 mV more positive than the
resting potential of the plasma membrane of the rest of the cell
[39]. The concentration of Ca2+ in cilia is approximately seven
times higher than in the cytoplasm of cultured human pigment
epithelial cells [39]. While the cytoplasm and cytoplasmic orga-
nelles serve as a sink for Ca2+, channels in the ciliary membrane
may import enough Ca2+ to maintain a concentration greater
than 500 nM.
Ciliary ion channels are essential components of the vision
and olfaction signal transduction pathways. Several other ciliary
ion channels, including polycystin 2 (PKD2) and PKD2‐like 1
(PKD2‐L1), are members of the transient receptor potential
(TRP) family of cation channels. Homomeric PKD2 conducts
Na+ and K+ and is impermeable to Ca2+, while PKD2‐L1 is per-
meable to Na+, K+, and Ca2+ [40, 41]. The ionic environment
defined by the channels contributes to cell and organismal
health. Mutations in PKD2 cause autosomal dominant polycys-
tic kidney disease (ADPKD) and deletion of the gene causes
embryonic lethality. Without PKD2‐L1, mice have reduced
responses to hedgehog signaling and mild hedgehog‐related
developmental phenotypes [39]. It is not yet clear yet which
cilia proteins are influenced by the unique ionic environment.
Metabolites can pass between the cilium and the cytoplasm;
however, evidence suggests that active processes can sustain
different metabolite concentrations in cilia relative to the cyto-
plasm as a whole. Several isoforms of adenylyl cyclase, which
generates cyclic adenosine monophosphate (cAMP) through a
cyclizing reaction of ATP, localize to primary cilia [39, 40].
cAMP concentrations in primary cilia are ~5 times higher than
in the cell body and the levels can be altered through ciliary
signaling pathways [42]. Ciliary Ca2+ concentration also regu-
lates adenylyl cyclase activity, and thus could regulate ciliary
signal responses.
CILIA‐MEDIATED SIGNALING
Primary cilia are positioned to sample the extracellular environ-
ment near the cell. The ability to detect defined molecules and
signal their presence to the rest of the cell is a feature primary
cilia share with motile cilia and flagella. Sensory functions may
have co‐evolved with motility in early organisms as suggested
by a study that revealed several cilia signaling pathway compo-
nents in organisms that are thought to share an ancient early
ancestor with modern animals [43].
Olfaction and vision are both sensory functions mediated by
primary cilia [44]. Rod outer segments are modified cilia filled
with membrane disks containing rhodopsin, a G protein‐
coupled receptor (GPCR). Smell is the consequence of an
action potential initiated by ligand binding to olfactory GPCRs
on the cilia of olfactory neurons. Stereocilia, which detect
sound during hearing, are actin‐based structures and therefore
are not true cilia. The hair cells that project stereocilia do have
a true cilium, called a kinocilium. Although the kinocilium is
not required for mechanosensation of sound, mutations that
prevent cilia formation on hair cells disrupt proper stereocilia
organization possibly through failures to signal polarity cues
during development [45].
Sensing and signaling performed by primary cilia are essen-
tial for differentiation and tissue morphogenesis during devel-
opment. Mutations that cause systemic loss of primary cilia
result in embryonic lethality accompanied by patterning
defects such as left–right asymmetry [46, 47]. Loss of cilia in
adult animals leads to kidney cyst formation and obesity due to
abnormal appetite signaling [48]. Engineered loss of cilia in
defined cell types causes many additional phenotypes (see, for
example, [49–51]). Some ciliary proteins may contribute to
signaling in cancer [52] and signaling through primary cilia
also affects response to injury [52, 53]. Primary cilia in diverse
tissues do not participate in a single, stereotyped signaling
pathway. Rather, the expression and localization of receptors
and downstream effectors varies by both cell type and life
stage. There is a growing list of genes that contribute to cili-
opathies; mutation of many of which alter signal perception or
transmission [1].
A key paradigm for ciliary signaling is through GPCRs.
Figure 5.3 summarizes some of the potential cilia GPCR signal
transduction outcomes described below. Similar to rhodopsin
and olfactory GPCRs, cilia targeting of many GPCRs is so effi-
cient that the concentration of the GPCR in the ciliary mem-
brane far exceeds the receptor concentration anywhere else in
the cell (Figure  5.2b). Sequence features on the intracellular
face GPCRs contribute to ciliary targeting [53–55]. These fea-
tures are recognized by tubby (Tub) family proteins including
Tub and tubby‐like protein 3 (TULP3). Tub and TULP3 associ-
ate with the plasma membrane through binding PI(4,5)P2,
where they capture GPCRs with cilia localization sequences.
Tub and TULP3 binding to IFT‐A facilitates cilia delivery.
Within the cilium there is little PI(4,5)P2, so Tub and TULP3
dissociate from the membrane and GPCR cargos are released
[56]. Loss of Tub, which is predominantly expressed in the
brain, prevents cilia localization of GPCRs critical for appetite
regulation and results in obesity [57].
Delivery of membrane proteins, including GPCRs and ion
channels, to the base of the cilium can be mediated by lateral
passage from the plasma membrane, direct targeting of Golgi‐
derived vesicles or indirect trafficking through endosome‐
derived vesicles [58]. Bardet–Biedl syndrome (BBS) is a
ciliopathy caused by mutation in proteins that form a complex
called the BBSome [59, 60]. GPCRs are absent from cilia of
 5:  Primary Cilia 55

(a) Without ligand

?

Rab8 Tub GPCR

?
BBSome
PI(4)P Adenylyl
cyclase
Tub GDP
PKA
Kinase
target

Ion
channel

(b) Upon ligand binding

Ligand

PI(4)P Adenylyl
cyclase
cAMP cAMP
+ GTP
cAMP
P
PKA
P
BBSome

Possible consequences of ligand binding

G Protein GTP
GDP

Adenylyl cyclase ATP or cAMP

GPCR cilium exit with BBSome

Kinases inactive ACTIVE

Channels closed open

Figure 5.3  Generalized perspective of signaling through primary cilia. (a) Many GPCRs localize to primary cilia. Trafficking of GPCRs to
primary cilia is mediated by Rab8 and the BBSome. Tub family proteins escort GPCRs into primary cilia and release them when they reach the
PI(4)P‐rich membrane (orange). In the absence of ligand, GPCRs can associate with G proteins. (b) Ligand binding to GPCRs can have many
potential consequences. Activated receptors can trigger replacement of Gα‐associated GDP with GTP. As a result, the GTP‐bound Gα subunit dissociates
from the G protein β/γ subunits and both propagate signaling cascades. One possibility shown here is that Gα can stimulate adenylyl cyclase, which
produces cAMP. cAMP can activate kinases such as PKA and could affect ion channel activity. Receptor activation also leads to BBSome‐dependent
internalization of GPCRs that may reset signaling pathways. As indicated at the bottom, signaling cascades could also have the opposite effects,
such as decreasing cAMP production, inactivating kinases, or closing channels.

primary cells cultured from animals lacking individual BBS
proteins [61]. The BBSome complex forms a coat structure that
can interact directly with both GPCRs and secretory pathway
directors such as Rab8 and facilitate delivery of GPCRs to
the cilium.
Ciliary GPCRs bind diverse ligands including bile acids,
nucleotides, neuropeptides, and proteins [62]. Upon ligand
binding, the receptors facilitate replacement of Gα‐associated
GDP with GTP. The trimeric G protein then dissociates from the
receptor and the Gα and Gβγ subunits both initiate downstream
signaling events. There are multiple isoforms of the G protein
subunits and, in general, different GPCRs are thought to recruit
G proteins containing a specific subtype. However, the GPCR
TGR5 associates with a different Gα isoform in ciliated versus
unciliated cells [63]. Ciliary G proteins can activate different,
and sometimes opposing, downstream effector pathways. For
example, once liberated, stimulatory Gα subunits (Gαs) increase
adenylyl cyclase activity while inhibitory Gαi subunits decrease
cAMP levels. cAMP acts as a second messenger and promotes
activity of downstream kinases including protein kinase A
(PKA). Ligand binding leads to retrieval of activated GPCRs
from the cilium [64–67]. Although additional signaling of
56 THE LIVER:  BUILDING, MAINTAINING, AND DISMANTLING CILIA
internalized GPCRs has been demonstrated in other contexts
[68], no activity has been reported for GPCRs internalized
from cilia.
Internalization of activated receptors could help reset
signaling pathways. The BBSome, which is implicated in
trafficking proteins to cilia, also facilitates exit of activated
GPCRs through the transition zone in complex with the
small G protein Arl6 [65]. Upon receptor activation, Gαi
signaling, which reduces adenylyl cyclase and thus PKA
activity, causes BBSome proteins, Arl6, and retrograde IFT
proteins to accumulate at the distal end where the proteins
form large IFT trains, which then bind to activated GPCRs,
transit to the base of the cilium and escort the GPCR through
the transition zone [69].
An alternate mechanism to remove GPCRs and reset sign-
aling pathways is through direct release of the GPCRs in
cilia‐derived extracellular vesicles (EVs) [70]. Rhodopsin
and the membrane disks of photoreceptor cells are proces-
sively removed from the distal end of the outer segment [71].
EVs containing primary cilia proteins have also been col-
lected in urine [72]. Actin, which is generally not polymer-
ized in cilia, can drive release of cilia membranes in response
to increased PI(4,5)P2 when INPP5E is disabled or removed
from cilia [73, 74].
Many critical signaling pathways center on primary cilia
[75]. Obesity generated in ciliopathies is caused by disruptions
in satiety signaling through neuropeptide‐binding GPCRs in
hypothalamus cilia. The hedgehog signaling pathway is critical
for developmental and can contribute to cancer. During chronic
liver disease, hedgehog signaling in liver progenitor cells con-
tributes to regeneration [76]. Most components and effectors of
this complicated signaling cascade function in cilia. In the cil-
ium, the activity of the hedgehog receptor Patched influences
the ciliary localization of the GPCR Smoothened. The activity
of Smoothened in the cilium affects processing of Gli transcrip-
tion factors. Another GPCR, Gpr161, also localizes to the cil-
ium and mediates hedgehog signaling [77].
Although it is clear that ligand binding to receptors in the
unique ciliary environment initiates signaling cascades, how
the physiological consequences of these cascades propagate
and cause changes in the rest of the cell is generally less clear.
An exception is the hedgehog signaling pathway: the Gli fam-
ily transcription factors localize to primary cilia and can be
processed and trafficked out of the cilium in response to path-
way activation [78, 79]. There are many other possible out-
comes in addition to transcription alterations; however, all of
these are subject to the same limitation that Delling and col-
leagues revealed regarding Ca2+: upon exiting, soluble mole-
cules will be significantly diluted [5]. Delivery of cilia
effectors into the nucleus could be streamlined by the overlap-
ping components of the cilia and nuclear import/export path-
ways. As the centrosome serves as both the basal body and the
microtubule organizing center, it is possible that molecules
exiting cilia have a transient local concentration high enough
to transmit signals to proteins recruited to the centrosome.
Association with carrier molecules or membrane vesicles
could be an additional strategy.
cAMP and cGMP generated by signaling cascades can stimu-
late ion channel activity and alter ciliary Ca2+ levels; this is a
critical part of signal transduction in olfaction and vision [80,
81]. Ciliary ion channels may have multiple effects on signaling
cascades. Adenylyl cyclase 3 activity is affected by Ca2+ [82]
and it has been hypothesized that changes in intracellular Ca2+ in
cilia might alter downstream signaling components such as ade-
nylyl cyclase [83]. In animal models of PKD, tissue levels of
cAMP are elevated [84]. Activation of ion channels could, if the
paradigm is true, exert global effects on any ciliary proteins that
are sensitive to changes in ion concentrations. Such a scenario
might explain why PKD2‐L1‐deficient mice have reduced
responses to Shh [39].
It is not clear whether cilia act as mechanosensitive flow
sensors. The prevailing paradigm that primary cilia bend-
ing communicates the presence of external fluid flow via
mechanosensitive activation of calcium ion channels in the
ciliary membrane has been reconsidered. Detailed studies
using a fluorescent Ca2+ sensor localized to primary cilia
revealed no changes in ciliary Ca2+ levels upon deflection
with flow from a micropipette [5]. If cilia are mechanosensi-
tive, the signal transduction does not appear to be mediated by
changes in ciliary calcium. In cartilage, chondrocyte cilia par-
ticipate in mechanotransduction, not as a mechanosensor, but
as a chemical signal transducer. Receptors in chondrocyte
cilia membranes bind ATP that is released by the cell body
upon compression [85].
Although signal reception in cilia is compared to radio signal
reception using antennae, cilia are likely capable of being more
than just a passive sensor. It is possible that, like other
EVs, cilia‐derived vesicles may be biologically active.
Chlamydomonas releases vesicles from flagella that facilitate
breakdown of the wall encapsulating the daughter cells after
mating [86]. In addition to releasing particles, primary cilia may
participate in signaling through direct touch. Cholangiocyte
cilia within the bile duct and other cilia that project into tubules
can adhere to one another [87], which might facilitate direct
intracellular communication. Primary cilia in many other tissues
are not in a lumen, but rather, can be enmeshed in extracellular
matrix, project between layers of cells, or entwined with other
extracellular processes (examples include cilia on chondrocytes,
mammary myoepithelial cells, and neurons) [88, 89]. In deep
tissue, it is possible that primary cilia could signal through
direct contact.
BUILDING, MAINTAINING,
AND DISMANTLING CILIA
As cells cycle and differentiate, primary cilia are deconstructed
and recreated. When maintained, cilia length can vary by an
order of magnitude between cell types (from ~2 to >20 μm) but
is typically uniform within a population. The presence and
length of a cilium define the ability of a cell to detect and
respond to extracellular cues, so the following questions are rel-
evant to understanding the role of cilia in health and develop-
ment: How do cilia form and how is biogenesis regulated? How
is cilium length determined and maintained? When and how are
cilia removed from the cell surface? Partial answers to these
questions are discussed below.
 5:  Primary Cilia 57
Ciliogenesis
How is the critical and unique ciliary environment created?
Primary cilia can grow from a centriole deep in the cytoplasm or
docked at the plasma membrane (Figure  5.4a,b; reviewed in
[90]). In both pathways microtubule doublets extend from the
triplet microtubules of the centriole, membrane is added, then
presumably differentiated as the transition zone forms. The inter-
nal pathway requires the recruitment of vesicles that anchor and
fuse to form a large ciliary vesicle docked to the mother centri-
ole. As microtubules extend from the centriole, membrane is
added and the ciliary vesicle deforms to ensheath the nascent
cilium. The final stage of internal ciliogenesis requires the mem-
brane around the tip of the growing cilium to fuse with the
plasma membrane. Ciliogenesis at the plasma membrane is
thought to begin when a centriole docks at the plasma mem-
brane. Membrane is added at the plasma membrane as the micro-
tubules extend to ensheath the cilium. The membrane‐associated
GTPases Rab11 and Rab8 target vesicles to nascent cilia to pro-
vide both lipids and key ciliary membrane proteins [91].
Every centriole is not capable of becoming a basal body.
Structures called distal appendages or transition fibers must be
added to the distal end of the ninefold symmetric centriole tri-
plet microtubules to recruit vesicles or dock at the plasma mem-
brane. Several components assemble to form distal appendages,
and recent super‐resolution imaging efforts have revealed the
organization of many key distal appendage proteins (Figure 5.2)
[92]. CEP164 and SCLT1 are components of the distinctive
blade structure that is visible in electron micrographs. In con-
trast, FBF1 localizes to the distal appendage matrix – a space
that appears empty in electron micrographs.
Distal appendage proteins contribute to several aspects of
ciliogenesis. Studies of initiation steps in intracellular ciliogen-
esis [93] revealed that fusion of small vesicles recruited to distal
appendages requires membrane tubulating proteins: EHD1 and
EHD3. Subsequently, CEP164 in the distal appendage blade
participates in both microtubule growth and membrane expan-
sion (reviewed in [94]). CEP164 recruits tau tubulin kinase 2
(TTBK2), a protein required for a key transformation at the cen-
trosome – removal of proteins that cap the plus ends of the cen-
trosome microtubules. TTBK2 facilitates removal of CP110
from the distal end of the centriole. Although it must be removed
prior to ciliogenesis, mice with no CP110 have fewer cilia. This
suggests that CP110 may be required for early steps in centro-
some remodeling and then be removed prior to microtubule
extension [95]. In the internal ciliogenesis pathway CEP164
interactions with both Rab8a and its GTPase‐exchange factor
Rabin 8 are thought to promote membrane docking of early cili-
ary membranes [96].
Targeting and transport of ciliary components is also essen-
tial for cilia formation. Mutation or deletion of IFT‐B complex
proteins, which facilitates anterograde transport, disrupts cilia
formation. FBF1 in the distal appendage matrix colocalizes with
IFT88 and facilitates entry of IFT complexes into cilia [92, 97].
Members of the IFT‐B complex bind tubulin and may escort it
through the permeability barrier where it is concentrated in the
cilium [28, 98]. Interference with the formation of centriolar
satellites, which are thought to function as a protein complex
assembly platform, can also prevent cilia formation. Several
cilia proteins, including BBS4 and CEP290, associate with the
key scaffolding protein pericentriolar material‐1 (PCM‐1) in
centriolar satellites. Impaired autophagy can also prevent cilio-
genesis because oral–facial–digital syndrome 1 proteins are not
degraded but instead accumulate at centriolar satellites [99].
The two centrioles inside each cell are different and only one
of them undergoes remodeling to become the basal body. During
each cell division cycle, the two centrioles duplicate and then
separate to form the spindle pole. Each daughter inherits a pair
of centrioles, a duplicate (called the daughter) and an older orig-
inal (called the mother). The centriole that becomes the basal
body is always the older centriole. Cilia formation can be differ-
ent between sister cells because the two mother centrioles they
inherit are not equivalent. In the preceding cell cycle one of the
mother centrioles had been a daughter. The sister cell that inher-
its the older mother centriole forms a cilium earlier, which could
influence perception of extracellular developmental cues and
downstream differentiation [100]. The older mother centriole
may be faster because it already has distal appendages and can
retain association of ciliary membranes [101].
Many polarized cells build cilia at the apical plasma mem-
brane and several lines of evidence indicate that polarization
may position and promote ciliogenesis (reviewed in [102]).
Proteins that localize to the boundary between apical and baso-
lateral membranes in polarized epithelial cells also localize to
the base of the primary cilium. Although how polarity proteins
influence cilia formation and maintenance is unclear, their
importance has been demonstrated. Deletion, mutation, or phar-
macological disruption of polarity proteins prevents ciliogene-
sis. PAR complex proteins  –  PAR3, PAR6  –  and an atypical
PKC (aPKC) associate with the anterograde kinesin‐2 compo-
nent Kif3a. The lipid ceramide influences ciliogenesis and is
required for association of aPKC with other PAR proteins [103].
Cdc42, another PAR complex component, interacts with and
promotes the ciliary localization of exocyst proteins. The exo-
cyst complex facilitates targeting and tethering of post‐Golgi
vesicles and can interact with Rab8 to direct trafficking in polar-
ized cells [104].
Ciliogenesis at the plasma membrane may be facilitated in
some way by the midbody remnant (Figure  5.4b). After the
cytokinetic bridge is severed during cell division, part of
the bundle of microtubules and membrane that made up the
bridge can remain associated with the plasma membrane of one
of the daughter cells. Movement of the remnant across the
apical surface of the cell to where the centriole is docked pre-
cedes cilia growth. Although cilia and midbodies share many
molecular components, it is not clear if any of these are directly
transferred to contribute to cilia growth [105].
Maintaining or altering cilia length
Cilia length is typically constant across a population but can
vary across cell types. This suggests that cellular parameters
such as protein expression levels or turnover rates influence the
cilia length setpoint. Several factors that alter the cilia length
setpoint have been identified. For example, tubulin concentra-
tions and modifications can affect cilia length [28, 106, 107]. In
addition, proteins involved in IFT are known to affect cilia
length: increased expression of IFT‐B components yields longer
58 THE LIVER:  BUILDING, MAINTAINING, AND DISMANTLING CILIA

(a)

Ciliary
Distal vesicle
appendages

Centriole

(b)

Midbody
remnant

Transition
? zone

(c)

Dissasembly Severing Resorption

Figure 5.4  Ciliogenesis and cilia elimination. (a) Internal ciliogenesis begins with maturation of the mother centriole. Although they are depicted
as discrete stages, initial vesicle recruitment and axoneme extension may occur simultaneously. As the axoneme extends, the transition zone forms
and additional membrane must be delivered. When the membrane encapsulating the cilium contacts the plasma membrane, the bilayers are thought
to fuse so the membrane adjacent to the microtubules becomes continuous with the plasma membrane. The outer membrane of the ciliary vesicle
may become part of the ciliary pocket membrane. (b) Ciliogenesis at the plasma membrane requires formation of distal appendages and movement
of the mother centriole to the cell surface. It is not currently clear how the midbody remnant promotes cilia assembly at the cell surface. Like in the
internal pathway, the transition zone forms and microtubule growth must be coupled to membrane addition to extend the nascent cilium. (c) Three
­possibly strategies that could all contribute to cilium disappearance are depicted. Microtubule disassembly may be stimulated by activation of
the tubulin deacetylase HDAC6. Membrane removal via endocytosis could also contribute. Severing of ciliary membrane is known to occur and has
been associated with cilium removal. A third strategy involves dissolving the barriers that separate the ciliary cytoplasm and membrane from the
rest of the cell. There has been some evidence to support each possibility.

cilia [108]. IFT‐B delivery of tubulin to the tip may increase the
local concentration of tubulin and drive polymerization.
Kinesin‐2 diffuses back to the base of the cilium after it carries
IFT trains to the tip. Because diffusion of kinesin‐2 back to the
base of the cilium where it can form new anterograde trains is
length dependent, kinesin‐2 concentration has been proposed as
a key regulator of cilia length [109, 110].
The actin cytoskeleton can also influence cilia length [111].
Cilia length increases in cells when F‐actin is destabilized by
treatment with cytochalasin D. Similar increases in cilia length
have been reported by depleting Arp3, which nucleates F‐actin
branching and by treatment with a microRNA that downregulates
four positive regulators of branched F‐actin [112]. Cilia elongation
upon treatment with cytochalasin D is accompanied by an influx
of membrane‐associated actin‐binding proteins in cilia [113].
Cdc42 and aPKC, components of the PAR complex, participate
in signaling to influence F‐actin [114]. How disrupting branched
F‐actin affects cilia length is not yet clear. Perhaps the actin
motor myosin Va, which can bind Rab8 and a transition zone
protein [115], participates in actin regulation of cilia length.
Several signaling pathways and disease conditions alter cilia
length. Inhibition of adenylyl cyclase by treatment with lithium
causes cilia elongation [116]. In physiological contexts, adeny-
lyl cyclase activity can be stimulated or inhibited by Gα proteins
 5:  Primary Cilia 59
upon activation of ciliary GPCRs. It is not yet clear whether
cAMP as a second messenger directly influences a length regu-
lating pathways described above, or if it promotes cilia exten-
sion through other binding partners. Another candidate
downstream of signaling pathways that may also influence cilia
length is the proteasome, which associates with a transition
zone protein and cleaves Gli transcription factors [117].
Cilia elimination
While some organisms retain cilia or flagella, mammalian cells
disassemble cilia in G1 phase of the cell cycle or at the begin-
ning of mitosis. Cilia removal could be achieved by: (i) sever-
ing; (ii) disassembly; and (iii) resorption upon decommissioning
the boundaries that segregate the cilium (Figure 5.4c). The three
mechanisms could work together and there is evidence to sug-
gest that all three processes can happen [93]. Electron micros-
copy of epithelial cells provided early support for the model that
cilia are resorbed into the cytoplasm [118]. Stimulated release
of vesicles from the tip can precipitate cilia disassembly [74].
Aurora A kinase, which is regulated by many factors and func-
tions as a part of the cell cycle, phosphorylates both INPP5E
and the deacetylase HDAC6. HDAC6 destabilization of acety-
lated microtubules may facilitate axoneme deconstruction. The
activity of microtubule‐depolymerizing kinesins associated
with the axoneme may also contribute to reducing axoneme
length. It is not clear if there is a role for endocytosis in recover-
ing membrane from cilia. Polo‐like kinase 1, a kinase regulated
by the cell cycle, promotes cilia disassembly and phosphoryl-
ates a component of the transition zone, possibly disrupting the
integrity of the ciliary gate (see [119]).
CONCLUSION
Because of their distinct structure in electron microscopy, primary
cilia were identified on many cell types throughout the twentieth
century. A small community of researchers built foundational
knowledge about the structure and function of primary cilia. The
discovery that proteins move along cilia, followed by insights into
the many links between primary cilia function and health stimu-
lated expansion of cilia research. Investigators in disparate fields
have found that cilia biology is central to many aspects of cell
signaling and development. These, in turn, have provided new
insights into fundamental aspects of cilia biology. Technological
advances, including super‐resolution microscopy, have provided
new insights into the movement and molecular architecture of
cilia. However, many exciting questions remain. Emerging studies
will likely provide additional insights into how defects in cilia
form and function contribute to disease and will hopefully lead to
novel interventions to alleviate the symptoms of ciliopathies.
ACKNOWLEDGMENT
The author thanks Christine Kettenhofen, Corey Valinsky, and
Shu‐Hsien Sheu for helpful suggestions on the text.
REFERENCES
1. Reiter, J.F. and Leroux, M.R. Genes and molecular pathways underpinning
ciliopathies. Nat Rev Mol Cell Biol, 2017;18(9):533–47.
2. Battle, C., Ott, C.M., Burnette, D.T. et al. Intracellular and extracellular forces
drive primary cilia movement. Proc Natl Acad Sci U S A, 2015;112(5): 1410–15.
3. Yeh, C., Li, A., Chuang, J.‐Z. et al. IGF‐1 activates a cilium‐localized nonca-
nonical Gβγ signaling pathway that regulates cell‐cycle progression. Dev Cell,
2013;26(4):358–68.
4. Li, Q., Montalbetti, N., Wu, Y. et al. Polycystin‐2 cation channel function is
under the control of microtubular structures in primary cilia of renal epithe-
lial cells. J Biol Chem, 2006;281(49):37566–75.
5. Delling, M., Indzhykulian, A.A., Liu, X. et al. Primary cilia are not calcium‐
responsive mechanosensors. Nature, 2016;531(7596):656–60.
6. Nachury, M.V. How do cilia organize signalling cascades? Philos Trans R
Soc Lond B Biol Sci, 2014;369(1650):20130465.
7. Wloga, D., Joachimiak, E., Louka, P., and Gaertig, J. Posttranslational modi-
fications of tubulin and cilia. Cold Spring Harb Perspect Biol,
2017;9(6):a028159.
8. Gluenz, E., Höög, J.L., Smith, A.E. et al. Beyond 9+0: noncanonical axo-
neme structures characterize sensory cilia from protists to humans. FASEB J,
2010;24(9):3117–21.
9. Lechtreck, K.F. IFT–cargo interactions and protein transport in cilia. Trends
Biochem Sci, 2015;40(12):765–78.
10. Kozminski, K.G., Johnson, K.A., Forscher, P., and Rosenbaum, J.L. A motil-
ity in the eukaryotic flagellum unrelated to flagellar beating. Proc Natl Acad
Sci U S A, 1993;90(12):5519–23.
11. Cole, D.G., Diener, D.R., Himelblau, A.L. et al. Chlamydomonas kinesin‐II‐
dependent intraflagellar transport (IFT): IFT particles contain proteins
required for ciliary assembly in Caenorhabditis elegans sensory neurons. J
Cell Biol, 1998;141(4):993–1008.
12. Stepanek, L. and Pigino, G. Microtubule doublets are double‐track railways
for intraflagellar transport trains. Science, 2016;352(6286):721–4.
13. Benmerah, A. The ciliary pocket. Curr Opin Cell Biol, 2012;25(1):78–84.
14. Gonçalves, J. and Pelletier, L. The ciliary transition zone: finding the pieces
and assembling the gate. Mol Cells, 2017;40(4):243–53.
15. Jana, S.C., Mendonça, S., Machado, P. et al. Differential regulation of transi-
tion zone and centriole proteins contributes to ciliary base diversity. Nat Cell
Biol, 2018;20(8):928–41.
16. Sang, L., Miller, J.J., Corbit, K.C. et  al. Mapping the NPHP‐JBTS‐MKS
protein network reveals ciliopathy disease genes and pathways. Cell,
2011;145(4):513–28.
17. Dean, S., Moreira‐Leite, F., Varga, V., and Gull, K. Cilium transition zone
proteome reveals compartmentalization and differential dynamics of ciliopa-
thy complexes. Proc Natl Acad Sci U S A, 2016;113(35):E5135–43.
18. Diener, D.R., Lupetti, P., and Rosenbaum, J.L. Proteomic analysis of isolated
ciliary transition zones reveals the presence of ESCRT oroteins. Curr Biol,
2015;25(3):379–84.
19. Yang, T.T., Su, J., Wang, W.‐J. et  al. Superresolution pattern recognition
reveals the architectural map of the ciliary transition zone. Sci Rep,
2015;5(1):14096.
20. Williams, C.L., Li, C., Kida, K. et al. MKS and NPHP modules cooperate to
establish basal body/transition zone membrane associations and ciliary gate
function during ciliogenesis. J Cell Biol, 2011;192(6):1023–41.
21. Craige, B., Tsao, C.‐C., Diener, D.R. et al. CEP290 tethers flagellar transi-
tion zone microtubules to the membrane and regulates flagellar protein con-
tent. J Cell Biol, 2010;190(5):927–40.
22. Awata, J., Takada, S., Standley, C. et al. NPHP4 controls ciliary trafficking
of membrane proteins and large soluble proteins at the transition zone. J Cell
Sci, 2014;127(Pt 21):4714–27.
23. Lambacher, N.J., Bruel, A.‐L., van Dam, T.J.P. et al. TMEM107 recruits cili-
opathy proteins to subdomains of the ciliary transition zone and causes
Joubert syndrome. Nat Cell Biol, 2016;18(1):122–31.
24. Shi, X., Garcia, G., III, Van De Weghe, J.C. et al. Super‐resolution micros-
copy reveals that disruption of ciliary transition‐zone architecture causes
Joubert syndrome. Nat Cell Biol, 2017;19(10):1178–88.
25. Yang, T.T., Hampilos, P.J., Nathwani, B. et  al. Superresolution STED
microscopy reveals differential localization in primary cilia. Cytoskeleton
(Hoboken), 2012;70(1):54–65.
26. Takao, D. and Verhey, K.J. Gated entry into the ciliary compartment. Cell
Mol Life Sci, 2016;73(1):119–27.
60 THE LIVER: REFERENCES
27. Kee, H.L., Dishinger, J.F., Blasius, T.L. et al. A size‐exclusion permeability
barrier and nucleoporins characterize a ciliary pore complex that regulates
transport into cilia. Nat Cell Biol, 2012;14(4):431–7.
28. Endicott, S.J. and Brueckner, M. NUP98 Sets the size‐exclusion diffusion
limit through the ciliary base. Curr Biol, 2018;28(10):1643–50.e3.
29. Dishinger, J.F., Kee, H.L., Jenkins, P.M. et al. Ciliary entry of the kinesin‐2
motor KIF17 is regulated by importin‐β2 and RanGTP. Nat Cell Biol,
2010;12(7):703–10.
30. Souto‐Padron, T. and de Souza, W. Freeze‐fracture localization of filipin‐
cholesterol complexes in the plasma membrane of Trypanosoma cruzi. J
Parasitol, 1983;69(1):129.
31. Cuevas, P. and Gutierrez Diaz, J.A. Absence of filipin‐sterol complexes from
the ciliary necklace of ependymal cells. Anat Embryol, 1985;172(1).
32. Vieira, O.V., Gaus, K., Verkade, P. et al. FAPP2, cilium formation, and com-
partmentalization of the apical membrane in polarized Madin‐Darby canine
kidney (MDCK) cells. Proc Natl Acad Sci U S A, 2006;103(49):18556–61.
33. Garcia‐Gonzalo, F.R., Phua, S.C., Roberson, E.C. et al. Phosphoinositides
regulate ciliary protein trafficking to modulate hedgehog signaling. Dev Cell,
2015;34(4):400–9.
34. Chávez, M., Ena, S., Van Sande, J. et  al. Modulation of ciliary phosphoi-
nositide content regulates trafficking and sonic hedgehog signaling output.
Dev Cell, 2015;34(3):338–50.
35. Bielas, S.L., Silhavy, J.L., Brancati, F. et al. Mutations in INPP5E, encoding
inositol polyphosphate‐5‐phosphatase E, link phosphatidyl inositol signaling
to the ciliopathies. Nat Genet, 2009;41(9):1032–6.
36. Jacoby, M., Cox, J.J., Gayral, S. et al. INPP5E mutations cause primary cil-
ium signaling defects, ciliary instability and ciliopathies in human and
mouse. Nat Genet, 2009;41(9):1027–31.
37. Fansa, E.K., Kösling, S.K., Zent, E. et  al. PDE6δ‐mediated sorting of
INPP5E into the cilium is determined by cargo‐carrier affinity. Nat Commun,
2016;7:11366.
38. Jensen, V.L., Li, C., Bowie, R.V. et al. Formation of the transition zone by
Mks5/Rpgrip1L establishes a ciliary zone of exclusion (CIZE) that compart-
mentalises ciliary signalling proteins and controls PIP2 ciliary abundance.
EMBO J, 2015;34(20):2537–56.
39. Delling, M., DeCaen, P.G., Doerner, J.F. et al. Primary cilia are specialized
calcium signalling organelles. Nature, 2013;504(7479):311–14.
40. Shen, P.S., Yang, X., DeCaen, P.G. et al. The structure of the polycystic kid-
ney disease channel PKD2 in lipid nanodiscs. Cell, 2016;167(3):763–73.
e11.
41. Liu, X., Vien, T., Duan, J. et al. Polycystin‐2 is an essential ion channel subu-
nit in the primary cilium of the renal collecting duct epithelium. Elife,
2018;7:E2363.
42. Moore, B.S., Stepanchick, A.N., Tewson, P.H. et al. Cilia have high cAMP
levels that are inhibited by Sonic Hedgehog‐regulated calcium dynamics.
Proc Natl Acad Sci U S A, 2016;113(46):13069–74.
43. Sigg, M.A., Menchen, T., Lee, C. et al. Evolutionary proteomics uncovers
ancient associations of cilia with signaling pathways. Dev Cell,
2017;43(6):744–62.e11.
44. Singla, V. and Reiter, J.F. The primary cilium as the cell’s antenna: signaling
at a sensory organelle. Science, 2006;313(5787):629–33.
45. Sipe, C.W. and Lu, X. Kif3a regulates planar polarization of auditory hair
cells through both ciliary and non‐ciliary mechanisms. Development,
2011;138(16):3441–9.
46. Huangfu, D., Liu, A., Rakeman, A.S. et al. Hedgehog signalling in the mouse
requires intraflagellar transport proteins. Nature, 2003;426(6962):83–7.
47. Nonaka, S., Tanaka, Y., Okada, Y. et al. Randomization of left‐right asym-
metry due to loss of nodal cilia generating leftward flow of extraembryonic
fluid in mice lacking KIF3B motor protein. Cell, 1998;95(6):829–37.
48. Davenport, J.R., Watts, A.J., Roper, V.C. et al. Disruption of intraflagellar
transport in adult mice leads to obesity and slow‐onset cystic kidney disease.
Curr Biol, 2007;17 (18):1586–94.
49. Haycraft, C.J., Zhang, Q., Song, B. et al. Intraflagellar transport is essential
for endochondral bone formation. Development, 2007;134(2):307–16.
50. Dinsmore, C. and Reiter, J.F. Endothelial primary cilia inhibit atherosclero-
sis. EMBO Rep, 2016;17(2):156–66.
51. Foerster, P., Daclin, M., Asm, S. et al. mTORC1 signaling and primary cilia
are required for brain ventricle morphogenesis. Development,
2017;144(2):201–10.
52. Liu, H., Kiseleva, A.A., and Golemis, E.A. Ciliary signalling in cancer. Nat
Rev Cancer, 2018;18(8):511–24.
53. Berbari, N.F., Johnson, A.D., Lewis, J.S. et  al. Identification of ciliary
localization sequences within the third intracellular loop of G protein‐cou-
pled receptors. Mol Biol Cell, 2008;19(4):1540–7.
54. Loktev, A.V. and Jackson, P.K. Neuropeptide Y family receptors traffic via
the Bardet‐Biedl syndrome pathway to signal in neuronal primary cilia. Cell
Rep, 2013;5(5):1316–29.
55. Corbit, K.C., Aanstad, P., Singla, V. et al. Vertebrate Smoothened functions
at the primary cilium. Nature, 2005;437(7061):1018–21.
56. Badgandi, H.B., Hwang, S.‐H., Shimada, I.S. et al. Tubby family proteins are
adapters for ciliary trafficking of integral membrane proteins. J Cell Biol,
2017;216(3):743–60.
57. Sun, X., Haley, J., Bulgakov, O.V. et al. Tubby is required for trafficking G
protein‐coupled receptors to neuronal cilia. Cilia, 2012;1(1):21.
58. Lu, L. and Madugula, V. Mechanisms of ciliary targeting: entering importins
and Rabs. Cell Mol Life Sci, 2017;75(4):597–606.
59. Nachury, M.V., Loktev, A.V., Zhang, Q. et al. A core complex of BBS pro-
teins cooperates with the GTPase Rab8 to promote ciliary membrane biogen-
esis. Cell, 2007;129(6):1201–13.
60. Khan, S.A., Muhammad, N., Khan, M.A. et al. Genetics of human Bardet–
Biedl syndrome, an updates. Clin Genet, 2016;90(1):3–15.
61. Berbari, N.F., Lewis, J.S., Bishop, G.A. et al. Bardet‐Biedl syndrome pro-
teins are required for the localization of G protein‐coupled receptors to pri-
mary cilia. Proc Natl Acad Sci U S A, 2008;105(11):4242–6.
62. Mykytyn, K. and Askwith, C. G‐Protein‐coupled receptor signaling in cilia.
Cold Spring Harb Perspect Biol, 2017;9(9):a028183.
63. Masyuk, A.I., Huang, B.Q., Radtke, B.N. et al. Ciliary subcellular localiza-
tion of TGR5 determines the cholangiocyte functional response to bile acid
signaling. Am. J Physiol Gastrointest Liver Physiol,
2013;304(11):G1013–24.
64. Green, J.A., Schmid, C.L., Bley, E. et al. Recruitment of β‐arrestin into neu-
ronal cilia modulates somatostatin receptor subtype 3 ciliary localization.
Mol Cell Biol, 2016;36(1):223–35.
65. Ye, F., Nager, A.R., and Nachury, M.V. BBSome trains remove activated
GPCRs from cilia by enabling passage through the transition zone. J Cell
Biol. 2018;217(5):1847–68.
66. Pal, K., Hwang, S.‐H., Somatilaka, B. et al. Smoothened determines β‐arres-
tin–mediated removal of the G protein–coupled receptor Gpr161 from the
primary cilium. J Cell Biol, 2016;212(7):861–75.
67. Mashukova, A., Spehr, M., Hatt, H., and Neuhaus, E.M. β‐Arrestin2‐medi-
ated internalization of mammalian odorant receptors. J Neurosci,
2006;26(39):9902–12.
68. Luttrell, L.M., Ferguson, S.S., Daaka, Y. et al. Beta‐arrestin‐dependent for-
mation of beta2 adrenergic receptor‐Src protein kinase complexes. Science,
1999;283(5402):655–61.
69. Ye, F., Nager, A.R., and Nachury, M.V. BBSome trains remove activated
GPCRs from cilia by enabling passage through the transition zone. J Cell
Biol, 2018;217(5):1847–68.
70. Wood, C.R. and Rosenbaum, J.L. Ciliary ectosomes: transmissions from the
cell’s antenna. Trends Cell Biol, 2015;25 (5):276–85.
71. LaVail, M.M. Rod outer segment disk shedding in rat retina: relationship to
cyclic lighting. Science, 1976;194(4269):1071–4.
72. Hogan, M.C., Manganelli, L., Woollard, J.R. et al. Characterization of PKD
protein‐positive exosome‐like vesicles. J Am Soc Nephrol,
2009;20(2):278–88.
73. Nager, A.R., Goldstein, J.S., Herranz‐Pérez, V. et al. An actin network dis-
patches ciliary GPCRs into extracellular vesicles to modulate signaling. Cell,
2017;168(1–2):252–63.e14.
74. Phua, S.C., Chiba, S., Suzuki, M. et al. Dynamic remodeling of membrane
composition drives cell cycle through primary cilia excision. Cell,
2017;168(1–2):264–79.e15.
75. Hilgendorf, K.I., Johnson, C.T., and Jackson, P.K. The primary cilium as a
cellular receiver: organizing ciliary GPCR signaling. Curr Opin Cell Biol,
2016;39:84–92.
76. Grzelak, C.A., Martelotto, L.G., Sigglekow, N.D. et al. The intrahepatic sig-
nalling niche of hedgehog is defined by primary cilia positive cells during
chronic liver injury. J Hepatol, 2014;60(1):143–51.
77. Mukhopadhyay, S. and Rohatgi, R. G‐protein‐coupled receptors, Hedgehog
signaling and primary cilia. Semin Cell Dev Biol, 2014;33:63–72.
78. Haycraft, C.J., Banizs, B., Aydin‐Son, Y. et al. Gli2 and Gli3 localize to cilia
and require the intraflagellar transport protein polaris for processing and
function. PLoS Genet, 2005;1(4):e53.
 5:  Primary Cilia 61
79. Liu, J., Zeng, H., and Liu, A. The loss of Hh responsiveness by a non‐ciliary
Gli2 variant. Development, 2015;142(9):1651–60.
80. Nakamura, T. and Gold, G.H. A cyclic nucleotide‐gated conductance in
olfactory receptor cilia. Nature, 1987;325(6103):442–4.
81. Fesenko, E.E., Kolesnikov, S.S., and Lyubarsky, A.L. Induction by cyclic
GMP of cationic conductance in plasma membrane of retinal rod outer
­segment. Nature, 1985;313(6000):310–13.
82. Choi, E.J., Xia, Z., and Storm, D.R. Stimulation of the type III olfactory
adenylyl cyclase by calcium and calmodulin. Biochemistry, 1992;31(28):
6492–8.
83. Pablo, J.L., DeCaen, P.G., and Clapham, D.E. Progress in ciliary ion channel
physiology. J Gen Physiol, 2017;149(1):37–47.
84. Torres, V.E. and Harris, P.C. Strategies targeting cAMP signaling in the
treatment of polycystic kidney disease. J Am Soc Nephrol, 2014;
25(1):18–32.
85. Wann, A.K.T., Zuo, N., Haycraft, C.J. et al. Primary cilia mediate mecha-
notransduction through control of ATP‐induced Ca2+ signaling in compressed
chondrocytes. FASEB J, 2012;26(4):1663–71.
86. Wood, C.R., Huang, K., Diener, D.R., and Rosenbaum, J.L. The cilium
secretes bioactive ectosomes. Curr Biol, 2013;23 (10):906–11.
87. Ott, C., Elia, N., Jeong, S.Y. et al. Primary cilia utilize glycoprotein‐depend-
ent adhesion mechanisms to stabilize long‐lasting cilia‐cilia contacts. Cilia,
2012;1(1):3.
88. McDermott, K.M., Liu, B.Y., Tlsty, T.D., and Pazour, G.J. Primary cilia
regulate branching morphogenesis during mammary gland development.
Curr Biol, 2010;20(8):731–7.
89. Jensen, C.G., Poole, C.A., McGlashan, S.R. et  al. Ultrastructural, tomo-
graphic and confocal imaging of the chondrocyte primary cilium in situ. Cell
Biol Int, 2004;28(2):101–10.
90. Sorokin, S.P. Reconstructions of centriole formation and ciliogenesis in
mammalian lungs. J Cell Sci, 1968;3(2):207–30.
91. Lu, L. and Madugula, V. Mechanisms of ciliary targeting: entering importins
and Rabs. Cell Mol Life Sci, 2017;75(4):597–606.
92. Yang, T.T., Chong, W.M., Wang, W.‐J. et al. Super‐resolution architecture of
mammalian centriole distal appendages reveals distinct blade and matrix
functional components. Nat Commun, 2018;9(1):2023.
93. Sánchez, I. and Dynlacht, B.D. Cilium assembly and disassembly. Nat Cell
Biol, 2016;18(7):711–17.
94. Garcia‐Gonzalo, F.R. and Reiter, J.F. Open sesame: how transition fibers and
the transition zone control ciliary composition. Cold Spring Harb Perspect
Biol, 2017;9(2):a028134.
95. Yadav, S.P., Sharma, N.K., Liu, C. et al. Centrosomal protein CP110 controls
maturation of mother centriole during cilia biogenesis. Development,
2016;143(9):1491–501.
96. Schmidt, K.N., Kuhns, S., Neuner, A. et al. Cep164 mediates vesicular dock-
ing to the mother centriole during early steps of ciliogenesis. J Cell Biol.,
2012;199(7):1083–101.
97. Wei, Q., Xu, Q., Zhang, Y. et al. Transition fibre protein FBF1 is required for
the ciliary entry of assembled intraflagellar transport complexes. Nat
Commun, 2013;4(1):2750.
98. Kubo, T., Brown, Jason M., Bellve, Karl et  al. Together, the IFT81 and
IFT74 N‐termini form the main module for intraflagellar transport of tubulin.
J Cell Sci, 2016;129(10):2106–19.
99. Park, S.M., Lim, J.S., Ramakrishina, S. et al. Brain somatic mutations in
MTOR disrupt neuronal ciliogenesis, leading to focal cortical dyslamina-
tion. Neuron, 2018;99(1):83–97.e7.
100. Anderson, C.T. and Stearns, T. Centriole age underlies asynchronous primary
cilium growth in mammalian cells. Curr Biol, 2009;19(17):1498–502.
101. Paridaen, J.T.M.L., Wilsch‐Bräuninger, M., and Huttner, W.B. Asymmetric
inheritance of centrosome‐associated primary cilium membrane directs
ciliogenesis after cell division. Cell, 2013;155(2):333–44.
102. Bernabé‐Rubio, M. and Alonso, M.A. Routes and machinery of primary
cilium biogenesis. Cell Mol Life Sci, 2017;74 (22):4077–95.
103. He, Q., Wang, G., Dasgupta, S. et al. Characterization of an apical cera-
mide‐enriched compartment regulating ciliogenesis. Mol Biol Cell, 2012;23(16):
3156–66.
104. Wu, B. and Guo, W. The exocyst at a glance. J Cell Sci, 2015;
128(16):2957–64.
105. Bernabé‐Rubio, M., Andrés, G., Casares‐Arias, J. et al. Novel role for the
midbody in primary ciliogenesis by polarized epithelial cells. J Cell Biol,
2016;214(3):259–73.
106. Sharma, N., Kosan, Z.A., Stallworth, J.E., Berbari, N.F., and Yoder, B.K.
Soluble levels of cytosolic tubulin regulate ciliary length control. Mol Biol
Cell, 2011;22(6):806–16.
107. Gadadhar, S., Dadi, H., Bodakuntla, S. et al. Tubulin glycylation controls
primary cilia length. J Cell Biol, 2017;216(9):2701–13.
108. Kim, S. and Dynlacht, B.D. Assembling a primary cilium. Curr Opin Cell
Biol, 2013;25(4):506–11.
109. Chien, A., Shih, S.M., Bower, R. et al. Dynamics of the IFT machinery at
the ciliary tip. Elife, 2017;6:979.
110. Hendel, N.L., Thomson, M., and Marshall, W.F. Diffusion as a ruler: mod-
eling kinesin diffusion as a length sensor for intraflagellar transport.
Biophys J, 2018;114(3):663–74.
111. Mirvis, M., Stearns, T., and Nelson, W.J. Cilium structure, assembly, and disas-
sembly regulated by the cytoskeleton. Biochem J, 2018;475(14):2329–53.
112. Yan, X. and Zhu, X. Branched F‐actin as a negative regulator of cilia formation.
Exp Cell Res, 2013;319(2):147–51.
113. Kohli, P., Höhne, M., Jüngst, C. et  al. The ciliary membrane‐associated
proteome reveals actin‐binding proteins as key components of cilia. EMBO
Rep, 2017;18(9):1521–35.
114. Drummond, M.L., Li, M., Tarapore, E. et al. Actin polymerization controls
cilia‐mediated signaling. J Cell Biol, 2018;217(9):3255–66.
115. Assis, L.H.P., Silva‐Junior, R.M.P., Dolce, L.G. et al. The molecular motor
Myosin Va interacts with the cilia‐centrosomal protein RPGRIP1L. Sci
Rep, 2017;7(1):43692.
116. Ou, Y., Ruan, Y., Cheng, M. et al. Adenylate cyclase regulates elongation of
mammalian primary cilia. Exp Cell Res, 2009;315(16):2802–17.
117. Gerhardt, C., Leu, T., Lier, J.M., and Rüther, U. The cilia‐regulated protea-
some and its role in the development of ciliopathies and cancer. Cilia,
2016;5(1):14.
118. Rieder, C.L., Jensen, C.G., and Jensen, L.C. The resorption of primary cilia
during mitosis in a vertebrate (PtK1) cell line. J Ultrastruct Res,
1979;68(2):173–85.
119. Seeger‐Nukpezah, T., Liebau, M.C., Höpker, K. et  al. The centrosomal
kinase plk1 localizes to the transition zone of primary cilia and induces
phosphorylation of nephrocystin‐1. PLoS One, 2012;7(6):e38838.
 Endocytosis in Liver Function
6 and Pathology
Micah B. Schott, Barbara Schroeder, and Mark A. McNiven
Department of Biochemistry and Molecular Biology, Division of Gastroenterology and Hepatology, Mayo Clinic,
Rochester, MN, USA
INTRODUCTION
A prominent function of the hepatocyte is the regulated uptake
of extracellular material for subsequent processing and/or trans-
port into bile. This process, known as endocytosis, depends on
elaborate vesicle trafficking machinery that is linked to specific
lipid–membrane subdomains and the cytoskeletal matrix. This
provides a mechanism to sequester and internalize transmem-
brane receptor/ligand complexes, such as epidermal growth fac-
tor, hepatocyte growth factor, and iron‐bound transferrin, and to
help maintain normal lipid serum levels through the endocytosis
of low‐density lipoproteins (LDLs). Of equal importance is the
fact that this highly evolved machinery can be “hijacked” by
many pathogens including bacteria, viruses, and parasites to
infect the liver, leading to inflammation and hepatitis. This
review will outline the molecular and cell biological mecha-
nisms that underlie endocytosis in the liver, including the diverse
roles of endocytic vesicles in the cytoplasm and how these path-
ways are altered in liver disease.
ENDOCYTIC VESICLE FORMATION
AT THE PLASMA MEMBRANE
Endocytic vesicles are formed at the plasma membrane as inward
budding events that invaginate toward the cytoplasm. These
membrane structures form vesicles that are packed with various
types of “cargo” such as integral membrane proteins, receptor–
ligand complexes, lipids, fluid, and nutrients. The cargo content
of endocytic vesicles differs considerably between the various
modes of uptake. For example, nonselective endocytic pathways,
such as macropinocytosis, mediate the import of nutrient‐rich
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
extracellular fluid, whereas selective pathways, such as recep-
tor‐mediated endocytosis, import specific soluble ligands and
transmembrane receptors. Regardless of the mechanism of
internalization, endocytic vesicles deliver cargo to the early
­
endosome (EE) for sorting to different destinations, including
recycling back to the cell surface, secretion to the extracellular
milieu, or degradation by late endosomes and lysosomes.
During receptor‐mediated endocytosis, extracellular ligands,
such as LDL, transferrin, epidermal growth factor (EGF), hor-
mones, and many others, bind with high affinity to specific
receptors at the plasma membrane to propagate intracellular
signaling cascades that lead to gene transcription and other
processes. In addition to signal transduction, receptor–ligand
binding also initiates endocytic uptake, an event that “desensi-
tizes” signaling by reducing receptor availability at the cell
surface. More recently, insights into “housekeeping” trophic
receptors, such as the transferrin receptor (TfR) and the LDL
receptor (LDLR), which have been assumed not to play a
prominent role in cell signaling, have been observed to activate
specific signaling cascades.
The formation of endocytic vesicles requires the coordination
of a wide variety of adaptor proteins that link cargo to the
­endocytic machinery and cytoskeleton. As a brief summary, the
various modes of endocytosis and their molecular machinery
are described in the following section.
CLATHRIN‐DEPENDENT ENDOCYTOSIS
Clathrin‐dependent endocytosis is a central and intensely studied
process by which most surface receptors are internalized. The initial
steps of this event require receptor–ligand interactions at the plasma
 6:  Endocytosis in Liver Function and Pathology 63

membrane, which trigger the recruitment of adaptor proteins that
orchestrate the assembly of a clathrin coat (Figure 6.1).
Clathrin‐coated vesicles are relatively uniform in size (100–
150 nm diameter) and are present in all cell types. They were first
observed by Roth and Porter, who noted that the endocytosis of
yolk protein into the oocyte of the mosquito was associated with
a marked increase in invaginations of the oocyte cell membrane,
which was coated on the cytoplasmic face with what they termed
a bristle coat [3]. Soon thereafter, clathrin was identified as the
major protein component of the coat [4].
Clathrin‐dependent endocytosis is a highly coordinated pro-
cess that remains a topic of great research interest. First, plasma
membrane subdomains undergo phospholipid remodeling at
sites of vesicle formation, a step that is critical for the recruit-
ment of adaptors that link surface receptors with clathrin and
other endocytic proteins. Next, inward membrane curvature is
achieved by the actin cytoskeleton and curvature‐sensing BAR
domain proteins. Clathrin is recruited from the cytosol to stabi-
lize the forming vesicle and aid in its displacement from the
plasma membrane. Forming vesicles are separated from the cell
surface by dynamin oligomers that induce vesicle scission by
generating constrictive force around the thin vesicle neck. Once
internalized, the clathrin coat is disassembled to allow for vesi-
cle fusion with downstream target endosomes. All of these steps
occur rapidly on the order of 1–5 minutes and require spatial
and temporal coordination of over 50 different adaptors [1].
A model of these events is depicted in Figure 6.1.
Coated pit/vesicle formation
and associated factors
The plasma membrane is markedly enriched in the phospholipid
phosphatidylinositol 4,5‐bisphosphate (PI(4,5)P2) compared to
other cellular compartments, which aids in recruitment of adap-
tors during the initial stages of vesicle formation. The mecha-
nisms that initiate receptor sequestration and membrane bending
at these sites are not clearly understood. At minimum, these
early steps seem to require the PI(4,5)P2‐binding protein Fes/
CIP4 homology (FCH) domain only 1 and 2 (FCHO1 and
FCHO2), as well as the adaptors epidermal growth factor recep-
tor (EGFR) pathway substrate 15 (Eps15) and intersectins 1 and
2 [1]. FCHO proteins also contain curvature‐sensing BAR
domains that likely aid in their localization at early budding
vesicles. Eps15, along with other cargo‐specific adaptors, helps
to recruit a critical endocytosis protein known as adaptin 2
(AP2), the main link between surface receptors and clathrin.
The role of receptor–ligand interactions in signaling clathrin‐
coated pit formation is unclear, but some studies suggest that
receptors may directly recruit a specific subset of endocytic
adaptors. For example, transferrin receptor contains a tyrosine

(a) Adaptor, cargo

recruitment Membrane bending Scission Uncoating

(b) (c) BAR-domain proteins

(endophilin, amphiphisin)
Synd
apin Cortactin
Actin
Dynamin
Arp2/3

AP2
Clathrin

GPCR RTK
AP2

Arrestin Clathrin

Figure 6.1  Model of clathrin‐mediated endocytosis. (a) Endocytosis occurs at plasma membrane sites of receptor–ligand cargo where clathrin and
clathrin adaptors are recruited. Invaginating membranes are separated from the plasma membrane by scission machinery in cooperation with the
cytoskeleton. Vesicle uncoating allows for participation in downstream endocytic trafficking routes. Cartoon adapted from [1] with permission of
Springer Nature. (b) Scanning electron micrograph of clathrin‐coated vesicles show an intracellular view of clathrin‐coated pits. Reprinted from [2]
with permission of Rockefeller University Press. (c) Components of clathrin‐coated vesicles, including fission machinery and associated actin
cytoskeleton. These components come together to complete the process of cargo sequestration, vesicle formation, and membrane scission.
64 THE LIVER: HEPATOCELLUAR TRAFFICKING OF ENDOCYTIC VESICLES
recognition motif that activates AP2 to recruit clathrin and
stimulate endocytosis [5]. Another example is β‐adrenergic
­ ­phagocytosis that imports extracellular pathogens, and caveolin‐­
receptors that signal the recruitment of the clathrin adaptor β‐
arrestin [6].
Clathrin is recruited from the cytosol to bind AP2 or other
cargo‐specific adaptors that are directly linked to surface recep-
tors destined for internalization. Consisting of both heavy and
light chains, clathrin forms heterodimers that further assemble
into three‐legged triskelions. Clathrin assembly can occur
around nascent vesicle pits, but has also been shown to assemble
at the plasma membrane prior to vesicle formation as flat, planar
lattices [7, 8]. Whereas clathrin heavy chains are capable of
binding to adaptins and other accessory proteins, the light chains
prevent premature lattice assembly in the cytosol and are regu-
lated by calcium binding and phosphorylation [8–13].
Clathrin does not provide sufficient force for inward
­membrane curvature. Thus, inward curvature on the plasma
membrane is thought to require actin polymerization that syner-
gizes with the clathrin coat and vesicle scission machinery. In
yeast, actin binds clathrin adaptors such as Sla2 and Ent1 to
“pull” coated pits inward, while actin polymerization and myo-
sin at the plasma membrane aids in constriction of the endocytic
vesicle neck [1]. BAR domain proteins, such as endophilin and
amphiphysin, also sense membrane curvature and facilitate
endocytic neck constriction prior to vesicle scission [14, 15].
Other adaptors are also known to link the actin cytoskeleton to
the endocytic vesicles, including the actin‐binding proteins pro-
filin, synapsin, syndapin, and cortactin, some of which interact
with the molecular machinery that drives vesicle scission.
Scission of clathrin‐coated vesicles marks their separation from
the plasma membrane. This is accomplished by a family of large
GTPase mechanoenzymes known as dynamins. Dynamin binds to
GTP near its N‐terminus, while membrane attachment is mediated
by a plekstrin homology (PH) domain. Dynamin proteins also bind
a variety of effectors via a C‐terminal proline‐rich domain (PRD)
[16–18], including BAR domain‐containing proteins and cytoskel-
etal adaptors. Dynamin has been referred to as a “pinchase”
because of its ability to generate discrete vesicles from invaginated
coated pits. Dynamin proteins can self‐assemble into oligomeric
rings along lipid vesicles [19] and membrane tubules [20]. In vitro
studies of various dynamin mutations have revealed that GTP
hydrolysis generates the constrictive force for vesicle scission [21].
Following vesicle scission, the clathrin coat is disassembled,
and nascent vesicles fuse with other peripheral endosomes.
Clathrin disassembly requires the ATPase Hsc70 and the remod-
eling of membrane PI(4,5)P2 phospholipids to phosphatidylino-
sitol 4‐phosphate (PI(4)P) by synaptojanin. Hsc70 works
together with its co‐chaperone GAK/auxilin that is recruited in
part by PI(4)P phospholipids [22, 23]. After GAK/auxilin
recruitment, the clathrin coat is released from the vesicle. It is
worth noting that auxilin/GAK activity is not restricted to clath-
rin uncoating but is also required for the dynamin‐dependent
constriction of coated pits [24].
Clathrin‐independent endocytosis
Clathrin‐independent endocytosis is a broad category of
­internalization pathways that can be dynamin dependent and
independent [25]. This includes fluid‐phase endocytosis
that  nonselectively imports extracellular fluid and nutrients,
mediated endocytosis. Although relatively little is known
regarding these mechanisms in comparison to clathrin‐mediated
endocytosis, perhaps the most broadly appreciated is the forma-
tion of caveolae, which are small flask‐shaped vesicles at the
plasma membrane that are enriched in sphingolipids and choles-
terol. In contrast to clathrin‐mediated endocytosis, caveolae are
relatively stable structures at the plasma membrane and are
internalized at a very slow rate [26, 27]. This has led to some
controversy regarding the role of caveolae in hepatocellular
endocytosis, especially given the moderate expression of caveo-
lin, the main structural component of caveolae, in the liver.
Nonetheless, hepatocyte caveolae are internalized by dynamin
scission machinery [28, 29], and caveolin has been shown to
play several important hepatic functions, including lipid metab-
olism and liver regeneration [30].
HEPATOCELLUAR TRAFFICKING
OF ENDOCYTIC VESICLES
Once internalized by either clathrin‐dependent or clathrin‐inde-
pendent endocytosis, endocytic vesicles fuse with EEs that
serve as a central hub for sorting and trafficking of endocytic
cargo. From here, cargo can be directed through multiple sort-
ing pathways, such as recycling back to the plasma membrane,
retrograde transport to the Golgi, degradation by lysosomes and
late endosomes, or secretion from endolysosomal vesicles that
fuse with the plasma membrane (Figure 6.2). Recycling requires
the sorting and concentration of cargo along membrane tubules
that extend from endosomal vesicles. Cargo that is not recycled
can be directly internalized within small intraluminal vesicles
(ILVs), forming distinct endosomal structures known as multi-
vesicular bodies (MVBs). These cargo‐rich ILVs are then
degraded by the late endosomal pathway, marked by the acidifi-
cation of MVBs by fusion with lysosomes. Hepatocytes may
also secrete the contents of MVBs, late endosomes, and lys-
osomes into the bile as these vesicles fuse with the apical plasma
membrane. These diverse events are dictated by the spatial and
temporal coordination of various protein complexes, post‐trans-
lational modifications, phospholipid dynamics, and cytoskeletal
regulators. In addition, it has become increasingly appreciated
that endosomal vesicles play diverse roles in cellular homeosta-
sis beyond that of cargo trafficking [31].
Cargo sorting at early endosomes
Newly formed endocytic vesicles converge at the early endo-
some (EE), also called the “sorting endosome,” that decides
the  fate of endocytosed cargo. The importance of this sorting
process cannot be overstated as it is essential to hepatocyte
function and insures the proper degradation, recycling, or
­transcytosis of trophic and growth factor receptors, bile acid
transporters and other integral membrane proteins [32]. EEs
contain multiple subdomains that are thought to mediate differ-
ent ­
sorting pathways [33]. For example, tubular membrane
extensions arising from specific endosomal subdomains are
 6:  Endocytosis in Liver Function and Pathology 65

Pl(4,5)P2 Basolateral membrane

Rab4
TJ
Rab11 Rab5
WASH
retromer
BC
ESCRT
EHD1
EHD3 Early/sorting
endosome Rab27b
EHD1 MVB
Rab11 EHD3
Endocytic
recycling Pl(3)P Rab27a
Pl(3,4)P2 Rab5
compartment
WASH
retromer Lys

Nucleus Golgi
Rab7
Late
Lys
endosome

Pl(3,5)P2

Figure 6.2  Hepatocyte endocytic trafficking routes and the distribution of Rab GTPases and phosphatidylinositides at the compartments within
the cell. Rab GTPases mediate trafficking from coated vesicles to early/sorting endosomes (Rab5) or late endosomes (Rab7) as well as recycling
back to the plasma membrane directly (Rab4) or indirectly via the endocytic recycling compartment (Rab11). Note that the plasma membrane and
the endosomal pathways are enriched in a subset of phosphatidylinositides that contribute to Rab targeting and additional specificity of the different
compartments. TJ, tight junction; BC, bile canaliculus; MVB, multivesicular body.

destined for recycling, whereas large vesicular subdomains are
thought to “mature” down the late endosomal pathway. Even
within a single EE, there is thought to be variation in pH and
phospholipid composition at different regions that aid in sorting
between recycling, secretory, and degradative routes.
The trafficking routes of endosomal vesicles is largely facili-
tated by a family of small GTPases known as Ras‐associated
binding (Rab) proteins. Approximately 60 different Rab pro-
teins play important roles in membrane and vesicle trafficking
pathways in humans, most of which are expressed in the hepato-
cyte [34, 35]. Rab proteins guide endocytic vesicle trafficking
by binding to effector proteins that modulate endocytic vesicle
functions. For EE function, Rab5 is predominant along with
binding effectors such as phosphoinositide 3‐kinase (PI3K)/
VPS34, EEA1, rabenosyn‐5, and others [36]. PI3K/VPS34
changes the phospholipid composition of EEs to phosphati-
dylinositol 3‐phosphate (PI(3)P), giving these vesicles a distinc-
tive membrane signature that aids in the recruitment of other
Rab5 effectors. EEA1 binds both Rab5 and PI(3)P and mediates
the fusion of endocytic vesicles by coordinating SNARE pro-
teins such as syntaxin6 and syntaxin13. Rabenosyn‐5 is thought
to regulate “fast recycling” of cargo such as the transferrin
receptor, and delivery of cargo to the endocytic recycling center
for “slow recycling” [37].
Cargo sorting is driven by dynamic membrane tubules that
extend from EE subdomains [38]. Although the mechanisms
that form these tubules are still unclear, several lines of evi-
dence suggest that cargo is first concentrated at specific EE
subdomains, followed by actin polymerization that generates
force for tubule initiation and extension. Membrane tubules
further extend along microtubule tracks, and scission machin-
ery separates the tubule from the EE, releasing tubular vesicle
carriers that transport the sorted cargo to the plasma
­membrane, endocytic recycling compartment (ERC), or Golgi
apparatus. Recent work has considerably advanced our under-
standing of the molecular mechanisms underlying each of
these steps.
In order for endocytic sorting to occur, endosomes must be
able to recognize and sequester specific cargo proteins. This
“cargo recognition” process appears to rely on a diverse class of
proteins known as the sorting nexins (Snx) [39]. Snx proteins
contain a conserved PX domain that binds PI(3)P phospholipids
enriched on EEs, as well as cargo‐specific binding domains that
vary between Snx family members. Snx proteins also contain
additional motifs such as BAR domains that sense membrane
curvature, FERM domains that aid in cargo recognition, and
others [40]. Sorting nexins work in concert with the retromer, a
trimeric protein complex composed of Vps26, Vps35, and
Vps29 that is critical to endocytic sorting. More recently, a new
retromer‐like complex known as retriever (consisting of DSCR3,
C160rf62, and VPS29) was also identified to work in concert
with Snx17 [41]. Retriever/Snx17 promotes the recycling of a
subset of cargo molecules that is distinct from retromer‐depend-
ent sorting. Thus, the ability of EEs to accurately decipher dif-
ferent types of cargo, and to organize cargo at endosomal
subdomains, is absolutely critical to hepatocellular vesicle
66 THE LIVER: HEPATOCELLUAR TRAFFICKING OF ENDOCYTIC VESICLES
transport, and involves an impressive synchronization between
cargo adaptors, protein complexes, and the cytoskeleton.
In addition to sorting cargo at endosomal subdomains,
Snx–retromer/retriever complexes coordinate actin dynamics
at these sites by recruiting the WASH complex [40, 42],
which consists of five individual proteins (WASH1, Fam21,
Strumpellin, SWIP, and CCDC53). WASH stimulates Arp2/3
for actin nucleation at EE sites where membrane tubulation
occurs. It is thought that actin generates the force required
for tubule formation, extension, and/or possibly scission.
Fam21 of the WASH complex seems to act as the main teth-
ering link to other pathways. For example, Fam21 is known
to bind Vps35 of the retromer complex [43]. Fam21 also
links the WASH complex to retriever by binding yet another
recently described protein complex known as CCC
(COMMD1, CCDC22, CCDC93) [41].
Early endosomal tubules elongate and move along microtu-
bule tracks. This requires tethering to molecular motor proteins
that travel toward (minus‐end dyneins) or away from (plus‐end
kinesins) the microtubule organizing center (MTOC). The
small GTPase Rab7 facilitates the tethering of EE tubules to
molecular motor proteins in cooperation with Snx‐BAR pro-
teins. The tethering of EE tubules to dyneins versus kinesins
plays an important role in dictating cargo destination. For
example, dynein‐tethered tubules are effectively transported
back to the perinuclear endocytic recycling compartment and/
or Golgi apparatus, whereas kinesin‐tethered tubules are
­transported to the plasma membrane. The scission machinery
that separates nascent tubules from EEs is currently unclear,
but likely involves actin, the ATPase EHD1, and/or dynamin
GTPases.
Regulation of cargo sorting by post‐
translational modifications
As described above, cargo sorting between recycling, degrada-
tion, and secretion pathways relies on various endosomal adap-
tors that bind specific cargo. These cargo–adaptor interactions,
as well as cargo fate, are guided by post‐translational modifica-
tions, such as ubiquitination and phosphorylation, that direct
downstream sorting pathways.
Ubiqutin (Ub) is an 8 kDa, highly conserved protein that can
be covalently linked to lysine residues on target proteins. This
can occur as a polyubiquitin chain or as a single monoubiquitin
attached to one or multiple lysine residues. Whereas ubiquitina-
tion of soluble proteins leads to their degradation by the protea-
some [44, 45], ubiquitination of endocytic cargo is a signal for
MVB internalization and degradation by the late endosomal
pathway (described in a later section). A prototypical example
of this process is the EGFR that is ubiquitinated following
ligand stimulation, which prevents EGFR recycling and pro-
motes its sorting to degradative pathways [46, 47].
Phosphorylation by nonreceptor tyrosine kinases (NRTK)
and by serine–threonine kinases also play important roles in
endocytic cargo trafficking. For example, the NRTK Src was
shown to phosphorylate the EGFR adaptor CIN85 to regulate
EGFR ubiquitination and degradation by the late endosomal
pathway [48]. In addition, the transferrin receptor (TfR) is
regulated by the NRTK c‐Abl kinase. Whereas TfR is normally
recycled back to the plasma membrane, inhibition of c‐Abl
kinase redirects TfR to late endosomes for degradation [49].
Serine–threonine kinases, such as cyclic AMP (cAMP)‐depend-
ent protein kinase A (PKA), are well‐described to regulate
GPCR sorting. For example, activation of the β‐adrenergic
receptor leads to multiple PKA phosphorylation events at its C‐
terminal cytoplasmic tail, which recruits β‐arrestins that induce
GPCR internalization and desensitization [6]. In addition, PKA
phosphorylation of the β‐adrenergic adaptor gravin is essential
for receptor recycling and resensitization back to the plasma
membrane [50–53]. Thus, it is clear that cargo‐specific phos-
phorylation by NRTKs and serine–threonine kinases drive many
important aspects of endocytic regulation, including cargo inter-
nalization and sorting between degradative and recycling
pathways.
Recycling endosomes and the endocytic
recycling compartment
Whereas early endosomal cargo can be recycled directly back to
the plasma membrane (fast recycling), a “slow recycling” path-
way also exists whereby cargo is shunted through tubular recy-
cling endosomes (RE) that are located at the cell periphery or
clustered within the perinuclear endocytic recycling compart-
ment (ERC). REs are enriched in Rab11, distinguishing these
vesicles from Rab5‐positive EEs.
The delivery of cargo from Rab5‐positive EEs to Rab11‐
positive REs seems to rely on a family of Eps15 homology
domain (EHD) proteins that play diverse roles at various stages
of endosomal sorting and recycling [54, 55]. EHD1 is thought
to cause scission of tubular REs arising from the ERC, and
perhaps EE tubules as well [56]. Indeed, EHD1‐depleted cells
show aberrant recycling of the transferrin receptor [57], as do
EHD3 and EHD4, which also aid in the transition of the TfR
from EE to RE [58, 59]. In addition to their well‐described
roles in the RE, EHD proteins also affect sorting at the EEs by
regulating Rab5 and its effectors rabenosyn‐5 and rabakyrin‐5
[58–60].
Tubular REs arising from the perinuclear ERC deliver cargo
back to the plasma membrane. The generation of these struc-
tures is still under investigation, but likely involves a variety of
mechanisms. Interestingly, the WASH complex, which is impor-
tant in the actin‐based budding of EE tubules, is also present on
REs and suggests its role in the generating tubular carriers from
the ERC [61]. However, EHD proteins also promote both the
scission of ERC‐derived tubular carriers (EHD1) and the stabi-
lization of these structures (EHD3). EHD1 seems to operate
within a protein complex containing the endosomal trafficking
adaptors MICAL‐L1 and syndapin2, both of which can generate
tubules from phosphatidylserine (PS)‐rich liposomes in vitro
[62]. The EHD1–MICAL‐L1–syndapin2 complex may also
work on EE‐derived tubular carriers, perhaps even through
binding to the Vps26 subunit of the retromer complex [60, 63,
64]. Thus, the generation of tubules is critical to cargo recycling
directly from EE vesicles and through the ERC, but the synergy
between the different molecular machineries on early endosomes
and REs is still unclear.
 6:  Endocytosis in Liver Function and Pathology 67
Multivesicular bodies and the late
endosomal pathway
Hepatocellular cargo that is not recycled will progress down
the  late endosomal pathway for degradation or secretion into
the  apical bile canaliculus. Due to the acidic pH of late
endosomes and their role in degradation of proteins and lipids,
late endosomes are more akin to lysosomes than EEs. Late
endosomes can be distinguished experimentally from EEs by
their acidic luminal pH, enrichment in PI(3,5)P2 phospholipids,
and association with Rab7 and other protein markers. Although
the transition from Rab5‐positive EEs to Rab7‐positive late
endosomes is not fully understood, the prevailing model is that
EEs themselves “mature” to become gradually more acidic by
fusion with smaller lysosomes that donate acid hydrolases
responsible for the breakdown of proteins, lipids, and nucleic
acids at low pH [65]. At the same time, endosomal cargo that is
destined for degradation will be invaginated within small, <100
nm ILVs. Early and late endosomes containing ILVs are
­identified as multivesicular bodies (MVBs) due to their unique
morphology. Cargo‐rich ILVs are degraded by lysosomal acidi-
fication via the late endosomal pathway, as is the case with the
EGFR, but MVBs may also fuse with the plasma membrane to
secrete ILVs as extracellular vesicles/exosomes [66].
Ubiquitination of endosomal cargo serves as the main signal for
sequestration into MVB ILVs [67]. Cargo that is destined for MVB
internalization relies on endosomal sorting complexes required
for transport (ESCRT‐0, ‐I, ‐II, and ‐III) that recognize ubiquitin‐
tagged membrane cargo for sorting into endosomal subdomains.
Ubiquitin recognition is achieved by ESCRT‐0, ESCRT‐I, and
ESCRT‐II components that contain ubiquitin‐binding motifs and
cluster cargo in cooperation with endosomal clathrin lattices and
the actin cytoskeleton. ESCRT‐III contains no ubiquitin recogni-
tion, but oligomerizes as a concentric spiral around clustered cargo
and is presumed to generate force, much like the action of a spring,
to push cargo inward into the lumen of the MVB [67, 68]. ESCRT‐
III filaments are subsequently disassembled by the action of the
ATPase Vps4, and this step is believed to mediate the final scission
of invaginated ILVs into the MVB lumen. It should be noted that
while ubiquitin modifications are important for cargo internaliza-
tion into MVB, recent studies indicate that ESCRTs also partici-
pate in ubiquitin‐independent cargo sorting. This may involve
direct binding of cargo to ESCRTs, ESCRT adaptors, and/or endo-
somal phospholipids [69].
Cargo‐rich ILVs are degraded by the late endosomal pathway,
but may be spared from degradation if the MVB fuses with the
plasma membrane. This unconventional secretory method
releases ILVs as extracellular vesicles (EVs), or “exosomes,”
into the extracellular milieu. Indeed, EVs have been detected in
blood, bile, and urine, indicating the potentially widespread
nature of this pathway to transmit cellular information through-
out the body. Various EV analytical studies have demonstrated
that these vesicles contain much more than integral membrane
cargo, but contain specific classes of lipids and soluble cytosolic
materials such as microRNAs and various protein enzymes. In
liver pathology, EVs are thought to play a role in tissue signal-
ing cell stress during non‐alcoholic steatohepatitis (NASH) and
other hepatic insults [70, 71]. Although it is clear that EVs play
an important role in tissue homeostasis and disease progression,
many questions remain regarding the mechanisms that package
soluble cytosolic materials into ILVs, the role of EVs in cell
communication, and their potential synthetic use in drug deliv-
ery systems.
Lysosomes/late endosomes
As described above, lysosomes and late endosomes share some
overlapping functions in hepatocytes. In conjunction with their
role in degrading proteins and lipids, lysosomes also serve
as  nutrient sensors and platforms for signal transduction.
Lysosomal nutrient sensing is largely mediated by the presence
or absence of amino acids that regulate the mammalian target of
rapamycin (mTOR), a nutrient‐sensing kinase that is critical in
metabolic liver diseases such as non‐alcoholic fatty liver disease
(NAFLD), NASH, and hepatocellular carcinoma [72]. mTOR
operates as part of two distinct complexes: mTOR complex 1
(mTORC1) and mTOR complex 2 (mTORC2). While both of
these complexes regulate specific facets of cell growth and pro-
liferation, mTORC1 is functionally linked to nutrient sensing at
the lysosome/late endosome. Amino acids stimulate the lysoso-
mal recruitment of mTORC1 to regulate enzymes involved in
protein synthesis (ribosomal p70 S6 kinase and 4‐EBP1), lipid
synthesis (SREBP1) and others. Active mTORC1 also sup-
presses “starvation” pathways such as autophagy (discussed in
more detail later) by the inhibitory phosphorylation of Ulk1 and
transcription factor EB (TFEB). In general, active mTORC1
signifies a “fed” state of nutrient abundance and thus activates
pathways related to cell growth and proliferation, while simulta-
neously suppressing pathways that are important during “starved”
states where nutrient supply is limited.
Lysosomes and late endosomes play important roles in cell
signaling by storing intracellular calcium (Ca2+) that serves as a
second messenger for signal transduction. Since Ca2+ concentra-
tions in the cytoplasm are very low, release of lysosomal cal-
cium by lysosomal calcium channels can trigger nearby
substrates. For example, lysosomal calcium release can activate
TFEB, a master transcriptional regulator of lysosome biogene-
sis, lipid catabolism, autophagy, and mitochondrial β‐oxidation
[73]. Under basal conditions, TFEB is phosphorylated by
mTORC1 and is sequestered in the cytosol. However, lysosomal
calcium release activates the phosphatase calcineurin, leading to
TFEB dephosphorylation and translocation into the nucleus. A
similar pathway was later described for TFE3, which is closely
related to TFEB, in the regulation of lysosome biogenesis and
energy metabolism [74]. Thus, the signaling events propagated
by lysosomes/late endosomes are closely related to their
­important roles as nutrient sensors in energy metabolism and
autophagy.
Hepatocyte polarity in endocytic trafficking
Hepatocytes possess a unique epithelial polarity that contributes
to specialized endocytic vesicle trafficking routes [75]. In tradi-
tional epithelium, cells possess an apical plasma membrane that
faces the luminal space, a lateral domain that contacts adjacent
epithelial cells, and a basal domain anchored to the basal ­lamina.
In hepatocytes, the apical (canalicular) domain faces the bile
canalicular lumen, and this domain is separated from the basal
68 THE LIVER:  AUTOPHAGY AND THE ENDOCYTIC MACHINERY
and lateral domains by tight junctions. The basolateral (sinusoi-
dal) domain interfaces with adjacent hepatocytes and the perisi-
nusoidal space of Disse that contains blood plasma, fenestrated
endothelial cells, and hepatic stellate cells.
A subset of endocytic cargo undergoes a unique hepatocyte‐
specific mode of transcytosis (Figure 6.2), whereby endocytic
vesicles formed at the basolateral domain eventually deposit
their contents at the apical bile canaliculus via specialized, sub-
apical endosomes [32]. Lysosomes also secrete cytosolic mate-
rials into the bile through an upstream pathway called autophagy
(discussed in the next section) [76]. Together with late
endosomes, these vesicles deliver bile acids, EVs, phospholip-
ids, cholesterol, polymeric immunoglobulin A (pIgA – part of
mucosal system), and other materials into the bile [31].
AUTOPHAGY AND THE ENDOCYTIC
MACHINERY
Beyond cargo internalization and sorting, endocytic vesicles
play additional roles in cell physiology [31]. A prominent
example of this is the intersection of late endosomes with the
process of autophagy, a “self‐eating” pathway that degrades
cytosolic materials (whole organelles, lipids, proteins, and
other materials) to meet energy demands during periods of low
nutrient availability [77, 78]. Autophagy also plays a critical
role in liver function by clearing damaged proteins and orga-
nelles that induce cellular stress. Dysregulation of autophagy
contributes to a wide range of liver diseases, including steato-
sis, steatohepatitis, cirrhosis, cancer, drug‐induced liver injury,
and others [reviewed in 79].
Autophagy mediates the transport of cytosolic materials to the
lysosome for degradation. While different subtypes of autophagy
exist, the canonical form, macroautophagy, is often referred to as
simply “autophagy.” This process can be activated in the liver by
prolonged fasting as well as glucagon stimulation, and is sup-
pressed by amino acids and/or insulin. These pathways are
thought to converge at the regulation of mTOR activity. mTOR
attenuates autophagy by inhibiting the Atg conjugation system
that drives the formation of double‐membrane phagophores
[80]. Phagophores are tubular membranes derived from the
endoplasmic reticulum (ER) and a variety of other cellular
sources such as Golgi, plasma membrane, and endosomal
vesicles [81–83]. These membranous structures (as well as
­ Autophagy is activated by excessive APAP to clear damaged
autophagosomes and autolysosomes) are decorated by a key
autophagy protein known as microtubule‐associated protein 1
light chain 3, or LC3, which is “activated” by the Atg conjuga-
tion system during autophagy initiation and links cytosolic
materials to autophagic membranes. LC3 accomplishes this task
by interacting with a family of autophagy receptors (p62/
SQSTM1, NBR1, OPTN, and others) that recognize ubiquity-
lated cargo in the cytoplasm, but LC3 may also directly bind to
autophagic cargo proteins containing LC3‐interacting regions
(LIR). The binding of LC3 to autophagy receptors aids in trap-
ping cytosolic materials to the phagophore, as this membrane
extends to envelope the autophagic cargo. When fully engulfed,
the phagophore closes to become a double‐membrane auto­
phagosome. Autophagosomes sequester autophagic cargo from
the cytoplasm prior to fusion with lysosomes/late endosomes.
This fusion results in an acidic “autolysosome” in which the
autophagic cargo is degraded.
Endosomal vesicles contribute to several stages of autophagy
[77]. Most critical is the contribution of acidic vesicles derived
from the endocytic pathway (i.e. lysosomes, late endosomes)
that fuse with autophagosomes to form autolysosomes that
degrade autophagic cargo. In hepatocytes, the late endosomal
Rab7, as well as the membrane scission protein dynamin2, are
essential autophagy proteins that mediate a process known as
autophagic lysosomal reformation (ALR), whereby nascent lys-
osomes are generated from recycling tubules that extend from
autolysosomes and late endosomes [84]. Rab7 has also been
implicated in the fusion of autophagosomes with lysosomes/late
endosomes [85]. In addition to late endosomes, other endoso-
mal vesicles are also known to play important roles in autophagic
progression. For example, EEs serve as one of the membrane
sources for phagophore formation [86]. Rab11‐positive REs are
known to harbor various components of the autophagic machin-
ery, and Rab11 itself has been implicated in the fusion of
autophagosomes with MVBs, forming a structure known as the
amphisome [87, 88].
Endosomal vesicles also facilitate other forms of autophagy
that are independent of autophagosome formation [89]. One
prominent example of this is chaperone‐mediated autophagy,
whereby lysosomes directly target cytosolic proteins that con-
tain a pentapeptide (KFERQ) motif that binds to the chaperone
Hsc70. This interaction facilitates the translocation of substrate
proteins into the lysosomal lumen via lysosome‐associated
membrane protein type 2A [90]. Dysregulation of this process
in mouse livers leads to gross metabolic dysregulation,
­especially the accumulation of hepatocellular lipids and the
depletion of glycogen stores [91].
Mitochondrial autophagy protects against
liver injury
A prominent function of the liver is clearance of toxins from the
blood by metabolizing blood alcohol, pharmaceutical drugs,
and other nutrients. Accordingly, liver injury is a prominent
concern in drug development and accounts for >50% of acute
liver failure [79]. The importance of autophagy in this process
has been observed particularly in the context of acetaminophen
(APAP) overdose, which induces severe liver injury [92].
mitochondria that produce reactive oxygen species (ROS), lead-
ing to hepatocellular death. The degradation of mitochondria by
autophagy is a protective mechanism, as pharmacological stim-
ulation of autophagy protects against APAP‐induced liver injury
in mice [93].
The selective autophagy of damaged mitochondria, termed
“mitophagy,” seems to utilize both autophagic and endosomal
machineries. The process is driven by mitochondrial fission that
produces smaller sized mitochondria that are more susceptible
to autophagic engulfment [94]. In canonical mitophagy,
damaged mitochondria accumulate PTEN‐induced putative
­
kinase 1 (PINK1) that recruits E3 ubiquitin ligase parkin to the
­mitochondrial outer membrane [95]. This generates ubiquitin
signatures on the mitochondrial surface that are recognized by
 6:  Endocytosis in Liver Function and Pathology 69
autophagy receptors that recruit LC3‐positive phagophores for
engulfment within the autophagosome [95, 96]. Recent reports
indicate that this engulfment may also be mediated directly
by  endosomal vesicles even in the absence of traditional
autophagosome‐based mitophagy [97, 98]. In this model,
­ and other endocytic machinery play important roles in this pro-
termed “endosomal mitophagy,” parkin recruits damaged mito-
chondria to Rab5‐positive EEs that utilize the ESCRT machin-
ery to bind ubiquitylated mitochondria. ESCRT recognition of
ubiquitylated mitochondria leads to an MVB‐like mitochondrial
engulfment and degradation by the late endosomal pathway.
In addition to mitophagy, endosomal vesicles have recently
been reported to play a novel role in mitochondrial fission and
fusion. These findings suggest that lysosomes make an intimate
connection with mitochondria, especially at sites of mitochon-
drial fission [99]. The tethering of mitochondria to lysosomes
requires active, GTP‐bound Rab7 on the lysosome, whereas
untethering requires a protein known as TBC1D15 that inacti-
vates Rab7 by facilitating GTP hydrolysis and dissociation.
Interestingly, TBC1D15 is recruited to the mitochondria to reg-
ulate Rab7 activity, suggesting a bidirectional mechanism
whereby mitochondria regulate lysosomal positioning via Rab7
inactivation, and lysosomes act on mitochondria to mediate
fission.
Autophagy and lipid droplet homeostasis
The liver is a central hub for the storage and regulation of neutral
lipids such as triglycerides and cholesterol esters. These lipids
are assembled from free fatty acids (FFA) that are synthesized de
novo or imported from extracellular sources (i.e. dietary fat,
lipids released from adipose tissue) and stored within specialized
organelles known as lipid droplets (LDs) [32]. Dysregulation of
these important hepatocellular functions leads to profound lipo-
dystrophies that impact both liver function and whole‐body lipid
homeostasis. Most common are fatty liver diseases that affect
20–30% of the world population and coincide with other meta-
bolic diseases (obesity, diabetes, metabolic syndrome) and/or
chronic alcohol consumption [100, 101]. Central to hepatic fat
storage and energy utilization, LDs store intracellular neutral
lipids surrounded by a phospholipid monolayer that associates
with proteins involved in lipid storage, breakdown, and fusion
[102–104]. In addition, several proteomic studies have shown
that LDs associate with numerous Rab GTPases, as well as
endosomal vesicle machinery found on endosomes, multivesicu-
lar bodies, and lysosomes [105–109].
LDs are catabolized by two different types of lipases – ­neutral
lipases that reside in the cytoplasm and acid lipases that reside
in lysosomes/late endosomes. Neutral lipases, such as adipose
triglyceride lipase (ATGL) and hormone‐sensitive lipase (HSL),
are recruited from the cytosol to the LD following β‐adrenergic
receptor activation of the cAMP/PKA pathway. It was once
thought that these neutral lipases played a more prominent role
in adipose tissue and skeletal muscle, but numerous studies have
now demonstrated their importance in liver function and meta-
bolic disease pathology [110–114]. Acid lipases such as lysoso-
mal acid lipase (LAL) are active within the low pH environment
of lysosomes/late endosomes, and thus cannot be recruited to
the LD surface in the same manner as the ­cytosolic lipases. LAL
gains access to LDs via lipophagy (Figure  6.3), a selective
autophagic mechanism for LD breakdown [115–117]. Although
the mechanisms that traffic hepatocellular LDs from the cyto-
plasm to the lumen of lysosomes and late endosomes are not
completely understood, it is clear that numerous Rab proteins
cess. For example, both Rab7 and Rab10 are activated on the
LD surface during periods of low nutrient availability. Rab10
was shown to form a trimeric complex with its effector Eps15
homology binding protein 1 (EHBP1) and EHD2 on the LD sur-
face, and this complex is important for phagophore recruitment
and extension around the LD [118]. Rab7 drives LD interaction
with multivesicular bodies and late endosomes marked by the
membrane tetraspanin CD63 [119]. The direct interaction of
LDs with MVBs/endosomes raises the possibility that alterna-
tive forms of “microlipophagy” may exist in parallel with tra-
ditional autophagosome‐based “macrolipophagy” (Figure  6.3)
Interestingly, Rab7 was also shown to be inhibited by chronic
ethanol exposure, which alters lysosomal morphology and
motility to perturb the lipophagic catabolism of LDs during
alcoholic fatty liver progression [120]. Other endocytic proteins
also appear to be involved in hepatocellular lipophagy, includ-
ing clathrin, dynamin2, and Vps4 [84, 119].
In addition to lipophagy, several studies now show a bidirec-
tional synergy between autophagy and cytosolic lipases. First,
various autophagy pathways are known to promote ATGL
­activity in hepatocytes to protect against fatty liver progression.
For  example, chaperone‐mediated autophagy was recently
shown to degrade LD proteins that inhibit ATGL‐mediated
lipolysis [121]. Macroautophagy also seems to facilitate
­hepatocyte lipolysis, as both HSL and ATGL contain several
LC3‐interacting regions that link these cytosolic lipases to
autophagosomes that function to deliver HSL and ATGL to the
surface of LDs [122]. Conversely, ATGL‐mediated lipolysis has
been reciprocally shown to induce autophagy at the transcrip-
tional level. In this model, ATGL activity releases FFAs
that  serve as signaling ligands for PGC‐1α and PPAR‐α. The
­mechanism for this relies on the activation of NAD‐dependent
deacetylase sirtuin 1, which facilitates the interaction of PPAR‐α
with deacetylated PGC‐1α to promote autophagic gene tran-
scription and other metabolic programs [123, 124].
VIRAL INFECTION
AND HEPATOCELLULAR
ENDOCYTOSIS
Human pathogens such as hepatitis B virus (HBV) and hepatitis
C virus (HCV) are a leading cause for liver diseases such as
steatohepatitis, cirrhosis, and hepatocellular carcinoma. These
viruses hijack normal hepatocellular processes, particularly host
endocytic pathways that are utilized for infection, propagation,
and  secretion to neighboring cells and tissues [reviewed in
125].  Five hepatitis viruses have been described that hijack
different  endocytic pathways for replication and secretion:
­
HBV  (Hepadnaviridae family), HDV (Deltaviridae), HCV
(Flaviviridae family), HAV  (Hepatovirus Picornaviridae),
and  HEV (Hepervirus Heperviridae). Although the genetic
and  molecular components of these hepatic viruses vary
70 THE LIVER: VIRAL INFECTION AND HEPATOCELLULAR ENDOCYTOSIS

(a) Plasma membrane

Clathrin-mediated
endocytosis Microlipophagy
LD
?

MVB/Late
EE endosome

Macrolipophagy

ALR
LD Lysosome
Dyn2
Phagophore Autophagosome Autolysosome

(b) (c)

Figure 6.3  Lipophagy in hepatocytes. (a) A working model of two distinct lipophagy pathways. Microlipophagy is proposed to occur by the direct
uptake of lipid droplets by endosomal vesicles such as multivesicular bodies (MVBs) and late endosomes for degradation by lysosomal enzymes.
Macrolipophagy utilizes traditional autophagy machinery whereby lipid droplets (LDs) are targeted by phagophores for complete engulfment within
autophagosomes which fuse with lysosomes to become degradative autolysosomes. The terminal stages of macrolipophagy incorporate autolyso-
some reformation (ALR), whereby membrane tubules extending from the autolysosome undergo scission by dynamin 2 (Dyn2). This generates
nascent lysosomes that contribute to autophagic and late endosomal degradation. (b) Transmission electron micrograph shows an LD (white arrow)
encased within an autophagic membrane in Huh7 human hepatoma cells. Reproduced from [118] with permission of American Association for the
Advancement of Science. (c) Electron micrograph of a siRNA Dyn2‐depleted Hep3B human hepatoma cell shows an autolysosome tubule (white
arrows) that is elongated in the absence of tubule scission machinery. Reproduced from [84] with permission of Rockefeller University Press.

considerably, each virus consists of a nucleotide genome of
ssRNA (D,C,A,E) or dsDNA (B) that is encapsulated by a pro-
teinaceous core and surrounded by an outer envelope consisting
of proteins and host‐derived lipids. A model for how each of
these viruses are assembled as they advance through hepatocyte
endocytic pathways is depicted in Figure 6.4.
Viral attachment and endocytosis
Hepatitis viruses attach to the cell surface and can move
­laterally along the plasma membrane prior to endocytosis. For
HBV, this interaction is facilitated by hepatocyte heparin sul-
fate proteoglycans (HSPGs) and by sodium taurocholate
cotransporting polypeptide (NTCP), both of which bind to
HBV envelope ­proteins. HCV attachment is also facilitated by
HSPGs, but has also been reported to interact with apolipopro-
teins, junctional proteins, surface receptors, and others that help
facilitate membrane attachment. HAV or HEV attachment to
the  ­hepatocyte is not well understood and seems to require
HSPGs, asialoglycoprotein receptor (ASPGR), and others
[reviewed in 125].
Hepatic viruses utilize clathrin‐mediated endocytosis for
hepatocellular uptake, but the contributions of other endocytic
routes such as caveolin‐mediated endocytosis, pinocytosis, and
phagocytosis cannot be ruled out [125]. Upon internalization,
these viruses traverse through hepatocyte endocytic pathways
prior to genome release into the nucleus via the cytoplasm. For
HBV, genome release into the nucleus is pH dependent and
requires early‐to‐late endosomal transitions mediated by Rab5
and Rab7 [126]. The mechanisms supporting these important
steps, including how these viruses evade late endosomal degra-
dation, remain unclear. In contrast to HBV and other hepatic
viruses, the HCV genome is translated in the cytoplasm follow-
ing release from EEs in a pH‐dependent manner [127].
 6:  Endocytosis in Liver Function and Pathology 71
Virus BCDE
C
HSPGs Clathrin-mediated
endocytosis
Endosome
B ACDE
RNA
Nucleus C
Golgi
Viral
genome
DNA
B C?
RNA
ER
Figure 6.4  Model depicting how the different hepatitis viruses (ABCDE) utilize common and distinct endocytic pathways in the hepatocyte
­during internalization, infection, maturation, and release. Modified from [125] with permission of John Wiley & Sons.
Viral assembly, replication,
and hepatocellular secretion
HBV‐infected cells secrete both infectious virions and non‐
infectious subviral particles (SVPs) made up of envelope pro-
teins that likely serve as “decoys” against the host immune
system. SVPs are assembled in the ER and packaged for secre-
tion within the ER–Golgi intermediate compartment (ERGIC)
[128]. For HCV, assembly occurs within ER‐derived double‐
membrane vesicles (~150 nm in diameter) that are constructed
in response to viral infection. Interestingly, HCV also stimulates
the formation of LDs that are contained within these vesicles
[129]. LDs are thought to facilitate HCV replication, as core and
nonstructural proteins are found on the LD surface. LDs may
also contain viral double‐stranded RNA, suggesting LDs are
also a site of genomic replication. Other work has shown that
LD‐associated proteins, such as Rab18 [130] and Plin3/Tip47
[131], also help facilitate HCV replication.
Hepatitis viruses utilize different endo‐membrane vesicle
pathways for secretion. For example, HCV secretion requires
trans‐Golgi machinery and Rab11, suggesting interplay between
the Golgi and RE compartments [132]. Other viruses, such as
HBV, utilize the MVB for uptake and secretion through a mech-
anism that requires the ESCRT machinery. It was recently found
that HBV induces the dramatic tubulation of MVBs and
autophagosomes by modulating Rab7 activity, which stimulates
lysosomal fusion and pH‐dependent secretion [133].
Recent studies reveal that hepatic viruses utilize late ­endocytic
compartments for maturation and release. Some of these viruses
(HAV, HEV) utilize MVB‐derived membranes to envelop nas-
cent virions as their genomes do not encode envelope proteins.
In addition to serving as a membrane source, MVBs also gener-
ate ILVs that support an exosome‐like shedding of hepatic
viruses. Trafficking of viruses to MVBs, as well as MVB fusion
with the plasma membrane, seems to require the participation of
several Rab proteins including Rab2b, Rab5a, Rab9b, Rab27a,
ABE
C?
Autophagosome
MVB
B
ESCRT:
Lysosome
ABE
and Rab27b. These later Rabs are particularly important in
MVB‐based secretion, as Rab27a is thought to mediate MVB
docking/fusion with the plasma membrane, whereas Rab27b
participates in the transfer of MVBs from microtubules to the
actin‐rich regions at the cell periphery [134].
In conclusion, hepatitis viruses utilize hepatocyte membrane
trafficking machinery for infection and propagation. Despite the
substantial progress that has been made toward understanding
how the host hepatocellular endocytic pathways are modified by
these viruses, future work will need to further define the syn-
ergy between the endosomal compartments, nucleus, ER, Golgi,
and the autophagic machinery in the viral life cycle.
FUTURE DIRECTIONS
New imaging and biochemical techniques have provided
detailed insights into the mechanisms of vesicle formation and
post‐endocytic trafficking. Nevertheless, the list of protein and
lipid components of the vesicle‐forming machinery continues
to expand while the functions of many endocytic components
in specific transport processes are yet to be established. It is
important to understand how the hepatocyte manages to
­coordinate and control a massive protein and lipid network that
supports vesicular trafficking events. In this context, ­challenges
for the future will be to understand how different endocytic
cargos are sequestered and sorted, how membrane subdomains
are maintained even within individual vesicles, and which
phospho‐substrates are targeted by kinase signaling cascades
to control endocytic trafficking events. Equally exciting will
be  the elucidation of alternative roles for endocytic vesicles
in  processes such as the autophagy of mitochondria and
LDs.  Insights into these processes will provide a foundation
for  understanding the function of hepatocytes in health and
disease.
72 THE LIVER:  REFERENCES

REFERENCES 27. Pelkmans, L., Burli, T., Zerial, M., and Helenius, A. Caveolin‐stabilized
membrane domains as multifunctional transport and sorting devices in endo-
cytic membrane traffic. Cell, 2004;118(6):767–80.
1. Kaksonen, M. and Roux, A. Mechanisms of clathrin‐mediated endocytosis.
28. Yao, Q., Chen, J., Cao, H. et al. Caveolin‐1 interacts directly with dynamin‐2.
Nature reviews Mol Cell Biol, 2018;19(5):313–26.
J Mol Biol, 2005;348(2):491–501.
2. Heuser, J. Effects of cytoplasmic acidification on clathrin lattice morphol-
29. Henley, J.R., Krueger, E.W., Oswald, B.J., and McNiven, M.A. Dynamin‐
ogy. J Cell Biol, 1989;108(2):401–11.
mediated internalization of caveolae. J Cell Biol, 1998;141(1):85–99.
3. Roth, T.F. and Porter, K.R. Yolk protein uptake in the oocyte of the mosquito
30. Fernandez‐Rojo, M.A. and Ramm, G.A. Caveolin‐1 function in liver physi-
Aedes aegypti. L. J Cell Biol, 1964;20:313–32.
ology and disease. Trends Mol Med, 2016;22(10):889–904.
4. Pearse, B.M. Clathrin: a unique protein associated with intracellular transfer
31. Schroeder, B. and McNiven, M.A. Importance of endocytic pathways in liver
of membrane by coated vesicles. Proc Natl Acad Sci U S A, 1976;73(4):
function and disease. Compr Physiol, 2014;4(4):1403–17.
1255–9.
32. Schulze, R.J. et al. The cell biology of the hepatocyte. J Cell Biol,
5. Kadlecova, Z., Spielman, S.J., Loerke, D. et  al. Regulation of clathrin‐­
2019;218(7):2096–112.
mediated endocytosis by hierarchical allosteric activation of AP2. J Cell
33. Short, B. Sorting out endosome form and function. J Cell Biol,
Biol, 2017;216(1):167–79.
2015;210(6):870.
6. Goodman, O.B., Jr., Krupnick, J.G., Santini, F. et al. Beta‐arrestin acts as a
34. Zhen, Y. and Stenmark, H. Cellular functions of Rab GTPases at a glance.
clathrin adaptor in endocytosis of the beta2‐adrenergic receptor. Nature,
J Cell Sci, 2015;128(17):3171–6.
1996;383(6599):447–50.
35. Stenmark, H. Rab GTPases as coordinators of vesicle traffic. Nat Rev Mol
7. Larkin, J.M., Donzell, W.C., and Anderson, R.G. Potassium‐dependent
Cell Biol, 2009;10(8):513–25.
assembly of coated pits: new coated pits form as planar clathrin lattices.
36. Jovic, M., Sharma, M., Rahajeng, J., and Caplan, S. The early endosome: a
J Cell Biol, 1986;103(6 Pt 2):2619–27.
busy sorting station for proteins at the crossroads. Histol Histopathol.
8. Heuser, J. Three‐dimensional visualization of coated vesicle formation in
2010;25(1):99–112.
fibroblasts. J Cell Biol, 1980;84(3):560–83.
37. Navaroli, D.M., Bellve, K.D., Standley, C. et  al. Rabenosyn‐5 defines the
9. Ungewickell, E. and Ungewickell, H. Bovine brain clathrin light chains
fate of the transferrin receptor following clathrin‐mediated endocytosis. Proc
impede heavy chain assembly in vitro. J Biol Chem, 1991;266(19):
Natl Acad Sci U S A, 2012;109(8):E471–80.
12710–14.
38. van Weering, J.R. and Cullen, P.J. Membrane‐associated cargo recycling by
10. Vigers, G.P., Crowther, R.A., and Pearse, B.M. Three‐dimensional structure
tubule‐based endosomal sorting. Semin Cell Dev Biol, 2014;31:40–7.
of clathrin cages in ice. EMBO J, 1986;5(3):529–34.
39. Cullen, P.J. Endosomal sorting and signalling: an emerging role for sorting
11. Nathke, I., Hill, B.L., Parham, P., and Brodsky, F.M. The calcium‐binding
nexins. Nat Rev Mol Cell Biol, 2008;9(7):574–82.
site of clathrin light chains. J Biol Chem, 1990;265(30):18621–7.
40. Wang, J., Fedoseienko, A., Chen, B. et al. Endosomal receptor trafficking:
12. Hill, B.L., Drickamer, K., Brodsky, F.M., and Parham, P. Identification of
retromer and beyond. Traffic, 2018;19(8):578–90.
the  phosphorylation sites of clathrin light chain LCb. J Biol Chem,
41. McNally, K.E. et al. Retriever is a multiprotein complex for retromer-inde-
1988;263(12):5499–501.
pendent endosomal cargo recycling. Nat Cell Biol, 2017;19(10):1214–25.
13. Schmid, S.L., Braell, W.A., Schlossman, D.M., and Rothman, J.E. A role for
42. Duleh, S.N. and Welch, M.D. WASH and the Arp2/3 complex regulate
clathrin light chains in the recognition of clathrin cages by ‘uncoating
endosome shape and trafficking. Cytoskeleton (Hoboken), 2010;67(3):
ATPase.’ Nature, 1984;311(5983):228–31.
193–206.
14. Meinecke, M., Boucrot, E., Camdere, G. et  al. Cooperative recruitment of
43. Harbour ME., Breusegem, S.Y., and Seaman, M.N. Recruitment of the endo-
dynamin and BIN/amphiphysin/Rvs (BAR) domain‐containing proteins leads
somal WASH complex is mediated by the extended “tail” of Fam21 binding
to GTP‐dependent membrane scission. J Biol Chem, 2013;288(9):6651–61.
to the retromer protein Vps35. Biochem J, 2012;442(1):209–20.
15. Dawson, J.C., Legg, J.A., and Machesky, L.M. Bar domain proteins: a role
44. Elsasser, S. and Finley, D. Delivery of ubiquitinated substrates to protein‐
in tubulation, scission and actin assembly in clathrin‐mediated endocytosis.
unfolding machines. Nat Cell Biol, 2005;7(8):742–9.
Trends Cell Biol, 2006;16(10):493–8.
45. Miller, J. and Gordon, C. The regulation of proteasome degradation by multi‐
16. Kim, Y. and Chang, S. Ever‐expanding network of dynamin‐interacting pro-
ubiquitin chain binding proteins. FEBS Lett, 2005;579(15):3224–30.
teins. Mol Neurobiol, 2006;34(2):129–36.
46. Huang, F., Goh, L.K., and Sorkin, A. EGF receptor ubiquitination is not
17. McNiven, M.A., Cao, H., Pitts, K.R., and Yoon, Y. The dynamin family
­necessary for its internalization. Proc Natl Acad Sci U S A, 2007;104(43):
of  mechanoenzymes: pinching in new places. Trends Biochem Sci,
16904–9.
2000;25(3):115–20.
47. Ravid, T., Heidinger, J.M., Gee, P., Khan, E.M., and Goldkorn, T. c‐Cbl‐
18. Praefcke, G.J. and McMahon, H.T. The dynamin superfamily: universal
mediated ubiquitinylation is required for epidermal growth factor receptor
membrane tubulation and fission molecules? Nat Rev Mol Cell Biol,
exit from the early endosomes. J Biol Chem, 2004;279(35):37153–62.
2004;5(2):133–47.
48. Schroeder, B., Srivatsan, S., Shaw, A., Billadeau, D., and McNiven, M.A.
19. Tuma, P.L. and Collins, C.A. Dynamin forms polymeric complexes in the
CIN85 phosphorylation is essential for EGFR ubiquitination and sorting into
presence of lipid vesicles. Characterization of chemically cross‐linked
multivesicular bodies. Mol Biol Cell, 2012;23(18):3602–11.
dynamin molecules. J Biol Chem, 1995;270(44):26707–14.
49. Cao, H., Schroeder, B., Chen, J., Schott, M.B., and McNiven, M.A. The
20. Takei, K., McPherson, P.S., Schmid, S.L., and De Camilli, P. Tubular mem-
endocytic fate of the transferrin receptor is regulated by c‐Abl kinase. J Biol
brane invaginations coated by dynamin rings are induced by GTP‐gamma S
Chem, 2016;291(32):16424–37.
in nerve terminals. Nature, 1995;374(6518):186–90.
50. Shih, M., Lin, F., Scott, J.D., Wang, H.Y., and Malbon, C.C. Dynamic com-
21. Sweitzer, S.M. and Hinshaw, J.E. Dynamin undergoes a GTP‐dependent
plexes of β2‐adrenergic receptors with protein kinases and phosphatases and
conformational change causing vesiculation. Cell, 1998;93(6):1021–9.
the role of gravin. J Biol Chem, 1999;274(3):1588–95.
22. Massol, R.H., Boll, W., Griffin, A.M., and Kirchhausen, T. A burst of auxilin
51. Lin, F., Wang, H., and Malbon, C.C. Gravin‐mediated formation of signaling
recruitment determines the onset of clathrin‐coated vesicle uncoating. Proc
complexes in beta 2‐adrenergic receptor desensitization and resensitization.
Natl Acad Sci U S A, 2006;103(27):10265–70.
J Biol Chem, 2000;275(25):19025–34.
23. Lee, D.W., Wu, X., Eisenberg, E., and Greene, L.E. Recruitment dynamics of
52. Fan, G.F., Shumay, E., Wang, H.Y., and Malbon, C.C. The scaffold protein
GAK and auxilin to clathrin‐coated pits during endocytosis. J Cell Sci,
gravin (cAMP‐dependent protein kinase‐anchoring protein 250) binds the
2006;119(Pt 17):3502–12.
β2‐adrenergic receptor via the receptor cytoplasmic Arg‐329 to Leu‐413
24. Sever, S., Skoch, J., Newmyer, S. et al. Physical and functional connection
domain and provides a mobile scaffold during desensitization. J Biol Chem,
between auxilin and dynamin during endocytosis. EMBO J, 2006;25(18):
2001;276(26):24005–14.
4163–74.
53. Tao, J., Wang, H.Y., and Malbon, C.C. Protein kinase A regulates AKAP250
25. Mayor, S., Parton, R.G., and Donaldson, J.G. Clathrin‐independent path-
(gravin) scaffold binding to the beta2‐adrenergic receptor. EMBO J,
ways of endocytosis. Cold Spring Harb Perspect Biol, 2014;6(6):a016758.
2003;22(24):6419–29.
26. Tagawa, A., Mezzacasa, A., Hayer, A. et  al. Assembly and trafficking of
54. Mintz, L., Galperin, E., Pasmanik‐Chor, M. et al. EHD1 – an EH‐domain‐
caveolar domains in the cell: caveolae as stable, cargo‐triggered, vesicular
containing protein with a specific expression pattern. Genomics, 1999;59(1):
transporters. J Cell Biol, 2005;170(5):769–79.
66–76.
 6:  Endocytosis in Liver Function and Pathology 73
55. Pohl, U., Smith, J.S., Tachibana, I. et al. EHD2, EHD3, and EHD4 encode
novel members of a highly conserved family of EH domain‐containing
­proteins. Genomics, 2000;63(2):255–62.
56. Xie, S., Bahl, K., Reinecke, J.B. et al. The endocytic recycling compartment
maintains cargo segregation acquired upon exit from the sorting endosome.
Mol Biol Cell, 2016;27(1):108–26.
57. Naslavsky, N., Boehm, M., Backlund, P.S., Jr., and Caplan, S. Rabenosyn‐5
and EHD1 interact and sequentially regulate protein recycling to the plasma
membrane. Mol Biol Cell, 2004;15(5):2410–22.
58. Naslavsky, N., Rahajeng, J., Sharma, M., Jovic, M., and Caplan, S.
Interactions between EHD proteins and Rab11‐FIP2: a role for EHD3 in
early endosomal transport. Mol Biol Cell, 2006;17(1):163–77.
59. Sharma, M., Naslavsky, N., and Caplan, S. A role for EHD4 in the regulation
of early endosomal transport. Traffic, 2008;9(6):995–1018.
60. Zhang, J., Reiling, C., Reinecke, J.B. et  al. Rabankyrin‐5 interacts with
EHD1 and Vps26 to regulate endocytic trafficking and retromer function.
Traffic, 2012;13(5):745–57.
61. Duleh, S.N. and Welch, M.D. WASH and the Arp2/3 complex regulate
­endosome shape and trafficking. Cytoskeleton, 2010;67(3):193–206.
62. Giridharan, S.S., Cai, B., Vitale, N., Naslavsky, N., and Caplan, S. Cooperation
of MICAL‐L1, syndapin2, and phosphatidic acid in tubular recycling endo-
some biogenesis. Mol Biol Cell, 2013;24(11):1776–90, S1–15.
63. Zhang, J., Naslavsky, N., and Caplan, S. EHDs meet the retromer: complex
regulation of retrograde transport. Cell Logist, 2012;2(3):161–5.
64. Gokool, S., Tattersall, D., and Seaman, M.N. EHD1 interacts with retromer
to stabilize SNX1 tubules and facilitate endosome‐to‐Golgi retrieval. Traffic,
2007;8(12):1873–86.
65. Scott, C.C., Vacca, F., and Gruenberg, J. Endosome maturation, transport and
functions. Semin Cell Dev Biol, 2014;31:2–10.
66. Hessvik, N.P. and Llorente, A. Current knowledge on exosome biogenesis
and release. Cell Mol Life Sci, 2018;75(2):193–208.
67. Schmidt, O. and Teis, D. The ESCRT machinery. Curr Biol, 2012;22(4):
R116–20.
68. Carlson, L.A., Shen, Q.T., Pavlin, M.R., and Hurley, J.H. ESCRT filaments
as spiral springs. Dev Cell, 2015;35(4):397–8.
69. Frankel, E.B. and Audhya, A. ESCRT‐dependent cargo sorting at multive-
sicular endosomes. Semin Cell Dev Biol, 2018;74:4–10.
70. Shen, J., Huang, C.K., Yu, H. et al. The role of exosomes in hepatitis, liver
cirrhosis and hepatocellular carcinoma. J Cell Mol Med, 2017;21(5):
986–92.
71. Maji, S., Matsuda, A., Yan, I.K., Parasramka, M., and Patel, T. Extracellular
vesicles in liver diseases. Am J Physiol Gastrointestinal Liver Physiol,
2017;312(3):G194–G200.
72. Cornu M., de Caudron de Coquereaumont, G., and Hall, M.N. mTOR signal-
ing in liver disease, in Signaling Pathways in Liver Diseases 3rd edn (eds. J.‐F.
Dufour and P.‐A. Clavien), John Wiley & Sons, Chichester, 2015, pp. 314–25.
73. Settembre, C., De Cegli, R., Mansueto, G. et al. TFEB controls cellular lipid
metabolism through a starvation‐induced autoregulatory loop. Nat Cell Biol,
2013;15(6):647–58.
74. Napolitano, G. and Ballabio, A. TFEB at a glance. J Cell Sci,
2016;129(13):2475–81.
75. Gissen, P. and Arias, I.M. Structural and functional hepatocyte polarity and
liver disease. J Hepatol, 2015;63(4):1023–37.
76. Sewell, R.B. et al. Pharmacologic perturbation of rat liver lysosomes: effects
on release of lysosomal enzymes and of lipids into bile. Gastroenterology,
1988;95(4):1088–98.
77. Tooze, S.A., Abada, A., and Elazar, Z. Endocytosis and autophagy: exploita-
tion or cooperation? Cold Spring Harbor Perspect Biol, 2014;6(5):a018358.
78. Mizushima, N. Autophagy: process and function. Genes Dev, 2007;21(22):
2861–73.
79. Czaja, M.J., Ding, W.X., Donohue, T.M., Jr. et al. Functions of autophagy in
normal and diseased liver. Autophagy, 2013;9(8):1131–58.
80. Alers, S., Loffler, A.S., Wesselborg, S., and Stork, B. Role of AMPK‐mTOR‐
Ulk1/2 in the regulation of autophagy: cross talk, shortcuts, and feedbacks.
Mol Cell Biol, 2012;32(1):2–11.
81. Mercer, T.J., Gubas, A., and Tooze, S.A. A molecular perspective of mam-
malian autophagosome biogenesis. J Biol Chem, 2018;293(15):5386–95.
82. Lamb, C.A., Yoshimori, T., and Tooze, S.A. The autophagosome: origins
unknown, biogenesis complex. Nat Rev Mol Cell Biol, 2013;14(12):759–74.
83. Reggiori, F. and Klionsky, D.J. Autophagosomes: biogenesis from scratch?
Curr Opin Cell Biol, 2005;17(4):415–22.
84. Schulze, R.J., Weller, S.G., Schroeder, B. et al. Lipid droplet breakdown
requires dynamin 2 for vesiculation of autolysosomal tubules in hepato-
cytes. J Cell Biol, 2013;203(2):315–26.
85. Hyttinen, J.M., Niittykoski, M., Salminen, A., and Kaarniranta, K.
Maturation of autophagosomes and endosomes: a key role for Rab7.
Biochim Biophys Acta, 2013;1833(3):503–10.
86. Ravikumar, B., Moreau, K., Jahreiss, L., Puri, C., and Rubinsztein, D.C.
Plasma membrane contributes to the formation of pre‐autophagosomal
structures. Nat Cell Biol, 2010;12(8):747–57.
87. Szatmari, Z., Kis, V., Lippai, M., et al. Rab11 facilitates cross‐talk between
autophagy and endosomal pathway through regulation of Hook localiza-
tion. Mol Biol Cell, 2014;25(4):522–31.
88. Fader, C.M. and Colombo, M.I. Autophagy and multivesicular bodies: two
closely related partners. Cell Death Differ, 2009;16(1):70–8.
89. Madrigal‐Matute, J. and Cuervo, A.M. Regulation of liver metabolism by
autophagy. Gastroenterology, 2016;150(2):328–39.
90. Tasset, I. and Cuervo, A.M. Role of chaperone‐mediated autophagy in
metabolism. FEBS J, 2016;283(13):2403–13.
91. Schneider, J.L., Suh, Y., and Cuervo, A.M. Deficient chaperone‐mediated
autophagy in liver leads to metabolic dysregulation. Cell Metab,
2014;20(3):417–32.
92. Chao, X., Wang, H., Jaeschke, H., and Ding, W.X. Role and mechanisms
of  autophagy in acetaminophen‐induced liver injury. Liver Int,
2018;38(8):1363–74.
93. Ni, H.M., Bockus, A., Boggess, N., Jaeschke, H., and Ding, W.X. Activation
of autophagy protects against acetaminophen‐induced hepatotoxicity.
Hepatology, 2012;55(1):222–32.
94. Gomes, L.C., Di Benedetto, G., and Scorrano, L. During autophagy mito-
chondria elongate, are spared from degradation and sustain cell viability.
Nat Cell Biol, 2011;13(5):589–98.
95. Lazarou, M., Sliter, D.A., Kane, L.A. et al. The ubiquitin kinase PINK1
recruits autophagy receptors to induce mitophagy. Nature, 2015;
524(7565):309–14.
96. Harper, J.W., Ordureau, A., and Heo, J.M. Building and decoding ubiquitin
chains for mitophagy. Nat Rev Mol Cell Biol, 2018;19(2):93–108.
97. Hammerling, B.C., Najor, R.H., Cortez, M.Q. et  al. A Rab5 endosomal
pathway mediates Parkin‐dependent mitochondrial clearance. Nat
Commun, 2017;8:14050.
98. Yamano, K., Wang, C., Sarraf, S.A. et al. Endosomal Rab cycles regulate
Parkin‐mediated mitophagy. eLife, 2018;7.
99. Wong, Y.C., Ysselstein, D., and Krainc, D. Mitochondria‐lysosome con-
tacts regulate mitochondrial fission via RAB7 GTP hydrolysis. Nature,
2018;554(7692):382–6.
100. Brunt, E.M., Wong, V.W., Nobili, V. et al. Nonalcoholic fatty liver disease.
Nat Rev Dis Primers, 2015;1:15080.
101. Spengler, E.K. and Loomba, R. Recommendations for diagnosis, referral
for liver biopsy, and treatment of nonalcoholic fatty liver disease and non-
alcoholic steatohepatitis. Mayo Clin Proc, 2015;90(9):1233–46.
102. Thiam, A.R., Farese, R.V., Jr., and Walther, T.C. The biophysics and cell
biology of lipid droplets. Nat Rev Mol Cell Biol, 2013;14(12):775–86.
103. Walther, T.C. and Farese, R.V., Jr. Lipid droplets and cellular lipid metabo-
lism. Annu Rev Biochem, 2012;81:687–714.
104. Guo, Y., Cordes, K.R., Farese, R.V., Jr., and Walther, T.C. Lipid droplets at
a glance. J Cell Sci, 2009;122(Pt 6):749–52.
105. Schmidt, C., Ploier, B., Koch, B., and Daum, G. Analysis of yeast lipid
droplet proteome and lipidome. Meth Cell Biol, 2013;116:15–37.
106. Zhang, H., Wang, Y., Li, J. et al. Proteome of skeletal muscle lipid droplet
reveals association with mitochondria and apolipoprotein a‐I. J Proteome
Res, 2011;10(10):4757–68.
107. Sato, S., Fukasawa, M., Yamakawa, Y. et  al. Proteomic profiling of lipid
droplet proteins in hepatoma cell lines expressing hepatitis C virus core
protein. J Biochem, 2006;139(5):921–30.
108. Cermelli, S., Guo, Y., Gross, S.P., and Welte, M.A. The lipid‐droplet pro-
teome reveals that droplets are a protein‐storage depot. Curr Biol,
2006;16(18):1783–95.
109. Brasaemle, D.L., Dolios, G., Shapiro, L., and Wang, R. Proteomic analysis
of proteins associated with lipid droplets of basal and lipolytically stimu-
lated 3T3‐L1 adipocytes. J Biol Chem, 2004;279(45):46835–42.
110. Schott, M.B., Rasineni, K., Weller, S.G. et  al. β‐Adrenergic induction of
lipolysis in hepatocytes is inhibited by ethanol exposure. J Biol Chem,
2017;292(28):11815–28.
74 THE LIVER:  REFERENCES
111. Ong, K.T., Mashek, M.T., Bu, S.Y., and Mashek, D.G. Hepatic ATGL
knockdown uncouples glucose intolerance from liver TAG accumulation.
FASEB J, 2013;27(1):313–21.
112. Ong, K.T., Mashek, M.T., Bu, S.Y., Greenberg, A.S., and Mashek, D.G.
Adipose triglyceride lipase is a major hepatic lipase that regulates triacylglyc-
erol turnover and fatty acid signaling and partitioning. Hepatology,
2011;53(1):116–26.
113. Wu, J.W., Wang, S.P., Alvarez, F. et al. Deficiency of liver adipose triglyc-
eride lipase in mice causes progressive hepatic steatosis. Hepatology,
2011;54(1):122–32.
114. Reid, B.N., Ables, G.P., Otlivanchik, O.A. et al. Hepatic overexpression of
hormone‐sensitive lipase and adipose triglyceride lipase promotes fatty
acid oxidation, stimulates direct release of free fatty acids, and ameliorates
steatosis. J Biol Chem, 2008;283(19):13087–99.
115. Schulze, R.J., Drizyte, K., Casey, C.A., and McNiven, M.A. Hepatic
lipophagy: new insights into autophagic catabolism of lipid droplets in the
liver. Hepatol Commun, 2017;1(5):359–69.
116. Singh, R. and Cuervo, A.M. Lipophagy: connecting autophagy and lipid
metabolism. Int J Cell Biol, 2012;2012:282041.
117. Singh, R., Kaushik, S., Wang, Y. et al. Autophagy regulates lipid metabo-
lism. Nature, 2009;458(7242):1131–5.
118. Li, Z., Schulze, R.J., Weller, S.G. et  al. A novel Rab10‐EHBP1‐EHD2
­complex essential for the autophagic engulfment of lipid droplets. Sci Adv,
2016;2(12):e1601470.
119. Schroeder, B., Schulze, R.J., Weller, S.G. et al. The small GTPase Rab7 as a cen-
tral regulator of hepatocellular lipophagy. Hepatology, 2015;61(6):1896–907.
120. Schulze, R.J., Rasineni, K., Weller, S.G. et  al. Ethanol exposure inhibits
hepatocyte lipophagy by inactivating the small guanosine triphosphatase
Rab7. Hepatol Commun, 2017;1:140–52.
121. Kaushik, S. and Cuervo, A.M. Degradation of lipid droplet‐associated pro-
teins by chaperone‐mediated autophagy facilitates lipolysis. Nat Cell Biol,
2015;17(6):759–70.
122. Martinez‐Lopez, N., Garcia‐Macia, M., Sahu, S. et  al. Autophagy in the
CNS and periphery coordinate lipophagy and lipolysis in the brown a­ dipose
tissue and liver. Cell Metab, 2016;23(1):113–27.
123. Sathyanarayan, A., Mashek, M.T., and Mashek, D.G. ATGL promotes
autophagy/lipophagy via SIRT1 to control hepatic lipid droplet catabolism.
Cell Rep, 2017;19(1):1–9.
124. Khan, S.A., Sathyanarayan, A., Mashek, M.T. et  al. ATGL‐catalyzed
­lipolysis regulates SIRT1 to control PGC‐1alpha/PPAR‐alpha signaling.
Diabetes, 2015;64(2):418–26.
125. Inoue, J., Ninomiya, M., Shimosegawa, T., and McNiven, M.A. Cellular
membrane trafficking machineries used by the hepatitis viruses.
Hepatology, 2018;68(2):751–76.
126. Macovei, A. et al. Regulation of hepatitis B virus infection by Rab5, Rab7,
and the endolysosomal compartment. J Virol, 2013;87(11):6415–27.
127. Coller, K.E. et al. RNA interference and single particle tracking analysis of
hepatitis C virus endocytosis. PLoS Pathog, 2009;5(12):e1000702.
128. Selzer, L. and Zlotnick, A. Assembly and release of hepatitis B virus. Cold
Spring Harb Perspect Med, 2015 Nov 9;5(12):pii: a021394.
129. Miyanari, Y. et al. The lipid droplet is an important organelle for hepatitis C
virus production. Nat Cell Biol, 2007;9:1089–97
130. Salloum, S. et al. Rab18 binds to hepatitis C virus NS5A and promotes
interaction between sites of viral replication and lipid droplets. PLoS
Pathog, 2013;9(8):e1003513.
131. Ploen, D. et al. TIP47 plays a crucial role in the life cycle of hepatitis C
virus. J Hepatol, 2013;58(6):1081–8.
132. Coller, K.E. et al. Molecular determinants and dynamics of hepatitis C
virus secretion. PLoS Pathog, 2012;8(1):e1002466.
133. Inoue, J., Krueger, E.W., Chen, J. et al. HBV secretion is regulated through
the activation of endocytic and autophagic compartments mediated by
Rab7 stimulation. J Cell Sci, 2015;128(9):1696–706.
134. Ostrowski, M. et al. Rab27a and Rab27b control different steps of the exo-
some secretion pathway. Nat Cell Biol, 2010;12:19–30.
 The Hepatocellular
7 Secretory Pathway
Catherine L. Jackson1 and Mark A. McNiven2
1
Institut Jacques Monod, UMR7592 CNRS Université Paris‐Diderot, Sorbonne Paris Cité, Paris, France
2
Department of Biochemistry and Molecular Biology, and The Center for Digestive Diseases, Mayo Clinic &
Foundation, Rochester, MN, USA
INTRODUCTION
Hepatocytes are known to secrete scores of proteins and lipid
particles into the sinusoidal space, and also play a key role in
lipid homeostasis in organisms. An emerging theme over the past
few years has been the intimate connection between membrane
trafficking and lipid metabolism, with important implications for
hepatocyte function. The secretory pathway assures processing
and delivery of proteins to their correct destination, relying on
highly organized vesicular trafficking machinery that includes
numerous enzymes, cytoskeletal proteins, molecular motors, and
coat proteins. This machinery is highly conserved in evolution,
and is found in virtually all eukaryotic cells. Specialized cell
types such as hepatocytes can have more specific secretion
­pathways. Very low density lipoprotein (VLDL) particles are
produced uniquely by liver cells, and use the ubiquitous vesicu­
lar trafficking machinery in addition to components specific to
this large cargo. Through conventional biochemical and molecu­
lar methods, the use of genetic model organisms, and recent
advancements in live cell imaging, much has been learned about
these trafficking pathways in hepatocytes and other epithelial
cells. This chapter will focus on the molecular mechanisms by
which nascent proteins and larger cargos are sequestered, pack­
aged into vesicle carriers, and targeted to specific hepatocellular
destinations during the secretory process.
THE SECRETORY PATHWAY
The secretory pathway begins with synthesis of proteins at the
endoplasmic reticulum (ER) [1]. Next, proteins are transported
to the Golgi apparatus (Figure 7.1), the central processing and
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
sorting station of the pathway [2]. Generally distributed
­throughout the cell and extending to the periphery, the ER is
characterized by a network of interconnected membrane struc­
tures that include cisternae and tubular networks [3, 4]. The
Golgi apparatus, in contrast, is localized in the perinuclear
region of the cell and forms a continuous ribbon in mammalian
cells [5]. This ribbon is composed of organized stacks of flat­
tened saccules, interconnected by more complex tubulo‐vesicu­
lar regions [5]. The first cis‐Golgi element and the trans‐Golgi
network (TGN) are tubular membrane meshworks at the entry
and exit sides of the Golgi apparatus, respectively.
Early studies of the secretory pathway made use of rat liver
(Figure 7.1), as well as exocrine pancreas cells, due to their high
levels of secretion [6]. These studies determined that during the
synthesis and transport of secreted proteins, the ER and Golgi
function sequentially, and prior to packaging of secretory cargo
into secretory granules or vesicles.
The mechanism by which proteins are transported from a
donor to an acceptor compartment in eukaryotic cells is via the
budding and fusion of vesicular carriers [7] (Figure  7.2). The
first step in formation of a vesicle is activation of a small G pro­
tein of the ADP ribosylation factor (Arf) family by a nucleotide
exchange factor, which leads to recruitment of proteins called
effectors to the membrane [10–12]. Cargo adaptors, which are
coupled to a vesicle coat, are a major class of effectors of the
Sar1 and Arf1 small G proteins that function at the ER and Golgi,
respectively. Both Arf and Sar proteins have an N‐terminal
amphipathic helix that inserts into the membrane, contributing to
deformation into a highly curved bud [13]. A variety of different
adaptors bind to the cargo, either transmembrane or luminal pro­
teins, to concentrate them into the forming vesicle [14]. Other
effectors of Sar1 and Arf1 include lipid‐modifying enzymes,
which change the composition of the membrane at the budding
76 THE LIVER:  EVOLUTIONARY ORIGINS OF VESICLE COATS
Figure 7.1  Elaborate vesicular processes in the hepatocyte. Thin‐­
section electron micrograph of a rat liver hepatocyte from the original
collections of, Dr. Keith Porter, about 1962. Deep in the cytoplasm,
parallel cisternae of the endoplasmic reticulum (ER) appear to give rise
to the flattened, stacked saccules of the Golgi apparatus (G). Scores of
small vesicular profiles can be seen in the Golgi region, as well as larger
lipoprotein‐filled secretory vesicles budding from the trans side of the
Golgi. Specific coats, accessory proteins, and address tags characterize
many of these vesicles. G, Golgi; ER, endoplasmic reticulum; Lys,
lysosome.
site, also contributing to membrane deformation [15–17].
Polymerization of the outer shell of the coat also aids in shaping
the membrane into a highly curved bud.
Once the bud has formed and undergone scission from the
donor compartment membrane, uncoating of the vesicle ensues.
The first step is hydrolysis of GTP on the Arf family small G
protein, which can commence even before the vesicle has under­
gone fission from the donor compartment [18]. The interaction
between vesicle cargo and the coat maintains the coat on the
vesicle membrane until the appropriate time for uncoating,
which must occur prior to fusion of the vesicle with its target
membrane.
Transport in the anterograde direction from the ER to the
Golgi apparatus is mediated by coat protomer II (COPII)‐coated
vesicles (Figure  7.2). On the ER membrane, subunits of the
COPII coat are recruited upon activation of Sar1 by its exchange
factor Sec12, to form buds and ultimately generate vesicles.
COPII vesicle formation occurs at specialized regions of the ER
that are called ER exit sites (ERES). In mammalian cells, COPII
vesicles undergo homotypic fusion, and the resulting compart­
ment recruits exchange factor Golgi brefeldin A resistant factor 1
(GBF1), its substrate Arf1, and the coat protomer I (COPI) coat,
which form COPI vesicles (Figure  7.2). The sorting compart­
ment formed upon GBF1–Arf1–COPI recruitment, called the
ER/Golgi intermediate compartment (ERGIC), is made up of a
complex network of tubules [19]. COPI‐coated vesicles are also
thought to mediate the retrograde transport of resident ER pro­
teins and Golgi enzymes from the Golgi apparatus back to the
ERGIC and to the ER [20, 21]. In addition to secretory proteins,
membrane proteins and lysosomal enzymes pass through the
secretory apparatus to reach their final destinations, which
reflects the important sorting role of the Golgi apparatus.
A particularly active sorting region of the Golgi is the TGN,
where several distinct types of vesicles containing specific cargo
proteins and harboring particular coat proteins arise [22–24].
Several of these coats contain cargo adaptor proteins (APs)
recruited to membranes by the Arf1 or Arf3 small G proteins.
The APs are four‐subunit complexes containing two large and
two small subunits. AP‐1, and probably AP‐3 as well, recruit
clathrin upon membrane binding to form a clathrin–adaptor
coat. However, this is not the case for AP‐4, which forms a non‐
clathrin coat [22, 23]. Another important set of adaptors at the
TGN are the Golgi‐localized γ‐ear‐containing, ADP ribosyla­
tion factor‐binding (GGA) proteins (GGA1, GGA2, and
GGA3), which bind clathrin and transport distinct cargo from
the TGN to endosomes. Secretory vesicles (or granules) carry­
ing constitutively secreted proteins from the TGN to the cell
surface are believed to be uncoated vesicles, as no coats have yet
been identified in mammalian cells [23].
EVOLUTIONARY ORIGINS
OF VESICLE COATS
The COPII, COPI, and AP–clathrin coats are all thought to
share a common origin, arising from a primordial coat com­
plex. All three vesicle coats have subunits containing a β‐pro­
peller domain followed by an α‐solenoid, also known as a
helix‐turn‐helix domain [8] (Figure  7.2). Strikingly, only
eukaryotes, and a few rare prokaryotes with rudimentary vesic­
ular trafficking pathways, have proteins containing a β‐propel­
ler motif fused to an α‐solenoid fold [8]. Hence this class of
protein is strictly correlated in evolution with the necessity to
bend membranes, such as in vesicle formation. This β‐
propeller–α‐solenoid structure is also found in nuclear pore
components that bind to the highly curved membrane of the
pore [25]. These observations led to the protocoatomer hypoth­
esis, which postulates that COPII, COPI, and clathrin–AP
coats, as well as some nucleoporins, all shared a primordial
ancestor [25]. Additional support for this hypothesis comes
from clear sequence homology between the β‐ and γ‐COP sub­
units of COPI and the large subunits of the AP‐1 and AP‐2
adaptors [8, 26, 27]. However, despite this homology, the
structures of the COPI and AP–clathrin coats are significantly
different (Figure 7.2). The β‐propeller–α‐solenoid subunits of
both coats polymerize due to interactions between α‐solenoids,
but in the case of COPI, these subunits make direct contact
with the membrane, and are positioned beside the other sub­
complex of COPI. In contrast, in the AP–clathrin coat, clathrin
forms a cage around the vesicle that does not directly touch the
membrane (Figure 7.2a). COPII resembles AP–clathrin in that
the β‐propeller–α‐solenoid subunit (Sec31), along with Sec13,
forms an outer cage that does not directly contact the mem­
brane (Figure 7.2a). However, the geometry of the COPII and
AP–clathrin coats is quite different based on in vitro structural
data (Figure  7.2a), and visualization of vesicles within cells
(Figure 7.2b).
 7:  The Hepatocellular Secretory Pathway 77

(a) β-propeller
α-solenoid
small GTPase
Sec23
Sec24
AP/COP large subunit (β)
AP/COP large subunit (EGADZ) Clathrin
AP/COP medium subunit
AP/COP small subunit
Transmembrane cargo

Sec13/Sec31 α-COP β’-COP

COPII COPI AP-1 + clathrin

(b)
COPII COPI Clathrin
A B C

D E F

Figure 7.2  Comparison of the COPII, COPI, and AP–clathrin coats. (a) All three coats contain β‐propeller–α‐solenoid components (red spheres
and blue rods), but they are arranged in different ways in each coat. In COPII and AP–clathrin coats, these subunits form an outer polyhedral layer.
In COPI, on the other hand, the β‐propeller–α‐solenoid‐containing subunits directly contact the membrane as well as the other COPI subunits. The
other subunits of COPI and the adaptor subunits share both sequence and structural homology. All three coats are recruited to membranes by a
member of the Arf small G protein (GTPase) family: Sar1 in the case of COPII, and Arf1 for both COPI and AP–clathrin coats. Reproduced from
[8] with permission of Elsevier. (b) Visualization in their native cellular environment of COPII (A,D), COPI (B,E) and AP‐1–clathrin (C,F)‐coated
vesicles by cryoelectron tomography. Imaging was performed on Chlamydomonas reinhardtii cells. Panels A–C show cross‐sections through each
vesicle; panels D–F are grazing slices through the coats at the tops of the same vesicles. Note that the geometry of the COPII and clathrin coats is
visible in the lower panels, revealing the triangular Sec13/31 COPII lattice (D) and clathrin triskelions (F). Scale bars, 50 nm. Reproduced from
Bykov 2017 [9], https://cdn.elifesciences.org/articles/32493/elife‐32493‐v2.pdf. Licensed under CCBY 4.0.
78 THE LIVER:  COPI‐COATED VESICLES
In addition to β‐propeller–α‐solenoid subunits, coats also
share similar features such as recruitment to membranes by a
member of the Arf family of small G proteins, as described
above. Other ancestral motifs found in all coat complexes are
longin domains (present in the medium and small COPI/AP
subunits, Figure  7.2a) and coiled coils [8]. This ensemble of
ancient conserved features supports the idea of an ancient pri­
mordial coat, which split into the three major coats of the secre­
tory pathway: COPII, COPI, and AP–clathrin.
COPI‐COATED VESICLES
COPI‐coated vesicles are required for intra‐Golgi trafficking
and for recycling from the Golgi apparatus back to the ER.
The COPI coat consists of seven subunits (α, β, β′, γ, δ, ε, ζ
COP) that form a 680 kDa complex and associate with the small
GTP‐binding protein Arf1. COPI sorts cargo proteins into form­
ing vesicles as it induces membrane curvature [28] (Figure 7.2).
COPI‐coated vesicles are best known for their function in the
retrieval of ER‐resident proteins from the Golgi. Selected COPI
subunits are known to associate with dilysine amino acid motifs
in the cytoplasmic domains of ER‐resident enzymes [20].
Furthermore, several COPI temperature‐sensitive mutants in yeast
have revealed defects in transport between the ER and the Golgi at
the restrictive temperature [21]. Further, COPI‐coated vesicles
have been shown to mediate retrograde transport of resident ER
proteins by using the KDEL (single‐letter code for amino acids)
receptor [29], confirming a retrieval role for COPI coat proteins.
Regulators of Arf–GTP binding and GTP hydrolysis control the
spatio‐temporal localization of Arf protein activation and down­
stream signaling. GDP release from Arf proteins is catalyzed by
Arf guanine nucleotide exchange factors (GEFs), allowing the
more abundant GTP to bind. This nucleotide exchange activity is
contained within the Sec7 domain, an evolutionarily conserved
sequence first identified in yeast Sec7p [30–32]. A Sec7 domain is
present in all Arf GEFs identified to date. The function of Sec7
domains was first identified in the yeast Gea1p protein [33]. The
human ortholog of yeast Gea1p, GBF1, was identified by P.
Melançon and colleagues [34]. GBF1 is a Golgi‐localized protein
whose overexpression was found to confer resistance to the fungal
toxin brefeldin A, which completely blocks secretion, and causes
disassembly of the Golgi apparatus and eventually its fusion with
the ER [10]. The Arf GTPase‐activating proteins (GAPs) catalyze
the hydrolysis of GTP on Arf family proteins, a function carried
out by a conserved GAP domain, which contains a zinc finger [10].
The primary sequence homology of the catalytic domains of the
Arf GEFs and GAPs accelerated their identification and allowed
studies of their evolution. Phylogenetic analyses of the Arf GAPs
has revealed that they have likely co‐evolved with their Arf sub­
strates [35], and a similar conclusion very likely holds for the Arf
GEFs [36]. From these results we can conclude that Arf proteins
function in a tightly coordinated manner with their regulators.
There are seven subfamilies of Arf GEFs in eukaryotic cells
[30]. Members of the GBF/Gea and BIG/Sec7 subfamilies of
Sec7 domain GEFs use class I Arf proteins as substrates [10],
and in addition, GBF1 possibly functions on class II Arf pro­
teins [34, 37]. In humans, there is one GBF/Gea member, GBF1,
and two members of the BIG/Sec7 subfamily, brefeldin A inhib­
ited GEF 1 (BIG1) and BIG2. These two subfamilies have a
similar domain structure, probably because they have descended
from a common ancestor, but have different steady‐state locali­
zations and functions [10, 30, 31]. In animal cells, including
hepatocytes, and yeast, the major steady‐state localization
of  GBF/Gea GEFs is at the early Golgi, whereas BIG/Sec7
GEFs are present predominantly at the late Golgi, ­including
the TGN [10, 31].
COPI vesicles mediate retrograde transport of components
that need to be recycled from the Golgi and ERGIC back to
earlier compartments and to the ER. Arf proteins (primarily
Arf1 and Arf3) at the TGN recruit the heterotetrameric clathrin
adaptor proteins, AP‐1, AP‐3, and AP‐4, and the three mono­
meric Golgi‐localized γ‐ear‐containing, ADP ribosylation fac­
tor‐binding proteins (GGAs 1–3) [24, 38]. To explain how one
or two Arf proteins can recruit so many different coats to differ­
ent membrane sites, several groups have found evidence that the
GEFs activating Arfs at different sites determine the coat that
will be recruited. GBF1 at ERGIC and the cis‐Golgi is responsi­
ble for COPI recruitment, and the BIG1 and BIG2 proteins at
the trans‐Golgi are responsible for recruiting the AP–clathrin
coats [10]. Demonstration of a direct interaction between GBF1
and COPI coat subunits provides a molecular explanation for
this specificity [39]. Interaction of GBF1 with COPI prior to
Arf1 activation ensures that COPI is stabilized on the membrane
where GBF1 activates Arf1 [39].
Removal of the coat from a vesicle (uncoating) is required for
fusion of the vesicle to the acceptor membrane. The first step in
vesicle uncoating occurs upon hydrolysis of GTP on the Arf
family G protein that maintains the coat on the membrane [40,
41]. Even after hydrolysis of GTP on Sar1 or Arf1, the coat (or
at least a portion of the coat) remains on the vesicle due to inter­
actions with other vesicle proteins such as cargo molecules [40,
42]. In the case of COPI vesicles, GTP hydrolysis on Arf1 is
mediated by Arf GAP proteins [28, 43, 44]. Arf GAP1 is
recruited only to highly curved membranes through its ALPS
motif, which ensures that the coat remains on the budding vesi­
cle until it has achieved a mature, highly curved morphology
[45, 46]. It was originally thought that the coat was lost quite
quickly after budding. However, interaction of the coat with a
tethering complex on the acceptor membrane has recently been
demonstrated for multiple types of vesicles [47, 48]. For COPI
vesicles transported to the ER, the Ds11 complex present on the
ER membrane acts as a vesicle docking site through direct inter­
action with the COPI coat [49–51]. Moreover, this coat–tether
interaction stimulates uncoating of the vesicle [51]. In addition
to fusing with the ER, COPI vesicles also mediate intra‐Golgi
retrograde trafficking. Interactions between COPI subunits and
the Golgi COG tethering complex [52, 53], and with the cis‐
Golgi tether p115 [54], are also likely to be involved in targeting
vesicles to these Golgi acceptor membranes.
GBF1 has been shown to have a crucial function in the
­replication of numerous viruses, including hepatitis C [55] and
hepatitis E [56]. Hepatitis A, B, C, D, and E viruses cause the
majority of liver disease across the globe, and represent a world­
wide health problem. For the well‐studied hepatitis C virus, lipid
metabolism is an important component in infection, and lipid
metabolism pathways are subverted for propagation of this virus
 7:  The Hepatocellular Secretory Pathway 79
[57]. The catalytic domain of GBF1 is required for HCV to form
functional viral replication complexes [58]. Arf4 and Arf5 are
also required for HCV replication, through their roles in lipid
homeostasis in hepatocytes [58]. Therefore, understanding the
mechanisms by which hepatitis viruses infect hepatocytes could
increase our knowledge of how the secretory pathway and lipid
homeostasis are coordinated in cells, in addition to providing
avenues for treatment of these important human pathogens.
COPII‐COATED VESICLES
COPII‐coated vesicles function in ER‐to‐Golgi
anterograde transport
Protein transport from the ER to the Golgi complex is mediated
by vesicular carriers that originate from morphologically
defined, specialized ribosome‐free regions of the ER called ER
exit sites (or ER export sites) (ERES) [59]. ER buds and an
abundance of small vesicular profiles at these sites, which are
distinct from smooth ER could be clearly visualized in the clas­
sic electron micrographs of pancreatic acinar cells by Jamieson
and Palade [60]. Studies using live cell imaging demonstrated
that these transitional areas are formed transiently and randomly
from the ER [61].
The cargo carriers that mediate ER‐to‐Golgi protein transport
have been subjects of intense study. As for many of the protein
trafficking steps along the secretory pathway, molecular compo­
nents were first identified in studies using the yeast Saccharomyces
cerevisiae as a model system. Yeast has provided a powerful
combination of genetics, molecular biology, and biochemistry to
elucidate mechanisms both in vitro and in vivo. Isolation of con­
ditional sec (secretion) mutants allowed the identification and
characterization of numerous components of trafficking path­
ways, including all essential subunits of the COPII coat [32]. By
using yeast membranes for ER vesicle budding in vitro, cell‐free
assays for vesicle budding and fusion could be reconstituted.
These systems enabled the isolation of ER‐derived transport ves­
icles that were 60–65 nm in diameter and exhibited a distinct
electron‐dense coat on their surface, termed COPII [62]. COPII
coat components were identified from the purified vesicles and
revealed to be distinct from those of COPI coats [62].
Early studies in various cell types including human hepatoma
HepG2 cells [63] reported that secretory cargo protein was present
at bud sites upon ER exit and within small 40–80 nm carrier vesicles
and tubules referred to as vesicular–tubular clusters (VTCs). In pan­
creatic cells, similar structures were shown to be COPII‐­positive.
The ER cisternae near VTCs often exhibited budding profiles that
were decorated with an electron‐dense, honeycomb‐patterned coat
(Figure 7.2b), which was identified as COPII by immunogold labe­
ling antibodies against COPII coat components [64].
COPII‐coated vesicle protein components
and vesicle formation
The COPII coat components, which are distinct from the
COPI coatomer, were postulated to induce vesicle budding of
ER‐derived transport vesicles because only the GTPase Sar1,
Sec13/31, and Sec23/24 were found to be necessary for the
generation of COPII‐coated vesicles from washed micro­
somes [62]. Sec12, a GTP‐GEF for Sar1 that is localized to
the ER initiates COPII vesicle biogenesis. After the activa­
tion of Sar1, the adaptor coat components Sec23/24 are
recruited to bind cargo. Subsequently, the structural β‐
propeller–α‐solenoid Sec13/31 coats are recruited to form
the coat and to contribute to membrane deformation. The
selection of cargo including transmembrane proteins and v‐
SNAREs (soluble NSF attachment receptors) by Sec23/24 is
mediated by three different sites for recognizing ER‐to‐Golgi
v‐SNAREs on Sec24. The interaction is regulated by the
assembly state of the SNAREs, which suggests that COPII
proteins can be involved in vesicle fusion specificity in the
ER‐to‐Golgi step [12].
In the final stages of COPII vesicle biogenesis, the GTP
used during vesicle formation is hydrolyzed by Sec23, a Sar1‐
specific GAP. Upon GTP binding by Sar1, membrane inser­
tion of the N‐terminal α helix of Sar1 deforms synthetic
liposomes into narrow tubules. Mutation of the helix led to a
defect in membrane curvature and the formation of vesicles
from native ER, although the recruitment of coat proteins
appeared to be normal [13]. Moreover, inhibition of GTP
hydrolysis by Sar1 resulted in COPII‐coated vesicles that
failed to detach from the ER, suggesting that regulation of the
N‐terminus of Sar1 by GTP binding and hydrolysis controls
COPII vesicle fission [65].
Although the minimal requirements for in vitro vesicle bio­
genesis are the five molecules listed above, other components
essential for COPII vesicle formation in different cell types have
been identified. Sec16 was first identified in yeast, and found to
be essential and required for COPII vesicle formation in vivo
[12]. Sec16 is present throughout eukaryotes, and is thought to
play a role in scaffolding other COPII components and organ­
izing ERES through interaction with other ERES components
[14]. Sec16 is also required for formation of stress granules con­
taining COPII components that are formed when cells are
starved and consequently shut down the secretory pathway [66].
Hence Sec16 has broad functions in regulation of COPII com­
ponents in various cellular physiological states, including cell
growth and starvation.
The TRK‐fused gene protein (TFK) and the cargo receptor
transport and Golgi organization 1 (TANGO1)/cutaneous T
cell lymphoma‐associated antigen (cTAGE5) proteins are
more recently identified ERES components that function in
ERES organization [67, 68]. TFK binds to the Sec23/24 inner
coat and contributes to uncoating of COPII vesicles [67]. As
described for COPI vesicles, the uncoating of COPII vesicles
also occurs late in a transport step, after targeting of the vesi­
cle to the a­ cceptor compartment [47, 48]. Interestingly, TFK
has an N‐­ terminally unstructured region in addition to a
COPII‐binding site, which confers properties of liquid droplet
phase separation in vitro. In the absence of TFK, COPII vesi­
cles disperse, suggesting that TFK might function to maintain
vesicles close to their target membranes though forming a liq­
uid droplet ­structure between ERES and ERGIC target mem­
branes [67]. In addition to TRK, the COPII vesicle coat
component Sec23 also interacts with the tethering complex
transport protein particle (TRAPP) at the acceptor
80 THE LIVER:  TRANS‐GOLGI NETWORK‐DERIVED VESICLES
compartment membrane prior to complete release of the
COPII coat and fusion [69].
TANGO1 was identified in a screen for genes affecting secre­
tion in flies, and was later demonstrated to play an essential role
in collagen secretion in keratinocytes and fibroblasts [68]. Like
TFK, TANGO1 interacts with the inner components of the
COPII coat, Sec23 and Sec24, but not the outer cage compo­
nents, and localizes to ERES [68]. TANGO1 is also expressed in
liver cells [70], where it functions in export of specific cargos,
as described below.
In mammalian cells, GST–hSec23‐binding columns allowed
the identification of a 125 kDa protein, p125, that is expressed
in liver, and shares sequence homology with p­ hosphatidic acid
(PA)‐preferring phospholipase A1, which produces lysophos­
pholipids [71]. Recently, the function of lysophospholipids in
lowering the rigidity of the ER membrane during COPII
­vesicle budding was demonstrated both in vitro and in cells
[72]. These results indicate the importance of membrane
­com­position and membrane biophysical properties in vesicle
formation.
LARGE COPII CARRIERS TRANSPORT
VLDL PARTICLES FROM THE ER
TO THE GOLGI
Recently, it has become clear that large cargos, too large to be
incorporated into classic 60 nm COPII vesicles, nevertheless
use COPII for their transport out of the ER to the Golgi in the
secretory pathway [73, 74]. In hepatocytes, VLDL particles are
formed from the ER membrane on the luminal side, and from
there are transported to the Golgi apparatus [70, 75]. However,
each VLDL particle is up to 90 nm in diameter, and especially
if multiple particles enter the same carrier, this cargo is too
large to be incorporated into classic COPII vesicles. However,
there is evidence for COPII involvement in ER–Golgi transport
of VLDL [76, 77]. In addition, the TANGO proteins are impor­
tant components in this process, acting as scaffolds for the
assembly of these larger COPII vesicular structures. TANGO1
was first identified as a transmembrane cargo receptor for
luminal procollagen, another cargo too large for classic COPII
vesicles. As a cargo receptor, TANGO1 was found to link lumi­
nal collagen to the COPII coat on the other side of the ER
membrane [68].
TANGO1 and an interacting partner called TALI are both
required for optimal secretion of VLDL particles in the hepato­
cyte cell line HepG2 [70]. TANGO1 and TALI not only interact,
but the complex interacts with apolipoprotein B (ApoB100),
required for production of VLDL particles via its interaction
with triglycerides [70]. All three proteins colocalize in HepG2
cells, and interaction of TANGO/TALI with apoB present on
VLDL particles results in their recruitment to ER exit sites,
where they are packaged into COPII carriers [70]. Interestingly,
ERGIC53, a marker of ERGIC membranes, colocalizes with
apoB‐containing structures at ER exit sites [70]. This and other
data support the conclusion that ERGIC is involved in forma­
tion of VLDL particles, as it is for collagen‐containing COPII
carriers [78].
TRANS‐GOLGI NETWORK‐DERIVED
VESICLES
Protein sorting and trafficking events
at the trans‐Golgi network
The TGN is a tubulovesicular network that acts primarily in
sorting proteins to their final destinations. Within this membra­
nous reticulum, cargo proteins are segregated efficiently into
distinct vesicles that will be targeted to various compartments,
including the endosome/lysosome system, the apical or basolat­
eral domains in polarized cells, and the secretory granule pool.
In regulated secretory cells, this granule pool awaits the appro­
priate extracellular stimulus to exocytose its contents.
The TGN is known to interact intimately and dynamically
with the endosomal apparatus, especially during events related
to receptor recycling, endosomal maturation, and the delivery of
lysosomal enzymes [23].
A large number of different types of vesicles are formed at
the trans‐Golgi and TGN, reflecting the intense sorting activity
that occurs at this exit face of the Golgi. The best characterized
are clathrin‐coated vesicles, which are linked to the membrane
through various APs, as described briefly above. Other non‐
clathrin‐coated vesicles are characterized by adaptor proteins
homologous to AP‐1, but which carry different cargo.
Formation of clathrin‐coated and clathrin‐
independent vesicles at the TGN
Clathrin was the first coat to be visualized in the Golgi region of
mammalian cells. An electron‐dense, spiked coat on TGN mem­
branes in liver hepatocytes could be seen clearly in the original
micrographs of Novikoff and Yam [80] (Figure 7.3a). The first‐
identified and best‐characterized cargo adaptors are heterotrim­
eric adaptor protein complexes AP‐1 to AP‐5, three of which
mediate sorting at the TGN (AP‐1, AP‐3, and AP‐4). All three of
these coats are recruited to membranes by Arf1 and/or Arf3.
AP‐1 in addition requires the phosphoinositide phosphatidylin­
ositol 4‐phosphate (PI(4)P) (Figure  7.4). TGN‐derived CCVs
are now known to have a coat composed of clathrin triskelia and
the AP complexes AP‐1 (made up of γ‐, β1‐, μ1‐, and σ1‐adaptin
subunits) [22] and AP‐3 (δ‐, β3‐, μ3‐, and σ3‐adaptin subunits)
[81]. AP‐4 does not bind to clathrin. The subunit μ1 of AP‐1 is
responsible for binding to YXXΦ motifs within cargo proteins,
whereas β1 interacts with dileucine‐based signals in cargo. The
γ‐adaptin subunit recruits accessory proteins to the site of vesi­
cle formation [23].
There are two forms of the μ1 subunit of AP‐1, μ1A and μ1B.
μ1A is ubiquitous, whereas μ1B is found only in polarized epi­
thelial cells [22, 23]. The two adaptor complexes AP‐1A and
AP‐1B, containing μ1A and μ1B, respectively, are both present
in polarized cells, and function in basolateral protein sorting.
AP‐1B is responsible for sorting of a specific subset of cargo
proteins at the TGN in epithelial cells to the basolateral plasma
membrane, including the LDL receptor and interleukin 6 recep­
tor β chain [82].
A second type of coat complex recruited to TGN membranes
by Arf1 and PI(4)P is the family of monomeric adaptors, the
 7:  The Hepatocellular Secretory Pathway 81

(a) (b)
VLDL transport vesicles
ER Golgi
VLDL apparatus
VLDL

Sar1b
Sec23/24
VLDL Sec13/31 VLDL
TANGO
VLDL

ApoB100 VLDL

Sar1a
Sec23/24
Sec13/31

Classic COPII vesicles

Figure 7.3  (a) VLDL particles are present in dilated elements (arrows) within the Golgi apparatus, and in secretory vesicles forming within the
trans‐Golgi and TGN. Rough ER is evident by the presence of ribosomes. G, Golgi; arrows, Golgi elements containing VLDL particles; tGE,
trans‐Golgi elements containing VLDL particles; C, clathrin‐coated vesicle; R, ribosomes; P, peroxisome. Magnification: 44,000×. Reproduced
with permission from [80]. ©1978 P.M. Novikoff and A. Yam. Originally published in Journal of Cell Biology. https://doi.org/10.1083/jcb.76.1.1
(b) Schematic diagram depicting VLDL trafficking from the ER to the Golgi in hepatocytes. VLDL particles form from the ER membrane on the
luminal side, and require apolipoprotein B 100 (apoB100). The particles bud into the ER lumen, then are packaged into VLDL transport vesicles
using the COPII machinery (Sec23/24, Sec13/31) as well as TANGO. The diameter of VLDL transport vesicles is approximately 110 nm [79], suf­
ficient to enclose VLDL‐sized particles, and larger than classic COPII vesicles (60–70 nm diameter on average). The latter contain nascent secretory
and transmembrane cargo proteins synthesized on ER‐localized ribosomes, and translocated into the ER lumen. Both VLDL transport vesicles and
classic COPII vesicles require COPII proteins for their budding from the ER membrane, but use different Sar1 small G proteins: Sar1b for VLDL
transport vesicles and Sar1a for classic COPII vesicles. Upon fission of the vesicle from the ER, both types of vesicles are targeted to the acceptor
compartment, uncoat, then fuse with the acceptor compartment membrane.

Basolateral membrane
6
11

Rab8
ab
ab

R
R

Golgi
4)P
PI( P2
)
,5
(4
PI b2
Ra BC
Rab9
LE/MVB 1
Rab
membrane
Apical

ER

Figure 7.4  Specific small Rab GTPases and phosphoinositides are localized to distinct membrane compartments along the hepatocyte secretory
pathway. Rab1 and Rab2 are believed to mediate the traffic of nascent proteins from the ER to the cis‐Golgi. Rab6 is a kinesin‐associated Rab
that mediates transport within the Golgi stacks, Rab11 regulates exit from the TGN, and Rab8 functions in the targeting of secretory vesicles from
the Golgi to the plasma membrane. Rab9 has a role in late endosome‐to‐TGN transport. PI(4)P is significantly enriched in the Golgi and is
believed to mediate the recruitment of multiple proteins that support vesicle formation at the TGN. PI(4,5)P2 at the Golgi functions with Arf1 and
phospholipase D (PLD) and is also involved in vesicle formation. LE/MVB, late endosome/multivesicular body; ER, endoplasmic reticulum; BC,
bile canalicular membrane.
82 THE LIVER:  COORDINATION BETWEEN THE SECRETORY PATHWAY AND LIPID METABOLISM
Golgi‐localized, γ‐ear‐containing, Arf‐binding proteins (GGAs).
There are three known GGA proteins in mammalian cells:
GGA1, GGA2, and GGA3 [23]. Each GGA protein contains
three domains: VHS (Vps27, Hrs, Stam), GAT (GGA and
TOM), and γ‐adaptin ear (GAE). Through the VHS domain, the
GGAs recognize DXXLL motifs in specific cargo such as the
mannose 6‐phosphate receptor (MPR) tail and thus provide a
sorting function between the TGN and endosomes [83]. The
GAT domain interacts with Arf; mutations in this domain abro­
gated recruitment of GGAs to the TGN as well as binding to Arf
[84]. The GAE domain is homologous to the GAE domain of
the AP‐1 complex‐recruiting accessory proteins [85].
CCVs generated from the TGN are involved in the transport
of newly synthesized lysosomal enzymes. Newly synthesized,
soluble lysosomal enzymes acquire mannose 6‐phosphate resi­
dues at the cis‐Golgi, and these residues are recognized by the
MPRs at the TGN [23]. The GGAs are essential for packaging
MPRs into CCVs and for transporting them from the TGN to
endosomes [86]. Another adaptor complex at the TGN, AP‐3, is
expressed ubiquitously, is localized to buds/vesicles associated
with the TGN, and can interact with clathrin and with sorting
signals in the cytoplasmic tails of lysosomal membrane proteins
[23]. Correct targeting of lysosomal‐associated membrane pro­
tein‐1 (LAMP‐1) and lysosomal integral membrane protein‐2
(LIMP‐2) are mediated by the AP‐3 adaptor complex [81]. In
addition to the heterotetrameric and GGA adaptor complexes,
additional adaptors have been identified, including epsin‐related
proteins [23].
In addition to Arf family small G proteins, another family of
small G proteins, called Rabs, are involved in each trafficking
step of the secretory pathway [87]. Rab1, along with Rab2, are
known to function in trafficking from the ER and ERGIC com­
partments [59, 87]. Other Rab GTPases, including Rab6,
Rab19, Rab30, Rab33, Rab36, Rab40, Rab41, and Rab43 func­
tion at the Golgi, and Rab8 (the ortholog of Sec4 in yeast) is
crucial in the exocytic pathway for late Golgi to plasma mem­
brane transport. Rab3, Rab26, Ra27, Rab37, and Rab2 have
also been implicated in TGN to plasma membrane trafficking
[87]. Rab8 and Rab11, along with Arf4 and Arf regulators,
function in trafficking from the TGN to cilia in photoreceptor
cells [88] (Figure 7.4).
Scission of TGN‐derived vesicles
Since the localization of the large GTPase dynamin to the Golgi
apparatus, in addition to its known cell surface distribution, was
reported, it has been suggested that vesicle fission at the Golgi
apparatus is mediated by this molecular pinchase. Dynamin’s
association with Golgi membranes was demonstrated bio­
chemically with specific antibodies that detected dynamin in
immunoblots of purified rat liver Golgi fractions and by the
immuno‐isolation of Golgi membranes on magnetic beads
coated with those dynamin antibodies [89]. Cortactin, an actin‐
binding protein, in a complex with dynamin, functions in post‐
Golgi transport in liver cells. By using in vitro or intact cell
experiments, it was shown that activation of Arf1 recruits actin,
cortactin, and dynamin to Golgi membranes. Disruption of cort­
actin–dynamin interactions reduced dynamin recruitment to the
Golgi and blocked the transit of nascent proteins from the TGN,
which indicates an essential role of the cortactin–dynamin com­
plex in TGN function [90].
The dynamins are believed to form large helical polymers from
which many interactive proline‐rich tail domains extend. These
domains bind to a variety of SH3‐domain‐containing proteins,
many of which appear to be actin‐binding proteins such as Abp1
and syndapin. These findings suggest that the dynamin family
acts as a “polymeric contractile scaffold” at the interface between
biological membranes and filamentous actin (Figure 7.4).
A number of non‐dynamin‐mediated mechanisms for scis­
sion of vesicles, including those formed at the TGN, have
recently been described [91]. In the case of Rab6 vesicles
formed at the TGN, cytoskeletal motors play a crucial role.
Active Rab6‐GTP recruits myosin II to membranes, where its
motor activity on filamentous actin creates the force necessary
to mediate vesicle fission [92].
COORDINATION BETWEEN
THE SECRETORY PATHWAY
AND LIPID METABOLISM
The secretory pathway is intimately linked to lipid metabolism
pathways [93]. The ER is one of the major sites within the cell for
lipid synthesis, with multiple pathways operating to shuttle lipids
to different destinations. Both the phospholipids that make up cel­
lular membranes, and neutral storage lipids (triglycerides and cho­
lesterol esters) are synthesized by enzymes in the ER membrane
[94, 95]. In addition to their synthesis, neutral lipids are also
thought to be packaged into lipid droplets in the ER membrane
[95]. A regulatory switch between channeling of fatty acids into
either phospholipids for secretory membrane production or into
neutral lipids for storage in lipid droplets (LDs) was first identified
in yeast [96]. Subsequent work in mammalian cells has uncovered
similar regulatory mechanisms. A critical control point in this
switch is the metabolism of phosphatidic acid (PA), which serves
as a precursor of both phospholipids and storage triglycerides [97].
Indeed, PA can either be dephosphorylated by PA phosphatases to
produce diglyceride (DAG), which is then fatty acid acylated to
form triglycerides, or converted by CDP‐DAG synthase to CDP‐
DAG, a precursor of all the major cellular phospholipids. Hence
the action of PA phosphatases is responsible for channeling PA
towards storage lipid synthesis. There is one PA phosphatase in
yeast, Pah1, and three in mammalian cells, lipin1, lipin2, and
lipin3. A lipin1‐knockout mouse has a lipodsystrophy phenotype,
and Pah1 is required for LD formation in yeast, showing the
importance of these enzymes for lipid storage [97]. Pah1 and lipins
are localized both to the nucleus and cytoplasm, and are involved
in transcriptional regulation of lipid metabolism genes [97].
In addition to ubiquitous regulatory pathways linking
the  secretory pathway and lipid metabolism, hepatocytes also
use  the secretory pathway for transport of VLDL particles
(Figure 7.3a), a central function in organismal lipid homeostasis
[75, 79, 98]. The liver plays a key role in the body both in stor­
age of lipids and in their distribution to other tissues. In the liver,
fatty acids from the diet, from triglyceride breakdown, and from
biosynthesis are incorporated into LDs for storage, and into
VLDL particles for distribution. Both types of particles have a
 7:  The Hepatocellular Secretory Pathway 83
hydrophobic lipid core made up of triglycerides surrounded by
a phospholipid monolayer, likely arising from deposition of
­triglycerides between the leaflets of the bilayer of the ER mem­
brane [75, 79]. LDs surrounded by cytosolic leaflet phospholip­
ids subsequently bud towards the cytoplasm, whereas VLDL
particles surrounded by phospholipids of the luminal leaflet bud
into the ER lumen. VLDL particles then enter the secretory
pathway, from where they are released from hepatocytes into
the bloodstream (Figure 7.3b).
In hepatocytes, overexpression of lipin1 has been shown to
repress VLDL secretion [98, 99]. The mechanism has not been
fully elucidated, but an attractive hypothesis is that high levels
of lipin1 channel PA to triglyceride synthesis rather than
­phospholipid synthesis, the latter being required for secretory
pathway activity. This could have important implications for the
pathogenesis of alcoholic fatty liver disease (FLD), as it has
been shown that ethanol increases lipin1 levels [99]. Moreover,
ethanol exposure inhibits VLDL secretion in mouse liver, and
the increase in lipin1 expression induced by ethanol is mediated
by AMPK‐SREBP1 signaling [75, 99]. In response to changing
nutritional signals, mTORC1 regulates lipin activity by control­
ling its nuclear localization in hepatocytes [100].
FUTURE DIRECTIONS
Major hepatocellular functions include secretion of nascent
proteins and VLDL particles into the circulation, and the forma­
tion of bile, processes for which protein and lipid trafficking are
essential. This chapter has described some of the molecular
components required to direct nascent proteins and more
­complex lipoprotein particles (VLDL) from the ER to the Golgi
apparatus and then to their final destination. The core machin­
ery is evolutionarily conserved and operates in all cells, but is
modified for the specific trafficking processes performed by the
liver, such as secretion of VLDL. Attaining a better understand­
ing of how these processes are utilized by the liver under nor­
mal and pathophysiological conditions is a current and future
challenge for liver cell biologists. An important development
over the past few years has been elucidation of the intimate con­
nection between membrane trafficking pathways and lipid
transport and metabolism. Both production of cytosolic lipid
droplets and luminal VLDL particles is closely coupled to
vesicular trafficking from the ER to the Golgi. An important
area for future work is in understanding this crosstalk as it
occurs in liver cells. Elucidating the mechanisms involved will
provide avenues for therapeutic intervention not only in liver
diseases such as alcoholic and non‐alcoholic FLD, but also in
countering infection by viruses (such as hepatitis C and E)
which subvert these processes for their propagation.
ACKNOWLEDGMENTS
This chapter is based on the chapter by Susan Chi and Mark
McNiven in the previous edition. CLJ was supported by grants
from the ANR (ANR-13-BSV2-0013) and the “Fondation pour
la Recherche Médicale” (DEQ20150934717), France.
REFERENCES
1. Rapoport, T.A., Li, L., and Park, E. Structural and mechanistic insights into
protein translocation. Annu Rev Cell Dev Biol, 2017;33:369–90.
2. Glick, B.S. and Nakano, A. Membrane traffic within the Golgi apparatus.
Annu Rev Cell Dev Biol, 2009;25:113–32.
3. Nixon‐Abell, J., Obara, C.J., Weigel, A.V. et  al. Increased spatiotemporal
resolution reveals highly dynamic dense tubular matrices in the peripheral
ER. Science, 2016;354(6311).
4. Shibata, Y., Hu, J., Kozlov, M.M., and Rapoport, T.A. Mechanisms shaping the
membranes of cellular organelles. Annu Rev Cell Dev Biol, 2009;25:329–54.
5. Rambourg, A. and Clermont Y. Three‐dimensional electron microscopy:
structure of the Golgi apparatus. Eur J Cell Biol, 1990;51(2):189–200.
6. Farquhar, M.G. and Palade, G.E. The Golgi apparatus (complex)‐(1954–1981)‐
from artifact to center stage. J Cell Biol, 1981;91(3 Pt 2):77s–103s.
7. Bonifacino, J.S. and Glick, B.S. The mechanisms of vesicle budding and
fusion. Cell, 2004;116(2):153–66.
8. Dacks, J.B. and Robinson, M.S. Outerwear through the ages: evolutionary
cell biology of vesicle coats. Curr Opin Cell Biol, 2017;47:108–16.
9. Bykov, Y.S., Schaffer, M., Dodonova, S.O. et al. The structure of the COPI
coat determined within the cell. eLife, 2017;6.
10. Donaldson, J.G. and Jackson, C.L. ARF family G proteins and their regula­
tors: roles in membrane transport, development and disease. Nat Rev Mol
Cell Biol, 2011;12(6):362–75.
11. Gillingham, A.K. and Munro S. The small G proteins of the Arf family and
their regulators. Annu Rev Cell Dev Biol, 2007;23:579–611.
12. Lee, M.C., Miller, E.A., Goldberg, J., Orci, L., and Schekman R. Bi‐­
directional protein transport between the ER and Golgi. Annu Rev Cell Dev
Biol, 2004;20:87–123.
13. Lee, M.C., Orci, L., Hamamoto, S. et  al. Sar1p N‐terminal helix initiates
membrane curvature and completes the fission of a COPII vesicle. Cell,
2005;122(4):605–17.
14. Barlowe, C. and Helenius, A. Cargo capture and bulk flow in the early secre­
tory pathway. Annu Rev Cell Dev Biol, 2016;32:197–222.
15. Brown, H.A., Gutowski, S., Moomaw, C.R., Slaughter, C., and Sternweis,
P.C. ADP‐ribosylation factor, a small GTP‐dependent regulatory protein,
stimulates phospholipase D activity. Cell, 1993;75(6):1137–44.
16. De Matteis, M.A., Wilson, C., and D’Angelo, G. Phosphatidylinositol‐
4‐phosphate: the Golgi and beyond. BioEssays, 2013;35(7):612–22.
17. Ktistakis, N.T., Brown, H.A., Waters, M.G., Sternweis, P.C., and Roth, M.G.
Evidence that phospholipase D mediates ADP ribosylation factor‐dependent
formation of Golgi coated vesicles. J Cell Biol, 1996;134(2):295–306.
18. Ambroggio, E., Sorre, B., Bassereau, P. et al. ArfGAP1 generates an Arf1
gradient on continuous lipid membranes displaying flat and curved regions.
EMBO J, 2010;29(2):292–303.
19. Bannykh, S.I., Rowe, T., and Balch, W.E. The organization of endoplasmic
reticulum export complexes. J Cell Biol, 1996;135(1):19–35.
20. Cosson, P. and Letourneur F. Coatomer interaction with di‐lysine endoplas­
mic reticulum retention motifs. Science, 1994;263(5153):1629–31.
21. Letourneur, F., Gaynor, E.C., Hennecke, S. et al. Coatomer is essential for
retrieval of dilysine‐tagged proteins to the endoplasmic reticulum. Cell,
1994;79(7):1199–207.
22. Bonifacino, J.S. Adaptor proteins involved in polarized sorting. J Cell Biol,
2014;204(1):7–17.
23. Guo, Y., Sirkis, D.W., and Schekman R. Protein sorting at the trans‐Golgi
network. Annu Rev Cell Dev Biol, 2014;30:169–206.
24. Manolea, F., Chun, J., Chen, D.W. et  al. Arf3 is activated uniquely at the
trans‐Golgi network by brefeldin A‐inhibited guanine nucleotide exchange
factors. Mol Biol Cell, 2010;21(11):1836–49.
25. Devos, D., Dokudovskaya, S., Alber, F. et al. Components of coated vesicles
and nuclear pore complexes share a common molecular architecture. PLoS
Biol, 2004;2(12):e380.
26. Duden, R., Griffiths, G., Frank, R., Argos, P., and Kreis, T.E. Beta‐COP,
a 110 kd protein associated with non‐clathrin‐coated vesicles and the
Golgi complex, shows homology to beta‐adaptin. Cell, 1991;64(3):
649–65.
27. Serafini, T., Stenbeck, G., Brecht, A. et al. A coat subunit of Golgi‐derived
non‐clathrin‐coated vesicles with homology to the clathrin‐coated vesicle
coat protein beta‐adaptin. Nature, 1991;349(6306):215–20.
28. Beck, R., Rawet, M., Wieland, F.T., and Cassel D. The COPI system:
­molecular mechanisms and function. FEBS Lett, 2009;583(17):2701–9.
84 THE LIVER:  REFERENCES
29. Orci, L., Stamnes, M., Ravazzola, M. et al. Bidirectional transport by distinct
populations of COPI‐coated vesicles. Cell, 1997;90(2):335–49.
30. Bui, Q.T., Golinelli‐Cohen, M.P., and Jackson, C.L. Large Arf1 guanine nucle­
otide exchange factors: evolution, domain structure, and roles in membrane
trafficking and human disease. Mol Genet Genomics, 2009;282(4):329–50.
31. Casanova, J.E. Regulation of Arf activation: the Sec7 family of guanine
nucleotide exchange factors. Traffic, 2007;8(11):1476–85.
32. Novick, P., Field, C., and Schekman R. Identification of 23 complementation
groups required for post‐translational events in the yeast secretory pathway.
Cell, 1980;21(1):205–15.
33. Peyroche, A., Paris, S., and Jackson, C.L. Nucleotide exchange on ARF
mediated by yeast Gea1 protein. Nature, 1996;384(6608):479–81.
34. Claude, A., Zhao, B.P., Kuziemsky, C.E. et al. GBF1: a novel Golgi‐associ­
ated BFA‐resistant guanine nucleotide exchange factor that displays speci­
ficity for ADP‐ribosylation factor 5. J Cell Biol, 1999;146(1):71–84.
35. Schlacht, A., Mowbrey, K., Elias, M., Kahn, R.A., and Dacks, J.B. Ancient
complexity, opisthokont plasticity, and discovery of the 11th subfamily of
Arf GAP proteins. Traffic, 2013;14(6):636–49.
36. Schlacht A. Evolution of the vesicle formation machinery. PhD thesis,
Department of Cell Biology, University of Alberta, Edmonton, Canada,
director Joel Dacks, 2015.
37. Lowery, J., Szul, T., Styers, M. et al. The Sec7 guanine nucleotide exchange
factor GBF1 regulates membrane recruitment of BIG1 and BIG2 guanine
nucleotide exchange factors to the trans‐Golgi network (TGN). J Biol Chem,
2013;288(16):11532–45.
38. Bonifacino, J.S. and Lippincott‐Schwartz, J. Coat proteins: shaping mem­
brane transport. Nat Rev Mol Cell Biol, 2003;4(5):409–14.
39. Deng, Y., Golinelli‐Cohen, M.P., Smirnova, E., and Jackson, C.L. A COPI
coat subunit interacts directly with an early‐Golgi localized Arf exchange
factor. EMBO Rep, 2009;10(1):58–64.
40. Sato, K. and Nakano A. Dissection of COPII subunit‐cargo assembly and
disassembly kinetics during Sar1p‐GTP hydrolysis. Nat Struct Mol Biol,
2005;12(2):167–74.
41. Tanigawa, G., Orci, L., Amherdt, M. et al. Hydrolysis of bound GTP by ARF
protein triggers uncoating of Golgi‐derived COP‐coated vesicles. J Cell Biol,
1993;123(6 Pt 1):1365–71.
42. Forster, R., Weiss, M., Zimmermann, T. et  al. Secretory cargo regulates the
turnover of COPII subunits at single ER exit sites. Curr Biol,
2006;16(2):173–9.
43. Kliouchnikov, L., Bigay, J., Mesmin, B. et  al. Discrete determinants in
ArfGAP2/3 conferring Golgi localization and regulation by the COPI coat.
Mol Biol Cell, 2009;20(3):859–69.
44. Weimer, C., Beck, R., Eckert, P. et  al. Differential roles of ArfGAP1,
ArfGAP2, and ArfGAP3 in COPI trafficking. J Cell Biol, 2008;183(4):
725–35.
45. Bigay, J., Casella, J.F., Drin, G., Mesmin, B., and Antonny, B. ArfGAP1
responds to membrane curvature through the folding of a lipid packing sen­
sor motif. EMBO J, 2005;24(13):2244–53.
46. Bigay, J., Gounon, P., Robineau, S., and Antonny, B. Lipid packing sensed
by ArfGAP1 couples COPI coat disassembly to membrane bilayer curvature.
Nature, 2003;426(6966):563–6.
47. Barrowman, J., Bhandari, D., Reinisch, K., and Ferro‐Novick S. TRAPP
complexes in membrane traffic: convergence through a common Rab. Nat
Rev Mol Cell Biol, 2010;11(11):759–63.
48. Schroeter, S., Beckmann, S., and Schmitt, H.D. Coat/tether interactions‐
exception or rule? Front Cell Dev Biol, 2016;4:44.
49. Andag, U. and Schmitt, H.D. Ds11p, an essential component of the Golgi‐
endoplasmic reticulum retrieval system in yeast, uses the same sequence
motif to interact with different subunits of the COPI vesicle coat. J Biol
Chem, 2003;278(51):51722–34.
50. Reilly, B.A., Kraynack, B.A., VanRheenen, S.M., and Waters, M.G. Golgi‐
to‐endoplasmic reticulum (ER) retrograde traffic in yeast requires Ds11p, a
component of the ER target site that interacts with a COPI coat subunit. Mol
Biol Cell, 2001;12(12):3783–96.
51. Ren, Y., Yip, C.K., Tripathi, A. et  al. A structure‐based mechanism for
vesicle capture by the multisubunit tethering complex Ds11. Cell,
­ from the endoplasmic reticulum (ER) is COPII‐dependent, and its lipidation
2009;139(6):1119–29.
52. Suvorova, E.S., Duden, R., and Lupashin, V.V. The Sec34/Sec35p complex,
a Ypt1p effector required for retrograde intra‐Golgi trafficking, interacts
with Golgi SNAREs and COPI vesicle coat proteins. J Cell Biol,
2002;157(4):631–43.
53. Zolov, S.N. and Lupashin, V.V. Cog3p depletion blocks vesicle‐mediated
Golgi retrograde trafficking in HeLa cells. J Cell Biol, 2005;168(5):
747–59.
54. Guo, Y., Punj, V., Sengupta, D., and Linstedt, A.D. Coat‐tether interaction in
Golgi organization. Mol Biol Cell, 2008;19(7):2830–43.
55. Goueslain, L., Alsaleh, K., Horellou, P. et  al. Identification of GBF1 as a
cellular factor required for hepatitis C virus RNA replication. J Virol,
2010;84(2):773–87.
56. Farhat, R., Ankavay, M., Lebsir, N. et al. Identification of GBF1 as a cellular
factor required for hepatitis E virus RNA replication. Cell Microbiol,
2018;20(1).
57. Lavie, M. and Dubuisson J. Interplay between hepatitis C virus and lipid
metabolism during virus entry and assembly. Biochimie, 2017;141:62–9.
58. Farhat, R., Seron, K., Ferlin, J. et al. Identification of class II ADP‐ribosylation
factors as cellular factors required for hepatitis C virus replication. Cell
Microbiol, 2016;18(8):1121–33.
59. Altan‐Bonnet, N., Sougrat, R., and Lippincott‐Schwartz, J. Molecular basis
for Golgi maintenance and biogenesis. Curr Opin Cell Biol,
2004;16(4):364–72.
60. Jamieson, J.D. and Palade, G.E. Synthesis, intracellular transport, and
­discharge of secretory proteins in stimulated pancreatic exocrine cells. J Cell
Biol, 1971;50(1):135–58.
61. Presley, J.F., Cole, N.B., Schroer, T.A. et al. ER‐to‐Golgi transport ­visualized
in living cells. Nature, 1997;389(6646):81–5.
62. Barlowe, C., Orci, L., Yeung, T. et al. COPII: a membrane coat formed by
Sec proteins that drive vesicle budding from the endoplasmic reticulum.
Cell, 1994;77(6):895–907.
63. Mizuno, M. and Singer, S.J. A soluble secretory protein is first concentrated
in the endoplasmic reticulum before transfer to the Golgi apparatus. Proc
Natl Acad Sci U S A, 1993;90(12):5732–6.
64. Orci, L., Ravazzola, M., Meda, P. et al. Mammalian Sec23p homologue is
restricted to the endoplasmic reticulum transitional cytoplasm. Proc Natl
Acad Sci U S A, 1991;88(19):8611–15.
65. Bielli, A., Haney, C.J., Gabreski, G. et al. Regulation of Sar1 NH2 terminus
by GTP binding and hydrolysis promotes membrane deformation to control
COPII vesicle fission. J Cell Biol, 2005;171(6):919–24.
66. Aguilera‐Gomez, A., Zacharogianni, M., van Oorschot, M.M. et al. Phospho‐
rasputin stabilization by Sec16 is required for stress granule formation upon
amino acid starvation. Cell Rep, 2017;20(4):935–48.
67. Hanna, M.G., Block, S., Frankel, E.B. et  al. TFG facilitates outer coat
­disassembly on COPII transport carriers to promote tethering and fusion
with ER‐Golgi intermediate compartments. Proc Natl Acad Sci U S A,
2017;114(37):E7707–E16.
68. Saito, K., Chen, M., Bard, F. et  al. TANGO1 facilitates cargo loading at
endoplasmic reticulum exit sites. Cell, 2009;136(5):891–902.
69. Cai, H., Yu, S., Menon, S. et al. TRAPPI tethers COPII vesicles by binding
the coat subunit Sec23. Nature, 2007;445(7130):941–4.
70. Santos, A.J., Nogueira, C., Ortega‐Bellido, M., and Malhotra V. TANGO1
and Mia2/cTAGE5 (TALI) cooperate to export bulky pre‐chylomicrons/
VLDLs from the endoplasmic reticulum. J Cell Biol, 2016;213(3):343–54.
71. Shimoi, W., Ezawa, I., Nakamoto, K. et al. p125 is localized in endoplasmic
reticulum exit sites and involved in their organization. J Biol Chem,
2005;280(11):10141–8.
72. Melero, A., Chiaruttini, N., Karashima, T. et al. Lysophospholipids facilitate
COPII vesicle formation. Curr Biol, 2018;28(12):1950–8 e6.
73. Brodsky, J.L., Gusarova, V., and Fisher, E.A. Vesicular trafficking of hepatic
apolipoprotein B100 and its maturation to very low‐density lipoprotein par­
ticles; studies from cells and cell‐free systems. Trends Cardiovasc Med,
2004;14(4):127–32.
74. Zanetti, G., Pahuja, K.B., Studer, S., Shim, S., and Schekman R. COPII and the
regulation of protein sorting in mammals. Nat Cell Biol, 2011;14(1):20–8.
75. Gluchowski, N.L., Becuwe, M., Walther, T.C., and Farese, R.V., Jr. Lipid
droplets and liver disease: from basic biology to clinical implications. Nat
Rev Gastroenterol Hepatol, 2017;14(6):343–55.
76. Gusarova, V., Brodsky, J.L., and Fisher, E.A. Apolipoprotein B100 exit
to  very low density lipoprotein occurs post‐ER. J Biol Chem,
2003;278(48):48051–8.
77. Shoulders, C.C., Stephens, D.J., and Jones, B. The intracellular transport of
chylomicrons requires the small GTPase, Sar1b. Curr Opin Lipidol,
2004;15(2):191–7.
 7:  The Hepatocellular Secretory Pathway 85
78. Santos, A.J., Raote, I., Scarpa, M., Brouwers, N., and Malhotra V. TANGO1
recruits ERGIC membranes to the endoplasmic reticulum for procollagen
export. eLife, 2015;4.
79. Tiwari, S. and Siddiqi, S.A. Intracellular trafficking and secretion of VLDL.
Arterioscler Thromb Vasc Biol, 2012;32(5):1079–86.
80. Novikoff, P.M. and Yam A. Sites of lipoprotein particles in normal rat hepat­
ocytes. J Cell Biol, 1978;76(1):1–11.
81. Peden, A.A., Rudge, R.E., Lui, W.W., and Robinson, M.S. Assembly and
function of AP‐3 complexes in cells expressing mutant subunits. J Cell Biol,
2002;156(2):327–36.
82. Guo, X., Mattera, R., Ren, X. et  al. The adaptor protein‐1 mu1B subunit
expands the repertoire of basolateral sorting signal recognition in epithelial
cells. Dev Cell. 2013;27(3):353–66.
83. Doray, B., Bruns, K., Ghosh, P., and Kornfeld, S. Interaction of the cation‐
dependent mannose 6‐phosphate receptor with GGA proteins. J Biol Chem,
2002;277(21):18477–82.
84. Puertollano, R., Randazzo, P.A., Presley, J.F., Hartnell, L.M., and Bonifacino,
J.S. The GGAs promote ARF‐dependent recruitment of clathrin to the TGN.
Cell, 2001;105(1):93–102.
85. Collins, B.M., Praefcke, G.J., Robinson, M.S., and Owen, D.J. Structural
basis for binding of accessory proteins by the appendage domain of GGAs.
Nat Struct Biol, 2003;10(8):607–13.
86. Puertollano, R., Aguilar, R.C., Gorshkova, I., Crouch, R.J., and Bonifacino,
J.S. Sorting of mannose 6‐phosphate receptors mediated by the GGAs.
Science, 2001;292(5522):1712–16.
87. Zhen, Y. and Stenmark H. Cellular functions of Rab GTPases at a glance.
J Cell Sci, 2015;128(17):3171–6.
88. Wang, J., Morita, Y., Mazelova, J., and Deretic, D. The Arf GAP ASAP1
provides a platform to regulate Arf4‐ and Rab11‐Rab8‐mediated ciliary
receptor targeting. EMBO J, 2012;31(20):4057–71.
89. Henley, J.R. and McNiven, M.A. Association of a dynamin‐like protein with
the Golgi apparatus in mammalian cells. J Cell Biol, 1996;133(4):761–75.
90. Cao, H., Weller, S., Orth, J.D. et al. Actin and Arf1‐dependent recruitment
of a cortactin‐dynamin complex to the Golgi regulates post‐Golgi transport.
Nat Cell Biol, 2005;7(5):483–92.
91. Renard, H.F., Johannes, L., and Morsomme, P. Increasing diversity of biologi­
cal membrane fission mechanisms. Trends Cell Biol, 2018;28(4):274–86.
92. Miserey‐Lenkei, S., Chalancon, G., Bardin, S. et al. Rab and actomyosin‐
dependent fission of transport vesicles at the Golgi complex. Nat Cell Biol,
2010;12(7):645–54.
93. Jackson, C.L., Walch, L., and Verbavatz, J.M. Lipids and their trafficking:
an integral part of cellular organization. Dev Cell, 2016;39(2):139–53.
94. Vance, J.E. Phospholipid synthesis and transport in mammalian cells.
Traffic, 2015;16(1):1–18.
95. Walther, T.C., Chung, J., and Farese, R.V., Jr. Lipid droplet biogenesis.
Annu Rev Cell Dev Biol, 2017;33:491–510.
96. Gaspar, M.L., Hofbauer, H.F., Kohlwein, S.D., and Henry, S.A.
Coordination of storage lipid synthesis and membrane biogenesis: evidence
for cross‐talk between triacylglycerol metabolism and phosphatidylinositol
synthesis. J Biol Chem, 2011;286(3):1696–708.
97. Siniossoglou, S. Phospholipid metabolism and nuclear function: roles of
the lipin family of phosphatidic acid phosphatases. Biochim Biophys Acta,
2013;1831(3):575–81.
98. Wang, Y., Viscarra, J., Kim, S.J., and Sul, H.S. Transcriptional regulation of
hepatic lipogenesis. Nat Rev Mol Cell Biol, 2015;16(11):678–89.
99. You, M., Jogasuria, A., Lee, K. et al. Signal transduction mechanisms of
alcoholic fatty liver disease: emerging role of lipin‐1. Curr Mol Pharmacol,
2017;10(3):226–36.
100. Peterson, T.R., Sengupta, S.S., Harris, T.E. et al. mTOR complex 1 ­regulates
lipin 1 localization to control the SREBP pathway. Cell, 2011;146(3):408–20.
 Mitochondrial Function,
8 Dynamics, and Quality Control
Marc Liesa, Ilan Benador, Nathanael Miller, and Orian S. Shirihai
Department of Medicine, Division of Endocrinology and Department of Molecular and Medical Pharmacology,
David Geffen School of Medicine at UCLA, Los Angeles, CA, USA

INTRODUCTION MITOCHONDRIAL OXIDATIVE

The word “mitochondrion” originates from the fusion of two
Greek words: mitos, meaning thread and chondros, meaning
grain [1]. This term was coined by Benda after visual inspection
of these organelles. It is worth noting that thread and grain are
the most common shapes of bacteria: coccus was coined for
grain‐shaped bacteria and bacillus for thread‐shaped bacteria.
Later, it was demonstrated that mitochondria originated from
bacteria engulfed by the ancestors of eukaryotic cells. The intra-
cellular parasite Rickettsia is widely accepted as the α‐proteo-
bacteria order from which mitochondria originated and is
morphologically classified as a cocco‐bacillus bacterial species.
Therefore, the endosymbiotic theory has both anatomical and
DNA conservation evidence in its favor. In this regard, the main
features of mitochondria that are present among all eukaryotic
species are conserved in bacteria, namely: (i) Mitochondria con-
sume oxygen to synthesize ATP, using a process named oxida-
tive phosphorylation (OXPHOS), similar to bacterial aerobic
ATP synthesis. (ii) Mitochondria have their own genome, which
encodes for 13 subunits of the respiratory complexes executing
OXPHOS, tRNA, and rRNAs required for their translation. (iii)
Mitochondria are motile organelles that can fuse or divide into
smaller organelles, similar to bacteria.
In this chapter, we will provide a summary of these three
conserved features of mitochondria, with a focus on hepato-
cyte mitochondrial OXPHOS function and the role of
­mitochondrial dynamics in hepatocytes determining OXPHOS
function.
PHOSPHORYLATION (OXPHOS)
Mitochondria are organelles formed by two membrane layers
with different compositions of lipids and proteins, separated
by an intermembrane space (IMS). Over 1500 proteins are esti-
mated to reside in human mitochondria, but only 13 are encoded
by the mitochondrial DNA (mtDNA). This low number of
­proteins means that protein import through the mitochondrial
membrane layers is a central process for mitochondrial biogen-
esis, function, and turnover.
The outer mitochondrial membrane is bidirectionally perme-
able to small solutes, including small peptides (<5 kDa), thanks
to the presence of a protein that acts as a channel, named porin/
VDAC. The family of TOM (translocase of the outer membrane)
proteins is responsible for protein transport across the outer
mitochondrial membrane [2, 3].
The permeability of the inner mitochondrial membrane is
more stringent, which limits the bidirectional diffusion of very
small molecules, including protons (H+, 1 Da). This stringency
means that the transport across the inner membrane of nutrients,
fuels, metabolites, and cofactors of polar nature is strictly
dependent on transporter proteins and their regulatory mecha-
nisms. The inner membrane is the location of proteins that exe-
cute oxidative phosphorylation (OXPHOS), a process briefly
defined as a chain of redox events involving electron transport
across protein multimers (complexes) that extrude protons to
the IMS, with proton re‐entry through ATP synthase producing
ATP. The stringent permeability to protons, together with high

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
 8:  Mitochondrial Function, Dynamics, and Quality Control 87
proton concentrations within the IMS limiting nutrient oxida-
tion rates, allows ATP synthase and thus ATP demand to control
OXPHOS rates (i.e. how much nutrient is oxidized and how
much oxygen is consumed). Moreover, protein import across
the inner membrane to build mitochondria is executed by the
family of TIM proteins (translocase of the inner membrane),
with energy‐demanding translocation of proteins also being
fueled by the difference in charge and pH across the inner mem-
brane generated by the proton gradient [4]. The dependency of
these two crucial processes on the proton gradient illustrate that
the stringent permeability of the inner membrane is essential for
proper mitochondrial function.
In this context, mitochondrial OXPHOS is constituted by
three major processes acting in coordination and ultimately
determined by ATP synthase activity: (i) nutrient and fuel oxida-
tion; (ii) electron transport chain and cytochrome c oxidase
activity; and (iii) mitochondrial ATP synthase activity.
Nutrient and fuel oxidation
Ninety percent of the oxygen that we breathe and transform to
water and CO2 is consumed by mitochondria. This can explain
why some nutrients are fully oxidized inside the mitochondria
(some amino acids and fatty acids). Moreover, nutrients whose
oxidation does not occur inside mitochondria can provide fuels/
intermediates to be oxidized inside mitochondria (i.e. glucose‐
derived glycerol‐3‐phosphate and pyruvate). Nutrient oxidation
reactions in the mitochondria mostly occur in the matrix, with a
few in the IMS (glycerol‐3‐phosphate dehydrogenase). These
reactions are catalyzed by enzymes named dehydrogenases that
strip electrons from nutrients and fuels and transfer them to
nicotine and flavin adenine dinucleotides (NAD+ and FAD+).
These adenine dinucleotides can accept two electrons each and
carry them for their transfer to the electron transport chain. In
the case of hepatocytes, mitochondrial nutrient and fuel oxida-
tion is highly regulated and complex. This complexity and high
regulation means that nutrient oxidation in the mitochondria is
not exclusively determined by nutrient availability outside or
inside hepatocytes, or even in the capacity of OXPHOS com-
plexes to transfer electrons coupled to ATP synthesis. There are
four underlying reasons for this complexity as describing in the
following.
Mitochondrial nutrient oxidation can sustain nutrient
synthesis in hepatocytes
Hepatocytes are responsible for synthesis and supply nutrients
to other tissues (preferentially the brain) under organismal fast-
ing and starvation, mostly glucose and ketone bodies. Under
fasting, hepatocytes do not oxidize glucose, they oxidize fatty
acids mostly in the mitochondria to sustain their ATP demand
and produce ketone bodies [5]. In the case of amino acids, their
oxidation during fasting can be used to provide ATP and carbon
intermediates for gluconeogenesis. Under the fed state, nutrient
oxidation preferences change, with mitochondria oxidizing
pyruvate resulting from glucose oxidation, a process that sup-
ports fatty acid synthesis. Thus, mitochondrial nutrient oxida-
tion and fuel/nutrient transport is specifically tuned to support
these context‐dependent hepatocyte functions, being sensitive
to hormonal regulation (which respond to fasting and feeding).
Oxidation products from different nutrients converge
in the TCA cycle
Glucose‐derived pyruvate, amino acid, and fatty acid oxidation in
the mitochondria result in intermediate and final metabolites that
can enter the citric acid/Krebs/TCA cycle. This redox cycle con-
sists of eight reactions catalyzed by: (i) four dehydrogenases (two
of which are decarboxylases), (ii) two synthetases, (iii) one
isomerase, and (iv) one hydratase that interconnects eight carbox-
ylic acids, with four to six carbons in their structure. The cycle
starts with the synthesis of citrate (six carbons) from the condensa-
tion of oxaloacetate (four carbons) and acetyl‐CoA (two carbons)
catalyzed by citrate synthase. Importantly, two carbons are lost as
two CO2 in one turn of the cycle, while generating three NADH,
one FADH2, one ATP, or GTP. This means that a minimal entry of
two carbons is required in each turn to replace the two carbons lost
as CO2, as no carboxylases are part of the core enzymes constitut-
ing the TCA cycle itself. As these eight TCA cycle reactions occur
within the mitochondrial matrix and the eight TCA acids are nega-
tively charged, their concentrations in the matrix need to be main-
tained within a range. This maintenance is not only required to
sustain the TCA cycle “spinning” to provide transferable electrons
and ATP, but also to preserve the energy contained in the differ-
ence in charge and pH generated by the proton gradient across the
inner membrane. The need for carbon supply in a context of a
limited capacity to change TCA cycle metabolite concentrations
in the matrix is widely accepted to explain the requirement for
anaplerotic and cataplerotic processes.
Anaplerosis and cataplerosis are required to sustain TCA
cycle rates and mitochondrial function
The carboxylic acids constituting the TCA cycle can be consumed
for de novo synthesis of multiple molecules, including glucose,
fatty acids, and heme, among others. The processes that replenish
TCA cycle intermediates after their consumption are named “ana-
plerotic.” In contrast, “cataplerosis” refers to the processes that
consume TCA cycle intermediates, which can be linked to biosyn-
thetic processes. Hepatocytes have a strong demand for anaplero-
sis to sustain cataplerotic processes that are regulated by fed and
fasting states, including synthesis of fatty acids, glucose, and
heme. Thus, anaplerosis and cataplerosis are tightly and hormo-
nally regulated in hepatocytes. The perfect example for an anaple-
rotic process in hepatocytes is the reaction catalyzed by
mitochondrial pyruvate carboxylase. This enzyme uses ATP
hydrolysis to add one carbon from bicarbonate (which can be
originated from CO2) to pyruvate. Under the fed state, pyruvate
derived from glucose is used by both pyruvate carboxylase and
pyruvate dehydrogenase to generate oxaloacetate and acetyl‐CoA,
respectively. In this way, pyruvate carboxylase activity replenishes
the oxaloacetate that left the TCA cycle, as citrate is used for fatty
acid synthesis or that left to be a carbon source for glucose synthe-
sis (cataplerosis). Another major cataplerotic process in liver is
heme synthesis, as the first step in its synthesis involves the con-
sumption of succinyl‐CoA (one of the eight TCA intermediates).
Glucose‐derived pyruvate and fatty acids compete
to be oxidized by mitochondrial OXPHOS
This competition facilitates the transition and therefore the
execution of the opposite roles in fat metabolism of hepatocyte
88 THE LIVER:  MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION (OXPHOS)
mitochondria in the fed versus the fasting state. In the post-
prandial state, glucose‐derived pyruvate is oxidized in the
mitochondrial matrix to generate citrate to be used for de novo
fatty acid synthesis, whereas during fasting, mitochondrial fatty
acid oxidation fuels glucose synthesis and ketone body produc-
tion. Key molecular mechanisms allowing this transition based
on competition can be summarized in two points:
• Product inhibition of TCA dehydrogenases and pyruvate
dehydrogenase (PDH): High levels of ATP, acetyl‐CoA, and
NADH during fatty acid oxidation in the matrix can slow
down PDH and TCA cycle oxidation rates by product inhibi-
tion, decreasing CO2 release [6, 7]. Slowing down the TCA
cycle allows balancing anaplerosis resulting from fatty acid
and amino acid oxidation, with cataplerosis consuming
oxaloacetate and acetyl‐CoA for glucose and ketone body
synthesis, respectively. This competitive mechanism in liver
and in other tissues such as muscle contributes to the decrease
in organismal CO2 production rates under fasting or when
fatty acid oxidation is the major nutrient being oxidized.
• Inhibition of long‐chain fatty acid entry into the mitochondria:
When mitochondrial oxidation of glucose‐derived pyruvate
increases in the fed state, cataplerosis mandates that elevated
citrate is exported out of the mitochondria. Exported citrate is
the precursor in the cytosol of malonyl‐CoA, a molecule that
inhibits activated fatty acid translocation into the mitochondria
by blocking the protein executing this transport, namely CPT1
(located in the outer membrane) [8]. The fact that glucose‐medi-
ated inhibition of fatty acid oxidation occurs by inhibiting the
entry of fatty acids supports the conclusion that, by default, fatty
acid oxidation will win the competition with pyruvate oxidation
in the mitochondrial matrix. It is also likely that it allows an
active and faster import of long‐chain fatty acids into the mito-
chondria to meet ATP demand and ketone body production
under fasting, which might not be possible by diffusion.
Electron transport chain and cytochrome c
oxidase activity
The electron transport chain is constituted by four multiprotein
complexes, named complexes I, II, III, and IV. Electrons carried
by NADH enter through complex I, a multiprotein complex also
known as NADH:ubiquinone oxidoreductase, whereas electrons
carried by FADH2 can enter via complex II, also known as the
TCA cycle enzyme succinate dehydrogenase. There are two
additional sites for electron entry from FADH2: inner mem-
brane‐bound glycerol‐3‐phosphate dehydrogenase and the elec-
tron‐transferring flavoprotein (ETF) and ETF:ubiquinone
oxidoreductase (ETF:QO). Of note, glycerol‐3‐phosphate is an
intermediate generated in the cytosol by glycolysis, while ETF
accepts the electrons from FADH2 generated by 11 different
mitochondrial dehydrogenases, mostly involved in mitochon-
drial fatty acid and amino acid oxidation processes. The destina-
tion of the electrons stripped from NADH and FADH2 is
complex IV, also named cytochrome c oxidase. Complex IV
uses these electrons to reduce O2 to water and, therefore, it is
directly responsible for respiration.
The process of electron transfer from the entry sites to com-
plex IV involves four steps, which are highly conserved in mito-
chondria from all different tissues and are shared:
• Ubiquinone reduction: Two electrons are used to reduce
ubiquinone to ubiquinol, which is inserted in the inner
­mitochondrial membrane. Ubiquinol transfers the electrons
to complex III and can freely diffuse within the inner mito-
chondrial membrane. This diffusion allows the transfer from
the accepting electron complexes to complex III.
• Complex III‐mediated reduction of cytochrome c: Complex III
is also named ubiquinol:cytochrome c oxidoreductase. Its role
is to transfer the electrons from ubiquinol to cytochrome c.
Cytochrome c is a small protein (104 amino acids) containing
a heme group and associated with the inner membrane. The
electrons are accepted by the Fe3+ within the heme group to be
reduced to Fe2+. Electrons from Fe2+‐cytochrome c molecules
are stripped by complex IV, to regenerate Fe3+‐cytochrome c
and transfer the electrons to O2.
• Proton translocation: Complexes I, III, and IV are the multi-
protein complexes. Complex I comprises 44–45 proteins,
complex III has 11 proteins, and complex IV has 13–14
­proteins (one protein initially thought to be part of complex
I  is now thought to be part of complex IV). The presence
of multiple proteins allows coupling of the electron transport
process and redox events to proton pumping to the IMS.
Accordingly, complexes transferring electrons and with lower
numbers of proteins do not pump protons. Complex II con-
tains 4 proteins, ETF/ETF:QO has 2 proteins, and glycerol‐3‐
phosphate dehydrogenase is a single protein. Although these
complexes do not pump protons themselves, the electrons
delivered to ubiquinone will still induce proton pumping when
they reach complexes III and IV. As these complexes can pro-
vide electrons much faster than complex I, it makes sense that
they do not pump protons to prevent a sudden increase in
membrane potential. Such an increase could even stop the
electron transport process itself, which increases the risk of
these electrons generating reactive oxygen species instead of
reaching complex IV to reduce oxygen and ­generate water.
• Supercomplex formation: The complexes that pump protons (I,
III, and IV) can be found in isolation but also form units named
“supercomplexes.” The exact mechanisms behind supercom-
plex assembly and their physiological regulation by feeding and
fasting in mitochondria have only been characterized recently,
and they are still subject to intense research [9]. Evidence sug-
gests that supercomplex regulation and composition might dif-
fer in mitochondria from different tissues (i.e. heart is different
from liver). However, there is agreement that supercomplex
formation can increase the efficiency of electron transfer by
cytochrome c from complex III to complex IV. Because
cytochrome c is a small protein loosely bound to the inner
membrane, the formation of a unit containing complexes III and
IV allows a pool of cytochrome c to be trapped in between
them. This trapped pool would allow electron transfer to be
faster, because it would not be solely dependent on random dif-
fusion of cytochrome c from complex III to complex IV. A simi-
lar explanation has been given to justify the interaction between
complexes I and III, but in this case it facilitates electron trans-
fer mediated by membrane‐inserted but diffusible ubiquinol.
Mitochondrial ATP synthase
Mitochondrial ATP synthase is a multiprotein complex formed
by 13 proteins divided into two structurally defined regions: a
 8:  Mitochondrial Function, Dynamics, and Quality Control 89
hydrophilic, matrix‐located region called F1 and a hydrophobic
region called F0 inserted in the inner membrane. ATP synthase is
also known as complex V, although it cannot transport electrons
itself. The rationale behind being named complex V is the func-
tional connection of ATP synthesis with the electron transport
chain, usually referred as “coupling.” This coupling is achieved
by the need for ATP synthase to translocate protons to harness
the energy required to phosphorylate ADP to ATP. This energy
transfer starts in the F0 region, which forms the proton pore
across the inner membrane and constitutes a motor that rotates
when protons translocate. The F0 motor directly interacts with
F1 region proteins, which as a result also rotate. The rotation of
F1 proteins causes conformational changes in other F1 proteins
to catalyze the phosphorylation of ADP to ATP in the matrix.
Mitochondrial ATP synthase activity and, therefore, cellular
ATP demand can determine and control nutrient oxidation and
respiration rates. The electron transport chain and ATP synthase
evolved in such a way that a large difference in the proton gradient
slows down electron transfer and high ATP levels in the mitochon-
drial matrix inhibit complex V‐mediated proton translocation
back to the matrix. Therefore, if the ATP/ADP ratio is high in the
matrix, protons cannot translocate, and electron transport will
stop, which will also prevent the complexes stripping electrons
from NADH and FADH2. The accumulation of NADH and
FADH2 will then decrease nutrient oxidation rates.
Importantly though, the mitochondrial proton gradient does
not exclusively serve ATP synthesis activity. In agreement with
this, nutrient and fuel oxidation in the mitochondria do not
exclusively serve either proton gradient maintenance or ATP
synthesis. Indeed, the intermediates or end products of mito-
chondrial nutrient/fuel oxidation are used to synthesize other
nutrients or essential cofactors such as heme, which may be
needed under conditions of high ATP/ADP ratio. Therefore,
other mechanisms are recruited to serve these additional pur-
poses of the proton gradient and mitochondrial nutrient oxida-
tion that are independent of changes in ATP demand. One of
these mechanisms is proton leak: some protons can still cross
the inner membrane when ATP synthase activity (i.e. ATP
demand) is low, but ATP synthase is still the major mechanism
of proton re‐entry to the mitochondria. This is indeed the reason
why it is called proton leak, as high ATP levels can still reduce
respiration rates, and the low respiration and nutrient oxidation
rates are possible thanks to the leak of protons (i.e. a small num-
ber). Uncoupling happens when respiration and nutrient oxida-
tion is increased because of another mechanism allowing proton
re‐entry at a higher capacity than the ATP synthase itself. This
uncoupling is not common in physiology; it is used to generate
heat in specialized tissues, such as brown adipose tissue.
Reverse ATP synthase activity
The reverse activity of mitochondrial ATP synthase reveals the
importance of the proton gradient beyond energizing ATP syn-
thesis. Reverse activity is the “default activity” of ATP synthase,
which is to hydrolyze ATP and extrude protons when the differ-
ence in membrane potential (proton gradient) is not large enough
[10]. This means that if the electron transport chain is damaged
and cannot sustain the proton gradient, the ATP synthase can
extrude protons to the IMS, energized by ATP hydrolysis. In this
way ATP synthase substitutes for the role of the electron
transport chain in maintaining the difference in charge generated
by the proton gradient. This reverse activity is thought to allow
protein import to replace dysfunctional components of the elec-
tron transport chain. Moreover, it still allows substrate and fuel
import for cataplerosis.
MITOCHONDRIAL DYNAMICS
AND TURNOVER, AND THEIR
RELATIONSHIP TO OXPHOS
In 1915, Lewis and Lewis showed that mitochondria inside cells
can move from one location to another, and even visualized
events supporting the existence of mitochondrial fusion and fis-
sion/division [11].
Regulation of mitochondrial dynamics
and quality control
Mitochondrial fusion is regulated by fusion proteins: mito-
fusins 1 and 2 (Mfn1 and Mfn2) fuse the outer mitochondrial
membrane, and optic atrophy 1 (OPA1) fuses the inner mito-
chondrial membrane (Figure  8.1). Conversely, mitochondrial
fission factor (Mff) and dynamin related protein 1 (Drp1) are
responsible for mitochondrial fission. These proteins tightly
regulate mitochondrial dynamics and morphology via hydroly-
sis of GTP to power their activities [12–14]. Mitochondrial
dynamics control mitochondrial architecture. Whereas fusion
and fission are ­controlled at the level of each mitochondrion in
the cell, an overriding signal may result in cellular reduction in
fusion or fission activity. Increased mitochondrial fission and
decreased fusion lead to fragmentation of the mitochondrial
network. An example of an overriding signal – the effect of cell
cycle on fusion and fission activity – is illustrated in Figure 8.2.
Mfn2 expression is decreased in different tissues by diet‐
induced obesity and contributes to tissue dysfunction [17–20].
Mfn2 has been implicated in the regulation of mitochondrial
function in addition to its role in mitochondrial fusion. Knockdown
of Mfn2 results in altered mitochondrial bioenergetic function, as
reduced glucose oxidation and decreased oxygen consumption in
primary rodent muscle and muscle‐derived cell lines is evident
[17]. Removal of Mfn2 altered substrate preference and impaired
thermogenesis in brown adipose tissue [21, 22]. These results col-
lectively demonstrate that mitochondrial morphology and tissue
function are intimately connected.
Mitochondrial morphology and dynamics control mitochon-
drial degradation via mitochondria‐specific autophagy (or
mitophagy) [23, 24], in part by constantly mixing and reorgan-
izing the mitochondrial content distribution, but also by gener-
ating small fission products that can be eliminated by mitophagy.
Fission events generate daughter mitochondria that can depo-
larize, leading to PINK1‐mediated phosphorylation of Mfn2
[25–28]. PINK1, which is normally degraded in functional,
polarized mitochondria, is stabilized on depolarized mitochon-
dria [29] due to inhibition of its protease. Stabilized PINK1
phosphorylates Mfn2, which is targeted by Parkin. Parkin ubiq-
uitinates Mfn2, whose ubiquitination designates mitochondria
destined for removal by mitophagy.
90 THE LIVER:  MITOCHONDRIAL DYNAMICS AND TURNOVER, AND THEIR RELATIONSHIP TO OXPHOS

Mitophagy

opa1↓
Fission mfn1↓
PINK1↑
parkin↑
2

Pre-autophagic pool
Fusion 3 4
Sustained
depolarization

1
Transient
depolarization

Figure 8.1  The mitochondria life cycle. Mitochondria go through continuous cycles of fusion and fission. Fusion is brief (1). It triggers a fission
event within a period of seconds to a few moments (2). A daughter mitochondrion may maintain an intact membrane potential (orange) or depolarize
(3, green). When depolarized, a subsequent fusion event is unlikely to take place, unless the mitochondrion repolarizes. Depolarized and solitary
mitochondria (4) remain for several hours in a pre‐autophagic pool (5) prior to removal by autophagy. Adapted from [15] with permission of Elsevier.

Initiation of
mitochondrial Hyperfusion of
network fusion mitochondrial
network
GLOBAL
Autophagy CONTROL

Mitochondrial
network begins
fragmentation
Recovery of
membrane
potential
Solitary
period

De- Intact
polarization membrane
potential Mitochondrial
fragmentation
and dispersion
to opposite
telomeres

Fission

LOCAL
CONTROL Fusion

Reorganization
Sequestration
@ 2009 VisuallyMedical.com

Figure 8.2  Mitochondrial life in the cell cycle. This diagram depicts the normal life cycle of an individual mitochondrion during the G0 phase of
the cell cycle. The mitochondrion undergoes fusion, fission, depolarization, and degradation by autophagy. This process is depicted as one of local
control whereby mitochondrial events are largely dictated by the local energetic status and associated local signals. During the cell cycle global
signals cause concerted changes in the mitochondrial population, as noted by hyperfusion in G1–S and fragmentation during M phase. These global
population effects are governed by the cell’s demand for energy required by cell division and the need for homogenization and sequestration of
cellular components during met‐phase. The cell cycle serves as an elegant example of the parities of local and global control. Adapted from [16]
with permission of Elsevier.
 8:  Mitochondrial Function, Dynamics, and Quality Control 91

The selective elimination of defective mitochondria, or
mitophagy, is a critical housekeeping process within a cell [25].
Mfn2 is also the target of ubiquitination by a second, redundant
pro‐mitophagy ubiquitin ligase, MUL1, whose loss exacerbates
the Parkin‐null phenotype [28]. This is a tightly regulated process,
where deubiquitinases such as ubiquitin C‐terminal hydrolase L1
(UCH‐L1) act in opposition to the ubiquintation of Mfn2 and other
proteins [30, 31]. Mitophagy is well‐controlled but can fail at any
of the above steps. Although the above mechanisms describe the
targeting of defective mitochondria to the mitophagic process,
mitophagy also requires proper lysosomal function, which has
been shown to be impaired in conditions of excess nutrients [32].
Loss of mitochondrial dynamics proteins leads to diverse
changes in mitochondrial function. For example, tissue-specific
ablation of Drp1 in liver protected against diet‐induced obesity
and glucose intolerance while increasing energy expenditure in
mice [33]. Similarly, liver-specific loss of mitofusin 1 increased
mitochondrial mass and shifted their fuel preference to the burn-
ing of fatty acids, while protecting from the effects of diet‐induced
obesity [34]. Conversely, loss of mitofusin 2 resulted in impaired
glucose metabolism, decreased lipid accumulation, glucose toler-
ance in liver and other tissues, and increased hepatic glucose
­production [20, 35]. Taken together, these data suggest a role of
morphology in controlling fuel preference of mitochondria.
SUBCELLULAR INDIVIDUALISM
AND EXCLUSIVE CLUBS
OF MITOCHONDRIA
The total number of mitochondria per cell varies in different tis-
sues and cell types, with reports estimating between 88 and
2000 mitochondria per cell. The discovery of mitochondrial
fusion, with the intracellular architecture looking like an inter-
connected network in some cell lines, supported the analogy
that OXPHOS is executed in all mitochondria as a choir of
88–2000 people singing a song in unison.
However, because different cellular processes with different
ATP demands can occur in different intracellular locations, it
would make physiological sense that mitochondria in different
locations might be functioning differently or executing different
anabolic processes. A classic example is the different mitochon-
drial subpopulations identified in muscle, with mitochondria
located between muscle fibers having a higher capacity to syn-
thesize ATP to support contraction compared to mitochondria
located below the fiber membrane (subsarcolemmal). In the adi-
pose tissue, a subpopulation of mitochondria that are uniquely
adherent to the lipid droplet has recently been identified
(Figure  8.3). These were named peridroplet mitochondria
(PDM) and have also been found in the liver (Figure 8.4).
What are the peridroplet mitochondria?
PDM are mitochondria that reside within 0.5 μm of lipid drop-
lets (LDs). Electron microscopy studies of adipose cells have
shown that PDM are characterized by an elongated morphol-
ogy, increased electron density in the mitochondria–LD contact
site, and cristae oriented perpendicular to the axis of mitochon-
dria–LD contact site. This unique morphology supported the
concept that PDM play a specialized role in fat metabolism.
However, it was unclear whether PDM promote fat oxidation,
fat storage, or both.
Recent studies have demonstrated that PDM are relatively
immobile and do not fuse with cytoplasmic mitochondria (CM)
and thereby maintain a distinct proteome and bioenergetic
­function. Compared to CM, PDM have higher levels of
ATP  synthase and cytochrome c oxidase protein levels and

BAT Low-speed High-speed

Dissection centrifugation centrifugation Peridroplet
Fat cake mitochondria (PDM)
(900 × g) (9,000 × g)

Supernatant Cytoplasmic
mitochondria (CM)

Nuclei and (Classical isolation)

unbroken cells

Figure 8.3  Isolation of peridroplet mitochondria by differential centrifugation. Schematic representation of peridroplet mitochondrial (PDM)
isolation technique. Tissue is dissected and homogenized with a glass–Teflon dounce homogenizer. Low‐speed centrifugation separates the fat cake
containing PDM from supernatant containing cytoplasmic mitochondria (CM). High‐speed centrifugation strips PDM from lipid droplets (LDs)
and pelleted CM mitochondria from the supernatant. Note that some mitochondrial isolation protocols discard the fat cake and/or begin with high‐
speed centrifugation step. LDs were marked by the neutral BODIPY 493/503 fluorescent dye (BODIPY) and mitochondria by MitoTracker deep
red dye (MitoTracker). Note the tubular structures staining positively for MitoTracker on LDs. Adapted from [36] with permission of Elsevier.
92 THE LIVER:  REFERENCES
Lean HFD
Figure 8.4  Mitochondrial association with lipid droplets can be observed in the liver. The association is particularly noticeable in livers from mice
on a high‐fat diet (HFD) and in ob/ob mice, although this has not been statistically validated. Adapted from [37] with permission of Springer Nature.
enzymatic capacity. Importantly, PDM have higher pyruvate
oxidation capacity but lower fatty acid oxidation capacity com-
pared to CM. Furthermore, cells with increased levels of PDM
have higher levels of LD expansion and triacyglyceride synthe-
sis. The current evidence supports the concept that PDM pro-
mote LD expansion through ATP‐dependent triacylglyceride
synthesis, which may protect mitochondria from lipotoxicity in
periods of lipid excess.
Peridroplet mitochondria in the liver
The liver is tasked with the rapid packaging of lipids into lipo-
proteins for distribution to other tissues in the body. However,
under conditions of nutrient excess, lipids accumulate within
hepatocytes, resulting in a chronic state of inflammation and
cirrhosis. The need for rapid and effective lipid synthesis in
hepatocytes suggests a potential role of PDM in maintaining
lipid homeostasis of hepatocytes.
Although PDM have not been specifically studied in liver,
evidence suggests the existence of PDM in hepatocytes. For
example, published electron microscopy images show levels of
PDM in both genetic and high‐fat diet mouse models of obesity
[37] (Figure 8.4). In addition, high‐fat feeding in both mice and
humans has been consistently associated with increased expres-
sion of perilipin5 (Plin5), an LD coat protein that strongly
recruits mitochondria to LDs [38]. Cell and animal models have
confirmed that overexpression of Plin5 in the liver expands LD
mass and protects against lipotoxic liver injury. This provides
further support for the concept that PDM support lipid homeo-
stasis in hepatocytes.
REFERENCES
1. Benda, C. Ueber die Spermatogenese der Vertebraten und höherer
Evertebraten. II. Theil: Die Histiogenese der Spermien. Verh Berl Physiol
Ges, 1898:393–8.
2. Koehler, C.M. Protein translocation pathways of the mitochondrion. FEBS
Lett, 2000;476(1–2):27–31.
3. Perry, A.J., Hulett, J.M., Likić, V.A., Lithgow, T., and Gooley, P.R. Convergent
evolution of receptors for protein import into mitochondria. Curr Biol,
2006;16(3):221–9.
ob/ob
4. Schmidt, O., Pfanner, N., and Meisinger, C.. Mitochondrial protein import:
From proteomics to functional mechanisms. Nat Rev Mol Cell Biol,
2010;11(9):655–67.
5. Boyd, M.E., Albright, E.B., Foster, D.W., and McGarry, J.D. In vitro reversal
of the fasting state of liver metabolism in the rat. Reevaluation of the roles of
insulin and glucose. J Clin Invest, 1981;68(1):142–52.
6. Garland, P.B., Shepherd, D., Nicholls, D.G., and Ontko, J. Energy‐dependent
control of the tricarboxylic acid cycle by fatty acid oxidation in rat liver
mitochondria. Adv Enzyme Regul, 1968;6:3–30.
7. Konig, T., Nicholls, D.G., and Garland, P.B. The inhibition of pyruvate and
Ls+‐isocitrate oxidation by succinate oxidation in rat liver mitochondria.
Biochem J, 1969;114(3):589–96.
8. McGarry, J.D., Takabayashi, Y., and Foster, D.W. The role of malonyl‐CoA
in the coordination of fatty acid synthesis and oxidation in isolated rat
­hepatocytes. J Biol Chem, 1978;253(22):8294–300.
9. Lapuente‐Brun, E., Moreno‐Loshuertos, R., Acin‐Perez, R. et  al.
Supercomplex assembly determines electron flux in the mitochondrial elec-
tron transport chain. Science, 2013;340(6140):1567–70.
10. Campanella, M., Casswell, E., Chong, S. et al. Regulation of mitochondrial
structure and function by the F1Fo‐ATPase inhibitor protein, IF1. Cell
Metab, 2008;8(1):13–25.
11. Lewis, M.R. and Lewis, W.H. Mitochondria (and other cytoplasmic
­structures) in tissue cultures. Am J Anat, 1915;17(3):339–401. https://doi.
org/10.1002/aja.1000170304
12. Sesaki, H. and Jensen, R.E. Division versus fusion: Dnm1p and Fz01p antag-
onistically regulate mitochondrial shape. J Cell Biol, 1999;147(4):699–706.
https://doi.org/10.1083/jcb.147.4.699
13. Shaw, J.M. and Nunnari, J. Mitochondrial dynamics and division in budding
yeast. Trends Cell Biol, 2002;12(4):178–84. https://doi.org/10.3174/ajnr.
A1256.Functional
14. Yaffe, M.P. The machinery of mitochondrial inheritance and behavior.
Science (New York), 1999;283(5407):1493–7.
15. Liesa, M. and Shirihai, O.S. Mitochondrial dynamics in the regulation of
nutrient utilization and energy expenditure. Cell Metab, 2013;17(4):
491–506.
16. Hyde, B.B., Twig, G., and Shirihai, O.S. Organellar vs cellular control of
mitochondrial dynamics. Semin Cell Dev Biol, 2010;21(6):575–81.
17. Bach, D., Pich, S., Soriano, F.X. et al. Mitofusin‐2 determines mitochondrial
network architecture and mitochondrial metabolism. A novel regulatory
mechanism altered in obesity. J Biol Chem, 2003;278(19):17190–7.
18. Le Blanc, S., Villarroel, P., Candia, V. et al. Type 2 diabetic patients and their
offspring show altered parameters of iron status, oxidative stress and genes
related to mitochondrial activity. Biometals, 2012;25(4):725–35.
19. Pich, S., Bach, D., Briones, P. et al. The Charcot‐Marie‐Tooth type 2A gene
product, Mfn2, up‐regulates fuel oxidation through expression of OXPHOS
system. Hum Mol Genet, 2005;14(11):1405–15.
20. Sebastian, D., Hernandez‐Alvarez, M.I., Segales, J. et al. Mitofusin 2 (Mfn2)
links mitochondrial and endoplasmic reticulum function with insulin signal-
ing and is essential for normal glucose homeostasis. Proc Natl Acad Sci U S
A, 2012;109(14):5523–8.
 8:  Mitochondrial Function, Dynamics, and Quality Control 93
21. Boutant, M., Kulkarni, S.S., Joffraud, M. et al. Mfn2 is critical for brown
adipose tissue thermogenic function. EMBO J, 2017;36(11):1543–58.
https://doi.org/10.15252/embj.201694914
22. Mahdaviani, K., Benador, I.Y., Su, S. et al. Mfn2 deletion in brown adipose
tissue protects from insulin resistance and impairs thermogenesis. EMBO
Rep, 2017;18(7):1123–38.
23. Dagda, R.K., Cherra, S.J., Kulich, S.M. et  al. Loss of PINK1 function
­promotes mitophagy through effects on oxidative stress and mitochondrial
fission. J Biol Chem, 2009;284(20):13843–55.
24. Twig, G., Elorza, A., Molina, A.J.A. et  al. Fission and selective fusion
­govern mitochondrial segregation and elimination by autophagy. EMBO J,
2008;27(January), 433–46.
25. Chen, Y. and Dorn, G.W.. PINK1‐phosphorylated mitofusin 2 is a Parkin
receptor for culling damaged mitochondria. Science (New York),
2013;340(6131):471–5.
26. Pickrell, A.M. and Youle, R.J. The roles of PINK1, parkin, and mitochon-
drial fidelity in Parkinson’s disease. Neuron, 2015;85(2):257–73. https://doi.
org/10.1016/j.neuron.2014.12.007
27. Song, M., Mihara, K., Chen, Y., Scorrano, L., and Ii, G.W.D. Mitochondrial
fission and fusion factors reciprocally orchestrate mitophagic culling in
mouse hearts and cultured fibroblasts. Cell Metab, 2015;21(2):273–85.
28. Yun, J., Puri, R., Yang, H. et al. MUL1 acts in parallel to the PINK1/parkin
pathway in regulating mitofusin and compensates for loss of PINK1/parkin.
ELife, 2014;3:e01958.
29. Narendra, D.P., Jin, S.M., Tanaka, A. et al. PINK1 is selectively stabilized on
impaired mitochondria to activate Parkin. PLoS Biol, 2010;8(1):e1000298.
30. Costes, S., Huang, C.J., Gurlo, T. et al. β‐cell dysfunctional ­ERAD/­ubiquitin/
proteasome system in type 2 diabetes mediated by islet amyloid polypeptide‐
induced UCH‐L1 deficiency. Diabetes, 2011;60(1):227–38.
31. Costes, S., Gurlo, T., Rivera, J.F., and Butler, P.C. UCHL1 deficiency exac-
erbates human islet amyloid polypeptide toxicity in B‐cells. Autophagy,
2014;10(6):1004–14.
32. Trudeau, K.M., Colby, A.H., Zeng, J. et al. Lysosome acidification by photo-
activated nanoparticles restores autophagy under lipotoxicity. J Cell Biol,
2016;214(1):25–34.
33. Wang, L., Ishihara, T., Ibayashi, Y. et al. Disruption of mitochondrial fission
in the liver protects mice from diet‐induced obesity and metabolic deteriora-
tion. Diabetologia, 2015;58(10):2371–80.
34. Kulkarni, S.S., Joffraud, M., Boutant, M. et al. Mfn1 deficiency in the liver
protects against diet‐induced insulin resistance and enhances the hypoglyce-
mic effect of metformin. Diabetes, 2016;65(12):3552–60.
35. Chen, X. and Xu, Y. Liver‐specific reduction of Mfn2 protein by RNAi
results in impaired glycometabolism and lipid homeostasis in BALB/c mice.
J Huazhong Uni Sci Technol Med Sci 2009;29(6):689–96.
36. Benador, I.Y., Veliova, M., Mahdaviani, K. et al. Mitochondria bound to lipid
droplets have unique bioenergetics, composition, and dynamics that support
lipid droplet expansion. Cell Metab, 2018;27(4),869–85.e6.
37. Arruda, A.P., Pers, B.M., Parlakgül, G. et al. Chronic enrichment of hepatic
endoplasmic reticulum‐mitochondria contact leads to mitochondrial dys-
function in obesity. Nature Med, 2014;20(12):1427–35.
38. Wang, C., Zhao, Y., Gao, X. et al. Perilipin 5 improves hepatic lipotoxicity
by inhibiting lipolysis. Hepatology, 2015;61(3):870–82.
 Nuclear Pore Complex
9 Michelle A. Veronin and Joseph S. Glavy
University of Texas at Tyler, Fisch College of Pharmacy, Tyler, TX, USA
OVERVIEW
In eukaryotic cells, the separation of nuclear DNA from
cytoplasmic organelles and cellular materials is governed by
the nuclear envelope. The nuclear envelope (NE) is defined
as a specialized endoplasmic reticulum (ER) membrane with
a double bilayer containing an inner and outer membrane
system. Although this partitioning protects the genome, it
also is needed to organize nuclear entry and exit for basic
cellular processes (Figure  9.1). Nuclear pore complexes
(NPCs), the gateways for macromolecular traffic into and
out of the nucleus, are set in circular openings of the double
membrane of the NE [1, 2]. There are highly regulated path-
ways that control nuclear entry and exit of molecules such as
transcription factors, RNAs, kinases, and viral particles. In
broad spectrum, proteins with their transport signaling
sequence to be imported or exported from the nucleus (i)
bind to transport receptors, (ii) are transported through
NPCs present in the NE, and (iii) translocate from the NPC
to intranuclear or cytoplasmic target sites [1, 3]. About a
thousand proteins are needed to build one NPC. NPCs are
composed of proteins termed nucleoporins or Nups, which
play a role in the structure of the NPC and in regulating the
translocation of molecules through the NPC [1, 3]. And the
composition of NPCs differs from species to species [4–6].
Abnormalities such as chromosomal translocations involv-
ing genes encoding Nups have been associated with many
forms of leukemia [7]. Proteins of the NPC have been asso-
ciated with cancer, Huntington’s disease, micronuclear for-
mation, viral infection, aging, triple A syndrome, and
primary biliary cirrhosis [7–10].
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
NUCLEAR PORE COMPLEX
The NPC is a large macromolecular assembly with an esti-
mated size of 110 MDa in vertebrates and 60 MDa in yeast
(Figure 9.1) [11–14]. NPCs possess a symmetric core with an
octahedral arrangement across the double membrane of the
nuclear envelope, resembling the spokes of a bicycle wheel. They
are composed of specialized proteins termed nucleoporins or
Nups. Proteomic studies in yeast and metazoa cataloged the
Nups of the NPC [4, 5]. Mammalian NPCs contained at least
seven additional Nups, including ALADIN, Nup358, Pom210,
Pom121, NDC1, Nup43, and Nup37 (Figure  9.2). So only
around 30 types of Nups exist in NPCs but they are present in
multiple copies, which adds to the puzzle of this large trans-
porter. Nups can be classified into six groups: (i) Y‐complex
(coat) Nups, (ii) adaptor Nups, (iii) channel Nups, (iv) cyto-
plasmic filament Nups, (v) nuclear basket Nups, (vi) mobile
Nup98, and (vii) POM (transmembrane) Nups (Figure  9.2).
Nups have structural designs to interlock (coiled coils, β‐pro-
pellers or disordered areas) while channel Nups contain flexi-
ble repeating units of phenylalanine‐glycine (FG) (forming sea
anemone‐like fingers within the NPC which are implicated in
active nuclear transport) [1, 15–17]. Nups are basically named
by their designated molecular weight and are organized into
macromolecular assemblies called subcomplexes. Nup sub-
complexes release from the NPC during open mitosis and are
the disassembly units of the NPC [18]. In the laboratory, inter-
phase subcomplexes can be isolated through biochemical
extraction of the NE with low percentage non‐ionic or zwitteri-
onic detergent treatment (such as 2% Triton X‐100 and/or 1%
CHAPS) [19–21]. A key point of NPCs is that these modular
 9:  Nuclear Pore Complex 95

(a) (b)

Figure 9.1  (a) Thin‐section electron microscopy illustrating the nuclear pore complex (NPC) in isolated HeLa cell nuclear envelope (NE).
Courtesy of Dr. Samuel Dales. (b) Artist’s rendition of NPC. Courtesy of Grace Glavy.

Cytoplasm

Nup358
Nup358

hCG1 hCG1
ALADIN ALADIN
Rae1 Rae1
Nup214 Nup214
Nup88 Nup98 Nup98 Nup88
Nup43 Nup160 Nup160 Nup43
Nup37 Nup133 Nup133 Nup37
Sec13 Nup107 Nup107 Sec13
Seh1Nup96 Nup96 Seh1
Nup75 Nup75
gp210 Nup205 Nup205 gp210
Nup188 Nup62 Nup62
Nup188
NDC1 Nup155 Nup58 Nup58 Nup155 NDC1
Nup93 Nup54 Nup54 Nup93
Nup Nup
pom121 35 pom121
35
Nup160 Nup160
Seh1 Nup133 Nup133 Seh1
Sec13 Nup107 Nup107 Sec13
Nup43 Nup96 Nup96 Nup43
Nup37 Nup75 Nup75 Nup37
Nup
153

Nup
153

Nup98 Nup98
Rae1 Nup50 Nup50 Rae1

Tpr Tpr

Nucleoplasm
Figure 9.2  Overall schematic of subcomplexes and unique non‐subcomplex proteins: Y‐Nups (dark blue), adaptor Nups (light blue), channel
Nups (violet), cytoplasmic filament Nups (green), pore membrane (POM) Nups (light brown), nuclear basket Nups (orange and dark brown), and
mobile Nup98/Rae1.

units are present in multiple copies arranged around two‐ and
eightfold axes of symmetry and form discrete structures within
the NPC [15]. Overall NPC architecture is conserved between
yeast and higher eukaryotes with an eightfold symmetry [15, 22].
Cryoelectron tomography (cryo‐ET) with three‐dimensional
reconstruction gives a resolution of 23 Å (Figure 9.3) [11, 12].
The nuclear basket extends nearly 60 nm into the nucleus
(Figure 9.3) [11–13, 23, 24]. The central core inner ring (IR) is
estimated at less than 50 nm while the envelope is approxi-
mately 50 nm [23]. The cytoplasmic rings (CRs) are believed to
act as a docking site for protein transport and bind to nuclear
rings (NRs) [11–13, 23, 24]. An early approach using a combi-
nation of proteomic data with a computational platform has
been applied to the architectures of the macromolecular assem-
blies of the NPC [25, 26]. The study concludes that the funda-
mental symmetry unit of the NPC is the spoke [25, 26]. Within
96 THE LIVER:  NUCLEOPORINS (NUPS)
the ringed structure are linked units and flexible repeat units that
both stabilize and are involved in transport [1]. The inner and
outer rings help to facilitate the membrane structure (Figure 9.3).
The complete architecture of the NPC includes some non‐Nup
proteins – the nuclear membrane proteins and the nuclear lam-
ina [15]. These non‐Nup proteins make up the surrounding
regions of the NPC and support its transport function.
NUCLEOPORINS (NUPS)
Y‐Nups
Y‐Nups are contained within the Y‐complex (Nup107 subcom-
plex) (Figure 9.2). The cytoplasmic Y‐complex has nine mem-
bers: Nup160, Nup133, Nup107, Nup96, Nup75, Nup43,
Nup37, Seh1, and Sec 13, whereas the nuclear Nup107 subcom-
plex contains a tenth member: ELYS (Figure 9.2) [11, 13, 21,
27–30]. This subcomplex has been classified as a keystone of
NPC assembly [31]. RNAi knockdown experiments of individ-
ual members within the Y‐complex affect select members of the
subcomplex plus some other channel Nups [20, 31]. These find-
ings show the interdependence of subcomplex members and the
overall NPC. Positioned at the curvatures of the membrane
embedding the NPC, the Y‐complex acts to stabilize these bends
in the membrane (Figure  9.3) [32]. In the “protocoatomer”
hypothesis, components of the Y‐complex serve as membrane‐
curving modules similar to the members of the COPI, COPII,
and clathrin complexes [33]. One member of the Y‐complex,
Sec13, is also found as key member of the COPII complex.
Furthermore, ArfGAP1 lipid packing sensor (ALPS) domains
are found within members of the subcomplex [25, 34]. ALPS
domains are classified as membrane‐binding amphipathic α‐
helix regions. Nup133 contains an ALPS domain shown to bind
to isolated membrane [34]. Additional Nups, including Nup107,
Nup160, and Nup188, contain predicted ALPS domains [25].
Nup205/
Nup155 Nup93 Nup62
Nup188
Figure 9.3  Cross‐section of the nuclear pore complex from the cytoplasm (top) to the nucleoplasm (bottom). Structure is based on cryoelectron
tomography. The curvature of the nuclear envelope (NE) is shown: outer nuclear membrane (ONM), luminal curve (LC), and inner nuclear mem-
brane (INM). The NPC is labeled: cytoplasmic rings (CRs) and nuclear rings (NRs).
Early reconstitution experiments in yeast were assembled
into a Y‐complex visualized through negative staining electron
microscopy [35]. The structural modules of the subcomplex are
a collection of α‐helical repeats and β‐propellers. X‐ray crystal
structures have been reported for several members. Analysis of
Nup133 revealed a seven‐bladed β‐propeller domain of the N‐
terminus [36]. The crystal structure of the α‐helical interaction
of the Nup107–133 shows an elongated structure forming a
compact interface in a tail‐to‐tail fashion [36]. This interaction
is a critical attachment point for Nup133 [36]. Another crystal
structure of Nup96–Sec13 forms a hetero‐octamer [37]. The N‐
terminus of Nup96 invades the six‐bladed β‐propeller structure
of Sec13, providing a seventh blade [37]. The remaining portion
of Nup96 forms an antiparallel α‐helical domain. The potential
interlocking modules of the subcomplex may be the meshwork
of the overall macromolecule. The potential coating modules
may form a cylinder layer that apposes the membrane [37].
Analysis of the Nup75–Seh1 X‐ray structure led to conclusions
that the scaffold Nups form a membrane‐bordering lattice, pro-
viding attachment sites for additional Nups such as channel
Nups, Nup98, and Nup155 [38].
Through X‐ray crystallography the NPC can be pieced
together bit by bit with the Y‐complex as the main component.
Larger crystal approaches has yielded great progress. A yeast
hexameric Y‐complex was achieved using a single‐domain
synthetic antibody crystallization chaperone [39]. Among the
information provided was further proof of an evolutionarily
conserved ring structure formed by the yeast Y‐complex. Our
cryo‐ET (approximately 23 Å) work has shown that the human
Y‐complex forms two reticulated rings on the cytoplasmic and
again on the nuclear side [12–14] (Figure 9.3). Numerous dis-
crete crystal structures of the Y‐complex fit directly into the
tomographic map, including the yeast hexameric structure [11–
14]. In our integrated approach, we coupled electron tomogra-
phy, single‐particle electron microscopy, and cross-linking mass
spectrometry to show that 32 copies of the Y‐complex amass
CRs
ONM
LC
INM
NRs
Outer Inner
Nup58 Nup54
Y complex Y complex
 9:  Nuclear Pore Complex 97
into two reticulated rings, one at each of the cytoplasmic and
nuclear faces of the NPC [11]. This twin‐ring organization of
the Y‐complex leaves an explanation for structural plasticity of
the NPC [11]. Flexible and spring‐shaped hinges provide for
large‐scale rearrangements that might be relevant for large unit
transport [11] (Figure 9.3). Furthermore, we connected the cyto-
plasmic filaments Nups to the Y‐complex [11–14] (see the cyto-
plasmic filaments section).
Adaptor Nups
The adaptor complex (formerly called the Nup93 subcom-
plex) includes five members: Nup93, Nup205, Nup188,
Nup155, and Nup35 [40]. The X‐ray crystal structure of
Nup93 reveals an elongated, α‐helical structure  [41]. This
form is evolutionarily conserved and therefore functionally
maintained [41]. Members of the adaptor complex contain
mainly α‐helical domains [41]. Just like the Y‐complex,
Nup93 is a highly abundant protein with 32 copies within the
NPC [30]. The Nup93 subcomplex aids in IR stabilization
[41] and is needed for correct nuclear pore assembly and
homeostasis of the NPC [41]. RNAi experiments suggest a
functional link between NE transmembrane NDC1, Nup93,
and Nup205. These results suggest an anchor point function
for the Nup93 subcomplex [42]. This has been confirmed by
structure studies that Nup93 is a key component to the IR.
Furthermore, Nup62 (channel Nup) has been shown to inter-
act with Nup93, illustrating interdependence between IR
Nups and the channel Nups [42]. Using an integrated
approach, we found in our studies that 32 copies of both
Nup188 and Nup205 (Figure  9.3, yellow), Nup93 (red),
Nup155 (green), and the channel Nups (next section: Nup62,
Nup58, and Nup54) fit into the IR with additional Nup155
protein reaching up and connecting to the outer ring
(Figure 9.3, green). As shown in Figure 9.3, the IRs comprise
rings similar to the Y‐complex, confirming their evolutionary
connection [12]. This resolves the subdomain of Nup205 and
Nup188 and the asymmetric structures of Nup93 into the IR
[12] and links channel Nups to the ring structure [12].
Channel Nups
Channel Nups have stretches of FG (Phe–Gly) repetitive resi-
dues, which are separated by polar spacer regions of variable
lengths (Figure 9.1b) [1, 16]. FG repeat domains form unstruc-
tured regions that form weak interactions with transporting pro-
teins called karyopherins (kaps) [1]. Channel Nups such as the
Nup62 subcomplex are located in the inner pore region or cen-
tral core IR [23, 24]. The Nup62 subcomplex includes Nup62,
Nup58–Nup45, and Nup54 (Figures 9.2 and 9.3) [15, 43]. This
subcomplex was referred to as the central plug region of the
NPC. Although these transport Nups line the inner NPC, it is
unlikely that they form a plug against transport, but rather a
dynamic transport area of the complex. The FG repeat domains
form tentacle‐like structures that emanate from and line the
channel of the pore (discussed further in the transport section).
In Figure 9.3, they are shown to line the IR region. Not shown
is  how they extend into the channel and would be ready to
receive cargo.
Nuclear basket Nups
Nup153 and Nup50, along with Tpr (translocated promoter
region protein), make up the nuclear basket and provide the sur-
face to utilize binding areas for transport. Tpr is an unusual Nup
(nucleoporin Tpr) with more filament protein properties than
most yet required for trafficking across the nuclear envelope.
Tpr acts as a framework component in the nuclear segment and
tethering chromatin to begin perinuclear heterochromatin exclu-
sion zones. Tpr is also believed to act as docking site for express-
ing genes interacting with select Nups. It participates in both
nuclear import and export pathways for proteins with or without
nuclear export sequences (NES) as well as the export of mRNA
[44, 45]. Tpr coiled‐coil domains help give a reliable support to
form and maintain the nuclear basket. Tpr plays a role in mitotic
spindle checkpoint signaling [44, 45]. Within the nuclear
basket, Nup153 is associated with Tpr and contains four zinc
fingers. These zinc fingers increase the local Ran concentration
to assist nuclear transport [46]. Along with Nup50, Nup153
helps to terminate karyopherin‐mediated transport (Figures 9.2
and 9.4) [1].
Cytoplasmic filament Nups
As in the case of channel Nups, cytoplasmic filament Nups pos-
sess FG regions. Their floppy tentacle nature is cohesive with
crystal formation. Thus far, only non‐FG regions of these Nups
have been reported. The X‐ray crystal structure of the non‐FG
repeat N‐terminus of Nup214 reveals a seven‐bladed β‐propel-
ler with a segment of its C‐terminus bound to the propeller [47].
Furthermore, X‐ray analysis of Nup58/45 revealed a possible
circumferential sliding mechanism to adjust the diameter of the
central transport channel [43]. The α‐helical region forms dis-
tinct tetramers with a hydrophobic interface. The residues are
laterally displaced in numerous tetramer confirmations, giving
the possibility of a sliding structure [43]. Selective knockdown
of Nup214 followed by cryo‐ET demonstrates that this subcom-
plex protrudes into the cytoplasmic ring region [11]. This posi-
tion secures FG repeats onto the framework of the rings to
facilitate nuclear transport [11]. Further application of gene‐
silencing/cryo‐ET uncovered a role for Nup358 to stabilize
solely the cytoplasmic reticulated double ring structure  [13].
These findings change the edges between Y and cytoplasmic
filament Nups [13].
Mobile Nup: Nup98
Nup98 is found both at the nuclear pore complex and within the
nucleus [27]. It has multiple functions and binding partners
[48]. Nup98 has roles in transport, mitotic progression, gene
expression, epigenetic changes, and viral infection [9, 48–52].
Nup98 arises from a Nup98‐–Nup96 precursor form that splits
by a self‐cleavage domain similar to those found in Drosophila
hedgehog and Flavobacterium glycosylasparaginase [27, 53]. It
is classified as a non‐subcomplex Nup with multiple locations
along both sides of the NPC (Figure  9.3) [27]. Nup98 ferries
between internuclear regions to NPCs while regulating tran-
scription of select genes implicated in development and growth
[9, 54–57]. The N‐terminus of Nup98 contains FG repeats used
98 THE LIVER:  NUCLEOPORINS (NUPS)

α-kap
D
1 T D
6
Importin T
β-kap
2

3
NLS-cargo

Cytoplasm

Nucleus

T
T RanGTP T
4
D RanGDP T
Exportin
β-kap

Figure 9.4  Karyopherin‐dependent transport cycle. Key components for the cycle are RanGTP, nuclear localization sequence (NLS)–cargo, kaps,
and nuclear pore complex (NPC). Key concept is that the concentration of RanGTP is much higher in the nucleus. (1) Formation of the α‐kap and
β‐kap transport complex. (2) α‐Kap recognition of NLS–cargo with β‐kap. (3) Selective transport through NPC. (4) Once inside, the high levels of
nuclear RanGTP increase the likelihood of interaction and hence, disassembly of the transport complex. (5) GTP‐bound β‐kap transport through
the NPC to the cytosol, while α‐kap forms a complex with GTP‐bound export β‐kap, then exits by way of the nearby NPC. Left behind is the NLS–
cargo. (6) Once back in the cytosol, GTP undergoes hydrolysis and kaps are recycled for the next round of transport.

as a docking site for kaps, but is not classified as an FG Nup but
rather as a mobile Nup [27, 58]. Nup98 facilitates mRNA export
from the nucleus [27, 59]. Mitotic phospho‐Nup98 is a deter-
mining factor in NPC disassembly [56, 57]. One of Nup98’s
interacting partners is Rae1; together they act as temporal regu-
lators of the anaphase‐promoting complex [60]. Vesicular sto-
matitis virus (VSV) matrix (M) protein binds Nup98 and inhibits
nuclear export of mRNA [59]. In some instances, inhibition has
been linked to activation of p53 [61]. Nup98 gene translocations
have been linked to several forms of leukemia (covered in
section on NPC and disease) [9]. Nup98 has also been impli-
cated in nuclear envelope breakdown (NEBD) (see later NEBD
section) [56, 62].
Pore membrane Nups
The NE is separated into three domains: the outer nuclear mem-
brane (ONM), the pore membrane (POM), and the inner nuclear
membrane (INM) [2]. The POM region is the sector of the NE
where INM and ONM fusion occurs. The POM curvatures are
lined with select membrane proteins believed to anchor NPCs
like the settings of a ring. POM proteins are involved in the ini-
tiation of pore complex formation, stabilization, release and ref-
ormation of the NPC. To date, four POM proteins have been
classified through proteomic and genetic screenings: gp210,
Pom121, NDC1, and Pom33. The largest POM protein, gp210
(Pom210), contains a single transmembrane domain as do the
rest of the POM proteins. NDC1 is found both in the POM
region and the spindle pole body (SPB) during mitosis. In yeast
it helps to anchor the SPB at the NE [63], but in human cells its
purpose is still unknown. RNAi experiments suggest a func-
tional link between NE membrane NDC1 and two members of
the adaptor complex: Nup93 and Nup205 [64]. These results
suggest an anchor point function for the Nup93 subcomplex
[42]. Elimination of Pom121 results in failure to assemble NPCs
and formation of continuous nuclear membranes [65]. Cross-
linking mass spectrometry (XL‐MS) reveals a direct linkage
between Nup133 and Pom121 at the POM interface [11].
Although no single POM or combination of POMs is com-
pletely indispensable for survival, implying structural/
 9:  Nuclear Pore Complex 99
functional redundancy of diverse transmembrane and soluble
Nups, nuclear import and soluble Nup localization is most
­perturbed in cells lacking NDC1 and Pom121 [64, 66, 67].
POM‐depletion studies point to the need for additional compo-
nents to restore overall POM function. NDC1 was shown to
form a linkage between the NE and soluble Nups [64, 68, 69]
and depletion of NDC1 shows markedly reduced channel Nup
staining compared to NDC1/control siRNA treatment [64, 69].
These results point to a crucial role for NDC1 in structure, func-
tion, NEBD, and assembly. POM protein “put‐back” experi-
ments fail to restore function, implying the need for additional
components as well as POM proteins [64, 69, 70]. We recognize
Pom121 is required for interphase assembly and localized with
Nups at forming pores [71]. In fact, a fragment of Pom121 has a
dominant‐negative effect on pore assembly [72]. Thus Pom121
plays an important role in assembly and nuclear pore biogene-
sis. The protein’s repeat‐containing POM domain is involved in
anchoring components of the pore complex to the pore mem-
brane (Figure 9.2) and when overexpressed in cells induces the
formation of cytoplasmic annulate lamellae [73, 74]. Interactions
of Pom121 with Nup155 and Nup160 are foreseen to contribute
to the formation of the nuclear pore as well as the fastening of
the NPC to the pore membrane [67].
The newest POM protein, Pom33, is required for proper NPC
distribution [75, 76] as well as assembly. But a percentage of
Pom33 resides in the ER [75, 76]. Although it is an element of
the tubular ER network, Pom33 (TMEM33) also acts as a POM
protein at NPCs [75, 76]. The C‐terminus of Pom33 contains
several amphipathic regions that both bind to Kap123 and serve
as attachment points, aiding in positioning the POM region
while N‐terminus transmembrane regions anchor the POM [76].
Phosphorylation of gp210 may be mediated by cyclin‐B–
cdc2, and depletion of cyclin B from C. elegans embryos also
leads to a nuclear‐twinning phenotype [77]. It has been shown
that gp210 is important for effective NPC disassembly and sug-
gests that phosphorylation of gp210 is an early event in NEBD
[76]. Yet, the role of this set of POM proteins at the onset of
disassembly is undetermined.
NUCLEAR LAMINA
In general, lamins fall into the category of either A‐type or B‐
type lamins, and these classifications are based on differing fea-
tures including protein structure and expression patterns [78].
For example, A‐type lamins contain an additional C‐terminal
domain consisting of a unique 90 amino acid sequence not pre-
sent in B‐type lamins. Additionally, B‐type lamins are expressed
in most cells, whereas A‐type lamins are expressed primarily in
differentiated cells [78]. Three lamin genes are encoded within
mammalian genomes: LMNA, LMNB1, and LMNB2. At least
seven protein isoforms are produced from alternative splicing;
the two major isoforms expressed from LMNA are lamin A and
C, which are ubiquitous in most differentiated cells [78]. Lamins
polymerize to form higher order structures, a process involving
dimerization from their α‐helical domain followed by head‐to‐
tail parallel association between dimers [78]. These higher order
structures form the structural foundation of the nuclear lamina
that contributes to nuclear stability [79]. The inner membrane of
the NE in mammalian cells contains a unique array of integral
membrane proteins including lamina‐associated polypeptides 1
and 2 (LAP1 and LAP2), emerin, MAN1, nesprin‐1, nesprin‐2,
and lamin B receptor (LBR) [2]. These proteins contain an LEM
domain (LAP2, emerin, and Man1 domain), which is a 40‐residue
motif that interacts specifically with lamin A/C [80].
When the NE breaks down during mitosis, the nuclear lamina
also disassembles, which occurs when lamins are phosphoryl-
ated by protein kinase C. During this disassembly, A‐type
lamins are dispersed throughout the cytoplasm while B‐type
lamins remain attached to the nuclear membrane. Because lamin
A is not farnesylated, it is more soluble than lamin B [78].
During NE reassembly, lamins are dephosphorylated by type 1
protein phosphatase to facilitate reassembly of the lamina, and
A‐type and B‐type lamins are transported into the nucleus [78].
Lamins have also been shown to play an important structural
role in determining nuclear morphology and stability. Multiple
studies have demonstrated that lowered expression of functional
lamins results in nuclei that are more fragile and prone to defor-
mation [78, 81]. Evidence for the platform role of lamins in the
formation of multiprotein complexes that function within the
nucleus is supported by their wide variety of interaction
partners, such as retinoblastoma 1 and c‐Fos [78]. Lamins are
also involved in facilitating the connection between the nucleus
and the cytoplasm through the LINC complex, which contains
lamin‐interacting proteins, as well as transducing signals from
the cytoskeleton into the nucleus through contact with NPCs
[78, 82].
In addition, it has been shown that lamins may be involved in
DNA repair. One study demonstrated that expression of disease‐
causing mutant lamins alters localization and expression of ATR
kinase, a regulator of DNA repair, which in turn may adversely
affect DNA repair pathways requiring the function of ATR as
well as the additional repair regulator ATM [83]. Furthermore,
the presence of mutant lamins was shown to reduce recruitment
of DNA repair factor 53BP1, a p53‐binding protein, in response
to DNA damage [83]. This suppression of 53BP1 may be con-
nected to degradation by cysteine protease cathepsin‐L (CTSL),
whose upregulation is associated with insufficient expression of
lamin A/C [79]. Substantial DNA damage, as well as alterations
in the genome including breaks in chromatids and chromo-
somes, increased aneuploidy, and telomere defects, have also
been observed in LMNA‐knockout mouse embryonic fibro-
blasts [79]. A‐type lamins have also been shown to participate in
DNA repair of double‐strand breaks by non‐homologous end
joining and homology‐directed repair [79].
NUCLEAR TRANSPORT CYCLE
Molecules smaller than 40 kDa can passively diffuse across the
NPC. With higher molecular weight proteins, an orchestrated
receptor‐mediated transport is needed to facilitate crossing into
the nucleus (Figure 9.4) [1, 16, 84, 85]. This process is highly
efficient, with evidence to support a potential import rate capac-
ity of 1000 molecules per second per NPC [1, 86, 87]. For
nucleocytoplasmic transport to be carried out via the NPC, the
100 THE LIVER:  NUCLEAR TRANSPORT CYCLE
NPC itself must be dynamic and undergo structural changes,
which may occur during varying stages of transport [1, 86]. The
NPC also constitutes a selective transport system between the
nucleus and cytoplasm composed of unstructured regions com-
prising FG repeats, which also act as docking sites for protein
cargo recognition molecules [88]. These recognition molecules,
known as karyopherins (kaps), bind to cargo in the cytosol and
carry it into the nucleus; conversely, they bind cargo in the
nucleus and deliver it to the cytoplasm. Kaps recognize their
cargo by binding short amino acid sequence segments of the
protein [89]. Cargos have nuclear localization sequences (NLS)
that are classically rich in basic residues for import [1, 84, 89,
90], and nuclear export sequences (NES) that are classically leu-
cine‐rich for export [1, 84]. An NLS or NES may be masked or
modified to allow for regulation of protein localization [90].
Cargo binding occurs through either a direct interaction between
β‐kaps and the cargo, or indirect interactions through α‐kap, an
adaptor protein, and β‐kap [88].
The majority of NPC traffic is achieved by coupling kap‐
cargo binding to a cycle of GTP‐binding and hydrolysis con-
ducted through the protein Ran [91]. Ran plays a key role in
maintaining proper transport directionality, which is determined
by a concentration gradient of RanGTP [88]. Kaps deliver NLS‐
cargo and transport it through the NPC through weak transient
interactions with channel Nups that extend into the NPC.
Because it is a small monomeric G protein, Ran has a low inher-
ent rate of GDP–GTP exchange and GTP hydrolysis [16, 84, 85,
89, 92], and it relies on interacting proteins to regulate its nucle-
otide state. Binding partners are separated according to their
role in the cycle. Specifically, guanine nucleotide exchange fac-
tor (GEF) for Ran is concentrated in the nucleus while the
GTPase‐activating protein (GAP) for Ran is concentrated in the
cytoplasm. As a result, nuclear Ran is in the GTP‐bound form
while the cytoplasm contains Ran that is predominantly in the
GDP‐bound form (Figure 9.4) [85, 89]. This established differ-
ence is then coupled to import through nucleotide‐dependent
binding of Ran to the kaps, of which the GTP‐bound form binds
α‐kaps. For import, RanGTP will dissociate its cargo, releasing
it into the nucleus [92]. Additionally, RanGTP facilitates the
formation of export complexes within the nucleus [88]. The
kap–RanGTP complex can then travel back through the NPC to
the cytoplasm where RanGTP is hydrolyzed and releases the
carrier protein for another cycle with NLS‐cargo binding and
import [92]. For export, this process occurs in the converse
manner.
Exporting kaps have the ability to bind when already bound
to RanGTP [92]. GTP hydrolysis causes dissociation of the ter-
nary complex in the cytoplasm, allowing the kaps to travel back
through the NPC for another cycle of export and bind to an
NLS‐cargo for import. This process is illustrated in Figure 9.4
[84, 85, 89]. The key concept is the Ran gradient and the con-
centration of RanGTP is much higher in the nucleus. Therefore,
once the complex has entered the nucleus, the high concentra-
tion allows for the disassembly and release of the components
(Figure 9.4) [84].
Inner nuclear membrane proteins are thought to use a similar
mechanism for movement across the POM and eventual target-
ing [2, 93]. NLSs of the integral proteins signal their transport
through the cargo‐based transport system [2]. This mechanism
implies an interaction between the integral protein and the
channel Nups [93]. The kap, along with its inner membrane
protein cargo, interacts with Nups across or through scaffolding
Nup barriers. Fully understanding this inner membrane transport
represents a significant challenge. Furthermore, the targeted
transport of POM membrane could occur through this cargo‐
based system. This may require a modification of the NLS or
additional adaptor proteins to abort transport.
In higher eukaryotes, integral membrane proteins can be
targeted to the inner nuclear membrane during mitotic NE
breakdown [94]. During reformation of the NE in telophase,
integral membranes may become trapped within the inner
nuclear membrane, and thus do not fully traverse the NPC.
Targeting mechanisms vary during interphase and for organisms
that undergo a closed mitosis [94]. One model proposed to
explain the process of membrane protein transport across the
NPC is the diffusion–retention model. This model underlines
the importance of interactions between membrane proteins
and components of the nucleus in membrane protein localiza-
tion, which has been observed in multiple organisms. During
transport, transmembrane domains are retained within the pore
membrane while soluble domains cross lateral channels, regions
between the pore membrane and NPC scaffold [94–97]. It has
been shown that membrane proteins with soluble domains larger
than 60–75 kDa are unable to be transported to the inner nuclear
membrane via the NPC, a limitation attributed to the size of the
narrow lateral channel within the NPC [94, 96, 98]. Another
model, supported by evidence of diminished accumulation of a
membrane reporter in the inner nuclear membrane following
depletion of ATP, suggests that transport occurs based on
continuous structural changes of the NPC that are energy
dependent [94, 99].
It has been shown through studies using yeast proteins Heh1
and Heh2 that these proteins are targeted to the inner nuclear
membrane in a process involving transport receptors Kap60 and
Kap95. Once the complex is formed between cargo, Kap60, and
Kap95, transport across the NPC is facilitated by interactions
with channel Nups. Once inside the nucleus, the release of cargo
from transport receptors is promoted by RanGTP [94]. Three
other membrane proteins have been discovered to contain an
NLS: human Pom121 and SUN2, and C. elegans Unc‐84 [73,
94, 100–102]. However, there have been some difficulties with
these findings, such as the connection between frequency of
NLSs in these membrane proteins and their chromatin‐binding
function, the necessity of multiple targeting signals and addi-
tional factors for localization of Unc‐84 and SUN2, respectively
[94, 102, 103]. In addition, it has been shown that varying trans-
membrane proteins exhibit different responses to decreased lev-
els of ATP and Ran [94, 102].
Heh1 and Heh2 have been shown to possess NLSs with a
high affinity for Kap60, known respectively as h1NLS and
h2NLS [94, 104]. These NLSs are separated from the trans-
membrane domain by an intrinsically disordered linker sequence
comprising 180 or 235 amino acids [94]. It has been demon-
strated that h2NLS, in combination with this linker sequence,
acts as a signal for a synthetic transmembrane segment as well
as for Sec61 to accumulate within the inner nuclear membrane.
Furthermore, the length of the linker sequence is associated
with inner nuclear membrane accumulation [94, 105]. A model
 9:  Nuclear Pore Complex 101
has been proposed based on a study that captured a reporter pro-
tein, Nsp1, at the anchor domain of a channel Nup. This model
states that upon binding to Kap60/95, h2NLS will travel through
the central channel of the NPC as interactions with channel
Nups facilitated by a linker sequence of adequate length occur.
In addition, transmembrane segments will diffuse through the
pore membrane [94]. It has also been shown through work with
Sec61 fusions that membrane reporters are co-translationally
inserted in the ER membrane rather than transported through the
NPC as soluble proteins. Accumulation of these reporters at the
inner nuclear membrane associated with an NLS and linker
sequence was observed [94].
GENES, TRANSPORT, AND THE NUCLEAR
PORE COMPLEX
Proposed by Günter Blobel, the “gene gating” hypothesis pos-
tulates the idea that transcribed genes are positioned close to
NPCs, thereby streamlining mRNA export [106]. Due to this
vicinity, NPCs potentially play a role in the adjustment of
genomic structure corresponding to different stages in devel-
opment, differentiation, and the cell cycle. It is an important
assumption that the genome of higher eukaryotes is organized
into multiple, distinct three‐dimensional structures. Each of
these structures reflects a given differentiated state of the cell.
Although DNA contains the information necessary for the for-
mation of these structures, the DNA alone would not be suffi-
cient for the entirety of such a process. Thus, NPCs contribute
to the interpretation of this information and subsequent struc-
tural organization. Additionally, the nuclear lamina is pro-
posed to participate in organization of compacted regions of
the genome [106]. NPCs are envisioned as gene gating orga-
nelles capable of interacting specifically with expressing genes
[106]. All transcripts of a particular gene are proposed to exit
the nucleus via the NPC to which the gene is gated [106]. The
nonrandom distribution of the NPC may mirror the underlying
genomic organization. Given that the specific conformation of
the genome is characteristic for each differentiated state, the
distribution of NPCs should be both unique to each differenti-
ated state as well as identical for all cells of the same differen-
tiated state and cell cycle stage [106]. Furthermore, it is
proposed that up to eight genes could be positioned due to the
eightfold symmetry of the structure [106]. Gene attachment is
suggested to be mediated by a DNA‐binding subunit that is an
identical feature of all NPCs [106].
NPCs have been implicated in the sheltering of euchromatin
regions from repressing factors [107–110]. Inside the nucleus,
the nuclear basket interacts with proteins involved in trafficking
of tRNA and mRNA [111, 112]. One example is the nuclear
basket protein Nup2p, which has been shown to interact with
the histone variant H2AZ part of the euchromatin boundary for-
mation and may position euchromatin in closer proximity to the
NPC [112–114]. Furthermore, studies combining peripheral
positioning data with quantitative polymerase chain reaction
(PCR) of mRNA levels correlate with expression of some genes
coupled to NPCs [115–117]. Studies have also found that
DNA and Nups immunoprecipitated with NPC components in
genome‐wide analysis screening [118, 119]. However it is not
yet known if every gene is gated, so more extensive work is
necessary to fully prove this hypothesis [105, 109].
NUCLEAR ENVELOPE BREAKDOWN
Cell division comprises two processes – mitosis and cytokine-
sis. Mitosis is the division of the genetic material into two equal
halves, whereas cytokinesis is the process by which the cell
divides itself into two daughter cells. Mitosis is achieved by
condensed chromosomes becoming attached in a bipolar fash-
ion to the microtubule‐based mitotic spindle and subsequent
pulling apart of the sister chromatids to opposite poles of the
cell. During interphase, NE surrounds the DNA but this barrier
may not remain throughout mitosis. It is a multifaceted barrier.
Because the mitotic spindles form outside the NE in higher
eukaryotes, the whole NE must disintegrate to progress through
the cell cycle [18]. This process is termed nuclear envelope
breakdown (NEBD) [18, 120]. In some lower eukaryotes, such
as yeast, the mitotic spindles embed in the inner NE and the
process occurs within an intact envelope yielding a “closed”
mitosis, while higher eukaryotic NEBD gives an “open” mitosis
[18]. The difference between the two forms has important impli-
cations in the events triggering the processes in both.
NEBD occurs with the G2/M phase transition at the very onset
of open mitosis and is a phosphorylation‐dependent process [18,
45, 121, 122]. A pivotal feature of NEBD leading to the opening
of the NE and the blending of cellular components is disassem-
bly of the NPC. Evidence indicates the mitotic disassembly of
the NE and the NPC is driven by reversible phosphorylation of a
subset of proteins [21, 42, 45, 123, 124], Nups, lamina, and NE
membrane proteins, which may disrupt structurally significant
interactions. It has been shown that mitosis‐activated cyclin‐
dependent kinase 1 (CDK1) activity is required for keeping
NPCs dissociated during mitosis, while reassembly shows a
phosphatase dependence [125]. In addition to kinase activity, a
microtubule‐based tearing process assists NE disassembly in
cells [126, 127]. A key component in this process is dynein,
which is recruited to the NE at the beginning of prophase, and
interacts with spindle microtubules [128]. This interaction and
resulting tension placed on the NE leads to its rupture [127, 128].
Microtubule‐targeting drugs have been shown to interfere with
the onset of this effect and thereby delay NEBD [127].
The combination of phosphorylation and microtubule‐based
tearing are believed to lead to the resulting NEBD [122]. Several
kinases lead to the hyperphosphorylation of Nup98 and Nup35
that may be contributing factors to NPC disassembly [56, 62,
129]. In addition, nuclear basket protein TPR may contribute to
NPC disassembly and localization through its phosphorylation
by mitotic kinases [45]. Although a general description of the
dynamic process of NEBD is starting to emerge, very little is
known about the molecular machinery behind it. Multiple sign-
aling pathways may be involved in the final triggering. For
example, it has been shown that high levels of RanGTP affect
the dynamics of the late steps of NEBD and may play a pivotal
role in mitotic entry [120]. The full extent of molecular compo-
nents required is not fully understood.
102 THE LIVER:  NUCLEAR PORE COMPLEX AND DISEASES
NUCLEAR PORE COMPLEX
AND DISEASES
Nups have emerged as factors associated with human diseases.
It has been well documented that chromosomal translocations
involving genes encoding Nups have been associated with many
forms of leukemia, including acute myelogenous leukemia
(AML), chronic myeloid leukemia (CML), myelodysplastic
syndrome (MDS), and T‐cell acute lymphoblastic leukemia
(ALL) [7, 9, 10, 130–133].
Cancer
The involvement of the NPCs in mitosis, particularly in assem-
bly and function of structures including kinetochores, mitotic
spindles, and centrosomes as well as the role they play in proper
chromosome segregation, indicates their importance in main-
taining genome integrity. Thus, changes that affect the ability of
the NPC to function in mitosis could contribute to cancer devel-
opment [134]. It has been shown that Nup358 (also known as
RanBP2), a component of the cytoplasmic filaments of the NPC
which functions in multiple cellular processes, including mito-
sis, is implicated in the development of colon cancer, the third
most common cancer worldwide. Specifically, Nup358 pro-
motes survival of colon cancer cells by contributing to the pre-
vention of mitotic cell death [134, 135]. A mutation that
generally correlates with a less favorable prognosis and is
observed in about 8–10% of patients with colon cancer is the
BRAF (V600) mutation [134–136]. It has been shown that
Nup358 ameliorates the effects of mitotic defects present within
BRAF‐like colon cancer cell lines. Furthermore, it has been
shown that knockdown of RANBP2 leads to mitotic defects in
colon cancer cells with this mutation, eventually resulting in cell
death specifically due to prolonged mitosis [134, 135, 137].
An important feature of nucleocytoplasmic transport through
the NPC is the recognition of NLSs and NESs on large proteins
by transport receptors [9]. Proteins that contain NLSs and NESs
include oncogenes and tumor suppressors that have nuclear
functions, such as p53 and FoxO [9]. Disruption of nucleocyto-
plasmic transport involving these proteins is associated with
tumor formation [9]. These proteins bind with Crm1, an expor-
tin that has been shown to exhibit overexpression in leukemias
as well as gliomas and osteosarcomas [9]. It is believed that this
overexpression promotes excessive export of tumor suppressors
out of the nucleus, thus decreasing their function [9, 138]. In
addition, a mutation in BRCA2, a tumor suppressor whose
mutant form is associated with the development of breast, ovar-
ian, and pancreatic cancers, has been connected to disruption of
nucleocytoplasmic transport. Interaction of a mutant BRCA2
and the 26S proteasome complex subunit DSS1 has been shown
to mask the NES of BRCA2 and allow recognition by Crm1,
which mislocalizes it to the cytoplasm [9, 139].
Several types of cancer are associated with intricate mecha-
nisms known as chromothripsis and kataegis that occur in a sin-
gle event and catastrophically alter the genome [140–143].
These mechanisms may be connected to DNA damage that
occurs as a result of NE rupture. For example, chromothripsis, a
cataclysmic incidence in the genome that leads to upwards of
hundreds of rearranged DNA fragments, is highly correlated
with NE rupture in micronuclei and chromatin bridges [141,
144, 145]. Such rearrangements may lead to circularization of
DNA, and these circular fragments may become amplified if
they carry oncogenes [141, 146].
Micronuclei and cancer
Micronuclei are small nuclei that form separately from the
­primary nucleus following the formation of an NE around mis‐
segregated chromosomes [147]. Some micronuclei remain
intact upon formation, but there is also evidence that a substan-
tial proportion of micronuclei experience NE rupture, which is
less likely to be repaired than NE rupture that occurs in primary
nuclei. This subsequently may lead to improper functioning of
the micronuclei as well as extensive DNA damage, which has
been observed in both cultured cancer cells as well as sections
of non‐small cell lung cancer (NSCLS) tumors [147]. Disrupted
micronuclei have also been observed in non‐transformed cell
lines, which indicates that micronuclei could potentially be use-
ful for detection in early cancer development [147–149]. There
is also evidence to show that defects are present within intact
micronuclei as well, including greater susceptibility to DNA
damage [147–149]. Disruption of micronuclei is correlated with
the development of some autoinflammatory diseases due to the
activation of cGAS, a DNA sensor, which is proposed to accu-
mulate on chromatin‐containing micronuclei following NE rup-
ture [141, 150–152].
Micronuclear disruption, specifically loss of compartmentali-
zation, does not appear to be caused by loss of the NPC [147]. It
has been shown that levels of core Nups detected by mAb414
did not greatly differ between intact and disrupted micronuclei.
However, there is evidence that Nup153 and TPR levels (both
basket Nups) are lower in disrupted micronuclei, although this
evidence is not conclusive regarding the connection between
loss of compartmentalization in micronuclei and stability of the
NPC [147].
Nup98
Nup98 regulates transcription of genes that have functions relat-
ing to development and the cell cycle [9, 54, 55]. Chromosomal
translocations involving Nup98 also alter expression of Nup96,
as both proteins are encoded by the same mRNA [9, 27].
Disrupted expression of Nup96 may serve as an additional con-
tributing factor to disease phenotypes that arise from Nup98
translocations with transcription factors, as Nup96 plays a role
in the regulation of export of mRNAs that are associated with
immunity and cell cycle regulation [9, 49, 153]. Nup98 translo-
cations are most frequently observed in AML, chronic myeloid
leukemia in blast crisis (CML‐bc), and myelodysplastic syn-
drome (MDS) [154]. Several mechanisms have been proposed
to explain the role of Nup98 fusion proteins in leukemic trans-
formations, including upregulation of HOXA genes, suppressed
differentiation, and increased self‐renewal [154, 155]. A study
focused on Nup98–HOXA9 revealed a biphasic effect of the
fusion protein on the growth of CD34+ hematopoietic cells, with
growth initially inhibited before continuous, long‐term prolif-
eration of primitive cells was observed [156]. This finding
 9:  Nuclear Pore Complex 103
suggests the development of AML from MDS, which has been
shown in transgenic mouse models containing the fusion protein
Nup98–HOXD13. In addition, Nup98–HOXA9 has been shown
to suppress hematopoietic differentiation and to increase primi-
tive self‐renewing cells [156].
Nup98 has been implicated in infection by influenza virus.
Specifically, downregulation of Nup98 by the virus nonstruc-
tural protein 1 (NS1) is associated with increased viral replica-
tion [9, 157]. This is correlated with evidence that viruses target
nuclear mRNA export in order to enhance replication, which
includes the export of mRNAs that encode antiviral factors. For
example, interaction occurs between Nup98 and the mRNA
export factor Rae1, which is targeted by the VSV matrix protein.
This inhibition of mRNA export has been shown to be reversible
through upregulation of Rae1 and Nup96–Nup98, which may
be achieved by treatment with interferons [157]. It has been
demonstrated that upon infection with the influenza virus in
293T and MDCK cells, Nup98 is actively degraded. This poten-
tially acts as a contributing factor to the suppression of nuclear
mRNA export [157]. It has also been shown that Nup98 is tar-
geted for degradation in cells infected with poliovirus, which is
likely facilitated by a viral 2A protease. Poliovirus additionally
targets two other Nups, Nup153 and Nup62, but cleavage of
Nup98 appears to occur more rapidly [158].
Triple A syndrome
Triple A or Allgrove syndrome, an adrenal insufficiency (acha-
lasia–addisonianism–alacrima), is a rare autosomal recessive
congenital disorder with three common features of achalasia,
Addison disease, and alacrima [159, 160]. In most cases, there
is no family history of inception. Triple A is caused by muta-
tions in the gene AAAS, encoding the Nup designated as
ALADIN [7, 161, 162]. The ALADIN nucleoporin is only
found within higher eukaryotes. Loss of ALADIN integration
into the NPC results in the disease state [163]. DNA repair pro-
tein transport may be diminished if ALADIN is absent. Lack of
DNA repair leads to instability and eventual cell death, espe-
cially in nerve cells [164]. POM protein NDC1 helps to target
ALADIN to the NPC [165–167]. Mutations in ALADIN affect
interactions and targeting, leading to changes in selective
nuclear protein import. Depletion of NDC1 results in mislocali-
zation of ALADIN and may be involved in the pathogenesis of
the disease [165–167].
Premature and accelerated aging
Function of the NPC has also been shown to have a role in aging
and age‐related deterioration. Deterioration occurs by means of
molecular damage that builds over time, and Nups have among
the greatest longevity of proteins within the mammalian brain
[46, 168, 169]. Nup damage is associated with increased nuclear
permeability, which poses the risk of infiltration by toxins and
cytoplasmic proteins [46, 170]. In some studies, leakage of the
cytoplasmic protein MAP2 has been observed within in vitro
models of Huntington’s disease (HD). Since aging is a signifi-
cant risk factor for neurodegeneration, this also demonstrates
the role of the NPC in the development of neurodegenerative
disorders [46]. A molecular study that observed changes in
aging brains of rats revealed that scaffold Nups have very slow
turnover rates compared to peripheral Nups, and additionally
showed that relative Nup composition of NPCs changed
throughout the aging process [155, 168–170]. This finding sug-
gests that slow Nup turnover rate may contribute to the accumu-
lation of damaged NPCs over time [155, 168].
A very rare and fatal premature aging disease called
Hutchinson–Gilford progeria syndrome has been associated
with the malformation of the protein lamin A [171–173].
Specifically, single‐base mutations of the LMNA gene lead to
activation of a cryptic splice site and subsequent farnesylation
of lamin A, producing a variant of the protein known as progerin
[172, 173]. Continuous farnesylation allows insertion of prog-
erin into the inner nuclear membrane, in which it accumulates,
inflicting damage upon aging cells [173, 174].
Huntington’s disease
HD, which is the most common heritable neurodegenerative
disorder, has been associated with mislocalization and aggrega-
tion of several Nups, as well as disruption of nucleocytoplasmic
transport [46]. HD is a member of a group of neurodegenerative
disorders known as polyQ diseases, which are all characterized
by repetitive CAG sequences that encode a long polyglutamine
(polyQ) tract within corresponding proteins [46, 175]. Nup
aggregates have been shown to colocalize with the mutant form
of the huntingtin protein, Htt. In addition, both the number and
size of these aggregates increase with pathological progression
of HD [46].
Disruption of nucleocytoplasmic transport as a contributing
factor of HD has been evidenced by disturbance of the Ran gra-
dient, which provides power for active transport and maintains
proper transport directionality through the interaction of the
protein Ran‐GTP with the transport receptor during nuclear
import, so cargo can be released. This gradient is maintained by
a GTPase‐activating protein located on the cytoplasmic fila-
ments of the NPC, RanGAP1 [46, 176]. Interactions between
RanGAP1 and RNAs that contain a hexanucleotide repeat
expansion (HRE) are linked to disruption of the Ran gradient.
These HRE mutations are correlated with some forms of HD
and are also commonly seen in amyotrophic lateral sclerosis
(ALS) and frontotemporal dementia (FTD) [177–180]. In addi-
tion, RanGAP1 and Nup88 have been shown to form aggregates
in a mouse model containing HD, and Nup62 has demonstrated
severe mislocalization in HD striatum tissue [46].
Primary biliary cirrhosis
Primary biliary cirrhosis (PBC) is an autoimmune disease of
the liver [8, 10]. Production of autoantibodies signals a loss of
tolerance and tissue damage [10]. In PBC, the gradual destruc-
tion of the bile ducts leads to the onset of cirrhosis of the liver.
Patients with PBC generate a panel of autoantibodies [10, 181,
182] against mitochondrial antigens (AMAs), and nuclear pro-
teins called antinuclear antibodies (ANAs) [10]. These ANAs
are directed against the NPC [183, 184] and generate a nuclear
rim staining by immunofluorescence. ANAs against Pom210
and Nup62 appear to be highly specific for PBC [10].
Combined with AMA detection, α‐gp210 and α‐Nup62
104 THE LIVER:  REFERENCES
antibodies may provide additional diagnostic and prognostic
tools [182, 185, 186]. Anti‐Pom210 ANAs in PBC specifically
recognize a 15 amino acid domain within the C‐terminus
[187]. It has been observed that increased ANA against
Pom210 is associated with more severe disease and a worse
prognosis [186], including greater vulnerability for progres-
sion to end‐stage liver disease [181, 188–190]. The persistence
of expression of AMAs and ANAs after liver transplants have
been described [10, 181]. In some patients, increased ANA
Pom210 levels persist years after transplantation. The question
remains as to the clinical relevance of anti‐NPC antibodies in
PBC and the likelihood that an autoimmune response may
ascend as a product of molecular mimicry [10, 181].
Understanding the organization of the membranous structure
of the NE and NPC is critical for defining the specificity of
ANAs in PBC. But it is not yet known how mechanistically
these Nups are involved in the disease state [7, 10, 185].
OUTLOOK
With advances in identifying the structural connections of the
NPC, we have gained valuable insights into the transporter. As
more puzzle pieces are fitted into place, we expect to link the
NPC with its evolution and significance in abnormal states of
nuclear function.
REFERENCES
1. Tran, E.J. and Wente, S.R. Dynamic nuclear pore complexes: life on the
edge. Cell, 2006;125(6):1041–53.
2. Lusk, C.P., Blobel, G., and King, M.C. Highway to the inner nuclear mem-
brane: rules for the road. Nat Rev Mol Cell Biol, 2007;8(5):414–20.
3. Weis, K. Regulating access to the genome: nucleocytoplasmic transport
throughout the cell cycle. Cell, 2003;112(4):441–51.
4. Rout, M.P., Aitchison, J.D., Suprapto, A. et al. The yeast nuclear pore com-
plex: composition, architecture, and transport mechanism. J Cell Biol,
2000;148(4):635–51.
5. Cronshaw, J.M., Krutchinsky, A.N., Zhang, W., Chait, B.T., and Matunis,
M.J. Proteomic analysis of the mammalian nuclear pore complex. J Cell
Biol, 2002;158(5):915–27.
6. DeGrasse, J.A., DuBois, K.N., Devos, D. et al. Evidence for a shared nuclear
pore complex architecture that is conserved from the last common eukary-
otic ancestor. Mol Cell Proteomics, 2009;8(9):2119–30.
7. Cronshaw, J.M. and Matunis, M.J. The nuclear pore complex: disease associa-
tions and functional correlations. Trends Endocrinol Metab, 2004;15(1):34–9.
8. Giorgini, A., Selmi, C., Invernizzi, P. et al. Primary biliary cirrhosis: solving
the enigma. Ann N Y Acad Sci, 2005;1051:185–93.
9. Mor, A., White, M.A., and Fontoura, B.M. Nuclear trafficking in health and
disease. Curr Opin Cell Biol, 2014;28:28–35.
10. Duarte‐Rey, C., Bogdanos, D., Yang, C.Y. et al. Primary biliary cirrhosis and
the nuclear pore complex. Autoimmun Rev, 2012;11(12):898–902.
11. Bui, K.H., von Appen, A., Diguilio, A.L. et al. Integrated structural analysis
of the human nuclear pore complex scaffold. Cell, 2013;155(6):1233–43.
12. Kosinski, J., Mosalaganti, S., von Appen, A. et al. Molecular architecture of
the inner ring scaffold of the human nuclear pore complex. Science,
2016;352(6283):363–5.
13. von Appen, A., Kosinski, J., Sparks, L. et al. In situ structural analysis of the
human nuclear pore complex. Nature, 2015;526(7571):140–3.
14. Hoelz, A., Glavy, J.S., and Beck, M. Toward the atomic structure of the
nuclear pore complex: when top down meets bottom up. Nat Struct Mol Biol,
2016;23(7):624–30.
15. Schwartz, T.U. Modularity within the architecture of the nuclear pore com-
plex. Curr Opin Struct Biol, 2005;15(2):221–6.
16. Lim, R.Y., Aebi, U., and Fahrenkrog, B. Towards reconciling structure and
function in the nuclear pore complex. Histochem Cell Biol, 2008;
129(2):105–16.
17. Kabachinski, G. and Schwartz, T.U. The nuclear pore complex – structure
and function at a glance. J Cell Sci, 2015;128(3):423–9.
18. Hetzer, M.W., Walther, T.C., and Mattaj, I.W. Pushing the envelope: struc-
ture, function, and dynamics of the nuclear periphery. Annu Rev Cell Dev
Biol, 2005;21:347–80.
19. Belgareh, N., Rabut, G., Bai, S.W. et al. An evolutionarily conserved NPC
subcomplex, which redistributes in part to kinetochores in mammalian cells.
J Cell Biol, 2001;154(6):1147–60.
20. Walther, T.C., Alves, A., Pickersgill, H. et al. The conserved Nup107–160
complex is critical for nuclear pore complex assembly. Cell, 2003;
113(2):195–206.
21. Glavy, J.S., Krutchinsky, A.N., Cristea, I.M. et  al. Cell‐cycle‐dependent
phosphorylation of the nuclear pore Nup107–160 subcomplex. Proc Natl
Acad Sci U S A, 2007;104(10):3811–16.
22. Aaronson, R.P. and Blobel G. On the attachment of the nuclear pore com-
plex. J Cell Biol, 1974;62(3):746–54.
23. Beck, M., Lucic, V., Forster, F., Baumeister, W., and Medalia O. Snapshots
of nuclear pore complexes in action captured by cryo‐electron tomography.
Nature, 2007;449(7162):611–15.
24. Beck, M., Forster, F., Ecke, M. et  al. Nuclear pore complex structure and
dynamics revealed by cryoelectron tomography. Science, 2004;306(5700):
1387–90.
25. Alber, F., Dokudovskaya, S., Veenhoff, L.M. et al. The molecular architecture
of the nuclear pore complex. Nature, 2007;450(7170):695–701.
26. Alber, F., Dokudovskaya, S., Veenhoff, L.M. et al. Determining the architectures
of macromolecular assemblies. Nature, 2007;450(7170):683–94.
27. Fontoura, B.M., Blobel, G., and Matunis, M.J. A conserved biogenesis path-
way for nucleoporins: proteolytic processing of a 186‐kilodalton precursor
generates Nup98 and the novel nucleoporin, Nup96. J Cell Biol,
1999;144(6):1097–112.
28. Loiodice, I., Alves, A., Rabut, G. et  al. The entire Nup107–160 complex,
including three new members, is targeted as one entity to kinetochores in
mitosis. Mol Biol Cell, 2004;15(7):3333–44.
29. Cristea, I.M., Williams, R., Chait, B.T., and Rout, M.P. Fluorescent proteins
as proteomic probes. Mol Cell Proteomics, 2005;4(12):1933–41.
30. Ori, A., Banterle, N., Iskar, M. et al. Cell type‐specific nuclear pores: a case
in point for context‐dependent stoichiometry of molecular machines. Mol
Syst Biol, 2013;9:648.
31. Boehmer, T., Enninga, J., Dales, S., Blobel, G., and Zhong, H. Depletion of a
single nucleoporin, Nup107, prevents the assembly of a subset of nucleoporins
into the nuclear pore complex. Proc Natl Acad Sci U S A, 2003;100(3):981–5.
32. Boehmer, T., Jeudy, S., Berke, I.C., and Schwartz, T.U. Structural and func-
tional studies of Nup107/Nup133 interaction and its implications for the
architecture of the nuclear pore complex. Mol Cell, 2008;30(6):721–31.
33. Devos, D., Dokudovskaya, S., Alber, F. et al. Components of coated vesicles
and nuclear pore complexes share a common molecular architecture. PLoS
Biol, 2004;2(12):e380.
34. Drin, G., Casella, J.F., Gautier, R. et al. A general amphipathic alpha‐helical
motif for sensing membrane curvature. Nat Struct Mol Biol, 2007;14(2):
138–46.
35. Lutzmann, M., Kunze, R., Buerer, A., Aebi, U., and Hurt, E. Modular self‐
assembly of a Y‐shaped multiprotein complex from seven nucleoporins.
EMBO J, 2002;21(3):387–97.
36. Berke, I.C., Boehmer, T., Blobel, G., and Schwartz, T.U. Structural and func-
tional analysis of Nup133 domains reveals modular building blocks of the
nuclear pore complex. J Cell Biol, 2004;167(4):591–7.
37. Hsia, K.C., Stavropoulos, P., Blobel, G., and Hoelz A. Architecture of a coat
for the nuclear pore membrane. Cell, 2007;131(7):1313–26.
38. Brohawn, S.G., Leksa, N.C., Spear, E.D., Rajashankar, K.R., and Schwartz,
T.U. Structural evidence for common ancestry of the nuclear pore complex
and vesicle coats. Science, 2008;322(5906):1369–73.
39. Stuwe, T., Correia, A.R., Lin, D.H. et al. Nuclear pores. Architecture of the
nuclear pore complex coat. Science, 2015;347(6226):1148–52.
40. Hawryluk‐Gara, L.A., Shibuya, E.K., and Wozniak, R.W. Vertebrate Nup53
interacts with the nuclear lamina and is required for the assembly of a
Nup93‐containing complex. Mol Biol Cell, 2005;16(5):2382–94.
 9:  Nuclear Pore Complex 105
41. Jeudy, S. and Schwartz, T.U. Crystal structure of nucleoporin Nic96 reveals
a novel, intricate helical domain architecture. J Biol Chem, 2007;
282(48):34904–12.
42. Antonin, W., Ellenberg, J., and Dultz E. Nuclear pore complex assembly
through the cell cycle: regulation and membrane organization. FEBS Lett,
2008;582(14):2004–16.
43. Melcak, I., Hoelz, A., and Blobel G. Structure of Nup58/45 suggests flexible
nuclear pore diameter by intermolecular sliding. Science, 2007;315(5819):
1729–32.
44. Rajanala, K. and Nandicoori, V.K. Localization of nucleoporin Tpr to the
nuclear pore complex is essential for Tpr mediated regulation of the export
of unspliced RNA. PLoS One, 2012;7(1):e29921.
45. Rajanala, K., Sarkar, A., Jhingan, G.D. et al. Phosphorylation of nucleoporin
Tpr governs its differential localization and is required for its mitotic func-
tion. J Cell Sci, 2014;127(Pt 16):3505–20.
46. Grima, J.C., Daigle, J.G., Arbez N et  al. Mutant Huntingtin disrupts the
nuclear pore complex. Neuron, 2017;94(1):93–107 e6.
47. Napetschnig, J., Blobel, G., and Hoelz A. Crystal structure of the N‐terminal
domain of the human protooncogene Nup214/CAN. Proc Natl Acad Sci U S
A, 2007;104(6):1783–8.
48. Tamura, K., Fukao, Y., Iwamoto, M., Haraguchi, T., and Hara‐Nishimura I.
Identification and characterization of nuclear pore complex components in
Arabidopsis thaliana. Plant Cell, 2010;22(12):4084–97.
49. Chakraborty, P., Wang, Y., Wei, J.H. et al. Nucleoporin levels regulate cell
cycle progression and phase‐specific gene expression. Dev Cell, 2008;
15(5):657–67.
50. Chakraborty, P., Seemann, J., Mishra, R.K. et al. Vesicular stomatitis virus
inhibits mitotic progression and triggers cell death. EMBO Rep, 2009;10(10):
1154–60.
51. Liang, Y., Franks, T.M., Marchetto, M.C., Gage, F.H., and Hetzer, M.W.
Dynamic association of NUP98 with the human genome. PLoS Genet,
2013;9(2):e1003308.
52. Franks, T.M., McCloskey, A., Shokirev, M.N. et al. Nup98 recruits the Wdr82‐
Set1A/COMPASS complex to promoters to regulate H3K4 trimethylation in
hematopoietic progenitor cells. Genes Dev, 2017;31(22):2222–34.
53. Rosenblum, J.S. and Blobel G. Autoproteolysis in nucleoporin biogenesis.
Proc Natl Acad Sci U S A, 1999;96(20):11370–5.
54. Kalverda, B., Pickersgill, H., Shloma, V.V., and Fornerod, M. Nucleoporins
directly stimulate expression of developmental and cell‐cycle genes inside
the nucleoplasm. Cell, 2010;140(3):360–71.
55. Capelson, M., Liang, Y., Schulte, R. et  al. Chromatin‐bound nuclear pore
components regulate gene expression in higher eukaryotes. Cell, 2010;140(3):
372–83.
56. Laurell, E., Beck, K., Krupina, K. et al. Phosphorylation of Nup98 by multi-
ple kinases is crucial for NPC disassembly during mitotic entry. Cell,
2011;144(4):539–50.
57. Laurell, E. and Kutay U. Dismantling the NPC permeability barrier at the
onset of mitosis. Cell Cycle, 2011;10(14).
58. Fontoura, B.M., Blobel, G., and Yaseen, N.R. The nucleoporin Nup98 is a
site for GDP/GTP exchange on ran and termination of karyopherin beta 2‐
mediated nuclear import. J Biol Chem, 2000;275(40):31289–96.
59. Enninga, J., Levy, D.E., Blobel, G., and Fontoura, B.M. Role of nucleoporin
induction in releasing an mRNA nuclear export block. Science, 2002;295(5559):
1523–5.
60. Jeganathan, K.B., Malureanu, L., and van Deursen, J.M. The Rae1‐Nup98
complex prevents aneuploidy by inhibiting securin degradation. Nature,
2005;438(7070):1036–9.
61. Singer, S., Zhao, R., Barsotti, A.M. et al. Nuclear pore component Nup98 is
a potential tumor suppressor and regulates posttranscriptional expression of
select p53 target genes. Mol Cell, 2012;48(5):799–810.
62. Linder, M.I., Kohler, M., Boersema, P. et al. Mitotic disassembly of nuclear
pore complexes involves CDK1‐ and PLK1‐mediated phosphorylation of
key interconnecting nucleoporins. Dev Cell, 2017;43(2):141–56 e7.
63. Kupke, T., Malsam, J., and Schiebel E. A ternary membrane protein complex
anchors the spindle pole body in the nuclear envelope in budding yeast. J
Biol Chem, 2017;292(20):8447–58.
64. Mansfeld, J., Guttinger, S., Hawryluk‐Gara, L.A. et al. The conserved trans-
membrane nucleoporin NDC1 is required for nuclear pore complex assem-
bly in vertebrate cells. Mol Cell, 2006;22(1):93–103.
65. Antonin, W. and Mattaj, I.W. Nuclear pore complexes: round the bend? Nat
Cell Biol, 2005;7(1):10–12.
66. Chen, J., Smoyer, C.J., Slaughter, B.D., Unruh, J.R., and Jaspersen, S.L. The
SUN protein Mps3 controls Ndc1 distribution and function on the nuclear
membrane. J Cell Biol, 2014;204(4):523–39.
67. Mitchell, J.M., Mansfeld, J., Capitanio, J., Kutay, U., and Wozniak, R.W.
Pom121 links two essential subcomplexes of the nuclear pore complex core
to the membrane. J Cell Biol, 2010;191(3):505–21.
68. Onischenko, E., Stanton, L.H., Madrid, A.S., Kieselbach, T., and Weis K.
Role of the Ndc1 interaction network in yeast nuclear pore complex assem-
bly and maintenance. J Cell Biol, 2009;185(3):475–91.
69. Stavru, F., Hulsmann, B.B., Spang, A. et al. NDC1: a crucial membrane‐inte-
gral nucleoporin of metazoan nuclear pore complexes. J Cell Biol,
2006;173(4):509–19.
70. Franz, C., Askjaer, P., Antonin, W. et al. Nup155 regulates nuclear envelope
and nuclear pore complex formation in nematodes and vertebrates. EMBO J,
2005;24(20):3519–31.
71. Doucet, C.M. and Hetzer, M.W. Nuclear pore biogenesis into an intact
nuclear envelope. Chromosoma, 2010;119(5):469–77.
72. Shaulov, L., Gruber, R., Cohen, I., and Harel A. A dominant‐negative form of
POM121 binds chromatin and disrupts the two separate modes of nuclear
pore assembly. J Cell Sci, 2011;124(Pt 22):3822–34.
73. Funakoshi, T., Clever, M., Watanabe, A., and Imamoto, N. Localization of
Pom121 to the inner nuclear membrane is required for an early step of inter-
phase nuclear pore complex assembly. Mol Biol Cell, 2011;22(7):1058–69.
74. Funakoshi, T., Maeshima, K., Yahata, K. et al. Two distinct human POM121
genes: requirement for the formation of nuclear pore complexes. FEBS Lett,
2007;581(25):4910–16.
75. Chadrin, A., Hess, B., San Roman, M. et al. Pom33, a novel transmembrane
nucleoporin required for proper nuclear pore complex distribution. J Cell
Biol, 2010;189(5):795–811.
76. Floch, A.G., Tareste, D., Fuchs, P.F. et al. Nuclear pore targeting of the yeast
Pom33 nucleoporin depends on karyopherin and lipid binding. J Cell Sci,
2015;128(2):305–16.
77. Galy, V., Antonin, W., Jaedicke, A. et al. A role for gp210 in mitotic nuclear‐
envelope breakdown. J Cell Sci, 2008;121(Pt 3):317–28.
78. Dittmer, T.A. and Misteli T. The lamin protein family. Genome Biol,
2011;12(5):222.
79. Graziano, S., Kreienkamp, R., Coll‐Bonfill, N., and Gonzalo S. Causes and
consequences of genomic instability in laminopathies: replication stress and
interferon response. Nucleus, 2018;9(1):258–75.
80. Stewart, C.L. and Roux, K.J., Burke B. Blurring the boundary: the nuclear
envelope extends its reach. Science, 2007;318(5855):1408–12.
81. Broers, J.L., Peeters, E.A., Kuijpers, H.J. et al. Decreased mechanical stiff-
ness in LMNA‐/‐ cells is caused by defective nucleo‐cytoskeletal integrity:
implications for the development of laminopathies. Hum Mol Genet,
2004;13(21):2567–80.
82. Crisp, M., Liu, Q., Roux, K. et al. Coupling of the nucleus and cytoplasm:
role of the LINC complex. J Cell Biol, 2006;172(1):41–53.
83. Manju, K., Muralikrishna, B., and Parnaik, V.K. Expression of disease‐caus-
ing lamin A mutants impairs the formation of DNA repair foci. J Cell Sci,
2006;119(Pt 13):2704–14.
84. Hoelz, A. and Blobel G. Cell biology: popping out of the nucleus. Nature,
2004;432(7019):815–16.
85. Weis K. Nucleocytoplasmic transport: cargo trafficking across the border.
Curr Opin Cell Biol, 2002;14(3):328–35.
86. Ribbeck, K. and Gorlich D. Kinetic analysis of translocation through nuclear
pore complexes. EMBO J, 2001;20(6):1320–30.
87. Yang, W., Gelles, J., and Musser, S.M. Imaging of single‐molecule transloca-
tion through nuclear pore complexes. Proc Natl Acad Sci U S A, 2004;
101(35):12887–92.
88. Hoelz, A., Debler, E.W., and Blobel G. The structure of the nuclear pore
complex. Annu Rev Biochem, 2011;80:613–43.
89. Stewart, M. Molecular mechanism of the nuclear protein import cycle. Nat
Rev Mol Cell Biol, 2007;8(3):195–208.
90. Lange, A., Mills, R.E., Lange, C.J. et al. Classical nuclear localization sig-
nals: definition, function, and interaction with importin alpha. J Biol Chem,
2007;282(8):5101–5.
91. Moore, M.S. and Blobel G. The GTP‐binding protein Ran/TC4 is
required for protein import into the nucleus. Nature, 1993;365(6447):
661–3.
92. Gorlich, D. and Kutay U. Transport between the cell nucleus and the
cytoplasm. Annu Rev Cell Dev Biol, 1999;15:607–60.
106 THE LIVER:  REFERENCES
93. King, M.C., Lusk, C.P., and Blobel G. Karyopherin‐mediated import of
integral inner nuclear membrane proteins. Nature, 2006;442(7106):1003–7.
94. Meinema, A.C., Poolman, B., and Veenhoff, L.M. The transport of integral
membrane proteins across the nuclear pore complex. Nucleus, 2012;3(4): 322–9.
95. Ellenberg, J., Siggia, E.D., Moreira, J.E. et al. Nuclear membrane dynamics
and reassembly in living cells: targeting of an inner nuclear membrane
protein in interphase and mitosis. J Cell Biol, 1997;138(6):1193–206.
96. Worman, H.J. and Courvalin, J.C. The inner nuclear membrane. J Membr
Biol, 2000;177(1):1–11.
97. Wu, W., Lin, F., and Worman, H.J. Intracellular trafficking of MAN1, an
integral protein of the nuclear envelope inner membrane. J Cell Sci,
2002;115(Pt 7):1361–71.
98. Soullam, B. and Worman, H.J. Signals and structural features involved in
integral membrane protein targeting to the inner nuclear membrane. J Cell
Biol, 1995;130(1):15–27.
99. Ohba, T., Schirmer, E.C., Nishimoto, T., and Gerace L. Energy‐ and tempera-
ture‐dependent transport of integral proteins to the inner nuclear membrane
via the nuclear pore. J Cell Biol, 2004;167(6):1051–62.
100. Yavuz, S., Santarella‐Mellwig, R., Koch, B. et  al. NLS‐mediated NPC
functions of the nucleoporin Pom121. FEBS Lett. 2010;584(15):3292–8.
101. Tapley, E.C., Ly, N., and Starr, D.A. Multiple mechanisms actively target
the SUN protein UNC‐84 to the inner nuclear membrane. Mol Biol Cell,
2011;22(10):1739–52.
102. Turgay, Y., Ungricht, R., Rothballer, A. et al. A classical NLS and the SUN
domain contribute to the targeting of SUN2 to the inner nuclear membrane.
EMBO J, 2010;29(14):2262–75.
103. Antonin, W., Ungricht, R., and Kutay U. Traversing the NPC along the
pore  membrane: targeting of membrane proteins to the INM. Nucleus,
2011;2(2):87–91.
104. Meinema, A.C., Laba, J.K., Hapsari, R.A. et  al. Long unfolded linkers
facilitate membrane protein import through the nuclear pore complex.
Science, 2011;333(6038):90–3.
105. Liu, D., Wu, X., Summers, M.D., Lee, A., Ryan, K.J., and Braunagel, S.C.
Truncated isoforms of Kap60 facilitate trafficking of Heh2 to the nuclear
envelope. Traffic, 2010;11(12):1506–18.
106. Blobel G. Gene gating: a hypothesis. Proc Natl Acad Sci U S A,
1985;82(24):8527–9.
107. Schneider, R. and Grosschedl, R. Dynamics and interplay of nuclear architecture,
genome organization, and gene expression. Genes Dev, 2007;21(23):3027–43.
108. Taddei A. Active genes at the nuclear pore complex. Curr Opin Cell Biol,
2007;19(3):305–10.
109. Lanctot, C., Cheutin, T., Cremer, M., Cavalli, G., and Cremer, T. Dynamic
genome architecture in the nuclear space: regulation of gene expression in
three dimensions. Nat Rev Genet, 2007;8(2):104–15.
110. Menon, B.B., Sarma, N.J., Pasula, S. et al. Reverse recruitment: the Nup84
nuclear pore subcomplex mediates Rap1/Gcr1/Gcr2 transcriptional activa-
tion. Proc Natl Acad Sci U S A, 2005;102(16):5749–54.
111. Dieppois, G., Iglesias, N., and Stutz F. Cotranscriptional recruitment to the
mRNA export receptor Mex67p contributes to nuclear pore anchoring of
activated genes. Mol Cell Biol, 2006;26(21):7858–70.
112. Ishii, K., Arib, G., Lin, C., Van Houwe, G., and Laemmli, U.K. Chromatin
boundaries in budding yeast: the nuclear pore connection. Cell,
2002;109(5):551–62.
113. Dilworth, D.J., Tackett, A.J., Rogers, R.S. et  al. The mobile nucleoporin
Nup2p and chromatin‐bound Prp20p function in endogenous NPC‐medi-
ated transcriptional control. J Cell Biol, 2005;171(6):955–65.
114. Marshall, I.C. and Wilson, K.L. Nuclear envelope assembly after mitosis.
Trends Cell Biol, 1997;7(2):69–74.
115. Casolari, J.M., Brown, C.R., Drubin, D.A., Rando, O.J., and Silver, P.A.
Developmentally induced changes in transcriptional program alter spatial
organization across chromosomes. Genes Dev, 2005;19(10):1188–98.
116. Taddei, A., Van Houwe, G., Hediger, F. et  al. Nuclear pore association
­confers optimal expression levels for an inducible yeast gene. Nature,
2006;441(7094):774–8.
117. Brickner, J.H. and Walter, P. Gene recruitment of the activated IN01 locus
to the nuclear membrane. PLoS Biol, 2004;2(11):e342.
118. Casolari, J.M., Brown, C.R., Komili, S. et al. Genome‐wide localization of
the nuclear transport machinery couples transcriptional status and nuclear
organization. Cell, 2004;117(4):427–39.
119. Casolari, J.M. and Silver, P.A. Guardian at the gate: preventing unspliced
pre‐mRNA export. Trends Cell Biol, 2004;14(5):222–5.
120. Kutay, U. and Hetzer, M.W. Reorganization of the nuclear envelope during
open mitosis. Curr Opin Cell Biol, 2008;20(6):669–77.
121. Dultz, E., Zanin, E., Wurzenberger, C. et al. Systematic kinetic analysis of
mitotic dis‐ and reassembly of the nuclear pore in living cells. J Cell Biol,
2008;180(5):857–65.
122. Champion, L., Linder, M.I., and Kutay, U. Cellular reorganization during
mitotic entry. Trends Cell Biol, 2017;27(1):26–41.
123. De Souza, C.P. and Osmani, S.A. Mitosis, not just open or closed. Eukaryot
Cell, 2007;6(9):1521–7.
124. Bodoor, K., Shaikh, S., Salina, D. et  al. Sequential recruitment of NPC
proteins to the nuclear periphery at the end of mitosis. J Cell Sci,
1999;112(Pt 13):2253–64.
125. Onischenko, E.A., Gubanova, N.V., Kiseleva, E.V., and Hallberg, E. Cdk1
and okadaic acid‐sensitive phosphatases control assembly of nuclear pore
complexes in Drosophila embryos. Mol Biol Cell, 2005;16(11):5152–62.
126. Salina, D., Bodoor, K., Eckley, D.M. et al. Cytoplasmic dynein as a facilita-
tor of nuclear envelope breakdown. Cell, 2002;108(1):97–107.
127. Beaudouin, J., Gerlich, D., Daigle, N., Eils, R., and Ellenberg J. Nuclear
envelope breakdown proceeds by microtubule‐induced tearing of the lam-
ina. Cell, 2002;108(1):83–96.
128. Busson, S., Dujardin, D., Moreau, A., Dompierre, J., and De Mey, J.R.
Dynein and dynactin are localized to astral microtubules and at cortical
sites in mitotic epithelial cells. Curr Biol, 1998;8(9):541–4.
129. Blethrow, J.D., Glavy, J.S., Morgan, D.O., and Shokat, K.M. Covalent cap-
ture of kinase‐specific phosphopeptides reveals Cdk1‐cyclin B substrates.
Proc Natl Acad Sci U S A, 2008;105(5):1442–7.
130. Romana, S.P., Radford‐Weiss, I., Ben Abdelali, R. et al. NUP98 rearrange-
ments in hematopoietic malignancies: a study of the Groupe Francophone
de Cytogenetique Hematologique. Leukemia, 2006;20(4):696–706.
131. Reader, J.C., Meekins, J.S., Gojo, I., and Ning Y. A novel NUP98‐PHF23
fusion resulting from a cryptic translocation t(11;17)(p15;p13) in acute
myeloid leukemia. Leukemia, 2007;21(4):842–4.
132. Panagopoulos, I., Kerndrup, G., Carlsen, N. et al. Fusion of NUP98 and the
SET binding protein 1 (SETBP1) gene in a paediatric acute T cell lymphoblas-
tic leukaemia with t(11;18)(p15;q12). Br J Haematol, 2007;136(2):294–6.
133. Suzuki, A., Ito, Y., Sashida, G. et al. t(7;11)(p15;p15) Chronic myeloid leu-
kaemia developed into blastic transformation showing a novel NUP98/
HOXA11 fusion. Br J Haematol, 2002;116(1):170–2.
134. Wong, R.W. and D’Angelo M. Linking nucleoporins, mitosis, and colon
cancer. Cell Chem Biol, 2016;23(5):537–9.
135. Vecchione, L., Gambino, V., Raaijmakers, J. et  al. A vulnerability of a
­subset of colon cancers with potential clinical utility. Cell, 2016;165(2):
317–30.
136. Popovici, V., Budinska, E., Tejpar, S. et al. Identification of a poor‐progno-
sis BRAF‐mutant‐like population of patients with colon cancer. J Clin
Oncol, 2012;30(12):1288–95.
137. Hashizume, C., Kobayashi, A., and Wong, R.W. Down‐modulation of
nucleoporin RanBP2/Nup358 impaired chromosomal alignment and
induced mitotic catastrophe. Cell Death Dis, 2013;4:e854.
138. Falini, B., Mecucci, C., Tiacci, E. et  al. Cytoplasmic nucleophosmin in
acute myelogenous leukemia with a normal karyotype. N Engl J Med,
2005;352(3):254–66.
139. Jeyasekharan, A.D., Liu, Y., Hattori, H. et al. A cancer‐associated BRCA2
mutation reveals masked nuclear export signals controlling localization.
Nat Struct Mol Biol, 2013;20(10):1191–8.
140. Maciejowski, J. and Imielinski M. Modeling cancer rearrangement land-
scapes. Curr Opin Syst Biol, 2017;1:54–61.
141. Hatch, E.M. Nuclear envelope rupture: little holes, big openings. Curr Opin
Cell Biol, 2018;52:66–72.
142. Alexandrov, L.B., Nik‐Zainal, S., Wedge, D.C. et al. Signatures of muta-
tional processes in human cancer. Nature, 2013;500(7463):415–21.
143. Roberts, S.A. and Gordenin, D.A. Hypermutation in human cancer
genomes: footprints and mechanisms. Nat Rev Cancer, 2014;14(12):
786–800.
144. Zhang, C.Z., Spektor, A., Cornils, H. et al. Chromothripsis from DNA dam-
age in micronuclei. Nature, 2015;522(7555):179–84.
145. Maciejowski, J., Li, Y., Bosco, N., Campbell, P.J., and de Lange, T.
Chromothripsis and kataegis induced by telomere crisis. Cell,
2015;163(7):1641–54.
146. Korbel, J.O. and Campbell, P.J. Criteria for inference of chromothripsis in
cancer genomes. Cell, 2013;152(6):1226–36.
 9:  Nuclear Pore Complex 107
147. Hatch, E.M., Fischer, A.H., Deerinck, T.J., and Hetzer, M.W. Catastrophic
nuclear envelope collapse in cancer cell micronuclei. Cell, 2013;154(1):
47–60.
148. Crasta, K., Ganem, N.J., Dagher, R. et al. DNA breaks and chromosome
pulverization from errors in mitosis. Nature, 2012;482(7383):53–8.
149. Terradas, M., Martin, M., Hernandez, L., Tusell, L., and Genesca A.
Nuclear envelope defects impede a proper response to micronuclear DNA
lesions. Mutat Res, 2012;729(1–2):35–40.
150. Harding, S.M., Benci, J.L., Irianto, J. et al. Mitotic progression following
DNA damage enables pattern recognition within micronuclei. Nature,
2017;548(7668):466–70.
151. Mackenzie, K.J., Carroll, P., Martin, C.A. et  al. cGAS surveillance of
micronuclei links genome instability to innate immunity. Nature,
2017;548(7668):461–5.
152. Dou, Z., Ghosh, K., Vizioli, M.G. et  al. Cytoplasmic chromatin triggers
inflammation in senescence and cancer. Nature, 2017;550(7676):402–6.
153. Faria, A.M., Levay, A., Wang, Y. et al. The nucleoporin Nup96 is required
for proper expression of interferon‐regulated proteins and functions.
Immunity, 2006;24(3):295–304.
154. Gough, S.M., Slape, C.I., and Aplan, P.D. NUP98 gene fusions and hemat-
opoietic malignancies: common themes and new biologic insights. Blood,
2011;118(24):6247–57.
155. Sakuma, S. and D’Angelo, M.A. The roles of the nuclear pore complex in cel-
lular dysfunction, aging and disease. Semin Cell Dev Biol, 2017;68 :72–84.
156. Takeda, A., Goolsby, C., and Yaseen, N.R. NUP98‐HOXA9 induces long‐term
proliferation and blocks differentiation of primary human CD34+ hemat-
opoietic cells. Cancer Res, 2006;66(13):6628–37.
157. Satterly, N., Tsai, P.L., van Deursen, J. et  al. Influenza virus targets the
mRNA export machinery and the nuclear pore complex. Proc Natl Acad Sci
U S A, 2007;104(6):1853–8.
158. Park, N., Katikaneni, P., Skern, T., and Gustin, K.E. Differential targeting
of nuclear pore complex proteins in poliovirus‐infected cells. J Virol,
2008;82(4):1647–55.
159. Allgrove, J., Clayden, G.S., Grant, D.B., and Macaulay, J.C. Familial glu-
cocorticoid deficiency with achalasia of the cardia and deficient tear pro-
duction. Lancet, 1978;1(8077):1284–6.
160. Sandrini, F., Farmakidis, C., Kirschner, L.S. et al. Spectrum of mutations of
the AAAS gene in Allgrove syndrome: lack of mutations in six kindreds
with isolated resistance to corticotropin. J Clin Endocrinol Metab,
2001;86(11):5433–7.
161. Krumbholz, M., Koehler, K., and Huebner A. Cellular localization of 17 natural
mutant variants of ALADIN protein in triple A syndrome – shedding light on
an unexpected splice mutation. Biochem Cell Biol, 2006;84(2):243–9.
162. Mukhopadhya, A., Danda, S., Huebner, A., and Chacko A. Mutations of the
AAAS gene in an Indian family with Allgrove’s syndrome. World J
Gastroenterol, 2006;12(29):4764–6.
163. Dusek, T., Korsic, M., Koehler, K. et  al. A novel AAAS gene mutation
(p.R194X) in a patient with triple A syndrome. Horm Res, 2006;65(4):
171–6.
164. Zimmer, V., Liebe, R., and Lammert F. Nuclear receptor variants in liver
disease. Dig Dis, 2015;33(3):415–19.
165. Cho, A.R., Yang, K.J., Bae, Y. et al. Tissue‐specific expression and subcel-
lular localization of ALADIN, the absence of which causes human triple A
syndrome. Exp Mol Med, 2009;41(6):381–6.
166. Kind, B., Koehler, K., Lorenz, M., and Huebner A. The nuclear pore complex
protein ALADIN is anchored via NDC1 but not via POM121 and GP210 in
the nuclear envelope. Biochem Biophys Res Commun, 2009;390(2):205–10.
167. Yamazumi, Y., Kamiya, A., Nishida, A. et al. The transmembrane nucleop-
orin NDC1 is required for targeting of ALADIN to nuclear pore complexes.
Biochem Biophys Res Commun, 2009;389(1):100–4.
168. Savas, J.N., Toyama, B.H., Xu, T., Yates, J.R., 3rd, and Hetzer, M.W.
Extremely long‐lived nuclear pore proteins in the rat brain. Science,
2012;335(6071):942.
169. Toyama, B.H., Savas, J.N., Park, S.K. et al. Identification of long‐lived pro-
teins reveals exceptional stability of essential cellular structures. Cell,
2013;154(5):971–82.
170. D’Angelo, M.A., Raices, M., Panowski, S.H., and Hetzer, M.W. Age‐
dependent deterioration of nuclear pore complexes causes a loss of nuclear
integrity in postmitotic cells. Cell, 2009;136(2):284–95.
171. De Sandre‐Giovannoli, A., Bernard, R., Cau, P. et al. Lamin a truncation in
Hutchinson‐Gilford progeria. Science, 2003;300(5628):2055.
172. Eriksson, M., Brown, W.T., Gordon, L.B. et  al. Recurrent de novo point
mutations in lamin A cause Hutchinson‐Gilford progeria syndrome. Nature,
2003;423(6937):293–8.
173. Gordon, L.B., Shappell, H., Massaro, J. et  al. Association of Lonafarnib
treatment vs no treatment with mortality rate in patients with Hutchinson‐
Gilford progeria syndrome. JAMA, 2018;319(16):1687–95.
174. Casasola, A., Scalzo, D., Nandakumar, V. et  al. Prelamin A processing,
accumulation and distribution in normal cells and laminopathy disorders.
Nucleus, 2016;7(1):84–102.
175. Finkbeiner, S. Huntington’s disease. Cold Spring Harb Perspect Biol,
2011;3(6).
176. Floch, A.G., Palancade, B., and Doye V. Fifty years of nuclear pores and
nucleocytoplasmic transport studies: multiple tools revealing complex
rules. Methods Cell Biol, 2014;122:1–40.
177. Freibaum, B.D., Lu, Y., Lopez‐Gonzalez, R. et al. GGGGCC repeat expan-
sion in C90rf72 compromises nucleocytoplasmic transport. Nature,
2015;525(7567):129–33.
178. Jovicic, A., Mertens, J., Boeynaems, S. et al. Modifiers of C90rf72 dipep-
tide repeat toxicity connect nucleocytoplasmic transport defects to FTD/
ALS. Nat Neurosci, 2015;18(9):1226–9.
179. Hensman Moss, D.J., Poulter, M., Beck, J. et al. C90rf72 expansions are the
most common genetic cause of Huntington disease phenocopies. Neurology,
2014;82(4):292–9.
180. Zhang, K., Donnelly, C.J., Haeusler, A.R. et  al. The C90rf72 repeat
expansion disrupts nucleocytoplasmic transport. Nature, 2015;
525(7567):56–61.
181. Nakamura, M., Shimizu‐Yoshida, Y., Takii, Y. et  al. Antibody titer to
gp210‐C terminal peptide as a clinical parameter for monitoring primary
biliary cirrhosis. J Hepatol, 2005;42(3):386–92.
182. Invernizzi, P., Selmi, C., Ranftler, C., Podda, M., and Wesierska‐Gadek J.
Antinuclear antibodies in primary biliary cirrhosis. Semin Liver Dis,
2005;25(3):298–310.
183. Worman, H.J. and Courvalin, J.C. Antinuclear antibodies specific for pri-
mary biliary cirrhosis. Autoimmun Rev, 2003;2(4):211–17.
184. Invernizzi, P., Podda, M., Battezzati, P.M. et  al. Autoantibodies against
nuclear pore complexes are associated with more active and severe liver
disease in primary biliary cirrhosis. J Hepatol, 2001;34(3):366–72.
185. Worman, H.J. Nuclear envelope protein autoantigens in primary biliary cir-
rhosis. Hepatol Res, 2007;37 Suppl 3:S406–11.
186. Sfakianaki, O., Koulentaki, M., Tzardi, M. et  al. Peri‐nuclear antibodies
correlate with survival in Greek primary biliary cirrhosis patients. World J
Gastroenterol, 2010;16(39):4938–43.
187. Nickowitz, R.E. and Worman, H.J. Autoantibodies from patients with pri-
mary biliary cirrhosis recognize a restricted region within the cytoplasmic
tail of nuclear pore membrane glycoprotein Gp210. J Exp Med,
1993;178(6):2237–42.
188. Itoh, S., Ichida, T., Yoshida, T. et al. Autoantibodies against a 210 kDa glyco-
protein of the nuclear pore complex as a prognostic marker in patients with
primary biliary cirrhosis. J Gastroenterol Hepatol, 1998;13(3): 257–65.
189. Bogdanos, D.P., Liaskos, C., Pares, A. et al. Anti‐gp210 antibody mirrors
disease severity in primary biliary cirrhosis. Hepatology, 2007;45(6):1583;
author reply ‐4.
190. Wesierska‐Gadek, J., Penner, E., Battezzati, P.M. et al. Correlation of initial
autoantibody profile and clinical outcome in primary biliary cirrhosis.
Hepatology, 2006;43(5):1135–44.
 Protein Maturation
10 and Processing at the
Endoplasmic Reticulum
Ramanujan S. Hegde
MRC Laboratory of Molecular Biology, Cambridge, UK
INTRODUCTION
The liver, and hepatocytes isolated from it, have long played a cen-
tral role in our understanding of cellular organization, protein traf-
ficking, and secretion. The abundance, accessibility, high secretory
capacity, and wide range of functions made the liver an ideal choice
for many of the earliest and seminal studies of cellular function.
The first electron microscopic studies revealed a remarkable com-
partmentalization to hepatocytes, including the highly abundant
membrane‐bound organelles of the endoplasmic reticulum (ER),
Golgi, mitochondria, peroxisomes, and others. Upon the develop-
ment of subcellular fractionation methods, each of these organelles
could be isolated and studied by classical biochemistry and
­enzymology [1]. The isolation of liver and hepatocyte membrane
fractions enriched in “rough” ER [2] was instrumental in the dis-
covery that ER‐bound ribosomes synthesize secretory and mem-
brane proteins [3]. These proteins were found to be selectively
imported into the ER or inserted into the ER membrane. Indeed,
the ER eventually proved to be the major site of secretory and
membrane protein biosynthesis and maturation. Twenty years later,
the ER was discovered to also be a major site of quality control, the
process whereby immature, defective, or otherwise incompletely
assembled proteins are triaged for selective degradation [4]. This
chapter will cover the basic principles of secretory and membrane
protein maturation and quality control in the ER.
THE SEGREGATION OF SECRETORY
AND MEMBRANE PROTEINS TO THE ER
Essentially all secretory and membrane proteins destined for
the cell surface, extracellular environment, Golgi, lysosomes, and
endosomes begin their biosynthesis at the ER. Secretory proteins
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
are completely translocated across the ER membrane into the
luminal space, while membrane proteins are woven into the ER
membrane, with specific regions exposed to the cytosol and other
portions exposed to the lumen. After achieving their correct
topology (i.e. orientation relative to the membrane), these pro-
teins undergo a series of maturation events in the ER to produce
a functional protein. This involves folding, various modifications,
and assembly with other proteins or cofactors. Only upon suc-
cessful maturation are the secretory and membrane proteins
allowed to exit the ER for their final destinations via vesicular
trafficking pathways. Thus, the ER can be considered an intracel-
lular factory for the biosynthesis and maturation of secretory and
membrane proteins. The key steps in the assembly line of this
factory are the selective targeting of secretory and membrane
proteins to the ER, their translocation across or into the ER mem-
brane, their folding into the correct conformation, and assembly
into a functional product.
Secretory and membrane proteins are
recognized via signal sequences
Cellular protein synthesis occurs on ribosomes located in the
cytosol, but over half of all proteins eventually reside in non‐
cytosolic locales within the highly compartmentalized eukary-
otic cell. Thus, it is obvious that cells must effectively sort and
segregate proteins among numerous intracellular organelles.
One of the major conceptual advances in understanding this
process was the articulation of a cogent hypothesis for the basis
of intracellular protein segregation. The “signal hypothesis”
postulated that specific regions of a newly synthesized polypep-
tide (i.e. signal sequences) provide unique codes that specify the
eventual destination for that protein [5]. This concept, originally
developed to explain how secretory proteins are selectively
 10:  Protein Maturation and Processing at the Endoplasmic Reticulum 109
Protein
Prolactin
TGF-β2
Growth hormone
N-terminal signals
Osteopontin
Leptin
Cleaved
Apo-A1
EGF-receptor
Glucagon
lnhibinβ
BiP
Choriogonadotropin
Rhodopsin
Transmembrane
CFTR
domains
ASGR I
Synaptotagmin II
Glycophorin C
Mannosidase II
Figure 10.1  Examples of signal sequences used for protein targeting to the endoplasmic reticulum. In many cases, the signal sequence is located
at the N‐terminus and is cleaved by the enzyme signal peptidase shortly after targeting has been accomplished. In other cases, the first transmem-
brane domain is used for targeting. These are not removed, and are part of the final protein product. In all cases, the defining feature is a stretch
of ~10 or more hydrophobic residues (underlined sequence).
segregated to the ER, eventually proved to be widely generaliz-
able. The signals for targeting to the ER, nucleus, mitochondria,
peroxisomes, and other locations have been identified. In each
case, the sorting signals are selectively recognized by dedicated
machinery that mediates correct targeting to the intended
destination.
For secretory and membrane proteins, the distinguishing
feature of their signal sequences essential for selective
­recognition is hydrophobicity (Figure  10.1). In many mem-
brane proteins, the first transmembrane domain (TMD), which
necessarily must be hydrophobic in order to eventually reside
stably in the lipid bilayer, serves as the signal sequence for
selective recognition. However, secretory proteins are soluble
in their mature form, and therefore do not typically contain
long stretches of hydrophobic residues in their final primary
sequence. Instead, they are usually made as precursors contain-
ing a cleavable signal sequence at the N‐terminus. The cleava-
ble signal is typically ~15–40 amino acids long and contains a
central hydrophobic region essential for its selective recogni-
tion and targeting to the ER. This signal is removed at the ER
by an enzyme called signal peptidase, and is therefore not part
of the final mature protein. Some membrane proteins also use a
cleavable N‐terminal signal for targeting to the ER. Thus, all
secretory and membrane proteins are recognized via a hydro-
phobic motif that sets them apart from other non‐ER targeted
proteins [6].
Multiple pathways of protein targeting
to the ER membrane
Of the ~20 000 proteins encoded in the human genome, ~7000
are targeted to the ER [7]. These proteins are highly diverse in
their biophysical characteristics, number and location of their
Targeting sequence
MNIKGSPWKGSLLLLLVSNLLLCQSVAP
MHYCVLSAFLILHLVTVAL. ..
MATGSRTSLLLAFGLLCLPWLQEGSA...
MRIAVICFCLLGITCA...
MHWGTLCGFLWLWPYLFYVQA...
MKAAVLTLAVLFLTGSQA...
MRPSGTAGAALLALLAALCPRA...
MKSIYFVAGLFVMLVQG...
MPLLWLRGFLLASCWllVRSSPTPGS...
MKLSLVAAMLLLLSAARA...
MEMFQGLLLLLLLSMGGTWA...
...WQFSMLAAYMFLLIVLGFPINFLTLYVTVQH...
...FFWRFMFYGIFLYLGEVTKAVQPLLLGRllA...
...KFRLSLLALAFNILLLVVICVVSSQSMQLQK...
...EINKIPLPPWALIAMAVVAGLLLLTCCFCIC...
...TIMDIVVIAGVIAAVAIVLVSLLFVMLRYMYR...
... RRRFALVICSGCLLVFLSLYllLNFAAPAATQ...
TMDs, and topology with respect to the ER membrane. This
diversity means that no single mechanism can recognize, target,
translocate, and insert all of these proteins. Hence, cells
have evolved different pathways to deal with different types of
proteins.
There are two general mechanisms of protein targeting to the
ER. In co‐translational targeting, proteins are recognized as
they are being translated (Figure 10.2a). Thus, the entire ribo-
some is targeted to the ER, where the nascent protein is trans-
located across or inserted into the membrane during its
synthesis. This mechanism has the advantage that an unfolded
polypeptide can be moved across or inserted into the mem-
brane before it can fold in the cytosol. In post‐translational
targeting, proteins are synthesized completely in the cytosol
before their delivery to the ER membrane for translocation or
insertion. This mechanism generally operates on proteins that
do not have large domains that need to be translocated across
the membrane.
In eukaryotic cells, the majority of proteins use co‐transla-
tional targeting [8]. However, some secretory proteins are very
small (less than ~100 amino acids) and their synthesis occurs
too quickly to be effectively recognized during translation.
These proteins, which include various hormones, cytokines, and
antibacterial peptides, are instead post‐translationally targeted
and translocated into the ER. Similarly, some membrane pro-
teins contain a single TMD located close to the C‐terminus [9].
This TMD serves as the signal for targeting such “tail‐anchored”
membrane proteins to the ER. Because a C‐terminal TMD
will  not emerge from the ribosome until after the protein has
been fully synthesized, these proteins are also targeted post‐­
translationally. The pathways for co‐translational and post‐
translational targeting involve different factors, and will be
considered in successive sections.
110 THE LIVER:  THE SEGREGATION OF SECRETORY AND MEMBRANE PROTEINS TO THE ER

(a)

Secretory
SRP- protein
receptor Translocon Membrane
protein

ER lumen

SRP Cytosol

Signal
sequence
Translocation Membrane
Targeting integration
Ribosome

Signal
recognition
mRNA
Initiation of
synthesis

(b) Translocation
Sec61
complex Signal Nascent
Plug
sequence protein ER membrane

Membrane
insertion

Lateral gate
Ribosome Ribosome

Inactive translocon Opened by a signal Active translocon

Figure 10.2  Segregation of secretory and membrane proteins to the endoplasmic reticulum (ER). (a) The key steps of co‐translational transloca-
tion including: signal sequence recognition by the signal recognition particle (SRP), targeting to the ER membrane via SRP receptor (SR), and
translocation through a translocon. Membrane proteins utilize the same basic steps and machinery, but undergo integration into the membrane rather
than complete translocation. (b) The Sec61 complex forms a protein‐conducting channel in the ER membrane. In the inactive state (left), a seam in
the channel (the lateral gate) is closed, and the pore is occluded by a plug. A signal sequence can open the Sec61 complex by intercalating into the
lateral gate (middle). In the active state (right), the channel can move polypeptides in two directions: across the membrane (translocation) or into
the lipid bilayer (membrane insertion).

Signal recognition particle mediates Selective ER targeting of SRP‐bound ribosomes is medi-

ated by the presence in the ER membrane of a specific recep-
co‐translational protein targeting tor for SRP. The interaction between SRP and SRP receptor
Nearly all N‐terminal signals and TMDs are recognized co‐­ (SR) not only mediates targeting, but also the transfer of the
translationally as they first emerge from the ribosome. A large, nascent polypeptide to a translocon formed by the Sec61
ribosome‐associated factor termed signal recognition particle complex (described in the next section). This unidirectional
(SRP) interacts directly with the hydrophobic signal sequence transfer of the nascent chain from SRP to the translocon is
[10]. Co‐translational SRP‐mediated recognition has three impor- energy dependent. Both SRP and SR are GTPases. SRP binds
tant consequences. First, SRP shields the hydrophobic TMD to a signal sequence in its GTP form, while SR binds to the
or signal from the bulk aqueous environment to prevent inappro- translocon in its GTP‐bound form. When the GTP‐bound
priate interactions or aggregation. In this sense, SRP can be forms of SRP and SR interact with each other, they hydrolyze
viewed as a chaperone of sorts that protects nascent secretory and their GTPs. This induces a series of conformational changes
membrane proteins during the earliest stages of their synthesis. that ultimately results in SRP dissociating from the signal
Second, the SRP–signal interaction modulates ribosome function sequence, SR, and the ribosome. The ribosome and the
to slow the rate of translation. This transient slowing provides ­liberated signal sequence now engage the translocon. In this
extra time for targeting to the ER membrane. And third, SRP is way,  nascent proteins containing a signal sequence or TMD
a  key distinguishing feature that allows some ribosomes to be are selectively delivered to translocons at the ER where they
targeted to the ER, while others remain in the cytosol. complete their synthesis.
 10:  Protein Maturation and Processing at the Endoplasmic Reticulum 111
Sec61 complex forms a translocation channel
across the ER
After targeting to the ER translocon, secretory proteins are
translocated through a channel across the membrane, while
membrane proteins are inserted into the lipid bilayer (discussed
in the next section). Both of these processes are mediated by a
translocation channel that can open perpendicularly across the
membrane for translocation and laterally toward the membrane
for insertion [11, 12]. This translocation channel is formed by
the three‐protein Sec61 complex.
The structure of the Sec61 complex [13, 14] shows it to be a
toroidal channel with a central hourglass‐shaped pore and a
seam along the side (Figure 10.2b, left). When the Sec61 com-
plex is quiescent, the pore is plugged and the front of the seam,
known as the lateral gate, is closed. Thus, ions and other small
molecules cannot leak into or out of the ER.
After SRP targets a translating ribosome to the translocon, the
Sec61 complex binds tightly to a site that is located adjacent to
the ribosomal tunnel through which the newly made protein
emerges. This positions the Sec61 complex such that its channel
is roughly co‐linear with the ribosomal tunnel. At this point, the
signal sequence of a secretory protein still needs to open
the  Sec61 channel to initiate translocation. This occurs when
the  signal sequence becomes embedded into the lateral gate
(Figure 10.2b, middle). This only occurs for bona fide signals,
thereby providing a proofreading mechanism that prevents pro-
teins lacking a signal sequence from opening the channel.
Embedding the signal into the lateral gate of Sec61 widens
the channel [15], which leads to displacement of the plug that
normally occupies the central pore. Thus, recognition of a signal
or TMD causes the Sec61 complex to open and makes it active
for translocation and membrane insertion (Figure 10.2b, right).
In this mode, the Sec61 complex is able to translocate proteins
across the membrane through the central channel, or insert
membrane proteins into the lipid bilayer via the lateral gate.
In the case of secretory proteins, the segment of polypeptide
downstream from the signal sequence enters the channel and
accesses the lumen of the ER. Continued translation by the
bound ribosome expels the growing polypeptide into the lumen,
where molecular chaperones (discussed later) bind to the
­nascent protein and ensure its translocation. Once translocation
has been initiated, the signal sequence is removed by an enzyme
called signal peptidase. Translocation continues until translation
is finished, at which point the last bit of the nascent protein
enters the ER lumen, the ribosome dissociates from the Sec61
complex, and the channel again returns to its quiescent state.
Co‐translational membrane protein insertion
by the Sec61 complex
If the protein targeted to the Sec61 translocon contains TMDs,
it needs to be inserted into the ER membrane [16]. The lateral
gate of the Sec61 complex provides the key point of transition
between the aqueous cytosol and hydrophobic lipid bilayer. As
a TMD emerges from the ribosome, it exploits this gate to enter
the membrane (Figure 10.2b, right).
The orientation (or topology) in which a TMD is inserted is
influenced by several parameters. If a protein has been targeted
by a signal sequence and a large soluble domain has been
­translocated into the lumen, then the topology has already been
committed. When a TMD then emerges from the ribosome in
this type of protein, it will insert with its N‐terminal‐flanking
domain in the lumen and C‐terminal‐flanking domain facing the
cytosol.
For proteins that are targeted via their first TMD, the determi-
nation of topology is more complex [17]. Here, the length of the
polypeptide preceding the TMD, the charged residues flanking
the TMD, the TMD length, and its overall hydrophobicity all
influence its final topology. In general, the N‐terminus is favored
to face the cytosol if it is long and contains basic residues adja-
cent to the TMD. Otherwise, the TMD will be oriented with the
N‐terminus facing the ER lumen.
For proteins containing multiple TMDs, each TMD after the
first takes on the opposite orientation of the TMD that precedes
it. This means that for multipass membrane proteins, topology
of the first TMD essentially fixes the topology of the remaining
ones [18]. Each TMD is thought to enter the membrane via the
lateral gate of Sec61 in succession. There may be accessory pro-
teins that facilitate TMD insertion, and there may be complex
membrane proteins whose insertion is not simply vectorial [16].
These aspects of how a multipass membrane protein is inserted
remain to be investigated in detail.
Post‐translational targeting of secretory
and membrane proteins
Proteins whose sole targeting signal is close to the C‐terminus
cannot be targeted co‐translationally [9]. The reason is that the
ribosomal tunnel through which nascent proteins emerge
houses ~40 amino acids. This means that at the very minimum,
~40 residues are needed beyond the targeting signal for it to be
recognized during synthesis. However, it probably takes
~5–10 seconds to carry out the reactions involved in co‐
translational targeting. At a translation speed of 6 residues per
second, this means that another ~30–60 residues are needed
after the targeting signal is exposed for co‐translational target-
ing to occur successfully. Thus, proteins whose targeting sig-
nal is located within the last ~90–120 residues must be targeted
post‐translationally.
In the case of secretory proteins, the hydrophobic N‐terminal
signal sequence must be shielded to prevent its aggregation,
while the remainder of the polypeptide must be prevented from
folding. The abundant cytosolic protein calmodulin is thought to
bind the signal sequence dynamically to prevent aggregation,
while still allowing opportunities for it to engage a translocon at
the ER [19]. Folding of the remainder of the protein is probably
prevented by chaperones, typically of the Hsp70 family, that
release the protein prior to its translocation.
The translocon for post‐translational translocation contains
the Sec61 complex together with the accessory proteins Sec62
and Sec63 [20]. This translocon can recognize the signal
sequences of post‐translationally translocated secretory proteins
by a mechanism that is probably similar to that described above
for co‐translational proteins. This recognition leads to the initial
insertion of the small secretory protein into the Sec61 translo-
cation channel, exposing a portion of the polypeptide to the
lumen. The ER luminal chaperone BiP, recruited to the site of
112 THE LIVER:  FOLDING AND MATURATION OF SECRETORY AND MEMBRANE PROTEINS
translocation by its interaction with Sec63, probably interacts
with this exposed segment of polypeptide to favor its transloca-
tion into the ER lumen [21].
Post‐translationally inserted membrane proteins use an
entirely different set of factors for targeting and insertion [22].
The targeting factor for such proteins is a cytosolic factor called
TRC40 (known as Get3 in yeast) that directly recognizes, binds,
and shields TMDs [23]. TRC40 has a specific receptor at the ER
membrane. This receptor dislodges the TMD from TRC40 and
facilitates its insertion into the membrane. TRC40 then returns
to the cytosol for another round of targeting. Some proteins have
TMDs that are not hydrophobic enough to bind TRC40. These
are kept soluble in the cytosol by calmodulin and inserted into
the ER by another complex called EMC (for ER membrane pro-
tein complex) [24]. EMC probably has additional functions dur-
ing co‐translational membrane protein insertion, but these are
poorly understood.
Mis‐targeted proteins are degraded by the cell
Despite the extensive machinery dedicated to protein targeting,
failures occur. A secretory or membrane protein that does not
reach the ER poses a threat to the cytosol. Such mis‐localized
proteins are prone to aggregation, are unlikely to fold in the
cytosolic environment, and can make inappropriate interactions
with other proteins. These adverse consequences are ­underscored
by the finding that inherited mutations in a signal sequence can
cause disease in humans [25] and transgenic mice [26]. Under
normal circumstances, such consequences are avoided by qual-
ity control pathways that recognize mis‐localized proteins and
promptly route them for degradation [27].
The cytosol contains a family of quality‐control factors that
have the capability of recognizing exposed signal sequences and
TMDs [28, 29]. When protein targeting works properly, these
hydrophobic regions are either removed (in the case of signal
sequences) or buried in the membrane (in the case of TMDs).
Thus, their exposure in the cytosol is a cue that targeting or
membrane insertion has failed. The quality‐control factors that
recognize these exposed targeting elements recruit ubiquitin
ligases that mark the mis‐localized protein with multiple ubiq-
uitins. The ubiquitinated protein is then delivered to the protea-
some for degradation. In this manner, the cytosol is kept free of
proteins that belong in other compartments of the cell.
FOLDING AND MATURATION
OF SECRETORY AND MEMBRANE
PROTEINS
Correct ER targeting and translocation of a secretory or mem-
brane protein are only the first steps of its eventual matura-
tion into a functional product. Many polypeptides are
processed, modified, and assembled with cofactors before
they leave the ER. Furthermore, the linear polypeptide must
be folded into a stable three‐dimensional structure, and in
many cases, assembled with other proteins. Each protein has
its own unique requirements; some are highly specialized in
their maturation, while others use ubiquitous machinery.
Although the mechanisms that coordinate the multistep matu-
ration events of complex proteins remain poorly understood,
many of the individual steps have been studied in detail and
are discussed here.
Numerous co‐ and post‐translational
modifications accompany protein maturation
Most secretory and membrane proteins are covalently modified
during their maturation. The most common modification is
asparagine‐linked (or N‐linked) glycosylation [30]. This usually
occurs co‐translationally, as a protein is entering the ER lumen
through the translocon. The enzyme that mediates glycosyla-
tion, oligosaccharyl transferase (OST), is tightly associated with
the translocon [31]. As certain sequences (Asn‐Xxx‐Ser or Asn‐
Xxx‐Thr, where Xxx is any amino acid except proline) pass
through the translocon into the ER lumen, OST transfers a fully
assembled 14‐sugar core glycan to the asparagine residue
(Figure 10.3, left).
Not all consensus sites are necessarily glycosylated, and
glycosylation is not uniformly efficient in all cell types or under
all conditions. Although the reasons for this heterogeneity are
poorly understood, it is likely to influence the protein’s function
and/or trafficking in subtle ways that are important in a tissue‐
specific manner. These subtleties notwithstanding, glycosyla-
tion as a whole is critical to the proper folding and maturation of
a wide range of proteins, and failure or even partial deficiency
of glycosylation can be disastrous for a cell. Indeed, impaired
processing of glycans causes a variety of severe inherited dis-
eases in humans [32].
Another frequent modification is signal sequence cleavage.
As mentioned above, N‐terminal signal sequences are tempo-
rary sequence elements that are used for correct targeting and
translocation of secretory and membrane proteins, but which
are not part of the final mature product. They are removed
after they have completed their function; this removal usually
happens co‐translationally at the translocon. A proteolytic
enzyme called signal peptidase interacts with the translocon in
such a manner that its active site is positioned adjacent to the
opening of the translocation channel in the ER lumen [33]. As
a nascent polypeptide begins translocation, the boundary
between a signal sequence and mature domain becomes
exposed to signal peptidase, which efficiently cuts the poly-
peptide at a precise position.
The third widely used modification is attachment of a glyco-
sylphosphatidylinositol (GPI) anchor to some proteins at their
C‐terminus [34]. GPI anchors are essentially phospholipid mol-
ecules containing a glycan (i.e. a glycolipid). Attachment of the
GPI anchor to an otherwise soluble protein therefore tethers it to
the membrane. This allows GPI‐anchored proteins to reside on the
cell surface within certain microdomains that are favored by the
lipids to which they are tethered. Furthermore, GPI anchors can
be cleaved by extracellular phospholipases, allowing for regu-
lated release (i.e. secretion) of certain proteins under specific
conditions [35].
GPI anchor addition occurs in the ER and is mediated by a
GPI–transamidase complex. The choice of which proteins will
be modified by the transamidase is determined by the presence
at the very C‐terminus of a hydrophobic “signal sequence” for
 10:  Protein Maturation and Processing at the Endoplasmic Reticulum 113

ER
export

UGGT

Calreticulin
Folded
ERP57 protein

Core N-linked Folding

glycan cycle
EDEM and
Man I
OST gluc. I
Misfolded
protein
gluc. II
Glucosidase II

ER lumen

Cytosol

Degradation

Figure 10.3  Glycan‐dependent protein folding by lectin chaperones in the ER. Many nascent proteins entering the endoplasmic reticulum (ER)
are modified co‐translationally with a 14‐sugar core glycan containing two N‐acetyl glucosamines (squares), nine mannoses (circles), and three
glucoses (triangles). The glucoses are rapidly trimmed by glucosidases (gluc. I and II), allowing the mono‐glucosylated glycan to recruit a lectin
chaperone (calreticulin), along with associated factors (such as the protein disulfide isomerase‐like molecule ERP57). Protein folding is mediated
by repeated cycles of binding and release from calreticulin, regulated by removal and re‐addition of the terminal glucose. The enzyme UGGT only
adds the glucose to non‐native proteins, ensuring that calreticulin only re‐binds if folding has not been achieved. If folding fails despite repeated
cycles of calreticulin binding, the mannoses get further trimmed (by mannosidases such as EDEM), and the glycan now becomes a target for bind-
ing by another lectin that routes the protein for degradation (see Figure 10.4). The same general scheme is used by other chaperones: they are
recruited to substrates early in their biosynthesis, and undergo cycles of binding and release until the polypeptide either achieves its correct folded
state, or is transferred to the degradation pathway. OST, oligosaccharyl transferase.

GPI anchor addition. The transamidase recognizes this signal,
and proteolytically removes it simultaneously with addition of
the GPI anchor (in a transamidation reaction).
Beyond these relatively universal ER‐specific modifications,
numerous substrate‐specific and cell type‐specific m
­ odifications
have been described. For example, proline and lysine residues in
collagen are hydroxylated, some secreted signaling molecules
and morphogens are lipid‐modified (e.g. with cholesterol), and
there are other proteolytic processing events that have been
described in the ER and other compartements of the secretory
pathway. In each case, the modifications are typically critical
for either the protein’s maturation, stability, and/or eventual
function.
Chaperones assist in protein maturation
Molecular chaperones are essential cellular factors that partici-
pate in the maturation of nearly every protein [36]. Chaperones
assist in the folding and assembly of nascent polypeptides by
direct binding, thereby preventing inappropriate interactions
with other proteins in the highly crowded cellular environment.
Of course, a nascent protein cannot fold when bound to a chap-
erone. Thus, chaperones periodically release the nascent protein
and transiently afford it an opportunity to fold. If folding is
unsuccessful, the chaperone quickly re‐binds. By such repeated
(and usually energy‐dependent) cycles of binding and release,
chaperones can facilitate protein folding before inappropriate
interactions lead to nonproductive aggregation.
Essential to chaperone function is their ability to selectively
bind immature, but not properly folded, conformations of a nas-
cent protein. This selectivity is thought to be achieved in several
ways, but typically involves the recognition of polypeptide seg-
ments that should normally be shielded in the final folded struc-
ture. For example, the buried core of a folded soluble protein
usually contains hydrophobic parts of the polypeptide. Exposure
of such hydrophobic patches, which necessarily means the pro-
tein is not properly folded, represents a common recognition
motif for chaperone binding. Similarly, hydrophilic residues
within some TMDs become buried when multiple TMDs
are assembled correctly. Exposure of such residues within the
hydrophobic lipid bilayer can therefore signify a non‐native
conformation. In the same way, exposed cysteines that
114 THE LIVER:  FOLDING AND MATURATION OF SECRETORY AND MEMBRANE PROTEINS
eventually should be disulfide bonded with other cysteines, can
recruit certain types of chaperones. By recognizing these and
other sequence elements, chaperones can distinguish between
non‐native and native folded states.
Chaperones are recruited almost immediately after a nascent
secretory or membrane protein begins translocation. This is
because proteins are translocated in an unfolded conformation,
which presents an ideal substrate for chaperone binding. There
are many chaperone systems in the ER (see next section), and
the choice of which chaperones are recruited depends on the
specific sequence elements within the particular translocating
protein [37]. The chaperones usually remain on the maturing
polypeptide well after protein synthesis and translocation are
completed, and different subsets of chaperones and maturation
factors may associate with the same polypeptide at different
stages of its maturation. This is especially likely if a protein’s
maturation is complex, although such non‐model proteins are
relatively poorly studied. Because these chaperones are ER resi-
dents, their repeated binding to immature secretory and mem-
brane proteins serves the dual function of preventing the
premature exit of nonfunctional products from the ER.
Multiple chaperone systems operate
in the ER lumen
Among the various ER chaperones, perhaps the most abundant
and well studied is a luminal protein called BiP. BiP is a mem-
ber of a very large and well‐conserved family of chaperones
(called the Hsp70 family) with homologs in all known organ-
isms and most cellular compartments [38]. These chaperones
are ATPases that utilize the energy of ATP hydrolysis to bind
and release from exposed hydrophobic patches on substrates.
The ADP‐bound state of the Hsp70s has high affinity for sub-
strate; upon ATP binding, substrates are released so they can
attempt to fold. The steps of ATP binding, ATP hydrolysis, and
exchange of ADP for ATP are all tightly regulated by additional
factors broadly termed co‐chaperones. Thus, BiP and its various
co‐chaperones form a universal and general chaperone system
in the ER lumen that recognizes and shields exposed hydropho-
bic patches of secretory and membrane proteins in the process
of folding and assembly.
Although less extensively studied, another very abundant
ATPase in the ER lumen termed GRP94 also functions as a
chaperone. Like BiP, GRP94 has a cytosolic homolog (termed
Hsp90), further illustrating the theme that similar conserved
mechanisms of protein folding are utilized in multiple cellular
compartments. Hsp90 interactions with its wide range of sub-
strates are regulated by its ATPase cycle and co‐chaperones
[39]. It is likely that GRP94 functions analogously to Hsp90,
and is important for the folding and maturation of a subset of
proteins that include integrins (cell surface adhesion molecules),
toll‐like receptors (cell surface receptors of the innate immune
system), and immunoglobulins [40].
Because most proteins transiting the ER become glyco-
sylated, eukaryotic cells have evolved to exploit this modifica-
tion for numerous purposes including protein stability, solubility,
protein folding, quality control, trafficking, and protein–protein
interactions [41]. In the ER, a major use for N‐linked glycans is
the recruitment of a class of factors known as lectins (a general
term which simply means carbohydrate‐binding). The two
­lectin‐based chaperones in the ER are called calnexin (CNX,
a  membrane protein) and calreticulin (CRT, a soluble
­luminal  protein). They function by a very similar mechanism
(Figure 10.3) and are recruited to substrates via interaction with
a specific isoform of the N‐linked glycan. As mentioned above,
the initial glycan added to a protein consists of 14 sugars: two
N‐acetyl‐glucosamines, nine mannoses, and three terminal
­glucoses. Shortly after addition of this core glycan, two of the
terminal glucoses are removed by the action of glucosidases,
resulting in a mono‐glucosylated form of the N‐glycan that is
recognized specifically by CNX or CRT. Upon release from
CNX/CRT, the remaining glucose is removed by a glucosidase
and the protein has an opportunity to fold.
If it fails to fold properly, a enzyme termed UDP‐
glucose:glycoprotein glucosyltransferase (UGGT) adds a single
glucose back to the N‐glycan, thereby making it a substrate for
CNX/CRT binding and initiating another round of attempted
folding. Thus, as with the general chaperones discussed above,
repeated cycles of chaperone binding and release permit a sub-
strate to attempt folding while being protected from inappropri-
ate interactions [42]. In this case, a reversible glucose tag is used
to modulate substrate–chaperone interactions, with UGGT
­acting as a “folding sensor” to determine when the substrate
has folded. If a substrate passes UGGT inspection, it is not re‐­
glucosylated and the glucose‐free N‐glycosylated protein is
released from the CNX/CRT folding cycle. These proteins are
deemed to be folded, and are exported out of the ER to the Golgi
and beyond.
Glycosylated proteins do not have an unlimited time to fold.
The ER also contains mannosidases such as the EDEM family
and mannosidase I. These are typically very slow, but over time,
can remove terminal mannose residues from the N‐glycan. Once a
particular terminal mannose has been removed, the protein is not
allowed to attempt folding any longer and is routed for d­ egradation.
The degradation pathway from the ER is discussed later.
A distinctive feature of the ER lumen is its oxidizing environ-
ment [43]. This means that cysteines in nascent proteins will
typically be oxidized into disulfide bonds with other cysteines
(either within the same protein, or in other proteins). Disulfide
bonds can greatly stabilize the folded state of a protein or pro-
tein complex, thereby facilitating its function in the harsh extra-
cellular environment. For example, antibodies (e.g. IgG) are
composed of multiple polypeptide chains that are held together
by a series of intra‐ and intermolecular disulfide bonds that lend
it remarkable stability and longevity after secretion.
The correct formation of disulfide bonds in a nascent protein
is mediated by a class of chaperone enzymes termed oxidore-
ductases, the most well known of which is protein disulfide
isomerase (PDI) [44]. These chaperones can interact directly
with substrates (often via exposed hydrophobic patches). Upon
substrate interaction, an internal disulfide bond within PDI can
be reduced concomitantly with oxidation of cysteines in the
substrate. Thus, a disulfide bond is effectively transferred from
PDI to a substrate with which it interacts. The PDI then becomes
re‐oxidized by other enzymes (e.g. the oxidase Ero1) for another
round of substrate oxidation. PDI can also function in reverse,
and has the capacity of reducing disulfide bonds within a sub-
strate. This means that by reducing and re‐oxidizing different
 10:  Protein Maturation and Processing at the Endoplasmic Reticulum 115
pairs of cysteines, PDI can “shuffle” disulfide bonds until the
correct protein fold is achieved (and hence the name disulfide
isomerase). Some oxidoreductases are recruited to substrates
indirectly via interaction with other chaperones. For example,
the oxidoreductase ERP57 interacts with CRT (Figure  10.3),
thereby allowing it to operate on glycoproteins [45]. Thus,
different members of the family of oxidoreductases in the
­ many substrates, by other maturation pathways. Indeed, the loss
ER  lumen probably operate on different subsets of client
substrates.
Although the above chaperone systems are the most abundant
and well‐studied components of the protein maturation machin-
ery in the ER, they are not the only ones. The sheer volume and
diversity of substrates that transit the ER necessitates numerous
other factors to ensure correct protein folding and maturation.
For example, peptidyl‐prolyl‐isomerases catalyze the change in
conformation of proline‐containing peptide bonds between the
cis and trans isomers. Other chaperones, such as tapasin, help in
the assembly of peptides onto MHC class I. Yet other factors are
likely to be specialized for the assembly of multimeric protein
complexes or specifically involved in membrane protein folding
and assembly, although both of these complex processes are
rather poorly understood. And finally, there are likely to exist
numerous highly specialized chaperones dedicated to specific
substrates with unique requirements. For example, collagen bio-
synthesis utilizes a chaperone called Hsp47 that is selectively
expressed in fibroblasts. In the liver, apolipoprotein B biosyn-
thesis uniquely requires a factor termed microsomal triglyceride
transfer protein (MTTP) that mediates non‐covalent lipid–
protein interactions needed for lipoprotein particle assembly.
Many proteins are assembled into
multimeric complexes
Numerous proteins function as part of multiprotein complexes.
Secretory and membrane protein complexes are often assem-
bled in the ER by processes that are relatively poorly under-
stood. In fact, some of the most important and highly expressed
proteins transiting the secretory pathway are multiprotein com-
plexes, including immunoglobulins, the T‐cell receptor, ion
channels, MHC complexes, and numerous others. It is thought
that non‐assembled subunits of a larger complex are associated
with chaperone(s) until they find and associate with the appro-
priate interacting partner(s). For example, BiP (IgG‐binding
protein) was originally discovered by its association with
incompletely assembled IgG heavy chains. How assembly is
coordinated is not known in most cases, and it is unclear whether
it is simply a stochastic process of subunits finding each other
by diffusion, or if there are more regulated mechanisms that
facilitate assembly.
From this discussion, it should be obvious that the ER lumen
is a remarkably complex protein‐folding and maturation factory.
Some features of this factory are universal: it is highly rich in
evolutionarily conserved chaperone systems that protect nascent
proteins from inappropriate interactions and aggregation. Other
aspects of the ER are unique to this organelle: it has N‐linked
glycosylation, lectin‐based chaperones, an oxidizing environ-
ment favoring disulfide bond formation, and a very high flux of
membrane proteins that need proper folding and maturation.
And yet other ER‐specific processes are highly specialized for
unique substrates in only certain cell types. All of these path-
ways operate simultaneously; some are probably purely parallel
processes, while most are partially overlapping and coordinated
events. Thus, the loss or reduced capacity of some maturation
pathways can likely be compensated, at least temporarily for
of a surprising number of the major chaperones (e.g. CRT,
CNX, GRP94, UGGT, and many co‐chaperones) is compatible
with cellular viability. Yet, in each case, the function of a com-
plex organism or differentiated tissues is severely compromised
(usually leading to embryonic lethality), illustrating the limits of
redundancy. Although many basic principles have been deline-
ated using model systems, our knowledge of protein maturation
should be considered fragmentary and incomplete at best.
QUALITY CONTROL AND THE CULLING
OF IMMATURE PROTEINS
The relegation of secretory and membrane protein biosynthesis
to an intracellular compartment (in contrast to the plasma mem-
brane as occurs in prokaryotes) provides many advantages to the
eukaryotic cell. Perhaps the most obvious and biggest benefit is
the opportunity for considerably more control over what does
and does not gain access to the cell surface. This control is
­manifested in many ways. For example, the cell can effectively
regulate the quantity of a secretory or membrane protein
released to the cell surface independently of the protein’s syn-
thesis. Thus, key regulatory proteins such as hormones, surface
receptors, or ion channels can be made and stored intracellu-
larly, then rapidly mobilized to the surface at a moment’s notice.
This can provide cells with exquisite temporal control and
responsiveness to changing environmental conditions.
The other main regulatory benefit is quality control [46]. By
making and inspecting all secretory and membrane proteins
intracellularly, the cell can ensure that only properly folded and
functional products have access to the surface. This is critical
because partially functional or misfolded proteins can often be
far worse than no protein at all. Thus, in highly complex
­multicellular organisms where intercellular communication via
secreted proteins and their receptors are essential for overall
­fitness, quality control is a key functional role of the ER.
The basic logic of ER quality control consists of five general
steps (Figure  10.4). First, the cell must have mechanisms to
­recognize a misfolded, misassembled, or otherwise damaged
protein. This task is thought to be carried out by chaperones,
which, as discussed above, can discriminate between mature
and non‐native protein structures. Second, once a potentially
misfolded protein is identified, it must be targeted to degrada-
tion machinery. Third, misfolded proteins must be retrotranslo-
cated back to the cytosol. Fourth, once the polypeptide is
exposed to the cytosol, it must be ubiquitinated, typically by the
same machinery that mediates retrotranslocation. And finally,
the ubiquitin is used as a handle for extraction of the protein
from the ER and delivery for degradation by the proteasome.
This entire process is often referred to as ER‐associated
116 THE LIVER:  QUALITY CONTROL AND THE CULLING OF IMMATURE PROTEINS

1 2 3
Misfolded
protein Recognition Targeting Retrotranslocation

Lectin Poly-
ubiquitin
Chaperone

Deglycosylation
Ubiquitination
E3 ligase
complex
4

Extraction and
degradation

5
Proteasome

Figure 10.4  The major steps in endoplasmic reticulum (ER)‐associated degradation (ERAD). Schematic depiction of the overall logic of ERAD
using a misfolded glycoprotein as an example. The misfolded protein is recognized by a combination of a lectin (such as OS‐9) and chaperone(s)
(step 1). The recognition complex is then targeted to a retrotranslocation site built around an ER‐resident E3 ubiquitin ligase complex (step 2). The
misfolded protein then dissociates from the targeting complex and is retrotranslocated through a channel that is formed, at least partially, by the E3
ligase complex (step 3). On the cytosolic side, the exposed misfolded protein is ubiquitinated (step 4). The poly‐ubiquitin serves as a handle for
extraction from the membrane by the p97 ATP complex (not depicted) and degradation by the proteasome (step 5). Non‐glycoproteins are thought
to use the same general steps, but do not utilize lectins for recognition.

degradation, or ERAD [47]. This core series of events applies to
both secretory and membrane proteins that fail to mature prop-
erly in the ER, although the specific factors involved in their
recognition and retrotranslocation may be different.
Although less common, there may be situations where mis-
folded proteins in the ER can be routed via alternative pathways
for degradation in lysosomes. One class of proteins that seem
to  use the lysosomal pathway are GPI‐anchored proteins. It
appears that after attachment of the GPI lipid anchor, the protein
cannot be efficiently retrotranslocated back to the cytosol.
Instead, these proteins use vesicular trafficking pathways to
access the lysosome for degradation [48].
Misfolded proteins are recognized
by chaperones
Nascent proteins entering the ER typically undergo repeated
cycles of chaperone binding and release in an attempt to achieve
their final folded and assembled state. If maturation nonetheless
fails, the cell must eventually triage the misfolded protein for
degradation. In addition to their role in folding, chaperones also
play an important role in degradation. Most mutant or otherwise
misfolded proteins are typically found associated with chaper-
ones. This is likely to be important for three reasons. First, chap-
erones prevent inappropriate interactions between a misfolded
protein and other cellular factors. Second, the chaperone
­maintains the protein in a largely unfolded state (or facilitates
unfolding in some cases), which may be important for its retro-
translocation out of the ER (see next section). And third, the
chaperone may interact with or deliver proteins to the machin-
ery for retrotranslocation.
In the case of glycoproteins, sugar trimming and ER lectins
play a key role in degradation [49] in addition to their role during
the folding cycle. After repeated deglucosylation and regluco-
sylation cycles through the CNX/CRT system, mannosidases
(including proteins termed EDEMs) remove terminal mannoses,
including the one on which glucosylation normally occurs. This
allows the substrate to exit the CNX/CRT cycle, and now makes
the mannose‐trimmed glycan a substrate for binding by another
lectin termed OS‐9. Mannose trimming by EDEMs and binding
by OS‐9 routes substrates into the degradation pathway via OS‐9
associations with the retrotranslocation machinery [50, 51].
A retrotranslocation pathway for misfolded
protein degradation in the cytosol
The export of misfolded secretory or membrane proteins to the
cytosol for degradation is termed retrotranslocation. This pro-
cess involves components in the ER lumen, within the mem-
brane, and in the cytosol [47]. At the heart of the retrotranslocation
 10:  Protein Maturation and Processing at the Endoplasmic Reticulum 117
reaction is a multiprotein complex centered around one of
several membrane‐embedded E3 ubiquitin ligases [50, 51].
­ AND ABUNDANCE
These ligase complexes are thought to directly bind to misfolded
membrane proteins via interactions within the membrane. In the
case of misfolded proteins in the ER lumen, chaperones and/or
lectins associated with the misfolded proteins deliver them to the
E3 ligase complex. Then, in a step that remains poorly under-
stood, the E3 ligase complex provides a channel through which
the misfolded protein can access the cytosol [52].
Once misfolded proteins are cytosolically exposed, they are
poly‐ubiquitinated by the E3 ligase complex. Ubiquitin is a
small protein whose covalent attachment to substrates marks
them for degradation. In addition to providing a mark, ubiquitin
may prevent sliding of the misfolded protein back into the ER
lumen. The ubiquitinated protein then recruits a large ATPase
complex containing a protein termed p97 (also called VCP or
Cdc48) [53]. Together with associated cofactors, the p97 com-
plex appears to recognize both substrate and ubiquitin, and uses
the energy of ATP hydrolysis to mediate extraction. The p97‐
associated poly‐ubiquitinated substrate is then delivered to the
proteasome, a large multicatalytic enzyme responsible for most
protein degradation in the cytosol.
Although the general scheme of recognition, retrotransloca-
tion, ubiquitination, and degradation is likely to apply for essen-
tially all misfolded secretory and membrane proteins, the details
may differ in a substrate‐specific manner. For example, it is
already clear that glycoproteins and non‐glycoproteins utilize
different (but overlapping) pathways. Similarly, soluble luminal
proteins have different requirements for their degradation than
integral membrane proteins. Furthermore, the specific location
of a misfolded domain (whether in the lumen, in the TMD, or
cytosol) may influence which retrotranslocation machinery is
utilized. Thus, just as multiple parallel and partially overlapping
pathways operate during protein maturation of different types of
substrates, quality control and retrotranslocation pathways are
similarly complex. The diversity of substrates is simply too vast
for a single unifying machinery to handle all potential cases.
Nevertheless, the steps of ubiquitination, p97‐dependent extrac-
tion, and proteasomal degradation appear to be universal to all
substrates.
Physiologic uses of quality control
Although quality control and degradation are typically considered
pathways of aberrant protein disposal, they are physiologically
exploited for regulatory control in some cases. Two examples of
central importance to liver physiology are the regulation of apoli-
poprotein B and HMG‐CoA reductase. Both proteins are exten-
sively degraded at the ER by the same quality‐control and ERAD
machinery used for misfolded proteins [54, 55]. Yet, changes in
the environment or physiologic state of a cell can rapidly and
selectively alter their fate from degradation to maturation. For
apolipoprotein B, triglyceride abundance negatively regulates its
ER‐associated degradation, leading to greater secretion of lipo-
protein particles. Analogously, reduced cholesterol levels prevent
HMG‐CoA reductase degradation, allowing its increased func-
tional expression in the cholesterol biosynthetic pathway. Thus,
general quality‐control pathways have been exploited for highly
selective physiologic regulation in these and other situations.
REGULATION OF ER HOMEOSTASIS
The ER is not a static organelle. Its abundance, composition,
and functions adapt to changes in functional need. Highly secre-
tory cells and tissues (such as hepatocytes, the exocrine pan-
creas, and antibody‐secreting B cells) are packed with ER in
amounts that greatly exceed the amounts found in nonsecretory
cells. In fact, this correlation between ER abundance and secre-
tory capacity first led to the hypothesis that the ER was involved
in secretion. It is therefore not surprising that cells have mecha-
nisms to sense changes in the load of secretory and membrane
proteins transiting the ER, and adapt accordingly.
The discovery of these pathways came about through the
study of how cells respond to excessive misfolded proteins in
the ER. The “unfolded protein response” (UPR) is activated
whenever the load of secretory and membrane proteins exceeds
the capacity of the ER to properly mature and/or metabolize
them, a situation generally termed ER stress [56]. The net result
of UPR activation is the initiation of signaling pathways that
simultaneously alleviate the load on the ER (temporarily), while
upregulating the biosynthetic and maturation machinery to
expand ER processing capacity. Failure to effectively adapt to
ER stress leads to chronic UPR activation, cellular dysfunction,
and, at some point, apoptotic cell death [57]. Chronic ER stress
and the consequences arising from it are increasingly appreci-
ated to play central roles in various diseases, including many
afflicting the liver.
The unfolded protein response
communicates ER stress
Non‐native proteins, whether in the process of folding or in
preparation for degradation, are typically chaperone bound.
Hence, exceeding the protein‐processing capacity of the ER
has two consequences: the availability of unoccupied chaper-
ones decreases and at least some nascent polypeptides will not
have chaperones bound to them. Both of these consequences
seem to be utilized for sensing ER stress and activating
the UPR.
There are three known ER transmembrane proteins that act as
UPR sensors: Ire1, PERK, and ATF6 [56]. The luminal domains
of Ire1 and PERK associate with BiP, which maintains these
proteins in a monomeric state that is inactive for signaling.
When the levels of unfolded proteins increase, BiP molecules
dissociate from Ire1 and PERK and associate instead with
unfolded proteins. Dissociation of BiP from the UPR sensors
frees them to be activated. In the case of Ire1 (and possibly
PERK), misfolded proteins then bind directly to the luminal
domain, acting as a “ligand” for its activation. ATF6 is also inac-
tive under normal conditions and activated during ER stress;
however, the mechanism of activation is not known. Upon acti-
vation, each UPR sensor initiates a different set of downstream
reactions that ultimately result in transcriptional activation of
genes that increase the protein‐folding capacity of the ER
(Figure 10.5). After homeostasis in the ER is restored and the
levels of unfolded proteins are reduced, the UPR sensors are
returned to their inactive states.
118 THE LIVER:  REGULATION OF ER HOMEOSTASIS AND ABUNDANCE

Normal conditions ER stress

Abortive
translocation
Endoplasmic reticulum Golgi

BiP

??

Ire1 PERK ATF6

Trafficking and
XbP1 splicing elF2α phosp. proteolysis

Translation
[general]

ATF4 Translation

Transcription
[genes for chaperones, ERAD, metabolism, redox, trafficking]

Figure 10.5  Endoplasmic reticulum (ER) stress initiates a multi‐tiered unfolded protein response. During unstressed normal conditions (left), the
signal transduction factors Ire1, PERK, and ATF6 are inactive. Ire1 and PERK are thought to be held in an inactive state by their binding to the ER
luminal chaperone BiP, while the mechanism of ATF6 repression is unknown. During ER stress (right), the chaperone BiP is engaged with the
excessive quantity of misfolded proteins. This relieves repression of Ire1, PERK, and ATF6, leading to their activation by different mechanisms.
Ire1 catalyzes the splicing of mRNA coding for Xbp1, resulting in production of Xbp1 protein. PERK phosphorylates eIF2α, which causes reduced
global translation and selectively increased translation of ATF4. Relieving the repression of ATF6 leads to its trafficking to the Golgi, where it is
proteolytically cleaved to release a cytosolic domain. Xbp1, ATF4, and ATF6 are all transcription factors that collectively modulate the expression
of hundreds of genes that act to improve the protein‐processing capacity of the ER.

Ire1 activation (via autophosphorylation) causes its cytosolic
domain to act as an endonuclease to mediate the splicing of
mRNA coding for Xbp1. Upon Ire1‐mediated removal of an
intron from Xbp1 mRNA, a functional transcription factor can
be translated.
PERK activation leads to its cytosolic kinase domain phospho-
rylating the translation initiation factor eIF2α. Phosphorylation of
eIF2α leads to a general reduction in overall translation (thereby
reducing the burden of new proteins requiring ER function), while
allowing increased translation of a select few messages, including
the transcription factor ATF4. The UPR sensor ATF6 trafficks to
the Golgi, where it is proteolytically processed within its trans-
membrane domain. This cleavage releases the cytosolic domain
of ATF6, which functions as a transcription factor.
Thus, UPR activation leads to the production of at least three
transcription factors (Xbp1, ATF4, and ATF6), which together
initiate complex changes in gene expression involving hundreds
of factors [58, 59]. The most obvious adaptations are the upreg-
ulation of chaperones (including BiP) needed for protein matu-
ration in the ER. In addition, ER components involved in protein
translocation, protein degradation, lipid biosynthesis, redox
homeostasis, and others are also increased. The net result is an
increase in protein‐processing capacity of the ER. Depending
on the extent of this increase, the ER itself may be expanded to
accommodate the higher demand on its function.
In addition to these transcriptional changes, other adaptations
are also induced in the more acute time frame. As already noted
above, the most prominent acute adaptation is a global reduction
in translation via phosphorylation of eIF2α by PERK. In addi-
tion, certain ER‐associated mRNAs are rapidly degraded during
ER stress due to the endonuclease activity of Ire1 [60].
Furthermore, the import of some secretory or membrane pro-
teins into the ER may be attenuated as a consequence of reduced
availability of chaperones needed for translocation [61]. This
translocational regulation is substrate‐selective, and may serve
to reduce the burden during acute stress of non‐essential pro-
teins that are particularly prone to misfolding. Thus, a combina-
tion of general and specific responses can temporarily reduce
substrate flux into the ER until the transcriptional response has
an opportunity to improve ER homeostasis.
Excessive ER stress leads to programmed
cell death
If cells experience ER stress for a prolonged time, they eventu-
ally undergo apoptosis (also termed programmed cell death)
[57]. Remarkably, the signals that initiate apoptosis derive from
Ire1 and PERK, the same molecules charged with alleviating ER
stress. In the case of Ire1, two mechanisms have been proposed.
First, Ire1 may associate with the signal transduction factor
 10:  Protein Maturation and Processing at the Endoplasmic Reticulum 119
TRAF2, whose activation can initiate a signaling pathway cul-
minating in apoptosis. Second, the endonuclease activity of Ire1
has been proposed to degrade the mRNAs coding for various
pro‐survival factors, thereby leading to cell death. Evidence for
these Ire‐mediated signaling pathways come from experiments
in cell culture, but their role in vivo remains to be established.
Cell death downstream of PERK activation is unambiguous.
The transcription factor downstream of PERK, ATF4, includes
CHOP as one of its transcriptional targets. CHOP is another tran-
scription factor which can induce numerous genes, including
those that mediate apoptosis. The longer CHOP is activated, the
higher its levels and the greater the probability of initiating cell
death pathways. Conversely, cells and animals lacking CHOP
are refractory to cell death in response to prolonged ER stress.
With both Ire1 and PERK, the signaling responses leading to
cell death are slower than those leading to an improvement of
ER homeostasis. This difference in timing means that with mod-
erate or short‐term stresses, ER homeostasis is restored and
signaling from Ire1 and PERK is attenuated before they trigger
cell death. Only with prolonged UPR activation is cell death
favored. It appears that in multicellular organisms, it is prefera-
ble to lose a chronically stressed cell than risk the adverse con-
sequences of its prolonged dysfunction or death by necrotic
mechanisms that may induce inflammation.
The UPR is used to expand the ER during
development
The transcriptional response downstream of the UPR sensors
increases expression of factors that build the ER. This includes
lipid biosynthesis enzymes, the various proteins in the ER mem-
brane and lumen, and the trafficking factors that mediate protein
export from the ER. These factors improve the protein‐process-
ing capacity of the ER in response to stress as discussed above.
Remarkably, the same signaling pathway is also used to expand
the ER in cell types that are specialized for the high level pro-
duction of secretory or membrane proteins.
The most dramatic examples of a vastly expanded ER are
antibody‐secreting plasma cells and the exocrine pancreas that
secretes vast quantities of digestive enzymes. Experiments in
mice have shown that Xbp1 (the downstream target of Ire1) is
essential for development of plasma cells [62] and exocrine tis-
sues [63]. It was initially thought that the high expression of
secretory proteins induced during development of these cell
types caused ER stress, leading to Ire1 signaling and ER expan-
sion. However, later experiments showed that ER expansion
does not depend on first inducing secretory protein load; instead,
the ER expands in preparation for the secretory proteins to come
[64]. Thus, in addition to its role as a stress sensor, Ire1 signaling
is utilized as an intracellular signaling pathway that is develop-
mentally regulated to control ER abundance. Whether the other
UPR sensors are also developmentally regulated is not known.
Role of the UPR in various diseases
As we have discussed, the UPR has the capacity to protect cells
from the effects of stress, but execute a cell death program if the
stress is prolonged. This key role in making life‐or‐death
decisions means that the UPR can have a major impact on
organism physiology. Hence, ER stress and UPR signaling have
been implicated in the pathogenesis of a variety of diseases.
There are two general ways the UPR impacts pathophysiology.
First, the UPR may cause the death of stressed cells that would
have been useful to the organism had they not died. Second, the
adaptive features of the UPR might allow dysfunctional cells to
live and grow despite their detrimental effects on the organism.
Examples of undesirable UPR‐induced cell death include
inherited diseases of protein misfolding and lifestyle‐induced
tissue damage. For example, Charcot–Marie–Tooth 1B neurop-
athy can be caused by a mutation in the P0 glycoprotein of
Schwann cells. Expression of this mutant protein causes ER
stress, cell death, and demyelination. Strikingly, deletion of
CHOP in a mouse model of this disease alleviates both cell
death and neuropathy [65]. In another example, type 2 diabetes
imposes a high demand for insulin production from pancreatic
beta cells. This causes chronic ER stress and eventual beta cell
death, the latter of which can be partially reversed by deletion of
CHOP [66]. Thus, excessive UPR signaling leading to CHOP
expression may eliminate cells that would otherwise retain par-
tial functionality.
Examples of undesirable UPR‐mediated protection include
certain types of cancer [67] and infectious diseases [68]. For
example, in multiple myeloma, an activated UPR facilitates the
high levels of antibody production and rapid cell growth by con-
stantly improving ER function. In virally infected cells, the pro-
duction of viral glycoproteins is also facilitated by UPR‐mediated
upregulation of ER functionality. In both of these cases, cell
death is a more desirable outcome of the UPR than adaptation.
The recent development of small molecule modulators of differ-
ent aspects of the UPR [69, 70] may hold promise for interven-
ing in at least some of the diseases where slightly more or less
UPR signaling is desirable.
CONCLUSIONS AND UNKNOWNS
It has been over 50 years since the ER was discovered as the site
of secretory and membrane protein biosynthesis. In that time, a
cohesive molecular framework has been developed to explain
the mechanisms of selective segregation of these proteins to the
ER, their translocation across or insertion into the ER mem-
brane, and their maturation by various chaperones and process-
ing enzymes. The major pathways for ER‐based quality control
and maintenance of ER homeostasis have been elucidated.
While the broad concepts and core machinery for most of these
pathways are now in hand, many gaps exist in our knowledge.
How are the processes of targeting and translocation to the ER
regulated, especially for complex proteins or under different
conditions? How are multi‐spanning membrane proteins relia-
bly inserted in the precisely desired topology and subsequently
folded and assembled into a functional product? What kind of
specialized maturation and quality‐control pathways operate in
different cell types, and how do they differ from the universal
ones that have been studied so far? How are the various steps in
a protein’s folding and maturation coordinated? How does the
cell distinguish between folding proteins en route to a mature
120 THE LIVER:  REFERENCES
product, and a misfolded protein that should be degraded? How
is the response to ER stress and the choice between adaptation
and death tuned differently among different cell types? And
what is the role of ER stress in the various diseases where it is
observed to be activated? The answers to these and numerous
other questions await further study and should occupy cell biol-
ogists for the foreseeable future.
REFERENCES
1. Claude, A. Fractionation of mammalian liver cells by differential centrifuga-
tion: II. Experimental procedures and results. J Exp Med, 1946;84:61–89.
2. Palade, G.E. and Siekevitz, P. Liver microsomes; an integrated morphologi-
cal and biochemical study. J Biophys Biochem Cytol, 1956;2:171–200.
3. Redman, C.M. and Sabatini, D.D. Vectorial discharge of peptides released by
puromycin from attached ribosomes. Proc Natl Acad Sci U S A,
1966;56:608–15.
4. Lippincott‐Schwartz, J., Bonifacino, J.S., Yuan, L.C., and Klausner, R.D.
Degradation from the endoplasmic reticulum: disposing of newly synthe-
sized proteins. Cell, 1988;54:209–20.
5. Blobel, G. Intracellular protein topogenesis. Proc Natl Acad Sci U S A,
1980;77:1496–500.
6. von Heijne, G. Signal sequences. The limits of variation. J Mol Biol,
1985;184:99–105.
7. UniProt Consortium. UniProt: the universal protein knowledgebase. Nucleic
Acids Res, 2018;46:2699.
8. Jan, C.H., Williams, C.C., and Weissman, J.S. Principles of ER cotransla-
tional translocation revealed by proximity‐specific ribosome profiling.
Science, 2014;80(346):716.
9. Kutay, U., Hartmann, E., and Rapoport, T.A. A class of membrane proteins
with a C‐terminal anchor. Trends Cell Biol, 1993;3:72–5.
10. Keenan, R.J., Freymann, D.M., Stroud, R.M., and Walter, P. The signal rec-
ognition particle. Annu Rev Biochem, 2001;70:755–75.
11. Rapoport, T.A., Li, L., and Park, E. Structural and mechanistic insights into
protein translocation. Annu Rev Cell Dev Biol, 2017;33:369–90.
12. Voorhees, R.M. and Hegde, R.S. Toward a structural understanding of co‐
translational protein translocation. Curr Opin Cell Biol, 2016;41:91–9.
13. Van den Berg, B., Clemons, W.M., Collinson, I. et al. X‐ray structure of a
protein‐conducting channel. Nature, 2004;427:36–44.
14. Voorhees, R.M., Fernández, I.S., Scheres, S.H.W., and Hegde, R.S. Structure
of the mammalian ribosome‐Sec61 complex to 3.4 Å resolution. Cell,
2014;157:1632–43.
15. Voorhees, R.M. and Hegde, R.S. Structure of the Sec61 channel opened by a
signal sequence. Science, 2016;351:88–91.
16. Shao, S. and Hegde, R.S. Membrane protein insertion at the endoplasmic
reticulum. Annu Rev Cell Dev Biol, 2011;27:25–56.
17. Higy, M., Junne, T., and Spiess, M. Topogenesis of membrane proteins at the
endoplasmic reticulum. Biochemistry, 2004;43:12716–22.
18. Wessels, H.P. and Spiess, M. Insertion of a multispanning membrane pro-
tein  occurs sequentially and requires only one signal sequence. Cell,
1988;55:61–70.
19. Shao, S. and Hegde, R.S. A calmodulin‐dependent translocation pathway for
small secretory proteins. Cell, 2011;147:1576–88.
20. Panzner, S., Dreier, L., Hartmann, E., Kostka, S., and Rapoport, T.A.
Posttranslational protein transport in yeast reconstituted with a purified com-
plex of Sec proteins and Kar2p. Cell, 1995;61:561–70.
21. Matlack, K.E., Misselwitz, B., Plath, K., and Rapoport, T.A. BiP acts as a
molecular ratchet during posttranslational transport of prepro‐alpha factor
across the ER membrane. Cell, 1999;97:553–64.
22. Hegde, R.S. and Keenan, R.J. Tail‐anchored membrane protein insertion into
the endoplasmic reticulum. Nat Rev Mol Cell Biol, 2011;12:787–98.
23. Stefanovic, S. and Hegde, R.S. Identification of a targeting factor
for  ­
posttranslational membrane protein insertion into the ER. Cell,
2007;128:1147–59.
24. Guna, A., Volkmar, N., Christianson, J.C., and Hegde, R.S. The ER mem-
brane protein complex is a transmembrane domain insertase. Science,
2018;359:470–3.
25. Karaplis, A.C., Lim, S.K., Baba, H., Arnold, A., and Kronenberg, H.M.
Inefficient membrane targeting, translocation, and proteolytic processing by
signal peptidase of a mutant preproparathyroid hormone protein. J Biol
Chem, 1995;270:1629–35.
26. Rane, N.S., Kang, S.‐W., Chakrabarti, O., Feigenbaum, L., and Hegde, R.S.
Reduced translocation of nascent prion protein during ER stress contributes
to neurodegeneration. Dev Cell, 2008;15:359–70.
27. Juszkiewicz, S. and Hegde, R.S. Quality control of orphaned proteins. Mol
Cell, 2018;71:443–57.
28. Hessa, T., Sharma, A., Mariappan, M. et al. Protein targeting and degradation
are coupled for elimination of mislocalized proteins. Nature, 2011;475:394–7.
29. Itakura, E., Zavodszky, E., Shao, S. et  al. Ubiquilins chaperone and triage
mitochondrial membrane proteins for degradation. Mol Cell, 2016;63:21–33.
30. Aebi, M. N‐linked protein glycosylation in the ER. Biochim Biophys Acta
Mol Cell Res, 2013;1833:2430–7.
31. Braunger, K., Pfeffer, S., Shrimal, S. et al. Structural basis for coupling pro-
tein transport and N‐glycosylation at the mammalian endoplasmic reticulum.
Science, 2018;360:215–19.
32. Ng, B.G. and Freeze, H.H. Perspectives on glycosylation and its congenital
disorders. Trends Genet, 2018;34:466–76.
33. Auclair, S.M., Bhanu, M.K., and Kendall, D.A. Signal peptidase I: cleaving
the way to mature proteins. Protein Sci, 2012;21:13–25.
34. Orlean, P. and Menon, A.K. Thematic review series: lipid posttranslational
modifications. GPI anchoring of protein in yeast and mammalian cells, or:
how we learned to stop worrying and love glycophospholipids. J Lipid Res,
2007;48:993–1011.
35. Fujihara, Y. and Ikawa, M. GPI‐AP release in cellular, developmental, and
reproductive biology. J Lipid Res, 2016;57:538–45.
36. Hartl, F.U., Bracher, A., and Hayer‐Hartl, M. Molecular chaperones in pro-
tein folding and proteostasis. Nature, 2011;475:324–32.
37. Molinari, M. and Helenius, A. Chaperone selection during glycoprotein
translocation into the endoplasmic reticulum. Science, 2000;288:331–3.
38. Mayer, M.P. and Bukau, B. Hsp70 chaperones: Cellular functions and
molecular mechanism. Cell Mol Life Sci, 2005;62:670–84.
39. Schopf, F.H., Biebl, M.M., and Buchner, J. The HSP90 chaperone machin-
ery. Nat Rev Mol Cell Biol, 2017;18:345–60.
40. Ansa‐Addo, E.A., Thaxton, J., Hong, F., et al. Clients and oncogenic roles of
molecular chaperone gp96/grp94. Curr Top Med Chem, 2016;16:2765–78.
41. Helenius, A. and Aebi, M. Roles of N‐linked glycans in the endoplasmic
reticulum. Annu Rev Biochem, 2004;73:1019–49.
42. Tannous, A., Pisoni, G.B., Hebert, D.N., and Molinari, M. N‐linked sugar‐
regulated protein folding and quality control in the ER. Semin Cell Dev Biol,
2015;41:79–89.
43. Bulleid, N.J. and van Lith, M. Redox regulation in the endoplasmic reticu-
lum. Biochem Soc Trans, 2014;42:905–8.
44. Gruber, C.W., Cemazar, M., Heras, B., Martin, J.L., and Craik, D.J. Protein
disulfide isomerase: the structure of oxidative folding. Trends Biochem Sci,
2006;31:455–64.
45. Oliver, J.D., van der Wal, F.J., Bulleid, N.J., and High, S. Interaction of the
thiol‐dependent reductase ERp57 with nascent glycoproteins. Science,
1997;275:86–8.
46. Sitia, R. and Braakman, I. Quality control in the endoplasmic reticulum pro-
tein factory. Nature, 2003;426:891–4.
47. Vembar, S.S. and Brodsky, J.L. One step at a time: endoplasmic reticulum‐
associated degradation. Nat Rev Mol Cell Biol, 2008;9:944–57.
48. Satpute‐Krishnan, P., Ajinkya, M., Bhat, S. et al. ER stress‐induced clear-
ance of misfolded GPI‐anchored proteins via the secretory pathway. Cell
2014;158:522–33.
49. Ruddock, L.W. and Molinari, M. N‐glycan processing in ER quality control.
J Cell Sci, 2006;119:4373–80.
50. Denic, V., Quan, E.M., and Weissman, J.S. A luminal surveillance complex
that selects misfolded glycoproteins for ER‐associated degradation. Cell,
2006;126:349–59.
51. Carvalho, P., Goder, V., and Rapoport, T.A. Distinct ubiquitin‐ligase com-
plexes define convergent pathways for the degradation of ER proteins. Cell
2006;126:361–73.
52. Wu, X. and Rapoport, T.A. Mechanistic insights into ER‐associated protein
degradation. Curr Opin Cell Biol, 2018;53:22–8.
53. Ye, Y., Meyer, H.H., and Rapoport, T.A. The AAA ATPase Cdc48/p97 and
its  partners transport proteins from the ER into the cytosol. Nature,
2001;414:652–6.
 10:  Protein Maturation and Processing at the Endoplasmic Reticulum 121
54. Wangeline, M.A., Vashistha, N., and Hampton, R.Y. Proteostatic tactics in
the strategy of sterol regulation. Annu Rev Cell Dev Biol, 2017;33:467–89.
55. Fisher, E., Lake, E., and McLeod, R.S.. Apolipoprotein B100 quality control
and the regulation of hepatic very low density lipoprotein secretion. J Biomed
Res, 2014;28:178–93.
56. Walter, P. and Ron, D. The unfolded protein response: from stress pathway to
homeostatic regulation. Science, 2011;80(334):1081–6.
57. Tabas, I. and Ron, D. Integrating the mechanisms of apoptosis induced by
endoplasmic reticulum stress. Nat Cell Biol, 2011;13:184–90.
58. Harding, H.P., Zhang, Y., Zeng, H. et al. An integrated stress response regu-
lates amino acid metabolism and resistance to oxidative stress. Mol Cell,
2003;11:619–33.
59. Travers, K.J., Patil, C.K., Wodicka, L. et al. Functional and genomic analy-
ses reveal an essential coordination between the unfolded protein response
and ER‐associated degradation. Cell, 2000;101:249–58.
60. Hollien, J., and Weissman, J.S. Decay of endoplasmic reticulum‐localized
mRNAs during the unfolded protein response. Science, 2006;313:104–7.
61. Kang, S.‐W., Rane, N.S., Kim, S.J. et al. Substrate‐specific translocational
attenuation during ER stress defines a pre‐emptive quality control pathway.
Cell, 2006;127:999–1013.
62. Iwakoshi, N.N., Lee, A.‐H., and Glimcher, L.H. The X‐box binding pro-
tein‐1 transcription factor is required for plasma cell differentiation and the
unfolded protein response. Immunol Rev, 2003;194:29–38.
63. Lee, A.‐H., Chu, G.C., Iwakoshi, N.N., and Glimcher, L.H. XBP‐1 is
required for biogenesis of cellular secretory machinery of exocrine glands.
EMBO J, 2005;24:4368–80.
64. Hu, C.‐C.A., Dougan, S.K., McGehee, A.M., Love, J.C., and Ploegh, H.L.
XBP‐1 regulates signal transduction, transcription factors and bone marrow
colonization in B cells. EMBO J, 2009;28:1624–36.
65. Pennuto, M., Tinelli, E., Malaguti, M. et al. Ablation of the UPR‐mediator
CHOP restores motor function and reduces demyelination in Charcot‐Marie‐
Tooth 1B mice. Neuron, 2008;57:393–405.
66. Song, B., Scheuner, D., Ron, D., Pennathur, S., and Kaufman, R.J. Chop ­deletion
reduces oxidative stress, improves β cell function, and promotes cell survival in
multiple mouse models of diabetes. J Clin Invest, 2008;118:3378–89.
67. Oakes, S.A. Endoplasmic reticulum proteostasis: a key checkpoint in cancer.
Am J Physiol Physiol, 2017;312:C93–C102.
68. Li, S., Kong, L., and Yu, X. The expanding roles of endoplasmic reticulum
stress in virus replication and pathogenesis. Crit Rev Microbiol,
2015;41:150–64.
69. Das, I., Krzyzosiak, A., Schneider, K. et al. Preventing proteostasis diseases
by selective inhibition of a phosphatase regulatory subunit. Science,
2015;348:239–42.
70. Sidrauski, C., Acosta‐Alvear, D., Khoutorsky, A. et  al. Pharmacological
brake‐release of mRNA translation enhances cognitive memory. Elife,
2013;2:e00498.
 Protein Degradation
11 and the Lysosomal System
Susmita Kaushik and Ana Maria Cuervo
Department of Developmental and Molecular Biology, Institute for Aging Research, Albert Einstein College
of Medicine, Bronx, NY, USA
INTRODUCTION TO INTRACELLULAR
PROTEOLYSIS
Intracellular proteins undergo continuous synthesis and degra-
dation [1, 2]. This constant renewal of the proteome assures its
stability and contributes to regulate its function. Altered or dam-
aged proteins are eliminated by the proteolytic systems before
their accumulation inside cells interferes with normal cell
function [1, 3]. Damaged proteins are first recognized by
molecular chaperones, which facilitate protein refolding/repairing
(Figure 11.1). However, if the damage is too extensive, or under
conditions unfavorable for protein repair, damaged proteins are
targeted for degradation. The proteolytic systems thus consti-
tute, along with intracellular chaperones, essential components
of the surveillance mechanisms responsible for cellular quality
control. Furthermore, the coordinated balance between protein
synthesis and degradation also allows cells to rapidly modify
intracellular levels of subsets of proteins to accommodate to a
changing extracellular environment or to particular cellular con-
ditions. Increased protein degradation can enhance the effect of
reduced protein synthesis to decrease the levels of particular
proteins inside cells. Conversely, shutdown of protein degrada-
tion is often used by cells to augment the cellular protein pool.
In addition to these roles in cellular quality control and pro-
tein homeostasis, proteolysis is also often used by cells as an
extra source of energy when nutrients are scarce [2, 4]. Thus, the
free amino acids resulting from the breakdown of cellular pro-
teins can be utilized for the synthesis of essential proteins dur-
ing starvation, but also as an additional source of energy through
their breakdown in the urea and creatinine cycles, or by their
transformation into glucose through gluconeogenesis in the case
of glucogenic amino acids. Despite the associated expenditure
of energy required to break down intracellular products, this
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
continuous recycling of the essential building blocks makes
protein degradation a very conservative and economic process
with a positive net energetic outcome. Lastly, protein degradation
is also necessary during major cellular remodeling (i.e. embryo-
genesis, morphogenesis, cell differentiation), and as a defensive
mechanism against harmful agents and pathogens [5–7].
Protein degradation is an essential cellular process active and
functional in all types of cells and conserved throughout evolu-
tion. Due to the very active metabolic nature of liver and the
unique properties of this organ as an experimental tool (acces-
sibility, relative cellular uniformity, and ease for morphological
analysis), most of the early studies in protein degradation were
performed in liver. In fact, the major proteolytic systems, their
essential components, and many of the regulatory mechanisms
of protein degradation were first identified in liver [8–12]. The
initial interest in liver proteolysis revolved around the role of
this process as a source of energy, in particular during starva-
tion. However, the importance of the proteolytic systems in
quality control in this organ has become increasingly appreciated
in recent years both in normal liver physiology and in certain
pathological conditions.
INTRACELLULAR PROTEOLYTIC
SYSTEMS
Hepatocytes, like most cells, contain a large variety of proteases
both in the cytosol and confined in intracellular compartments.
Some of these proteases, such as caspases, calpains, or different
secretases, contribute to partial cleavage of cellular proteins, for
the most part with regulatory purposes (reviewed in [13]). For
example, many enzymes are synthesized as inactive prozymogens
 11:  Protein Degradation and the Lysosomal System 123

Life-cycle of proteins
(c)
(a) Folding/ CHAPERONES
assembling
Cytosol (b) REPAIR
Protein ENZYMES
Aggregation
Chaperone
Ribosomes
Refolding/
repair
TRANSLATION
MACHINERY CHAPERONES

(d)
Degradation Proteasome

Disaggregation

ER Lysosomes

PROTEOLYTIC SYSTEMS

Figure 11.1  Quality‐control mechanisms in liver. Two major systems contribute to quality control in liver: chaperones and the proteolytic
systems. Chaperones guarantee proper folding of all proteins after synthesis (a) and act as their companions throughout their biological cycle,
assisting in the unfolding/refolding required for translocation into different subcellular compartments or for assembly into protein complex.
Alterations in protein folding (b, c) are first detected by chaperones, which often facilitate their refolding back into a functional protein. However,
if proper refolding is not possible, the altered proteins are eliminated from the cell by degradation through the proteolytic systems, namely the
ubiquitin–proteasome system and the lysosomes (d).

that need to undergo partial cleavage in order to become active.
Similarly, cleavage of particular regions in certain proteins
results in changes in their intracellular location while still
­preserving full functionality. However, the term “proteolytic
system,” the main focus of this chapter, has been reserved to
define the group of intracellular proteases, assisting components,
and intracellular compartments that participate in complete deg-
radation of proteins into their constitutive amino acids. Two
major proteolytic systems are responsible for most of the intra-
cellular protein degradation: the ubiquitin–proteasome system
(UPS) and the lysosomes [2]. Although the main focus of this
chapter is on the lysosomal system, we briefly review here the
characteristics of the UPS and refer interested readers to recent
reviews on this topic for details.
THE UBIQUITIN–PROTEASOME SYSTEM
The proteasome is a multicatalytic complex with a proteolytic
core, known as the 20S proteasome, which in high eukaryotes
results from the association of 28 subunits in four rings stacked
as a cylinder‐like structure [3, 14–16]. Three major types of pro-
teolytic activity have been described in 20S proteasomes, but
because some of the subunits of this core are exchangeable, pro-
teasomes with different catalytic activities coexist in all cells.
The activity of the catalytic core is modulated through the
assembly of different regulatory subunits (19S or 11S) that dock
on one or both sides of the 20S proteasome to form several
proteasome species, likely to participate in different cellular
processes (Figure 11.2) [3, 15, 16]. The components of the regu-
latory subunits are primarily chaperones, ATPases, and enzymes
able to remove degradation tags from the substrate proteins
delivered to the proteasome [17, 18]. Although degradation of
substrate proteins directly by the nude 20S proteasome has been
described, degradation of a vast number of substrates involves
the participation of the regulatory complexes [19–21]. The subunits
of this complex mediate substrate recognition and unfolding,
and often act as the driving force that opens the proteasome bar-
rel and “pushes” substrate proteins into the catalytic core [22].
The proteasome can degrade untagged proteins, but most
substrates of this proteolytic system are selectively targeted for
degradation through the covalent linkage of ubiquitin, a small
(8 kDa) heat‐stable protein that also undergoes self‐conjugation,
resulting in the formation of poly‐ubiquitin chains bound to a
lysine in the candidate substrate [14, 20, 21, 23]. Linkage of
ubiquitin to the cargo proteins is mediated by a series of
enzymes – generically known as E ligases – that act sequentially
124 THE LIVER:  THE LYSOSOMAL SYSTEM

E4 Polyub
26S
19S

E3 β

20S β
α
Ubiquitination
Substrate
19S
Ubiquitin

E1
E2

Figure 11.2  The ubiquitin–proteasome system. The proteasome is a multicatalytic complex formed by a catalytic core (20S) and regulatory subunits
(19S and 11S). The catalytic core contains four rings of alpha and beta subunits assembled as a cylinder. Ubiquitin is attached to proteasome
substrates through a process mediated by regulatory enzymes known as ubiquitin ligases (E). E1 activates ubiquitin for assembly to the substrates,
while E2 transfers the activated ubiquitin to the substrate presented by E3. Multiple rounds through this cycle lead to the formation of poly‐ubiquitin
chains covalently linked to the targeted protein. Removal of the ubiquitin tag by the de‐ubiquitinating subunits located in the regulatory 19S particle
is required for internalization of the substrate into the catalytic core. Subunits of the regulatory 19S also modulate the aperture of the cylinder‐like
structure of the 20S proteasome to facilitate substrate access.

to activate ubiquitin, present it to the substrate, and catalyze the
conjugation [24]. Repeated cycles of ubiquitination result in the
formation of the poly‐ubiquitin chain recognized by the chaper-
ones and the ubiquitin‐binding subunits of the regulatory com-
plex of the proteasome. In many instances, ubiquitination is
preceded by phosphorylation of the substrate protein, which
often favors the exposure of the lysine residue that will be used
for conjugation of the ubiquitin by specific E‐recognizing
enzymes [20, 21, 23]. Ubiquitination is a universal form of pro-
tein tagging used not only as a marker for protein degradation
through the proteasome but also for degradation through other
proteolytic systems, intracellular targeting of proteins to par-
ticular cellular compartments, signaling, enzyme activation, and
regulation of membrane dynamics, among other things [24, 25].
How the same tagging mechanism could be involved in so dif-
ferent cellular processes has been, for a long time, one of the
burning questions in the field. The way in which ubiquitin is
conjugated to the proteins is decisive for the function of the tag
[26, 27]. Ubiquitin contains seven different lysine residues that
can potentially be used for conjugation to the substrates.
Linkage through particular residues has been shown now to
determine their recognition by different auxiliary proteins
involved in different cellular pathways, and hence to mediate
the different fates of the conjugated proteins [23].
The UPS plays a critical role as part of the cellular protein
quality‐control system for both cytosolic and secretory proteins
[3, 14, 23, 25]. Unfolded proteins, largely newly synthesized
cytosolic proteins that cannot reach their proper folded confor-
mation, and a large subset of post‐translationally damaged pro-
teins (oxidized, glycated, etc.) are removed from the cytosol via
the UPS system [28, 29].
Added to the participation of the UPS system in cellular
homeostasis and protein quality control by the selective removal
of altered proteins, this proteolytic complex also contributes to
the regulatory degradation of essential intracellular proteins,
usually in a very rapid manner. This fast degradation of factors
involved in cell cycle progression, cell division, transcription,
and cell signaling confers a critical regulatory role to the UPS in
multiple essential cellular processes [14, 30].
Malfunctioning of the UPS severely impairs cell viability.
Primary defects of components of this system have been identified
in different types of aggregopathies, establishing a link between
the UPS and cellular degeneration [23, 25, 31–35]. The toxic con-
sequences of proteasome blockage have been extensively explored
as anticancer treatment since functional proteasome is required for
cell cycle progression and cell division [36, 37].
THE LYSOSOMAL SYSTEM
Lysosomes are single‐membrane organelles containing a wide
variety of hydrolases, including proteases, lipases, glycosidases,
and nucleotidases, which make them able to degrade all kinds of
macromolecules. This high concentration of enzymes was actu-
ally what determined their original discovery as a cellular frac-
tion with high acid phosphatase activity [38]. These initial
biochemical observations by the group of de Duve were shortly
followed by the ultrastructural studies of Novikoff and col-
leagues, who performed the first electron microscopy study in
the isolated fraction and confirmed the presence of single‐membrane
vesicles of 0.1–0.5 μm diameter and spherical shape [39]. Most
 11:  Protein Degradation and the Lysosomal System 125
of the main criteria that defined a lysosome have persisted
despite advances in the understanding of the molecular compo-
nents and functional relevance of this lytic compartment. The
analysis of the different routes followed by the cargo (compo-
nents to be degraded by the lysosomes) in order to reach the
lysosomal lumen has given rise to the description of a series of
lysosome‐related compartments, such as endosomes, phago-
somes, and autophagosomes, all of which will be described in
the following sections.
Cytosolic vesicles are catalogued as lysosomes, also referred
to as secondary lysosomes once they have received cargo, and
distinguished from the other lysosome‐related vesicles when
they fulfill the following criteria: single membrane, pH in the
range of 4.5–5.5, active (cleaved) hydrolases, presence of lyso-
some membrane proteins, and absence of typical endosomal
markers such as mannose‐6‐phosphate receptor or particular
endosome‐associated Rab proteins [13, 38, 40, 41].
The enzymatic machinery
Lysosomal hydrolases are synthesized as pre‐proenzymes that
get activated through proteolytic cleavage as the pH decreases.
After synthesis in the endoplasmic reticulum (ER), all lysoso-
mal enzymes traffic through the Golgi, where they become gly-
cosylated and covalently tagged with a mannose‐6‐phosphate
residue. This tag is selectively recognized by the mannose‐6‐
phosphate receptor at the trans‐Golgi network (TGN), which
helps concentrate lysosomal enzymes inside small vesicles bud-
ding from the Golgi and targeted to endosomes. Sorting at the
Golgi requires the participation of adaptor proteins such as AP1
and of the GGAs or Golgi‐localized gamma‐adaptin ear homol-
ogy domain proteins [42, 43]. As the acidification of this
compartment increases with maturation, the hydrolases dis-
sociate from the receptor and are delivered to lysosomes,
whereas the receptor is recycled back to the Golgi. A percentage
of lysosomal enzymes escape this sorting step, despite the tag-
ging, and follow the secretory pathway to be released in the
extracellular media. Even though there is growing evidence that
lysosomal enzymes can play important functions in extracellu-
lar matrix remodeling, cell defense, and maintenance of the
extracellular environment [44], some of the released enzymes
are internalized back into the cell via endocytosis after interact-
ing with a variant of the mannose‐6‐phosphate receptor present
at the plasma membrane. Double knockdown of the two
mannose‐6‐phosphate receptors has revealed the presence of
an as yet poorly understood mannose‐6‐phosphate‐independent
­targeting [45].
Although more than 50 different lysosomal hydrolases have
been described in the lysosomal lumen, lysosomal pro-
teases – also known as cathepsins – are of particular interest for
this chapter. Cathepsins can act both as endo‐ and exopeptidases
(cleaving directly the internal residues of the amino acid
sequence of the cargo proteins, or only N‐terminal or C‐terminal
residues, respectively) and belong to the families of serine,
cysteine, and aspartic proteases [46–48]. It is generally accepted
that degradation of substrate proteins is initiated by endoproteo-
lytic cleavage that generates peptides amenable to exoprotease
degradation. Lysosomal proteases reach maximal activity at the
very acidic pH of the lysosomal lumen, but many of them still
have considerable residual activity at neutral pH. The particular
redox conditions of the lysosomal lumen and the low pH facili-
tate unfolding of internalized cytosolic proteins, which allows
proteases to gain access to internal residues. Changes in the
redox potential and in the luminal pH modulate the proteolytic
activity inside lysosomes [49, 50].
Defective lysosomal hydrolysis has been associated with
severe human disorders known generically as lysosomal storage
disorders (LSDs). These disorders differ in the particular hydro-
lase affected, but all of them share the presence of engorged
lysosome‐related compartments in the cytosol of the cell, with
the consequent cell and organ expansion (i.e. hepatomegaly and
splenomegaly are features common to many LSDs). Pathology
arises both as a result of the defective processing of the substrate
for the hydrolase affected in each LSD, which will limit recy-
cling of particular essential components, and as a consequence
of an abnormally expanded lysosomal system filled with unde-
graded products that alter the lysosomal lumenal conditions.
Thus, accumulation of these products inside lysosomes reduces
their, pH, changes their redox status, and alters the dynamics of
proteins between lysosomal lumen and membrane, as well as
the ability of lysosomes to fuse with other lysosomes (homo-
typic fusion) or with other vesicular compartments in the cell
(heterotypic fusion). Readers are referred to reviews for detailed
descriptions of the different LSDs and current therapeutic
advances for these disorders [51–54].
Proteins at the lysosomal membrane
The lysosomal membrane is far from a mere static barrier to
contain the highly active luminal hydrolases away from the
cytosol. On the contrary, proteins at the lysosomal membrane
mediate the essential functions of this organelle. Thus, trans-
porters at the membrane allow trafficking of cytosolic compo-
nents in and out of the lysosomal compartment, a proton pump
maintains the low luminal pH, and integral membrane proteins
and the membrane’s associated partners facilitate the fusion of
lysosomes with other vesicular compartments [42, 49, 55, 56].
The most abundant of the integral proteins at the lysosomal
membrane are the lysosome‐associated membrane proteins
(LAMPs) [55–58]. Two different but highly homologous pro-
teins, LAMP‐1 and LAMP‐2, are the most prominent members
of this family. They are single‐span membrane proteins with a
very heavily glycosylated luminal region and a very short cyto-
solic tail. Glycosylation constitutes about 60% of the mass of
these proteins and is required to preserve their stability, likely
by preventing luminal hydrolases from gaining access to their
peptide core. Despite their high homology, studies in animal
models knocked out for LAMP‐1 or LAMP‐2 have revealed
marked functional differences between these two proteins.
Knockout of LAMP‐1 has a very mild phenotype and upregula-
tion of LAMP‐2 [59], supporting the supposition that deficiency
of LAMP‐2 can be compensated for by LAMP‐2. In contrast,
loss of LAMP‐2 expression results in a dramatic phenotype with
major alterations in lysosomal biogenesis, problems in clear-
ance of autophagic compartments, and altered vesicular traffick-
ing [60]. This diversity of effects can be attributed to the
existence in most cells of three isoforms of LAMP‐2 resulting
from the alternative splicing of the Lamp2 gene [61]. These
126 THE LIVER:  THE LYSOSOMAL SYSTEM
isoforms  –  LAMP‐2A, LAMP‐2B, and LAMP‐2C  –  have
identical luminal regions but distinctive transmembrane and
cytosolic tails. As described below, LAMP‐2A is essential for a
selective type of autophagy of cytosolic proteins [62], whereas
LAMP‐2B seems to play a role in vesicular fusion between lys-
osomes and autophagic vesicles [60, 63]. Lastly, LAMP‐2C has
been described as a receptor for DNA and RNA uptake and deg-
radation in lysosomes [64].
The other abundant type of lysosomal membrane protein is
the lysosomal integral membrane protein (LIMP). The two
members of this family described to date, LIMP1 and LIMP2,
insert into the lysosomal membrane in a hairpin fashion, with
N‐ and C‐termini exposed in the cytosol. LIMP1 has been
shown to contribute to fusion events in secretory organelles,
whereas LIMP2 participates in lysosomal biogenesis through its
interactions with vesicle fusion and fission components [65].
Although less abundant than LAMPs and LIMPs, the con-
served vacuolar proton pump (V‐type H+ ATPase) has been
extensively studied for its essential role in maintaining the acid
lysosomal lumen [41]. This pump uses the energy derived from
ATP hydrolysis to promote translocation of protons through the
coordinated action of a cytosolic region with ATPase activity
and a transmembrane region that acts as a translocon.
Different transporters are present at the lysosomal membrane
to mediate translocation of hydrolysis byproducts from the lyso-
somal lumen toward the cytosol for recycling [66]. The only
amino acid transporter cloned to date is the cysteine transporter
or cystinosin, a seven‐span transmembrane protein that, when
mutated, gives rise to the LSD cystinosis [67]. A monosaccha-
ride transporter has also been identified at the lysosomal mem-
brane, and mutations leading to functional loss of this transporter
have been found in the LSD Salla disease. Lysosomal efflux of
other products, such as oligosaccharides, small peptides, and
free fatty acids, has been shown experimentally, but little is
known about the transporters that modulate their exit from the
lysosome. The role of lysosomal membrane transporters in
hepatic homeostasis of metals, such as copper, is also receiving
growing attention [68].
Lastly, some lysosomal enzymes are located at the lysosomal
membrane rather than the lumen. Typical examples of these
membrane enzymes are the 70 kDa lysosomal apyrase‐like
protein, which contributes to the metabolism of tri‐ and diphos-
phate nucleotides, and the acetyl‐CoA:α‐glucosaminide N‐
acetyltransferase, which transfers acetyl groups to heparin
sulfate. Part of the lysosomal acid phosphatase originally used
to identify this compartment in the early studies by de Duve is
also localized in the lysosomal membrane as an inactive single‐
span protein that is released into the lumen by proteolytic
processing.
As better understanding is gained on the function of different
lysosomal membrane proteins, the number of diseases related to
alterations in these proteins keeps growing. In addition to the
above‐mentioned LSD due to mutations in transporters, genetic
alterations in the lamp2 gene have also been connected to a
muscle vacuolopathy known as Danon disease [63]. Danon dis-
ease has been classified as a lysosomal glycogen storage disor-
der with normal acid maltase activity in which autophagic
vacuoles (AVs) accumulate in all tissues, including liver.
Patients suffer fatal cardiomyopathy due to the weakening effect
of the vacuoles on the cardiac muscle. LAMP‐2 deficiency is the
primary defect in this disease, which is associated with altered
lysosomal biogenesis and malfunctioning of the autophagic
system.
Lysosomal pathways for proteolysis
Lysosomes are the common final compartment for degradation
of components that originate from both the inside of the cell
(autophagy) and from the plasma membrane or extracellular
media (heterophagy). Because of the focus of this chapter on
intracellular protein degradation, only a brief description of the
heterophagic pathways is presented in this section. Interested
readers are referred to recent reviews on endocytosis for more
details about this fundamental process [69–75].
Heterophagic pathways
Hepatocytes, like any other cell type, require continuous sam-
pling of their surroundings in order to accommodate changes in
the extracellular environment. This constant exchange of infor-
mation is attained through endocytosis, which often leads to
activation of complex signaling mechanisms originating from
the plasma membrane or from the membrane of the internalized
vesicular compartments (endosomes) [76, 77]. The extracellular
components, or cargo, can be internalized through different
endocytic mechanisms, which mainly differ in the manner in
which this cargo is recognized (Figure  11.3). Through fluid‐
phase endocytosis or pinocytosis, small portions of the extracel-
lular media, including mainly soluble macromolecules and
micronutrients as well as small particles, are continuously inter-
nalized into small vesicles. The capability of this process is rela-
tively low but because of its constant nature it can result in the
internalization of as much as 30% of the plasma membrane per
minute. This forces a coordinate balance between the endocytic
process and the hepatocyte secretory mechanisms that will
return most of this membrane back to the cell surface. Fluid‐
phase endocytosis is a nonsaturable process with relatively low
capacity, used often for internalization of molecules such as
sucrose [78, 79]. One further step in the selectivity and effi-
ciency of cargo internalization is achieved through a second
type of endocytosis known as absorptive endocytosis [80]. In
this case, the internalized molecules interact weakly with the
plasma membrane before being endocytosed, resulting in some
degree of cargo concentration in the particular region of the
plasma membrane that undergoes invagination. In contrast to
fluid‐phase endocytosis, absorptive endocytosis is saturable as it
depends on the amount of cellular surface available for cargo
interaction, but the concentration step makes it about 100 times
more efficient than fluid‐phase endocytosis [81]. Many proteins
in the blood are internalized in hepatocytes through this mecha-
nism. The maximal example of selectivity and efficiency in
cargo internalization occurs through receptor‐mediated endocy-
tosis, where cargo molecules first bind directly to particular
receptor proteins at the plasma membrane, which triggers a
series of events leading to the concentration of receptor cargo
in particular regions of the plasma membrane, internalization
in vesicles that form through a complex process of adaptor protein
assembly, and their delivery to lysosomes through a regulated
 11:  Protein Degradation and the Lysosomal System 127

Fluid phase Absorption Receptor-mediated

Cargo interacts
Cargo binds to
Cargo in solution weakly with the
specific receptors
plasma membrane

Nonsaturable Saturable Saturable

Low capacity 2-100 times higher 106-109 higher

Hepatocytes Hepatocytes Hepatocytes

Sucrose Aldolase Insulin-like growth factor

Figure 11.3  Types of endocytosis. Hepatocytes, like most cells in the organism, continuously sample and receive contents and information from
the extracellular environment via endocytosis. The three types of endocytosis simultaneously active in hepatocytes are displayed here. Left:
Fractions of the extracellular media are internalized in a nonselective manner through fluid‐phase endocytosis, a very low‐capability endocytic type.
Middle: The interaction of extracellular components with the membrane prior to internalization explains the higher capability of absorptive endo-
cytosis. This type of endocytosis can be saturated once all the surface of the cell membrane is occupied by interacting molecules. Right: Receptor‐
mediated endocytosis is the most selective type of endocytosis, in which extracellular components interact with particular receptors at the cell
surface. The high capability of this pathway is attained through the concentration of the cargo‐loaded receptor proteins in particular regions of the
plasma membrane where internalization occurs. Examples of extracellular molecules internalized by each pathway are depicted at the bottom.

series of vesicular fusion and fission events [75, 77, 82–85].
Like absorptive endocytosis, receptor‐mediated endocytosis is a
saturable process, but its capability is 106–109 times higher than
the former.
Independent of the mechanisms that mediate internalization,
lysosomes constitute the final destination of the cargo‐contain-
ing vesicles. These endocytic vesicles, or endosomes, serve also
as sorting compartments for molecules such as the receptors that
are spared from degradation and are instead delivered back to
the plasma membrane to continue their function [75]. Endosomes
mature into acidic lysosomes through fusion and fission events
regulated by pairs of proteins located in the membranes of both
acceptor and donor vesicles [83]. Once a certain degree of acidi-
fication is attained, recycling is no longer possible and the cargo
is completely degraded by the lysosomal hydrolases. Readers
are referred to these comprehensive reviews for a detailed
description of the molecular components that participate in
endocytosis [69, 70, 77, 83, 84].
Lastly, macrophagic cells in the liver such as Kupffer cells
are capable of internalization of extracellular pathogens and
other large particulate components by phagocytosis, a special-
ized form of heterophagy [83, 84]. In contrast to the discrete
invaginations of the plasma membrane characteristic of endocy-
tosis, during phagocytosis these professional defense cells
undergo major deformity of their plasma membrane and
cytoskeletal rearrangements to assure engulfment and internali-
zation of the oversized cargo. Phagocytic vesicles or phago-
somes fuse with secondary lysosomes through similar
mechanisms to the ones followed by endocytic vesicles for com-
plete degradation of their cargo. Readers are referred to reviews
and other chapters in this book addressing the importance of the
phagocytic activity of Kupffer cells in chronic liver inflamma-
tion and in the innate immune response [85–87].
Autophagic pathways
The term autophagy is commonly reserved for the group of dif-
ferent pathways that mediates the lysosomal degradation of
components present inside cells, including all soluble proteins,
organelles, cellular subcompartments, and particulate protein
deposits. However, autophagic cargo does not always belong to
the cell, as extracellular components internalized by phagocyto-
sis can also be the target of degradation by the autophagic
­system through a process described as xenophagy (to point out
the exogenous nature of the cargo) [88, 89].
Although autophagic degradation has been known since the
early discovery of lysosomes and its morphological features
and hormonal regulation have been extensively analyzed using
liver as a study model, a complete molecular dissection of some
of the autophagic pathways was obtained only in the last 15
years. The main drive for this “rediscovery” of autophagy has
been the three different mutagenesis screenings in yeast, which
led to the identification of more than 35 genes generically
known as autophagy‐related genes (ATG) [90–92]. Genetic
manipulations of these genes  –  knockouts, knockdowns, and
overexpressions – have helped in identifying novel physiologi-
cal roles for autophagy and have linked dysfunction of this
lysosomal pathway with prominent human diseases such as
cancers, neurodegenerations, myopathies, and different metabolic
disorders [93–96].
128 THE LIVER:  THE LYSOSOMAL SYSTEM

Three major types of autophagy have been described in
liver and in almost all types of mammalian cells: macroau-
tophagy, microautophagy, and chaperone‐mediated autophagy
(CMA) [1, 93, 97] (Figure  11.4). Since the molecular
mechanisms and physiological relevance of each autophagic
pathway are different, the following sections provide sep-
arate reviews of the major advances in the current under-
standing of each.
Macroautophagy
The original description of a type of nonselective degradation in
which whole regions of the cytosol sequestered into double‐
membrane vesicles were delivered to lysosomes upon fusion of
the two vesicular compartments gave rise to the term “macroau-
tophagy,” to differentiate it from a smaller scale degradation
resulting from small invaginations in the lysosomal membrane,
or “microautophagy” [8, 9] (Figure  11.4). Macroautophagy is
quantitatively the most important form of autophagy and also
the best characterized as a result of the recent identification of
the ATG genes. The protein products of these genes organize in
four major groups: two conjugation cascades, an initiation com-
plex, and a negative regulatory complex [1, 98]. The conjuga-
tion events are required for the formation of the limiting
membrane that elongates and seals itself, forming a double‐
membrane vesicle or autophagosome [99] (Figure  11.5).
Elongation occurs through a complex conjugation of a protein
(Atg5) to another protein (Atg12) and of a protein (Atg8 or LC3
in mammals) to a lipid (phosphatidylethanolamine) [100]. The
series of events required for conjugation and the similarity of
the enzymes that modulate this process to E ligases make
autophagic conjugation resemble, to some extent, the tagging of
proteasome substrates by ubiquitination. Activation of conjuga-
tion, required for the formation of the autophagosome, occurs
through the activation complex. The main components of this
complex are Atg6/beclin‐1, a member of the phosphatidylinosi-
tol‐3‐kinase type III (PI3K‐III), and several other exchangeable
proteins that allow this complex to regulate both autophagic and
endocytic pathways [101].
Although levels of the components of the initiation complex
were initially proposed as good markers for autophagic activity,
changes in the stoichiometry of their interactions, as well as in
the intracellular location of these complexes, actually determine
macroautophagy activity [102]. Once the autophagosome is
formed, sequestering inside the cytosolic cargo, it is delivered to
lysosomes in a microtubule‐dependent manner. Both vesicles
fuse, allowing lysosomal enzymes to gain access to the cargo
(Figure  11.5). Although most of the autophagosomes fuse
directly to secondary lysosomes, forming autophagolysosomes,
in almost all cells direct fusion of autophagosomes to late
endosomes, forming amphisomes, has also been described
[103–105]. This interaction of the autophagic and endocytic
pathways does not usually contribute much to the autophagic
degradation but, under certain conditions when autophago-
some–lysosome fusion is impaired, it becomes the main route
for autophagosome clearance.

Chaperone-mediated autophagy
Microautophagy

LAMP-2A

Macroautophagy

CMA substrate-
chaperone
Lysosome

Autophagosome

Limiting membrane

Figure 11.4  Autophagic pathways in liver. Three major types of autophagy coexist in hepatocytes and almost all mammalian cells: macroau-
tophagy, microautophagy, and chaperone‐mediated autophagy (CMA). In macroautophagy, whole regions of the cytosol, including organelles and
soluble proteins, are sequestered by a limiting membrane that seals to form an autophagosome. This double‐membrane vesicle delivers cargo to
lysosomes for degradation through vesicle‐to‐vesicle fusion. In microautophagy, whole regions of the cytosol are internalized into the lysosomal
lumen in invaginations or projections of the lysosomal membrane that seal to form single‐membrane tubules and vesicles. CMA allows selective
degradation of soluble cytosolic proteins. Substrates are identified by a chaperone/co‐chaperone complex that delivers them to the membrane of the
lysosome where they interact with a receptor protein, the lysosome‐associated membrane protein type 2A (LAMP‐2A). Substrate proteins unfold
and are translocated into the lumen, assisted by a luminal chaperone.
 11:  Protein Degradation and the Lysosomal System 129

(a)

(b) (c)

Figure 11.5  Ultrastructure of the autophagic system. (a) Electron micrographs of sections of mouse livers after 6 hours of starvation. Autophagic
vacuoles – autophagosomes and autophagolysosomes – are indicated with arrows. Bottom panels show details of different stages of maturation of
autophagic vacuoles, from early or immature vesicles (left) when cargo is still recognizable, to late or mature vesicles (right) where only degraded
material and membranous structures are observed in the lumen. (b, c) Electron micrographs of lysosomal fractions isolated from the same livers.
Autophagic vacuoles (b) and secondary lysosomes (c) are shown. Insets show representative examples of the type of autophagic/lysosome structure
present in each fraction.

The fourth subset of Atg proteins acts as a negative regulator
of macroautophagy and has as its central component mTOR,
one of the major nutrient‐sensing intracellular kinases [106–
108]. mTOR integrates the information received through the
insulin‐signaling pathway and the amino acid receptors, which
informs this kinase of the energetic status of the cell. In the pres-
ence of nutrients or energy excess, mTOR is activated to exert a
negative effect on macroautophagy. However, when nutrients
are sparse, mTOR is inactivated, resulting in activation of mac-
roautophagy. The downstream effectors of mTOR on the
autophagic pathway are well characterized in yeast, but their
mammalian homologs are yet to be identified.
The hormonal regulation of macroautophagy in liver was
well characterized even before the molecular components of
this pathway were identified. The insulin/glucagon balance in
blood has a direct effect on macroautophagic activity in liver [9, 11,
109]. In the postprandial period, the high levels of insulin circu-
lating in blood repress activation of macroautophagy, whereas
the gradual increase in circulating levels of glucagon, as the
time after nutrient consumption progresses, results in activation
130 THE LIVER:  THE LYSOSOMAL SYSTEM
of macroautophagy to provide the energy and the essential com-
ponents required for the synthesis of proteins necessary during
the nutritional stress.
Besides an alternative source for cellular energy, macroau-
tophagy is also a major component of the systems responsible
for intracellular quality control [1]. Although selective removal
of soluble proteins cannot be attained through this pathway,
macroautophagy contributes to the continuous turnover of orga-
nelles, proteins adopting irreversible insoluble conformations
(such as protein inclusions and aggregates), and it is at the fore-
front of the cellular response to organelle stress, from ER stress
to mitochondrial or Golgi stress [110–112]. In contrast to the
lack of selectivity of this pathway for soluble cytosolic proteins,
removal of organelles or particulate structures from the cytosol
via macroautophagy is a discriminatory process. Thus, only
those structures identified as nonfunctional or damaged are
sequestered inside autophagosomes for their lysosomal deliv-
ery. Cargo recognition is mediated by a family of soluble
autophagy receptors (i.e. p62, optineurin, NBR2, Nix) that rec-
ognize specific “marks” in the substrate (i.e. ubiquitination,
exposure of specific proteins in the surface of the organelle,
etc.) [111, 113].
Although the mechanisms mediating this selective recogni-
tion are under intensive investigation, the importance of macro-
autophagy in quality control has been underscored as a result of
studies with conditional mouse model knockout for macroau-
tophagy genes in different tissues. Conditional knockout of the
essential macroautophagy gene ATG7 in liver [114], the first
established model of this type, revealed, as expected, a lower
ability of the liver of these animals to accommodate proteolysis
rates to the cellular energetic needs during nutrient deprivation.
Unexpectedly, damaged organelles and protein inclusions accu-
mulate in the livers of these animals even when maintained under
normal nutritional conditions, supporting a critical role for mac-
roautophagy in the maintenance of cellular homeostasis through
the continuous removal of intracellular components [114].
Similar consequences have been observed upon conditional
blockage of macroautophagy in other tissues and organs [115–
117]. This subset of studies underscores a prevalent role for the
basal activity of macroautophagy, previously classified as only a
stress‐induced pathway. Whether or not basal and inducible mac-
roautophagy have similar molecular requirements and regulatory
mechanisms is currently under intensive investigation.
The high degrading capability of macroautophagy is the
basis for its participation in processes requiring major cellular
remodeling, such as embryogenesis, development, cell differ-
entiation, wound healing, and tissue regeneration [93, 96,
118]. In addition, as explained above, macroautophagic
sequestration of pathogens, free in the cytosol after escaping
the endocytic system or still contained in compartments of the
endocytic and phagocytic pathway (xenophagy), confers mac-
roautophagy an important role in cellular defense against
external aggressors [119].
The wide variety of functions attributable to macroau-
tophagy explains the negative consequences of the malfunc-
tioning of this pathway and its association with different
human pathologies. Recent connections established between
macroautophagy failure and liver diseases will be reviewed in
later sections.
Microautophagy
Described initially as a constitutive mechanism for cytosolic
content turnover in lysosomes in liver, microautophagy is
still the less understood form of autophagy. The classic defi-
nition of this autophagic pathway resulted from the morpho-
logical observation of secondary lysosomes containing
multiple vesicular structures in their lumen filled with cyto-
solic content [8, 120] (Figure 11.4). Further studies, mainly
in yeast, have demonstrated that the vacuole, equivalent to
the lysosome in yeast, can trap whole regions of cytosol,
including both soluble proteins and organelles, through
invaginations, tubulations, and projections from this com-
partment’s membrane. This uptake of cytosolic content has
been reproduced recently in vitro using isolated yeast vacu-
oles [121, 122]. These studies have revealed that deforma-
tion of the membrane requires cytoskeletal components and
is regulated by a subset of gene products specific to this
process. In addition, blockage of particular ATG genes also
inhibits microautophagy in yeast, suggesting that macro‐
and microautophagy share some common components.
Based on the mechanism of cargo sequestration, microau-
tophagy has been considered a nonselective type of
autophagy, but, as in the case of macroautophagy, some level
of selectivity has been described in relation to the removal
of organelles. Thus, after exposure to conditions associated
with peroxisome proliferation in yeast, microautophagy
contributes to the restoration of cellular homeostasis by
selectively removing the excess of peroxisomes [123].
Lysosomal elimination of peroxisomes has also been
described in liver after clofibrate‐induced proliferation of
these organelles [124]. Whether or not this process takes
place via microautophagy requires future elucidation.
Microautophagy of complete nuclear regions by the vacuole
has recently been described under the name “piecemeal micro-
autophagy of the nucleus” [125]. In this case, targeted interac-
tions of proteins in the nuclear envelope and in the vacuole
membrane induce the formation of a nuclear bleb that is cap-
tured, pinched off, and degraded by the vacuole. Interestingly,
genetic material is consistently excluded from the protruding
regions of the nucleus.
Most of the microautophagy‐specific genes do not seem
to be conserved in higher species, which initially limited the
study of this process in mammals. However, more recently a
microautophagy‐like process has been described first in
mammalian cells [126] and lately its presence has also been
confirmed in Drosophila [127]. An important difference
with yeast autophagy is that this process takes place in
endosomes and uses the ESCRT machinery, responsible for
the formation of multivesicular bodies in this compartment,
for internalization of substrates, hence the name of endoso-
mal microautophagy attributed to this pathway. Endosomal
microautophagy can occur in bulk, by trapping of small por-
tions of the cytosol, or through selective targeting of pro-
teins previously identified by the constitutive cytosolic
chaperone hsc70 [126, 128]. The physiological relevance of
endosomal microautophagy remains still unclear, although a
recent report proposes a role for this pathway in the rapid
response to starvation [129].
 11:  Protein Degradation and the Lysosomal System 131
Chaperone‐mediated autophagy
In contrast to the autophagic pathways described in previous
sections, in CMA, soluble cytosolic proteins are delivered to the
lysosomal lumen in a selective manner after crossing the lysoso-
mal membrane [97, 130] (Figure 11.4). The selectivity of this
pathway is conferred by a cytosolic chaperone, hsc70, the con-
stitutive member of the hsp70 family of chaperones, which
identifies in the substrate proteins a pentapeptide motif (bio-
chemically related to the pentapeptide KFERQ) and targets
them to the lysosomal surface [131]. The substrate proteins bind
there to monomeric forms of LAMP‐2A [62], thus driving the
multimerization of this receptor into the high molecular weight
protein complex required for translocation [132]. After unfold-
ing, and assisted by a chaperone located in the lysosomal lumen,
substrates cross the lysosomal membrane and are rapidly
degraded [133, 134]. Membrane levels of the lysosomal recep-
tor are limiting for this pathway and determine overall the capa-
bility of particular lysosomes to perform CMA [135].
Transcriptional changes, regulated cleavage, and membrane
subcompartmentalization of the lysosomal receptor contribute
to regulate the lysosomal levels of LAMP‐2A and consequently
the cellular CMA activity [136]. Regulation of this process
occurs in part at the lysosomal membrane through a coordinated
balance between phosphorylation and dephosphorylation events
mediated by the TORC2/Akt/PHLPP axis [137]. Other intracel-
lular signaling pathways modulating CMA activity include reti-
noic acid receptor alpha, NRF2, NFAT and humanin, a small
mitochondrial peptide [138–141].
Basal levels of CMA can be detected in almost all mamma-
lian cells, but this pathway is maximally activated under condi-
tions of stress [97, 130]. Activation of CMA fulfills two types of
cellular needs: maintenance of a supply of amino acids neces-
sary for protein synthesis under conditions of prolonged starva-
tion and selective removal of damaged proteins from the cytosol.
Although macroautophagy is activated in liver during the first
hours of starvation to provide the amino acids and other macro-
molecules required for different anabolic processes under this
condition, as starvation persists, macroautophagic activity
decreases concomitant with a progressive increase in CMA
activity.
Activation of CMA late in starvation likely allows selective
degradation of non‐essential proteins to provide amino acids for
the synthesis of proteins essential under this stress condition.
CMA is also upregulated during mild oxidative stress and after
exposure of cells to toxic compounds known to induce irrevers-
ible conformational changes in proteins [142]. Altered proteins
can thus be eliminated selectively via CMA without affecting
nearby functional proteins.
Blockage of CMA in cultured cells does not result in major
changes in cellular homeostasis under basal conditions, as pro-
teolysis rates are maintained by a compensatory activation of
macroautophagy [143]. However, although these pathways can
compensate for one another, they are not redundant, and the lack
of CMA becomes noticeable during stress conditions. Exposure
of cells with impaired CMA to different types of stressors
reveals their higher sensitivity to stress and a marked increase in
apoptotic cell death, despite macroautophagy being perfectly
functional [143]. Similar compensation of macroautophagy was
observed in mice with hepatic CMA blockage [144]. These
results support the theory that the selectivity of CMA might be
important under these conditions.
PROTEIN DEGRADATION IN LIVER
DISEASE AND AGING
The intrinsic high metabolic activity of the liver makes fast
adaptation to changes in nutritional conditions and tight quality
control indispensable for its proper functioning. Consequently,
deregulation of the major proteolytic systems described in this
chapter results inevitably in altered liver function and has been
shown to contribute to the pathogenesis of common liver dis-
eases (Figure 11.6) [95, 96].
Altered proteasome function contributes, at least in part, to
the formation of Mallory bodies, one of the best‐studied types of
hepatic protein inclusion, observed in diverse chronic liver dis-
eases such as alcoholic and non‐alcoholic steatohepatitis,
chronic cholestasis, metabolic disorders, and hepatocellular
neoplasms [145]. Direct inhibition of the proteasome by differ-
ent agents, such as ethanol in the case of alcoholic steatohepati-
tis, favors the accumulation of cytokeratins, different
proapoptotic factors, and negative regulators of cytokine signal-
ing, which contribute to the chronic disease [146].
Dysfunction of both the proteasome and the autophagic sys-
tem has been described in α‐1‐antitrypsin deficiency, a protein
conformational disorder resulting from point mutations in a
secretory liver protein that accumulates in hepatocytes, lead-
ing to chronic liver inflammation and carcinogenesis [147].
The soluble form of the mutant protein is usually degraded by
the proteasome, whereas the semi‐aggregate and aggregate
forms that accumulate in the ER depend on macroautophagy
for their removal [148]. Furthermore, macroautophagy also
contributes to the removal of the altered mitochondrion abun-
dant in the α‐1‐antitrypsin‐deficient liver [149]. A direct con-
nection between the carcinogenesis associated with this
disorder and autophagy has not been established yet. However,
there is extensive evidence for an involvement of autophagy in
hepatocarcinogenesis [150]. Hepatic macroautophagy has
been proposed to have anti‐oncogenic functions, because
reduced levels of essential autophagy genes triggers tumori-
genesis in liver and a decrease in autophagy‐related proteins is
also common in hepatocarcinoma patients and is associated
with poor prognosis [151].
Because reduced protein degradation and upregulation of
protein synthesis are favorable conditions for rapid cell division,
induction of macroautophagy in cancer cells often slows down
their replication. However, macroautophagy is not completely
shut down in these cells because it still constitutes a major
defense against adverse conditions, such as the poor nutritional
environment of the center of hypovascularized tumors or
the damage induced by anti‐oncogenic treatments [119].
Furthermore, growing evidence supports that macroautophagy
can also be utilized by hepatic cancer cells for tumor pro-
gression and an increase in autophagy markers has also been
associated with poor prognosis and higher rates of recurrence
after surgery [152]. Modulators of macroautophagy are
132 THE LIVER:  PROTEIN DEGRADATION IN LIVER DISEASE AND AGING

Alpha-1-antitrypsin deficiency HCV Hepatitis C virus

• Soluble protein degraded by • Promotes ER stress

proteasome ER • Blocks autophagosome
• Aggregates in the ER lysosome fusion
degraded by macroautophagy
• Altered mitochondria

Fatty liver LYS

• Lipid droplet coat proteins, MIT

Cancer
enzymes of lipogenesis are
degraded by CMA Decreased macroautophagy
• Lipid droplets degraded by in many cancers
macroautophagy LD AV
• Chronic lipid stimulus
inhibits macroautophagy
and CMA leading to fatty
liver Aging

• Decreased macroautophagy
• Decreased CMA

Figure 11.6  Autophagy and liver disease. Recent studies have revealed numerous connections between autophagy malfunctioning and liver
pathology. Examples of some of these liver disorders and the changes in the autophagic system under these pathological conditions are shown here.
In certain pathologies, for example α‐1‐antitrypsin deficiency, the soluble protein is degraded by the proteasome, while once the protein forms
aggregates – in this case in the ER – macroautophagy is more adept at degrading these. Sustained lipid challenge leads to blockage of macroau-
tophagy and CMA, which normally degrade the lipid stores, resulting in fatty liver disease. Decreased activity of the proteolytic systems is respon-
sible, in part, for the adverse abnormalities observed in aging liver, while it is beneficial for the cancer cells. Bacteria and viruses subvert the
autophagy machinery for their own replication. AV, autophagic vacuole; ER, endoplasmic reticulum; HCV, hepatitis C virus; LD, lipid droplet;
LYS, lysosomes; MIT, mitochondria.

currently being considered as possible enhancers of anticancer
therapies [153, 154].
As in other cells, macroautophagy also plays a defensive role
against liver pathogens. Bacteria and viruses are internalized
and degraded in lysosomes. However, some pathogens can
develop ways to overcome the lysosomal system. For example,
infection by the hepatitis C virus has been shown to block mac-
roautophagy, but not at the level of autophagosome formation
[155]. Instead, this virus induces ER stress, which promotes
­formation of a large number of autophagosomes but at the same
time blocks their fusion with lysosomes. Blockage of autophago-
some formation prevents virus replication, supporting the t­ heory
that the virus might use these double‐membrane vesicles for
assembly, as previously described for other viruses in other
cell types.
Connections between liver disease and alterations in
autophagy are not limited to the role of this pathway in quality
control, but instead extend to the function of this catabolic pro-
cess in the maintenance of the cellular energetic balance.
Studies support the idea that a defect in macroautophagy could
contribute to the pathogenesis of fatty liver [156–158]. Thus,
basal macroautophagy activity has been demonstrated to con-
tribute to the regulation of lipid stores in all cells.
Macroautophagy of lipid droplets or lipophagy is particularly
important in liver, as this is the organ that receives the highest
content of circulating lipids from lipolysis in the adipose tis-
sue. However, a maintained lipogenic challenge, such as that
induced by high‐fat diet, has a negative effect on liver macro-
autophagy [156]. The inability of the liver under these con-
ditions to reduce the size of the lipid stores through
macroautophagic degradation results in abnormal growth of
lipid droplets inside hepatocytes, initiating the progression
toward non‐alcoholic fatty liver. It is also likely that defective
macroautophagy as a result of insulin resistance in conditions
such as aging may be the basis of fatty liver being associated
with the metabolic syndrome of aging. For lipophagy to occur,
the lipid droplet coat proteins need to be removed first from
the surface of the lipid droplets. This selective removal is per-
formed by CMA [159]. Consequently, liver‐specific CMA‐
null mice present marked steatosis, and derangements in lipid
metabolism. Several lipogenesis enzymes and lipid carrier
proteins are also CMA substrates, explaining the severe phe-
notype of the CMA‐deficient mice [144]. Similar to the effect
on macroautophagy, acute lipid challenge activates CMA
whereas excess dietary fat inhibits this pathway [160], high-
lighting the reciprocal modulatory effect of lipids on
autophagy, but excess lipids overwhelming and inhibiting the
pathways, mimicking scenarios observed with aging.
 11:  Protein Degradation and the Lysosomal System 133
A decrease with age in total rates of protein degradation
has been described in liver of almost all organisms analyzed,
supporting a negative effect of aging on the intracellular pro-
teolytic systems [161]. The degradation of short‐lived pro-
teins, the main substrate of the UPS, is better preserved in old
livers, suggesting that the impairment of this system is less
pronounced than that of the lysosomal system. In fact, early
studies in livers from young and old rodents showed no net
differences in proteasome‐dependent degradation [162, 163].
Moreover, the analysis of the three different proteolytic activ-
ities of the 20S proteasome revealed a decrease in some of its
activities but an increase in others, suggesting that rather than
dramatic quantitative changes, qualitative changes account
for most of the altered degradation of proteasome substrates
with age [162–164]. Changes in the 20S proteolytic activities with
age may result in part from changes in their subunit composi-
tion [165]. Proteomic analyses have revealed the coexistence
of different subpopulations of proteasomes as cells age.
Defects in ubiquitination steps or in the complex processes
that mediate recognition of ubiquitinated proteins by the pro-
teasome might also occur with age, but there is still poor
understanding of these changes.
Age‐related changes in the liver lysosomal system at the
morphological level have been extensively reported in the lit-
erature. Expansion of the lysosomal vacuolar compartment
and accumulation of undigested products inside lysosomes in
the form of an autofluorescent pigment known as lipofuscin
are often used as typical biomarkers of aging [166]. Accumulation
of lipofuscin originates from changes in the proteolytic
susceptibility of the intracellular components and from defects
in autophagy [167]. A reduction in the induction of macroau-
tophagy in the period in between meals, as well as problems
with the clearance of the already‐formed AVs by lysosomes,
seem to be behind the failure of this autophagic pathway with
age. Because of the important role of this autophagic pathway
in the maintenance of organelle homeostasis, the decline with
age in macroautophagy has been proposed to be behind the
increased number of dysfunctional mitochondria with age; this
organelle is tightly linked to aging [167]. Defective macroau-
tophagy induction results from the negative effect on this path-
way of the constitutive signaling through the insulin receptor,
even in the absence of insulin, characteristic of the metabolic
syndrome of age [168, 169]. Although the molecular basis of
this defect remains unknown, which prevents any targeted
restorative effort over macroautophagy, certain interventions
have proven beneficial in the prevention of the age‐dependent
macroautophagic decline. Thus, caloric restriction, the only
intervention known to slow down aging, anti‐lipolytic drugs,
and intermittent fasting preserve macroautophagy activity in
livers of old rodents at levels comparable to those in young
livers [169–171].
A decrease in CMA activity has been well characterized in
livers of old rodents [135]. The main reason for the CMA
decline is a reduction in the lysosomal levels of the LAMP‐2A
receptor with age. Reduced levels of LAMP‐2A are not caused
by a transcriptional decrease, defects in translation, or altered
trafficking to lysosome. Instead, low lysosomal levels of
LAMP‐2A originate from increased instability of this protein
once it reaches the lysosomes [172]. Changes in the lipid
composition of the lysosomal membrane seem to be behind the
reduced stability of LAMP‐2A. This decrease is initially com-
pensated for by an increase in the amount of lysosomal chaper-
one that promotes translocation across the membrane but at late
stages this compensation is not enough and CMA activity is
severely impaired. In CMA‐null mice, the absence of CMA is
initially compensated by upregulation of macroautophagy;
however, with age, it is not sufficient for maintaining the proteo-
stasis in response to different stressors [173]. Studies in a trans-
genic mouse model in which the decrease in levels of LAMP‐2A
in liver was prevented demonstrated that old mice with pre-
served CMA activity contain lower levels of oxidized and
aggregate proteins than the livers of their wild‐type littermates
[174]. They also show improved ability to respond to stress and
preserved liver function until late in life. These studies support
the theory that a decline in CMA activity with age could be
responsible, at least in part, for the accumulation of damaged
proteins with age, lower resistance to stress, and functional
decline in this organ [174]. On a positive note, restoration of
only one of the different proteolytic systems seems enough to
induce a major improvement, probably because of the existing
cross‐talk among these systems. As more insights are gained
into the reasons for the functional decrease in other proteolytic
pathways and as further attempts to correct these defects are
undertaken, it may be possible to prevent other age‐related
changes in liver and utilize them in the fight against diseases of
the aging liver.
CONCLUSION
The proteolytic systems play a pivotal role in the maintenance
of liver homeostasis and its energetic balance, and in the fight
against intracellular and extracellular aggressors. The differ-
ent mechanisms for protein degradation can no longer be con-
sidered as independent units. Increasing numbers of reports
support the existence of continuous cross‐talk among the dif-
ferent autophagic pathways and with the proteasome system.
Compensation of one system for another is not complete but
it is often enough to preserve cellular homeostasis at least
under basal nonstressed conditions. Failure of the proteolytic
systems has been identified as the basis of the pathogenesis of
different liver diseases. The recent advances in our under-
standing of the molecular mechanisms behind the different
intracellular proteolytic systems, their regulation, and the cel-
lular consequences of their blockage or altered function could
set the basis for novel therapeutic approaches for the treat-
ment of liver diseases.
ACKNOWLEDGMENTS
We thank the numerous colleagues in the field of autophagy
who through their animated discussions have helped shape this
chapter. Work in our laboratory is supported by NIH/NIA
(AG021904, AG031782), NIH/NIDDK (DK098408), and the
generous support of Robert and Renée Belfer.
134 THE LIVER: REFERENCES

REFERENCES 31. Chen, Y., Zhang, Y., and Guo, X. Proteasome dysregulation in human cancer:
implications for clinical therapies. Cancer Metastasis Rev, 2017;36(4):
703–16.
1. Mizushima, N., Levine, B., Cuervo, A., and Klionsky D. Autophagy fights
32. Barac, Y.D., Emrich, F., Krutzwakd‐Josefson, E. et al. The ubiquitin‐proteasome
disease through cellular self‐digestion. Nature, 2008;451:1069–75.
system: a potential therapeutic target for heart failure. J Heart Lung
2. Ciechanover, A. Intracellular protein degradation: From a vague idea thru the
Transplant, 2017;36(7):708–14.
lysosome and the ubiquitin‐proteasome system and onto human diseases and
33. Gilda, J.E. and Gomes, A.V. Proteasome dysfunction in cardiomyopa-
drug targeting. Best Pract Res Clin Haematol, 2017;30(4):341–55.
thies. J Physiol, 2017;595(12):4051–71.
3. Goldberg, A.L. Protein degradation and protection against misfolded or
34. Bonet‐Costa, V., Pomatto, L.C., and Davies, K.J. The proteasome and oxida-
damaged proteins. Nature, 2003;18:895–9.
tive stress in Alzheimer’s disease. Antioxid Redox Signal, 2016;25(16):
4. Singh, R. and Cuervo, A.M. Autophagy in the cellular energetic balance.
886–901.
Cell Metab, 2011;13:495–504.
35. Kammerl, I.E. and Meiners, S. Proteasome function shapes innate and adap-
5. Di Bartolomeo, S., Nazio, F., and Cecconi, F. The role of autophagy during
tive immune responses. Am J Physiol Lung Cell Mol Physiol, 2016;311(2):
development in higher eukaryotes. Traffic, 2010;11(10):1280–9.
L328–36.
6. Cuervo, A.M. and Macian, F. Autophagy and the immune function in aging.
36. Vigneron, N., Abi Habib, J., and Van den Eynde, B.J. Learning from the
Curr Opin Immunol, 2014;29C:97–104.
proteasome how to fine‐tune cancer immunotherapy. Trends Cancer,
7. Mizushima, N. and Levine, B. Autophagy in mammalian development and
2017;3(10):726–41.
differentiation. Nat Cell Biol, 2010;12(9):823–30.
37. Manasanch, E.E. and Orlowski, R.Z. Proteasome inhibitors in cancer ther-
8. Ahlberg, J. and Glaumann, H. Uptake – microautophagy – and degradation
apy. Nat Rev Clin Oncol, 2017;14(7):417–33.
of exogenous proteins by isolated rat liver lysosomes. Effects of pH, ATP,
38. Novikoff, A., Essner, E., and Quintana, N. Golgi apparatus and lysosomes.
and inhibitors of proteolysis. Exp Mol Pathol, 1985;42:78–88.
Fed Proc, 1964;23:1010–22.
9. Deter, R.L., Baudhuin, P., and De Duve, C. Participation of lysosomes in
39. Maggi, V. and Hart, K. Lysosomes and lysosomal enzymes in hearts of ham-
cellular autophagy induced in rat liver by glucagon. J Cell Biol, 1967;35(2):
sters (BIO 14.6 and BBL x7) with congenital cardiomyopathy. Recent Adv
C11–16.
Stud Cardiac Struct Metab, 1973;3:489–95.
10. Marzella, L., Ahlberg, J., and Glaumann, H. Isolation of autophagic vacuoles
40. Jefferies, K.C., Cipriano, D.J., and Forgac, M. Function, structure and regu-
from rat liver: morphological and biochemical characterization. J Cell Biol,
lation of the vacuolar (H+)‐ATPases. Arch Biochem Biophys, 2008;476(1):
1982;93(1):144–54.
33–42.
11. Mortimore, G.E. and Poso, A.R. Intracellular protein catabolism and its con-
41. Braulke, T. and Bonifacino, J.S. Sorting of lysosomal proteins. Biochim
trol during nutrient deprivation and supply. Annu Rev Nutr, 1987;7:539–64.
Biophys Acta, 2009;1793(4):605–14.
12. Mortimore, G.E. and Ward, W.F. Behavior of the lysosomal system during
42. Ghosh, P., Dahms, N., and Kornfeld, S. Mannose 6‐phosphate receptors: new
organ perfusion. An inquiry into the mechanism of hepatic proteolysis. Front
twists in the tale. Nat Rev Mol Cell Biol, 2003;4:202–12.
Biol, 1976;45:157–84.
43. Obermajer, N., Jevnikar, Z., Doljak, B., and Kos, J. Role of cysteine cath-
13. Dice, J. Lysosomal Pathways of Protein Degradation. Landes Bioscience,
epsins in matrix degradation and cell signalling. Connect Tissue Res,
Austin, TX, 2000.
2008;49:193–6.
14. Ciechanover, A. and Stanhill, A. The complexity of recognition of ubiquit-
44. Reczek, D., Schwake, M., Schröder, J. et al. LIMP‐2 is a receptor for lysoso-
inated substrates by the 26S proteasome. Biochim Biophys Acta,
mal mannose‐6‐phosphate‐independent targeting of beta‐glucocerebrosi-
2014;1843(1):86–96.
dase. Cell, 2007;131(4):770–83.
15. Kato, K. and Satoh, T. Structural insights on the dynamics of proteasome
45. Pillay, C.S., Elliott, E., and Dennison, C. Endolysosomal proteolysis and its
formation. Biophys Rev, 2018;10(2):597–604.
regulation. Biochem J, 2002;363(Pt 3):417–29.
16. Pickart, C. and Cohen, R. Proteasomes and their kin: proteases in the
46. Zavasnik‐Bergant, T. and Turk, B. Cysteine proteases: destruction ability
machine age. Nat Rev Mol Cell Biol, 2004;5:177–87.
versus immunomodulation capacity in immune cells. Biol Chem. 2007;388:
17. Rosenzweig, R. and Glickman, M.H. Chaperone‐driven proteasome assem-
1141–9.
bly. Biochem Soc Trans, 2008;36(Pt 5):807–12.
47. Yorimitsu, T., Nair, U., Yang, Z., and Klionsky, D.J. Endoplasmic reticulum
18. Schmidt, M., Hanna, J., Elsasser, S., and Finley, D. Proteasome‐associated
stress triggers autophagy. J Biol Chem, 2006;281(40):30299–304.
proteins: regulation of a proteolytic machine. Biol Chem, 2005;386(8):
48. Dell’Angelica, E.C., Mullins, C., Caplan, S., and Bonifacino, J.S. Lysosome‐
725–37.
related organelles. FASEB J, 2000;14(10):1265–78.
19. Rousseau, A. and Bertolotti, A. Regulation of proteasome assembly and
49. Arunachalam, B., Phan, U.T., Geuze, H.J., and Cresswell, P. Enzymatic
activity in health and disease. Nat Rev Mol Cell Biol, 2018;19:697–712.
reduction of disulfide bonds in lysosomes: characterization of a gamma‐
20. Zhou, P. Targeted protein degradation. Curr Opin Chem Biol,
interferon‐inducible lysosomal thiol reductase (GILT). Proc Natl Acad Sci U
2005;9(1):51–5.
S A, 2000;97(2):745–50.
21. Huibregtse, J.M. UPS shipping and handling. Cell, 2005;120(1):2–4.
50. Settembre, C., Fraldi, A., Jahreiss, L. et al. A block of autophagy in lysoso-
22. Schmidt, M., Lupas, A.N., and Finley, D. Structure and mechanism of ATP‐
mal storage disorders. Hum Mol Genet, 2008;17(1):119–29.
dependent proteases. Curr Opin Chem Biol, 1999;3(5):584–91.
51. Platt, F.M. Emptying the stores: lysosomal diseases and therapeutic strate-
23. Myeku, N. and Duff, K.E. Targeting the 26S proteasome to protect against
gies. Nat Rev Drug Discov, 2018;17(2):133–50.
proteotoxic diseases. Trends Mol Med, 2018;24(1):18–29.
52. Ferreira, C.R. and Gahl, W.A. Lysosomal storage diseases. Transl Sci Rare
24. Ravid, T. and Hochstrasser, M. Diversity of degradation signals in the ubiquitin‐
Dis, 2017;2(1–2):1–71.
proteasome system. Nat Rev Mol Cell Biol, 2008;9(9):679–90.
53. Platt, F.M., d’Azzo, A., Davidson, B.L., Neufeld, E.F., and Tifft, C.J.
25. Mishra, R., Upadhyay, A., Prajapati, V.K., and Mishra, A. Proteasome‐
Lysosomal storage diseases. Nat Rev Dis Primers, 2018;4(1):27.
mediated proteostasis: novel medicinal and pharmacological strategies for
54. Ozen, H. Glycogen storage diseases: new perspectives. World J Gastroenterol,
diseases. Med Res Rev, 2018;38(6):1916–73.
2007;13:2541–53.
26. Li, W. and Ye, Y. Polyubiquitin chains: functions, structures, and mecha-
55. Andrejewski, N., Punnonen, E.L., Guhde, G. et al. Normal lysosomal mor-
nisms. Cell Mol Life Sci, 2008;65(15):2397–406.
phology and function in LAMP‐1‐deficient mice. J Biol Chem,
27. Ikeda, F. and Dikic, I. Atypical ubiquitin chains: new molecular signals.
1999;274(18):12692–701.
‘Protein Modifications: Beyond the Usual Suspects’ review series. EMBO
56. Winchester, B. Lysosomal membrane proteins. Eur J Paediatr Neurol,
Rep, 2008;9(6):536–42.
2001;5(Suppl A):11–19.
28. Goldberg, A.L. Functions of the proteasome: from protein degradation and immune
57. Tanaka, Y., Guhde, G., Suter, A. et al. Accumulation of autophagic vacuoles
surveillance to cancer therapy. Biochem Soc Trans, 2007;35(Pt 1):12–17.
and cardiomyopathy in Lamp‐2‐deficient mice. Nature, 2000;406:902–6.
29. Hanna, J. and Finley, D. A proteasome for all occasions. FEBS Lett,
58. Konecki, D., Foetisch, K., Zimmer, K., Schlotter, M., and Lichter‐Konecki,
2007;581(15):2854–61.
U. An alternatively spliced form of the human lysosome‐associated mem-
30. Wolf, D.H. and Hilt, W. The proteasome: a proteolytic nanomachine of cell
brane protein‐2 gene is expressed in a tissue‐specific manner. Biochem
regulation and waste disposal. Biochim Biophys Acta, 2004;1695(1–3):
Biophys Res Commun, 1995;215:757–67.
19–31.
 11:  Protein Degradation and the Lysosomal System 135
59. Cuervo, A. and Dice, J. A receptor for the selective uptake and degradation
of proteins by lysosomes. Science, 1996;273:501–3.
60. Nishino, I., Fu, J., Tanji, K. et  al. Primary LAMP‐2 deficiency causes X‐
linked vacuolar cardiomyopathy and myopathy (Danon disease). Nature,
2000;406:906–10.
61. Tabuchi, N., Akasaki, K., Sasaki, T., Kanda, N., and Tsuji, H. Identification
and characterization of a major lysosomal membrane glycoprotein, LGP85/
LIMP II in mouse liver. J Biochem, 1997;122(4):756–63.
62. Chou, H., Vadgama, J., and Jonas, A. Lysosomal transport of small mole-
cules. Biochem Med Metab Biol, 1992;48:179–93.
63. Kalatzis, V. and Antignac, C. New aspects of the pathogenesis of cystinosis.
Pediatr Nephrol, 2003;18:207–15.
64. Fujiwara, Y., Hase, K., Wada, K., and Kabuta, T. An RNautophagy/
DNautophagy receptor, LAMP2C, possesses an arginine‐rich motif that
mediates RNA/DNA‐binding. Biochem Biophys Res Commun, 2015;460(2):
281–6.
65. Geldner, N. and Jurgens, G. Endocytosis in signalling and development.
Curr Opin Plant Biol, 2006;9:589–94.
66. Fukuda, T., Ewan, L., Bauer, M. et  al. Dysfunction of endocytic and
autophagic pathways in a lysosomal storage disease. Ann Neurol, 2006;59(4):
700–8.
67. Kramer, H. Sorting out signals in fly endosomes. Traffic, 2002;3(2):87–91.
68. Polishchuk, E.V. and Polishchuk, R.S. The emerging role of lysosomes in
copper homeostasis. Metallomics, 2016;8(9):853–62.
69. Kaksonen, M. and Roux, A. Mechanisms of clathrin‐mediated endocytosis.
Nat Rev Mol Cell Biol, 2018;19(5):313–26.
70. Merrifield, C.J. and Kaksonen, M. Endocytic accessory factors and regula-
tion of clathrin‐mediated endocytosis. Cold Spring Harb Perspect Biol,
2014;6(11):a016733.
71. Traub, L.M. and Bonifacino, J.S. Cargo recognition in clathrin‐mediated
endocytosis. Cold Spring Harb Perspect Biol, 2013;5(11):a016790.
72. Fares, H. and Grant, B. Deciphering endocytosis in Caenorhabditis elegans.
Traffic, 2002;3(1):11–19.
73. Saksena, S., Sun, J., Chu, T., and Emr, S.D. ESCRTing proteins in the endo-
cytic pathway. Trends Biochem Sci, 2007;32(12):561–73.
74. Cullen, P.J. Endosomal sorting and signalling: an emerging role for sorting
nexins. Nat Rev Mol Cell Biol, 2008;9(7):574–82.
75. Rappoport, J.Z. Focusing on clathrin‐mediated endocytosis. Biochem J,
2008;412(3):415–23.
76. Predescu, S.A., Predescu, D.N., and Malik, A.B. Molecular determinants of
endothelial transcytosis and their role in endothelial permeability. Am J
Physiol Lung Cell Mol Physiol, 2007;293(4):L823–42.
77. Meijer, D.K., Scholtens, H.B., and Hardonk, M.J. The role of the liver in
clearance of glycoproteins from the general circulation, with special refer-
ence to intestinal alkaline phosphatase. Pharm Weekbl Sci,
1982;4(3):57–70.
78. Mukherjee, S., Ghosh, R.N., and Maxfield, F.R. Endocytosis. Physiol Rev,
1997;77:759–803.
79. Koenig, J. and Edwardson, J.M. Endocytosis and recycling of G protein‐cou-
pled receptors. Trends Pharmacol Sci, 1997;18:276–87.
80. Seaman, M.N. Endosome protein sorting: motifs and machinery. Cell Mol
Life Sci, 2008;65(18):2842–58.
81. Komada, M. Controlling receptor downregulation by ubiquitination and deu-
biquitination. Curr Drug Discov Technol, 2008;5(1):78–84.
82. Casey, C.A., Lee, S.M., Aziz‐Seible, R., and McVicker, B.L. Impaired recep-
tor‐mediated endocytosis: its role in alcohol‐induced apoptosis. J
Gastroenterol Hepatol. 2008;23(Suppl 1):S46–9.
83. Huynh, K.K., Kay, J.G., Stow, J.L., and Grinstein, S. Fusion, fission, and
secretion during phagocytosis. Physiology (Bethesda), 2007;22:366–72.
84. Gao, B., Jeong, W.I., and Tian, Z. Liver: an organ with predominant innate
immunity. Hepatology, 2008;47(2):729–36.
85. Bilzer, M., Roggel, F., and Gerbes, A.L. Role of Kupffer cells in host defense
and liver disease. Liver Int, 2006;26(10):1175–86.
86. Krenkel, O. and Tacke, F. Liver macrophages in tissue homeostasis and dis-
ease. Nat Rev Immunol, 2017;17(5):306–21.
87. Ferre, N. and Claria, J. New insights into the regulation of liver inflammation
and oxidative stress. Mini Rev Med Chem, 2006;6(12):1321–30.
88. Levine, B. and Klionsky, D. Development by self‐digestion: molecular mecha-
nisms and biological functions of autophagy. Dev Cell, 2004;6:463–77.
89. Deretic, V. Autophagosome and phagosome. Methods Mol Biol,
2008;445:1–10.
90. Levine, B. and Klionsky, D.J. Autophagy wins the 2016 Nobel Prize in
Physiology or Medicine: breakthroughs in baker’s yeast fuel advances in
biomedical research. Proc Natl Acad Sci U S A, 2017;114(2):201–5.
91. Galluzzi, L., Baehrecke, E.H., Ballabio, A. et al. Molecular definitions of
autophagy and related processes. EMBO J, 2017;36(13):1811–36.
92. Dikic, I. and Elazar, Z. Mechanism and medical implications of mamma-
lian autophagy. Nat Rev Mol Cell Biol, 2018;19(6):349–64.
93. Mizushima, N. A brief history of autophagy from cell biology to physiol-
ogy and disease. Nat Cell Biol, 2018;20(5):521–7.
94. Thorburn, A. Autophagy and disease. J Biol Chem, 2018;293(15):
5425–30.
95. Madrigal‐Matute, J. and Cuervo, A.M. Regulation of liver metabolism by
autophagy. Gastroenterology, 2016;150(2):328–39.
96. Schneider, J.L. and Cuervo, A.M. Liver autophagy: much more than just
taking out the trash. Nat Rev Gastroenterol Hepatol, 2014;11(3):187–200.
97. Kaushik, S. and Cuervo, A.M. The coming of age of chaperone‐mediated
autophagy. Nat Rev Mol Cell Biol, 2018;19(6):365–81.
98. Bento, C.F., Renna, M., Ghislat, G. et al. Mammalian autophagy: how does
it work? Annu Rev Biochem, 2016;85:685–713.
99. Ohsumi, Y. and Mizushima, N. Two ubiquitin‐like conjugation systems
essential for autophagy. Semin Cell Dev Biol, 2004;15(2):231–6.
100. Mizushima, N., Noda, T., Yoshimori, T. et al. A protein conjugation system
essential for autophagy. Nature, 1998;395(6700):395–8.
101. Liang, X., Jackson, S., Seaman, M. et al. Induction of autophagy and inhi-
bition of tumorigenesis by beclin 1. Nature, 1999;402:672–6.
102. Itakura, E., Kishi, C., Inoue, K., and Mizushima, N. Beclin 1 forms two
distinct phosphatidylinositol 3‐kinase complexes with mammalian Atg14
and UVRAG. Mol Biol Cell, 2008;19(12):5360–72.
103. Eskelinen, E.L., Illert, A.L., Tanaka, Y. et al. Role of LAMP‐2 in lysosome
biogenesis and autophagy. Mol Biol Cell, 2002;13(9):3355–68.
104. Seglen, P.O., Berg, T.O., Blankson, H. et al. Structural aspects of autophagy.
Adv Exp Med Biol, 1996;389:103–11.
105. Seglen, P.O., Gordon, P.B., and Holen, I. Non‐selective autophagy. Semin
Cell Biol, 1990;1(6):441–8.
106. Beugnet, A., Tee, A., Taylor, P., and Proud, C. Regulation of targets of
mTOR signalling by intracellular amino acid availability. Biochem J,
2003;372(Pt 2):555–66.
107. Dann, S., Selvaraj, A., and Thomas, G. mTOR Complex1‐S6K1 signaling: at the
crossroads of obesity, diabetes and cancer. Trends Mol Med, 2007;13:252–9.
108. Dennis, P.B., Jaeschke, A., Saitoh, M. et al. Mammalian TOR: a homeo-
static ATP sensor. Science, 2001;294(5544):1102–5.
109. Meijer, A.J. Amino acid regulation of autophagosome formation. Methods
Mol Biol, 2008;445:89–109.
110. Chu, C.T. Mechanisms of selective autophagy and mitophagy: implications
for neurodegenerative diseases. Neurobiol Dis, 2019;122:23–34.
111. Gatica, D., Lahiri, V., and Klionsky, D.J. Cargo recognition and degradation
by selective autophagy. Nat Cell Biol, 2018;20(3):233–42.
112. Anding, A.L. and Baehrecke, E.H. Cleaning house: selective autophagy of
organelles. Dev Cell, 2017;41(1):10–22.
113. Khaminets, A., Behl, C., and Dikic, I. Ubiquitin‐dependent and independ-
ent signals in selective autophagy. Trends Cell Biol, 2016;26(1):6–16.
114. Komatsu, M., Waguri, S., Ueno, T. et al. Impairment of starvation‐induced
and constitutive autophagy in Atg7‐deficient mice. J Cell Biol,
2005;169(3):425–34.
115. Hara, T., Nakamura, K., Matsui, M. et al. Suppression of basal autophagy
in neural cells causes neurodegenerative disease in mice. Nature,
2006;441:885–9.
116. Komatsu, M., Waguri, S., Chiba, T. et  al. Loss of autophagy in the central
nervous system causes neurodegeneration in mice. Nature, 2006;441:880–4.
117. Nakai, A., Yamaguchi, O., Takeda, T. et al. The role of autophagy in cario-
myocytes in the basal state and in response to hemodynamic stress. Nat
Med, 2007;13:619–24.
118. Mizushima, N. and Komatsu, M. Autophagy: renovation of cells and
­tissues. Cell, 2011;147(4):728–41.
119. Levine, B. Cell biology: autophagy and cancer. Nature, 2007;446(7137):745–7.
120. Marzella, L., Ahlberg, J., and Glaumann, H. Autophagy, heterophagy,
microautophagy and crinophagy as the means for intracellular degradation.
Virchows Archiv B Cell Pathol, 1981;36(2–3):219–34.
121. Uttenweiler, A., Schwarz, H., and Mayer, A. Microautophagic vacuole
invagination requires calmodulin in a Ca2+‐independent function. J Biol
Chem, 2005;280:33289–97.
136 THE LIVER: REFERENCES
122. Kunz, J., Schwarz, H., and Mayer, A. Determination of four sequential
stages during microautophagy in vitro. J Biol Chem, 2004;279:9987–96.
123. Sakai, Y., Koller, A., Rangell, L., Keller, G., and Subramani, S. Peroxisome
degradation by microautophagy in Pichia pastoris. Identification of specific
steps and morphological intermediates. J Cell Biol, 1998;141:625–36.
124. Yokota, S. Degradation of normal and proliferated peroxisomes in rat
hepatocytes: regulation of peroxisomes quantity in cells. Microsc Res Tech,
2003;61(2):151–60.
125. Roberts, P., Moshitch‐Moshkovitz, S., Kvam, E. et al. Piecemeal microautophagy
of nucleus in Saccharomyces cerevisiae. Mol Biol Cell, 2003;14:129–41.
126. Sahu, R., Kaushik, S., Clement, C.C. et  al. Microautophagy of cytosolic
proteins by late endosomes. Dev Cell, 2011;20(1):131–9.
127. Mukherjee, A., Patel, B., Koga, H., Cuervo, A.M., and Jenny, A. Selective
endosomal microautophagy is starvation‐inducible in Drosophila.
Autophagy, 2016;12(11):1984–99.
128. Tekirdag, K. and Cuervo, A.M. Chaperone‐mediated autophagy and endosomal
microautophagy: joint by a chaperone. J Biol Chem, 2018;293(15):5414–24.
129. Mejlvang, J., Olsvik, H., Svenning, S. et al. Starvation induces rapid degra-
dation of selective autophagy receptors by endosomal microautophagy. J
Cell Biol, 2018;217(10):3640–55.
130. Dice, J. Chaperone‐mediated autophagy. Autophagy, 2007;3:295–9.
131. Chiang, H., Terlecky, S., Plant, C., and Dice, J. A role for a 70 kDa heat
shock protein in lysosomal degradation of intracellular protein. Science,
1989;246:382–5.
132. Bandhyopadhyay, U., Kaushik, S., Vartikovsky, L., and Cuervo, A.M.
Dynamic organization of the receptor for chaperone‐mediated autophagy at
the lysosomal membrane. Mol Cell Biol, 2008;28:5747–63.
133. Agarraberes, F., Terlecky, S., and Dice, J. An intralysosomal hsp70 is
required for a selective pathway of lysosomal protein degradation. J Cell
Biol, 1997;137:825–34.
134. Cuervo, A.M., Dice, J.F., and Knecht, E. A population of rat liver lys-
osomes responsible for the selective uptake and degradation of cytosolic
proteins. J Biol Chem, 1997;272(9):5606–15.
135. Cuervo, A.M. and Dice, J.F. Age‐related decline in chaperone‐mediated
autophagy. J Biol Chem, 2000;275:31505–13.
136. Cuervo, A. and Dice, J. Regulation of lamp2a levels in the lysosomal
­membrane. Traffic, 2000;1:570–83.
137. Arias, E., Koga, H., Diaz, A. et  al. Lysosomal mTORC2/PHLPP1/Akt
­regulate chaperone‐mediated autophagy. Mol Cell, 2015;59(2):270–84.
138. Anguiano, J., Garner, T.P., Mahalingam, M., Das, B.C., Gavathiotis, E., and
Cuervo, A.M. Chemical modulation of chaperone‐mediated autophagy by
retinoic acid derivatives. Nat Chem Biol, 2013;9(6):374–82.
139. Gong, Z., Tasset, I., Diaz, A. et al. Humanin is an endogenous activator of
chaperone‐mediated autophagy. J Cell Biol, 2018;217(2):635–47.
140. Pajares, M., Rojo, A.I., Arias, E. et al. Transcription factor NRF2 modu-
lates chaperone‐ mediated autophagy through the regulation of LAMP2A.
Autophagy, 2018;14(8):1310–22.
141. Valdor, R., Mocholi, E., Botbol, Y. et al. Chaperone‐mediated autophagy
regulates T cell responses through targeted degradation of negative regula-
tors of T cell activation. Nat Immunol, 2014;15(11):1046–54.
142. Kiffin, R., Christian, C., Knecht, E., and Cuervo, A. Activation of chaperone‐
mediated autophagy during oxidative stress. Mol Biol Cell, 2004;15:4829–40.
143. Massey, A.C., Kaushik, S., Sovak, G., Kiffin, R., and Cuervo, A.M.
Consequences of the selective blockage of chaperone‐mediated autophagy.
Proc Natl Acad Sci U S A, 2006;103:5905–10.
144. Schneider, J.L., Suh, Y., and Cuervo, A.M. Deficient chaperone‐mediated
autophagy in liver leads to metabolic dysregulation. Cell Metab, 2014;20(3):
417–32.
145. Strnad, P., Zatloukal, K., Stumptner, C., Kulaksiz, H., and Denk, H.
Mallory‐Denk‐bodies: lessons from keratin‐containing hepatic inclusion
bodies. Biochim Biophys Acta, 2008;1782:764–74.
146. Osna, N. and Donohue, T.J. Implication of altered proteasome function in
alcoholic liver injury. World J Gastroenterol, 2007;13:4931–7.
147. Perlmutter, D.H. The cellular response to aggregated proteins associated
with human disease. J Clin Invest, 2002;110:1219–20.
148. Kamimoto, T., Shoji, S., Hidvegi, T. et al. Intracellular inclusions contain-
ing mutant alpha1‐antitrypsin Z are propagated in the absence of autophagic
activity. J Biol Chem, 2006;281:4467–76.
149. Teckman, J., An, J.K., Blomenkamp, K., Schmidt, B., and Perlmutter, D.
Mitochondrial autophagy and injury in the liver in alpha 1‐antitrypsin defi-
ciency. Am J Physiol Gastrointest Liver Physiol, 2004;286(5):G851–G62.
150. Galluzzi, L., Pietrocola, F., Bravo‐San Pedro, J.M. et  al. Autophagy in
malignant transformation and cancer progression. EMBO J, 2015;34(7):
856–80.
151. Ding, Z.B., Shi, Y.H., Zhou, J. et al. Association of autophagy defect with a
malignant phenotype and poor prognosis of hepatocellular carcinoma.
Cancer Res, 2008;68(22):9167–75.
152. Yang, J.D., Seol, S.Y., Leem, S.H. et al. Genes associated with recurrence
of hepatocellular carcinoma: integrated analysis by gene expression and
methylation profiling. J Korean Med Sci, 2011;26(11):1428–38.
153. Song, J., Guo, X., Xie, X. et al. Autophagy in hypoxia protects cancer cells
against apoptosis induced by nutrient deprivation through a Beclin1‐
dependent way in hepatocellular carcinoma. J Cell Biochem, 2011;112(11):
3406–20.
154. Liu, Y.L., Yang, P.M., Shun, C.T. et al. Autophagy potentiates the anti‐can-
cer effects of the histone deacetylase inhibitors in hepatocellular carci-
noma. Autophagy, 2010;6(8):1057–65.
155. Sir, D., Chen, W., Choi, J. et  al. Induction of incomplete autophagic
response by hepatitis C virus via the unfolded protein response. Hepatology,
2008;48:1054–61.
156. Singh, R., Kaushik, S., Wang, Y. et al. Autophagy regulates lipid metabo-
lism. Nature, 2009;458(7242):1131–5.
157. Inami, Y., Yamashina, S., Izumi, K. et  al. Hepatic steatosis inhibits
autophagic proteolysis via impairment of autophagosomal acidification
and cathepsin expression. Biochem Biophys Res Commun, 2011;412(4):
618–25.
158. Fukuo, Y., Yamashina, S., Sonoue, H. et  al. Abnormality of autophagic
function and cathepsin expression in the liver from patients with non‐alco-
holic fatty liver disease. Hepatol Res, 2014;44(9):1026–36.
159. Kaushik, S. and Cuervo, A.M. Degradation of lipid droplet‐associated pro-
teins by chaperone‐mediated autophagy facilitates lipolysis. Nat Cell Biol,
2015;17(6):759–70.
160. Rodriguez‐Navarro, J.A., Kaushik, S., Koga, H. et al. Inhibitory effect of
dietary lipids on chaperone‐mediated autophagy. Proc Natl Acad Sci U S A,
2012;109(12):E705–14.
161. Ward, W. Protein degradation in the aging organism. Prog Mol Subcell Biol,
2002;29:35–42.
162. Shibatani, T., Nazir, M., and Ward, W. Alteration of rat liver 20S proteas-
ome activities by age and food restriction. J Gerontol A Biol Sci Med Sci,
1996;51:B316–22.
163. Shibatani, T. and Ward, W. Effect of age and food restriction on alkaline
protease activity in rat liver. J Gerontol, 1996;51:B316–22.
164. Gray, D., Tsirigotis, M., and Woulfe, J. Ubiquitin, proteasomes, and the
aging brain. Sci Aging Knowledge Environ, 2003;2003:RE6.
165. Keller, J., Gee, J., and Ding, Q. The proteasome in brain aging. Ageing Res
Rev, 2002;1:279–93.
166. Hochshild, R. Lysosomes, membranes and aging. Exp Gerontol,
1970;6:153–66.
167. Brunk, U. and Terman, A. Lipofuscin: mechanisms of age‐related accu-
mulation and influence on cell function. Free Rad Biol Med, 2002;33:
611–19.
168. Bergamini, E. and Kovacs, J. Exploring the age‐related changes in hor-
mone‐regulated protein breakdown by the use of a physiologic model of
stimulation of liver autophagy, in Protein Metabolism in Aging Modern
Aging Research, vol. 9 (eds. H. Segal et al.), Wiley‐Liss, New York, 1990,
pp. 361–70.
169. Donati, A., Cavallini, G., Paradiso, C. et  al. Age‐related changes in the
autophagic proteolysis of rat isolated liver cells: effects of antiaging dietary
restrictions. J Gerontol, 2001;56:B375–83.
170. Donati, A., Cavallini, G., Carresi, C. et al. Anti‐aging effects of anti‐lipol-
ytic drugs. Exp Gerontol, 2004;39:061–7.
171. Martinez‐Lopez, N., Tarabra, E., Toledo, M. et al. System‐wide benefits of
intermeal fasting by autophagy. Cell Metab, 2017;26(6):856–71 e5.
172. Kiffin, R., Kaushik, S., Zeng, M. et al. Altered dynamics of the lysosomal
receptor for chaperone‐mediated autophagy with age. J Cell Sci,
2007;120:782–91.
173. Schneider, J.L., Villarroya, J., Diaz‐Carretero, A. et  al. Loss of hepatic
chaperone‐mediated autophagy accelerates proteostasis failure in aging.
Aging Cell, 2015;14(2):249–64.
174. Zhang, C. and Cuervo, A.M. Restoration of chaperone‐mediated autophagy
in aging liver improves cellular maintenance and hepatic function. Nat
Med, 2008;14:959–65.
 Peroxisome Assembly,
12 Degradation, and Disease
Rong Hua1 and Peter K. Kim1,2
1
Cell Biology Program, Hospital for Sick Children, Toronto, ON, Canada
2
Department of Biochemistry, University of Toronto, Toronto, ON, Canada
INTRODUCTION
Peroxisomes are single membrane‐bound organelles that are
ubiquitous to most eukaryotic cells (Figure  12.1). Since their
initial description by J. Rhodin in 1954 [1] and biochemical
characterization by C. de Duve and P. Baudhuin in 1966 [2],
much more has become known about their physiological roles
in various organisms. Enclosed by a single membrane, their
matrix contains a diverse array of enzymes involved in multiple
metabolic reactions. In mammalian cells, the main functions of
peroxisomes include β‐oxidation of very‐long‐chain fatty acids
(VLCFAs), detoxification of hydrogen peroxide, as well as the
biosynthesis of a number of lipid‐related molecules, such as bile
acids and plasmalogens.
The importance of peroxisomes in these catabolic and ana-
bolic reactions is best illustrated in a group of genetically
­inherited diseases known collectively as the peroxisomal disor-
ders. These peroxisomal metabolic diseases are categorized into
two general groups: single metabolic enzyme disorders, which
are caused by a defect in single enzymes required for peroxiso-
mal functions, and peroxisomal biogenesis disorders (PBDs),
which result from a defect in peroxisome formation [3]. PBDs
represent a spectrum of disorders which include the Zellweger
syndrome spectrum (PBD‐ZSD) (Zellweger syndrome, neona-
tal adrenoleukodystrophy, and infantile Refsum disease) and
rhizomelic chondrodysplasia punctate type 1 (RCDP1). These
diseases were named based solely on the clinical description,
but not on biochemical or genetic analysis, with Zellweger
­syndrome being the most severe clinical manifestation of PBD
and infantile Refsum disease showing the mildest symptoms
[3]. RCDP1, however, presents a different biochemical and
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
clinical phenotype compared to PBD‐ZSD [3]. PBDs result
from ­mutations in any of the 14 peroxins, encoded by PEX
genes, which are involved in peroxisome assembly and forma-
tion. The characterization of peroxins has been greatly assisted
by both genetic and proteomic studies in yeast and mutated
Chinese hamster ovary (CHO) cells with defects in peroxisome
number [4, 5]. In mammalian cells, 14 peroxins have been iden-
tified (Table 12.1), whereas there are at least 32 and 22 peroxins
in yeast and plants respectively [6]. RCDP1 is due to mutations
in PEX7, whereas mutations in the other 13 peroxins lead to
ZSD (Table 12.1).
Like other organelles, peroxisomes need to interact and com-
municate with their surroundings, including other intracellular
compartments, for their proper formation and maintenance, as
well as for their metabolic functions. A common method of
intracellular communication between organelles is via the mem-
brane contact site (MCS), where the opposing membranes of the
two organelles are physically tethered. The close proximity
between organelles allows for the exchange of small molecules
and signals [7]. Peroxisomes have long been seen to be juxta-
posed to other organelles in electron micrographs, especially the
endoplasmic reticulum (ER). However, the molecular mecha-
nisms underlying these peroxisome contact sites and their
implications for peroxisome biogenesis and function are only
recently starting to be explored [8].
The estimated half‐life for peroxisomes is approximately
2–3  days, suggesting that a balance between peroxisome
­biogenesis and their degradation is crucial for maintaining per-
oxisome homeostasis. Compared to what is known about the bio-
genesis of peroxisomes, less is known about the mechanism of
their degradation. It is thought that peroxisomes are mainly
138 THE LIVER:  PEROXISOME BIOGENESIS

(a) (b)

Figure 12.1  Peroxisomes in mammalian cells. (a) Live cell confocal image of COS7 cell expressing the peroxisome marker UB‐RFP‐SKL (red).
DAPI (blue), nucleus. (b) Live cell confocal image of COS7 cell expressing the ER marker ssGFP‐KDEL (green) and the peroxisome marker
UB‐RFP‐SKL (red). The white box indicates the magnified area shown in the right panels. Scale bars, 10 μm.

Table 12.1  Identified mammalian PEX genes involved in peroxisome biogenesis

Gene Function in biogenesis Molecular function Disease
PEX1 Matrix protein import ATP‐dependent dislocation of PEX5 ZS, NALD, IRD
PEX2 Matrix protein import E3 ubiquitin ligase for pexophagy ZS, IRD
PEX3 PMP targeting; de novo formation Membrane anchor of PEX19 ZS
PEX5 Matrix protein import PTS1 receptor ZS, NALD
PEX6 Matrix protein import ATP‐dependent dislocation of PEX5 ZS, NALD
PEX7 Matrix protein import PTS2 receptor RCDP type 1
PEX10 Matrix protein import PEX5 recycling ZS, NALD
PEX11 Proliferation Elongation of peroxisome ZS
PEX12 Matrix protein import PEX5 recycling ZS, NALD, IRD
PEX13 Matrix protein import Docking complex ZS, NALD
PEX14 Matrix protein import Docking complex ZS
PEX16 PMP targeting; de novo formation PMP receptor at ER and peroxisome ZS
PEX19 PMP targeting; de novo formation Group II PMP receptor and chaperone ZS
PEX26 Matrix protein import Membrane anchor of PEX1/PEX6 ZS, NALD, IRD
IRD, infantile Refsum disease; PMP, peroxisome membrane protein; PTS, peroxisomal targeting signal; NALD, neonatal adrenoleukodystrophy; RCDP, rhizomelic chondrodys-
plasia punctata; ZS, Zellweger syndrome.

degraded through an autophagy pathway called pexophagy, in PEROXISOME BIOGENESIS

which damaged or superfluous peroxisomes are wrapped around
by a double‐membrane vesicle called a autophagosome. This then
fuses with a lysosome and the engulfed content is degraded [9].
Thus far, several key players that induce and regulate pexophagy
have been identified, and a peroxisome‐specific autophagy
­adaptor protein (NBR1) has recently been characterized [10].
This chapter will focus largely on our understanding of per-
oxisomes in the mammalian system. In particular, the chapter
will cover the current models for peroxisome biogenesis, the
import mechanisms for both peroxisomal membrane and matrix
proteins, their communication with other organelles, especially
in the context of lipid transport, and the degradation of peroxi-
somes. In order to appreciate the importance of peroxisome
regulation (formation and degradation), their metabolic func-
tions and related diseases will also be discussed.
Growth and division model
Since their first discovery, the origin of peroxisomes has been a
subject of debate. Based on electron micrographs of hepatocytes
in the early 1960s, peroxisomes were reported to be in juxtapo-
sition with the ER, suggesting the involvement of the ER in the
origin of peroxisomes. This led C. de Duve and P. Baudhuin to
propose in their early review that peroxisomes were formed by
budding from the ER [2]. This view was later challenged
by Paul B. Lazarrow and Yukio Fujiki in the mid 1980s, when
they introduced the “growth and division” model in their author-
itative review [11]. This postulation was based on the seminal
finding that peroxisomal proteins, both the matrix and mem-
brane proteins, were synthesized on free polyribosomes in the
 12:  Peroxisome Assembly, Degradation, and Disease 139
cytosol and then imported post‐translationally into peroxisomes.
Peroxisomes were therefore considered to be autonomous
­organelles, like mitochondria and chloroplasts, that proliferate
by the fission of pre‐existing ones.
Today, we know that peroxisomes are formed via two mecha-
nisms: de novo biogenesis and growth and division. De novo
biogenesis involves the emergence of pre‐peroxisomal vesicles
from the ER, which then fuse together to form mature peroxi-
somes. The growth and division mechanism is mediated by
elongation and fission machinery, consisting of the PEX11
­family, dynamin‐like protein 1 (DLP1/Drp1), MFF, and fission
1 (FIS1). The PEX11 family proteins induce the elongation of
the peroxisomal membrane and also help to recruit the fission
machinery to their site of action, leading to the final scission of
the membrane [12, 13] (Figure 12.2).
Model for de novo peroxisome biogenesis
This view of peroxisomes as an autonomous organelle came
under question due to a number of key observations. First, rather
than multiplying solely by growth and division from pre‐­
existing ones, peroxisomes were found to be formed de novo in
peroxisome‐deficient cells. In cells with mutation or deletion in
any of the three peroxisome biogenesis factors – PEX3, PEX16,
and PEX19 – no peroxisomal membrane ghosts were detected.
Upon reintroduction of the wild‐type peroxin, peroxisomal
membrane structure could be detected within a short time frame,
followed by the restoration of matrix protein import [14–16].
This raises the possibility that peroxisomes could also be formed
from yet‐to‐be‐identified pre‐peroxisomal structures or derived
de novo from other organelles. In fact, such small pre‐peroxisomal
structures/vesicles have been identified in multiple yeast spe-
cies, including Yarrowia lipolytica, Saccharomyces cerevisiae,
and Hansenula polymorpha [17–19]. These pre‐peroxisomal
vesicles were shown to contain distinct pools of membrane pro-
teins and undergo heterotypic fusion to form new peroxisomes.
Interestingly, these structures were rapidly degraded via
autophagy in pex mutant cells if they did not mature into fully
functional peroxisomes [19].
Although similar pre‐peroxisomal structures have not yet
been identified in mammalian cells, recent studies have pointed
to the ER as a contributor for the de novo formation of mam-
malian peroxisomes. Using electron microscopy, peroxisomes
in mouse dendritic cells were found to be surrounded by a dou-
ble‐membrane structure that is continuous with the rough ER
[20]. Similar structures were also observed in CHO cells lack-
ing PEX16 [21] and hepatocytes of PEX5 knockout mice [22].
Moreover, as discussed in detail below, most mammalian
­peroxisomal membrane proteins (PMPs) can be sorted to per-
oxisomes via the ER. The initial ER recruitment and the subse-
quent ER‐to‐peroxisome trafficking of these PMPs appear to
depend on PEX16 [23]. Therefore, the ER is likely the mem-
brane source for these pre‐peroxisomal structures in both yeast
and mammals, although the precise mechanism involved
remains unclear.
Conceptually, de novo formation of peroxisomes requires the
action of three groups of peroxins involved in (i) membrane
modulation, (ii) peroxisomal membrane assembly/PMP import,
and (iii) matrix protein import (Figure  12.2). The de novo
biogenesis of peroxisome starts with the recruitment of mem-
brane modulators that induce membrane curvature and the
assembly of pre‐peroxisomal structures/vesicles at the ER.
PEX16 is one of the most likely candidates to be involved in this
process, since it is capable of recruiting a variety of PMPs to the
ER, including the early biogenesis factor PEX3 and the PEX11
family proteins, which are known membrane modulators that
induce peroxisomal membrane elongation [23]. These PMPs
together with PEX16 concentrate at specialized subdomains on
the ER and thereby lead to the formation of the pre‐peroxisomal
structures. These pre‐peroxisomal structures are then released
from the ER via an unknown mechanism. One aspect worth
mentioning is that the release of these pre‐peroxisomal struc-
tures is unlikely via the secretory pathway, as neither COPI or
COPII is required for the de novo peroxisome formation in
Pex3‐deficient cells upon reintroduction of wild‐type PEX3
[24]. These ER‐derived structures will continuously import
PMPs via the action of PEX3 and PEX19, leading to the assem-
bly of the matrix protein import machinery and eventually give
rise to the formation of new mature peroxisomes.
Contribution of de novo biogenesis versus
fission in peroxisome formation
It is now generally accepted that the ER contributes to the
de novo formation of peroxisomes, especially in cells without
pre‐existing ones. However, the extent that the two mecha-
nisms – ER‐derived de novo formation versus growth and divi-
sion  –  are utilized in normal mammalian cells is still under
debate. One hypothesis is that both mechanisms exist, but they
are activated by different metabolic signals. The constitutive
formation of peroxisomes may be mediated by the de novo path-
way, whereas the rapid peroxisome proliferation is achieved by
growth and division. Rodent liver cells respond rapidly to clofi-
bric acid treatment by doubling and in some cases, quadrupling
the number of peroxisomes within days. This rapid proliferation
of peroxisomes is mediated by the activation of peroxisome
­proliferator‐activated receptor α (PPARα), which is a nuclear
hormone receptor that regulates the expression of genes associ-
ated with lipid metabolism [25]. One of the genes that was
shown to be upregulated by PPARα activation is PEX11α, which
is a tissue‐selective gene expressed most prominently in the
liver [26]. Since the PEX11 family proteins are involved in per-
oxisome elongation and fission, it is likely that under conditions
of rapid peroxisome proliferation, peroxisomes multiply mainly
via growth and division. At the same time, the de novo biogen-
esis pathway is the constitutive mechanism for maintaining
­peroxisome number during normal cell growth.
On the other hand, it is also possible that the two mechanisms
work in concert with each other. As discussed in detail below,
mammalian PMPs can be targeted to peroxisomes either directly
from the cytosol or via the ER, suggesting a role of the ER in
constitutively providing PMPs to pre‐existing peroxisomes for
their maintenance. Although the exact mechanism remains
unclear, it is likely that the ER‐targeted PMPs are sorted to per-
oxisomes via vesicular transport between the two organelles.
Therefore, the ER‐derived pre‐peroxisomal vesicles could either
mature into newly formed peroxisomes (i.e. de novo formation),
140 THE LIVER:  PEROXISOMAL PROTEIN TARGETING

Endoplasmic
reticulum
i. Membrane
remodeling

ii. PMP import

de novo biogenesis
Membrane modulators

Peroxins for PMP import Pre-peroxisome

Peroxins for matrix protein import

Other PMPs iii. Matrix protein

import
Matrix proteins
PEX5/PEX7

Mature
peroxisome

Elongation
(PEX11β)

Fission
Constriction
(FIS1/MFF/DLP1)
(FIS1/MFF/DLP1)

Growth and division

Figure 12.2  Peroxisome biogenesis. Mammalian peroxisomes can propagate via growth and division from pre‐existing ones, or be formed de
novo from other organelles, such as the ER. The de novo formation of peroxisomes starts when membrane modulators are recruited to specialized
subdomains on the ER (i). The peroxisome biogenesis factor PEX16 is likely involved in this process, in which it recruits other peroxisome mem-
brane proteins (PMPs) to the, ER, including the early biogenesis factor PEX3 and the known membrane modulator, the PEX11 family. The accu-
mulation of these PMPs together with PEX16 on the ER leads to the formation of the pre‐peroxisomal structures, which are then released via an
unknown mechanism. These ER‐derived structures will continuously import PMPs via the action of PEX3 and PEX19 (ii), leading to the assembly
of the matrix protein import machinery (iii), and eventually mature into fully functional peroxisomes. On the other hand, peroxisomes can also
propagate via growth and division from existing ones. Peroxisome proliferation is initiated by membrane elongation and followed by constriction
and final scission of the membrane. The PEX11 family induce membrane elongation and recruit the fission machinery proteins (DLP1/Drp1, FIS1,
and MFF) to their site of action, leading to the fission of peroxisomes.

or fuse with pre‐existing ones to provide proteins and/or lipid
content for their growth and division. One possible reason for
the existence of this specialized ER‐to‐peroxisome trafficking
pathway in the mammalian system, as well as perhaps in plants,
is that these organisms need to maintain a much larger popula-
tion of peroxisomes than most yeasts. Mammalian and plant
cells typically contain 100–1000 peroxisomes, whereas only
2–10 peroxisomes are found in most yeasts. Hence, the rela-
tively large numbers of peroxisomes imply that there is more
demand on the ER in terms of providing both protein and lipid
content required for maintaining the steady‐state number of
­peroxisomes in both mammals and plants.
PEROXISOMAL PROTEIN TARGETING
Targeting of peroxisomal membrane proteins
The targeting of most if not all PMPs to peroxisomal mem-
branes requires the action of three peroxins named PEX3,
PEX16, and PEX19 [23]. Mutation in any of these three
­peroxins results in the absence of any detectable peroxisomal
membrane structures; whereas other PMPs are either rapidly
degraded in the cytosol or mislocalized to other organelles, such
as the mitochondria [14, 15, 27]. Currently, PMPs are thought to
be targeted to peroxisomes via two distinct but not mutually
 12:  Peroxisome Assembly, Degradation, and Disease 141
exclusive pathways, the Group I pathway (i.e. indirectly via
the ER) and the Group II pathway (i.e. directly from the cytosol)
(Figure 12.3a).
In the direct Group II pathway, most PMPs are thought to be
synthesized on free ribosomes in the cytosol and targeted post‐
translationally to peroxisomes. PEX19 acts as a soluble receptor
that binds to these newly synthesized PMPs and stabilizes them
in a soluble state in the cytosol [28]. The cargo–PEX19 complex
is then guided to peroxisomal membranes via the interaction
between PEX19 and the docking factor PEX3. The targeting of
PEX3 itself also depends on PEX19 as well as PEX16 on the
membranes. However, how PMPs are inserted into the peroxiso-
mal lipid bilayers remains unclear.
The indirect Group I pathway is thought to depend on PEX16,
which is a PMP with dual localization in both peroxisomes and
the ER [23]. PEX16 is shown to target to peroxisomes exclu-
sively via the ER, as the domains required for its co‐translational
insertion to the ER and subsequent ER‐to‐peroxisome targeting
have been recently identified [23]. At the ER membranes,
PEX16 is able to recruit a wide range of PMPs that differ
in membrane topology and function to the ER. Interestingly,
both PEX3 and PEX19 are dispensable for PEX16‐mediated
recruitment of PMPs to the ER, implying mechanistic differ-
ences between PMP targeting to the ER and to the peroxisomes
[23]. However, how the ER‐targeted PMPs are subsequently
transported to the existing peroxisomes remains largely unclear.
One possibility is that this protein‐targeting pathway is medi-
ated by vesicular transport between the two organelles, in which
pre‐peroxisomal vesicles budding from the ER fuse with
­existing peroxisomes to provide them with essential protein
and/or lipid content.
Import of peroxisomal matrix proteins
In contrast to the protein import pathways of the ER and mito-
chondria, peroxisomal matrix proteins translocate across the
lipid bilayer as fully folded proteins, and in some cases as oligo-
meric complex [29, 30]. Therefore, the mechanism of matrix
protein import to peroxisomes is more closely related to the
nucleus, except that no stable translocon or channel is involved
in the process. The peroxisomal matrix protein import cascade
can be divided into four steps: (i) cargo binding; (ii) docking;
(iii) translocation and cargo release; and (iv) receptor ubiquit-
ination and recycling [31] (Figure 12.3b).
Cargo binding
Peroxisomal matrix proteins contain one of the two peroxisomal
targeting signals called PTS1 or PTS2. Most peroxisomal matrix
proteins exhibit PTS1, which consists of a tripeptide sequence
(S/A/C)‐(K/R/H)‐(L/M) at the extreme C‐terminal end [32].
The PTS1 signal is recognized and bound by the soluble ­receptor
PEX5, which contains a helix bundle at its N‐terminus, and a
series of tetratricopeptide repeats (TPRs) within the C‐terminal
half of the protein. The TPRs are involved in PTS1 signal bind-
ing; while the helix bundle is involved in binding to the docking
complex (PEX13 and PEX14) at the peroxisomal membrane [33].
The less common PTS2 is a nonapeptide with the consensus
sequence (R/K)‐(L/V/I)‐X5‐(H/Q)‐(L/A) located at the N‐terminus
of the protein. The cytosolic receptor for PTS2‐containing
proteins is PEX7, which contains six WD40 repeats involved in
binding to PTS2 sequences [34]. Unlike PEX5, PEX7 requires
species‐specific accessory proteins to carry out its function. For
example, the accessory protein for PEX7 is the orthologous
redundant Pex18p/Pex21p in yeast S. cerevisiae, whereas in
mammalian and plant cells it is the longer splicing isoform of
PEX5 (i.e. PEX5L) [35]. In addition, a small number of peroxi-
somal matrix proteins do not contain a PTS1 or PTS2 sequence.
These proteins are thought to be imported into peroxisomal
matrix either by binding to PEX5 in a PTS1‐independent
­manner or via a piggy‐back mechanism by binding to other
PTS‐containing cargos [29, 36].
Docking
The docking complex at the peroxisomal membranes consists of
two peroxins, PEX13 and PEX14. These proteins bind with
each other and have binding affinity for both PTS receptors.
Interestingly, they contain several binding sites for PEX5,
­suggesting a dynamic stepwise binding of the cargo–receptor
complex. Moreover, PEX14 is believed to present with the
­initial binding site for the cargo–receptor complex, as it has a
higher binding affinity than PEX13 [31, 37].
Translocation and cargo release
The precise mechanism of cargo protein translocation across the
peroxisomal membrane is not fully understood. The current
results favor a model by which peroxisomal matrix proteins
translocate across the membranes through a protein‐conducting
channel or a translocon. In this model, a transient pore/
protein‐conducting channel is assembled upon the docking of
the cargo–receptor complex at the peroxisomal membrane, and
disassembled after cargo translocation across the membrane
[38]. One candidate protein that may be involved in the forma-
tion of such transient pores is PEX5, as an aqueous pore up to
9 nm in diameter can be formed when Pex5p‐containing com-
plex purified from yeast peroxisomal membranes is reconsti-
tuted into liposomes [39]. It is unclear whether a distinct pore
exists for the PTS2 cargos.
The cargo protein needs to be separated from the receptor
before translocating across the membrane. However, the mecha-
nism of cargo–receptor dissociation is still a matter of debate.
The import receptor PEX5 comprises two structurally and func-
tionally autonomous parts. The N‐terminal part interacts with
the docking complex at the peroxisomal membranes during
cargo translocation, whereas the C‐terminal part binds to the
PTS1 cargos. Upon binding of its N‐terminal part to the docking
complex, the C‐terminal segment of PEX5 is thought to undergo
a conformational change, thereby inducing cargo release [40].
Receptor ubiquitination and recycling
After cargo release, the receptors are recycled back to the cyto-
sol where they either enter another import cycle or are targeted
for proteasomal degradation. This process requires the action of
the RING complex (PEX2/PEX10/PEX12) and AAA‐type
ATPase complex (PEX1/PEX6/PEX26). During the import
cycle, the PTS1‐receptor PEX5 is regulated by different types
of  ubiquitination [41]. Mono‐ubiquitination of PEX5 at the
142 THE LIVER:  PEROXISOMAL LIPID TRAFFICKING

(a)
19
Chaperone?
PMP

PMP PMP

19 19
PMP
3 3
16 PMP
16 PMP

Peroxisome ER lumen
lumen

Direct pathway Indirect pathway

(b)
PTS1

PTS1

5 5 Ub

II
IV

Ub 6
PTS1

13 14 14 5 5 14 2 10 12 5 26

III

PTS1

Figure 12.3  Peroxisomal protein import. (a) Peroxisomal membrane proteins can be targeted to peroxisomes either directly from the cytosol (i.e.
the direct Group II pathway) or via the ER (i.e. the indirect Group I pathway). The newly synthesized PMPs is recognized by the soluble receptor
PEX19, which is then guided to the peroxisomal membrane via its interaction with the docking factor PEX3. PEX3 itself is targeted to peroxisomes
via PEX16. The ER targeting of PMPs in mammalian cells is proposed to depend on PEX16. (b) Peroxisomal matrix protein import occurs via a
multistep cycle that is divided into four steps: cargo binding (I); docking (II); translocation and cargo release (III); and receptor ubiquitination and
recycling (IV). The peroxisomal matrix protein containing either PTS1 or PTS2 signal is recognized and bound by the soluble receptor PEX5 or
PEX7, respectively. The cargo–receptor complex is then docked at the peroxisomal membrane via the docking complex (PEX13 and PEX14). A
transient pore that is mainly constituted by PEX5 is assembled upon the docking of cargo–receptor complex, and thereby allowing the translocation
of cargo across the membrane and its subsequent release into the peroxisomal matrix. After cargo release, the receptors are then recycled back to
the cytosol with the action of the RING complex (PEX2/PEX10/PEX12) and AAA‐type ATPase complex (PEX1/PEX6/PEX26).

conserved N‐terminal cysteines is found to be required for PEROXISOMAL LIPID TRAFFICKING

receptor cycling, while poly‐ubiquitination on lysine residues is
a signal for proteasomal degradation. The RING complex is
thought to be responsible for receptor ubiquitination. The ubiq-
uitinated PEX5 is recognized by the two AAA‐type ATPases
PEX1 and PEX6, which are anchored to the peroxisomal mem-
brane via their interaction with the integral membrane protein
PEX26. ATP hydrolysis by PEX1 and PEX6 provides the driv-
ing force that is required for the extraction of receptors from the
membrane, resulting in the recycling of receptors back into the
cytosol [31, 38].
Besides proteins, peroxisomes also need to exchange lipid mol-
ecules with the surroundings to regulate and balance their pro-
liferation and cellular functions. Peroxisomes must acquire
membrane lipids for their growth and proliferation from another
donor organelle, as they lack most of the enzymes required for
phospholipid biosynthesis. Peroxisomes also need to exchange
metabolites with other organelles for their cellular functions.
For example, metabolites are transferred between peroxisome
and the mitochondria for β‐oxidation of VLCFAs. However, the
 12:  Peroxisome Assembly, Degradation, and Disease 143
precise mechanism by which lipids are exchanged between
­peroxisomes and other organelles is not fully understood.
In general, lipid exchange between various cellular compart-
ments can occur via either the vesicular transport mechanism or
nonvesicular mechanisms. The bulk transfer of lipids occurs
when vesicles budding from the donor membrane fuse with the
receiving membrane. In the nonvesicular trafficking pathways,
specific lipid molecules are usually transferred at the MCSs,
which are the regions where membranes of two organelles are in
close proximity (within 10–30 nm) [42]. The implication of
various lipid‐sorting mechanisms for lipid transfer between
­peroxisomes and other organelles will be discussed.
Interplay between peroxisomes and the ER
As discussed earlier, the ER contributes to the de novo biogen-
esis as well as the growth and division of existing peroxisomes,
as it provides peroxisomes with a subset of PMPs. Moreover, as
the ER is the major site of lipid synthesis, another likely role of
the ER in peroxisome biogenesis is to provide membrane lipids
for the development of peroxisomal endomembrane system.
Much of our knowledge about lipid transfer between peroxi-
somes and the ER is gained from studies done in yeast; and
indeed, a nonvesicular pathway is suggested to be involved in
this process [43]. Although the molecular mechanism underly-
ing this nonvesicular lipid transport is not known, close contact
(i.e. MCSs) between the two organelles are suggested to be
required for this transfer.
So far, two protein complexes have been identified in yeast to
allow for the tethering between peroxisomes and the ER. The
first tethering complex was identified in the yeast Pichia
­pastoris, consisting of the integral peroxisomal membrane
­protein Pex30p and the ER reticulons Rnt1, Rnt2, and Yop1.
The ER–peroxisome contact sites mediated by this complex
were shown to be involved in peroxisome proliferation [44]. In
the budding yeast S. cerevisiae, a tethering complex, consisting
of Pex3p and the peroxisome inheritance factor Inp1p, was
identified to play a role in peroxisome inheritance. Inp1p acts as
a molecular hinge between the ER‐localized Pex3p and the per-
oxisomal Pex3p, thereby anchoring peroxisomes to the ER in
the mother cells. On the other hand, peroxisomes that are
enriched for Inp2p are transported to the bud along microtu-
bules via the myosin motor protein 2 (Myo2).
Although no known mammalian homolog of Pex30p or Inp1p
has been identified, a novel tethering complex has recently been
identified in mammalian cells that is mediated by the direct
interaction between the ER‐resident VAMP‐associated proteins
A and B (VAPA and VAPB) and the peroxisomal membrane
protein ACBD5 [45, 46]. Both VAP proteins were found to con-
centrate at specific loci on the ER close to peroxisomes, where
they interact with the FFAT motif‐containing protein ACBD5
on peroxisomal membranes to facilitate the formation of ER–­
peroxisome contact sites. These contact sites bring the two orga-
nelles close to each other, allowing for the transfer of materials
between them. Lipid molecules that may be transferred at these
contact sites include membrane lipids and precursors of choles-
terol and plasmalogens, as disruption of this VAP–ACBD5 con-
tact was shown to interfere with both peroxisome growth and
cellular cholesterol and plasmalogen homeostasis. However, it
is not known whether ACBD5 binds to lipid molecules directly
via its acyl‐CoA‐binding domain, or whether other lipid‐­binding
proteins are recruited to these contact sites and help to s­ huttle
lipids between the two organelles.
Interplay between peroxisomes
and mitochondria
Peroxisomes and mitochondria share many common charac-
teristics and functions. For example, the biogenesis of both
organelles are under the control of the pro‐lipid metabolism
nuclear receptors PPARs. They also cooperate with each
other in the β‐oxidation of VLCFAs in mammalian cells. Lipid
exchange between the two organelles is suggested to occur at
their MCSs. A tethering complex consisting of Pex11p and the
ERMES protein Mdm34 has been identified in yeast, allowing
for the efficient transfer of metabolites between peroxisomes
and mitochondria [47]. In mammalian cells, one of the pro-
teins involved in the interaction between peroxisomes and
mitochondria is identified to be ABCD1. Defects in ABCD1
are associated with the peroxisomal disorder X‐linked adre-
noleukodystrophy (X‐ALD), which is characterized by the
­accumulation of VLCFAs [48].
On the other hand, lipid transfer between peroxisomes and
mitochondria may also occur via vesicular transport, as a novel
vesicular trafficking pathway between them has recently been
reported in the mammalian system. Mitochondria‐derived vesi-
cles (MDVs) have been shown to fuse with a subpopulation
(~10% of total) of peroxisomes [49]. However, the function and
physiological significance of these MDVs remains unclear.
Whether they contribute to peroxisome biogenesis from the
mitochondria or they are involved in lipid exchange between
the two organelles needs to be further investigated.
Interplay between peroxisomes
and lipid droplets
Lipid droplets (LDs) are the lipid‐storage organelles found in all
organisms due to their ability to accumulate neutral lipids such
as triacylglycerols (TAGs) and cholesterol ester. An imbalance
in lipid storage and lipolysis or secretion will lead to hepatic
steatosis, which is a condition of abnormal lipid accumulation
in LDs in the liver. This condition can be induced by an increase
in fatty acid uptake/synthesis and LD biogenesis; at the same
time, a decrease in LD catabolism (such as β‐oxidation of fatty
acid) and impaired secretion can also lead to steatosis [50].
There is growing evidence that rather than solely being a storage
organelle, LDs also have a multifunctional nature, being
­implicated in protein degradation and pathogen replication as
well [51].
The close proximity between peroxisomes and LDs has long
been seen in electron micrographs, and was later confirmed in
multiple organisms, including the mammalian COS7 cells [52].
The association between these two metabolic organelles is
thought to be linked to lipolysis in LDs and β‐oxidation of fatty
acids in peroxisomes. The fact that accumulation of LDs was
seen in mouse hepatocytes lacking peroxisomes reinforced
a  connection between the two organelles [53]. Although the
144 THE LIVER:  PEROXISOME DEGRADATION: PEXOPHAGY
precise mechanism for peroxisome–LD communication is still
unclear, hemifusion between LD and peroxisomal membranes
(i.e. pexopodia) has been proposed to mediate the direct interac-
tion between the two organelles in yeast, which helps to facili-
tate the transfer of fatty acids from LDs to peroxisome for
β‐oxidation [54]. However, whether a similar mechanism exists
in other organisms is not known.
Remarkably, the insertion of UBXD8, a hairpin protein with
dual localization to the ER and LD, into the cytoplasmic leaflet
of the ER bilayer was suggested to depend on PEX3 and PEX19.
The shared protein import machinery between PMPs and LD‐
destinated hairpin proteins suggests a coordinated relationship
between the two organelles in the ER [55]. This mutual control
for both peroxisome and LD biogenesis will allow the cell to
tightly regulate energy levels in response to metabolic stimuli,
thereby maintaining lipid homeostasis in the liver.
Interplay between peroxisomes and lysosomes
For a long time, the interaction between peroxisomes and lys-
osomes were mainly focused on pexophagy, a process by which
peroxisomes are selectively degraded in lysosomes via the
autophagy pathway. However, a novel lysosome–peroxisome
contact site (LPMC) has recently been reported in mammalian
cells, and its implication in cholesterol transport was discussed
[56]. Mammalian cells can acquire exogenous cholesterol
through receptor‐mediated endocytosis of plasma low‐density
lipoprotein (LDL). The resulting free cholesterol emerges in the
late endosome/lysosome, where it is further transported to vari-
ous downstream organelles, including the ER, plasma mem-
brane, and mitochondria, via an unknown mechanism [57]. The
newly identified LPMC is shown to be mediated by the interac-
tion between the lysosomal membrane protein synaptotagmin
VII (Syt7) and phosphatidylinositol 4,5‐bisphosphate (PI(4,5)
P2) on the peroxisomal membrane. These LPMCs are proposed
to serve as a potential mechanism for cholesterol efflux from
lysosomes, as evidenced by the accumulation of cholesterol in
cultured fibroblasts from patients with peroxisomal disorders.
After reaching peroxisomes, it is unclear whether the LDL‐
derived cholesterol is further processed in peroxisomes or sub-
sequently transported to other cellular compartments.
PEROXISOME DEGRADATION:
PEXOPHAGY
Peroxisome homeostasis in cells is maintained by the balance
between peroxisome biogenesis (i.e. new peroxisomes are
formed de novo or via growth and division) and the degradation
of abnormal or dysfunctional peroxisomes. Peroxisomes can be
turned over via three mechanisms: (i) matrix protein degradation
by the Lon protease, (ii) 15‐lipoxygenese (15‐LOX)‐mediated
autolysis, and (iii) the autophagy pathway (i.e. pexophagy) [58].
Lon proteases are characterized as ATP‐dependent proteases
that have been found in eukaryotic mitochondria, chloroplasts,
and peroxisomes. The peroxisome‐specific isoform of Lon pro-
tein (pLon) in rat liver was shown to play a role in processing
and activating several PTS1‐containing peroxisomal enzymes
involved in β‐oxidation, including acyl‐CoA oxidase (ACOX)
[59]. The pLon protease may also be involved in the degradation
of induced β‐oxidation enzymes upon peroxisomal proliferation
[60]. In yeast, the peroxisomal Lon promotes degradation of
misfolded or damaged peroxisomal matrix proteins. Interestingly,
in the plant Arabidopsis, peroxisome degradation via autophagy
(i.e. pexophagy) is enhanced in cells that lack Lon [61].
In the 15‐LOX‐mediated autolysis, the peroxisomal mem-
brane is permeabilized, leading to the release of its matrix con-
tent. 15‐LOX belongs to the lipoxygenase enzyme family which
converts polyunsaturated fatty acids into conjugated hydroper-
oxides, leading to the lipoperoxidation of organellar membranes
and the synthesis of signaling molecules [62]. As a result of
membrane lipid peroxidation, the matrix content is released and
thereby degraded by the cytosolic proteases [63]. 15‐LOX was
found to be localized to some, but not all, peroxisomes in rat
hepatocytes, suggesting a potential role of 15‐LOX in pro-
grammed degradation of peroxisomes in these cells [64].
Moreover, a novel role of 15‐LOX in regulating autophagy has
been proposed, as cells deficient in 12/15‐LOX (a homolog of
15‐LOX) exhibit defects in autophagy [65]. Furthermore, a
LOX‐derived oxidized phospholipid was found to act as an
effective substrate for key proteins (such as yeast Atg8 and
mammalian LC3) required for autophagy, implying a linkage
between phospholipid oxidation with the cellular turnover
machinery, autophagy [65].
The principal machinery for peroxisome degradation is the
autophagy‐related process called pexophagy. Like autophagy,
pexophagy can occur in two modes: macropexophagy and
micropexophagy [66, 67]. During macropexophagy, the peroxi-
some is wrapped around by a double‐membrane vesicle called
the autophagosome, to form a new structure named the pex-
ophagosome. The mature pexophagosome then moves along the
microtubules and fuses with a lysosome to degrade the seques-
tered peroxisome. In contrast, during micropexophagy, the
engulfment of cargo (i.e. peroxisomes) occurs by invagination
of the lysosomal membrane, resulting in the budding of vesicles
into the vacuole lumen and subsequent degradation of its
engulfed content. Macropexophagy occurs in all the eukaryotic
cells that have been examined, including human, whereas
micropexophagy has only been observed in two fungi species,
P. pastoris and Aspergillus nidulans. Therefore, only macropex-
ophagy will be discussed in this section.
Ubiquitination‐mediated peroxisome
recognition for pexophagy
How pexophagy is triggered or how a peroxisome is recog-
nized and targeted for pexophagy is not fully understood.
Recent data suggest that ubiquitination of membrane proteins
of specific organelles is required for selective autophagy.
Consistent with this, ubiquitination of PMPs was shown to
induce pexophagy. For example, overexpression of two
PMPs – PMP34 and PEX3 – that are fused with ubiquitin in
the cytosolic side, dramatically induced pexophagy in mam-
malian cells [68]. Interestingly, expression of a PEX3 mutant
that is defective in ubiquitination could also induce peroxi-
some ubiquitination and degradation, implying that ubiquit-
ination of PEX3 is dispensable for pexophagy [69]. This result
 12:  Peroxisome Assembly, Degradation, and Disease 145
also suggests that other unidentified endogenous PMPs are
ubiquitinated upon PEX3 overexpression.
Another PMP that can be ubiquitinated, thereby triggering
pexophagy, is mammalian PEX5, the soluble receptor for the
PTS1‐containing peroxisomal matrix proteins. During the
import cycle, the PTS1 receptor PEX5 is regulated by different
types of ubiquitination [41]. The mono‐ubiquitination of PEX5
on a conserved N‐terminal cysteine is recognized by the AAA‐
type ATPase complex (PEX1/PEX6/PEX26) that extracts PEX5
from the peroxisomal membrane, thereby recycling it back into
the cytosol for further rounds of protein import. However, when
the mono‐ubiquitination‐mediated receptor recycling pathway
is blocked, poly‐ubiquitination of PEX5 on lysine residues
occurs, followed by its extraction from the membranes and
rapid degradation by the 26S proteasome. A potential role for
mono‐ubiquitination of PEX5 as a quality‐control mechanism
for peroxisome degradation has been proposed in recent studies.
For example, overexpression of an export‐deficient PEX5
mutant was shown to trigger pexophagy in mouse embryonic
fibroblasts [70]. This PEX5 mutant was generated by fusing a
bulky C‐terminal tag to the PEX5 protein. The bulky C‐terminal
tag does not interfere with PEX5 mono‐ubiquitination, but
inhibits its export from the peroxisomal membranes, thereby
allowing the mono‐ubiquitinated PEX5 to be recognized by the
autophagy machinery [70]. Moreover, mono‐ubiquitination of
PEX5 could also be induced under conditions of ROS activation
[71]. It has been shown that in response to ROS activation,
PEX5 is phosphorylated at Ser141 by ataxia telangiectasia‐
mutated (ATM) kinase. This, in turn, promotes PEX5 mono‐
ubiquitination at K209, allowing it to be recognized by the
autophagy adaptor protein p62, thereby inducing pexophagy
[71]. However, whether ubiquitination of PMPs happens under
normal physiological conditions is not known.
E3 ubiquitin ligase for pexophagy
Peroxisomes have four putative E3 ubiquitin ligases (PEX2,
PEX10, PEX12, and TRIM37) that are all shown to act on the
ubiquitination of PEX5 for its removal from the peroxisomal
membrane during matrix protein import [72, 73]. Of these four
peroxisomal E3 ubiquitin ligases, only PEX2, which is a RING
finger E3 ligase, has been shown to be involved in the ubiquitina-
tion reaction required for pexophagy. PEX2 is usually rapidly
degraded and, thereby, maintained at low levels under normal
conditions. However, this protein is stabilized in stress condi-
tions, for example during amino acid starvation [74]. The stabili-
zation of PEX2 protein expression results in a rapid increase in
PMP ubiquitination which is a signal for pexophagy.
AAA ATPase defect and pexophagy
The most common mutations in PBDs are found in the peroxi-
somal AAA ATPase, which comprises PEX1, PEX6, and PEX26
[75]. Mutations in these peroxins are believed to reduce peroxi-
some numbers by affecting peroxisomal biogenesis. However,
there is growing evidence suggesting that the AAA ATPase may
also act in peroxisomal quality control. The main function of the
peroxisomal AAA ATPase is to remove ubiquitinated PEX5. As
such, it has been shown that the loss of AAA ATPase function,
due to either mutation or depletion of its expression, leads to
increased pexophagy activity and loss of peroxisomes [76].
Inhibiting autophagy resulted in not only increase in peroxi-
some number but also increase in peroxisomal functions [76].
Together, these findings suggest that the loss of peroxisomes in
AAA ATPase‐mutant cells may be due to increased degradation
of peroxisomes and not the lack of their biogenesis.
Autophagic receptors for pexophagy
Once a peroxisome is activated for degradation by accumulation
of ubiquitin on the cytosolic side of the membrane, it will be
recognized by a selective autophagy cargo receptor/adaptor
[77]. The autophagy adaptor links the substrate (i.e. peroxi-
some) to a nascent autophagosome by binding to both the ubiq-
uitin on the substrate and the autophagy effector LC3 on the
autophagosome. At present, the mammalian autophagy recep-
tors that have been shown to play a role in pexophagy are p62
and NBR1 [10]. These two proteins share similar domain com-
positions, including an N‐terminal PB1 domain that allows for
interaction with another p62 molecule, a UBA domain at the
C‐terminus that binds to ubiquitin on the cargo, a ZZ‐type zinc
finger in the middle, and an LIR motif that binds to the autophagy
factor LC3 [58]. NBR1, but not p62, is shown to be necessary
and sufficient for the turnover of endogenous peroxisomes [10].
P62 is not required for pexophagy when NBR1 is excess, but it
works cooperatively with NBR1 to increase the efficiency of
NBR1‐mediated pexophagy [10].
The difference in specificity between these two adaptors may
be explained by the presence of an additional JUBA domain
(a  novel membrane‐interacting, amphipathic α‐helix region
­preceding the UBA domain) in NBR1 [10]. The NBR1 mutants
that lack the JUBA domain (NBR1ΔJ and NBR1ΔJΔUBA)
did  not localize to peroxisome, suggesting an essential role
of  the JUBA domain for its proper subcellular localization.
Interestingly, the mutant that lacks only the UBA domain
(NBR1ΔUBA) still localized to peroxisomes, but exhibited
increased colocalization with other organelles, such as the mito-
chondria. Taken together, these results suggest that while both
the JUBA and UBA domains are required for its specificity for
peroxisomes, the UBA domain acts as a regulator of NBR1
binding to peroxisomes.
PEROXISOME FUNCTIONS AND
DISORDERS IN THE LIVER
Fatty acid β‐oxidation
Beta‐oxidation is the process of metabolizing fatty acids to
acetyl‐CoA for ATP synthesis. While peroxisomes are the
sole  site of β‐oxidation in plants and yeast, it occurs in both
mitochondria and peroxisomes in mammalian cells. The subset
of lipids that can only be β‐oxidized in peroxisomes include
VLCFAs, long‐chain dicarboxylic acids, leukotrienes, prosta-
glandins, isoprenoid‐derived fatty acids, and pristanic acid
(a  by‐product of α‐oxidation). The four steps involved
in  ­peroxisomal β‐oxidation are similar to those found in
146 THE LIVER:  PEROXISOME FUNCTIONS AND DISORDERS IN THE LIVER
mitochondria; however, the specific enzymes involved are dif-
ferent. The fatty acid substrates are first activated to CoA esters,
before being transported across the peroxisomal membrane
by  one of the three ABC transporters (ABCD1–3) [78].
Peroxisomal β‐oxidation occurs via a sequential cycle of dehy-
drogenation (by acyl‐CoA oxidase 1–3, ACOX), hydration,
dehydrogenation (by multifunctional protein 1 and 2, MFP),
and thiolytic cleavage (by sterol carrier protein x, SCPx). The
resulting chain‐shortened (by 2 carbons) fatty acids either
enter a new cycle of β‐oxidation or are transported out of the
peroxisome [79, 80].
Liver dysfunction was often reported in patients with single
peroxisomal β‐oxidation enzyme deficiencies; however, the
clinical presentation is diverse. For example, patients with
mutations in ABCD1 (also known as X‐lined adrenoleukod-
ystropy, which is the most frequent peroxisomal disease) [81]
and SCPx did not develop liver abnormalities [82]. Also, no
liver disease was reported in patients with mild deficiency in
MFP2 and ACOX1, who survived into adulthood [83, 84].
However, patients with severe mutations in MFP2 possess both
hepatic and neuronal abnormalities and died within the first 2
years of life without achieving any developmental milestones.
Microscopic analysis on liver biopsies and post‐mortem liver
showed enlarged peroxisomes that are reduced in number in
these patients [85]. Moreover, a single patient with a defect in
ABCD3 was reported to present with hepatosplenomegaly and
severe liver disease at the age of 1.5 years. This patient was born
with no neonatal problems and achieved normal developmental
milestones. Microscopic analysis of cultured primary skin fibro-
blasts revealed that peroxisomes appeared to be reduced in
number and enlarged in size. The liver disease progressed and
the patient developed decompensated liver cirrhosis associated
with hepatopulmonary syndrome. The patient died shortly after
liver transplantation due to respiratory complications at 4 years
of age [86].
Fatty acid α‐oxidation
Fatty acids with the presence of a methyl group at the
3‐position, including phytanic acid, cannot be metabolized
by the β‐oxidation pathway. Instead, they need to be first
α‐oxidized in peroxisomes to pristanic acid, which then can be
further metabolized via the β‐oxidation pathway in peroxi-
somes and mitochondria. Phytanic acid is found in almost all
animals, including humans; however, it is not endogenously
synthesized in animals but rather is a constituent of their diet.
Phytanic acid can be efficiently absorbed by animals in the
form of either phytanic acid itself or its precursors, such as
phytol. Phytol is widely found in the green leaves of plants and
trees, and it is linked to the porphytin ring of chlorophyll via
an ester bond. It cannot be digested by humans, as it can only
be liberated from chlorophyll by bacteria found in the rumen
of ruminant animals. Therefore, humans obtain phytanic acid
primarily from the consumption of dairy products and meat
originating from ruminant animals [80, 87].
To initiate the α‐oxidation pathway, phytanic acid is first
activated to phytanoyl‐CoA by the acyl‐CoA synthetases
­ Analysis of plasma bile acids from patients with peroxisomal
ACSL1 (found in mitochondria, the ER, and cytosol) and
ACSVL1 (found in the ER and peroxisomes). Phytanoyl‐CoA
then undergoes sequential reactions catalyzed by the p­ eroxisomal
enzymes phytanoyl‐CoA 2‐hydroxylase (PHYH), 2‐hydroxy-
phytanoyl‐CoA lyase (HACL1), and pristanal dehydrogenase
(PrDH), resulting in chain shortening of one carbon to produce
a 2‐methyl fatty acid, pristanoyl‐CoA. Defect in α‐oxidation in
humans leads to an autosomal recessive neurological disease
called Refsume disease, which is characterized by the accumu-
lation of phytanic acid in the plasma and tissues including the
liver [80, 87]. It is caused by mutations in the PHYH gene
encoding the peroxisomal enzyme involved in the first step of
α‐oxidation, or in PEX7 gene encoding the receptor for PTS2‐
containing proteins such as PHYH [88–90]. Compared to other
peroxisomal disorders, the onset of Refsume disease occurs
later in childhood with a progressive course. The clinical fea-
tures include retinitis pigmentosa, anosmia, deafness, chronic
polyneuropathy, and ichthyosis. Refsum disease is one of the
few peroxisomal disorders for which a therapy has been devel-
oped. The current therapy calls for controlling or eliminating
phytanic acid from the diet [80, 87].
Bile acid synthesis
In addition to the metabolism of fatty acids, peroxisomal β‐oxi-
dation also plays an important role in the synthesis of bile acids,
as it is required for converting the C27 bile acid intermediates to
the mature C24 bile acids. Cholesterol is converted into bile
acids via sequential reactions catalyzed by multiple enzymes
that are predominantly expressed in the liver. After the ring
modification of cholesterol followed by oxidation and shorten-
ing of the sterol side‐chain, the resulting C27 bile acid interme-
diates di‐ and trihydroxycholestanoic acid (DHCA and THCA)
are first activated to their CoA esters by the very‐long‐chain
acyl‐CoA synthetase (VLCS) found at both the ER and peroxi-
somes or the bile acyl‐CoA synthetase (BACS) found exclu-
sively at the ER (and hepatic) [91]. The subsequent uptake of
these CoA esters into peroxisomes is likely via ABCD3, as a
patient with a defect in ABCD3 presented with hepatospleno-
megaly and severe liver disease with a significant increased
level of peroxisomal C27 bile acid intermediates in plasma [86].
In Abcd3−/− mice, a reduction of C24 bile acids and an accumu-
lation of C27 bile acid intermediates were also found in liver,
bile, and intestine, indicating a role of ABCD3 in transporting
C27 bile acids into the peroxisomes [86]. After transporting into
the peroxisomes, the CoA‐ester form of DHCA and THCA (i.e.
DHC‐CoA and THC‐CoA) are first converted to chenodeoxy-
choloyl‐CoA (CDC‐CoA) and choloyl‐CoA (CA‐CoA), respec-
tively, via the enzyme α‐methylacyl‐CoA racemase (AMACR)
[92], followed by the β‐oxidation of these substrates. The CoA‐
ester of the C27 bile acid intermediate is oxidized by the
branched‐chain acyl‐CoA oxidase (BCOX2) [93], and the sub-
sequent steps of β‐oxidation are catalyzed by MFP2 and SCPx
[82, 85, 94]. The last step of bile acid synthesis is catalyzed by
the bile acyl‐CoA amino acid N‐acyltransferase (BAAT), which
conjugates the newly formed primary bile acids to taurine and
glycine [95, 96]. The conjugated bile acids are then transported
out of hepatic peroxisomes and excreted into the bile.
disorders, including PBDs and peroxisomal β‐oxidation disor-
ders, invariably shows increased levels of C27 bile acid. As C27
 12:  Peroxisome Assembly, Degradation, and Disease 147
bile acids are only partially conjugated by BAAT and therefore
less efficiently excreted in the bile as compared to C24 bile
acids [96], fat malabsorption is a common clinical feature in
these patients. Interestingly, the extent of peroxisome deficiency
has a clear correlation with the extent of the deficiency in bile
acid synthesis, as patients with the less severe forms of PBD
(i.e. infantile Refsum disease and neonatal adrenoleukodystro-
phy) presented less cholestasis and lower levels of bile acid
­precursors in serum than those with the most severe form (i.e.
Zellweger syndrome) [97]. Bile acid abnormalities are sug-
gested to contribute to the liver pathology in patients with
­peroxisomal disorders. At physiological concentrations during
cholestasis, bile acids did not cause apoptosis or necrosis [98].
Instead, they are thought to directly activate a signaling pathway
in hepatocytes that promotes hepatic inflammation during
­cholestasis, thereby leading to liver injury [99].
Ether phospholipid biosynthesis
Ether phospholipids are a specialized group of phospholipids
with an ether bond at the sn‐1 position of the glycerol backbone.
The ether bond can be either a regular ether bond or a vinyl–
ether bond that is present in plasmanyl‐phopholipids or plasme-
nyl‐phospholipids, respectively. The most abundant form of
plasmenyl‐phospholipids are plasmalogens which contain an
ethanolamine or a choline moiety in the head group of the glyc-
erol backbone. They are enriched in the heart, kidney, lung, and
skeletal muscle. Especially in the brain, they represent up to
90% of the phosphatidylethanolamine fraction [100, 101]. The
vinyl–ether bond at the sn‐1 position and enrichment of polyun-
saturated fatty acids at the sn‐2 position provide plasmalogens
with unique features that allow them to function as: (i) mediators
for maintaining membrane physical bilayer properties; (ii)
­sacrificial oxidants; and (iii) reservoirs for biologically active
lipid mediators [100].
Plasmalogens are synthesized in concert between peroxi-
somes and the ER. The peroxisomal matrix enzyme glycerone-
phosphate O‐acyltransferase (GNPAT) initiates the biosynthesis
pathway by catalyzing the acylation of dihydroxyacetine phos-
phate (DHAP) at the sn‐1 position, followed by the exchange of
acyl group for an alkyl group by alkylglycerone phosphate syn-
thase (AGPS). Further modifications are completed in the ER,
resulting in the formation of mature plasmalogens [100]. Defect
in plasmalogen synthesis leads to the peroxisomal disorder rhi-
zomelic chondrodysplasia punctate (RCDP), which impairs the
normal development of multiple organs of the body, including
bone, brain, lens, lung, kidney, and heart [100, 101]. Depending
on which gene is mutated, there are three forms of RCDP:
RCDP type 1 (mutation in PEX7, encoding the receptor for
PTS2‐containing proteins, such as APGS) [102], RCDP type 2
(mutation in GNPAT, encoding the enzyme involved in the first
step of the plasmalogen biosynthesis) [103], and RCDP type 3
(mutation in APGS, encoding the enzyme involved in the second
step of the plasmalogen biosynthesis) [104]. Interestingly,
reduced levels of plasmalogens have also been reported in
nonperoxisomal disorders, such as Alzheimer’s disease.
­ PCa, is also shown to be essential for optimal proliferation of
However, whether the reduction in plasmalogens is the cause of
these diseases or a downstream effect needs to be further
­investigated [105]. Plasmalogen replacement therapy may be a
potential treatment for RCDP patients. However, the low trans-
port efficiency through lactation and low incorporation into
brain of the target plasmalogen limit the beneficial effects of
this treatment in the mouse studies [106, 107].
PEROXISOMES IN ANTIVIRAL
INNATE IMMUNITY
Besides playing a central role in various metabolic activities as
described above, peroxisomes have also been proposed to
­participate in the innate immune response by recent research.
Upon infection, microorganisms are detected by the pattern
­recognition receptors (PRRs) that activate multiple signaling
transduction pathways, leading to the activation of nuclear
factor‐κB (NF‐κB), mitogen‐activated protein kinase (MARKs),
and interferon regulatory factors (IRFs). These transcription
factors will then induce the expression of various proinflamma-
tory factors, including the cytokines interleukin‐1, TNF, and
type I and III interferons (IFNs), thereby creating an antiviral
state within the cell [108].
The adaptor protein for the RIG‐I‐like receptor (RLR) family,
mitochondrial antiviral signaling protein (MAVS), has been
recently identified to be present on both mitochondria and per-
oxisomes in mammalian cells, including human hepatocytes
[109]. During viral infection, the peroxisomal MAVS was
shown to trigger an immediate induction of antiviral gene
expression in a type I IFN‐independent manner to provide a
short‐term antiviral effect. In contrast, the mitochondrial MAVS
activates an interferon‐dependent signaling pathway, but with
a  delayed kinetics, for a sustained protection later during
­infection [109]. The coordinate between the peroxisomal and
mitochondrial MAVS is proposed to occur at a subdomain on
the ER called mitochondria‐associated membranes (MAMs)
[110]. During viral infection, peroxisomes and mitochondria are
found to interact with each other at the MAM, which acts as an
“innate immune synapse” that coordinates the role of peroxiso-
mal and mitochondrial MAVS.
PEROXISOMES IN CANCER
So far, not much is known about the role of peroxisomes in
human cancer development. Reduced peroxisome abundance
has been observed in multiple cancer cells, including hepatocel-
lular carcinoma [111], colon carcinoma [112], breast cancer
[113], and renal cell carcinoma [114]. However, the mechanism
that leads to the loss of peroxisomes in these cancer cells is still
unclear. Moreover, the protein levels of peroxisomal enzymes
involved in branched‐chain fatty acid β‐oxidation, α‐methyla-
cyl‐CoA racemase (AMACR) and peroxisomal multifunctional
protein 2, are found to be upregulated in human prostate cancer
(PCa) [115]. AMACR, one of the established biomarkers in
some PCa cell lines in vitro [116]. Monocarboxylate transporter
2 (MCT2), a putative biomarker for PCa, has recently been
shown to localize mainly at peroxisomes in PCa cells. MCT2
148 THE LIVER:  REFERENCES
expression was upregulated significantly from nonmalignant to
malignant cells, and this increase in protein expression is
directly correlated with its peroxisomal localization, implying a
possible role of peroxisomes in prostate malignant transforma-
tion [117]. However, it remains unclear whether peroxisome
abundance is also increased in PCa.
REFERENCES
1. Rhodin, J. Correlation of ultrastructural organization and function in normal
and experimentally changed proximal convoluted tubule cells of the mouse
kidney. Thesis, Aktiebolaget Godvil Stockholm, Karolinska Institute, 1954.
2. De Duve, C. and Baudhuin, P. Peroxisomes (microbodies and related parti-
cles). Physiol Rev, 1966;46(2):323–57.
3. Steinberg, S.J., Dodt, G., Raymond, G.V. et al. Peroxisome biogenesis disor-
ders. Biochim Biophys Acta, 2006;1763(12):1733–48.
4. Erdmann, R., Veenhuis, M., Mertens, D., and Kunau WH. Isolation of per-
oxisome‐deficient mutants of Saccharomyces cerevisiae. Proc Natl Acad Sci
U S A, 1989;86(14):5419–23.
5. Tsukamoto, T., Yokota, S., and Fujiki Y. Isolation and characterization of
Chinese hamster ovary cell mutants defective in assembly of peroxisomes. J
Cell Biol, 1990;110(3):651–60.
6. Fujiki, Y., Okumoto, K., Mukai, S., Honsho, M., and Tamura, S. Peroxisome
biogenesis in mammalian cells. Front Physiol, 2014;5:307.
7. Prinz, W.A. Bridging the gap: membrane contact sites in signaling, metabo-
lism, and organelle dynamics. J Cell Biol, 2014;205(6):759–69.
8. Shai, N., Schuldiner, M., and Zalckvar, E. No peroxisome is an island  –
Peroxisome contact sites. Biochim Biophys Acta, 2016;1863(5):1061–9.
9. Cho, D.H., Kim, Y.S., Jo, D.S., Choe, S.K., and Jo, E.K. Pexophagy: molecu-
lar mechanisms and implications for health and diseases. Mol Cells,
2018;41(1):55–64.
10. Deosaran, E., Larsen, K.B., Hua, R. et al. NBR1 acts as an autophagy recep-
tor for peroxisomes. J Cell Sci, 2013;126(Pt 4):939–52.
11. Lazarow, P.B. and Fujiki Y. Biogenesis of peroxisomes. Annu Rev Cell Biol,
1985;1:489–530.
12. Koch, J. and Brocard, C. PEX11 proteins attract Mff and human Fis1 to
coordinate peroxisomal fission. J Cell Sci, 2012;125(Pt 16):3813–26.
13. Delille, H.K., Dodt, G., and Schrader M. Pex11pbeta‐mediated maturation of
peroxisomes. Commun Integr Biol, 2011;4(1):51–4.
14. Honsho, M., Tamura, S., Shimozawa, N. et al. Mutation in PEX16 is causal
in the peroxisome‐deficient Zellweger syndrome of complementation group
D. Am J Hum Genet, 1998;63(6):1622–30.
15. Matsuzono, Y., Kinoshita, N., Tamura, S. et al. Human PEX19: cDNA clon-
ing by functional complementation, mutation analysis in a patient with
Zellweger syndrome, and potential role in peroxisomal membrane assembly.
Proc Natl Acad Sci U S A, 1999;96(5):2116–21.
16. Ghaedi, K., Tamura, S., Okumoto, K., Matsuzono, Y., and Fujiki, Y. The per-
oxin pex3p initiates membrane assembly in peroxisome biogenesis. Mol Biol
Cell, 2000;11(6):2085–102.
17. Titorenko, V.I., Chan, H., and Rachubinski RA. Fusion of small peroxisomal
vesicles in vitro reconstructs an early step in the in vivo multistep peroxisome
assembly pathway of Yarrowia lipolytica. J Cell Biol, 2000;148(1):29–44.
18. van der Zand, A., Gent, J., Braakman, I., and Tabak, H.F. Biochemically
distinct vesicles from the endoplasmic reticulum fuse to form peroxisomes.
Cell, 2012;149(2):397–409.
19. Knoops, K., Manivannan, S., Cepinska, M.N. et al. Preperoxisomal vesicles
can form in the absence of Pex3. J Cell Biol, 2014;204(5):659–68.
20. Geuze, H.J., Murk, J.L., Stroobants, A.K. et al. Involvement of the endoplas-
mic reticulum in peroxisome formation. Mol Biol Cell, 2003;14(7):2900–7.
21. Hashiguchi, N., Kojidani, T., Imanaka, T. et al. Peroxisomes are formed from
complex membrane structures in PEX6‐deficient CHO cells upon genetic
complementation. Mol Biol Cell, 2002;13(2):711–22.
22. Baes, M., Gressens, P., Baumgart, E. et  al. A mouse model for Zellweger
syndrome. Nat Genet, 1997;17(1):49–57.
23. Hua, R., Gidda, S.K., Aranovich, A., Mullen, R.T., and Kim PK. Multiple
domains in PEX16 mediate its trafficking and recruitment of peroxisomal
proteins to the ER. Traffic, 2015;16(8):832–52.
24. South, S.T., Sacksteder, K.A., Li, X., Liu, Y., and Gould, S.J. Inhibitors of
COPI and COPII do not block PEX3‐mediated peroxisome synthesis. J Cell
Biol, 2000;149(7):1345–60.
25. Latruffe, N. and Vamecq, J. Peroxisome proliferators and peroxisome prolif-
erator activated receptors (PPARs) as regulators of lipid metabolism.
Biochimie, 1997;79(2–3):81–94.
26. Shimizu, M., Takeshita, A., Tsukamoto, T., Gonzalez, F.J., and Osumi, T.
Tissue‐selective, bidirectional regulation of PEX11 alpha and perilipin genes
through a common peroxisome proliferator response element. Mol Cell Biol,
2004;24(3):1313–23.
27. Muntau, A.C., Mayerhofer, P.U., Paton, B.C., Kammerer, S., and Roscher,
A.A. Defective peroxisome membrane synthesis due to mutations in human
PEX3 causes Zellweger syndrome, complementation group G. Am J Hum
Genet, 2000;67(4):967–75.
28. Jones, J.M., Morrell, J.C., and Gould, S.J. PEX19 is a predominantly
­cytosolic chaperone and import receptor for class 1 peroxisomal membrane
proteins. J Cell Biol, 2004;164(1):57–67.
29. McNew, J.A. and Goodman, J.M. An oligomeric protein is imported into
peroxisomes in vivo. J Cell Biol, 1994;127(5):1245–57.
30. Walton, P.A., Hill, P.E., and Subramani, S. Import of stably folded proteins
into peroxisomes. Mol Biol Cell, 1995;6(6):675–83.
31. Platta, H.W. and Erdmann, R. Peroxisomal dynamics. Trends Cell Biol,
2007;17(10):474–84.
32. Brocard, C. and Hartig, A. Peroxisome targeting signal 1: is it really a simple
tripeptide? Biochim Biophys Acta, 2006;1763(12):1565–73.
33. Stanley, W.A., Filipp, F.V., Kursula, P. et al. Recognition of a functional per-
oxisome type 1 target by the dynamic import receptor pex5p. Mol Cell,
2006;24(5):653–63.
34. Lazarow, P.B. The import receptor Pex7p and the PTS2 targeting sequence.
Biochim Biophys Acta, 2006;1763(12):1599–604.
35. Schliebs, W. and Kunau, W.H. PTS2 co‐receptors: diverse proteins with
common features. Biochim Biophys Acta, 2006;1763(12):1605–12.
36. van der Klei, I.J. and Veenhuis, M. PTS1‐independent sorting of peroxisomal
matrix proteins by Pex5p. Biochim Biophys Acta, 2006;1763(12):
1794–800.
37. Williams, C. and Distel, B. Pex13p: docking or cargo handling protein?
Biochim Biophys Acta, 2006;1763(12):1585–91.
38. Meinecke, M., Bartsch, P., and Wagner, R. Peroxisomal protein import pores.
Biochim Biophys Acta, 2016;1863(5):821–7.
39. Meinecke, M., Cizmowski, C., Schliebs, W. et al. The peroxisomal importomer
constitutes a large and highly dynamic pore. Nat Cell Biol, 2010;12(3):273–7.
40. Freitas, M.O., Francisco, T., Rodrigues, T.A. et al. PEX5 protein binds mon-
omeric catalase blocking its tetramerization and releases it upon binding the
N‐terminal domain of PEX14. J Biol Chem, 2011;286(47):40509–19.
41. Platta, H.W., Hagen, S., Reidick, C., and Erdmann, R. The peroxisomal
receptor dislocation pathway: to the exportomer and beyond. Biochimie,
2014;98:16–28.
42. Helle, S.C., Kanfer, G., Kolar, K. et al. Organization and function of mem-
brane contact sites. Biochim Biophys Acta, 2013;1833(11):2526–41.
43. Raychaudhuri, S. and Prinz, W.A. Nonvesicular phospholipid transfer
between peroxisomes and the endoplasmic reticulum. Proc Natl Acad Sci U
S A, 2008;105(41):15785–90.
44. David, C., Koch, J., Oeljeklaus, S. et al. A combined approach of quantita-
tive interaction proteomics and live‐cell imaging reveals a regulatory role for
endoplasmic reticulum (ER) reticulon homology proteins in peroxisome bio-
genesis. Mol Cell Proteomics, 2013;12(9):2408–25.
45. Hua, R., Cheng, D., Coyaud, E. et al. VAPs and ACBD5 tether peroxisomes
to the ER for peroxisome maintenance and lipid homeostasis. J Cell Biol,
2017;216(2):367–77.
46. Costello, J.L., Castro, I.G., Hacker, C. et  al. ACBD5 and VAPB mediate
membrane associations between peroxisomes and the ER. J Cell Biol,
2017;216(2):331–42.
47. Mattiazzi Usaj, M., Brloznik, M., Kaferle, P. et al. Genome‐wide localiza-
tion study of yeast pex11 identifies peroxisome‐mitochondria interactions
through the ERMES complex. J Mol Biol, 2015;427(11):2072–87.
48. McGuinness, M.C., Lu, J.F., Zhang, H.P. et  al. Role of ALDP (ABCD1)
and  mitochondria in X‐linked adrenoleukodystrophy. Mol Cell Biol,
2003;23(2):744–53.
49. Neuspiel, M., Schauss, A.C., Braschi, E. et  al. Cargo‐selected transport
from  the mitochondria to peroxisomes is mediated by vesicular carriers.
Curr Biol, 2008;18(2):102–8.
 12:  Peroxisome Assembly, Degradation, and Disease 149
50. Gluchowski, N.L., Becuwe, M., Walther, T.C., and Farese, R.V., Jr. Lipid
droplets and liver disease: from basic biology to clinical implications.
Nat Rev Gastroenterol Hepatol, 2017;14(6):343–55.
51. Beller, M., Thiel, K., Thul, P.J., and Jackle, H. Lipid droplets: a dynamic
organelle moves into focus. FEBS Lett, 2010;584(11):2176–82.
52. Schrader, M. Tubulo‐reticular clusters of peroxisomes in living COS‐7 cells:
dynamic behavior and association with lipid droplets. J Histochem Cytochem,
2001;49(11):1421–29.
53. Dirkx, R., Vanhorebeek, I., Martens, K. et  al. Absence of peroxisomes in
mouse hepatocytes causes mitochondrial and ER abnormalities. Hepatology,
2005;41(4):868–78.
54. Binns, D., Januszewski, T., Chen, Y. et al. An intimate collaboration between
peroxisomes and lipid bodies. J Cell Biol, 2006;173(5):719–31.
55. Schrul, B. and Kopito RR. Peroxin‐dependent targeting of a lipid‐droplet‐
destined membrane protein to ER subdomains. Nat Cell Biol,
2016;18(7):740–51.
56. Chu, B.B., Liao, Y.C., Qi, W. et al. Cholesterol transport through lysosome‐
peroxisome membrane contacts. Cell, 2015;161(2):291–306.
57. Chang, T.Y., Chang, C.C., Ohgami, N., and Yamauchi Y. Cholesterol sensing,
trafficking, and esterification. Annu Rev Cell Dev Biol, 2006;22:129–57.
58. Zientara‐Rytter, K. and Subramani S. Autophagic degradation of peroxi-
somes in mammals. Biochem Soc Trans, 2016;44(2):431–40.
59. Omi, S., Nakata, R., Okamura‐Ikeda, K., Konishi, H., and Taniguchi H.
Contribution of peroxisome‐specific isoform of Lon protease in sorting
PTS1 proteins to peroxisomes. J Biochem, 2008;143(5):649–60.
60. Yokota, S., Haraguchi, C.M., and Oda T. Induction of peroxisomal Lon pro-
tease in rat liver after di‐(2‐ethylhexyl)phthalate treatment. Histochem Cell
Biol, 2008;129(1):73–83.
61. Bartel, B., Farmer, L.M., Rinaldi, M.A. et al. Mutation of the Arabidopsis
LON2 peroxisomal protease enhances pexophagy. Autophagy,
2014;10(3):518–19.
62. Maccarrone, M., Melino, G., and Finazzi‐Agro, A. Lipoxygenases and their
involvement in programmed cell death. Cell Death Differ, 2001;8(8):776–84.
63. van Leyen, K., Duvoisin, R.M., Engelhardt, H., and Wiedmann M. A func-
tion for lipoxygenase in programmed organelle degradation. Nature,
1998;395(6700):392–5.
64. Yokota, S., Oda, T., and Fahimi, H.D. The role of 15‐lipoxygenase in disrup-
tion of the peroxisomal membrane and in programmed degradation of per-
oxisomes in normal rat liver. J Histochem Cytochem, 2001;49(5):613–22.
65. Morgan, A.H., Hammond, V.J., Sakoh‐Nakatogawa, M. et al. A novel role
for 12/15‐lipoxygenase in regulating autophagy. Redox Biol, 2015;4:40–7.
66. Feng, Y., He, D., Yao, Z., and Klionsky, D.J. The machinery of macroau-
tophagy. Cell Res, 2014;24(1):24–41.
67. Gatica, D., Lahiri, V., and Klionsky, D.J. Cargo recognition and degradation
by selective autophagy. Nat Cell Biol, 2018;20(3):233–42.
68. Kim, P.K., Hailey, D.W., Mullen, R.T., and Lippincott‐Schwartz, J. Ubiquitin
signals autophagic degradation of cytosolic proteins and peroxisomes. Proc
Natl Acad Sci U S A, 2008;105(52):20567–74.
69. Yamashita, S., Abe, K., Tatemichi, Y., and Fujiki Y. The membrane peroxin
PEX3 induces peroxisome‐ubiquitination‐linked pexophagy. Autophagy,
2014;10(9):1549–64.
70. Nordgren, M., Francisco, T., Lismont, C. et al. Export‐deficient monoubiq-
uitinated PEX5 triggers peroxisome removal in SV40 large T antigen‐­
transformed mouse embryonic fibroblasts. Autophagy, 2015;11(8):1326–40.
71. Zhang, J., Tripathi, D.N., Jing, J. et  al. ATM functions at the peroxisome
to  induce pexophagy in response to ROS. Nat Cell Biol, 2015;
17(10):1259–69.
72. Okumoto, K., Noda, H., and Fujiki, Y. Distinct modes of ubiquitination of
peroxisome‐targeting signal type 1 (PTS1) receptor Pex5p regulate PTS1
protein import. J Biol Chem, 2014;289(20):14089–108.
73. Wang, W., Xia, Z.J., Farre, J.C., and Subramani, S. TRIM37, a novel E3
ligase for PEX5‐mediated peroxisomal matrix protein import. J Cell Biol,
2017;216(9):2843–58.
74. Sargent, G., van Zutphen, T., Shatseva, T. et  al. PEX2 is the E3 ubiquitin
ligase required for pexophagy during starvation. J Cell Biol,
2016;214(6):677–90.
75. Nazarko, T.Y. Pexophagy is responsible for 65% of cases of peroxisome bio-
genesis disorders. Autophagy, 2017;13(5):991–4.
76. Law, K.B., Bronte‐Tinkew, D., Di Pietro, E. et  al. The peroxisomal AAA
ATPase complex prevents pexophagy and development of peroxisome bio-
genesis disorders. Autophagy, 2017;13(5):868–84.
77. Johansen, T. and Lamark, T. Selective autophagy mediated by autophagic
adapter proteins. Autophagy, 2011;7(3):279–96.
78. Morita, M. and Imanaka, T. Peroxisomal ABC transporters: structure, func-
tion and role in disease. Biochim Biophys Acta, 2012;1822(9):1387–96.
79. Baes, M. and Van Veldhoven, P.P. Hepatic dysfunction in peroxisomal
­disorders. Biochim Biophys Acta, 2016;1863(5):956–70.
80. Waterham, H.R., Ferdinandusse, S., and Wanders, R.J. Human disorders
of  peroxisome metabolism and biogenesis. Biochim Biophys Acta,
2016;1863(5):922–33.
81. Kemp, S., Berger, J., and Aubourg, P. X‐linked adrenoleukodystrophy:
clinical, metabolic, genetic and pathophysiological aspects. Biochim
Biophys Acta, 2012;1822(9):1465–74.
82. Ferdinandusse, S., Kostopoulos, P., Denis, S. et al. Mutations in the gene
encoding peroxisomal sterol carrier protein X (SCPx) cause leukencepha-
lopathy with dystonia and motor neuropathy. Am J Hum Genet,
2006;78(6):1046–52.
83. Lines, M.A., Jobling, R., Brady, L. et al. Peroxisomal D‐bifunctional pro-
tein deficiency: three adults diagnosed by whole‐exome sequencing.
Neurology, 2014;82(11):963–8.
84. Ferdinandusse, S., Barker, S., Lachlan, K. et al. Adult peroxisomal acyl‐
coenzyme A oxidase deficiency with cerebellar and brainstem atrophy. J
Neurol Neurosurg Psychiatry, 2010;81(3):310–12.
85. Ferdinandusse, S., Denis, S., Mooyer, P.A. et al. Clinical and biochemical spec-
trum of D‐bifunctional protein deficiency. Ann Neurol, 2006;59(1):92–104.
86. Ferdinandusse, S., Jimenez‐Sanchez, G., Koster, J. et al. A novel bile acid
biosynthesis defect due to a deficiency of peroxisomal ABCD3. Hum Mol
Genet, 2015;24(2):361–70.
87. Wanders, R.J., Komen, J., and Ferdinandusse, S. Phytanic acid metabolism
in health and disease. Biochim Biophys Acta, 2011;1811(9):498–507.
88. Jansen, G.A., Waterham, H.R., and Wanders, R.J. Molecular basis of
Refsum disease: sequence variations in phytanoyl‐CoA hydroxylase
(PHYH) and the PTS2 receptor (PEX7). Hum Mutat, 2004;23(3):209–18.
89. Mihalik, S.J., Morrell, J.C., Kim, D. et  al. Identification of PAHX, a
Refsum disease gene. Nat Genet, 1997;17(2):185–9.
90. van den Brink, D.M., Brites, P., Haasjes, J. et al. Identification of PEX7 as
the second gene involved in Refsum disease. Am J Hum Genet,
2003;72(2):471–7.
91. Mihalik, S.J., Steinberg, S.J., Pei, Z. et al. Participation of two members of
the very long‐chain acyl‐CoA synthetase family in bile acid synthesis and
recycling. J Biol Chem, 2002;277(27):24771–9.
92. Schmitz, W., Albers, C., Fingerhut, R., and Conzelmann, E. Purification
and characterization of an alpha‐methylacyl‐CoA racemase from human
liver. Eur J Biochem, 1995;231(3):815–22.
93. Vanhove, G.F., Van Veldhoven, P.P., Fransen, M. et al. The CoA esters of
2‐methyl‐branched chain fatty acids and of the bile acid intermediates di‐
and trihydroxycoprostanic acids are oxidized by one single peroxisomal
branched chain acyl‐CoA oxidase in human liver and kidney. J Biol Chem,
1993;268(14):10335–44.
94. Novikov, D., Dieuaide‐Noubhani, M., Vermeesch, J.R. et  al. The human
peroxisomal multifunctional protein involved in bile acid synthesis: activity
measurement, deficiency in Zellweger syndrome and chromosome map-
ping. Biochim Biophys Acta, 1997;1360(3):229–40.
95. He, D., Barnes, S., and Falany, C.N. Rat liver bile acid CoA:amino acid
N‐acyltransferase: expression, characterization, and peroxisomal localiza-
tion. J Lipid Res, 2003;44(12):2242–9.
96. Solaas, K., Ulvestad, A., Soreide, O., and Kase BF. Subcellular organiza-
tion of bile acid amidation in human liver: a key issue in regulating the
biosynthesis of bile salts. J Lipid Res, 2000;41(7):1154–62.
97. Van Eldere, J.R., Parmentier, G.G., Eyssen, H.J. et al. Bile acids in peroxi-
somal disorders. Eur J Clin Invest, 1987;17(5):386–90.
98. Zhang, Y., Hong, J.Y., Rockwell, C.E. et al. Effect of bile duct ligation on
bile acid composition in mouse serum and liver. Liver Int,
2012;32(1):58–69.
99. Allen, K., Jaeschke, H., and Copple BL. Bile acids induce inflammatory
genes in hepatocytes: a novel mechanism of inflammation during obstruc-
tive cholestasis. Am J Pathol, 2011;178(1):175–86.
100. Braverman, N.E. and Moser, A.B. Functions of plasmalogen lipids in health
and disease. Biochim Biophys Acta, 2012;1822(9):1442–52.
101. da Silva, T.F., Sousa, V.F., Malheiro, A.R., and Brites, P. The importance of
ether‐phospholipids: a view from the perspective of mouse models. Biochim
Biophys Acta, 2012;1822(9):1501–8.
150 THE LIVER:  REFERENCES
102. Purdue, P.E., Skoneczny, M., Yang, X., Zhang, J.W., and Lazarow, P.B.
Rhizomelic chondrodysplasia punctata, a peroxisomal biogenesis disorder
caused by defects in Pex7p, a peroxisomal protein import receptor: a mini-
review. Neurochem Res, 1999;24(4):581–6.
103. Ofman, R., Hettema, E.H., Hogenhout, E.M. et  al. Acyl‐CoA:
dihydroxyacetonephosphate acyltransferase: cloning of the human cDNA
and resolution of the molecular basis in rhizomelic chondrodysplasia punc-
tata type 2. Hum Mol Genet, 1998;7(5):847–53.
104. de Vet, E.C., Ijlst, L., Oostheim, W., Wanders, R.J., and van den Bosch, H.
Alkyl‐dihydroxyacetonephosphate synthase. Fate in peroxisome biogene-
sis disorders and identification of the point mutation underlying a single
enzyme deficiency. J Biol Chem, 1998;273(17):10296–301.
105. Han, X., Holtzman, D.M., and McKeel, D.W., Jr. Plasmalogen deficiency in
early Alzheimer’s disease subjects and in animal models: molecular charac-
terization using electrospray ionization mass spectrometry. J Neurochem,
2001;77(4):1168–80.
106. Brites, P., Ferreira, A.S., da Silva, T.F. et al. Alkyl‐glycerol rescues plasm-
alogen levels and pathology of ether‐phospholipid deficient mice. PLoS
One, 2011;6(12):e28539.
107. Das, A.K., Holmes, R.D., Wilson, G.N., and Hajra, A.K. Dietary ether lipid
incorporation into tissue plasmalogens of humans and rodents. Lipids,
1992;27(6):401–5.
108. Odendall, C. and Kagan, J.C. Peroxisomes and the antiviral responses of
mammalian cells. Subcell Biochem, 2013;69:67–75.
109. Dixit, E., Boulant, S., Zhang, Y. et al. Peroxisomes are signaling platforms
for antiviral innate immunity. Cell, 2010;141(4):668–81.
110. Horner, S.M., Liu, H.M., Park, H.S., Briley, J., and Gale, M., Jr.
Mitochondrial‐associated endoplasmic reticulum membranes (MAM) form
innate immune synapses and are targeted by hepatitis C virus. Proc Natl
Acad Sci U S A, 2011;108(35):14590–5.
111. Litwin, J.A., Beier, K., Volkl, A., Hofmann, W.J., and Fahimi, H.D.
Immunocytochemical investigation of catalase and peroxisomal lipid beta‐
oxidation enzymes in human hepatocellular tumors and liver cirrhosis.
Virchows Arch, 1999;435(5):486–95.
112. Lauer, C., Volkl, A., Riedl, S., Fahimi, H.D., and Beier, K. Impairment of
peroxisomal biogenesis in human colon carcinoma. Carcinogenesis,
1999;20(6):985–9.
113. el Bouhtoury, F., Keller, J.M., Colin, S., Parache, R.M., and Dauca M.
Peroxisomal enzymes in normal and tumoral human breast. J Pathol,
1992;166(1):27–35.
114. Frederiks, W.M., Bosch, K.S., Hoeben, K.A., van Marle, J., and Langbein,
S. Renal cell carcinoma and oxidative stress: the lack of peroxisomes. Acta
Histochem, 2010;112(4):364–71.
115. Zha, S., Ferdinandusse, S., Hicks, J.L. et al. Peroxisomal branched chain
fatty acid beta‐oxidation pathway is upregulated in prostate cancer.
Prostate, 2005;63(4):316–23.
116. Zha, S., Ferdinandusse, S., Denis, S. et  al. Alpha‐methylacyl‐CoA race-
mase as an androgen‐independent growth modifier in prostate cancer.
Cancer Res, 2003;63(21):7365–76.
117. Valenca, I., Pertega‐Gomes, N., Vizcaino, J.R. et al. Localization of MCT2
at peroxisomes is associated with malignant transformation in prostate can-
cer. J Cell Mol Med, 2015;19(4):723–33.
 Organelle–Organelle
13 Contacts: Origins and
Functions
Uri Manor
Waitt Advanced Biophotonics Center, Salk Institute for Biological Studies, La Jolla, CA, USA
INTRODUCTION
Eons ago, before the rise of the eukaryotic cell, single‐celled
protobacteria and archaeal cells dominated the biosphere. The
revolutionary theory of endosymbiosis, first posited by Lynn
Margulis, dictates that what we today call mitochondria were
actually prokaryotic cells (“protomitochondria”) that were
engulfed by a larger prokaryotic host cell in a symbiotic rela-
tionship [1]. Similarly, the nucleus has been proposed to be the
evolutionary progeny of an ancient endosymbiont that was
engulfed by our ancestral host cell. The precise ordering or
mechanism of these ancient endosymbiotic marriages is unclear:
Was it first mitochondria or the nucleus? How did the archaeal
cell engulf other cells long before the evolution of phago/endo-
cytotic machinery? Nonetheless, these critical events are now
accepted to be a prerequisite and the basis for the evolution of
multicellular life and all its complexities. It would be difficult to
overstate the impact endosymbiosis had on life and, indeed, the
planet Earth. Today, all eukaryotic life is both shaped and con-
strained by these early evolutionary events in ways that have
only recently started to become apparent. For this chapter, I will
first explain how recent work shows that early endosymbiosis as
traditionally understood requires a simple but profound modifi-
cation; it turns out that “ectosymbiosis” provides a much more
feasible and parsimonious model [2, 3]. Then I will explain how
this novel model for the evolution of the modern cell sets the
stage for everything we currently know about organelle biology,
in particular how interactions between the endoplasmic reticu-
lum (ER) and the rest of our membrane‐bound organelles (i.e.
mitochondria, lysosomes, peroxisomes, endosomes, and the
Golgi) reflect our ancestral archaea’s evolution into a modern
eukaryotic cell. With this context, we will then explore the
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
implications of this evolutionary history and the insight this pro-
vides towards the mechanisms for organelle–organelle interac-
tions in physiology and pathophysiology.
THE ORIGIN STORY OF THE ER, THE
CHIEF OPERATING OFFICER OF
THE NUCLEUS
As briefly mentioned above, until recently, all endosymbiotic
origin stories of the modern eukaryotic cell traditionally cast
the nucleus and mitochondria as the descendants of smaller
ancestral cells engulfed by a larger ancestral cell. This could be
described as an “outside‐in” model, in which endosymbionts
are engulfed by and evolved within the host cell. A recent study
strongly supports a conceptual inversion of this model to an
“inside‐out” model, wherein “ectosymbiotic” protomitochon-
dria closely interacted and exchanged materials with membrane
protrusions on the surface of a larger archaeal host cell [2]. For
many generations these two cell types continually exchanged
materials in mutually beneficial transactions that reinforced
symbiotic coevolution. That archaeal host cell is homologous
to what we now call the nucleus (aka “protonucleus”), with the
bleb extrusion sites representing what are now nuclear pore
complexes. The ancestral archaeal cell already evolved proteo-
lytic, N‐glycosylation, membrane scission, and cytoskeletal
machineries necessary for the cargo trafficking and modifica-
tion that exemplifies the Golgi and lysosomal compartments
[2]. The protonucleus and protomitochondria continued to
interact and the protonucleus’s blebs eventually expanded and
fused around the protomitochondria until they all formed an
152 THE LIVER:  THE ORIGIN STORY OF THE ER, THE CHIEF OPERATING OFFICER OF THE NUCLEUS

interconnected web that became what we now call the ER, This model is appealing in both its simplicity and its parsi-
mitochondria, cytoplasm, endomembrane system, and the mony. It is a more plausible model, in that it no longer requires
plasma membrane (Figure  13.1) [2]. In short, the modern the highly complicated, unlikely, and energetically expensive
eukaryotic cell. engulfment and subsequent endosymbiosis of a separate cell

(a) (b)
Epibiotic bacterium
(future mitochondrion)

Eocyte
(future nucleus)

(c) (d)

(e) (f)

original weakened half full nuclear

S-layer S-layer pore pore

Figure 13.1  Inside‐out model for the evolution of eukaryotic cell organization. Model showing the stepwise evolution of eukaryotic cell organiza-
tion from (a) an eocyte ancestor with a single bounding membrane and a glycoprotein rich cell wall (S‐layer) interacting with epibiotic α‐proteo-
bacteria (protomitochondria). (b) The eocyte cell forms protrusions, aided by protein–membrane interactions at the protrusion neck. These
protrusions facilitated material exchange with protomitochondria. (c) Selection for a greater area of contact between the symbionts would have led
to bleb enlargement and the eventual loss of the S‐layer from the protrusions. (d) Blebs would have then been further stabilized by the development
of a symmetric nuclear pore outer ring complex (Figure 13.2) and through the establishment of LINC complexes that, following the gradual loss of
the S‐layer, physically connected the original cell body (the nascent nuclear compartment) to the inner bleb membranes. (e) With the expansion of
blebs to enclose the protomitochondria, a process that would have facilitated the acquisition of bacterial lipid biosynthesis machinery by the host,
the site of cell growth would have progressively shifted to the cytoplasm, facilitated by the development of regulated traffic through the nuclear
pore. At the same time, the spaces between blebs would have enabled the gradual maturation of proteins secreted into the environment via the peri-
nuclear space through glycosylation and proteolytic cleavage. (f) Finally, bleb fusion would have connected cytoplasmic compartments and driven
the formation of an intact plasma membrane, perhaps through a process akin to phagocytosis whereby one bleb enveloped the whole. This simple
topological transition would have isolated the endoplasmic reticulum from the outside world, driven the full development of a system of vesicular
trafficking, and established strict vertical transmission of mitochondria, leading to a cell with modern eukaryotic cell organization. Reproduced
from Baum 2014 [2], https://bmcbiol.biomedcentral.com/articles/10.1186/s12915‐014‐0076‐2. Licensed under CCBY 4.0.
 13:  Organelle–Organelle Contacts: Origins and Functions 153

type for the protonucleus. Instead, the host cell is the nucleus,
meaning the only symbiotic partner left to account for is the
mitochondria. The inside‐out model of the coevolution of the
host cell and protomitochondria also makes more sense from an
evolutionary perspective: The exchange of materials with ectos-
ymbiotic protomitochondria across generations could facilitate
incremental advantages on evolutionary time‐scales necessary
for natural selection to develop a strong symbiotic relationship.
Most relevant to this chapter and textbook, the inside‐out model
is parsimonious because it contextualizes so much of modern
cell and organelle biology.
Conceptualizing the ancestral ER as an outgrowth of the host
cell’s probing interactions with its environment, it is not difficult
to see how it became the chief operating officer of our ancestral
host cell, what we today call the nucleus. The fact that the
nuclear membrane is continuous with the ER, that all the orga-
nelles are almost always in contact with the ER (Figure 13.2),
and that the ER plays a key role in mitochondrial and endo-
membrane dynamics are natural outcomes of this evolutionary
­history. Similarly, the essential roles of ER–organelle contacts
make much more sense in the context of having existed since the
eukaryotic primordial era. The largest organelle in the cell, the
ER physically occupies more of the cell than any other
­organelle  –  approximately 35%. When considering ER move-
ment over time, over 90% of the cellular volume is in contact
with the ER at one point or another in less than 5 minutes [4, 5].
Thus, it is not terribly surprising that the ER interacts with all of
the other organelles in the cell. Nonetheless, the apposition of
the ER with other organelles ultimately represents an opportu-
nity for these organelle pairs to synergistically carry out novel
functions that would otherwise be impossible. With this frame-
work in mind, let us explore the physiological roles of these
ER–organelle contacts.
THE ER AND MITOCHONDRIA:
AN ANCIENT MARRIAGE
Perhaps not coincidentally, the most well‐characterized orga-
nelle–organelle contact in cell biology is the ER–mitochondria
junction (commonly referred to as mitochondria‐associated
(ER) membranes, or “MAMs”). When considering the inside‐
out origin story of the eukaryotic cell, it is clear that ER–­
mitochondria interactions were preordained long before they
existed in their current form – protomitochondria and the host
cell (i.e. the protonucleus or, in this specific context, proto‐ER)
were interacting long before eukaryotic cells or organelles
existed. Thus, it should be unsurprising these two organelles
are highly coordinated, working together to control key physi-
ological processes such as lipid synthesis, calcium signaling,
and even cell death. But what was not obvious until recently is

Plasma
membrane

ER tubules

ER sheet

Endosomes

Mitochondria

Peroxisomes

Lipid droplets

Golgi

MCS

Nuclear envelope

Figure 13.2  Structure of endoplasmic reticulum (ER) membrane contact sites (MCSs). The ER consists of the nuclear envelope (outlined with a
dashed line) and the peripheral ER, which spreads into the cytosol as a network of sheets and tubules. The peripheral ER forms MCSs with the plasma
membrane, mitochondria, endosomes, peroxisomes, lipid droplets, and the Golgi. Reproduced from [16] with permission of Springer Nature.
154 THE LIVER:  THE ER AND MITOCHONDRIA: AN ANCIENT MARRIAGE
the ER also plays a key role in regulating mitochondrial
­biogenesis and motility, dictating when and where new mito-
chondria are formed, and where they are going [6].
In both yeast and animal cells, MAMs have been character-
ized with high‐resolution microscopy, revealing an average dis-
tance of 10–30 nm, often with protein tethers visible by electron
microscopy appearing at these junctions [7]. Live cell micros-
copy reveals that ER and mitochondria will often move together,
illustrating that these two organelles can remain tethered as they
move about the cell [4]. The majority of MAMs involve extended
regions of membrane apposition, in which the two membranes
(ER and the mitochondrial outer membrane) are parallel to one
another. A smaller fraction of MAMs involve ER tubules wrap-
ping around mitochondria orthogonally to their longer axis (most
mitochondria adopt a tubular morphology). The two most well‐
studied functions of MAMs – lipid exchange and calcium signaling –
are thought to be largely mediated at the parallel types of
contacts, by virtue of the larger amount of contact area available
in this conformation. The orthogonally conjoined MAMs were
only recently discovered to play a key role in mitochondrial
dynamics and will be discussed at the end of this section.
Phospholipid synthesis
The vast majority of lipid synthesis enzymes are localized to the
ER membrane, but several key enzymes reside within the mito-
chondrial outer membrane. Some phospholipid synthesis
requires several key enzymes, some of which are on the ER, and
others on mitochondria. Long before microscopes were capable
of live imaging of MAMs, biochemical analyses revealed the
presence of phosphatidylserine (PS) synthase at MAMs [8]. Two
additional phospholipids – phosphatidylcholine (PC) and phos-
phatidylethanolamine (PE) – are also known to be dependent on
and synthesized at MAMs. PE and PC synthesis illustrates the
tight coordination and exchange of material between the ER and
mitochondria: PS, first synthesized in the ER, is then transferred
to the mitochondrial outer membrane and then from there to the
mitochondrial inner membrane, whereupon it is enzymatically
converted to PE. To generate PC, PE is then translocated from
the mitochondrial inner membrane through the outer membrane
and then back to the ER, whereupon ER‐resident enzymes enzy-
matically convert PE to PC. Fascinatingly, PC is also found on
the mitochondrial outer membrane, indicating that PC is translo-
cated from ER back to mitochondria [8–12]. This ping‐pong pat-
tern of phospholipid translocation and modification likely began
to occur prior to the evolution of the eukaryotic cell and may
have served as a key selective factor for the eventual ectosymbi-
otic relationship between protomitochondria and the protonuclei.
In fact, evidence exists that all eukaryotic phospholipids were
inherited from protomitochondria [13]. Thus, today the steady‐
state levels of key phospholipids in each of our organelles is
clearly demarcated by MAM machinery that was constructed
during the courting period of the eukaryotic cell’s ancestors [2].
The manifold roles of phospholipids in many human diseases are
beyond the scope of this chapter other than to point out that this
is largely a consequence of the foundational importance of the
proper functioning of MAMs, which in turn is a consequence of
their evolutionary history, without which the modern eukaryotic
cell as we know it would not exist.
Calcium signaling
One of the most profound manifestations of the integral role of
mitochondria in the evolution of multicellular life is apoptosis,
wherein mitochondria can control whether a cell will live or die.
This is essential for multicellular life and a classic example is in
development wherein “pruning” removes structures unnecessary
later on in life. That apoptosis is manifested by what was once an
entirely different organism raises deep philosophical questions
about the nature of the modern cell. Who is really in charge? Are
we but slaves to our protomitochondrial ancestors, whose only
function is to gather resources and protect our mitochondrial
overlords? Although the precise mechanisms are beyond the
scope of this chapter, the essential role of apoptosis in health and
disease highlights yet another key role of MAMs  –  regulating
calcium transfer from the ER to mitochondria, which in turn can
trigger apoptosis. Interestingly, calcium influx from the ER to
mitochondria appears to occur in specific subdomains, highlight-
ing the existence of a highly conserved, highly organized molec-
ular interaction that facilitates the process [14, 15].
There are at least three known physiological purposes for
calcium transfer from the ER to mitochondria. The calcium
concentration in the ER is usually somewhere between 100 and
500 μM, whereas cytoplasmic calcium is usually around 100 nM.
Thus, MAMs can provide a source of highly concentrated
­calcium, which may be necessary for the activation of calcium‐
dependent mitochondrial proteins and processes [6, 16]. For
example, elevated calcium levels in the mitochondrial matrix
are required to activate the tricarboxylic acid cycle for the gen-
eration of ATP, and the highly selective mitochondrial calcium
uniporter on the mitochondrial inner membrane is unlikely to
pass high enough concentrations of calcium anywhere other
than near MAMs that facilitate direct transfer of calcium from
the ER [6, 16–18].
As mentioned above, calcium influx from the ER to the mito-
chondria also plays a role in apoptosis. Calcium stimulates the
opening of the mitochondrial permeability transition pore,
which leads to release of cytochrome c, which then leads to the
apoptotic caspase signaling cascade. Blocking the ER calcium
channel Ins(1,4,5)P3R confers apoptotic resistance in several
cell lines [19, 20]. Blocking key ER–mitochondria tethering
complexes, such as the mitochondrial outer membrane protein
FIS1 (also implicated in mitochondrial fission) and the ER pro-
tein BAP31, also confers protection against apoptosis [21].
Finally, the key GTPase protein involved in mitochondrial fis-
sion, dynamin‐related protein 1 (Drp1), is also important in
mitochondrial outer membrane permeabilization (MOMP) that
is required for cytochrome c release and consequent apoptosis
[22]. When calcium is released from the ER to mitochondria,
the apoptotic proteins BAX and BAK facilitate MOMP while
also facilitating stable localization of Drp1 to the outer mito-
chondrial membrane [23–25]. Taken together, these findings
demonstrate a strong link between calcium release, apoptosis,
and mitochondrial fission.
Mitochondrial fission
Interestingly, mitochondrial fission (a precursor to both mito-
chondrial biogenesis and degradation) is also mediated by
 13:  Organelle–Organelle Contacts: Origins and Functions 155
calcium [26]. The precise mechanisms regulating mitochondrial
fission are both beyond the scope of this chapter, and not yet
completely understood. However, recent work has revealed an
entirely new role for ER–mitochondria interactions in regulat-
ing mitochondrial fission. The first key finding was that ER
tubules are almost always orthogonally oriented with respect to
mitochondrial fission sites, and that these intersections repre-
sent mitochondrial constrictions [27]. These constrictions are
thought to be necessary for Drp1 oligomers to circumscribe
mitochondria since Drp1 oligomer rings are ~110 nm in diam-
eter, whereas a non‐constricted mitochondrion has a diameter
closer to 350 nm [27]. Interestingly, the actin cytoskeleton was
soon shown to play a key role in this process [26, 28–31], both
by providing the force necessary for constriction (how else
could a 50 nm ER tubule produce the force necessary to con-
strict the double‐membraned 350 nm mitochondrion?) and by
providing a scaffold for and activation of Drp1 GTPase activity
[32]. These studies also led to the observation that, prior to
mitochondrial fission, Drp1 proteins are not just diffusing
around in the cytoplasm as previously thought, but rather are
localized to the cytoplasmic surface of ER tubules [33]. The
actin polymerization involved in this process is regulated by an
ER‐anchored member of the formin protein family called INF2
(which is implicated in Charcot–Marie–Tooth disease and focal
segmental glomerulosclerosis) and the mitochondrial outer
membrane‐anchored protein Spire1C [26, 28, 30, 34].
The processes described above delineate a (relatively) clear
mechanism for outer membrane fission, but how inner mem-
brane fission is regulated by the ER is less clear in this scenario.
Until you consider calcium. Follow‐up studies quickly showed
that calcium influx from the ER to mitochondria via the mito-
chondrial calcium uniporter (MCU) was able to drive mitochon-
drial inner membrane fission, even in the absence of outer
membrane fission [26]. Interestingly, this calcium exchange
appears to be somewhat dependent on the presence of INF2,
indicating that both inner and outer membrane fission can be
controlled by a complementary set of proteins. In a process that
remains less well understood, it appears that mitochondrial
DNA (mtDNA) nucleoids also appear to divide at ER–mito-
chondria fission sites, providing an ER‐regulated mechanism
for coupling mtDNA and mitochondrial biogenesis, or, in the
case of mitophagy, the degradation of dysfunctional mtDNA
and mitochondria components while preserving functional
mitochondria [35]. In conclusion, mitochondrial fission is an
intricately coordinated process that involves multiple compo-
nents with complementary, yet completely independent func-
tions in each step, all dependent on interactions between the ER,
mitochondrial inner and outer membranes and mtDNA, and the
actin cytoskeleton.
THE ER AND ENDOSOMES: HINTS OF
A UNIVERSAL MECHANISM
Given that the ER interacts with every organelle in the cell, it
should not be surprising that the ER interacts with endosomes in
multiple contexts. Endosomes are membrane vesicles involved
in the trafficking of cargos to/from the plasma membrane and
extracellular space to/from the Golgi and the lysosome. Contacts
between the ER and endosomes play multiple roles in regulating
endosomal function, depending on the context and subcellular
location. The ER contains phosphatases that can dephosphoryl-
ate important cargo such as epidermal growth factor receptor
(EGFR) [36]. ER–endosome contacts can regulate endosomal
motility in a cholesterol‐dependent fashion [37], and cholesterol
may be trafficked from endosomes to the ER at contact sites
[38]. Given the importance of these cellular processes and in
particular cholesterol, it is perhaps not surprising that multiple
diseases are caused by mutations in ER–endosome contact pro-
teins [39–42]. Along the same lines, targeting ER–endosome
contacts presents an attractive target for potential therapies for
metabolic disorders.
More recently, it was shown that, similarly to mitochondria,
endosomal fission is also mediated by ER tubules wrapping
around and constricting endosomes at their fission sites [43].
The multi‐destination character of endosomal trafficking neces-
sitates the existence of spatiotemporal regulation of endosomal
fission such that cargo with different destinies are segregated at
the appropriate location and time. The specific destination and
direction of cargo traffic is of course regulated by many molecu-
lar components, but the precise spatial location of endosomal
fission was recently shown to be regulated by no less than the
ER, in a fashion quite similar to that observed for mitochondria:
ER tubules wrap around endosomes at fission sites, at which
point actin regulatory proteins on the endosomal surface (which
includes endosomal isoforms of Spire family proteins) and
dynamin proteins (related to Drp1) cooperate to drive endoso-
mal fission [43]. It has not yet been shown whether ER‐anchored
INF2 similarly plays a role in Spire‐mediated endosomal fis-
sion, but disruption of either of these proteins does appear to
result in decreased endosomal fission (unpublished data), sug-
gesting a conserved molecular mechanism for ER‐mediated
­fission of both mitochondria and endosomes. Complicating the
story are studies that indicate that endosome‐associated Rab
GTPase proteins also regulate mitochondrial fission [44, 45],
directly coupling the endolysosomal machinery to the mito-
chondrial fission machinery. At this point, it is perhaps reason-
able to wonder whether puppet‐master ER brings all the
necessary components together to regulate fission of all the
organelles.
THE ER AND GOLGI: CLOSE BUT
MYSTERIOUS
The Golgi complex is the central hub of the cellular secretory
pathway. Immediately adjacent to the ER, the Golgi is known
to receive proteins and lipid cargos from the ER, process those
cargos further with post‐translational modifications, and then
sort them on towards their final destination. The Golgi is com-
pletely dependent on input from the ER; blocking traffic from
the ER to the Golgi eventually results in the total disintegra-
tion of the Golgi. Trafficking is bidirectional and cargos can
be transported from the Golgi to the ER as well [46–51].
These well‐known, well‐characterized functional interactions
between the Golgi and ER are one of the first things we are
156 THE LIVER:  ER, MITOCHONDRIA, AND PEROXISOMES: SHARED CUSTODY?
taught when introduced to cell biology. So you might expect
that the physical interactions between Golgi and ER mem-
branes are among the most well characterized. But you would
be wrong.
What is known about direct ER–Golgi membrane interac-
tions is that they facilitate lipid exchange. One well‐studied
example of this is the transfer of ceramide from the ER to the
Golgi, as a first step towards glycosphingolipid and sphingomy-
elin synthesis in the Golgi [52]. Although the proteins involved
in ceramide transfer from Golgi to ER are known, whether they
are the key proteins involved in maintaining ER–Golgi contacts
remains unclear. Nor do we know how ER–Golgi contacts are
regulated. Interestingly, regulation of the cytoskeleton by INF2
and the ER seem to be important for regulating normal Golgi
structure [53], once again tempting one to think there are con-
served mechanisms involved in ER‐ and actin‐mediated regula-
tion of organelle structure and dynamics.
THE ER AND AUTOPHAGOSOMES:
ANOTHER UNIVERSAL MECHANISM
Autophagy is the main mechanism by which the cell degrades
intracellular components. The autophagosome is assembled at
least in part by the ER, often triggered by stress. Assembly of
the autophagosome depends on the coordinated recruitment of
membrane proteins to the ER at the autophagosome assembly
site. Interestingly, ER–mitochondria contact sites were recently
shown to be a preferred location for the recruitment of these
proteins – mitochondria even appear to contribute membrane to
autophagosomes [54–58], and this process is likely to play a
role in maintaining homeostasis in hepatocytes [59–61]. It is
perhaps only fitting that the same organelle involved in dictat-
ing the timing of the cell’s death (i.e. apoptosis) also controls
the cell’s self‐cannibalism. That said, while autophagy is often
associated with cell death, it is also known to promote survival,
and thus so‐called “autophagic cell death” may in fact be
reversing causality and blaming ambulances for accidents. ER–
mitochondria contacts are not necessary for autophagosome
assembly: ER–plasma membrane contacts have also been
shown to be the site of autophagosome biogenesis [62, 63].
Interestingly, VAPs, one of the key protein families mediating
ER–autophagosome contacts and autophagosome biogenesis,
also mediate contacts with multiple other organelles, including
the plasma membrane, mitochondria, lysosomes, endosomes,
and Golgi [64, 65].
Although the physical interaction of ER and autophagosomes is
clear – they literally serve as hubs for autophagosome biogenesis –
the functional purpose (or origin) of this intimate relationship
is less so. The answer maybe lies in the fact that the ER is
­perhaps the central player in the cell stress response. This fram-
ing is supported by the finding that disrupting autophagy leads
to increased ER stress and subsequent cell death. It is thus
thought that ER stress‐induced autophagy may be a last‐ditch
effort by the cell to survive before succumbing to stress‐induced
apoptosis [66].
ER AND LIPID DROPLETS: OLD
“BUD” DIES?
Lipid droplets (LDs) were thought to be passive lipid‐storage
facilities until the discovery of multiple proteins and active
functions for LDs. Although LDs have been shown to make con-
tacts with the ER in electron microscopy, much like ER–
mitochondria contacts, it is thought that ER–LD contacts are
unique for a number of reasons, mainly that LDs are known to
originate from the ER. It is thought that when lipid esters in the
ER–lipid bilayer reach a critical concentration, they are no
longer stable, and cause the lipid membrane to bulge and even-
tually bud off into a nascent LD [67]. Thus, ER–LD contacts
have been speculated to contain continuous membranes at some
point in the LD life cycle but have not yet been directly observed.
New high‐resolution cellular cryoelectron microscopy techniques
may finally be able to conclusively answer this question.
The importance of ER–LD contacts is highlighted by the fact
that mutations in one of the key ER–LD tethering proteins,
seipen, cause Berardinelli–Seip congenital lipodystrophy. This
disease causes a lack of adipose tissue and fat deposition in the
liver [68]. Cells with mutations in seipin have an abnormally
high number of cells with either small LDs or greatly oversized
LDs [69]. Fascinatingly, sometimes LDs are observed inside the
nucleus, contacting the inner nuclear reticular membrane [70],
highlighting once more the historical evolutionary framework
viewing the ER as an extension of the nucleus itself.
Taking us even deeper into our evolutionary framework, we
must also consider recent work demonstrating the importance of
contacts between LDs and mitochondria (LD–mitos). Although
LD–mito contacts have been observed in brown adipose tissue
(BAT) [71], heart [72], and type I skeletal muscle [73], their
function remained unclear. Two recent studies shed new light on
how LD–mitos cooperate to regulate metabolism and energetic
flux in the cell. First, Rambold et al. in 2015 showed that fatty
acids are trafficked from LDs to mitos during starvation, ena-
bling the cell to switch from glycolysis to β‐oxidation for ATP
energy production [74]. More recently, an elegant study from
the Shirihai lab showed that mitochondria directly contacting
LDs have distinct features that facilitate LD biogenesis and tria-
cylglyceride synthesis [75]. Although this leaves some ambigu-
ity about whether LD–mito contacts facilitate catabolism or
biogenesis of LDs (it almost certainly depends heavily on vari-
ations in cell and tissue type and context), the role of LD–mito
contacts in metabolism is unimpeachable.
ER, MITOCHONDRIA, AND
PEROXISOMES: SHARED CUSTODY?
Although specific tethers between peroxisomes and the ER or
mitochondria are not clearly defined, a discussion on organelle–
organelle interactions would be incomplete without discussing
peroxisome biogenesis, which has been proposed to originate
both from the ER and from mitochondria [76, 77]. That the ER
contacts all organelles at high spatiotemporal frequency makes
 13:  Organelle–Organelle Contacts: Origins and Functions 157

it difficult to discern specific functions of ER–organelle con-
tacts. This difficulty is further confounded by the fact that all
membrane‐bound organelles will contain proteins that at one
point or another were localized on the ER. So‐called pre‐peroxi-
somal vesicles are thought to shuttle proteins from the ER to
peroxisomes, presumably facilitating peroxisomal biogenesis
without any direct contacts or tethering [76, 77]. However, ER–
peroxisome tethers have indeed been reported (mediated by our
aforementioned friends in the VAP family), and these tethers are
now known to be important for phospholipid synthesis, regula-
tion of cholesterol levels, and peroxisome growth [78].
At the same time, new evidence is emerging that peroxisomes
are derived from both the ER and mitochondria. The acceptance
of this model, almost exclusively pioneered by the McBride
laboratory [76, 77], was severely hampered by yeast studies that
showed no involvement of mitochondria. Intriguingly, this both
highlights a key evolutionary step in higher eukaryotes and also
explains the physiological role of peroxisomes. Specifically,
peroxisomes have been posited to be indispensable for regulat-
ing mitochondrial reactive oxygen species (ROS) generated by
fatty acid metabolism. Notably, yeast mitochondria long ago
discarded their ability to β‐oxidize fatty acids – hence the dis-
crepancy. In mammalian cells, however, all evidence currently
points to a model wherein peroxisomes evolved from a need to
keep mitochondrial ROS generation in check. Thus, the physical
biogenesis of peroxisomes from both ER and mitochondrial
membranes provides a strong testament to the entire theory of
symbiosis, in particular our favored “inside‐out” model. For an
interesting discussion of the symbiotic theory of peroxisome
evolution, see the 2017 essay by D. Speijer [79].
SOME FINAL REMARKS
Given the number of organelle types, the number of permuta-
tions of transorganelle contacts can get high, especially when
considering tri‐ or even quad‐organelle contacts (e.g. LDs, ER
and mitochondria, or endosomes, ER, autophagosomes, and
mitochondria). There is new evidence for co‐transport of orga-
nelles by “piggybacking” on one another [80, 81], and it is
highly likely there will be new instances of such noncanonical
transport revealed in the near future. So this chapter is by no
means comprehensive. The ER–plasma membrane contacts
were barely mentioned, and I intentionally focused on mamma-
lian‐specific organelles, so no mention of chloroplasts either.

Table 13.1  List of known organelle membrane contact proteins in metazoans

Contact site Protein name Description Reference
ER–PM VAPs ER receptors for numerous proteins containing FFAT motif [65]
STIM1, Orai Dynamic tether: PM Ca2+ channel Orai binds to ER integral STIM1 at low [83]
luminal Ca2+
E‐Syt1/2/3 SMP domain‐containing tethers; E‐Syt1 is a Ca2+‐dependent dynamic tether [84]
Junctophylin1/2/3/4 ER residents, bind PM via MORN domains [85]
DHPR, RyR PM and ER Ca2+ channels, interact and function in a concerted way [86]
ORP5, ORP8 LTPs, dynamic tethers, contain TMD and PH domain [87]
ER–mitochondria MFN1/2 Mitochondrial fusion GTPase; an ER MFN2 pool mediates tethering by [88]
interacting with mitochondrial MFN1/2
IP3R, VDAC, Grp75 ER Ca2+ release channel IP3R and mitochondrial metabolite channel VDAC [89]
are connected. Grp75 may be involved
Fis1, BAP31 Mitochondrial Fis1 and ER BAP31 interact for transmission of apoptotic [21]
signals
PTPIP51, VAP PTPIP51 is a mitochondrial LTP structurally equipped for tethering via [90]
VAP binding
ER–endosome VAPs ER receptors for numerous proteins containing FFAT motif [65]
StARD3, StARD3NL Integral endosomal proteins, interact with ER proteins VAP via FFAT motif [91]
ORP1L, ORP5 LTPs active at the ER–endosome interface [37, 38]
PTP1B, EGFR, Annexin A1 Components mediating interplay between ER and multivesicular bodies [36, 92]
Protrudin, Rab7 ER protrudin interacts with Rab7 and phosphatidylinositol‐3‐phosphate on [93]
late endosomes
ER–Golgi VAPs ER receptors for numerous proteins containing FFAT motif [65]
OSBP LTP, dynamic phosphatidylinositol‐4‐phosphate‐dependent tether, contains [94]
FFAT motif and PH domain
CERT LTP, putative dynamic tether, contains FFAT motif and PH domain [95]
FAPP2 LTP, structurally equipped for tethering (FFAT motif, PH domain) [52]
Nir2 Phosphatidylinositol transfer protein, contains FFAT motif [95]
Lysosome–peroxisome Synaptotagmin‐7 Mediates lysosome–peroxisome tethering important for cholesterol transfer [96]
ER–lipid droplet DGAT2, FATP1 Lipid droplet‐resident DGAT2 and ER‐resident FATP1 interact and [97]
coordinate lipid droplet biogenesis
Mitochondria–lipid Perilipin‐5 Lipid droplet scaffold protein involved in interaction with mitochondria [98]
droplet
Mitochondrial Mic60/27/26/25/19/10, Mitochondrial contact site and cristae organizing system (MICOS), an [99]
IM–OM Qil1 integral protein complex of the inner membrane
SAMM50, Metaxin 1/2 MICOS inteaction partners in the outer mitochondrial membrane
Reproduced from [64] with permission of Elsevier.
FFAT, phenylalanine in an acidic tract; SMP, synaptotagmin‐like mitochondrial lipid‐binding protein; MORN, membrane occupation and recognition nexus; LTP, lipid transfer
protein; TMD, transmembrane domain; PH, pleckstrin homology; IM, mitochondrial inner membrane; OM, mitochondrial outer membrane.
158 THE LIVER:  REFERENCES
Needless to say, the transorganelle contact field is large and still
in its infancy. Indeed, there is at least one entire book dedicated
to organelle contact sites [82]. In lieu of discussing all the
known proteins involved in transorganelle contacts in this one
chapter, readers can refer to Table 13.1 (taken from [64]) and the
references therein for additional details [21, 36–38, 52, 65,
83–99]. Regularly searching for “organelle contacts,” “transor-
ganelle,” and “membrane contact sites” on PubMed or Google
Scholar will likely reveal new publications every week. Novel
proteomics methods with increasing spatiotemporal resolution
promise to allow for mapping of novel proteins involved in
organelle–organelle contacts in different conditions (e.g. cell
stress, in the presence of disease‐associated mutations, etc.)
[100], and subtractive analysis thereof will provide detailed
insight towards how these complexes regulate and are regulated
by physiological cues. The bad news is this means we have
much more work to do. The good news is this means novel drug
targets and therapies are likely to multiply in the coming years.
With some luck and much effort, this golden age of organelle
biology will lead to a corresponding revolution in medicine.
REFERENCES
1. Sagan, L. On the origin of mitosing cells. J Theor Biol, 1967;14(3):255–74.
2. Baum, D.A. and Baum, B. An inside‐out origin for the eukaryotic cell. BMC
Biol, 2014;12:76.
3. Dey, G., Thattai, M., and Baum, B. On the archaeal origins of eukaryotes and
the challenges of inferring phenotype from genotype. Trends Cell Biol,
2016;26(7):476–85.
4. Valm, A.M. et al. Applying systems‐level spectral imaging and analysis to
reveal the organelle interactome. Nature, 2017;546(7656):162–7.
5. Cohen, S., Valm, A.M., and Lippincott‐Schwartz, J. Interacting organelles.
Curr Opin Cell Biol, 2018;53:84–91.
6. Rowland, A.A. and Voeltz, G.K. Endoplasmic reticulum‐mitochondria
contacts: function of the junction. Nat Rev Mol Cell Biol,
­ 36. Eden, E.R. et al. Membrane contacts between endosomes and ER provide
2012;13(10):607–25.
7. Csordas, G. et al. Structural and functional features and significance of the
physical linkage between ER and mitochondria. J Cell Biol, 2006;
174(7):915–21.
8. Vance, J.E. Phospholipid synthesis in a membrane fraction associated with
mitochondria. J Biol Chem, 1990;265(13):7248–56.
9. Voelker, D.R. Interorganelle transport of aminoglycerophospholipids.
Biochim Biophys Acta, 2000;1486(1):97–107.
10. van Meer, G., Voelker, D.R., and Feigenson, G.W. Membrane lipids: where
they are and how they behave. Nat Rev Mol Cell Biol, 2008;9(2):112–24.
11. Osman, C., Voelker, D.R., and Langer, T. Making heads or tails of phospho-
lipids in mitochondria. J Cell Biol, 2011;192(1):7–16.
12. Stone, S.J. and Vance, J.E. Phosphatidylserine synthase‐1 and ‐2 are localized
to mitochondria‐associated membranes. J Biol Chem, 2000;275(44):
34534–40.
13. Lykidis, A. Comparative genomics and evolution of eukaryotic phospholipid
biosynthesis. Prog Lipid Res, 2007;46(3–4):171–99.
14. Rizzuto, R. et al. Close contacts with the endoplasmic reticulum as determi-
nants of mitochondrial Ca2+ responses. Science, 1998;280(5370):1763–6.
15. Rizzuto, R. et  al. Microdomains with high Ca2+ close to IP3‐sensitive
channels that are sensed by neighboring mitochondria. Science, 1993.
­ regulate mitochondrial fission via RAB7 GTP hydrolysis. Nature,
262(5134):744–7.
16. Phillips, M.J. and Voeltz, G.K. Structure and function of ER membrane con-
tact sites with other organelles. Nat Rev Mol Cell Biol, 2016;17(2):69–82.
17. Rizzuto, R. et al. Mitochondria as sensors and regulators of calcium signal-
ling. Nat Rev Mol Cell Biol, 2012;13(9):566–78.
18. Baughman, J.M. et al. Integrative genomics identifies MCU as an essential
component of the mitochondrial calcium uniporter. Nature, 2011;
476(7360):341–5.
19. Jayaraman, T. and Marks, A.R. T cells deficient in inositol 1,4,5‐trisphosphate
receptor are resistant to apoptosis. Mol Cell Biol, 1997;17(6):3005–12.
20. Khan, A.A. et al. Lymphocyte apoptosis: mediation by increased type 3 ino-
sitol 1,4,5‐trisphosphate receptor. Science, 1996;273(5274):503–7.
21. Iwasawa, R. et al. Fis1 and Bap31 bridge the mitochondria‐ER interface to
establish a platform for apoptosis induction. EMBO J, 2011;30(3):556–68.
22. Chan, D.C. Fusion and fission: interlinked processes critical for mitochon-
drial health. Annu Rev Genet, 2012;46:265–87.
23. Lucken‐Ardjomande, S. and Martinou, J.C. Regulation of Bcl‐2 proteins and
of the permeability of the outer mitochondrial membrane. C R Biol,
2005;328(7):616–31.
24. Montessuit, S. et al. Membrane remodeling induced by the dynamin‐related
protein Drp1 stimulates Bax oligomerization. Cell, 2010;142(6):889–901.
25. Wasiak, S., Zunino, R., and McBride, H.M. Bax/Bak promote sumoylation
of DRP1 and its stable association with mitochondria during apoptotic cell
death. J Cell Biol, 2007;177(3):439–50.
26. Chakrabarti, R. et al. INF2‐mediated actin polymerization at the ER stimu-
lates mitochondrial calcium uptake, inner membrane constriction, and divi-
sion. J Cell Biol, 2018;217(1):251–68.
27. Friedman, J.R. et  al. ER tubules mark sites of mitochondrial division.
Science, 2011;334(6054):358–62.
28. Korobova, F., Ramabhadran, V., and Higgs, H.N. An actin‐dependent step in
mitochondrial fission mediated by the ER‐associated formin INF2. Science,
2013;339(6118):464–7.
29. Moore, A.S. et al. Dynamic actin cycling through mitochondrial subpopula-
tions locally regulates the fission‐fusion balance within mitochondrial net-
works. Nat Commun, 2016;7:12886.
30. Manor, U. et  al. A mitochondria‐anchored isoform of the actin‐nucleating
spire protein regulates mitochondrial division. Elife, 2015;4.
31. Korobova, F., Gauvin, T.J., and Higgs, H.N. A role for myosin II in mam-
malian mitochondrial fission. Curr Biol, 2014;24(4):409–14.
32. Ji, W.K. et  al. Actin filaments target the oligomeric maturation of the
dynamin GTPase Drp1 to mitochondrial fission sites. Elife, 2015;4:e11553.
33. Ji, W.K. et  al. Receptor‐mediated Drp1 oligomerization on endoplasmic
reticulum. J Cell Biol, 2017;216(12):4123–39.
34. Curchoe, C.L. and Manor, U. Actin cytoskeleton‐mediated constriction of
membrane organelles via endoplasmic reticulum scaffolding. ACS Biomater
Sci Eng, 2017;3(11):2727–32.
35. Lewis, S.C., Uchiyama, L.F., and Nunnari, J. ER‐mitochondria contacts cou-
ple mtDNA synthesis with mitochondrial division in human cells. Science,
2016;353(6296):aaf5549.
sites for PTP1B‐epidermal growth factor receptor interaction. Nat Cell Biol,
2010;12(3):267–72.
37. Rocha, N. et al. Cholesterol sensor ORP1L contacts the ER protein VAP to
control Rab7‐RILP‐p150 Glued and late endosome positioning. J Cell Biol,
2009;185(7):1209–25.
38. Du, X. et al. A role for oxysterol‐binding protein‐related protein 5 in endo-
somal cholesterol trafficking. J Cell Biol, 2011;192(1):121–35.
39. Ueno, S. et al. The gene encoding a newly discovered protein, chorein, is
mutated in chorea‐acanthocytosis. Nat Genet, 2001;28(2):121–2.
40. Belvedere, R. et al. Annexin A1 contributes to pancreatic cancer cell pheno-
type, behaviour and metastatic potential independently of formyl peptide
receptor pathway. Sci Rep, 2016;6:29660.
41. Kondo, I. et al. COH1 analysis and linkage study in two Japanese families
with Cohen syndrome. Clin Genet, 2005;67(3):270–2.
42. Hashimoto, Y. et al. Protrudin regulates endoplasmic reticulum morphology
and function associated with the pathogenesis of hereditary spastic paraple-
gia. J Biol Chem, 2014;289(19):12946–61.
43. Rowland, A.A. et al. ER contact sites define the position and timing of endo-
some fission. Cell, 2014;159(5):1027–41.
44. Wong, Y.C., Ysselstein, D., and Krainc, D. Mitochondria‐lysosome contacts
2018;554(7692):382–6.
45. Landry, M.C. et al. A functional interplay between the small GTPase Rab11a
and mitochondria‐shaping proteins regulates mitochondrial positioning and
polarization of the actin cytoskeleton downstream of Src family kinases. J
Biol Chem, 2014;289(4):2230–49.
46. Lippincott‐Schwartz, J. et al. Rapid redistribution of Golgi proteins into the
ER in cells treated with brefeldin A: evidence for membrane cycling from
Golgi to ER. Cell, 1989;56(5):801–13.
 13:  Organelle–Organelle Contacts: Origins and Functions 159
47. Lippincott‐Schwartz, J. et al. Microtubule‐dependent retrograde transport of
proteins into the ER in the presence of brefeldin A suggests an ER recycling
pathway. Cell, 1990;60(5):821–36.
48. Lippincott‐Schwartz, J. Bidirectional membrane traffic between the endo-
plasmic reticulum and Golgi apparatus. Trends Cell Biol, 1993;3(3):81–8.
49. Presley, J.F. et al. ER‐to‐Golgi transport visualized in living cells. Nature,
1997;389(6646):81–5.
50. Ward, T.H. et al. Maintenance of Golgi structure and function depends on the
integrity of ER export. J Cell Biol, 2001;155(4):557–70.
51. Sengupta, P. et al. ER trapping reveals Golgi enzymes continually revisit the
ER through a recycling pathway that controls Golgi organization. Proc Natl
Acad Sci U S A, 2015;112(49):E6752–61.
52. D’Angelo, G. et al. Glycosphingolipid synthesis requires FAPP2 transfer of
glucosylceramide. Nature, 2007;449(7158):62–7.
53. Ramabhadran, V. et  al. Splice variant‐specific cellular function of the
formin  INF2 in maintenance of Golgi architecture. Mol Biol Cell,
2011;22(24):4822–33.
54. Hamasaki, M. et al. Autophagosomes form at ER‐mitochondria contact sites.
Nature, 2013;495(7441):389–93.
55. Gomez‐Suaga, P. et  al. The ER‐mitochondria tethering complex VAPB‐
PTPIP51 regulates autophagy. Curr Biol, 2017;27(3):371–85.
56. Rambold, A.S. and Lippincott‐Schwartz, J. Mechanisms of mitochondria
and autophagy crosstalk. Cell Cycle, 2011;10(23):4032–8.
57. Hailey, D.W. et  al. Mitochondria supply membranes for autophagosome
­biogenesis during starvation. Cell, 2010;141(4):656–67.
58. Rambold, A.S. and Lippincott‐Schwartz, J. Starved cells use mitochondria
for autophagosome biogenesis. Cell Cycle, 2010;9(18):3633–4.
59. Kang, S.W. et al. AMPK activation prevents and reverses drug‐induced mito-
chondrial and hepatocyte injury by promoting mitochondrial fusion and
function. PLoS One, 2016;11(10):e0165638.
60. Fu, D. et al. Coordinated elevation of mitochondrial oxidative phosphoryla-
tion and autophagy help drive hepatocyte polarization. Proc Natl Acad Sci U
S A, 2013;110(18):7288–93.
61. Fu, D., Lippincott‐Schwartz, J., and Arias, I.M. Increased mitochondrial
fusion and autophagy help isolated hepatocytes repolarize in collagen sand-
wich cultures. Autophagy, 2013;9(12):2154–5.
62. Zhao, Y.G. et  al. The ER contact proteins VAPA/B interact with multiple
autophagy proteins to modulate autophagosome biogenesis. Curr Biol,
2018;28(8):1234–45 e4.
63. Zhao, Y.G. and Zhang, H. The ER‐localized autophagy protein EPG‐3/
VMP1 regulates ER contacts with other organelles by modulating ATP2A/
SERCA activity. Autophagy, 2018;14(2):362–3.
64. Eisenberg‐Bord, M. et al. A tether is a tether is a tether: tethering at mem-
brane contact sites. Dev Cell, 2016;39(4):395–409.
65. Murphy, S.E. and Levine, T.P. VAP, a versatile access point for the endoplas-
mic reticulum: review and analysis of FFAT‐like motifs in the VAPome.
Biochim Biophys Acta, 2016;1861(8 Pt B):952–61.
66. Molino, D. et al. ER‐driven membrane contact sites: evolutionary conserved
machineries for stress response and autophagy regulation? Commun Integr
Biol, 2017;10(5–6):e1401699.
67. Khandelia, H. et al. Triglyceride blisters in lipid bilayers: implications for
lipid droplet biogenesis and the mobile lipid signal in cancer cell mem-
branes. PLoS One, 2010;5(9):e12811.
68. Magre, J. et  al. Identification of the gene altered in Berardinelli‐Seip
congenital lipodystrophy on chromosome 11q13. Nat Genet, 2001;
­ lum and the Golgi complex requires the VAP proteins and is essential for
28(4):365–70.
69. Szymanski, K.M. et al. The lipodystrophy protein seipin is found at endo-
plasmic reticulum lipid droplet junctions and is important for droplet mor-
phology. Proc Natl Acad Sci U S A, 2007;104(52):20890–5.
70. Ohsaki, Y. et al. PML isoform II plays a critical role in nuclear lipid droplet
formation. J Cell Biol, 2016;212(1):29–38.
71. Boutant, M. et  al. Mfn2 is critical for brown adipose tissue thermogenic
function. EMBO J, 2017;36(11):1543–58.
72. Wang, H. et al. Analysis of lipid droplets in cardiac muscle. Methods Cell
Biol, 2013;116:129–49.
73. Tarnopolsky, M.A. et al. Influence of endurance exercise training and sex on
intramyocellular lipid and mitochondrial ultrastructure, substrate use, and
mitochondrial enzyme activity. Am J Physiol Regul Integr Comp Physiol,
2007;292(3):R1271–8.
74. Rambold, A.S., Cohen, S., and Lippincott‐Schwartz, J. Fatty acid traffick-
ing in starved cells: regulation by lipid droplet lipolysis, autophagy, and
mitochondrial fusion dynamics. Dev Cell, 2015;32(6):678–92.
75. Benador, I.Y. et al. Mitochondria bound to lipid droplets have unique bio-
energetics, composition, and dynamics that support lipid droplet expansion.
Cell Metab, 2018;27(4):869–85 e6.
76. Mohanty, A. and McBride, H.M. Emerging roles of mitochondria in the
evolution, biogenesis, and function of peroxisomes. Front Physiol,
2013;4:268.
77. Sugiura, A. et al. Newly born peroxisomes are a hybrid of mitochondrial
and ER‐derived pre‐peroxisomes. Nature, 2017;542(7640):251–4.
78. Hua, R. et al. VAPs and ACBD5 tether peroxisomes to the ER for peroxi-
some maintenance and lipid homeostasis. J Cell Biol, 2017;216(2):367–77.
79. Speijer, D. Evolution of peroxisomes illustrates symbiogenesis. Bioessays,
2017;39(9).
80. Salogiannis, J. and Reck‐Peterson, S.L. Hitchhiking: a non‐canonical mode
of microtubule‐based transport. Trends Cell Biol, 2017;27(2):141–50.
81. Salogiannis, J., Egan, M.J., and Reck‐Peterson, S.L. Peroxisomes move by
hitchhiking on early endosomes using the novel linker protein PxdA. J Cell
Biol, 2016;212(3):289–96.
82. Tagaya, M. Organelle Contact Sites: From Molecular Mechanism to
Disease. Springer, New York, 2017.
83. Liou, J. et  al. Live‐cell imaging reveals sequential oligomerization and
local plasma membrane targeting of stromal interaction molecule 1 after
Ca2+ store depletion. Proc Natl Acad Sci U S A, 2007;104(22):9301–6.
84. Giordano, F. et al. PI(4,5)P(2)‐dependent and Ca2+‐regulated ER‐PM interac-
tions mediated by the extended synaptotagmins. Cell, 2013;153(7):1494–509.
85. Takeshima, H., Hoshijima, M., and Song, L.S. Ca2+ microdomains organ-
ized by junctophilins. Cell Calcium, 2015;58(4):349–56.
86. Rebbeck, R.T. et  al. The beta(1a) subunit of the skeletal DHPR binds to
skeletal RyR1 and activates the channel via its 35‐residue C‐terminal tail.
Biophys J, 2011;100(4):922–30.
87. Chung, J. et  al. Intracellular transport. PI4P/phosphatidylserine counter-
transport at ORP5‐ and ORP8‐mediated ER‐plasma membrane contacts.
Science, 2015;349(6246):428–32.
88. de Brito, O.M. and Scorrano, L. Mitofusin 2 tethers endoplasmic reticulum
to mitochondria. Nature, 2008;456(7222):605–10.
89. Szabadkai, G. et al. Chaperone‐mediated coupling of endoplasmic reticu-
lum and mitochondrial Ca2+ channels. J Cell Biol, 2006;175(6):901–11.
90. De Vos, K.J. et al. VAPB interacts with the mitochondrial protein PTPIP51
to regulate calcium homeostasis. Hum Mol Genet, 2012;21(6):1299–311.
91. Alpy, F. et al. STARD3 or STARD3NL and VAP form a novel molecular
tether between late endosomes and the ER. J Cell Sci, 2013;126(Pt
23):5500–12.
92. Eden, E.R. et al. Annexin A1 tethers membrane contact sites that mediate
ER to endosome cholesterol transport. Dev Cell, 2016;37(5):473–83.
93. Raiborg, C. et  al. Repeated ER‐endosome contacts promote endosome
translocation and neurite outgrowth. Nature, 2015;520(7546):234–8.
94. Mesmin, B. et  al. A four‐step cycle driven by PI(4)P hydrolysis
directs  sterol/PI(4)P exchange by the ER‐Golgi tether OSBP. Cell,
2013;155(4):830–43.
95. Peretti, D. et al. Coordinated lipid transfer between the endoplasmic reticu-
Golgi‐mediated transport. Mol Biol Cell, 2008;19(9):3871–84.
96. Chu, B.B. et al. Cholesterol transport through lysosome‐peroxisome mem-
brane contacts. Cell, 2015;161(2):291–306.
97. Xu, N. et al. The FATP1‐DGAT2 complex facilitates lipid droplet expan-
sion at the ER‐lipid droplet interface. J Cell Biol, 2012;198(5):895–911.
98. Wang, H. et  al. Perilipin 5, a lipid droplet‐associated protein, provides
physical and metabolic linkage to mitochondria. J Lipid Res,
2011;52(12):2159–68.
99. Zerbes, R.M. et al. Mitofilin complexes: conserved organizers of mitochon-
drial membrane architecture. Biol Chem, 2012;393(11).
100. Hung, V. et al. Spatially resolved proteomic mapping in living cells with the
engineered peroxidase APEX2. Nat Protoc, 2016;11(3):456–75.
 Gap and Tight Junctions
14 in Liver: Structure, Function,
and Pathology
John W. Murray1,2 and David C. Spray1,3,4
1
Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, New York, NY, USA
2
Department of Anatomy and Structural Biology, Albert Einstein College of Medicine, Bronx,
New York, NY, USA
3
Department of Neuroscience, Albert Einstein College of Medicine, Bronx, New York, NY, USA
4
Department of Medicine, Albert Einstein College of Medicine, Bronx, New York, NY, USA

STRUCTURE OF JUNCTIONS
IN THE LIVER
This chapter is an update that integrates new findings on both
tight junctions and gap juntions that have emerged since the pre­
vious edition of this volume [1]. For more detailed historical
background on gap junctions in particular, the reader is referred
to an even earlier edition [2].
Thin‐section and freeze‐fracture electron microscopy applied
to liver has revealed elaborate junctional complexes near the
apical cellular domains at appositional areas between hepato­
cytes (Figure 14.1). Tight junctions surround the bile canaliculi,
where they seal the paracellular spaces between hepatocytes,
regulating movement of solutes, ions, and water through the
extracellular space, and act as a fence to block lateral diffusion
of membrane‐embedded molecules between basolateral and
apical membrane, thereby establishing and maintaining cellular
polarity. In thin section, tight junctions appear as small, very
close sites of membrane contact or “kisses” between the cell
surfaces. In replicas of freeze‐fractured membranes, tight junc­
tions appear as continuous, branching, or anastomosing strands
of 10 nm particles in the P‐fracture (protoplasm, i.e. cytosol)
face (Figure  14.1a), with complementary grooves in the
E‐(extracellular) face [4]. The resistance of the tight junctional
barrier is proportional to the number of tight junction strands,
which in liver can be quite high.
Gap junctions are found in the basolateral membrane
nearby the tight junctions. In contrast to the linear strands of
intramembrane particles that form tight junctions, gap junc­
tion particles are clustered together to form discoid areas or
gap junction plaques (Figure  14.1). Gap junctions occupy a
large fraction, as much as 3%, of the total surface area of
hepatocytes [5]. In stained thin sections of liver tissue (and of
hepatocyte cell pairs), gap junctions are recognized as septi­
laminar linear membrane appositions separated by an extra­
cellular gap (Figure 14.1b,d). The seven lamina consist of the
transparent extracellular gap sandwiched on either side by
each cell’s three‐lamina membranes, consisting of two stained
and thus electron‐opaque lipid head groups on each side of the
electron‐lucent membrane interior. The overall thickness of
this double membrane specialization is approximately 15–18
nm, and the extracellular space in the region of gap junctional
contact is approximately 2–4 nm wide.
In freeze‐fracture images, gap junctions of liver and hepato­
cyte cell pairs are recognizable as arrays or plaques of intram­
embranous particles approximately 9 nm across present in the
P‐fracture face with complementary pits on the E fracture face.
These plaques are generally round or oval and can be quite large
in hepatocytes (Figure 14.1c,d), commonly exceeding 1 μm in
diameter and containing more than 10 000 particles. The parti­
cles seen in freeze‐fracture replicas and the bridges across the
extracellular space seen in thin section are believed to represent
channels with hydrophilic walls extending from the cytoplasmic
aspect of one cell to that of another (Figure 14.1).
High‐resolution ultrastructural studies on isolated liver gap
junctions using techniques of X‐ray diffraction and low‐angle

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
 14:  Gap and Tight Junctions in Liver: Structure, Function, and Pathology 161

(a) (b) (e)

(c) (d)

(f) (g)

Figure 14.1  Fine structure of appositional membranes between hepatocytes, showing tight junctions and gap junctions. (a, c, f, g) Freeze‐fracture;
(b, d) thin sections; (e) immunofluorescence. (a) Below the bile canaliculi, tight junction webs are seen that seal off the canalicular surfaces from
the appositional membrane, as seen by freeze‐fracture electron micrograph. (b) Thin‐section micrograph illustrates tight junctions (Tj) sealing the
canaliculus (arrowhead) and a large gap junction (Gj) below it. Freeze‐fracture (c) and thin‐section (d) views show larger gap junction plaques
isolated from tight junctional strands. (e) Immunofluorescence of mouse liver showing ZO‐1 (tight junction protein‐1, green) lining the bile
­canaliculi of the liver and the gap junction protein, connexin 32, outside of but in close proximity to the bile canaliculi. White arrows show bile
­canaliculi. Nuclei are shown in blue. (f) Freeze‐fracture micrographs of the adluminal plasma membrane where tight junction strands form well‐
developed networks and small gap junction plaques (g), observed peripherally or within the tight junction network. Bars = 0.1 nm or as indicated.
Modified from [3] with permission of Rockefeller University Press.

scattering provided early details to the model of the gap junction
channel [6]. It is now generally accepted that liver gap junction
channels are composed of 12 subunits, six contributed by each
cell, forming what is termed a hemichannel or connexon
(Figure 14.2). The subunits are radially symmetrical around the
central pore, and each is believed to be tilted slightly relative to
the plane of the membrane. Cryoelectron microscopy and com­
puter modeling have provided evidence that the formation of rat
liver gap junctions requires a 30 degree rotation between
hemichannels for proper docking [7]. The structure of a gap
junction channel consisting of connexin 26 in an open confirma­
tion has been determined to 0.35 nm resolution [8]. This depicts
the Cx26 gap junction connexon as a positively charged hexago­
nal funnel that narrows to a 0.14 nm central pore, which then
interconnects with the connexon of an apposing cell through an
interdigitated double‐walled extracellular cavity.
MOLECULAR COMPONENTS AND GENES
OF TIGHT AND GAP JUNCTIONS
Tight junctions
Rodent liver membrane fractions served as the source for the
initial isolation of tight junction‐associated proteins. The first of
these was a high molecular weight (~225 kDa) molecule, which
was termed zonula occludens 1 (ZO‐1, also known as tight
162 THE LIVER:  MOLECULAR COMPONENTS AND GENES OF TIGHT AND GAP JUNCTIONS

(a)
Tight junction proteins Gap junction proteins

1 2 3 4 1 2 3 4 1 2 3 4 1 1 2 3 4 1 2 3 4

CLDN1 CLDN2 OCLN JAM1

GJB1 GJB2

ZO1

(b) (c)
Tight junction channel Gap junction channel
(CLDN2)

Cell 1

Cell 2

Figure 14.2  Molecular components and topologies of liver tight and gap junctions. (a) Tight junction proteins are primarily tetraspan claudins
(Cldn1/2) and occludin (OCLN), but single‐pass proteins such as junction adhesion molecule (JAM1) are also components. Gap junction proteins
between healthy hepatocytes are primarily connexins 32 and 26 (Cx32 or GJB1 and Cx26 or GJB2), although Cx43 (GJA1) is also present in some
cells and may be induced following injury. The zonula occludens proteins ZO‐1 (shown) and ZO‐2/3, which do not contain membrane‐spanning
domains, are also prominent in liver and bind to both tight junction and gap junction proteins. Diagrams represent amino acid sequences of human
isoforms from Protter software. (b) Secondary structure of tight and gap junctions reveal contrasting ways in which these proteins form paracellular
and intercellular channels. On left is a conceptual drawing of a tight junction formed of Cldn2, illustrating the permeability pathway between the
cells proposed to be formed by hydrophilic residues in the extracellular beta barrels. To the right are paired hexameric connexons or hemichannels
forming bidirectional permeation pathways connecting interiors of adjacent cells.

junction protein 1, Tjp1), and discovery of the other family
members ZO‐2 and ZO‐3 (Tjp2 and 3) quickly followed (see
[4]). These proteins were soon realized to be members of the
membrane‐associated guanylate kinase (MAGUK) superfamily,
which contains domains (termed SH‐2, GUK, and PDZ) that are
specialized for binding to other proteins. Although ZO‐1/2/3 are
commonly associated with tight junctions, they are actually
scaffolding proteins that also bind to protein components of
other junctional types, including gap junctions. The first true
core component of tight junctions was identified as a 65 kDa
protein in junctional membrane fractions purified from chicken
liver and termed occludin (Ocln) [9]. Ocln was shown to form
tight junction‐like webs of intramembrane particles in trans­
fected cells and it seemed to satisfy criteria as the universal
transmembrane protein of tight junctions. It thus came as a
major surprise that the Ocln‐null mouse still had similar struc­
tures [10]. Re‐examining the purified chick membrane fractions
led to discovery of two additional somewhat smaller proteins
(~23 kDa), termed claudin‐1 and claudin‐2 (Cldn1/2), that
became the founding members of what is now known to be a
large family of 24 members in the human genome.
Although liver tight junctions are mainly composed of Ocln
and members of the Cldn family, other core molecular tight junc­
tion components include the junctional adhesion molecules
(JAMs), coxsackie adenovirus receptor (CAR), tricellulin, and a
few other identified or suspected molecules [11]. Tight junctions
in various tissues have distinct complements of these core pro­
teins, which can form both homo‐ and heterophilic interactions.
This diversity reflects in part a tissue‐specific organization, but it
also achieves various degrees of “tightness,” a topic considered
 14:  Gap and Tight Junctions in Liver: Structure, Function, and Pathology 163
in more detail in subsequent sections dealing with normal and
pathological functions.
The major tight junction components occludin and the claudins
are tetraspan proteins with intracellular N‐ and C‐termini and
two  extracellular loop domains (Figure  14.2). By contrast, the
JAMs and CAR are single‐pass membrane proteins. The first
loop of Cldns is twice the length of the second loop and its amino
acid sequence influences the paracellular charge selectivity,
whereas the second extracellular loop acts as a receptor for a bac­
terial toxin, and the C‐terminus binds cytoplasmic proteins,
including cytoskeletal elements via a PDZ motif. Discrete
­residues within the first extracellular loop of Cldn1 in liver are
critical for hepatitis C virus (HCV) entry [12, 13]. Like claudins,
Ocln possesses intracellularly located PDZ‐binding domains for
scaffold association proteins such as ZO‐1 that link it to the
cytoskeleton.
Numerous tight junction proteins are expressed in murine
­livers, including claudin‐1, ‐2, ‐3, ‐5, ‐7, ‐8, ‐12, ‐14, occludin,
JAM‐A, CAR and tricellulin; claudin‐1, ‐2, ‐3 are expressed in
the bile canaliculus region of hepatocytes [14, 15]. Claudin‐2
immuno‐localization shows a lobular gradient increasing from
periportal to pericentral hepatocytes, whereas claudin‐1 and ‐3
are expressed uniformly throughout the liver lobule. Further­
more, claudin‐2 expression induces cation‐selective channels in
tight junctions of epithelial cells, as do claudin‐7 and ‐15 [16].
Gap junctions
The first gap junction protein isolated from detergent‐­solubilized
gap junctions displayed an apparent molecular weight of
27 kDa, although a 21 kDa protein was also present (for histori­
cal overview, see [4]). Cloning of the cDNA encoding the more
slowly migrating protein [2, 17] indicated that its predicted
molecular weight is about 32 000, and this protein is generally
referred to as connexin32 (Cx32). In an alternative nomencla­
ture this gap junction protein is referred to as beta1 and its gene
as Gjb1 in rodents and GJB1 in humans [18]. The cDNA encod­
ing the 21 kDa protein was subsequently cloned and found to
encode a protein of predicted molecular weight of 26 000 [19].
This connexin is now termed Cx26 (or beta2) and its gene is
Gjb2 or GJB2.
Cx32 and Cx26, the core components of hepatocyte gap junc­
tions, are also found in a variety of other cell types (for review,
see [20]). Moreover, another gap junction protein, Cx43 (or
alpha1; gene Gja1 or GJA1) is prominent between other liver
cell types, including stellate cells, Kupffer cells, and endothelial
cells (for review, see [21]). The connexin gene family now com­
prises at least 20 proteins in vertebrates. Connexins are absent in
nonchordate genomes, where gap junction channels are formed
by a separate gene family encoding the innexin proteins [20].
Although chordate homologs of innexin proteins have been
identified, only one of these so‐called pannexin (Panx1/2/3)
proteins has been shown to form channels, and the channels that
Panx1 forms appear to be exclusively involved in formation of
nonjunctional, rather than junctional channels [22]. Pannexins
are found in the liver, where they have been proposed to play
roles in acetaminophen‐induced injury and other disorders [23].
Like Ocln and Cldns, connexins are tetraspan membrane pro­
teins. They exhibit very high homology among family members,
particularly in extracellular and transmembrane domains.
Homologous extracellular domains are thought to account for
the ability of many connexins to form heterotypic channels (i.e.
cells expressing one type of connexin forming gap junctions
with cells expressing a different connexin). The existence of
heteromeric connexons (i.e. hexameric connexons that contain
different connexin proteins) in the liver was initially suggested
from biochemistry experiments that fractionated detergent‐solu­
bilized gap junctions [24], and this was supported by the
electrophysiological properties of hepatocytes isolated from
­
Cx32‐deficient and wild‐type mice [25], and by studies on
reconstituted vesicles in which differences in permeability for
second messenger molecules were obtained when the ratio of
Cx32 to Cx26 was varied [26]. The greatest dissimilarities
among connexins occur in the amino acid sequence of the C‐ter­
minus, which is the domain thought to play the major role in
regulation of gap junction channels. It is within this cytoplasmic
region of the proteins that most of the potential phosphorylation
sites are found and against which most of the useful connexin‐
specific antibodies have been raised. Moreover, the connexin
C‐terminus contains binding sites for numerous protein p­artners,
including kinases and structural proteins in a scaffolding
­complex that has been termed the Nexus [27]. The tight junc­
tion‐associated protein ZO‐1 interacts directly with the extreme
C‐terminus of Cx43, utilizing the second PDZ domain of ZO‐1,
and Cx32 interacts with the SH‐3 domain of the PDZ‐scaffold­
ing protein discs large homolog 1 (Dlgh 1) in the liver [28, 29].
Furthermore, small gap junction plaques are associated with
tight junction strands in certain cell types, including hepatocytes
[5], and Cx32 is partially colocalized with Ocln and Cldn1 in
hepatocytes [30]. Such protein–protein interactions likely pro­
vide platforms localizing intracellular signal transduction to
contacts between the cells. In addition to the binding of
cytoskeletal proteins to liver connexins, immunoprecipitation
pulldown studies have identified a number of mitochondrial
proteins bound to Cx32, suggesting the possibility of gap junc­
tion involvement in mitochondrial signaling in the liver [31].
FUNCTIONS AND REGULATION
OF TIGHT JUNCTIONS IN THE LIVER
The barrier: permeability regulation
Tight junctions between hepatocytes provide a high‐resistance
barrier to leakage of water and solutes into and out of the bile
canaliculus, termed the blood–biliary barrier. The number of
tight junctional strands encircling the bile canaliculus is nor­
mally high, indicating that this permeability is low, which is
validated by resistance measurements from microelectrodes
inserted into the canalicular space [32]. As a consequence, this
barrier functions to maintain the composition of the canalicular
fluid, allowing the accumulation of high concentrations of bile
acids, phospholipids, and other preferentially secreted organic
anions within this intercellular compartment.
Approaches commonly used to investigate the barrier func­
tion of tight junctions involve measurement of the permeability
of the paracellular pathway to ions and uncharged hydrophilic
164 THE LIVER:  FUNCTIONS AND REGULATION OF TIGHT JUNCTIONS IN THE LIVER

(a) Fence Barrier

(d1) (e)

I1
V1
V2
(d2) –50

100 pA
I2

(b1) 5 sec
(c)
I1

(f)
(b2)

20 pA
I2

2 sec
Figure 14.3  Functions of tight and gap junctions. (a) Function and molecular components of tight junctions. The major tight junction functions
are termed “fence,” where membrane‐embedded molecules are segregated to apical and basolateral domains, and “barrier,” where aqueous mole­
cules in canalicular fluid are excluded from the blood. (b) Fence function images of tight junctions in liver. The hepatocytes are labeled with
BODIPY‐sphingomyelin, which is effectively retained from flowing past the apical domain (b1, arrow). In hepatocytes treated with 3 mM EGTA
for 5 minutes, the probe diffused through the tight junction, labeling the basolateral face (b2, arrow). (c) Barrier function illustrated with thin‐sec­
tion image. Tight junctions flanking the microvillus‐filled canaliculus exclude lanthanum, which had been introduced into the extracellular spaces
via the vascular system and is seen to penetrate the extracellular spaces but is denied access to the canalicular lumen by the tight junctions (arrows).
(d) Gap junctions permit diffusion of molecules up to a size limit of about 1 kDa between cell cytoplasms. Light (d1) and epifluorescence (d2)
photos showing that Lucifer Yellow CH (5% wt/vol) injected iontophoretically into one cell of a pair of hepatocytes passes detectably to a coupled
neighboring cell within 30 seconds. (e) Voltage‐clamp experiment on a pair of hepatoma cells stably transfected with Cx32. A 50 mV hyperpolariza­
tion (V2) of one cell produces an initial current in the other cell (I1), which decreases over a period of seconds. This voltage sensitivity is charac­
teristic of liver gap junction channels. (f) Single gap junction channel openings and closures can be seen in this example of a high‐again recording
from a poorly coupled pair of SKHep1 cells transfected with rat Cx32. In response to a driving force of 50 mV, channels continuously open and
close, as indicated by abrupt transitions of equal size but opposite polarity recorded in each cell’s voltage‐clamp circuit.

macromolecules. Such permeability is visualized using elec­
tron‐dense (e.g. ruthenium red, lanthanum) or fluorescent dyes
and can also be quantified as the transepithelial resistance (TER)
measured between the apical and basolateral compartments
[33]. Determination of the crystal structure of Cldn15 revealed
that the β‐helical extracellular domain forms a dimer that could
provide a paracellular permeation pathway if lined with hydro­
philic amino acid residues but would otherwise be impermeable
to charged molecules [34]. Of particular relevance to the liver,
Cldn2 has such hydrophilic residues in its extracellular domain,
whereas Cldn1 does not. Thus, Cldn2 provides an ion permea­
tion pathway across the paracellular barrier that is analogous to
the cell–cell channels connecting cell interiors provided by gap
junctions (Figure 14.2b). In the liver Cldn2 shows a distribution
that is graded across the lobule. This expression profile is
hypothesized to generate directional bile flow from the center to
the periphery of the lobule, as evidenced by deficient flow in the
Cldn2‐null mouse [35].
The fence: establishing and maintaining
cell polarity
Tight junctions also separate the apical from the basolateral cell
surface domains, thereby isolating membrane proteins to one
region or the other and establishing and maintaining cell polar­
ity. This is termed the “fence” function of tight junctions, and it
is quantified through measurements of ratio of proteins in the
two domains. One straightforward assay involves the use of
fluorescent lipids or lipophilic probes (such as BODIPY‐­
sphingomyelin). When added to medium bathing specifically
the canalicular or basolateral surface, the lipids insert into the
outer leaflet of the plasma membrane. Restriction of dye to one
compartment or the other can be compared under conditions in
which tight junctions are present or disrupted by Ca2+ chelation
(Figure 14.3b) [30].
Regulation of tight junctions
Both barrier and fence functions of tight junctions are highly
dependent on the integrity of the cytoskeleton. For example,
under optimal culture conditions, primary rat hepatocytes dis­
play well‐developed networks of tight junction strands, with
circumferential actin filaments near the tight junction regions
[30] (Figure 14.1). The tight junction molecules occludin, clau­
din‐1, ZO‐1, and ZO‐2 are observed by immunofluorescence at
the cell borders, and the fence function of tight junctions in the
cells is well maintained, as evidenced by retention of labeled
sphingomyelin (Figure  14.3b). Treatment of the cells with an
actin depolymerizing drug (mycalolide B) caused disappear­
ance of both the circumferential actin filaments and occludin,
while tight junction strands remained virtually intact. These
results provide evidence that occludin may be especially critical
 14:  Gap and Tight Junctions in Liver: Structure, Function, and Pathology 165
for providing strong linkage between the actin cytoskeleton and
tight junctions in hepatocytes [30].
A number of signaling molecules have been associated with
the regulation of tight junctions, including tyrosine kinases,
cAMP, Ca2+, protein kinase C (PKC), heteromeric G proteins,
and phospholipase C [36, 37]. A protocol of switching between
Ca2+‐containing and Ca2+‐free solutions has demonstrated that in
monolayers incubated in low calcium medium, tight junction
proteins are disassembled, dephosphorylated, and are less
tightly associated with actin filaments [30]. This indicates that
the function of tight junctions may be locally regulated by sign­
aling events within the tight junction plaque or may regulate
some aspect of intracellular signaling. Cytoplasmic signaling
pathways are believed to directly control the barrier function of
tight junctions. ATP‐depletion experiments result in altered
tight junction structure, decreased TER, and an increased asso­
ciation of tight junction proteins with the actin filaments [38].
The PKC‐activating phorbol esters induce a rapid decrease in
tight junction permeability and alter perijunctional actomyosin
[39]. Viral and bacterial pathogens have exploited core tight
junction proteins and their adaptors to disrupt the paracellular
barrier, and these strategies are in many examples associated
with a reorganization of perijunctional actin [13].
Tight junction disease
Hepatocyte tight junctions are believed to provide the primary
intercellular barrier between the sinusoidal and the canalicular
space. This barrier, often referred to as the blood–bile barrier,
is compromised under experimentally induced liver injury in
mice, with accompanying impaired bile secretion and jaun­
dice [40]. In tissue thin sections, the tight junctions of hepato­
cytes can be seen to block access of the experimentally
administered extracellular tracer (lanthanum) to the canalicu­
lar lumen (Figure  14.3c). Irregularities in the structure and
distribution of tight junctions accompanied by increased
blood–bile barrier permeability have been observed in experi­
mental mouse models including extrahepatic cholestasis after
common bile duct ligation, intrahepatic choleostasis after
ethinylestradiol treatment, partial hepatectomy, choline‐defi­
cient diet, hepatocyte injury following hepatotoxic drug
administration, and experimental colitis [30, 40]. Loss of gap
junctions, leaky tight junctions, and disorganized actin bun­
dles are considered to be sufficient to cause cholestasis and the
eventual development of jaundice. With advances in studies of
tight junctions, it has become clear that these three causes are
never independent of each other. Disruption of actin filaments
clearly deteriorates barrier function of tight junctions, and the
interactions among tight junction proteins has not been fully
clarified. Tight junction proteins have also been shown to
mediate entry of numerous hepatotropic viruses, and func­
tional and structural abnormalities of tight junctions are impli­
cated in liver cancer.
Regarding genetic diseases of human tight junction proteins,
missense mutations in ZO‐2 have been identified in patients
with familial hypercholanemia (FHCA: MIM phenotype 607748)
and in progressive familial intrahepatic cholestasis 4 (PFIC4:
615878). In a syndrome associating ichthyosis and neonatal
sclerosing cholangitis (NISCH syndrome: 607626), mutations
of claudin‐1 are also observed, and the lack of claudin‐1 may
lead to increased paracellular permeability between bile duct
epithelial cells [41].
HCV gains access to hepatocytes through binding to co‐
receptors on hepatocytes that include extracellular loop domains
of Cldn1 and Ocln and the tetraspanin membrane protein CD81
[12, 13]. The potential utility of targeting claudin tight junction
proteins to block infection by HCV is highlighted by a report
that monoclonal Cldn1 antibodies can clear HCV infection in
humanized mice [42]. Other tight junction components may
also bind viruses, and HCV appears to target multiple claudin
protein family members. Moreover, reovirus coat protein binds
to extracellular JAM‐A homodimerization domains, and CAR is
related to immunoglobulin junctional adhesion molecules and is
localized to junctional domains.
FUNCTIONS OF GAP JUNCTIONS
IN THE LIVER
As described above, hepatocytes express high levels of Cx32
and somewhat lower levels of Cx26. Kupffer cells, stellate cells,
cells of Glisson’s capsule, cholangiocytes, and sinusoidal
endothelial cells all express Cx43. Interestingly, chronic or
acute liver injury can induce Cx43 expression in hepatocytes but
the significance of this is unclear [43]. Endothelium of the por­
tal vein and arteries, however, express Cx37 and Cx40, which
are commonly expressed in endothelia and smooth muscle.
Hepatocytes and Kupffer cells also express hemichannel‐form­
ing Panx1. As mentioned above, these are related to the inverte­
brate gap junction proteins innexins, but in the chordate animal
phylum (which includes vertebrates) it appears that gap junction
function has been taken over by connexins and that pannexins
do not form communicating channels between cells [21, 22].
Gap junctions closely align with the functional structure of
hepatocytes and therefore to the functional structure of the
entire liver. Gap junctions form in close association with tight
junctions and adherens junctions, and these all come together to
organize the apical membrane of the hepatocyte, the bile canali­
culus (Figure  14.1). As is seen elsewhere in this volume, bile
canaliculi take the form of narrow tubules that cross the lateral
surface of hepatocytes. They contain microvilli on their luminal
surface that provide high surface area to the crowded bile cana­
licular interior. Bile canaliculi give a characteristic chickenwire
appearance across the liver acini when seen by fluorescence
microscopy (Figures 14.1 and 14.4). Cx32 and Cx26 are seen to
align along this chickenwire, forming irregular disk‐shaped
plaques, rather than tubules. Since gap junctions are associated
with the apical junctions of hepatocytes [5], injuries or disease
that can cause loss of hepatocyte polarity or loss of functional
organization of the hepatic acinus can also alter gap junction
localization and expression [21].
Zonation of the liver
Gap junctions couple cells electrically by allowing ions to pass
between the channels such that a change membrane potential
experienced by a cell can be shared by its neighbor through a
166 THE LIVER:  FUNCTIONS OF GAP JUNCTIONS IN THE LIVER

Hepatic lobules

P.triad

P.triad

Figure 14.4  Exposure to toxic doses of acetaminophen leads to centrilobular necrosis and increased expression of connexin 43. Immunofluorescence
microscopy images of connexins 32 and 43 in C57BL6 mouse livers treated with 500 mg kg−1 acetaminophen (APAP) or control solvent alone (Ctl).
Twenty‐four hours after IP administration of APAP, livers were removed, frozen and subsequently sliced and co‐stained with antibodies to Cx32
and Cx43 followed by fluorescent secondary antibodies to allow visualization of both proteins in the same microscopy fields. Control livers (top
panels) show punctate plaques of Cx32 (left panels) that dot the borders of the hepatocytes in a chickenwire‐like appearance, while Cx43 (right
panels) staining is minimal except in the fibroblast cells of Glisson’s capsule (white arrow). APAP‐treated livers (lower panels) show autofluores­
cent necrotic cells (nec.) and punctate staining of both Cx32 and Cx43 throughout the liver. The cartoon at the right depicts the lobular organization
of the liver along with its metabolic gradient zones (1, 2, 3) and portal triads (P. Triad), central veins (CV), and centrilobular necrosis (nec.) that
occurs with APAP toxicity. The approximate locations of these zones of the liver are also indicated in the lower panels.

flow of ions. In neurons, gap junctions are referred to as elec­
trical synapses, and they function to transmit impulses and syn­
chronize the cells. The physiological consequences of high
levels of gap junction expression that is seen in hepatocytes
includes not only ionic coupling, but also transfer of cellular
signals, especially those involved in growth control and metab­
olism. Relaying of metabolic signals is important not only
because of the anatomical arrangement of the hepatic lobule
but also because many physiologic functions show gradations
along the lobule, from the periportal region (containing the
portal triad) to pericentral regions (containing the terminal
hepatic venule). That is, the liver contains functional metabolic
zones 1, 2, and 3 that reflect the level of oxygen and the flow
blood through the sinusoid from perportal region to the central
vein (Figure  14.4). Interestingly, bile flows in the opposite
direction. Many metabolic enzymes exhibit ­asymmetric zona­
tion. For example, hepatocytes with higher gluconeogenic
activity are found periportally rather than pericentrally [44,
45], and glutamine synthetase, which removes ammonia from
blood, is strictly localized to hepatocytes surrounding the cen­
tral vein (zone 3). Connexin 26 also exhibits asymmetric
zonation, with stronger staining in the periportal region (zone
1), but the significance of this is still unclear [46]. Zonal locali­
zation of metabolic enzymes and the activity of gap junctions
presents a conceptual dilemma; if substrates can freely diffuse
through the liver via gap junctions then why would it be benefi­
cial to localize enzymes to specific zones? There may be
numerous answers to this question depending the enzymes
being examined, but one explanation is that the substrates and
various molecules can still form gradients within the acinus
even though they can diffuse between cells. For instance, it has
been shown that changing the direction of flow of an isolated
perfused liver will change the zonal location of gluconeogen­
esis [47]. Nevertheless, the strong electrical coupling mediated
by gap junctions serves to equilibrate the membrane potential
among hepatocytes. In isolated rat livers, perfusion with gluca­
gon leads to hyperpolarization of all hepatocytes across the
hepatic acinus. However, in the presence of octanol, a gap
junction blocker, glucagon‐induced hyperpolarization is higher
in periportal hepatocytes [48] Furthermore, heptanol blockade
of gap junctions has been shown to abolish metabolic and
hemodynamic effects of nerve stimulation within the liver [49].
 14:  Gap and Tight Junctions in Liver: Structure, Function, and Pathology 167
Calcium signaling
Calcium is a major intracellular signaling molecule and the
­permeability of Ca2+ ions and the second messenger, inositol
trisphosphate (IP3), has important roles in coordinating the
­ in the gap junctions, but by 26 hours most hepatocytes have lost
response of hepatocytes to environmental fluctuations. Imaging
studies on hepatocyte pairs has shown that injection of Ca2+ into
one cell leads to a rise in Ca2+ in the adjacent cell [50]. This
intercellular spread can be blocked by treatment of the cells
with the gap junction inhibitor heptanol, indicating that move­
ment is by way of gap junction channels. Although both IP3 and
Ca2+ ions can pass through gap junctions, most models indicate
that IP3 is the primary propagating agent. When injected into a
cell, IP3 increases intracellular Ca2+ in injected cells and their
coupled partners. Because the Ca2+ rise in the recipient cell can
occur far from the junctional region, and because the traveling
waves are very fast, diffusion of Ca2+ itself cannot account for its
propagation. In fact, Ca2+ is strongly buffered by the intracellu­
lar environment and cannot easily diffuse. Instead the system
relies on signaling molecules such as IP3 and its metabolites
[51] that can trigger rapid release in distal areas. Intercellular
Ca2+ signaling is commonly used by cells to coordinate and
regulate a wide range of cellular functions including cell growth
and differentiation [52]. Ca2+ waves also contribute to the vari­
ous functions of liver associated with vasopressin and norepi­
nephrine stimulation [53], and increased Ca2+ induces bile
canalicular contraction and secretion of bile by hepatocytes and
cholangiocytes [54]. Gap junction‐mediated regenerative waves
of elevated Ca2+ appear to provide long‐range signaling for the
propagation of these contractions.
In addition to transmission through gap junction channels,
Ca2+ waves may also propagate via rises in extracellular nucleo­
tides that are detected by adjacent and non‐adjacent cells. This
latter mechanism utilizes the activation of purinergic receptors
by ATP and related nucleotides that are released from the stimu­
lated cells at least partly through Panx1 channels [55]. This type
of signaling can be inhibited by purinergic receptor antagonists
or apyrase, and hepatocytes can be desensitized to this signaling
by inclusion of ATP itself [56]. Cholangiocytes also utilize
purinergic signaling for bile secretion, and this appears to be
activated by ATP released from within bile [54]. Intracellular
Ca2+ signaling within the liver represents a confluence for
responses to changes in the environment, and gap junctions as
well as purinergic signaling participate in how these responses
are propagated.
Growth control
Control over the growth of tissues has long been proposed as an
important function of gap junctions. Evidence supporting this
association include loss of functional coupling in liver tumor
cells, reduced gap junction expression in hepatomas compared
with that in adjacent normal tissue, reduced rate of tumor growth
in cells transfected with connexins, and loss of gap junction
communication in response to chemical tumor promoters [57].
Models that have been useful for understanding the association
between gap junction expression and growth control in liver
include the regeneration of liver following surgical or chemical
insults. Following partial hepatectomy, whereby approximately
two thirds of the liver is surgically removed, gap junctions in the
recovering liver undergo a strikingly coordinated cycle of disap­
pearance and reappearance over a time course of 48 hours (for
reviews see [5, 58]). For the first 20–24 hours there is no change
these junctions. By 36 hours, gap junctions begin to reappear
and then fully recover by 48 hours. The disappearance corre­
sponds to a peak in mitosis. Electrical measurements indicate
that junctional conductance decreases and increases with a simi­
lar time course, although increased nonjunctional resistance
contributes to the coupling measurements. Using immunoblot
techniques, Cx32 was also found to undergo a similar cycle of
decrease and reappearance, with the decrease coinciding with a
time of maximal DNA synthesis. However, Cx26 mRNA was
found to be selectively increased before the onset of S‐phase
after partial hepatectomy as well as after bile duct ligation.
Freshly isolated primary hepatocytes, which generally do not
undergo cell division, also undergo changes in the expression of
gap junction proteins and electrical coupling that are similar to
those that are seen following tissue injury in situ; gap junctions
initially disappear and then reappear over a time course of days
[59].The disappearance of gap junctions in hepatocyte cell cul­
ture can be delayed by treatment with cAMP or glucagon or
with protein synthesis inhibitors, and connexin mRNA stability
is increased under these conditions. These findings indicate that
that reduction in gap junction expression and function correlates
with cellular reprogramming and is permissive for proliferation
in the liver, though not absolutely required. In this regard it is
also possible that loss of gap junctions could provide a selective
advantage for preneoplastic liver cells as they develop into rap­
idly proliferating tumor cells.
Regulation of coupling between hepatocytes
The extent to which signals spread in the liver depends on the
number of open gap junction channels. Gap junction channels
between hepatocytes are closed by, among other things, cyto­
plasmic acidification, cellular injury, high extracellular Ca2+,
CC14, certain alcohols, and large voltage gradients (for reviews,
see [2, 60]).
Phosphorylation of gap junctions
Another important avenue of gap junction regulation is phos­
phorylation. All connexins, with the exception of Cx26, appear
to be substrates for phosphorylation, and Cx43 has been shown
to be phosphorylated at 16 different sites [57]. Connexin 32 is
phosphorylated in its cytoplasmic tail at Ser233 by cAMP‐
dependent protein kinase (PKA), and this correlates with
increased conductance [61]. Ca2+‐calmodulin‐dependent protein
kinase II and protein kinase C can also phosphorylate Cx32, and
these can produce tryptic digestion patterns that are distinct
from those obtained with cAMP‐dependent protein kinase.
Phosphorylation by protein kinase C appears to inhibit some
types of proteolysis [61]. Although in other cell types inhibition
of junctional communication has been associated with the
expression of tyrosine kinases, Cx32 in isolated rat liver gap
junctions is not phosphorylated by purified pp60v‐src tyrosine
kinase or the insulin receptor.
168 THE LIVER:  FUNCTIONS OF GAP JUNCTIONS IN THE LIVER
Oligomerization of gap junctions
As with other transmembrane proteins, gap junction proteins are
co‐translationally inserted into the endoplasmic reticulum (ER).
The nascent connexin protein must fold correctly and oli­
gomerize with five other subunits to form a hexameric complex,
the connexon. Connexins have their N‐ and C‐termini located
on the cytoplasmic face and recently, compatibility motifs have
been identified that roughly predict which connexins can het­
ero‐oligomerize. The compatibility motifs were identified by
amino acid sequence homology clustering, and allow categori­
zation of connexins as either R (arginine) or W (tryptophan) or
other, based on the amino acids that are present. The motif is
found at the junction between the cytosolic intracellular loop
and the third transmembrane domain. Connexins 26 and 32 are
W types, whereas Cx43 is R type. Cx43 is unusual in that its
monomer form is stabilized in the ER, whereas Cx32 and Cx26
appear to oligomerize within the ER. These form hetero‐
oligomers as is predicted by their compatibility motif.
Oligomerization and trafficking through the various compart­
ments in the cell appear to be assisted by chaperones, and the
regulation of oligomerization in the ER and Golgi apparatus
suggests the presence of patrolling complexes that prevent inter‐
connexin association [62].
Trafficking of and cytoskeletal interactions
of gap junctions
Gap junction trafficking pathways beyond the ER and Golgi
compartments are not clearly defined and it is plausible that
connexons made up of different connexin subunits use alternate
routes. Growing evidence is solidifying the view that microtu­
bules play a major role in regulation of connexin trafficking
[63]. The dependence of gap junction regulation on cytoskeletal
components such as microtubules and microfilaments presents a
mechanism for directed trafficking within the cell. In most cell
types, microtubules stretch across the length of the cell and form
railroad tracks for the movement of organelles and macromole­
cules that may otherwise be trapped by the crowded nature of
cytoplasm. In hepatocytes, like in other polarized cell types,
microtubules extend from the microtubule organizing center as
well as from the apical domain itself. Therefore, the slow‐
growing minus ends of microtubules are found near the bile
canaliculus and the fast‐growing plus ends are found near the
basolateral plasma membrane, although some microtubules are
found in other orientations. Studies from our group have confirmed
that Cx32‐containing vesicles isolated from liver and within
cultured cells traverse along microtubules in the plus end direc­
tion using the microtubule motor kinesin 1 [64]. This suggests
that connexins are delivered to the basolateral membrane, where
they form large gap junction plaques that localize to the borders
of the apical–basolateral domains (see Figure 14.1). Connexins,
and Cx43 in particular have been shown to not only bind micro­
tubules, but to traffic to adherens junction‐associated mem­
branes containing N‐cadherin and beta‐catenin. Knockdown of
EB1 protein has been shown to reduce gap junction plaque for­
mation, and EB1 may function to deliver connexins and other
junctional proteins to the plasma membrane through its micro­
tubule plus end tracking activity [63]. Furthermore, embryonic
fibroblasts from Cx43‐knockout mice show aberrant cell
­polarity and reduced wound closure, and transfection with dom­
inant‐negative Cx43 that lacks a putative microtubule‐binding
domain reconstitutes this phenotype [65]. Interestingly, in
hepatocytes vesicle delivery of biosynthetic components such
as lipoprotein occurs at the sinusoidal face, near to plus ends of
microtubules, while gap junction plaques are found in close
proximity to the apical domain where microtubule minus ends
are abundant. Liver injury or other conditions that can disrupt
hepatocyte polarity can induce expression of Cx43 (Figure 14.4),
and the above findings suggest that Cx43 may take advantage of
altered cell polarity during such situations.
Microfilaments form another important cytoskeletal compo­
nent for junctional proteins. The filamentous actin network is
integral to the structure of cell junctions and the shape, polarity,
and contractile activity of cells in general. In hepatocytes, bile
canaliculi are frequently identified by the abundance of actin
that surrounds them. For delivery of junctional proteins, actin is
unlikely to serve as a railroad track, since actin filaments are
much shorter than microtubules and frequently form an iso­
tropic network. However, actin may participate in anchoring
vesicles containing connexins to particular regions of the cell
and they may provide a directional bias to allow vesicles and
various macromolecules to form contacts and, for example, pro­
mote fusion of vesicles with the plasma membrane. Additionally,
there is considerable literature linking gap junction protein
expression with the regulation of cell motility, cell adhesion,
and migration. These studies have primarily focused on Cx43,
which has an unusually long C‐terminal tail, but recent studies
also implicate Cx32 and Cx26 in cell migration [66]. Neurons
from developing mouse brain as well as embryonic fibroblasts
from Cx43‐knockout animals exhibit reduced and disorgan­
ized cell migration, whereas overexpression of Cx43 has led
to  enhanced migration of multiple cell types in both wound
healing and transwell migration assays. Short hairpin RNA
(shRNA)‐mediated knockdown of Cx43 has also led to disor­
ganized cell polarity and reduced migration. Although the
mechanisms by which connexins may affect cell polarity and
motility are not understood, it has been suggested that Cx43 and
other connexins can act as dynamic scaffolding proteins that
organize and attract actin‐modifying proteins to cell–cell junc­
tions and to sites of cell adhesion and protrusion [67]. As indi­
cated earlier, Cx43 has also been shown to bind microtubules,
and the actin‐organizing and microtubule‐binding activities may
work together to control cell shape and movement. In this regard
it is interesting that adult hepatocytes, which are typically non‐
migratory and do not express Cx43, can be induced to express
this connexin during tissue injury, whereas migratory cells of
the liver, such as Kupffer and stellate cells, express Cx43.
Once within the plasma membrane, connexons are believed
to diffuse laterally and dock with opposing connexons from
others cells. Hepatocytes form gap junctions with all neighbor­
ing hepatocytes, making approximately six cell–cell contacts.
It is unclear whether this process is active or passive; however,
cell adhesion molecules are known to play a role in the forma­
tion of gap junctions. Studies using tetracysteine and GFP‐
tagged Cx43 have demonstrated that the insertion to membrane
plaques occurs laterally onto the edges, whereas removal takes
place from the center of the plaque [68, 69] in a process that
 14:  Gap and Tight Junctions in Liver: Structure, Function, and Pathology 169
may involve clathrin/actin‐mediated endocytosis. Because
docked connexons cannot be separated by most physiological
conditions, junctional removal appears to be a phagocytic‐
like  event which involves the formation of double membrane
vesicles, termed annular junctions. It is unclear whether this is
the only form of internalization for gap junctions since annu­
lar membrane profiles are not commonly observed in normal
liver [5].
The half‐lives of connexins are unusually short for trans­
membrane proteins, with measurements of 3–6 hours for mouse
liver connexins [2]. It has been shown that internalized gap
junctions are degraded by lysosomes and proteasomes, and via
autophagic processes [68]. Connexins have also been found to
inhibit autophagy, whereas nutrient starvation can promote
release of this inhibition resulting in degradation of connexins
and further upregulation of autophagy [70].
PATHOLOGY ASSOCIATED WITH GAP
JUNCTIONS AND TIGHT JUNCTIONS
IN LIVER
Junctional proteins within the liver are critical for proper organ
development and maintenance of biological function and over­
all health. Genetic deficiencies in gap junction proteins can
result in deafness, neuropathy, cataracts and other eye disorders,
and heart, skin, and connective tissue disease [71]. However, the
liver does not appear to absolutely require communication
through gap junctions in the adult. Although, as described ear­
lier in this chapter, hepatocytes of the liver primarily express
connexins Cx32 and Cx26, they also express pannexins [20].
Pannexins have only been observed to form hemichannels,
which link the cytosol to the cell exterior rather than another
cell, but they may still allow indirect cell–cell communication
through the extracellular space.
Genetic animal models of connexin deficiency
Gap junction beta 1 (GJB1) is notable as a genetic target for
mutations that cause X‐linked Charcot–Marie–Tooth disease
(CMTX). GJB1 encodes Cx32 and more than 300 distinct muta­
tions of Cx32 have been found to cause CMTX [72]. In an
attempt to mimic CMTX disease in a rodent model, Cx32‐­
deficient knockout mice (Cx32‐KO) were generated through
homologous recombination [73]. Although these mice demon­
strated progressive demyelination typical of the human disease,
they also showed deficiencies that have not been observed in
CMTX patients. Liver function was compromised, as evidenced
by reduced glucose release from liver in response to s­ ympathetic
nerve stimulation or hormone stimulation [74]. Cx26 levels
were decreased, as were the areas of gap junction plaques;
­electrophysiological conductance and permeability to IP3 was
also decreased compared to wild type [75]. Additionally, the
proliferation rate of Cx32‐KO hepatocytes in vivo was high, and
spontaneous and chemically induced liver tumors were more
prevalent [76], and the mice showed increased carcinogenesis
following X‐ray radiation [77]. Our laboratory compared cul­
tured hepatocytes from Cx32‐KO and wild‐type mice using
serum‐free medium and found a markedly higher proliferation
rate of Cx32‐KO hepatocytes [78]. Interestingly, however, the
proliferation rate of Cx32‐KO hepatocytes after partial hepatec­
tomy was significantly lower [79]. Experiments on transgenic
mice expressing a dominant‐negative mutant of Cx32 have
found delayed liver regeneration and an increase of susceptibil­
ity to chemical hepatocarcinogen [80]. Despite all of this, non­
functioning Cx32 in CMTX patients is not associated with liver
abnormalities. The lack of reported liver complications in
CMTX patients could be due to greater innervation of human
liver compared to that of the mouse, which perhaps overcomes
some of the functions of gap junctions, but it is also unclear
whether liver cancer rates or other properties have been meas­
ured in these patients.
Connexin deficiency and liver cancer
Studies using Cx32‐KO mice have supported the notion that
intact gap junction cellular communication functions to sup­
press tumors and that loss of gap junction communication can
lead to tumors [21]. Aged Cx32‐KO mice have also been shown
to have more spontaneous liver tumors in male animals [81]. As
we have noted, hepatocellular carcinoma is typically accompa­
nied by reduced gap junction activity and loss of Cx26 or accu­
mulation of Cx32 in the cytosol rather than plasma membrane
can increase the cancer potential of liver cells [22]. Connexins
themselves can become cytoplasmically localized instead of
forming cell–cell junctions and there is evidence that cytoplas­
mic localization and downregulation may promote hepatocel­
lular carcinoma, invasion, and metastasis [82]. Furthermore, in
important early studies on gap junction function in liver, loss of
intercellular coupling was shown to be a hallmark of cancer
cells [57]. Induction of connexin expression in many cancer
cell lines has been shown to reduce cell growth and invasion
phenotypes and regulate epithelial‐to‐mesenchymal transition
[57]. Mouse models of cancer have shown variable effects. In
some cases it appears that gap junctions can promote tumori­
genesis by increasing blood vessel growth and distributing
chemokines [57], while many studies now indicate that
increased gap junction expression in late‐stage tumors can
­promote metastatic features.
Non‐gap junction activities are also thought to play a role in
possible tumor‐promoting phenotypes of connexins. When
unpaired, connexins form hemichannels at the cell surface that
can allow distribution of cell contents into the extracellular
space. Under normal conditions, hemichannels are believed to
remain closed. However, under situations of decreased extracel­
lular Ca2+ or pHi, hemichannels may open and for instance
release ATP and other signaling molecules that can promote
inflammation, which in turn can promote carcinogenesis.
Gap junctions exchange small molecules between cells and
therefore are expected to contribute to cancer treatments that
involve small molecules or any soluble effectors that can pass
through gap junctions (i.e. that have a molecular weight of
approximately 1500 Da or less). This includes ions, nucleotides,
metabolites, bile acids, and even peptides and miRNAs.
However, the effect of gap junctions toward cell survival to toxic
small molecules is complicated by the “kiss of death, kiss of
life” conundrum, sometimes referred to as the “bystander”
170 THE LIVER:  PATHOLOGY ASSOCIATED WITH GAP JUNCTIONS AND TIGHT JUNCTIONS IN LIVER
versus “Good Samaritan” effect [83]: gap junctions might serve
to distribute toxins such that all cells are affected, or they might
serve to distribute toxins such that the toxin is no longer c­ ritically
damaging. Additionally, gap junctions can distribute protective
antioxidants, such as glutathione, that can detoxify cells. In the
liver, for instance, a toxic drug might distribute through a hepatic
cord in such a way as to allow cytochrome P450 or conjugating
enzymes to modify the drug and allow it to be excreted or fur­
ther metabolized. Some studies have shown that gap junction
inhibition can block paracrine signaling between cancer cells
and non‐cancer cells, and mouse liver and lung metastases have
been reduced and survival prolonged following treatment with
the gap junction inhibitor oleamide in combination with an anti‐
VEGF antibody [84].
It remains an important question whether therapeutics to pro­
mote or inhibit gap junction function could be beneficial for the
treatment of cancers, including liver cancers. It appears that
such treatments must take into effect the complex role that gap
junctions are now thought to play in the cancer process.
Historically, gap junctions have been considered tumor suppres­
sors. Changes in gap junctions during fibrosis progression have
also been well documented in the liver. Hepatic stellate cells
(also known as Ito cells) play a major role in fibrosis of the liver
under many settings and a reduction in their activation and
growth in the liver represents a major goal in the prevention of
liver fibrosis. During stellate cell activation the level of Cx43
increases on these cells as well as on hepatocytes and blockage
of gap junction and hemichannel signaling can reduce fibrosis in
mice [85]. However, induced expression of Cx43 in hepatocytes
appears to be beneficial during liver injury and its inhibition can
be detrimental [21]. Also interestingly, there is a significant
interaction of connexin proteins with mitochondria. In models
of ischemic injury in the heart and brain, Cx43 has been found
associated with mitochondria and provides protection during
reperfusion. Cx32 has also been found associated with rat liver
mitochondria and may provide a tethering link between the
hepatocyte plasma membrane and mitochondria [31]. It there­
fore appears that understanding the precise function of gap junc­
tions in specific cells and at specific times during injury will be
critical for potentially alleviating liver injury. With regard to
uncontrolled cell growth and cancer, gap junction communica­
tion has for many years been viewed as limiting and tumor sup­
pressive. This view is now being replaced by more complex
models where gap junctions may be tumor suppressors in early
stages of cancer but can become tumor promoters at later stages.
The role of gap junctions in drug‐induced
liver injury
Drug‐induced liver injury is the most common cause of acute
liver failure in the United States and a large portion of the cases
involve the widely available pain and fever‐reducing drug aceta­
minophen (acetyl‐para‐aminophenol or APAP). Liver injury is
also a primary concern in drug development and a principal rea­
son for the removal of drugs from the clinical pipeline. Drugs that
cause harm to the liver are described as either intrinsically toxic,
with predictable dose‐dependent effects, or idiopathically toxic,
with rare and unpredictable effects. Idiopathic drugs are difficult
to study and likely dependent on multiple converging variables,
whereas the study of intrinsically toxic drugs has been crucial in
understanding how liver toxicity unfolds. Some of the most stud­
ied intrinsically toxic drugs in rodents are acetaminophen,
thioacetamide, carbon tetrachloride, dimethylnitrosamine, and
d‐galactosamine, with acetaminophen being the most prominent
and directly clinically relevant [86]. Acetaminophen is nontoxic
at moderate dose but severely toxic and potentially life threaten­
ing at high dose (e.g. >6 g in adults). Like many drugs, acetami­
nophen is metabolized by the liver where it reacts with
cytochromes and is conjugated to glutathione, sulfate, and glucu­
ronic acid, allowing for further metabolism and excretion.
However, at high doses glutathione becomes depleted, and a toxic
reactive metabolite, N‐acetyl‐p‐benzoquinoneimine (NAPQI),
accumulates and can instigate a chain reaction of hepatocyte
necrosis that initiates around central veins of liver lobules
(Figure 14.4). Notably, cytochromes CYP2E1 and CYP3A4 that
catalyze the conversion of acetaminophen into NAPQI also
strongly localize around central veins. The propagating cycle of
necrosis involves reactive oxygen species (ROS), mitochondrial
oxidative stress and dysfunction, activation of Jun N‐terminal
kinase, and an inflammatory response [21, 86].
Gap junctions and gap junction proteins are expected to con­
tribute to this in multiple ways. For instance, although APAP
can enter cells by diffusion, gap junction channels could distrib­
ute cellular signals, ions, and critical molecules, and hemichan­
nels could extrude inflammatory signals such as ATP and other
and damage‐associated molecular pattern molecules. On the
other hand, gap junctions could also distribute detoxifying com­
pounds such as glutathione or UDP‐glucuronic acid to hepato­
cytes where these may be depleted. A critical step in the
progression of toxicity appears to be the spreading of centri­
lobular necrosis to the rest of the liver. In 2004 Asamoto and
colleagues [87] described the generation of transgenic rats that
express hepatocyte‐specific dominant‐negative Cx32 under
control of the albumin promoter. This resulted in decreased gap
junction activity, reduced Cx32 and Cx26 membrane localiza­
tion, and resistance to the hepatic toxins carbon tetrachloride
and d‐galactosamine. A second study then found that these
rats were resistant to acetaminophen toxicity, showing lowered
serum aminotransferase levels and improved liver histology
compared to wild type [88]. In addition, it was found that aceta­
minophen‐induced cell death in isolated hepatocytes from
Cx32‐KO mice was no longer synchronized as it is in wild type.
Connexin expression was also required to allow cell‐to‐cell pro­
tection of coupled hepatocytes from female to male animals.
Another study observed that Cx32‐KO mice themselves were
resistant to acetaminophen and thioacetamide toxicity as
observed by aminotransferase levels, histology, and death of the
animals. Toxicity could also be reduced by the gap junction‐
blocking drug 2‐aminoethoxy‐diphenyl‐borate (2‐APB) [89].
Although these studies appeared compelling, a number of
doubts have been raised. As described earlier in the chapter,
Cx32‐KO mice display increased levels of chemically induced
liver tumors, suggesting that they have increased susceptibility
to drugs. Another study indicated that Cx32‐KO mice have
reduced levels of centrilobular glutathione, and that they are in
fact more susceptible to acetaminophen. Maes and colleagues,
who have studied acetaminophen toxicity extensively, weighed
 14:  Gap and Tight Junctions in Liver: Structure, Function, and Pathology 171
in on these issues by performing their own studies. They found
that Cx32‐KO mice have similar levels of cell death, inflamma­
tion, and oxidative stress in response to acetaminophen, with
somewhat lower levels of protein adduct formation [90]. The
gap junction blocker 2‐APB was also shown to exert protective
effects through inhibition of CYP450s and c‐Jun N‐terminal
kinase. Follow‐up reports by some of these investigators have
now found that connexin and pannexin hemichannels may be
important, as specific inhibition of hemichannels provided pro­
tection from drug‐induced liver injury [91]. Collectively, these
studies indicate that all of the contributors to acetaminophen
toxicity are not well resolved. Many technical details can influ­
ence the results, such as the solubilizing agent DMSO, and the
species of mouse or rodent; for instance, rats are less susceptible
than mice [92]. Connexin proteins and the activity of gap junc­
tions (and also Panx1) appear to feature in these processes, but
at present a clear protective or damaging role in this toxicity
cannot be assumed and it remains to be seen whether gap block­
age can be useful in the clinical setting.
PROSPECTS AND PERSPECTIVES
The two most prominent junctional types in liver are the gap
junctions, which provide direct intercellular communication,
and the tight junctions, which serve to partition membrane
domains of individual cells and to occlude extracellular space,
restricting pericellular diffusion. The integral membrane pro­
teins of gap junctions, the connexins, have been structurally
well characterized, but their newly recognized roles in binding
to and organizing cytoplasmic proteins suggest that they may
form nucleation sites for intracellular as well as extracellular
signaling. Connexins interact with actin and microtubules as
well as mitochondria, and they participate in recovery from
injury in many tissues, although their precise role as both gap
junctions and hemichannels (where Panx1 may also contribute)
in drug‐induced liver injury remains enigmatic. By contrast,
peripheral tight junction proteins were identified prior to the
core proteins, and one area of intense interest is the diversity of
“tightness” profiles provided by the large claudin family of
integral membrane proteins. Although gap and tight junctions
perform different functions, there are numerous points at which
these junctional types appear to overlap. Indeed, the findings
that the traditionally tight junction‐associated protein ZO‐1
also binds to connexins, and that occludin colocalizes with
Cx32 indicate the possibility for either coordinate or reciprocal
regulation of macromolecular complexes containing gap and
tight junction proteins. Studies of protein–protein interactions
and of coordinate and subordinate regulation of gene families
may elucidate the intricacies of inter‐ and intracellular signal­
ing concerning growth control by gap junctions and mainte­
nance of the “blood–biliary barrier” formed by tight junctions
[30, 93]. As with the HCV receptor CD81, the major gap junc­
tion components (connexins) and the major tight junction com­
ponents (occludin and claudins) are tetraspan proteins with
critical intracellular and extracellular domains. The appropria­
tion of tight junction proteins as portals for virus entry suggests
special properties for these proteins. As tight junctions actively
form the apical–basolateral junction, it is possible that this loca­
tion is a protective niche for viruses or that viruses can exploit
temporary disruptions of the blood–biliary barrier for their
entry. Future studies are likely to reveal new functions of
connexins, occludin, and claudins that affect the processes
­
of  viral entry as well as downstream signaling and eventual
induction of hepatitis.
REFERENCES
1. Kojima T., Sawada, N., Yamaguchi H., Fort A.G., and Spray D.C. Gap and
tight junctions in liver: composition, regulation, and function, in The Liver:
Biology and Pathobiology, 5th edn (eds. I.M. Arias et al.), Wiley‐Blackwell,
Chichester, 2009, pp. 201–20.
2. Spray, D.C., Saez, J.C., Herzberg, E.L. et al. Gap junctions in liver: composi­
tion, function and regulation, in The Liver: Biology and Pathobiology, 3rd
edn (eds. I.M. Arias et al.), Raven Press, New York, 1994, pp. 951–67.
3. Spray, D.C., Ginzberg, R.D., Morales, E.A., Gatmaitan, Z., and Arias, I.M.
Electrophysiological properties of gap junctions between dissociated pairs of
rat hepatocytes. J Cell Biol, 1986;103(1):135–44.
4. Franke, W.W. Discovering the molecular components of intercellular
­junctions  –  a historical view. Cold Spring Harb Perspect Biol, 2009;1(3):
a003061.
5. Revel, J.P., Yancey, S.B., and Meyer, D.J. Cellular architecture and intercel­
lular communication in rat liver. JAMA, 1981;245(9):958–9.
6. Unwin, P.N. and Ennis, P.D. Two configurations of a channel‐forming mem­
brane protein. Nature, 1984;307(5952):609–13.
7. Perkins, G.A., Goodenough, D.A., and Sosinsky, G.E. Formation of the gap
junction intercellular channel requires a 30 degree rotation for interdigitating
two apposing connexons. J Mol Biol, 1998;277(2):171–7.
8. Maeda, S., Nakagawa, S., Suga, M. et al. Structure of the connexin 26 gap
junction channel at 3.5 A resolution. Nature, 2009;458(7238):597–602.
9. Furuse, M., Hirase, T., Itoh, M. et  al. Occludin: a novel integral mem­
brane  protein localizing at tight junctions. J Cell Biol, 1993;123(6 Pt 2):
1777–88.
10. Saitou, M., Fujimoto, K., Doi, Y. et al. Occludin‐deficient embryonic stem
cells can differentiate into polarized epithelial cells bearing tight junctions.
J Cell Biol, 1998;141(2):397–408.
11. Chiba, H., Osanai, M., Murata, M. et al. Transmembrane proteins of tight
junctions. Biochim Biophys Acta, 2008;1778(3):588–600.
12. Evans, M.J., von Hahn, T., Tscherne, D.M. et al. Claudin‐1 is a hepatitis C virus
co‐receptor required for a late step in entry. Nature, 2007;446(7137):801–5.
13. Guttman, J.A. and Finlay, B.B. Tight junctions as targets of infectious
agents. Biochim Biophys Acta, 2009;1788(4):832–41.
14. Rahner, C., Mitic, L.L., and Anderson, J.M. Heterogeneity in expression and
subcellular localization of claudins 2, 3, 4, and 5 in the rat liver, pancreas,
and gut. Gastroenterology, 2001;120(2):411–22.
15. Yamamoto, T., Kojima, T., Murata, M. et al. p38 MAP‐kinase regulates func­
tion of gap and tight junctions during regeneration of rat hepatocytes.
J Hepatol, 2005;42(5):707–18.
16. Amasheh, S., Meiri, N., Gitter, A.H. et  al. Claudin‐2 expression induces
cation‐selective channels in tight junctions of epithelial cells. J Cell Sci,
2002;115(Pt 24):4969–76.
17. Paul, D.L. Molecular cloning of cDNA for rat liver gap junction protein.
J Cell Biol, 1986;103(1):123–34.
18. Kumar, N.M. and Gilula, N.B. The gap junction communication channel.
Cell, 1996;84(3):381–8.
19. Zhang, J.T. and Nicholson, B.J. Sequence and tissue distribution of a second
protein of hepatic gap junctions, Cx26, as deduced from its cDNA. J Cell
Biol, 1989;109(6 Pt 2):3391–401.
20. Dahl, G. and Muller, K.J. Innexin and pannexin channels and their signaling.
FEBS Lett, 2014;588(8):1396–402.
21. Maes, M., Crespo Yanguas, S., Willebrords, J., Cogliati, B., and Vinken M.
Connexin and pannexin signaling in gastrointestinal and liver disease. Transl
Res, 2015;166(4):332–43.
22. Sosinsky, G.E., Boassa, D., Dermietzel, R. et al. Pannexin channels are not
gap junction hemichannels. Channels (Austin), 2011;5(3):193–7.
172 THE LIVER: REFERENCES
23. Willebrords, J., Maes, M., Crespo Yanguas, S., and Vinken M. Inhibitors of
connexin and pannexin channels as potential therapeutics. Pharmacol Ther,
2017;180:144–60.
24. Stauffer, K.A. The gap junction proteins beta 1‐connexin (connexin‐32) and
beta 2‐connexin (connexin‐26) can form heteromeric hemichannels. J Biol
Chem, 1995;270(12):6768–72.
25. Valiunas, V., Niessen, H., Willecke, K., and Weingart, R. Electrophysiological
properties of gap junction channels in hepatocytes isolated from connexin32‐
deficient and wild‐type mice. Pflugers Arch, 1999;437(6):846–56.
26. Locke, D., Liu, J., and Harris, A.L. Lipid rafts prepared by different methods
contain different connexin channels, but gap junctions are not lipid rafts.
Biochemistry, 2005;44(39):13027–42.
27. Duffy, H.S., Delmar, M., and Spray, D.C. Formation of the gap junction
nexus: binding partners for connexins. J Physiol Paris, 2002;96(3–4):
243–9.
28. Duffy, H.S., Fort, A.G., and Spray, D.C. Cardiac connexins: genes to nexus.
Adv Cardiol, 2006;42:1–17.
29. Duffy, H.S., Iacobas, I., Hotchkiss, K. et al. The gap junction protein con­
nexin32 interacts with the Src homology 3/hook domain of discs large
homolog 1. J Biol Chem, 2007;282(13):9789–96.
30. Kojima, T., Yamamoto, T., Murata, M. et al. Regulation of the blood‐biliary
barrier: interaction between gap and tight junctions in hepatocytes. Med
Electron Microsc, 2003;36(3):157–64.
31. Fowler, S.L., Akins, M., Zhou, H., Figeys, D., and Bennett, S.A. The liver
connexin32 interactome is a novel plasma membrane‐mitochondrial
­signaling nexus. J Proteome Res, 2013;12(6):2597–610.
32. Graf, J. and Boyer, J.L. The use of isolated rat hepatocyte couplets in
­hepatobiliary physiology. J Hepatol, 1990;10(3):387–94.
33. Gunzel, D., Zakrzewski, S.S., Schmid, T. et al. From TER to trans‐ and para­
cellular resistance: lessons from impedance spectroscopy. Ann N Y Acad Sci,
2012;1257:142–51.
34. Suzuki, H., Nishizawa, T., Tani, K. et al. Crystal structure of a claudin pro­
vides insight into the architecture of tight junctions. Science, 2014;344(6181):
304–7.
35. Matsumoto, K., Imasato, M., Yamazaki, Y. et  al. Claudin 2 deficiency
reduces bile flow and increases susceptibility to cholesterol gallstone disease
in mice. Gastroenterology, 2014;147(5):1134–45 e10.
36. Buckley, A. and Turner, J.R. Cell biology of tight junction barrier regulation
and mucosal disease. Cold Spring Harb Perspect Biol, 2018;10(1).
37. Fasano, A., Fiorentini, C., Donelli, G. et  al. Zonula occludens toxin
­modulates tight junctions through protein kinase C‐dependent actin reor­
ganization, in vitro. J Clin Invest, 1995;96(2):710–20.
38. Tsukamoto, T. and Nigam, S.K. Tight junction proteins form large com­
plexes and associate with the cytoskeleton in an ATP depletion model for
reversible junction assembly. J Biol Chem, 1997;272(26):16133–9.
39. Turner, J.R., Angle, J.M., Black, E.D. et  al. PKC‐dependent regulation of
transepithelial resistance: roles of MLC and MLC kinase. Am J Physiol,
1999;277(3 Pt 1):C554–62.
40. Pradhan‐Sundd, T., Vats, R., Russell, J.M. et al. Dysregulated bile transport­
ers and impaired tight junctions during chronic liver injury in mice.
Gastroenterology, 2018;155(4):1218–32.e24.
41. Gissen, P. and Arias, I.M. Structural and functional hepatocyte polarity and
liver disease. J Hepatol, 2015;63(4):1023–37.
42. Mailly, L., Xiao, F., Lupberger, J. et al. Clearance of persistent hepatitis C
virus infection in humanized mice using a claudin‐1‐targeting monoclonal
antibody. Nat Biotechnol, 2015;33(5):549–54.
43. Crespo Yanguas, S., Willebrords, J., Maes, M. et al. Connexins and pannex­
ins in liver damage. EXCLI J, 2016;15:177–86.
44. Gumucio, J.J. and Miller, D.L. Functional implications of liver cell heteroge­
neity. Gastroenterology, 1981;80(2):393–403.
45. Jungermann, K. and Katz N. Functional specialization of different hepato­
cyte populations. Physiol Rev, 1989;69(3):708–64.
46. Berthoud, V.M., Iwanij, V., Garcia, A.M., and Saez, J.C. Connexins and
glucagon receptors during development of rat hepatic acinus. Am J Physiol,
1992;263(5 Pt 1):G650–8.
47. Kinugasa, A. and Thurman, R.G. Differential effect of glucagon on gluco­
neogenesis in periportal and pericentral regions of the liver lobule. Biochem
J, 1986;236(2):425–30.
48. Lee, S.M. and Clemens, M.G. Subacinar distribution of hepatocyte mem­
brane potential response to stimulation of gluconeogenesis. Am J Physiol.,
1992;263(3 Pt 1):G319–26.
49. Seseke, F.G., Gardemann, A., and Jungermann K. Signal propagation via gap
junctions, a key step in the regulation of liver metabolism by the sympathetic
hepatic nerves. FEBS Lett, 1992;301(3):265–70.
50. Saez, J.C., Connor, J.A., Spray, D.C., and Bennett, M.V. Hepatocyte gap
junctions are permeable to the second messenger, inositol 1,4,5‐trisphos­
phate, and to calcium ions. Proc Natl Acad Sci U S A, 1989;86(8):2708–12.
51. Nathanson, M.H. and Burgstahler, A.D. Coordination of hormone‐induced
calcium signals in isolated rat hepatocyte couplets: demonstration with con­
focal microscopy. Mol Biol Cell, 1992;3(1):113–21.
52. Clapham, D.E. Calcium signaling. Cell, 2007;131(6):1047–58.
53. Bartlett, P.J., Gaspers, L.D., Pierobon, N., and Thomas, A.P. Calcium‐
dependent regulation of glucose homeostasis in the liver. Cell Calcium,
2014;55(6):306–16.
54. Guerra, M.T. and Nathanson, M.H. Calcium signaling and secretion in chol­
angiocytes. Pancreatology, 2015;15(4 Suppl):S44–8.
55. Scemes, E. and Spray, D.C. Extracellular K+ and astrocyte signaling via con­
nexin and pannexin channels. Neurochem Res, 2012;37(11):2310–16.
56. Schlosser, S.F., Burgstahler, A.D., and Nathanson, M.H. Isolated rat hepato­
cytes can signal to other hepatocytes and bile duct cells by release of nucleo­
tides. Proc Natl Acad Sci U S A, 1996;93(18):9948–53.
57. Aasen, T., Mesnil, M., Naus, C.C., Lampe, P.D., and Laird, D.W. Gap junc­
tions and cancer: communicating for 50 years. Nat Rev Cancer,
2016;16(12):775–88.
58. Yamasaki, H., Krutovskikh, V., Mesnil, M. et al. Role of connexin (gap junc­
tion) genes in cell growth control and carcinogenesis. C R Acad Sci III,
1999;322(2–3):151–9.
59. Saez, J.C., Gregory, W.A., Watanabe, T. et al. cAMP delays disappearance of
gap junctions between pairs of rat hepatocytes in primary culture. Am J
Physiol, 1989;257(1 Pt 1):C1–11.
60. Spray, D.C., Rozental, R., and Srinivas, M. Prospects for rational develop­
ment of pharmacological gap junction channel blockers. Curr Drug Targets,
2002;3(6):455–64.
61. Saez, J.C., Nairn, A.C., Czernik, A.J. et al. Phosphorylation of connexin 32,
a hepatocyte gap‐junction protein, by cAMP‐dependent protein kinase, pro­
tein kinase C and Ca2+/calmodulin‐dependent protein kinase II. Eur J
Biochem, 1990;192(2):263–73.
62. Koval, M., Isakson, B.E., and Gourdie, R.G. Connexins, pannexins and
innexins: protein cousins with overlapping functions. FEBS Lett,
2014;588(8):1185.
63. Epifantseva, I. and Shaw, R.M. Intracellular trafficking pathways of Cx43
gap junction channels. Biochim Biophys Acta, 2018;1860(1):40–7.
64. Fort, A.G., Murray, J.W., Dandachi, N. et al. In vitro motility of liver con­
nexin vesicles along microtubules utilizes kinesin motors. J Biol Chem,
2011;286(26):22875–85.
65. Francis, R., Xu, X., Park, H. et al. Connexin43 modulates cell polarity and
directional cell migration by regulating microtubule dynamics. PLoS One,
2011;6(10):e26379.
66. Polusani, S.R., Kalmykov, E.A., Chandrasekhar, A., Zucker, S.N., and
Nicholson, B.J. Cell coupling mediated by connexin 26 selectively con­
tributes to reduced adhesivity and increased migration. J Cell Sci,
2016;129(23):4399–410.
67. Matsuuchi, L. and Naus, C.C. Gap junction proteins on the move: connexins,
the cytoskeleton and migration. Biochim Biophys Acta, 2013;1828(1):
94–108.
68. Falk, M.M., Bell, C.L., Kells Andrews, R.M., and Murray, S.A. Molecular
mechanisms regulating formation, trafficking and processing of annular gap
junctions. BMC Cell Biol, 2016;17(Suppl 1):22.
69. Gaietta, G., Deerinck, T.J., Adams, S.R. et al. Multicolor and electron micro­
scopic imaging of connexin trafficking. Science, 2002;296(5567):503–7.
70. Bejarano, E., Yuste, A., Patel, B. et al. Connexins modulate autophagosome
biogenesis. Nat Cell Biol, 2014;16(5):401–14.
71. Laird, D.W. Syndromic and non‐syndromic disease‐linked Cx43 mutations.
FEBS Lett, 2014;588(8):1339–48.
72. Scherer, S.S. and Kleopa, K.A. X‐linked Charcot‐Marie‐Tooth disease.
J Peripher Nerv Syst, 2012;17(Suppl 3):9–13.
73. Nelles, E., Butzler, C., Jung, D. et al. Defective propagation of signals gener­
ated by sympathetic nerve stimulation in the liver of connexin32‐deficient
mice. Proc Natl Acad Sci U S A, 1996;93(18):9565–70.
74. Stumpel, F., Ott, T., Willecke, K., and Jungermann K. Connexin 32 gap junc­
tions enhance stimulation of glucose output by glucagon and noradrenaline
in mouse liver. Hepatology, 1998;28(6):1616–20.
 14:  Gap and Tight Junctions in Liver: Structure, Function, and Pathology 173
75. Niessen, H. and Willecke K. Strongly decreased gap junctional permeability
to inositol 1,4, 5‐trisphosphate in connexin32 deficient hepatocytes. FEBS
Lett, 2000;466(1):112–14.
76. Schwarz, M., Wanke, I., Wulbrand, U., Moennikes, O., and Buchmann A.
Role of connexin32 and beta‐catenin in tumor promotion in mouse liver.
Toxicol Pathol, 2003;31(1):99–102.
77. King, T.J. and Lampe, P.D. Altered tumor biology and tumorigenesis in
irradiated and chemical carcinogen‐treated single and combined con­
nexin32/p27Kip1‐deficient mice. Cell Commun Adhes, 2005;12(5–6):
293–305.
78. Kojima, T., Fort, A., Tao, M., Yamamoto, M., and Spray, D.C. Gap junction
expression and cell proliferation in differentiating cultures of Cx43 KO
mouse hepatocytes. Am J Physiol Gastrointest Liver Physiol, 2001;281(4):
G1004–13.
79. Temme, A., Ott, T., Dombrowski, F., and Willecke, K. The extent of synchro­
nous initiation and termination of DNA synthesis in regenerating mouse liver
is dependent on connexin32 expressing gap junctions. J Hepatol, 2000;32(4):
627–35.
80. Dagli, M.L., Yamasaki, H., Krutovskikh, V., and Omori, Y. Delayed liver
regeneration and increased susceptibility to chemical hepatocarcinogenesis
in transgenic mice expressing a dominant‐negative mutant of connexin32
only in the liver. Carcinogenesis, 2004;25(4):483–92.
81. Igarashi, I., Makino, T., Suzuki, Y. et  al. Background lesions during a
­24‐month observation period in connexin 32‐deficient mice. J Vet Med
Sci, 2013;75(2):207–10.
82. Li, Q., Omori, Y., Nishikawa, Y. et  al. Cytoplasmic accumulation of con­
nexin32 protein enhances motility and metastatic ability of human hepatoma
cells in vitro and in vivo. Int J Cancer, 2007;121(3):536–46.
83. Spray, D.C., Hanstein, R., Lopez‐Quintero, S.V. et  al. Gap junctions and
bystander effects: Good Samaritans and executioners. Wiley Interdiscip Rev
Membr Transp Signal, 2013;2(1):1–15.
84. Zibara, K., Awada, Z., Dib, L. et al. Anti‐angiogenesis therapy and gap junc­
tion inhibition reduce MDA‐MB‐231 breast cancer cell invasion and metas­
tasis in vitro and in vivo. Sci Rep, 2015;5:12598.
85. Crespo Yanguas, S., da Silva, T.C., Pereira, I.V.A. et al. TAT‐Gap19 and car­
benoxolone alleviate liver fibrosis in mice. Int J Mol Sci, 2018;19(3).
86. Iorga, A., Dara, L., and Kaplowitz, N. Drug‐induced liver injury: cascade of
events leading to cell death, apoptosis or necrosis. Int J Mol Sci, 2017;18(5).
87. Asamoto, M., Hokaiwado, N., Murasaki, T., and Shirai, T. Connexin 32
dominant‐negative mutant transgenic rats are resistant to hepatic damage by
chemicals. Hepatology, 2004;40(1):205–10.
88. Naiki‐Ito, A., Asamoto, M., Naiki, T. et al. Gap junction dysfunction reduces
acetaminophen hepatotoxicity with impact on apoptotic signaling and con­
nexin 43 protein induction in rat. Toxicol Pathol, 2010;38(2):280–6.
89. Patel, S.J., Milwid, J.M., King, K.R. et al. Gap junction inhibition prevents
drug‐induced liver toxicity and fulminant hepatic failure. Nat Biotechnol,
2012;30(2):179–83.
90. Maes, M., McGill, M.R., da Silva, T.C. et al. Connexin32: a mediator of aceta­
minophen‐induced liver injury? Toxicol Mech Methods, 2016;26(2):88–96.
91. Maes, M., Crespo Yanguas, S., Willebrords, J. et al. Connexin hemichannel
inhibition reduces acetaminophen‐induced liver injury in mice. Toxicol Lett,
2017;278:30–7.
92. Maes, M. and Vinken M. Connexin‐based signaling and drug‐induced hepa­
totoxicity. J Clin Transl Res, 2017;3(Suppl 1):189–98.
93. Lee, N.P. The blood‐biliary barrier, tight junctions and human liver diseases.
Adv Exp Med Biol, 2012;763:171–85.
 Ribosome Biogenesis and its
15 Role in Cell Growth and
Proliferation in the Liver
Katherine I. Farley‐Barnes1 and Susan J. Baserga1,2,3
1
Department of Molecular Biophysics & Biochemistry, Yale University School of Medicine, New Haven, CT, USA
2
Department of Genetics, Yale University School of Medicine, New Haven, CT, USA
3
Department of Therapeutic Radiology, Yale University School of Medicine, New Haven, CT, USA

INTRODUCTION Saccharomyces cerevisiae. Only recently have scientists made

Ribosomes, the cellular machinery responsible for all protein
synthesis, are essential for cells to grow and to divide. The bio-
genesis of ribosomes is thus a highly regulated process that
coordinates a number of cellular cues and depends especially
upon nutrient availability. Nowhere is the relationship between
nutrient availability, cellular signaling, and ribosome biogenesis
more critical than in the liver. Liver size decreases dramatically
after fasting in rats and mice [1, 2]. When refed, however, the
liver size recovers within 24 hours [1, 2]. Total liver proteins and
RNA follow the same fluctuation, indicating that the liver
responds to limited nutrients by downregulating ribosome pro-
duction and upregulating it again once the organism is presented
with an adequate diet [1–3].
Understanding ribosome biogenesis is not only essential to
understanding basic cellular biology, but also to probing the
pathogenesis of a number of human diseases. Such diseases
include cancer, where an increase in cell growth and prolifera-
tion goes hand‐in‐hand with an increase in ribosome biogenesis
(reviewed in [4]). Ribosome dysfunction has also been
implicated in neurodegenerative diseases (reviewed in [5]).
­ by RNA polymerase I (RNAPI). As the rDNA is being tran-
Additionally, aberrant ribosome biogenesis causes a number of
tissue‐specific disorders, termed ribosomopathies (reviewed in
[6, 7]). It is intriguing that dysregulation of such an essential
process like ribosome biogenesis would result in a viable organ-
ism. Indeed, how changes in ribosome biogenesis, a process that
is required for every cell type, would create tissue‐specific
pathologies is an outstanding question in the field.
Much of the current knowledge of the steps in ribo-
some ­biogenesis comes from studies in the budding yeast,
advances into understanding how this process works in humans.
Additionally, how ribosome biogenesis changes among differ-
ent tissues is largely unknown. One of the best models for
understanding ribosome biogenesis at the tissue level, however,
has been the liver. Foundational studies on the liver have pro-
vided insights into how ribosomes are made and how the cell
responds to various stimuli to regulate the production of ribo-
somes. Indeed, nucleoli, the non‐membrane‐bound organelles
responsible for ribosome production, were first biochemically
isolated from liver cells in 1956 [8]. It is likely that in the future,
the liver will continue to play a large role in our understanding
of ribosome biogenesis and its relation to human disease.
OVERVIEW OF RIBOSOME BIOGENESIS
Ribosome biogenesis begins in a non‐membrane‐bound orga-
nelle inside the cell nucleus called the nucleolus. It starts with
transcription of the tandemly repeated ribosomal DNA (rDNA)
scribed, the nucleolus forms around it. Thus, the repeats of
rDNA on the chromosomes are called nucleolar organizing
regions, or NORs. The NORs are located on five different chro-
mosomes in humans (13, 14, 15, 21, and 22). Therefore, human
cells have the potential to form 10 nucleoli per cell, although far
fewer nucleoli are often observed [9]. The number of nucleoli in
a diploid mouse hepatocyte ranges from 3 to 6 (mice have the
potential to form up to 12 nucleoli per cell), with 3 nucleoli per
nucleus being seen most often [10].

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
 15:  Ribosome Biogenesis and its Role in Cell Growth and Proliferation in the Liver 175
The human nucleolus consists of three different compart-
ments: the fibrillar center (FC), the dense fibrillar component
(DFC), and the granular component (GC) (Figure 15.1). A distinct
function in making ribosomes can be ascribed to each of these
compartments. The nucleolus forms around the FC, and at the
interface of the FC and the DFC, where rDNA transcription
takes place. After the transcription of the rDNA, modification of
the pre‐ribosomal RNA (pre‐rRNA), processing of the pre‐
rRNA, and pre‐ribosome assembly proceed outwards from the
DFC into the third nucleolar compartment, the GC. In the GC,
ribosomal proteins (r‐proteins) assemble with the pre‐rRNA to
form the pre‐small subunit (pre‐SSU) and pre‐large subunit
(pre‐LSU) of the ribosome [11].
One explanation for the tripartite organization of the nucleo-
lus uses the physics of liquid–liquid phase separations. Just like
an oil droplet in water, the proteins of the nucleolus separate
into multiple compartments due to differences in surface tension
and hydrophobicity. For example, a structure similar to the
DFC/GC compartments can be replicated in vitro by mixing
fibrillarin (FBL), a box C/D small nucleolar ribonucleoprotein
(snoRNP) methyltransferase enzyme that is enriched in the
DFC, and nucleophosmin (NPM1), a protein with a number of
functions that is enriched in the GC [12].
At the FC/DFC interface, the rDNA is transcribed as a 47S
polycistronic precursor that contains 3 of the 4 the ribosomal
RNAs (rRNAs) which eventually are incorporated into the small
(18S) and large (5.8S and 28S) subunits of the ribosome. To
form the mature rRNAs, the pre‐rRNA must be modified,
cleaved, and processed. Two types of RNA modifications are
primarily used: 2′‐O‐methylation and pseudouridylation
(reviewed in [13]). Generally, the modifications are placed at
functionally important regions of the ribosome, such as the
tRNA‐binding sites or at the interface of the small and large
subunits (reviewed in [13]). These modifications aid in folding
the pre‐rRNA into the correct secondary and tertiary structures.
Most of these modifications are made by snoRNPs which direct
their guide RNA to base pair with the corresponding site to be
modified and position the enzyme to perform the modification.
The majority of pre‐rRNA modifications are carried out by two
types of snoRNPs: box C/D, which perform the 2′‐O‐methylations,
and box H/ACA, which perform the pseudouridylations, both
named for conserved snoRNA sequences.
FC
DFC
GC
Figure 15.1  The morphology of the nucleolus in mammalian cells.
On the left is the nucleolus of a cell actively engaged in ribosome bio-
genesis. On the right is the nucleolus of a cell undergoing inhibition of
rRNA transcription. Indicated are the three nucleolar compartments:
the fibrillar center (FC), the dense fibrillar component (DFC), and the
granular component (GC). See text for details.
In addition to the modifications made to the pre‐rRNA, the pre‐
rRNA is also extensively cleaved and processed. In humans, pre‐
rRNA processing occurs through multiple pathways (Figure 15.2).
The difference between the two major processing pathways
begins with the choice of cleavage at either site 1 or site 2. If site
1 is cleaved first, the 41S pre‐rRNA is made first until a later
cleavage at site 2 separates the 21S and 32S pre‐rRNAs. If site 2
is cleaved first, the small and large subunit pre‐rRNAs are sepa-
rated into the 30S and 32S pre‐rRNAs, respectively (Figure 15.2,
and reviewed in [14]). Regardless of the pathway used, the end
result is the mature 18S, 5.8S, and 28S rRNAs.
In addition to the snoRNAs and r‐proteins, a number of trans‐
acting factors are also important for optimal pre‐rRNA process-
ing. Best characterized in yeast, these factors include endo‐ and
exonucleases, helicases, AAA‐ATPases, GTPases, and more [15].
Many of the functions of these enzymes are still being defined. In
yeast, over 200 proteins are involved in ribosome assembly [15],
a number that is likely to be greatly increased in humans. Indeed,
a series of studies are currently striving to define all of the factors
necessary for this process in humans [16–19].
Ribosome assembly concludes in the cytoplasm, after the
addition of the 5S rRNA which is transcribed from chromosome
1 by RNA polymerase III (RNAPIII) in humans. The pre‐SSU
and pre‐LSU are both exported through the nuclear pore com-
plex, and export of each subunit requires common factors as
well as subunit‐specific factors. Exportin 1 (XPO1, or Crm1 in
yeast) is one of the best characterized proteins involved in both
large and small subunit export [20]. Additionally, the final steps
of pre‐rRNA processing occur in the cytoplasm, including the
processing of the 18SE to the mature 18S rRNA at site 3.
Overall, human ribosome biogenesis begins in the nucleolus but
encompasses the nucleus and cytoplasm as well, leaving many
steps in the process open to regulation.
REGULATION OF RIBOSOME BIOGENESIS
Our understanding of the intricacies of the regulation of human
ribosome biogenesis is ever increasing. Regulation begins at the
level of the rDNA with the organization and silencing of select
rDNA repeats. RNAPI transcription of the rDNA is also
­influenced by many cellular signals. Correct processing of the
pre‐rRNA is controlled by a number of molecules that signal to
modulate the transcription of r‐proteins, snoRNPs, and other
assembly and processing factors. Finally, nuclear export of the
pre‐ribosomal subunits can be monitored and regulated by
the cell. In all, ribosome biogenesis requires all three RNA poly-
merases and hundreds of other factors, so there are many ways
for the cell to control steps in this process. Just a few of these
mechanisms are described in detail below.
mTOR and the regulation of ribosomal
protein translation
One key regulator of ribosome biogenesis is the mechanistic tar-
get of rapamycin (mTOR) pathway. Essential for cell growth,
the mTOR pathway responds to various cellular signals to
upregulate the transcription of r‐proteins and to ultimately
176 THE LIVER:  REGULATION OF RIBOSOME BIOGENESIS

A′ A0 1 3 2a 2 4

47S pre-rRNA 5′ 18S 5.8S 28S 3′

5′ETS ITS1 ITS2 3′ETS

A0 1 3 2a 2 4

45S 18S 5.8S 28S

site 2 cleavage site 1 cleavage

A0 1 3 2a 4 3 2a 2 4

30S 18S 32S 5.8S 28S 18S 5.8S 28S 41S

3 2a 3 2a 4

21S 18S 21S 18S 5.8S 28S 32S

18SE 18S 12S 5.8S 18SE 18S 12S 5.8S

18S 5.8S 28S 18S 5.8S 28S

Figure 15.2  Schematic depicting the two major pre‐rRNA processing pathways in human cells. The 47S pre‐rRNA is processed and cleaved
through multiple pathways to form the mature 18S, 5.8S, and 28S rRNAs that are incorporated into the synthesized ribosome. ETS and ITS
­represent external transcribed spacers and internal transcribed spacers, respectively.

increase global protein synthesis. There are two different
mTOR‐containing cytoplasmic complexes, termed mTORC1
and mTORC2. The mTORC1 complex contains mTOR, RPTOR
(regulatory associated protein of MTOR complex 1), AKT1S1
(AKT substrate 1), DEPTOR (DEP domain containing mTOR‐
interacting protein), and MLST8 (mTOR‐associated protein,
LST8 homolog). mTORC2 contains mTOR, RICTOR (RPTOR‐
independent companion of MTOR complex 2), PRR5 (proline‐
rich 5), MAPKAP1 (mitogen‐activated protein kinase‐associated
protein 1), DEPTOR, and MLST8. Although mTORC1 and
mTORC2 have different complex members and substrates,
mTOR functions as a serine/threonine kinase in both complexes.
Importantly for ribosome biogenesis, mTORC1 signals to RPS6
kinases 1 and 2 (S6K1 and S6K2). S6K1 phosphorylates several
substrates, ultimately promoting increased translation initiation
and elongation, as well as increased ribosome biogenesis
through the upregulation of rDNA transcription (reviewed in
[21]). mTORC1 also phosphorylates eukaryotic translation ini-
tiation factor 4B pseudogene 1 (4EBP1). 4EBP1 affects cap‐
dependent translation, and its phosphorylation prevents its
binding to eukaryotic translation initiation factor 4E (EIF4E).
This leaves EIF4E free to bind eukaryotic initiation factor 4
gamma proteins (EIF4Gs), which then bind 5′ caps of mRNAs
and recruit other factors required for translation (reviewed in
[22]). Of note, some of mTORC1’s function is inhibited by the
immunosuppressant rapamycin [23].
mTOR plays an important role in ribosome biogenesis
through control of the translation of mRNAs encoding
r‐proteins. In higher eukaryotes, mRNAs that code for r‐proteins
are denoted by a 5′ terminal oligopyrimidine (5′TOP) motif
[24]. This motif begins with a C and contains, on average, 12.2
nucleotides in humans [25]. When cells are stimulated with
nutrients, 5′TOP mRNAs are recruited to ribosomes for
increased translation [26]. Rapamycin treatment prevents the
translation of 5′TOP mRNAs, and mutation of the 5′TOP
­renders the mRNA’s translation insensitive to rapamycin treat-
ment [27–29]. More recent studies have shown that mTORC1
affects 5′TOP translation through its phosphorylation of
EIF4E‐binding proteins (4E‐BPs) [30]. Phosphorylation by
active mTORC1 allows EIF4E to bind EIF4G1 [30]. 5′TOP
mRNAs require this EIF4E–EIF4G1 interaction for their cap
binding and translation, while other mRNAs do not, providing
an explanation for mTOR’s specific role in 5′TOP translation
after nutrient stimulation [30].
Regarding feeding and nutrient stimulation, the liver relies
heavily on mTORC1 signaling. Nutrient uptake, as well as
uptake of amino acids such as leucine, stimulates ribosome
­biogenesis in the liver [31]. In the livers of rats administered
leucine, mTOR signaling is upregulated and translation of
5′TOP‐containing mRNAs is increased [32]. This response can
be inhibited using the drug rapamycin [32]. Translation of
ribosomal protein mRNAs is also upregulated during liver
­
 15:  Ribosome Biogenesis and its Role in Cell Growth and Proliferation in the Liver 177
regeneration in rats, although the precise mechanism of action
is unknown [33, 34].
Additionally, in the liver, mTOR signaling acts as an impor-
tant regulatory mechanism in the body’s response to ethanol.
Chen et  al. found decreased levels of the protein DEPTOR,
a  known inhibitor of mTORC1 phosphorylation, in both a
chronic‐plus‐binge ethanol‐fed mouse model of alcoholic stea-
tosis and in patients with alcoholic liver disease [35]. Decreased
DEPTOR leads to increased levels of 4EBP1 and S6K phospho-
rylation [35]. Phosphorylated S6K in turn phosphorylates sterol
regulator element‐binding transcription factor 1 (SREBF1),
which translocates to the nucleus and upregulates the expres-
sion of proteins involved in fatty acid biosynthesis [36, 37].
Phosphorylated mTORC1 can also phosphorylate lipin 1, pro-
moting the transcription of SREBF1 which contributes to the
build up of lipids in alcoholic steatosis [38]. Overall, mTOR
signaling highlights the importance of regulating ribosome
­biogenesis for proper liver health.
Regulation of RNAPI transcription
RNAPI transcriptional regulation is critical in the liver, since the
liver responds so dramatically to nutrient availability. Upon
feeding, it has been shown that RNAPI activity quickly increases
[2]. Similarly, RNAPI activity increases during liver regenera-
tion, in addition to the increased r‐protein translation previously
mentioned (cited many times, including in [39, 40]).
The tandemly repeated rDNA is transcribed by RNAPI as a
large (47S) polycistronic precursor that contains the 18S, 5.8S,
and 28S mature rRNAs. Only about half of the rDNA repeats
are active at any given time, and RNAPI transcription is tightly
regulated to ensure generation of an adequate number of ribo-
somes for cell growth or maintenance. RNAPI transcription ini-
tiation begins with the binding of upstream binding transcription
factor (UBTF, commonly referred to as UBF) to the upstream
control element (UCE) (Figure 15.3). UBF binding then helps to
recruit selectivity factor 1 (SL1), a species‐specific complex
containing the TATA‐binding protein (TBP) and three TBP‐
associated ­factors (TAFs) in humans [41]. The pre‐initiation
complex is formed when SL1 binds to the core promoter (CP) of
rRNA genes along with RNAPI. The interaction between SL1
and RNAPI is mediated by TIF‐1A (Rrn3 in S. cerevisiae) [42].
Multiple signaling pathways regulate rDNA transcription.
Among these is again mTOR signaling, which alters RNAPI
transcription through multiple mechanisms. These mechanisms
include S6K phosphorylation which in turn phosphorylates the
C‐terminal tail of UBF, aiding its recruitment of SL1 and the
rest of the RNAPI transcription machinery [43]. mTOR also
works to activate TIF‐1A by promoting phosphorylation of
SL1/TIF-IB PoI I
UBF TIF-IA A43
UCE CP
Figure 15.3  Schematic representation of the transcription initiation
complex at the promoter of the rRNA gene. See text for details.
residue S44 and decreasing the amount of S199 phosphorylation,
ultimately increasing TIF‐1A’s interaction with SL1 [44].
Additionally, mTOR may bind directly to the RNAPI
promoter [45].
The MAPK signaling pathway also has been shown to play a
role in rDNA transcriptional regulation through multiple mech-
anisms. In the MAPK pathways, extracellular stimulation by
factors such as epidermal growth factor (EGF) triggers a cas-
cade of protein kinases which in turn promote factors necessary
for growth and development. Like mTOR signaling, MAPK
phosphorylation of two residues on TIF‐1A increases RNAPI
initiation [46]. Additionally, UBF is phosphorylated via the
MAPK signaling cascade. UBF phosphorylation may then
enhance UBF’s interaction with SL1 and/or disrupt formation of
the enhanceosomes that are promoted by UBF and slow down
RNAPI [43, 47]. Both options increase RNAPI activity.
Signaling through the retinoblastoma protein (pRB), a pocket
family protein and tumor suppressor, can also affect ribosome
biogenesis via modulation of RNAPI transcription. This occurs
both directly through the RNAPI transcription machinery and
indirectly through the interaction of pRB with other factors.
pRB, as well as the pocket family protein p130, directly affects
rDNA transcription by disrupting the interaction between SL1
and UBF [48, 49]. Indirectly, pRB modulates rDNA transcrip-
tion through its interactions with E2F1. When pRB is hyper-
phosphorylated, it cannot bind to E2F transcription factors.
These E2Fs are then free to activate transcription of several
­target genes that move the cell cycle successfully from G1 to S
phase. Conversely, when the cell is stressed, pRB becomes
hypophosphorylated, is able to bind and inhibit E2Fs, and the
cell cycle is stopped, ultimately feeding back on ribosome bio-
genesis (reviewed in [50]). Additionally, at low expression of
E2F1, E2F1 is able to bind directly to the rDNA promoter, acti-
vating RNAPI transcription [51]. At high E2F1 levels, p14ARF is
expressed, and p14ARF binds to E2F1 and inhibits its transcrip-
tional activity so that rDNA transcription is decreased [52].
In  turn, decreased rDNA transcription feeds back on E2F1
expression to reduce it [53].
Regulation of ribosome biogenesis by MYC
MYC acts as a transcription factor to regulate expression of
numerous proteins required for ribosome biogenesis. It has been
studied regarding hepatocyte proliferation and growth, but
MYC also plays a role in many other tissues and is especially
significant in cancer [54]. While the MYC transcription factor
family contains multiple MYC proteins (including c‐MYC,
n‐MYC, and l‐MYC), c‐MYC is the best studied with regards to
ribosome biogenesis. c‐MYC modulates the transcription medi-
ated by all three RNA polymerases. To enhance RNAPI activity,
c‐MYC can regulate UBF and TIF‐1A expression [55, 56].
c‐MYC is also located in the nucleolus and can directly a­ ssociate
with SL1 to stimulate RNAPI transcription [57, 58]. Additionally,
c‐MYC and n‐MYC regulate RNAPII transcription of r‐proteins
[59–61]. c‐MYC regulates transcription of the 5S rRNA as well
by interacting with the RNAPIII transcription machinery [62].
MYC’s role in ribosome biogenesis is critical for proper liver
function. Overexpression of c‐MYC in hepatocytes results in
enlarged nucleoli as well as increased expression of several
178 THE LIVER:  REGULATION OF RIBOSOME BIOGENESIS
proteins required for ribosome biogenesis [63]. Additionally,
c‐MYC plays a role in several liver diseases [54]. For example,
c‐MYC expression is upregulated in the livers of patients with
alcoholic liver disease (ALD) and after ethanol feeding in a
mouse model that mimics the early stages of ALD [64].
Interestingly, combined c‐MYC overexpression and ethanol
feeding results in decreased p53 levels [64], indicating a possi-
ble mechanism for the liver dysplasia observed in ALD patients.
Nucleolar stress regulated through p53
Regulation of ribosome biogenesis is also controlled by the
tumor suppressor p53 in a process known as nucleolar stress
(reviewed in [65, 66]). In a normally functioning cell, r‐proteins
are engaged in the process of translation as structural compo-
nents of ribosomes. In this scenario, MDM2 ubiquitinates the
p53 protein, targeting p53 for degradation (Figure  15.4).
However, when there are perturbations in ribosome biogenesis,
r‐proteins become disengaged. Two of these r‐proteins, uL18
(RPL5) and uL5 (RPL11), along with the 5S rRNA, form the 5S
ribonucleoprotein complex (5S RNP). The 5S RNP then
binds to MDM2, blocking MDM2’s ability to ubiquitinate p53,
­ultimately resulting in stabilization of p53 levels (Figure 15.4)
[67, 68]. p53 stabilization causes cellular senescence and
­apoptosis through regulating the transcription of several known
proteins [69].
p53 also plays a key role in a number of liver diseases
(reviewed in [70]). Notably, p53 is important in liver regenera-
tion, where it must first be stabilized after injury to remove the
affected tissue, and then downregulated to allow for prolifera-
tion of new tissue [71]. p53 stabilization also occurs in the liv-
ers  of mice during nutrient deprivation [72]. While scientists
have proposed a 5′ adenosine monophosphate kinase (AMPK)‐
dependent mechanism of p53 stabilization, AMPK‐depleted
Normal growing cell
MDM2
Functional ribosomes
Nucleolar stress
uL18
uL18 p14ARF
5S rRNA uL5
5S rRNA
uL5
uL18
Free ribosomal proteins 5S rRNA
uL5
Figure 15.4  The nucleolar stress response in human cells. During normal growth conditions, MDM2 ubiquitinates p53, marking it for degradation
and allowing cell growth and proliferation (top). Under conditions of nucleolar stress, r‐proteins are no longer engaged in functional ribosomes. The
5S RNP containing uL18, uL5, and the 5S rRNA binds MDM2, either with or without p14ARF, to prevent MDM2 from ubiquitinating p53. p53 levels
are therefore stabilized, triggering cell cycle arrest and apoptosis (bottom). Reproduced from [65] with permission of Portland Press Limited.
cells are still able to stabilize p53, leaving open the possibility
of an additional aspect of the nucleolar stress response in the
liver during starvation [72].
Regulation of ribosome biogenesis is
coordinated with circadian rhythms
In mice, humans, and birds, it has been shown that liver size
fluctuates according to daily rhythms [73–76]. These daily
rhythms, governed by the organism’s internal circadian clock,
control a number of cellular cues that result in liver size changes.
For example, mice generally eat at night, during their most
active phase. Throughout the night, therefore, mouse liver size
increases and, conversely, liver size decreases during the day
when mice are not eating (resting phase) [75]. Interestingly,
findings by Sinturel et al. have shown that when an organism
eats is crucial for regulating this liver mass cycle, since mice fed
during the day (not during their normal feeding time at night) do
not undergo these changes [75].
At the biochemical level, liver size fluctuations correlate with
the changes in protein translation controlled by the number of
ribosomes present in the hepatocytes (i.e. more ribosomes leads
to larger cell size and thus larger liver mass) [75]. Interestingly,
the increased number of ribosomes is not due to the production
of more ribosomes by increased RNAPI transcription, but rather
to the decreased ability to eliminate excess ribosomes. During
the active phase of the mice, more small subunit r‐proteins
are  translated, likely through increased mTOR signaling as
­discussed above [75, 77–79]. During the resting phase, these
r‐proteins are not made, resulting in excess pre‐18S rRNA.
To reduce the number of ribosomes in the cell, pre‐18S rRNAs
are polyadenylated [80]. In the resting liver cells, this polyade-
nylation then targets these RNAs for degradation by the
Ub Cell growth and
proliferation
p53 (steady state
p53 levels)
MDM2
p14ARF
Cell cycle arrest
p53 and apoptosis
(p53 levels
increased)
MDM2
 15:  Ribosome Biogenesis and its Role in Cell Growth and Proliferation in the Liver 179
catalytic component of the exosome, the 3′‐to‐5′ exonuclease
EXOSC10, and fewer ribosomes are produced overall [75, 80].
This strategy for controlling ribosome biogenesis may be espe-
cially useful in understanding the tissue‐specific intricacies of
the liver diseases discussed below.
ROLE OF RIBOSOME BIOGENESIS
IN LIVER DISEASE
North American Indian childhood cirrhosis
North American Indian childhood cirrhosis (NAIC, OMIM
604901) is a liver‐specific ribosomopathy that results from the
mutation of the ribosome biogenesis factor UTP4 (previously
Cirhin) [81, 82]. Identified in the Ojibway‐Cree First Nations
children from Quebec, Canada, NAIC is characterized by
­transient neonatal jaundice that progresses to biliary cirrhosis
[81, 83]. Thus far, the only effective treatment for this disease is
liver transplantation [83].
In this autosomal recessive disorder, all patients have a
homozygous missense mutation on chromosome 16 (16q22) in
the UTP4 gene (R565W) [81, 84]. The UTP4 protein is a mem-
ber of the t‐Utp subcomplex, which is required for pre‐rRNA
transcription and processing [85, 86]. Specifically, human UTP4
is required for small subunit pre‐rRNA processing at sites A′,
A0, 1, and 2a (Figure 15.2) [87]. It is intriguing, therefore, that
a defect in the function of this ubiquitous and essential protein
leads to liver‐specific pathology.
Few models have been developed for studying NAIC. In a
zebrafish model of NAIC, no pre‐rRNA processing defects were
observed in zebrafish depleted of UTP4 using a morpholino tar-
geting either the start site or the splice acceptor site between
exons 14 and 15 [88], despite the established role of UTP4 in
yeast and human pre‐rRNA processing [85–87]. However, mod-
est increases in p53 levels were observed in the UTP4 zebrafish
model, and the biliary defects seen were abrogated in a p53‐
mutated background [88]. While NAIC is the only disorder of
ribosome biogenesis, or ribosomopathy, known to affect liver
function specifically, this mechanism of p53 stabilization also
occurs in other ribosomopathies [65]. It has also been reported
that the Utp4 homozygous knockout mouse is embryonic lethal,
while the heterozygous mouse develops normally [89], although
these results have not been fully described. Further definition of
the role of UTP4 in liver pathogenesis is therefore needed.
Shwachman–Diamond syndrome
Shwachman–Diamond syndrome (SDS, OMIM 260400) is a
ribosomopathy that usually presents in early childhood with
bone marrow failure (neutropenia) and/or pancreatic insuffi-
ciency manifested as failure to thrive [90, 91]. Ninety percent of
patients have mutations in the Shwachman–Bodian–Diamond
syndrome gene (SBDS; SDO1 in yeast), though recently muta-
tions have also been found in DNAC21 and EFL1 [90, 92]. All
of the proteins encoded by these genes participate in the late
steps in cytoplasmic maturation of the LSU (60S) subunit of the
ribosome. SDS can progress to myelodysplastic syndrome and
acute myelogenous leukemia with increasing patient age
[90, 91, 93]. The median survival of a cohort of SDS patients
was recently found to be only 41 years [93].
SDS patients often have liver abnormalities such as hepato-
megaly, elevated transaminases and mildly elevated bile acid
levels, particularly early in childhood [94]. Most often the
transaminase levels normalize after infancy though the bile acid
levels often remain elevated. In some patients over the age of 30,
liver microcysts, but not fatty liver or cirrhosis, can be visual-
ized by magnetic resonance imaging. Taken together, these find-
ings suggest a connection between the liver and SDS. This is a
ripe topic for further follow‐up using new liver imaging tech-
niques. In addition, the development of the gut microbiome in
children may also play an important role [95].
Hepatitis B/C and hepatocellular carcinoma
For hundreds of years, pathologists have used the nucleolus to
prognosticate cancer without truly understanding the connec-
tion between them [96]. What is known is that many diverse
cancer cells have increased numbers and size of nucleoli, and
that these findings correlate with a worse cancer prognosis
(reviewed in [97]).
In hepatocellular carcinoma (HCC), increased nucleolar size
has been shown to predict the development of HCC in patients
with chronic liver disease [98–100]. This correlation is espe-
cially strong in patients infected with hepatitis B virus (HBV).
Interestingly, the HBx oncoprotein of HBV has been known to
directly affect nucleolar function by binding to the protein
NPM1, which aids in bringing HBx to the nucleolus [101].
Once in the nucleolus, HBx–NPM1 works to increase rDNA
transcription, leading to increased cellular proliferation and
transformation [101].
The function of the hepatitis C virus (HCV) is also linked to
ribosome biogenesis. rDNA transcription is upregulated in
HCV‐infected cells [102, 103] via UBF, where UBF is phospho-
rylated after HCV activation of cell cycle proteins cyclin D1
(CCND1) and cyclin‐dependent kinase 4 (CDK4). Additionally,
NS5B, the RNA polymerase required for HCV replication, also
interacts with the nucleolar phosphoprotein nucleolin (NCL)
[104]. There is thus a strong link between the nucleolus and
viruses responsible for chronic liver disease and HCC.
Fragile X syndrome
Fragile X syndrome is caused by mutation in the FMR1 gene on
the X chromosome that results in a greater frequency of a CGG
repeated sequence, leading to transcriptional silencing of FMR1
gene expression [105]. Loss of the FMR1 protein (FMRP)
causes a range of symptoms including intellectual disability and
obesity [106]. The cell responds to FMRP loss by increasing
mTOR and ERK signaling, both of which are crucial to ribo-
some biogenesis and protein synthesis [107, 108]. Decreases in
eIF4E phosphorylation cause a concomitant increase in the
translation of a protein important for neuronal function, matrix
metalloproteinase 9 (MMP‐9). Thus, FMRP loss in fragile X
causes an increase in MMP‐9, which in turn results in intellec-
tual disability.
180 THE LIVER:  REFERENCES
Although fragile X syndrome is not a liver‐specific disorder,
recent studies have suggested that it may be treated with met-
formin, a drug that affects liver function [109, 110]. Metformin
is prescribed to manage high blood sugar in patients with type 2
diabetes [111]. Metformin was tested as a therapeutic agent for
fragile X syndrome because it inhibits the mTORC1 and MAPK/
ERK pathways that are activated in fragile X patients and that are
important for liver development [112]. Interestingly, in Frm1−/y
mice, metformin treatment did restore many of the cognitive
defects, including increasing social preference and decreasing
repetitive behaviors. Additionally, defects in insulin signaling
and in circadian rhythms were also rescued by metformin treat-
ment in a Drosophila model of fragile X syndrome [113].
Interestingly, metformin treatment did not rescue levels of
ERK phosphorylation in the livers of Frm1−/y mice, despite
rescuing these levels in the brain [109]. Additionally, it is pos-
sible that metformin treatment is influencing other pathways
that are affected by FMRP loss, such as gluconeogenesis and
the gut microbiome, and it is the rescue of these other path-
ways that contributes to metformin’s effect on the signs and
symptoms of fragile X syndrome [109]. In liver cancer cells,
the drug metformin has recently been suggested to have an
anticancer effect through the DEPTOR–mTOR pathway [114],
suggesting an unexplored mechanism of metformin action as
related to FMRP loss.
CONCLUSION
Several liver‐specific disorders are directly influenced by ribo-
some biogenesis in humans. Understanding the cellular biology
of human ribosome biogenesis therefore has far‐reaching conse-
quences for treating liver disease. However, we have only a
modest understanding of the complex process of ribosome bio-
genesis itself in humans, and we are only beginning to under-
stand the layers of regulation needed to modulate ribosome
biogenesis at the tissue level. Indeed, there are many questions
remaining, including how a such a ubiquitous process as ribo-
some biogenesis may cause liver‐specific defects. Regardless, it
is clear that the process of making ribosomes and the regulation
of this process are important for liver metabolism and disease
pathogenesis. Therefore, targeting the nucleolus for therapeutic
intervention may be a viable option in the future of treatment of
liver‐associated pathologies.
REFERENCES
1. Anand, P. and Gruppuso, P.A. Rapamycin inhibits liver growth during
­refeeding in the rat via control of ribosomal protein translation but not cap‐
dependent translation initiation. J Nutr, 2006;136(1):27–33.
2. Conde, R.D. and Franze‐Fernandez, M.T. Increased transcription and
decreased degradation control and recovery of liver ribosomes after a period
of protein starvation. Biochem J, 1980;192(3):935–40.
3. Addis, T., Poo, L.J., and Lew, W. The rate of protein formation in the
organs  and tissues of the body: I. after casein refeeding. J Biol Chem,
1936;116(1):343–52.
4. Derenzini, M., Montanaro, L., and Trere, D. Ribosome biogenesis and cancer.
Acta Histochem, 2017;119(3):190–7.
5. Parlato, R. and Kreiner, G. Nucleolar activity in neurodegenerative
diseases:  a missing piece of the puzzle? J Mol Med (Berl),
­
2013;91(5):541–7.
6. McCann, K.L. and Baserga, S.J. Genetics. Mysterious ribosomopathies.
Science, 2013;341(6148):849–50.
7. Danilova, N. and Gazda, H.T. Ribosomopathies: how a common root can
cause a tree of pathologies. Dis Models Mech, 2015;8(9):1013–26.
8. Monty, K.J., Litt, M., Kay, E.R., and Dounce, A.L. Isolation and properties
of liver cell nucleoli. J Biophys Biochem Cytol, 1956;2(2):127–45.
9. Farley, K., Surovtseva, Y., Merkel, J., and Baserga, S. Determinants of mam-
malian nucleolar architecture. Chromosoma, 2015;124(3):323–31.
10. Shea, J.R. and Leblond, C.P. Number of nucleoli in various cell types of the
mouse. J Morphol, 1966;119(4):425–33.
11. Krüger, T., Zentgraf, H., and Scheer, U. Intranucleolar sites of ribosome
­biogenesis defined by the localization of early binding ribosomal proteins.
J Cell Biol, 2007;177(4):573.
12. Feric, M., Vaidya, N., Harmon, T.S. et al. Coexisting liquid phases underlie
nucleolar subcompartments. Cell, 2016;165(7):1686–97.
13. Sloan, K.E., Warda, A.S., Sharma, S. et al. Tuning the ribosome: The influ-
ence of rRNA modification on eukaryotic ribosome biogenesis and function.
RNA Biol, 2017;14(9):1138–52.
14. Henras, A.K., Plisson‐Chastang, C., O’Donohue, M.‐F., Chakraborty, A.,
and Gleizes, P.‐E. An overview of pre‐ribosomal RNA processing in eukary-
otes. Wiley Interdiscip Rev RNA, 2015;6(2):225–42.
15. Woolford Jr., J.L. and Baserga, S.J. Ribosome biogenesis in the yeast
Saccaromyces cerevisiae. Genetics [Internet], 2013;195:1–39.
16. Farley‐Barnes, K.I., McCann, K.L., Ogawa, L.M. et al. Diverse regulators of
human ribosome biogenesis discovered by changes in nucleolar number.
Cell Rep, 2018;22(7):1923–34.
17. Badertscher, L., Wild, T., Montellese, C. et al. Genome‐wide RNAi screen-
ing identifies protein modules required for 40S subunit synthesis in human
cells. Cell Rep, 2015;13(12):2879–91.
18. Wild, T., Horvath, P., Wyler, E. et al. A protein inventory of human ribosome
biogenesis reveals an essential function of exportin 5 in 60S subunit export.
PLoS Biol, 2010;8(10):e1000522.
19. Tafforeau, L., Zorbas, C., Langhendries, J.L. et al. The complexity of human
ribosome biogenesis revealed by systematic nucleolar screening of pre‐
rRNA processing factors. Mol Cell, 2013;51(4):539–51.
20. Zemp, I. and Kutay, U. Nuclear export and cytoplasmic maturation of
­ribosomal subunits. FEBS Lett, 2007;581(15):2783–93.
21. Gentilella, A., Kozma, S.C., and Thomas, G. A liaison between mTOR
signaling, ribosome biogenesis and cancer. Biochim Biophys Acta,
­
2015;1849(7):812–20.
22. Hay, N. and Sonenberg, N. Upstream and downstream of mTOR. Genes Dev,
2004;18(16):1926–45.
23. Thoreen, C.C., Kang, S.A., Chang, J.W. et  al. An ATP‐competitive mam-
malian target of rapamycin inhibitor reveals rapamycin‐resistant functions of
mTORC1. J Biol Chem, 2009;284(12):8023–32.
24. Yoshihama, M., Uechi, T., Asakawa, S. et al. The human ribosomal protein
genes: sequencing and comparative analysis of 73 genes. Genome Res,
2002;12(3):379–90.
25. Perry, R.P. The architecture of mammalian ribosomal protein promoters.
BMC Evolutionary Biol, 2005;5(1):15.
26. Meyuhas, O. and Kahan, T. The race to decipher the top secrets of TOP
mRNAs. Biochim Biophys Acta, 2015;1849(7):801–11.
27. Jefferies, H.B., Reinhard, C., Kozma, S.C., and Thomas, G. Rapamycin
selectively represses translation of the “polypyrimidine tract” mRNA family.
Proc Natl Acad Sci U S A, 1994;91(10):4441–5.
28. Terada, N., Patel, H.R., Takase, K. et  al. Rapamycin selectively inhibits
translation of mRNAs encoding elongation factors and ribosomal proteins.
Proc Natl Acad Sci U S A, 1994;91(24):11477–81.
29. Jefferies, H.B., Fumagalli, S., Dennis, P.B., Reinhard, C., Pearson, R.B., and
Thomas, G. Rapamycin suppresses 5’TOP mRNA translation through
­inhibition of p70s6k. EMBO J, 1997;16(12):3693–704.
30. Thoreen, C.C., Chantranupong, L., Keys, H.R. et al. A unifying model
for mTORC1‐mediated regulation of mRNA translation. Nature,
2012;485:109.
31. Anthony, T.G., Anthony, J.C., Yoshizawa, F., Kimball, S.R., and Jefferson,
L.S. Oral administration of leucine stimulates ribosomal protein mRNA
translation but not global rates of protein synthesis in the liver of rats. J Nutr,
2001;131(4):1171–6.
 15:  Ribosome Biogenesis and its Role in Cell Growth and Proliferation in the Liver 181
32. Reiter, A.K., Anthony, T.G., Anthony, J.C., Jefferson, L.S., and Kimball, S.R.
The mTOR signaling pathway mediates control of ribosomal protein mRNA
translation in rat liver. Int J Biochem Cell Biol, 2004;36(11):2169–79.
33. Aloni, R., Peleg, D., and Meyuhas, O. Selective translational control and
nonspecific posttranscriptional regulation of ribosomal protein gene expres-
sion during development and regeneration of rat liver. Mol Cell Biol,
1992;12(5):2203–12.
34. Yo‐ichi, N. and Kikuo, O. Stimulation of the synthesis of ribosomal pro-
teins  in regenerating rat liver with special reference to the increase in the
amounts of effective mRNAs for ribosomal proteins. Eur J Biochem,
1980;107(2):323–9.
35. Chen, H., Shen, F., Sherban, A. et al. DEPTOR suppresses lipogenesis and
ameliorates hepatic steatosis and acute‐on‐chronic liver injury in alcoholic
liver disease. Hepatology, 2018.
36. Porstmann, T., Santos, C.R., Griffiths, B. et al. SREBP activity is regulated
by mTORC1 and contributes to Akt‐dependent cell growth. Cell Metab,
2008;8(3–3):224–36.
37. Düvel, K., Yecies, J.L., Menon, S. et al. Activation of a metabolic gene regu-
latory network downstream of mTOR complex 1. Mol Cell,
2010;39(2):171–83.
38. Peterson, T.R., Sengupta, S.S., Harris, T.E. et al. mTOR complex 1 regulates
lipin 1 localization to control the SREBP pathway. Cell, 2011;146(3):408–20.
39. Tsukada, K. and Lieberman, I. Synthesis of ribonucleic acid by liver nuclear
and nucleolar preparations after partial hepatectomy. J Biol Chem,
1964;239(9):2952–6.
40. Dabeva, M.D. and Dudov, K.P. Transcriptional control of ribosome produc-
tion in regenerating rat liver. Biochem J, 1982;208(1):101–8.
41. Comai, L., Tanese, N., and Tjian, R. The TATA‐binding protein and associ-
ated factors are integral components of the RNA polymerase I transcription
factor, SL1. Cell, 1992;68(5):965–76.
42. Miller, G., Panov, K.I., Friedrich, J. et al. hRRN3 is essential in the SL1‐
mediated recruitment of RNA Polymerase I to rRNA gene promoters. EMBO
J, 2001;20(6):1373–82.
43. Hannan, K.M., Brandenburger, Y., Jenkins, A. et al. mTOR‐dependent regu-
lation of ribosomal gene transcription requires S6K1 and is mediated by
phosphorylation of the carboxy‐terminal activation domain of the nucleolar
transcription factor UBF. Mol Cell Biol, 2003;23(23):8862–77.
44. Mayer, C., Zhao, J., Yuan, X., and Grummt, I. mTOR‐dependent activation
of the transcription factor TIF‐IA links rRNA synthesis to nutrient availabil-
ity. Genes Dev, 2004;18(4):423–34.
45. Tsang, C.K., Liu, H., and Zheng, X.F.S. mTOR binds to the promoters of
RNA polymerase I‐ and III‐transcribed genes. Cell Cycle, 2010;9(5):953–7.
46. Zhao, J., Yuan, X., Frödin, M., and Grummt, I. ERK‐dependent phosphoryla-
tion of the transcription initiation factor TIF‐IA is required for RNA poly-
merase I transcription and cell growth. Mol Cell, 2003;11(2):405–13.
47. Stefanovsky, V., Langlois, F., Gagnon‐Kugler, T., Rothblum, L.I., and Moss,
T. Growth factor signaling regulates elongation of RNA polymerase I tran-
scription in mammals via UBF phosphorylation and r‐chromatin remode-
ling. Mol Cell, 2006;21(5):629–39.
48. Hannan, K.M., Hannan, R.D., Smith, S.D. et al. Rb and p130 regulate RNA
polymerase I transcription: Rb disrupts the interaction between UBF and
SL‐1. Oncogene, 2000;19(43):4988–99.
49. Cavanaugh, A.H., Hempel, W.M., Taylor, L.J. et al. Activity of RNA poly-
merase I transcription factor UBF blocked by Rb gene product. Nature,
1995;374(6518):177–80.
50. Donati, G., Montanaro, L., and Derenzini, M. Ribosome biogenesis and con-
trol of cell proliferation: p53 is not alone. Cancer Res, 2012;72(7):1602.
51. Ayrault, O., Andrique, L., and Séité, P. Involvement of the transcriptional
factor E2F1 in the regulation of the rRNA promoter. Exp Cell Res,
2006;312(7):1185–93.
52. Eymin, B., Karayan, L., Seite, P. et al. Human ARF binds E2F1 and inhibits
its transcriptional activity. Oncogene, 2001;20(9):1033–41.
53. Donati, G., Brighenti, E., Vici, M. et al. Selective inhibition of rRNA tran-
scription downregulates E2F‐1: a new p53‐independent mechanism linking
cell growth to cell proliferation. J Cell Sci, 2011;124(17):3017.
54. Zheng, K., Cubero, F.J., and Nevzorova, Y.A. c‐MYC – making liver sick:
role of c‐MYC in hepatic cell function, homeostasis and disease. Genes,
2017;8(4):123.
55. Poortinga, G., Wall, M., Sanij, E. et  al. c‐MYC coordinately regulates
­ribosomal gene chromatin remodeling and Pol I availability during granulo-
cyte differentiation. Nucl Acids Res, 2011;39(8):3267–81.
56. Poortinga, G., Hannan, K.M., Snelling, H. et al. MAD1 and c‐MYC regulate
UBF and rDNA transcription during granulocyte differentiation. EMBO J,
2004;23(16):3325–35.
57. Arabi, A., Wu, S., Ridderstråle, K. et al. c‐Myc associates with ribosomal
DNA and activates RNA polymerase I transcription. Nat Cell Biol,
2005;7:303.
58. Grandori, C., Gomez‐Roman, N., Felton‐Edkins, Z.A. et al. c‐Myc binds to
human ribosomal DNA and stimulates transcription of rRNA genes by RNA
polymerase I. Nat Cell Biol, 2005;7(3):311–18.
59. Boon, K., Caron, H.N., van Asperen, R. et al. N‐myc enhances the expres-
sion of a large set of genes functioning in ribosome biogenesis and protein
synthesis. EMBO J, 2001;20(6):1383–93.
60. Coller, H.A., Grandori, C., Tamayo, P. et al. Expression analysis with oligo-
nucleotide microarrays reveals that MYC regulates genes involved in growth,
cell cycle, signaling, and adhesion. Proc Natl Acad Sci U S A,
2000;97(7):3260–5.
61. Menssen, A. and Hermeking, H. Characterization of the c‐MYC‐regulated
transcriptome by SAGE: Identification and analysis of c‐MYC target genes.
Proc Natl Acad Sci U S A, 2002;99(9):6274–9.
62. Gomez‐Roman, N., Grandori, C., Eisenman, R.N., and White, R.J. Direct
activation of RNA polymerase III transcription by c‐Myc. Nature,
2003;421:290.
63. Kim, S., Li, Q., Dang, C.V., and Lee, L.A. Induction of ribosomal genes and
hepatocyte hypertrophy by adenovirus‐mediated expression of c‐Myc in
vivo. Proc Natl Acad Sci U S A, 2000;97(21):11198–202.
64. Nevzorova, Y.A., Cubero, F.J., Hu, W. et al. Enhanced expression of c‐myc
in hepatocytes promotes initiation and progression of alcoholic liver disease.
J Hepatol, 2016;64(3):628–40.
65. Farley, K.I. and Baserga, S.J. Probing the mechanisms underlying human
diseases in making ribosomes. Biochem Soc Trans, 2016;44(4):1035–44.
66. Woods, S.J., Hannan, K.M., Pearson, R.B., and Hannan, R.D. The nucleolus
as a fundamental regulator of the p53 response and a new target for cancer
therapy. Biochim Biophys Acta, 2015;1849(7):821–9.
67. Sloan Katherine, E., Bohnsack Markus, T., and Watkins Nicholas, J. The 5S
RNP couples p53 homeostasis to ribosome biogenesis and nucleolar stress.
Cell Rep, 2013;5(1):237–47.
68. Donati, G., Peddigari, S., Mercer, C.A., and Thomas, G. 5S ribosomal RNA
is an essential component of a nascent ribosomal precursor complex that
regulates the Hdm2‐p53 checkpoint. Cell Rep, 2013;4(1):87–98.
69. Fischer, M. Census and evaluation of p53 target genes. Oncogene,
2017;36:3943.
70. Krstic, J., Galhuber, M., Schulz, T., Schupp, M., and Prokesch, A. p53 as a
dichotomous regulator of liver disease: the dose makes the medicine. Int J
Mol Sci, 2018;19(3):921.
71. Fan, X., Chen, P., Tan, H. et  al. Dynamic and coordinated regulation of
keap1‐NRF2‐ARE and p53/p21 signaling pathways is associated with aceta-
minophen injury responsive liver regeneration. Drug Metab Dispos,
2014;42(9):1532.
72. Prokesch, A., Graef, F.A., Madl, T. et al. Liver p53 is stabilized upon starva-
tion and required for amino acid catabolism and gluconeogenesis. FASEB J,
2017;31(2):732–42.
73. Leung, N.W.Y., Farrant, P., and Peters, T.J. Liver volume measurement by
ultrasonography in normal subjects and alcoholic patients. J Hepatol,
1986;2(2):157–64.
74. Wilson, W.O. and McFarland, L.Z. Diurnal changes in livers and digestive
systems of Coturnix as related to three photoperiodic regimens. Poultry Sci,
1969;48(2):477–82.
75. Sinturel, F., Gerber, A., Mauvoisin, D. et al. Diurnal oscillations in liver mass
and cell size accompany ribosome assembly cycles. Cell, 2017;169(4):651–
63.e14.
76. Fisher, H.I. and Bartlett, L.M. Diurnal cycles in liver weights in birds. The
Condor, 1957;59(6):364–72.
77. Jouffe, C., Cretenet, G., Symul, L. et  al. The circadian clock coordinates
ribosome biogenesis. PLoS Biol, 2013;11(1):e1001455.
78. Janich, P., Arpat, A.B., Castelo‐Szekely, V., Lopes, M., and Gatfield, D.
Ribosome profiling reveals the rhythmic liver translatome and circadian
clock regulation by upstream open reading frames. Genome Res,
2015;25(12):1848–59.
79. Atger, F., Gobet, C., Marquis, J. et al. Circadian and feeding rhythms dif-
ferentially affect rhythmic mRNA transcription and translation in mouse
liver. Proc Natl Acad Sci U S A, 2015;112(47):E6579–88.
182 THE LIVER:  REFERENCES
80. Shcherbik, N., Wang, M., Lapik, Y.R., Srivastava, L., and Pestov, D.G.
Polyadenylation and degradation of incomplete RNA polymerase I tran-
scripts in mammalian cells. EMBO Rep, 2010;11(2):106–11.
81. Betard, C., Rasquin‐Weber, A., Brewer, C. et al. Localization of a recessive
gene for North American Indian childhood cirrhosis to chromosome
region  16q22‐and identification of a shared haplotype. Am J Hum Genet,
2000;67(1):222–8.
82. Weber, A.M., Tuchweber, B., Yousef, I. et al. Severe familial cholestasis in
North American Indian children: a clinical model of microfilament dysfunc-
tion? Gastroenterology, 1981;81(4):653–62.
83. Drouin, E., Russo, P., Tuchweber, B., Mitchell, G., and Rasquin‐Weber, A.
North American Indian cirrhosis in children: a review of 30 cases. J Pediatr
Gastroenterol Nutr, 2000;31(4):395–404.
84. Chagnon, P., Michaud, J., Mitchell, G. et al. A missense mutation (R565W)
in cirhin (FLJ14728) in North American Indian childhood cirrhosis. Am J
Hum Genet, 2002;71(6):1443–9.
85. Gallagher, J.E., Dunbar, D.A., Granneman, S. et al. RNA polymerase I tran-
scription and pre‐rRNA processing are linked by specific SSU processome
components. Genes Dev, 2004;18(20):2506–17.
86. Prieto, J.L. and McStay, B. Recruitment of factors linking transcription and
processing of pre‐rRNA to NOR chromatin is UBF‐dependent and occurs
independent of transcription in human cells. Genes Dev, 2007;
21(16):2041–54.
87. Freed, E.F., Prieto, J.L., McCann, K.L., McStay, B., and Baserga, S.J.
NOL11, implicated in the pathogenesis of North American Indian childhood
cirrhosis, is required for pre‐rRNA transcription and processing. PLoS
Genet, 2012;8(8):e1002892.
88. Wilkins, B.J., Lorent, K., Matthews, R.P., and Pack, M. p53‐mediated biliary
defects caused by knockdown of cirh1a, the zebrafish homolog of the gene
responsible for North American Indian childhood cirrhosis. PLoS One,
2013;8(10):e77670.
89. Yu, B., Mitchell, G.A., and Richter, A. Cirhin up‐regulates a canonical NF‐
kappaB element through strong interaction with Cirip/HIVEP1. Exp Cell
Res, 2009;315(18):3086–98.
90. Warren, A.J. Molecular basis of the human ribosomopathy Shwachman‐
Diamond syndrome. Adv Biol Regul, 2018;67:109–27.
91. Myers, K.C., Bolyard, A.A., Otto, B. et al. Variable clinical presentation of
Shwachman‐Diamond syndrome: update from the North American
Shwachman‐Diamond Syndrome Registry. J Pediatr, 2014;164(4):866–70.
92. Stepensky, P., Chacon‐Flores, M., Kim, K.H. et al. Mutations in EFL1, an
SBDS partner, are associated with infantile pancytopenia, exocrine pancre-
atic insufficiency and skeletal anomalies in aShwachman‐Diamond like syn-
drome. J Med Genet, 2017;54(8):558–66.
93. Alter, B.P., Giri, N., Savage, S.A., and Rosenberg, P.S. Cancer in the National
Cancer Institute inherited bone marrow failure syndrome cohort after fifteen
years of follow‐up. Haematologica, 2018;103(1):30–9.
94. Toiviainen‐Salo, S., Durie, P.R., Numminen, K. et al. The natural history of
Shwachman‐Diamond syndrome‐associated liver disease from childhood to
adulthood. J Pediatr, 2009;155(6):807–11 e2.
95. Leung, D.H. and Yimlamai, D. The intestinal microbiome and paediatric
liver disease. Lancet Gastroenterol Hepatol, 2017;2(6):446–55.
96. Pianese, G. Beitrag zur Histologie und Aetiologie der Carcinoma.
Histologische und experimentelle Untersuchungen. Beitr Pathol Anat Allg
Pathol, 1896;142:1–193.
97. Derenzini, M., Montanaro, L., and Trere, D. What the nucleolus says to a
tumour pathologist. Histopathology, 2009;54(6):753–62.
98. Trere, D., Borzio, M., Morabito, A. et al. Nucleolar hypertrophy correlates
with hepatocellular carcinoma development in cirrhosis due to HBV infec-
tion. Hepatology, 2003;37(1):72–8.
99. Derenzini, M., Trerè, D., Oliveri, F. et al. Is high AgNOR quantity in hepat-
ocytes associated with increased risk of hepatocellular carcinoma in
chronic liver disease? J Clin Pathol, 1993;46(8):727–9.
100. Borzio, M., Trere, D., Borzio, F. et  al. Hepatocyte proliferation rate is a
powerful parameter for predicting hepatocellular carcinoma development
in liver cirrhosis. Mol Pathol, 1998;51(2):96–101.
101. Ahuja, R., Kapoor, N.R., and Kumar, V. The HBx oncoprotein of hepatitis
B virus engages nucleophosmin to promote rDNA transcription and cellu-
lar proliferation. Biochim Biophys Acta, 2015;1853(8):1783–95.
102. Raychaudhuri, S., Fontanes, V., Barat, B., and Dasgupta, A. Activation of
ribosomal RNA transcription by hepatitis C virus involves upstream
­binding factor phosphorylation via induction of cyclin D1. Cancer Res,
2009;69(5):2057–64.
103. Kao, C.F., Chen, S.Y., and Lee, Y.H. Activation of RNA polymerase I
transcription by hepatitis C virus core protein. J Biomed Sci,
­
2004;11(1):72–94.
104. Hirano, M., Kaneko, S., Yamashita, T. et  al. Direct interaction between
nucleolin and hepatitis C virus NS5B. J Biol Chem, 2003;278(7):5109–15.
105. Kremer, E., Pritchard, M., Lynch, M. et al. Mapping of DNA instability at
the fragile X to a trinucleotide repeat sequence p(CCG)n. Science,
1991;252(5013):1711–14.
106. Rajaratnam, A., Shergill, J., Salcedo‐Arellano, M. et al. Fragile X s­ yndrome
and fragile X‐associated disorders. F1000Res, 2017;6:2112.
107. Gkogkas, C.G., Khoutorsky, A., Cao, R. et al. Pharmacogenetic inhibition
of eIF4E‐dependent Mmp9 mRNA translation reverses fragile X s­ yndrome‐
like phenotypes. Cell Rep. 2014;9(5):1742–55.
108. Hou, L., Antion, M.D., Hu, D. et al. Dynamic translational and proteasomal
regulation of fragile X mental retardation protein controls mGluR‐depend-
ent long‐term depression. Neuron, 2006;51(4):441–54.
109. Gantois, I., Khoutorsky, A., Popic, J. et  al. Metformin ameliorates core
deficits in a mouse model of fragile X syndrome. Nature Med,
2017;23(6):674–7.
110. Dy ABC, Tassone, F., Eldeeb, M., Salcedo‐Arellano, M.J., Tartaglia, N.,
and Hagerman, R. Metformin as targeted treatment in fragile X syndrome.
Clin Genet, 2018;93(2):216–22.
111. Samocha‐Bonet, D., Debs, S., and Greenfield, J.R. Prevention and t­ reatment
of type 2 diabetes: a pathophysiological‐based approach. Trends Endocrinol
Metab, 2018;29(6):370–9.
112. Soares, H.P., Ni, Y., Kisfalvi, K., Sinnett‐Smith, J., and Rozengurt, E.
Different patterns of Akt and ERK feedback activation in response to rapa-
mycin, active‐site mTOR inhibitors and metformin in pancreatic cancer
cells. PLoS One, 2013;8(2):e57289.
113. Monyak, R.E., Emerson, D., Schoenfeld, B.P. et  al. Insulin signaling
­misregulation underlies circadian and cognitive deficits in a Drosophila
fragile X model. Mol Psychiatry, 2017;22(8):1140–8.
114. Obara, A., Fujita, Y., Abudukadier, A. et  al. DEPTOR‐related mTOR
­suppression is involved in metformin’s anti‐cancer action in human
liver cancer cells. Biochem Biophys Res Commun, 2015;460(4):
1047–52.
 miRNAs and Hepatocellular
16 Carcinoma
Yusuke Yamamoto1, Isaku Kohama1, and Takahiro Ochiya1,2
1
Division of Molecular and Cellular Medicine, National Cancer Center Research Institute, Tsukiji,
Tokyo, Japan
2
Institute of Medical Science, Tokyo Medical University, Shinjuku, Tokyo, Japan
INTRODUCTION
Several types of small noncoding RNAs have been identified in
eukaryotic cells, including microRNAs (miRNAs), small inter-
fering RNAs (siRNAs), and Piwi‐interacting RNAs (piRNAs)
[1, 2]. Revolutionary advances in next‐generation sequencing
have made substantial contributions to the discovery of new
classes of small RNAs and the identification of a large number
of these nucleotides. miRNAs have a length of 18–24 n­ ucleotides,
are endogenously expressed in almost all types of eukaryotic
cells, and primarily silence gene expression at the post‐
transcriptional level. In general, miRNAs are incorporated into
the RNA‐induced silencing complex (RISC) that functions as an
RNA silencer and is located in the cytoplasm of cells [3]. Based
on a computational prediction model, the expression of approxi-
mately 30% of all genes is influenced by miRNA‐induced gene
silencing [4]. Notably, miRNAs play important roles in regulat-
ing nearly every physiological cellular process, and their dys-
regulation is strongly correlated with various types of disease,
including cancer [3].
The expression patterns and levels of miRNAs are tightly
regulated in different types of cells during development; for
example, miR‐122 expression is limited to hepatocytes [5]. A
number of miRNAs have been shown to regulate cellular dif-
ferentiation and proliferation under physiological conditions
[6, 7]. Likewise, many miRNAs have been reported to play
key roles in cancer initiation and progression, including inva-
sion and metastasis processes in almost all types of cancers
[8–10]. The purpose of this chapter is to provide an overview
of the functions of miRNAs in hepatocellular carcinoma
(HCC) and their clinical applications, such as diagnostics
and therapeutics.
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
THE BASICS OF miRNA BIOLOGY
AND BIOGENESIS
The sequences of miRNAs are highly conserved across species
[11]. Primary miRNA genes (pri‐miRNAs) are transcribed by
the RNA polymerase II (Pol II) transcription units, and the
transcripts are called pri‐miRNAs (Figure 16.1). The pri‐miRNAs
possess a cap and a poly‐A structure, as well as protein‐coding
genes. The size of the pri‐miRNA usually exceeds 1 kilobase,
and it contains an RNA hairpin loop structure with a mature
miRNA sequence [12]. In most cases, one pri‐miRNA contains
one mature miRNA; however, some pri‐miRNAs include sev-
eral mature miRNAs, such as the miR‐17–92 cluster and miR‐
106a–363 cluster [13]. Some miRNA families, particularly the
let‐7 family, are widely conserved in many species and include
more than 10 mature miRNAs in humans [14]. Some miRNAs
are also located in the intron regions of protein‐coding genes,
and intronic RNAs become pri‐miRNAs during the process of
splicing; thus, the expression levels of these types of miRNAs
correlate with the parent gene expression [15].
During miRNA biogenesis, pri‐miRNAs, which are transcribed
by a Pol II RNA polymerase, contain hairpin loop structures
(approximately 70 nucleotides) that are recognized by a
DiGeorge syndrome critical region 8 (DGCR8) nuclear protein
in the nucleus. DGCR8 interacts with Drosha [16], an RNase III
endonuclease, to form the microprocessor complex. In the com-
plex, Drosha cleaves the pri‐miRNAs at the base of stem‐loop
structures and creates 60–70 nucleotide stem‐loop intermedi-
ates, called a precursor miRNA (pre‐miRNA) [17]. The pre‐miRNAs
in the nucleus are transported to the cytoplasm by the nucleo-
cytoplasmic shuttler exportin‐5 (Exp‐5) [18]. In the cytoplasm, the
hairpin structure of pre‐miRNA is cleaved by the RNase III
184 THE LIVER:  miRNAs IN HEPATOCELLULAR CARCINOMA

DNA

Transcribed by Pol RNA polymerase

Nucleus Pri-miRNA
(> 1 kb)
Processed by Drosha, DGCR8

Pre-miRNA
(60–70 nt)
Exported by Exp5
Cytoplasm

Mature-miRNA (22 nt) RISC

5’UTR 3’UTR

Protein coding sequence AAAAAAA

Translational inhibition
or mRNA degradation

Figure 16.1  The biogenesis and function pathways of miRNAs. Primary miRNA genes (pri‐miRNAs) are transcribed by the Pol II transcription
units, and the transcripts are called pri‐miRNAs with a cap and a poly‐A structure. In the process of miRNA biogenesis, pri‐miRNAs are recognized
by a DGCR8 nuclear protein in the nucleus. DGCR8 interacts with Drosha, an RNase III endonuclease, which cleaves the pri‐miRNAs at the base
of stem‐loop structures and creates a pre‐miRNA. The pre‐miRNAs in the nucleus are transported to the cytoplasm by Exp‐5. In the cytoplasm,
Dicer, an RNase III enzyme cleaves pre‐miRNA, yielding an imperfect miRNA duplex of approximately 22 nucleotides. The “guide” strand of the
duplex incorporates into RISC. Based on the “seed” sequence in miRNAs, the expression of target genes are inhibited by either translational
inhibition (partial complementarity) or mRNA degradation (perfect complementarity).

enzyme Dicer, yielding an imperfect miRNA duplex of approxi-
mately 22 nucleotides [17]. The imperfect miRNA duplex con-
tains a “guide” strand that is the antisense strand of the target
gene sequence and a “passenger” strand. Basically, the “guide”
strand of the duplex incorporates into the RNA‐induced silenc-
ing complex (RISC), although “passenger” strand of the duplex
might also function as an active miRNA in some cases. The
RISC contains primary miRNA‐binding proteins, namely, a
member of the argonaute protein family. In mouse and human,
four Ago (Ago I–IV) proteins are conserved [19]. Ago proteins
possess a PAZ domain that is responsible for binding both sin-
gle‐stranded and duplexed RNA; Ago II is a major protein in the
RISC with helicase and endonuclease activities [20]. In the pro-
cess of RISC assembly, the “passenger” strand of the imperfect
miRNA duplex that was cleaved by Dicer is removed and even-
tually degraded. The “guide” strand is selectively incorporated
into the RISC and acts as a gene silencer. When the “guide”
strand of the miRNA tightly binds to the Ago protein in the
RISC, the miRNA recognizes target genes through homologous
sequences at the 5′ end of the miRNA, which is designated the
“seed” sequence (generally, 6–8 nucleotides) [18]. Primarily,
miRNAs negatively regulate the expression target genes through
an interaction between the “seed” sequence at their 5′ ends and
the 3′ untranslated region (3′ UTR) of their target genes. Based
on the complementarity of the “seed” sequence with the target
gene sequence, the mature miRNA causes either translational
inhibition (partial complementarity) or mRNA degradation
(perfect complementarity) (Figure  16.1). A single miRNA is
considered to possess the capacity to control hundreds of target
genes and influence a wide range of gene pathways. These
unique characteristics indicate that miRNAs are potentially
critical modulators of nearly all physiological events and the
development of various diseases. Thus, one of the key issues in
miRNA biology is to identify the target genes. The initial pro-
cess used to predict target genes is to observe the pairing of the
3′ UTR of target genes and miRNA “seed” sequence located
from nucleotides 2 to 8 at the 5′ end of the miRNA. Several
online bioinformatics tools have been developed to predict the
miRNA target genes: Target Scan 7.1 (http://www.targetscan.
org/vert_72/), miRD (http://mirdb.org/), miRBase (http://www.
mirbase.org/help/targets.shtml), and miRWalk2.0 (http://zmf.
umm.uni‐heidelberg.de/apps/zmf/mirwalk2/). These websites
predict both (i) the genes that are potentially targeted by a specific
miRNA and (ii) miRNAs that bind to a specific gene.
miRNAs IN HEPATOCELLULAR
CARCINOMA
Hepatocellular carcinoma
The primary malignancies of the liver are mainly HCC, cholan-
giocarcinoma, and hepatic angiocarcinoma. More than 90% of
 16: 
miRNAs AND HEPATOCELLULAR CARCINOMA 185
liver cancer cases are HCC, and it frequently develops in patients
with hepatitis B virus (HBV) and hepatitis C virus (HCV)
infections and cirrhosis (80–90% of the HCC cases) [21]. Although
the treatment outcomes are continuously improving, HCC is the
third leading cause of cancer‐related death worldwide. Due to the
higher prevalence rates of HBV and HCV, the incidence rates of
HCC are higher in Asia and Africa than elsewhere; the total
number of the HCC cases worldwide is expected to increase in
the next few years. HCV is the most common cause of HCC in
Japan and Western countries, but in other Asian countries and
Africa, HBV is the most common cause of HCC development.
The main cause of hepatitis is a virus infection, but it is also
caused by alcoholism (alcoholic liver disease, ALD) and steato-
sis (non‐alcoholic steatohepatitis, NASH). HBV and HCV
infections cause chronic hepatitis, leading to abnormal cell pro-
liferation and death. Hepatitis also gradually results in genetic
alterations in the liver. Many people who become infected with
these viruses do not even know it. Some become asymptomatic
long‐term HBV or HCV carriers and develop liver fibrosis,
which subsequently leads to liver cirrhosis within 20 years. In
the last stage of the disease, the cirrhotic liver eventually devel-
ops HCC. Approximately 80% of the HCC cases develop from
liver cirrhosis. In the process of HCC development from hepati-
tis via liver cirrhosis, hepatocytes repeatedly grow and undergo
cell death, leading to the accumulation of genetic aberrations,
such as point mutations, genomic deletions, and amplifications,
which are the main causes of HCC development [22].
When the patients have a sufficient hepatic function reserve, the
optimal treatment for HCC is partial surgical resection. Although
the 5‐year survival rates after treatment have substantially improved
over the past few decades, the recurrence rates is still very high:
approximately 70%. Another HCC treatment is liver transplanta-
tion. Generally, patient who are eligible for liver transplantation
have multiple liver dysfunctions. Additionally, many HCC cases
are detected at an advance stage; thus, less than 40% of patients
with HCC are eligible for surgery and transplantation.
In general, radiological approaches, such as ultrasonography
and computed tomography, are used for the diagnosis and sur-
veillance of HCC. However, these methods are not optimal to
detect a small lesion in the liver. Other approaches for HCC
screening are serological tests of tumor biomarkers, such as α‐
fetoprotein (AFP) and protein induced by vitamin K absence or
antagonist‐II (PIVKA‐II). These methods are less invasive and
can be readily applied for HCC screening; however, the low sen-
sitivity and specificity of serological biomarkers are critical
issues, particularly for the detection of early‐stage HCC [23,
24]. Recently, expression profiling of miRNAs has been consid-
ered as a novel biomarker to detect cancer. The first studies were
performed using an miRNA microarray, which provided sub-
stantial contributions to miRNA biology and cancer biology. In
these studies, miRNAs whose expression levels in cancer tis-
sues were decreased compared with patient‐matched normal tis-
sues, target oncogenes, and thus the miRNAs originally function
as tumor suppressors. In contrast, miRNAs whose expression
levels are increased in cancer target tumor suppressor genes and
function as oncogenes in cancer development. Importantly,
these differential expression profiles of miRNAs are closely
associated with the clinical outcomes in patients with many
types of cancer, including HCC.
Although the earliest studies were mainly based on the
miRNA microarray platform, recent technological advances
in next‐generation sequencing have also contributed to
miRNA research in the field of cancer biology. Using both
comprehensive approaches, several miRNAs have been
identified as oncogenic and tumor‐suppressive miRNAs in
HCC, and some of these miRNAs are also considered diag-
nostic and prognostic biomarkers. As mentioned earlier,
these miRNAs likely affect various pathways by inhibiting
hundreds of genes. Although many studies have been per-
formed to investigate miRNA functions and their targets in
other types of cancer, not specifically HCC, theoretically,
the target genes identified for certain miRNAs in other can-
cers are potentially the same targets in HCC and thus the
pathways affected by certain miRNAs would also be similar.
Therefore, miRNA research in the field of HCC will likely
contribute not only to improving our understanding of
miRNA functions in HCC but also to development of inno-
vative and feasible therapeutic methods and novel diagnostic
strategies for patients with HCC, particularly for the early
and accurate detection of HCC.
Oncogenic roles of miRNAs in HCC
A comprehensive analysis of miRNA profiling in HCC cases
provides clear evidence that miR‐21 functions as an oncogene
during HCC development. Notably, miR‐21 is one of the miR-
NAs with the highest expression in primary HCCs, and its abun-
dant expression is a hallmark of various types of cancers. As
miR‐21 expression is substantially increased in various types of
cancers, it is designated as an “oncomiR.” One of the major
functions of miR‐21 in HCC is to inhibit the expression of a
tumor suppressor gene called phosphatase and tensin homolog
deleted from chromosome 10 (PTEN), and its expression is
closely associated with the poor differentiation of tumor cells.
The PTEN gene has empirically been confirmed to be a direct
target of miR‐21. PTEN is a phosphatase that dephosphorylates
focal adhesion kinase (FAK). When FAK is phosphorylated, it is
activated. The status of FAK activation correlates with aggres-
sive tumor behavior in HCC [25]. Therefore, miR‐21‐mediated
PTEN downregulation results in an increase in FAK phospho-
rylation and an aggressive phenotype of HCC characterized by
invasion. The inhibition of miR‐21 decreases the growth rate of
HCC cells in soft agar and promotes apoptotic cell death [26].
As expected, miR‐21 overexpression promotes the growth,
migration, and invasion of HCC cells [27]. The level of the
miR‐21 transcript is increased by signal transducer and activator
of transcription 3 (STAT3), a major mediator of interleukin 6
(IL‐6) signaling. STAT3 is involved in tumor transformation by
suppressing apoptotic signaling and directly binds to the pro-
moter region of the miR‐21 primary transcript to induce its
­transcription [28].
Another well‐known oncogenic miRNA in HCC is the
miR‐17–92 polycistronic cluster. This miRNA cluster contains
six miRNAs: miR‐17, miR‐18a, miR‐19a, miR‐20a, miR‐19b,
and miR‐92a, which are also considered “oncomiRs.”
Overexpression of the miR‐17–92 cluster was confirmed in
100% of human HCC cases, and some of these miRNAs, such
as miR‐20a, were expressed in high levels in cirrhotic liver
186 THE LIVER:  miRNAs IN HEPATOCELLULAR CARCINOMA
tissues, suggesting that they might be important for the early
stage of HCC development [29]. The expression of the miR‐17–
92 polycistronic cluster is positively regulated by the c‐Myc
oncogene. The direct binding of c‐Myc to the promoter region
in the miR‐17–92 polycistronic cluster has been reported [30].
The function of the miR‐17–92 cluster in vivo was investigated
using liver‐specific miR‐17–92 transgenic mice in combination
with a hepatic carcinogen (diethylnitrosamine) [31]. In this
mouse model, liver‐specific miR‐17–92 transgenic mice devel-
oped a significantly greater number of HCC lesions than the
control mouse group treated with diethylnitrosamine. Similar
data were also obtained from in vitro experiments; this research
group revealed that increased expression of the miR‐17–92
cluster clearly promoted the growth, colony formation, and
invasion of human HCC cell lines [31].
Profiling of miRNA expression in HCC cell lines derived
from chronic carriers of HBV and HCV and patients with
nonviral‐associated HCC identified the differential expres-
sion of several miRNAs: upregulation of miR‐222, miR‐221,
and miR‐31, and downregulation of miR‐223, miR‐126, and
miR‐122a. Upregulated miRNAs might function as onco-
genes, and in contrast, downregulated miRNAs might func-
tion as tumor suppressors. Further analyses showed that the
deregulated expression patterns of miR‐223 and miR‐222
clearly differentiated HCC cases from noncancerous liver tis-
sues, regardless of the viral infection status [32]. In another
study of HCV‐infected HCC cases, the miRNA profiling in a
set of 52 human primary liver tumors consisting of premalig-
nant dysplastic liver nodules and HCCs using quantitative
real‐time polymerase chain reaction (qRT‐PCR), identified 10
upregulated and 19 downregulated miRNAs compared to nor-
mal liver tissues [33]. A further analysis of these miRNAs
confirmed that expression levels of miR‐122, miR‐100, and
miR‐10a were significantly increased in HCV‐infected HCC
cases, indicating that they functioned as putative oncogenic
miRNAs. Likewise, miRNA profiling of 89 HCC samples
using a ligation‐mediated amplification method identified
subclasses of HCC (three main clusters: the wingless‐type
MMTV integration site, interferon‐related, and proliferation)
through an unsupervised clustering analysis. In a subset of the
proliferation subclass, poorly characterized miRNAs from
chr19q13.42 were expressed at high levels. Among these
miRNAs, enhanced expression of miR‐517a and miR‐520c
promoted the growth, migration, and invasion of the HCC
cells in vitro. In particular, miR‐517 enhanced tumorigenesis
and metastasis in an in vivo model. Thus, they are also consid-
ered oncogenic miRNAs [34].
As a new paradigm of cancer biology, many researchers
have focused on tumor‐initiating cells (TICs) or cancer stem
cells (CSCs), which are considered to possess a higher self‐
renewal capability and drug‐ and radio‐resistance. Ma et  al.
[35] identified a CD133‐positive population of HCC cells as
CSCs in HCC (approximately 1.3–13.6% of the cells in the
tumor bulk) and also observed higher levels of miR‐130b
expression in the CD133‐positive CSC population than in the
CD133‐negative main population. Importantly, when miR‐
130b was overexpressed in the CD133‐negative HCC cells
from a lentivirus vector, the cells exhibited a higher resistance
to chemotherapeutic agents and a significant increase in
tumorigenicity in vivo. In this study, TP53INP1 was identified
as one of the miR‐130b target genes. Thus, miR‐130b is one of
the key miRNAs regulating cancer stemness, including the
self‐renewal capability and drug resistance, indicating that it is
also an oncogenic miRNA [35].
The major issue in curing cancer is finding methods to pre-
vent cancer metastasis and recurrence after curative treat-
ment. Investigations of the molecular mechanisms of these
processes will provide new insights into clinical applications
for patients with HCC. In the field of miRNA biology, several
miRNAs have been reported to function as upstream regula-
tors of key molecules responsible for migration, invasion, and
metastasis, which have important roles in determining liver
cancer outcomes. Some oncogenic miRNAs have been
described as pro‐metastatic miRNAs. For example, overex-
pressed oncogenic miR‐17–5p promotes the migration of
HCC cells in vitro and in vivo by activating the p38 mitogen‐
activated protein kinase (MAPK) pathway and increasing the
phosphorylation of heat shock protein 27 (HSP27) [36]. Ding
et  al. [37] identified 22 miRNAs located at amplified or
deleted genomic DNA regions in HCC. Among them,
miR‐151, which is frequently amplified on chromosome
8q24.3 and is coexpressed with its host gene FAK, is closely
correlated with the intrahepatic metastasis of HCC.
Overexpression of miR‐151 induced the migration and inva-
sion of HCC cells in vitro and in vivo by inhibiting the expres-
sion of RhoGDIA, a putative metastasis suppressor in HCC
[37]. Another pro‐metastatic miRNA, miR‐143, has been
reported, and fibronectin type III domain‐containing 3B
(FNDC3B) is one of its direct target genes [38]. In this study,
the intratumor administration of miR‐143 significantly
enhanced HCC metastasis in an HCC cell‐xenografted mouse
model. Additionally, a study using p21‐HBx transgenic mice
showed that miR‐143 inhibition significantly blocked local
liver metastasis and distant lung metastasis [38]. Oncogenic
miR‐517 is also defined as a pro‐metastatic miRNA whose
overexpression causes metastatic dissemination in vivo [34].
A study profiling the expression of 156 miRNAs in HCC
cases revealed that miR‐222 upregulation was commonly
observed in an HCC cohort. The suppression of miR‐222
expression led to reduced cell motility by negatively regulat-
ing AKT signaling. In this study, protein phosphatase 2A
subunit B (PPP2R2A) was identified as one of the miR‐222
target genes using an in silico assay and luciferase reporter
assay of the PPP2R2A 3′UTR. Based on these data, miR‐222
is also a pro‐metastatic miRNA [39].
The oncogenic miRNAs have also been reported to be key
factors in intrahepatic cholangiocarcinoma (ICC). For exam-
ple, miR‐191 plays an important role in tumorigenesis. Next‐
generation sequencing technology with five pairs of ICC and
matched‐to‐normal bile duct tissues revealed higher levels of
miR‐191 expression in ICC than in adjacent normal bile duct
tissues. Overexpression of miR‐191 promotes the growth,
invasion, and migration of ICC cells in vitro and in vivo.
Moreover, ten–eleven translocation 1 (TET1) was identified
as a direct target gene of miR‐191, and reduced TET1 expres-
sion causes p53 promoter methylation, suppressing p53 tran-
scription. The expression of miR‐191 correlates with poor
prognosis for patients with ICC [40].
 16: 
miRNAs AND HEPATOCELLULAR CARCINOMA 187
Tumor‐suppressive effect of miRNAs on HCC
In the processes of cancer initiation and progression, both onco-
genes and tumor suppressor genes are essential components of
the mechanism regulating the cell cycle. In miRNA biology, a
number of reports identified specific miRNAs that negatively
regulate the cell cycle and function as tumor suppressors. Most
of tumor‐suppressive miRNAs target cyclin–cyclin‐dependent‐
kinase (cyclin–CDK) complexes, a class of positive modulators
of the cell cycle. A class of miRNAs were reported to function
as major negative regulators of cell cycle progression; this group
includes miR‐1, miR‐22, miR‐34a, miR‐122, miR‐375, and
let‐7. The most interesting of these miRNAs is miR‐122,
because it is abundantly expressed in the liver and comprises
over 70% of the total amount of miRNAs.
The expression of miR‐122 is specifically and significantly
downregulated in human HCC cases compared with adjacent
normal liver tissues [41]. As shown in the study by Tsai et al.,
miR‐122 functions as a tumor suppressor in HCC, and the
authors experimentally identified 32 target genes related to cell
movement, cell morphology, cell–cell signaling, and transcrip-
tion [41]. The miR‐122 expression level is downregulated in
most HCC cases [42]. In this study, cyclin G1 was one of the
target genes of miR‐122; the expression levels of cyclin G1
inversely correlated with miR‐122 expression. Obviously, cyc-
lin G1 is a positive regulator of cell cycle progression, leading to
p53 downregulation and the induction of hepatocarcinogenesis,
and as miR‐122 inhibits cyclin G1, it functions as a tumor sup-
pressor in HCC. In contrast, the aberrant expression of miR‐122
has also been reported in patients with both HCC and HCV
infection [33], presumably because miR‐122 expression in
hepatocytes is essential for HCV RNA accumulation [43].
During liver development, miR‐122 expression is associated
with four liver‐selective transcription factors, hepatocyte nuclear
factor 1α (HNF1α), HNF3β, HNF4α, and CCAAT/enhancer‐
binding protein α (C/EBPα), and regulates the balance between
the proliferation and differentiation of hepatocytes by directly
inhibiting CUTL1, a transcriptional repressor of genes specify-
ing terminal differentiation in multiple cell lineages [44].
Therefore, miR‐122 is thought to generally function as a tumor
suppressor in HCC, but its roles in the specific circumstances
remain unclear.
Primary HCC exhibits a variety of genomic abnormalities,
such as chromosomal instability, CpG hypermethylation, DNA
rearrangements associated with HBV integration, DNA hypo-
methylation, and, to a lesser extent, microsatellite instability
[45]. The hypermethylation of the promoter regions of tumor‐
suppressive miRNAs causes decreased expression and/or silenc-
ing in cancer cells. The expression of miR‐1 is downregulated in
primary HCC cases compared with patient‐matched liver tis-
sues, and miR‐1 is also a well‐known tumor suppressor miRNA.
Initially, miR‐1 was identified as a methylated miRNA in HCC;
thus, its expression is inhibited. Following treatment with 5‐
azacytidine (DNA hypomethylating agent) and/or trichostatin
A (histone deacetylase inhibitor), miR‐1 expression is restored
in HCC cells. Overexpression of miR‐1 suppresses cell growth
and induces apoptotic cell death by directly targeting the FoxP1,
MET, and HDAC4 genes [46]. Likewise, miR‐148 was origi-
nally identified as a hepatospecific miRNA that was expressed
at high levels in the mature liver and controlled the differentiation
of fetal hepatoblasts by directly targeting DNA methyltrans-
ferase 1 (DNMT1), a major enzyme responsible for epigenetic
silencing [47]. The authors also reported a beneficial effect of
miR‐148a on HCC suppression as a tumor suppressor miRNA.
Regardless of the DNMT1 expression level, miR‐148a func-
tions as a tumor suppressor by controlling the expression of the
c‐Met oncogene [47].
The miR‐375 expression level was also downregulated in
HCC tissues compared with adjacent nontumor tissues from
patients with HCC. Yes‐associated protein (YAP) is a potent
oncogenic driver, and miR‐375 overexpression decreases the
transcriptional activity of YAP, thus exerting a tumor‐suppres-
sive effect on HCC cells [48]. The research group also identified
stathmin 1 as one of downstream targets of miR‐223 using lucif-
erase reporter assay with the 3′ UTR; stathmin 1 expression
inversely correlates with miR‐223 expression in HCC cells [32].
miR‐34a is one of the most well‐known tumor suppressor
miRNAs whose transcription is directly regulated by p53, the
guardian of the genome. A qRT‐PCR analysis of 83 HCC for-
malin‐fixed paraffin‐embedded (FFPE) tissue samples revealed
decreased miR‐34a expression in HCC samples compared to the
patient‐matched liver tissues. As expected, ectopic expression
of miR‐34a inhibits cell proliferation and induces apoptosis in
HCC cells by influencing phospho‐ERK1/2 and phospho‐
STAT5 signaling. Thus, miR‐34a also functions as a tumor sup-
pressor in HCC [49].
Tumor suppressor miRNAs also function as anti‐metastatic
miRNAs to some extent. These miRNAs mainly modulate the
epithelial‐to‐mesenchymal transition (EMT). As a typical EMT‐
related miRNA, the transcription of the miR‐200 family is posi-
tively regulated by p53, which suppresses tumor progression
and metastasis. The upregulation of the miR‐200 family by p53
was confirmed by a profiling analysis of 92 primary HCCs and
9 HCC cell lines. The miR‐200 family targets ZEB1 and ZEB2,
which are master regulators of the EMT. Thus, p53 inhibits
EMT through miR‐200 family‐mediated ZEB1 and ZEB2
repression. Additionally, p53‐mediated activation of the
miR‐192 family suppresses ZEB2 expression. These miRNAs
could function as anti‐metastatic miRNAs by repressing the
EMT phenotype [50]. c‐Met is a transmembrane tyrosine kinase
receptor and an inducer of the EMT. Other tumor suppressor
miRNAs, such as miR‐34a and miR‐23b, have also been
reported to possess anti‐metastatic functions; miR‐34a sup-
presses tumor invasion and migration by directly targeting
c‐Met in HepG2 cells [51], and miR‐23b expression leads to
urokinase‐type plasminogen activator (uPA) and c‐Met down-
regulation and subsequently reduces cancer invasion and metas-
tasis [52]. Notably, miR‐122 has been reported to function as an
anti‐metastatic miRNA by directly inhibiting disintegrin and
metalloproteinase domain‐containing protein 10 (ADAM10)
[53] and ADAM17 [41]. ADAM family genes are closely asso-
ciated with cancer metastasis; thus, miR‐122 functions as an
anti‐metastatic miRNA. Recently, Su et al. identified miR‐217,
whose expression level was decreased in highly invasive
MHCC‐97H HCC cells and metastatic HCC tissues. Ectopic
expression of miR‐217 suppressed the invasion of MHCC‐97H
cells by directly repressing E2F3 [54]. More recently, miR‐
501–3p has been described to be involved in the metastasis of
188 THE LIVER:  ROLES OF miRNAs IN HEPATITIS VIRUS INFECTION
HCC. Significantly lower expression of miR‐501–3p was
observed both in metastatic HCC cell lines and recurrent and
metastatic HCC tissue samples. Similar to other research reports
of anti‐metastatic miRNAs, ectopically enhanced expression of
miR‐501‐3p suppresses the growth, migration, invasion and
EMT of metastatic HCC cells. Lin‐7 homolog A (LIN7A) was
identified as a direct target of miR‐501‐3p, and its inhibition
suppressed metastasis in HCC cells [55].
These tumor suppressor and/or anti‐metastatic miRNAs rep-
resent potentially useful prognostic biomarkers and treatment
strategies for HCC.
ROLES OF miRNAs IN HEPATITIS VIRUS
INFECTION
Although miRNAs are involved in the process of HCC initiation
and progression, they are also closely associated with infections
with the hepatitis viruses HBV and HCV. HBV is a nuclear
DNA virus, and no miRNAs have been identified in the HBV
genome, although one miRNA sequence in the HBV genome
was computationally predicted [56]. In contrast, HBV influ-
ences the expression of miRNAs in the host cells to create a
favorable environment for HBV replication and immune eva-
sion. HBV infection is basically characterized by two phases:
the acute and chronic phases. In the acute phase, HBV actively
replicates while evading the host immune attack. In the chronic
phase, the HBV infection becomes dormant and the virus repli-
cates stably and escapes from the immune system. In both infec-
tion phases, miRNAs in the host cells play important roles in the
interaction between the host and virus. The most well‐studied
miRNA involved in HBV infection is miR‐122, which is a liver‐
specific miRNA that is abundantly expressed in hepatocytes.
Although miR‐122 is essential for the process of HCV infec-
tion, the loss of miR‐122 unexpectedly promotes the replication
of HBV [57, 58]. miR‐122 directly binds to the HBV genome
sequence, which is highly conserved, and suppresses the expres-
sion of the viral genes. Moreover, HBV X protein (HBx), which
is essential for the transcription of the HBV genome, binds per-
oxisome proliferator‐activated receptor‐gamma (PPARγ) and
suppresses the transcription of miR‐122, presumably inducing
HBV replication [59].
The transfection of miRNA mimics showed that overexpres-
sion of miR‐1 increases HBV replication and upregulates HBV
core promoter transcription, antigen expression, and progeny
secretion. These effects of miR‐1 on HBV replication are not
directly mediated by targeting the HBV genome, as revealed by
bioinformatics and luciferase reporter analyses. Thus, miR‐1
regulates the expression of several host genes to enhance HBV
replication and reverse the cancer cell phenotype [60]. Moreover,
a screen with an miRNA library containing 2048 miRNAs iden-
tified 39 miRNAs that repress HBV replication by examining
the intracellular and extracellular DNA and HBsAg levels. In
particular, miR‐204 substantially decreases both HBV DNA and
HBsAg levels in HCC cells. Rab22a was identified as one of the
targets of miR‐204, and the loss of Rab22a also suppresses
intracellular and extracellular HBV DNA expression [61]. On
the other hand, HBx regulates miRNA expression in the host
cells. Notably, let‐7a has been identified to be negatively regu-
lated by HBx; their expression levels are inversely correlated.
As let‐7a directly targets STAT3, HBx‐mediated downregula-
tion of let‐7a and upregulation of STAT3 support cell prolifera-
tion, leading to hepatocarcinogenesis [62]. Additionally, an
miRNA microarray analysis revealed 10 miRNAs that were dif-
ferentially expressed between a stable HBV‐producing cell line
(HepG2.2.15) and its control cell line (HepG2). High expres-
sion of miR‐501 has been observed in HBV‐producing cells.
The loss of miR‐501 expression suppresses HBV replication,
but does not affect the growth of HBV‐producing cells.
According to the results of a luciferase reporter assay, HBXIP,
an inhibitor of HBV replication, is a potential target of miR‐501.
Thus, miR‐501 would be an important miRNA regulating HBV
replication in the host cells [63].
miRNAs are involved in the physiology and function of the
immune system. Host innate antiviral immunity is the first line
of defense against hepatitis virus infection. The defense mecha-
nism is precisely controlled by a number of genes at multiple
stages. The expression of miRNAs in the host cells was consist-
ently shown to play a key regulatory role in the process of HBV
infection and its defense. Notably, miR‐155 is upregulated dur-
ing HBV infection and affects the host immune response;
miR‐155 regulates the acute inflammatory response after the
recognition of pathogens by toll‐like receptors. Overexpression
of miR‐155 induces the expression of several interferon (IFN)‐
inducible antiviral genes and inhibits suppressor of cytokine
signaling 1 (SOCS1) expression, causing STAT1 and STAT3
phosphorylation in human hepatoma cells. Additionally,
miR‐155 overexpression partially inhibits HBx gene expression
in vitro [64]. Likewise, a study of 98 patients with HBV‐related
HCC revealed that higher expression of miR‐200c blocked
HBV‐mediated PD‐L1 expression by directly targeting the
3′UTR of PD‐L1 and resulted in a better prognosis. PD‐L1
expression correlates negatively with miR‐200c expression in
HBV‐related HCC cases [65].
In particular, the natural history of HBV infection in young
children frequently exhibits a shift from an acute to chronic
infection as the virus becomes dormant in the host cells: infected
hepatocytes. When HBV forms a covalently closed circular
DNA (cccDNA) in the nucleus of the host cells, it stably sur-
vives until the eventual reactivation of its life cycle, leading to
the chronic infection phase [66]. Some miRNAs have been
reported to be required for the chronic HBV infection.
Methylation of the CpG islands in the cccDNA by DNA meth-
yltransferase 1 (DNMT1) prevents viral gene expression.
DNMT1 was also identified as a direct target of miR‐152, whose
expression is frequently downregulated in HBV‐related HCCs
[67]. Thus, since the inhibition of miR‐152 causes global DNA
hypermethylation and increases the methylation levels of two
tumor suppressor genes, glutathione S‐transferase pi 1 (GSTP1)
and E‐cadherin 1 (CDH1), miR‐152 functions as a tumor sup-
pressor of the epigenetic aberrations associated with HBV‐
related HCC [67]. In addition, a computational analysis of the
HBV genome has identified seven sites that are potential targets
of human liver miRNAs. These miRNA target sites are located
in the cluster of a 995 bp segment within the viral polymerase
open reading frame (ORF) and the overlapping surface antigen
 16: 
miRNAs AND HEPATOCELLULAR CARCINOMA 189
ORF and are conserved among the most common HBV sub-
types [68]. The validation experiment using a luciferase reporter
assay identified a direct interaction between miR‐125a‐5p and
the HBV sequence that clearly inhibited the reporter activity of
the surface antigen [68].
Liver‐specific miRNAs also participate in the HCV infection
process. HCV is a single‐stranded RNA virus that infects hepat-
ocytes and develops persistent infections. The major miRNA in
the liver, miR‐122, is essential for HCV replication. The HCV
RNA genome comprises a 5′ UTR, a long polyprotein ORF, and
a 3′ UTR. A genetic interaction between miR‐122 and the 5′
UTR of the viral RNA genome was identified in a bioinformat-
ics search, based on an analysis of the predicted miRNA‐bind-
ing sites [57]. Notably, HCV RNA replicates in Huh‐7 cells
expressing miR‐122, but not in the HepG2 cells that do not
express miR‐122. When miR‐122 is inactivated with a 2′‐O‐
methylated RNA oligonucleotide, the expression of the HCV
replicon and core protein are significantly suppressed [57]; thus,
miR‐122 represents an attractive target for the development of
antiviral treatments.
miRNAs AS THERAPEUTIC STRATEGIES
FOR HCC
The biological significance of miRNAs in HCC initiation and
progression, as well as HBV and HCV viral replication, could
provide new opportunities for developing the therapeutic tar-
gets for HCC. When considering a therapeutic strategy for
HCC, liver cirrhosis offers a better window for therapeutics
because complete recovery of cirrhosis may prevent hepato-
carcinogenesis. All three members of the miR‐29‐family are
significantly downregulated in the cirrhotic liver and are asso-
ciated with TGF‐β‐mediated fibrosis [69]. Overexpression of
miR‐29b in mouse hepatic stellate cells (HSCs) decreases the
expression of α‐SMA, collagen I, and TIMP‐1. Overexpression
of miR‐29b in activated HSCs suppresses cell viability and
colony formation, and causes cell cycle arrest in G1 phase by
downregulating cyclin D1 and p21cip1. Therefore, the admin-
istration of miR‐29 might prevent hepatic fibrogenesis by tar-
geting activated HSCs in the liver [70]. On the other hand, the
HCV replication process is also targeted by miRNA‐based
therapeutics. Chronically infected chimpanzees were treated
with a locked nucleic acid (LNA)‐modified oligonucleotide
(SPC3649) complementary to miR‐122 and exhibited long‐
lasting suppression of HCV viremia, without side‐effects or
viral resistance in the chimpanzees. The prolonged effects of
SPC3649 are promising results for its use as an antiviral ther-
apy for HCV infection [43].
Regarding miRNA therapeutics for HCC, a number of target
miRNAs have been proposed in both experimental and preclini-
cal settings. Basically, two options for the target have been iden-
tified: tumor suppressor miRNAs and oncogenic miRNAs.
When a target miRNA is a tumor suppressor, its expression level
is generally decreased in HCC tissues and increased in normal
liver tissues. Thus, the administration of the miRNA into HCC
tissues may exert therapeutic effects. In contrast, when a target
miRNA is an oncogene, its expression level is increased in HCC
and thus the inhibition of the miRNA in HCC represents a
potential therapeutic strategy.
Kota and colleagues used miR‐26a as the target of HCC
therapy and reported the efficacy of an miRNA replacement in
HCC. The expression of miR‐26a is normally decreased in
HCC, and overexpression of miR‐26a induces cell cycle arrest
by directly targeting cyclins D2 and E2. The systemic adminis-
tration of miR‐26a with an adeno‐associated virus (AAV) inhib-
its cancer growth and induces tumor‐specific apoptosis in a
mouse model of HCC. The results of the delivery of miRNAs
provided a new therapeutic strategy for HCC [71]. For the thera-
peutics focusing on oncogenic miRNAs, one of the targets is
miR‐191. It has been reported as a therapeutic candidate miRNA
for HCC therapy. Indeed, miR‐191 expression is increased by
TCDD (2,3,7,8‐tetrachlorodibenzo‐p‐dioxin), a known liver
carcinogen, and it regulates a number of cancer‐related path-
ways. The administration of anti‐miR‐191 into the orthotopic
HCC xenograft inhibits the growth of tumor cells and reduces
the tumor mass [72]. Another target is miR‐21, which is overex-
pressed in HCC and designated as an onco‐miR. Wagenaar and
colleagues developed potent and specific single‐stranded oligo-
nucleotide inhibitors of miR‐21 (anti‐miR‐21). Treatment with
anti‐miR‐21 significantly decreases cell viability in the majority
of HCC cell lines by inducing apoptosis and necrosis. Similar to
the in vitro experiments, the effect of anti‐miR‐21 was con-
firmed in HCC tumor xenograft models [73]. Moreover, onco-
genic miRNAs of the miR‐17 family were also targeted for
pharmacological inhibition in an HCC model. Notably, miR‐17
itself was blocked by a tough decoy inhibitor in HCC cell lines,
leading to a global derepression of direct targets of miR‐17. A
lipid nanoparticle represented one of the most advanced plat-
forms for the systemic delivery of an oligonucleotide to tumor
tissues in vivo and was used to encapsulate a potent anti‐miR‐17
family oligonucleotide, which was designated as RL01‐17(5).
RL01‐17(5) was systemically administered to orthotopic HCC
xenografts, and significant tumor suppression was observed.
Thus, these findings potentially represent proof‐of‐concept evi-
dence for HCC therapies targeting the miR‐17 family [26].
Therapeutic strategies based on miRNAs or anti‐miRNAs
hold great promise due to their abilities to regulate a large num-
ber of genes and pathways. Prior to the therapeutic use of miR-
NAs or anti‐miRNAs in the clinic, target identification and
validation should be performed. Technological advances in the
in vivo delivery systems, such as nanoparticles, virus vectors,
and use of exosomes, will enable efficient and safe miRNA‐ or
anti‐miRNA‐based gene therapy in HCC.
EXOSOMAL miRNAs IN HCC
In 2007, an exemplary study by Valadi and colleagues reported
that miRNAs and mRNAs were packaged in extracellular vesi-
cles (EVs) [74]. These EVs are designated as exosomes,
microvesicles, apoptotic bodies, and other extracellular parti-
cles, according to their size, density, and secretion mechanism.
Secreted EVs contain cytoplasmic components; they can form
at the plasma membrane by direct budding into the extracellular
environment and produce large (100–1000 nm), irregularly
190 THE LIVER:  EXOSOMAL miRNAs IN HCC

shaped microvesicles. In contrast, another type of particle,
exosomes, are thought to be released from multivesicular bodies
(MVBs), whose size is 30–100 nm. EVs are secreted from cells
and transferred to another cell [74]. They are secreted from
almost all types of cell and play a key role in intercellular com-
munication by transporting RNA molecules and proteins
(Figure 16.2). Thus, EVs are a novel tool for the exchange of
genetic information between cells, such as miRNAs.
Notably, EVs have been detected in body fluids, including
blood, saliva, urine, and breast milk. Although RNases are pre-
sent in large amounts in body fluids such as the blood, circu-
lating extracellular RNAs are protected by the bilayer lipid
membrane of EVs. Prior to this research, miRNAs were con-
sidered to function intracellularly, but they are even active and
function in the extracellular space via EVs. Indeed, cancer
cell‐derived EVs containing miRNAs are transferred into
endothelial cells and regulate angiogenesis [75]. Currently,
many studies have proposed that RNAs, including miRNAs,
function as novel humoral factors in intercellular communica-
tion, and some have focused on the effect of EV‐derived miR-
NAs in cancer biology [76].
Cancer cells generally express oncogenic miRNAs at higher
levels and tumor‐suppressive miRNAs at lower levels, and EVs
from the cancer cells basically reflect the miRNA expression
profiles of the cells themselves. For example, p53‐mutated
colon cancer cells secrete oncogenic miR‐1246 via EVs. EVs
carrying miR‐1246 are delivered to the neighboring mac-
rophages for reprogramming into a cancer‐promoting state, cre-
ating a favorable microenvironment for the cancer cells [77]. In
contrast, miR‐143, which functions as a tumor suppressor, is
expressed at high levels in normal prostate cells, and the EVs
from normal prostate cells contain miR‐143. EVs from normal
prostate cells are transferred to cancer cells to block cancer cell
growth [78]. To date, a large number of studies have reported
that extracellular particles are an important source of miRNAs
in the circulation and suggest that EV‐derived miRNAs also
play key roles in carcinogenesis, including HCC initiation and
progression.
According to Kogure and colleagues, HCC cells secrete EVs
with distinct profiles of both RNAs and proteins compared with
their cells of origin. Similar to EVs derived from other cancer
cells, the HCC cell‐derived EVs were confirmed to be internal-
ized by other cells and modulate gene expression in recipient
cells [79]. Liver‐specific miR‐122, a tumor suppressor miRNA,
is also transferred via EVs between Huh7 and HepG2 human
hepatoma cells. In cell culture models, EV‐derived miR‐122
from Huh7 cells was internalized by miR‐122‐deficient HepG2
cells and subsequently downregulated miR‐122 target mRNAs,
indicating that intercellular communication via EVs occurred
between neighboring cells [80].
EVs secreted from HCC cells are also transferred to the sur-
rounding fibroblasts. Cancer‐associated fibroblasts (CAFs) are
well‐known to play a critical role in positively regulating the
tumor microenvironment. The miRNAs sequencing of CAF‐
derived exosomes from patients with HCC revealed a signifi-
cant loss of miR‐320a expression from CAF‐derived exosomes
[81]. An exogenous miRNA transfection experiment revealed
that CAFs transferred miRNAs to HCC cells. Notably,

Fibroblast
Conversion to CAFs
Macrophage, etc.

Block cell
growth

Endothelial cells
1. Increased vascular
EVs containing permeability
HCC cells miRNAs 2. Angiogenesis

Figure 16.2  Intercellular communication in hepatocellular carcinoma via extracellular vesicles. HCC‐derived EVs modulate the properties of
microenvironmental cells such as immune cells, endothelial cells, fibroblasts, epithelial cells, and mesenchymal stem cells. Cancer cells use EVs to
build a favorable environment to form tumors. For example, HCC‐derived EVs containing miRNAs control CAF conversion in an endocrine manner
[82], increase vascular permeability in a paracrine manner [83], and induce angiogenesis [84]. In contrast, HCC cells also receive EVs from
surrounding cells. Macrophage‐derived EVs containing miRNAs block the proliferation of HCC cells. EVs could work as an intercellular
communication tool between neighboring cells in an autocrine manner [80].
 16: 
miRNAs AND HEPATOCELLULAR CARCINOMA 191
miR‐320a functions as a tumor suppressor that inhibits HCC
cell proliferation, migration, and metastasis by directly target-
ing the PBX3 homeobox gene. Therefore, the transfer of normal
stromal cell‐derived miR‐320a via EVs is probably important
for the prevention of HCC development; however, CAF‐medi-
ated HCC tumor progression is at least partially caused by the
loss of the tumor‐suppressive miR‐320a from exosomes [81].
Since exosomes function locally and systemically, EV‐derived
miR‐1247–3p from highly metastatic HCC cells has been
described to regulate metastatic niche formation in the lung by
converting normal fibroblasts to CAFs [82]. EV‐derived miR‐
1247–3p suppresses B4GALT3 expression, resulting in the
activation of β1‐integrin–NF‐κB signaling in fibroblasts.
Interestingly, serum miR‐1247‐3p levels in EVs correlate with
lung metastasis in patients with HCC [82].
Furthermore, EV‐derived miR‐103 has been reported to cor-
relate with increased vascular permeability required for metas-
tasis. Deep sequencing and qRT‐PCR validation identified a
correlation between a high serum miR‐103 level and a high
metastatic capacity in patients with HCC. When endothelial
cells were retreated with EVs derived from hepatoma cells
expressing miR‐103 at high levels, their permeability signifi-
cantly increased, facilitating the transendothelial invasion of
tumor cells [83]. Cancer cell‐derived miR‐103 transfer decreases
the integrity of endothelial junctions by directly targeting VE‐
cadherin (VE‐Cad), p120‐catenin (p120) and zonula occludens
1 [83]. In addition, an evaluation of serum miRNA levels
revealed a high level of miR‐210‐3p in exosomes isolated from
the sera of patients with HCC, and the miR‐210 level positively
correlated with the microvessel density in HCC tissues [84]. By
directly targeting the SMAD4 and STAT6 genes, exosomal
miR‐210 enhanced angiogenesis, as confirmed by an in vitro
tubulogenesis assay using endothelial cells [84]. On the other
hand, macrophage‐derived EVs are taken up by HCC cells. Two
miRNAs, miR‐142 and miR‐223, are contained in EVs, and
transfer of these miRNAs contributes to the inhibition of the
proliferation of HCC cells by inhibiting the expression of stath-
min‐1 and insulin‐like growth factor‐1 receptor (IGF1R). Thus,
the transfer of EVs carrying specific miRNAs from immune
cells represents a potentially new defense system protecting
against HCC initiation and progression [85].
miRNAs AS DIAGNOSTIC BIOMARKERS
IN HCC
At the dawn of the use of miRNA profiling for cancer diagnos-
tics, whole transcriptomic analyses were applied to investigate
the molecular classification of the HCC tissues by determining
the clinical and genetic properties of the tumor [34]. Within the
last decade, a number of studies clearly indicated that miRNA
profiling provides a specific miRNA expression fingerprint of
tissues stratified according to malignancy, risk factors, and
oncogene/tumor suppressor gene alterations, displaying the
clinical and pathological features of HCC initiation and pro-
gression; thus, miRNA profiles could be exploited as potential
cancer biomarkers [86]. Because the miRNA signature likely
represents the HCC status, an understanding of miRNA profiles
would be useful for establishing new diagnostic and/or prognos-
tic tools for HCC. Although the molecular landscapes of tumors
were initially established using surgical or biopsy specimens,
more recently, an analysis of circulating nucleic acids, known as
a liquid biopsy strategy, has received increasing attention as a
diagnostic tool for solid tumors. The liquid biopsy has several
advantages; for example, it is less invasive and can be applied to
any methods, such as the detection of cell‐free DNA (cfDNA)
and miRNAs. A single biopsy provides spatially and temporally
limited information about the tumors; however, the liquid biopsy
provides the average genetic information and may be able to
reflect intratumor heterogeneity [87, 88].
Because the detection of tumor‐associated genomic DNA
mutations is useful for diagnostic decisions regarding treatment
strategies, such as the selection of molecular target drugs, the
methods for detecting cfDNA were established for preclinical
use (e.g., CancerSEEK, which distinguishes eight cancer types,
even at early stages) [89]. In addition to cfDNA, the liquid
biopsy approach is highly effective at detecting miRNAs, pro-
teins, and lipids packaged in EVs in the body fluids. An early
analysis of miRNAs detected in blood samples from patients
with HCC revealed that oncofetal miR‐500 is present at high
levels [90]. Therefore, the detection and/or quantification of
serum miRNA levels using liquid biopsy represents a poten-
tially feasible application for the diagnosis of HCC, assessments
of prognosis, surveillance of recurrent tumors, and prediction of
drug responses.
Deep sequencing followed by qRT‐PCR validation was
applied to identify serum biomarkers for HCC and identified
three miRNAs – miR‐25, miR‐375, and let‐7f – as biomarkers
for HCC. The receiver operating characteristic (ROC) curve
analysis of these miRNAs yielded an area under the ROC curve
(AUC) of 0.997, with a 97.9% sensitivity and 99.1% specificity
in distinguishing HCC cases from controls [91]. A logistic
regression model was established with seven miRNAs (miR‐122,
miR‐192, miR‐21, miR‐223, miR‐26a, miR‐27a, and miR‐801)
by randomly separating samples into a training cohort and a
validation cohort to discover a plasma miRNA combination that
discriminated HBV‐related HCC cases. The miRNA combina-
tion discriminated the HCC cases from healthy individuals
(AUC: 0.941), patients with chronic hepatitis B (AUC: 0.842),
and patients with cirrhosis (AUC: 0.884) [92]. Significantly
higher plasma levels of a single miRNA, miR‐21, were observed
in patients with HCC than in patients with chronic hepatitis and
healthy individuals. Interestingly, the ROC analysis of plasma
miR‐21 levels displayed an AUC of 0.773, with a 61.1% sensi-
tivity and 83.3% specificity in distinguishing HCC from chronic
hepatitis, and an AUC of 0.953 with an 87.3% sensitivity and
92.0% specificity in distinguishing HCC from healthy individu-
als [93]. These results are better than a conventional plasma
HCC biomarker, AFP, suggesting that miR‐21 is a promising
miRNA biomarker [93].
In 2015, Lin and colleagues reported a comprehensive study
of the serum miRNA profile for HCC diagnosis in five groups:
patients with HCC, patients with HBV‐related cirrhosis, patients
with chronic hepatitis (HBV), inactive HBsAg carriers, and
healthy individuals. Using six serum samples from patients with
HCC and eight serum samples from patients with chronic hepa-
titis (control), 19 highly expressed miRNAs were identified
192 THE LIVER:  REFERENCES
with the miRNA microarray. Using a combination of four mod-
els, such as a linear support vector machine, nonlinear support
vector machine, linear discriminant analysis, and logistical
regression, the authors developed an miRNA classifier (seven
miRNAs: miR‐29a, miR‐29c, miR‐133a, miR‐143, miR‐145,
miR‐192, and miR‐505) for detecting HCC [94]. This miRNA
classifier showed higher accuracy than the conventional AFP
method and detected small‐sized, early‐stage, and α‐fetopro-
tein‐negative HCC cases in at‐risk patients. Moreover, the circu-
lating miRNA signature was applied to estimate the risk of HCC
in patients with cirrhosis [95]. By comparing serum miRNA
profiles between 330 cirrhotic liver samples and 42 early‐stage
HCC samples, the calculated score based on the expression lev-
els of five miRNAs in serum predicted patients with cirrhosis
who were at high and low risk of developing HCC. The follow‐
up study (median period: 752 days) showed a good prediction
accuracy (AUC: 0.725, P < 0.001) [95].
As described above, miRNA levels in blood samples are valu-
able and useful for the preclinical detection of HCC, providing
patients with an opportunity to undergo curative resection and
achieve a better prognosis [94]. Recently, diagnostic models
based on serum miRNA levels have reported the successful dis-
crimination of specific cancer types from other types of cancers
[96] and the prediction of cancer metastasis [97], as well as the
differentiation of malignancy from benign tissues. Thus, the liq-
uid biopsy approach is a very promising method and a technol-
ogy that will likely be applied from the bench to the bedside in
the near future.
SUMMARY AND PERSPECTIVES
This chapter focused on the functions and potential clinical
applications, such as therapeutics and diagnostics, of miRNAs
in liver cancer. A brief overview of researchers’ progress in
improving our understanding of the function of miRNAs in hep-
atocarcinogenesis has also been provided. The research field of
miRNA biology in HCC has developed quickly and broadly
from basic studies to clinical studies. The importance of miR-
NAs in basic biology for the promotion and inhibition of HCC
development is based on the unique and fascinating features of
miRNAs, which influence thousands of genes and regulate a
variety of pathways. For instance, a liver‐specific miRNA,
miR‐122, constitutes approximately 70% of all the miRNA
molecules in the liver. As expected, miR‐122 is important for
liver development, and surprisingly is essential for the replica-
tion of HCV. In contrast, it functions as a tumor suppressor dur-
ing HCC development. Thus, one miRNA spatially and
temporally exerts multiple functions in different circumstances.
Due to their abilities to regulate complex gene networks, miR-
NAs represent potential therapeutic targets in HCC, as evi-
denced by findings from studies overexpressing miR‐26 and
inhibiting miR‐17 [71, 73]. The recent and most critical break-
through in the miRNA research field is the discovery of extra-
cellular RNAs, such as EV‐packaged miRNAs. The detection of
circulating miRNAs in the blood is performed using a microar-
ray or deep sequencing, resulting in the diagnosis of HCC. As
shown, studies with a large cohort have identified HCC
biomarkers. Nevertheless, the published data are still inconsist-
ent. Several explanations for this discrepancy are likely, such as
the use of different detection platforms, normalization methods,
sample preparation methods, and sample types (serum or
plasma), as well as variability in the cohort size. Therefore,
standardization is necessary to obtain a reliable liquid biopsy
test for HCC. Furthermore, EVs have the potential to serve as a
natural vector; thus, they could be applied to deliver specific
miRNA molecules, proteins, or drugs to the disease site. In the
next decade, we will look ahead to the future of miRNA research
towards clinical applications of circulating miRNAs as diagnos-
tic, prognostic, and therapeutic tools.
ACKNOWLEDGMENTS
This work was supported in part by the Japan Agency for
Medical Research and Development (AMED): P‐CREATE;
17cm0106217h0002 and Development and New Energy
and  Industrial Technology Development Organization;
16ae0101011h0003, a research grant from the Uehara Memorial
Foundation and a research grant from the Naito Foundation.
REFERENCES
1. Groszhans, H. and Filipowicz, W. Molecular biology: the expanding world
of small RNAs. Nature, 2008;451(7177):414.
2. Aravin, A.A., Hannon, G.J., and Brennecke, J. The piwi‐piRNA pathway
provides an adaptive defense in the transposon arms race. Science,
2007;318(5851):761–64.
3. Bartel, D.P. MicroRNAs: target recognition and regulatory functions. Cell,
2009;136:215–33.
4. Lim, L.P., Lau, N.C., Garrett‐Engele, P. et al. Microarray analysis shows that
some microRNAs downregulate large numbers of target mRNAs. Nature,
2005;433(7027):769.
5. Chang, J., Nicolas, E., Marks, D. et al. miR‐122, a mammalian liver‐specific
microRNA, is processed from hcr mRNA and may downregulate the high
affinity cationic amino acid transporter CAT‐1. RNA Biol, 2004;1(2):106–13.
6. Schratt, G.M., Tuebing, F., Nigh, E.A. et  al. A brain‐specific microRNA
regulates dendritic spine development. Nature, 2006;439(7074):283.
7. Chen, C.Z., Li, L., Lodish, H.F., and Bartel, D.P. MicroRNAs modulate
hematopoietic lineage differentiation. Science, 2004;303(5654):83–86.
8. Lu, J., Getz, G., Miska, E.A. et al. MicroRNA expression profiles classify
human cancers. Nature, 2005;435:834–38.
9. Voorhoeve, P.M. A genetic screen implicates miRNA‐372 and miRNA‐373
as oncogenes in testicular germ cell tumors. Cell, 2006;124:1169.
10. Mayr, C., Hemann, M.T., and Bartel, D.P. Disrupting the pairing between
let‐7 and Hmga2 enhances oncogenic transformation. Science, 2007;
315:1576.
11. Farh, K.K. Grimson, A., Jan, C. et al. The widespread impact of mammalian
microRNAs on mRNA repression and evolution. Science, 2005;310(5755):
1817–21.
12. Lee, Y., Jeon, K., Lee, J.T., Kim, S., and Kim, V.N. MicroRNA maturation:
stepwise processing and subcellular localization. EMBO J, 2002;21:
4663–70.
13. Truscott, M., Islam, A.B., and Frolov, M.V. Novel regulation and functional
interaction of polycistronic miRNAs. RNA, 2016;22:129–38.
14. Pasquinelli, A.E., Reinhart, B.J., Slack, F. et al. Conservation of the sequence
and temporal expression of let‐7 heterochronic regulatory RNA. Nature,
2000;408(6808):86.
15. Ruby, J.G., Jan, C.H., and Bartel, D.P. Intronic microRNA precursors that
bypass Drosha processing. Nature, 2007;448(7149):83.
 16: 
miRNAs AND HEPATOCELLULAR CARCINOMA 193
16. Gregory, R.I., Yan, K.P., Amuthan, G. et  al. The microprocessor complex
mediates the genesis of microRNAs. Nature, 2004;432(7014):235–40.
17. Lee, Y., Ahn, C., Han, J. et al. The nuclear RNaseIII Drosha initiates micro-
RNA processing. Nature, 2003;425:415–19.
18. Meister, G. and Tuschl, T. Mechanisms of gene silencing by double‐stranded
RNA. Nature, 2004;431:343.
19. Chen, P.Y. and Meister, G. microRNA‐guided posttranscriptional gene regu-
lation. Biol Chem, 2005;386(12):1205–18.
20. Song, J.J., Smith, S.K., Hannon, G.J., and Joshua‐Tor, L. Crystal structure of
argonaute and its implications for RISC slicer activity. Science, 2004;305: 1434.
21. Siegel, R.L., Miller, K.D., and Jemal, A. Cancer statistics 2017. CA Cancer
J Clin, 2017;67:7–30.
22. Forner, A., Llovet, J.M., and Bruix, J. Hepatocellular carcinoma. The Lancet,
2012;379:1245–55.
23. Oka, H., Tamori, A., Kuroki, T., Kobayashi, K., and Yamamoto, S.
Prospective study of alpha‐fetoprotein in cirrhotic patients monitored for
development of hepatocellular carcinoma. Hepatology, 1994;19(1):61–6.
24. Seo, S.I., Kim, H.S., Kim, W.J. et  al. Diagnostic value of PIVKA‐II and
alpha‐fetoprotein in hepatitis B virus‐associated hepatocellular carcinoma.
World J Gastroenterol, 2015;21:3928–35.
25. Meng, F., Henson, R., Wehbe‐Janek, H. et  al. MicroRNA‐21 regulates
expression of the PTEN tumor suppressor gene in human hepatocellular can-
cer. Gastroenterology, 2007;133:647–58.
26. Wagenaar, T.R., Zabludoff, S., Ahn, S.M. et  al. Anti‐miR‐21 suppresses
hepatocellular carcinoma growth via broad transcriptional network deregula-
tion. Mol Cancer Res, 2015;13:1009–21.
27. Xu, G, Zhang, Y., Wei, J. et al. MicroRNA‐21 promotes hepatocellular carci-
noma HepG2 cell proliferation through repression of mitogen‐activated pro-
tein kinase‐kinase 3. BMC Cancer, 2013;13:469.
28. Iliopoulos, D., Jaeger, S.A., Hirsch, H.A. et al. STAT3 activation of miR‐21
and miR‐181b‐1, via PTEN and CYLD, are part of the epigenetic switch
linking inflammation to cancer. Mol Cell, 2010;39:493–506.
29. Connolly, E., Melegari, M., Landgraf, P. et  al. Elevated expression of the
miR‐17–92 polycistron and miR‐21 in hepadnavirus‐associated hepatocel-
lular carcinoma contributes to the malignant phenotype. Am J Pathol,
2008;173:856–64.
30. Hayashita, Y., Osada, H., Tatematsu, Y. et  al. A polycistronic microRNA
cluster, miR‐17–92, is overexpressed in human lung cancers and enhances
cell proliferation. Cancer Res, 2005;65:9628–32.
31. Zhu, H., Han, C., and Wu, T. MiR‐17‐92 cluster promotes hepatocarcinogen-
esis. Carcinogenesis, 2015;36:1213–22.
32. Wong, Q.W., Lung, R.W., Law, P.T. et  al. MicroRNA‐223 is commonly
repressed in hepatocellular carcinoma and potentiates expression of
Stathmin1. Gastroenterology, 2008;135:257–69.
33. Varnholt, H., Drebber, U., Schulze, F. et al. MicroRNA gene expression pro-
file of hepatitis C virus‐associated hepatocellular carcinoma. Hepatology,
2008;47:1223–32.
34. Toffanin, S., Hoshida, Y., Lachenmayer, A. et al. MicroRNA‐based classifi-
cation of hepatocellular carcinoma and oncogenic role of miR‐517a.
Gastroenterology, 2011;140:1618–28.
35. Ma, S., Tang, K.H., Chan, Y.P. et  al. miR‐130b promotes CD133+ liver
tumor‐initiating cell growth and self‐renewal via tumor protein 53‐induced
nuclear protein 1. Cell Stem Cell, 2010;7: 694–707.
36. Yang, F., Yin, Y., Wang, F. et al. miR‐17‐5p promotes migration of human
hepatocellular carcinoma cells through the p38 mitogen‐activated protein
kinase‐heat shock protein 27 pathway. Hepatology, 2010;51:1614–23.
37. Ding, J., Huang, S., Wu, S. et al. Gain of miR‐151 on chromosome 8q24.3
facilitates tumour cell migration and spreading through downregulating
RhoGDIA. Nat Cell Biol, 2010;12:390–9.
38. Zhang, X., Liu, S., Hu, T. et al. Up‐regulated microRNA‐143 transcribed by
nuclear factor kappa B enhances hepatocarcinoma metastasis by repressing
fibronectin expression. Hepatology, 2009;50:490–9.
39. Wong, Q.W., Ching, A.K., Chan, A.W. et al. MiR‐222 overexpression con-
fers cell migratory advantages in hepatocellular carcinoma through enhanc-
ing AKT signaling. Clin Cancer Res, 2010;16:867–75.
40. Li, H., Zhou, Z.Q., Yang, Z.R. et  al. MicroRNA‐191 acts as a tumor pro-
moter by modulating the TET1‐p53 pathway in intrahepatic cholangiocarci-
noma. Hepatology, 2017;66:136–51.
41. Tsai, W.C., Hsu, P.W., Lai, T.C. et al. MicroRNA‐122, a tumor suppressor
microRNA that regulates intrahepatic metastasis of hepatocellular carci-
noma. Hepatology, 2009;49:1571–82.
42. Gramantieri, L., Ferracin, M., Fornari, F. et al. Cyclin G1 is a target of miR‐
122a, a microRNA frequently down‐regulated in human hepatocellular car-
cinoma. Cancer Res, 2007;67:6092–9.
43. Lanford, R.E., Hildebrandt‐Eriksen, E.S., Petri, A. et al. Therapeutic silenc-
ing of microRNA‐122 in primates with chronic hepatitis C virus infection.
Science, 2010;327:198–201.
44. Xu, H., He, J.H., Xiao, Z.D. et al. Liver‐enriched transcription factors regu-
late microRNA‐122 that targets CUTL1 during liver development.
Hepatology, 2010;52:1431–42.
45. Herath, N.I., Leggett, B.A., and MacDonald, G.A. Review of genetic and
epigenetic alterations in hepatocarcinogenesis. J Gastroenterol Hepatol,
2006;21:15–21.
46. Datta, J., Kutay, H., Nasser, M.W. et al. Methylation mediated silencing of
MicroRNA‐1 gene and its role in hepatocellular carcinogenesis. Cancer Res,
2008;68:5049–58.
47. Gailhouste, L., Gomez‐Santos, L., Hagiwara, K. et al. miR‐148a plays a piv-
otal role in the liver by promoting the hepatospecific phenotype and suppress-
ing the invasiveness of transformed cells. Hepatology, 2013;58:1153–65.
48. Liu, A.M., Poon, R.T., and Luk, J.M. MicroRNA‐375 targets Hippo‐signal-
ing effector YAP in liver cancer and inhibits tumor properties. Biochem
Biophys Res Commun, 2010;394:623–7.
49. Dang, Y., Luo, D., Rong, M. et al. Underexpression of miR‐34a in hepatocel-
lular carcinoma and its contribution towards enhancement of proliferating
inhibitory effects of agents targeting c‐MET. PLoS One, 2013;8:e61054.
50. Kim, T., Veronese, A., Pichiorri, F. et al. p53 regulates epithelial–mesenchy-
mal transition through microRNAs targeting ZEB1 and ZEB2. J Exp Med,
2011;208:875–83.
51. Li, N., Fu, H., Tie, Y. et  al. miR‐34a inhibits migration and invasion by
down‐regulation of c‐Met expression in human hepatocellular carcinoma
cells. Cancer Lett, 2009;275:44–53.
52. Salvi, A., Sabelli, C., Moncini, S. et al. MicroRNA‐23b mediates urokinase
and c‐met downmodulation and a decreased migration of human hepatocel-
lular carcinoma cells. FEBS J, 2009;276:2966–82.
53. Bai, S., Nasser, M.W., Wang, B. et al. MicroRNA‐122 inhibits tumorigenic
properties of hepatocellular carcinoma cells and sensitizes these cells to
sorafenib. J Biol Chem, 2009;284:32015–27.
54. Su, J., Wang, Q., Liu, Y. et al. miR‐217 inhibits invasion of hepatocellular
carcinoma cells through direct suppression of E2F3. Mol Cell Biochem,
2014;392:289–96.
55. Luo, C., Yin, D., Zhan, H. et al. MicroRNA‐501‐3p suppresses metastasis
and progression of hepatocellular carcinoma through targeting LIN7A. Cell
Death Dis, 2018;9:535.
56. Jin, W.B., Wu, F.L., Kong, D. et al. HBV‐encoded microRNA candidate and
its target. Comput Biol Chem, 2007;31:124–6.
57. Jopling, C.L., Yi, M., Lancaster, A.M. et al. Modulation of hepatitis C virus
RNA abundance by a liver‐specific microRNA. Science, 2005;309:1577–81.
58. Wang, S., Qiu, L., Yan, X. et  al. Loss of microRNA 122 expression in
patients with hepatitis B enhances hepatitis B virus replication through cyc-
lin G(1)‐modulated P53 activity. Hepatology, 2012;55:730–41.
59. Song, K., Han, C., Zhang, J. et al. Epigenetic regulation of MicroRNA‐122
by peroxisome proliferator activated receptor‐gamma and hepatitis b virus X
protein in hepatocellular carcinoma cells. Hepatology, 2013;58:1681–92.
60. Zhang, X., Zhang, E., Ma, Z. et al. Modulation of hepatitis B virus replica-
tion and hepatocyte differentiation by microRNA‐1. Hepatology,
2011;53:1476–85.
61. Naito, Y., Hamada‐Tsutsumi, S., Yamamoto, Y. et al. Screening of microR-
NAs for a repressor of hepatitis B virus replication. Oncotarget,
2018;9:29857–68.
62. Wang, Y., Lu, Y., Toh, S.T. et al. Lethal‐7 is down‐regulated by the hepatitis
B virus x protein and targets signal transducer and activator of transcription
3. J Hepatol, 2010;53:57–66.
63. Jin, J., Tang, S., Xia, L. et al. MicroRNA‐501 promotes HBV replication by
targeting HBXIP. Biochem Biophys Res Commun. 2013;430:1228–33.
64. Su, C., Hou, Z., Zhang, C. et  al. Ectopic expression of microRNA‐155
enhances innate antiviral immunity against HBV infection in human
hepatoma cells. Virol J, 2011;8:354.
65. Sun, C., Lan, P., Han, Q. et al. Oncofetal gene SALL4 reactivation by hepa-
titis B virus counteracts miR‐200c in PD‐L1‐induced T cell exhaustion. Nat
Commun, 2018;9:1241.
66. Ganem, D. and Prince, A.M. Hepatitis B virus infection – natural history and
clinical consequences. N Engl J Med, 2004;350:1118–29.
194 THE LIVER:  REFERENCES
67. Huang, J., Wang, Y., Guo, Y. et al. Down‐regulated microRNA‐152 induces
aberrant DNA methylation in hepatitis B virus‐related hepatocellular carci-
noma by targeting DNA methyltransferase 1. Hepatology, 2010;52:60–70.
68. Potenza, N., Papa, U., Mosca, N. et al. Human microRNA hsa‐miR‐125a‐5p
interferes with expression of hepatitis B virus surface antigen. Nucleic Acids
Res, 2011;39:5157–63.
69. Roderburg, C., Urban, G.W., Bettermann, K. et  al. Micro‐RNA profiling
reveals a role for miR‐29 in human and murine liver fibrosis. Hepatology,
2011;53:209–18.
70. Wang, J., Chu, E.S., Chen, H.Y. et al. MicroRNA‐29b prevents liver fibrosis
by attenuating hepatic stellate cell activation and inducing apoptosis through
targeting PI3K/AKT pathway. Oncotarget, 2015;6:7325–38.
71. Kota, J., Chivukula, R.R., O’Donnell, K.A. et  al. Therapeutic microRNA
delivery suppresses tumorigenesis in a murine liver cancer model. Cell,
2009;137:1005–17.
72. Elyakim, E., Sitbon, E., Faerman, A. et al. hsa‐miR‐191 is a candidate onco-
gene target for hepatocellular carcinoma therapy. Cancer Res, 2010;70:
8077–87.
73. Huang, X., Magnus, J., Kaimal, V. et al. Lipid nanoparticle‐mediated deliv-
ery of anti‐miR‐17 family oligonucleotide suppresses hepatocellular carci-
noma growth. Mol Cancer Ther, 2017;16:905–13.
74. Valadi, H., Ekström, K., Bossios, A. et  al. Exosome‐mediated transfer of
mRNAs and microRNAs is a novel mechanism of genetic exchange between
cells. Nat Cell Biol, 2007;9:654–9.
75. Kosaka, N., Iguchi, H., Hagiwara, K. et  al. Neutral sphingomyelinase 2
(nSMase2)‐dependent exosomal transfer of angiogenic microRNAs regulate
cancer cell metastasis. J Biol Chem, 2013;288:10849–59.
76. Kosaka, N., Yoshioka, Y., Fujita, Y. et al. Versatile roles of extracellular vesi-
cles in cancer. J Clin Invest, 2016;126:1163–72.
77. Cooks, T., Pateras, I.S., Jenkins, L.M. et al. Mutant p53 cancers reprogram
macrophages to tumor supporting macrophages via exosomal miR‐1246. Nat
Commun, 2018;9:771.
78. Kosaka, N., Iguchi, H., Yoshioka, Y. et al. Competitive interactions of cancer
cells and normal cells via secretory microRNAs. J Biol Chem,
2012;287:1397–405.
79. Kogure, T., Lin, W.L., Yan, I.K. et  al. Intercellular nanovesicle‐mediated
microRNA transfer: a mechanism of environmental modulation of hepato-
cellular cancer cell growth. Hepatology, 2011;54:1237–48.
80. Basu, S. and Bhattacharyya, S.N. Insulin‐like growth factor‐1 prevents
miR‐122 production in neighbouring cells to curtail its intercellular transfer
to ensure proliferation of human hepatoma cells. Nucleic Acids Res,
2014;42:7170–85.
81. Zhang, Z., Li, X., Sun, W. et al. (2017) Loss of exosomal miR‐320a from
cancer‐associated fibroblasts contributes to HCC proliferation and metasta-
sis. Cancer Lett, 2017;397:33–42.
82. Fang, T., Lv, H., Lv, G. et al. Tumor‐derived exosomal miR‐1247–3p induces
cancer‐associated fibroblast activation to foster lung metastasis of liver can-
cer. Nat Commun, 2018;9:191.
83. Fang, J.H., Zhang, Z.J., Shang, L.R. et al. Hepatoma cell‐secreted exosomal
microRNA‐103 increases vascular permeability and promotes metastasis by
targeting junction proteins. Hepatology, 2018;68:1459–75.
84. Lin, X.J., Fang, J.H., Yang, X.J. et al. Hepatocellular carcinoma cell‐secreted
exosomal microRNA‐210 promotes angiogenesis in vitro and in vivo. Mol
Ther Nucleic Acids, 2018;11:243–52.
85. Aucher, A., Rudnicka, D., and Davis, D.M. MicroRNAs transfer from human
macrophages to hepato‐carcinoma cells and inhibit proliferation. J Immunol,
2013;191:6250–60.
86. Ladeiro, Y., Couchy, G., Balabaud, C. et al. MicroRNA profiling in hepato-
cellular tumors is associated with clinical features and oncogene/tumor sup-
pressor gene mutations. Hepatology, 2008;47:1955–63.
87. Crowley, E., Di Nicolantonio, F., Loupakis, F., and Bardelli, A. Liquid
biopsy: monitoring cancer‐genetics in the blood. Nat Rev Clin Oncol,
2013;10:472–84.
88. Siravegna, G., Marsoni, S., Siena, S. et al. Integrating liquid biopsies into the
management of cancer. Nat Rev Clin Oncol, 2017;14:531–48.
89. Cohen, J.D., Li, L., Wang, Y. et al. Detection and localization of surgically
resectable cancers with a multi‐analyte blood test. Science, 2018;359:
926–30.
90. Yamamoto, Y., Kosaka, N., Tanaka, M. et  al. (2009) MicroRNA‐500 as a
potential diagnostic marker for hepatocellular carcinoma. Biomarkers,
2009;14:529–38.
91. Li, L.M., Hu, Z.B., Zhou, Z.X. et  al. Serum microRNA profiles serve as
novel biomarkers for HBV infection and diagnosis of HBV‐positive hepato-
carcinoma. Cancer Res, 2010;70:9798–807.
92. Zhou, J., Yu, L., Gao, X. et al. Plasma microRNA panel to diagnose hepatitis
B virus‐related hepatocellular carcinoma. J Clin Oncol, 2011;29:4781–8.
93. Tomimaru, Y., Eguchi, H., Nagano, H. et al. Circulating microRNA‐21 as a
novel biomarker for hepatocellular carcinoma. J Hepatol, 2012;56:167–75.
94. Lin, X.J., Chong, Y., Guo, Z.W. et al. A serum microRNA classifier for early
detection of hepatocellular carcinoma: a multicentre, retrospective, longitu-
dinal biomarker identification study with a nested case‐control study. Lancet
Oncol, 2015;16:804–15.
95. Huang, Y.H., Liang, K.H., Chien, R.N. et al. A circulating microRNA signa-
ture capable of assessing the risk of hepatocellular carcinoma in cirrhotic
patients. Sci Rep, 2017;7:523.
96. Yokoi, A., Matsuzaki, J., Yamamoto, Y. et al. Integrated extracellular micro-
RNA profiling for ovarian cancer screening. Nat Commun, 2018;9:4319.
97. Shiino, S., Matsuzaki, J., Shimomura, A. et al. Serum miRNA‐based predic-
tion of axillary lymph node metastasis in breast cancer. Clin Cancer Res,
2019;25(6):1817–27.
 Hepatocyte Apoptosis:
17 Mechanisms and Relevance
in Liver Diseases
Harmeet Malhi and Gregory J. Gores
College of Medicine, Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN, USA
INTRODUCTION
Apoptosis is a ubiquitous form of cell death occurring in human
liver diseases (Figure  17.1). It has historically been defined
morphologically by the presence of cytoplasmic shrinkage
(pyknosis), chromatin condensation, nuclear fragmentation
(karyorhexis), the presence of plasma membrane blebbing, and
the maintenance of an intact plasma membrane that retains its
integrity as the cell fragments into apoptotic bodies. Indeed,
apoptotic bodies were first described in the liver in patients with
yellow fever where they were referred to as Councilman bodies.
A more current biochemical description of apoptosis is caspase‐
dependent cell death [1]. Caspases are cysteine proteases that
cleave proteins at sites next to aspartic acid residues. The above
described apoptosis morphology is due to the cellular effects of
caspase activation. Caspase activation, and hence apoptosis, is a
highly regulated form of cell death, with multiple checkpoints
and molecular mediators, activated via two distinct pathways,
the extrinsic pathway and the intrinsic pathway. The extrinsic
pathway is initiated via death receptor activation and the intrin-
sic pathway by intracellular perturbations that result in caspase
activation (Figure  17.2). In hepatocytes, both pathways con-
verge on mitochondria.
Multiple intracellular molecules transmit apoptotic signals
and regulate the apoptotic signaling cascades, upstream and
downstream of mitochondria [2]. Mitochondrial permeabiliza-
tion is not only requisite but also sufficient for apoptosis; there-
fore, intracellular regulators downstream of mitochondrial
permeabilization such as caspase inhibitors cannot prevent cell
death [3]. Unlike developmental apoptosis, which is carefully
regulated in a spatiotemporal pattern and does not invoke
secondary events, pathologic apoptosis activates secondary
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
signaling events and can be massive, which can result in organ
failure. Secondary signaling events include pathologic apoptosis‐
induced tissue inflammation, injury, and fibrosis. For example,
in acute liver injury apoptosis is massive and correlates with
outcome (i.e. liver transplantation or death) [4]. In chronic liver
injury apoptosis is continuous, modulates the inflammatory
response and promotes fibrogenesis, resulting in cirrhosis [5, 6].
Hepatocyte apoptosis is evident in liver injury related to viral
hepatitis, metabolic diseases, alcoholic steatohepatitis, autoim-
mune hepatitis, and drug‐induced liver injury [7–9], emphasiz-
ing the shared pathogenic role of hepatocyte apoptosis in
liver injury from multiple, varied, acute, and chronic insults.
Apoptosis of other cellular compartments, such as sinusoidal
endothelial cells and stellate cells, also plays a role in liver
injury. Here we discuss the signaling mediators and regulators
of hepatocyte apoptosis and the inclusion of injury stimulus‐
specific information within each mechanism.
THE EXTRINSIC PATHWAY
Death receptors are cell surface transmembrane proteins that
belong to the tumor necrosis factor/nerve growth factor (TNF/
NGF) receptor superfamily, and are defined on the basis of
ligand specificity (i.e. their affinity for tumor necrosis factor
alpha (TNF‐α), Fas ligand (FasL), or tumor necrosis factor‐
related apoptosis‐inducing ligand (TRAIL) [10]. The extracel-
lular N‐terminal domain binds their respective ligands; there is
a membrane‐spanning region and then the intracellular C‐terminal
domain, which contains a conserved sequence known as the
death domain (DD). The ligand‐bound trimerized receptor
196 THE LIVER:  THE EXTRINSIC PATHWAY
(a)
(b)
Figure 17.1  Hepatocyte apoptosis in non‐alcoholic steatohepatitis.
(a) Photomicrograph of a hematoxylin and eosin‐stained liver section
from a dietary mouse model of obesity‐associated non‐alcoholic steato-
hepatitis, demonstrating the presence of hepatocyte apoptosis. The
black arrow points to an apoptotic hepatocyte. (b) The TUNEL stain
detects DNA strand breaks, a feature of cell death, in a corresponding
liver section from this mouse model of non‐alcoholic steatohepatitis.
complex brings together the DD, allowing recruitment of other
adaptor proteins to form a death‐inducing signaling complex
(DISC). For death signaling, Fas‐associated protein with death
domain (FADD) must be recruited to the DISC complex [10].
FADD contains a death effector domain (DED) through which it
binds inactive initiator caspases 8 and 10, in their procaspase
form. The procaspases form homodimers and undergo autopro-
teolytic cleavage with formation of active caspase 8 or 10 [11].
This process of activation of initiator caspases is referred to as
the induced‐proximity model of caspase activation. Activation
of initiator caspase 8 is essential for the transmission of apop-
totic signals from death receptors to mitochondria. This occurs
via the Bcl‐2 family protein Bid and is discussed below.
In hepatocytes, mitochondrial permeabilization with
amplification of the apoptotic cascade occurs in death recep-
tor‐initiated apoptosis. This involves release of mitochondrial
mediators of apoptosis which leads to the eventual activation
of the effector caspases 3 and 7, with positive feedback
amplification of caspase 8 activation. The requirement of
mitochondrial amplification categorizes hepatocytes as type
II cells, in contrast to type I cells in which caspase 8 or 10 can
directly activate caspase 3 and 7 without mitochondrial
involvement [12]. Caspase 8 proteolytically cleaves the proa-
poptotic BH3‐only protein of the Bcl‐2 family Bid to tBid
(truncated Bid), which leads to activation of Bax and Bak
(proapoptotic multidomain members of the Bcl‐2 family), and
pore formation in the outer mitochondrial membrane [13].
Multiple levels of signal transduction and amplification pre-
sent opportunities for regulation of death receptor‐mediated
apoptosis at many levels. Availability of cell surface receptor
and ligand is one level, for example the hepatocyte growth
factor (HGF) receptor Met associates with and regulates the
availability of Fas for binding its ligand [14]. Moreover, the
basal expression of death receptors in hepatocytes is low, and
known to increase in both acute and chronic liver diseases.
Cellular caspase 8 (FLICE)‐like inhibitory protein (cFLIP)
can inhibit cytotoxic signaling by death receptors [15]. cFLIP
is an enzymatically inactive homolog of caspase 8 with con-
served structural homology in the DED that allows binding to
FADD. This binding precludes maximal cellular activation of
caspase 8. Pro‐ and antiapoptotic members of the Bcl‐2 fam-
ily regulate the extrinsic pathway by modulating the ability of
tBid to activate Bax and Bak (discussed later) [15].
Tumor necrosis factor‐α
TNF‐α is a circulating cytokine, primarily produced by cells of
the immune system, including Kupffer cells in the liver, although
it can also be produced by other cell types, such as hepatocytes.
Hepatocytes express both tumor necrosis factor receptor 1
(TNFR1), a 55 kDa protein, and tumor necrosis factor receptor
2 (TNFR2), a 75 kDa protein, but their functional significances
differ [16]. TNFR1 is thought to mediate most of the biologic
effects of TNF‐α; it expresses a cytoplasmic death domain (DD)
and executes the apoptotic program by interacting with adaptor
proteins [17] (Figure 17.3). Ligand‐activated TNFR1 generates
sequential signaling events referred to as complex I and
complex II. On binding TNF‐α, TNFR1 recruits the adaptor
protein tumor necrosis factor receptor‐associated death domain
(TRADD). Signaling then proceeds in two steps. The first step,
or complex I, involves recruitment of tumor necrosis factor
receptor‐associated protein (TRAF‐2) and receptor‐interacting
protein 1 (RIP1), leading to rapid activation of nuclear factor κB
(NFκB) [18]. NFκB transcriptionally activates expression of
prosurvival (e.g. Bcl‐xL, A1, XIAP, and cFLIP) and proinflam-
matory genes (e.g. interleukin 6). Complex I also activates the
c‐Jun N‐terminal kinase (JNK) pathways. Complex I is internal-
ized with dissociation of TRAF2, RIP1, and TRADD from the
ligated receptor. TRADD then recruits FADD and procaspase 8
to initiate apoptotic signaling; this signaling pathway is referred
to as complex II. TRADD does not interact with TNFR2, nor
does FADD directly interact with TNFR1. Therefore, TNF‐α/
TNFR1 signaling first leads to NFκB‐mediated transcriptional
activation of prosurvival and proinflammatory genes prior to the
activation of apoptosis signaling. In cells resistant to NFκB, or
in the presence of a transcriptional inhibitor such as actinomy-
cin D, which prevents the synthesis of prosurvival proteins, the
apoptotic effect of TNF‐α is unmasked.
 17:  Hepatocyte Apoptosis: Mechanisms and Relevance in Liver Diseases 197

Figure 17.2  The extrinsic and intrinsic pathways of hepatocyte apoptosis. Mitochondrial permeabilization is required for hepatocyte apoptosis.
The extrinsic pathway is mediated by death receptors. Fas or TRAIL, upon ligation with their cognate receptors, activate events leading to mito-
chondrial permeabilization. The death‐inducing signaling complex is formed on the intracellular domain of ligated homotrimerized receptors in
conjunction with adaptor proteins, leading to caspase 8 activation, Bid cleavage, and activation of Bax and Bak. TNF‐α signaling pathway can
promote apoptosis by Bid‐induced lysosomal permeabilization. Intracellular perturbations such as ER stress, lysosomal permeabilization, or JNK
activate the intrinsic pathway of cell death. ER stress‐induced apoptosis is partly mediated by the transcription factor CHOP, which can upregulate
TRAIL‐R2 or Bim expression. JNK activation can be induced by TNF‐α, ER stress, or reactive oxygen species. These pathways are regulated by
the proapoptotic and antiapoptotic proteins of the Bcl‐2 family.

Necroptosis is a form of kinase‐regulated, caspase‐independ-
ent cell death best described following activation of TNFR1 by
TNF‐α, but characterized by morphology similar to necro-
sis – hence the name [19]. Necroptosis is known to occur only
under conditions of caspase 8 absence or inhibition and is medi-
ated by the receptor‐interacting protein (RIP) family kinases,
specifically RIP1 and RIP3. RIP1 has a DD and a caspase
recruitment domain which allows interactions at the DISC with
the adaptor protein TRADD. That is why RIP1 is present in
complex I. When caspase 8 is absent or inhibited, RIP1 is deu-
biquitinated and can interact with RIP3 to form the necrosome,
which also recruits mixed‐lineage kinase domain‐like pseudoki-
nase (MLKL), which is phosphorylated by RIP3 and translo-
cates to the plasma membrane where its trimerized form
mediates calcium influx and necroptosis [20]. When RIP1 is
polyubiquinated it promotes cell survival and inflammation.
The pathophysiologic contribution of necroptosis to the liver is
controversial due to the lack of RIP3 in normal hepatocytes.
Liver nonparenchymal cells and infiltrating immune cells
express RIP3, and this may account for some of the incongruous
observations in mouse models. However, RIP1 has many roles,
including kinase activity‐dependent and scaffolding function‐
dependent roles that can promote apoptosis or inflammation.
TNF‐α has pleiotropic effects in vivo, including hepatocyte
proliferation, liver inflammation, and modulation of hepatocyte
apoptosis. In a murine model of TNF‐α‐induced liver injury
(TNF‐α + d‐galactosamine), liver injury is Bax‐dependent [21].
TNF‐α‐associated caspase 8 activation can also cause lysosomal
permeabilization with release of intralysosomal cathepsin B
into the cytosol, which causes mitochondrial dysfunction [22].
Mice deficient in cathepsin B are protected from the injuri-
ous effects of TNF‐α [23]. c‐Jun N‐terminal kinase (JNK), a
stress‐activated kinase, is activated by TNF‐α. Sustained activa-
tion of JNK can lead to apoptosis by modulation of the Bcl‐2
family of proteins. JNK can also transcriptionally activate death
receptor expression (i.e. TRAIL receptor 2/death receptor 5).
Furthermore, JNK can promote TNF‐α‐induced apoptotic sign-
aling at complex II by facilitating degradation of cFLIP, thus
198 THE LIVER:  THE EXTRINSIC PATHWAY

Figure 17.3  Complex I and complex II of tumor necrosis factor alpha signaling. Tumor necrosis factor receptor 1 (TNFR1), upon binding TNF‐α
on its extracellular domain, activates complex I and complex II. Complex I is formed by the adaptor proteins TNFR1‐associated death domain
protein (TRADD) and receptor‐interacting protein (RIP), which recognize and bind via their death domains (DD) and TNF receptor‐associated
factor (TRAF2) via its kinase domain or an intermediate domain. Complex I mediates the activation of nuclear factor κB (NFκB) and transient c‐Jun
N‐terminal kinase (JNK) activation. NFκB translocates to the nucleus transcriptionally, activating antiapoptotic and inflammatory genes, such as
cellular FLICE‐like inhibitory protein (cFLIP), Bcl‐xL, Mcl‐1, A1, and XIAP, which regulate apoptosis at multiple levels. Sustained JNK activation
requires the adaptor protein RIP and is mediated in part by oxidative stress. Complex II is formed by receptor dissociation of TRADD, RIP, and
TRAF2 and ligand‐independent recruitment of Fas‐associated death domain (FADD) via its DD. FADD contains a death effector domain (DED),
leading to recruitment and activation of procaspase 8. In select conditions in the absence or inhibition of caspase 8 signaling, receptor‐interacting
protein 1 (RIP1) interacts with RIP3, leading to the recruitment and phosphorylation of the mixed‐lineage kinase domain‐like pseudokinase
(MLKL), which leads to necroptosis by translocating to the cell membrane.

antagonizing an antiapoptotic TNF‐α‐induced NFκB target
gene. Similarly, loss of cellular inhibitors of apoptosis proteins
1 and 2, also antiapoptotic NFκB target genes, sensitizes carci-
noma cells to TNF‐α‐mediated cytotoxicity [24]. TNF‐α can
lead to superoxide formation and caspase‐independent cell
death by TRADD and RIP1‐mediated activation of Nox1
NADPH oxidase leading to reactive oxygen species formation
[25]. This process is independent of FADD, and caspase 8 acti-
vation. Thus, a multitude of complex processes contribute to
TNF‐α cytotoxicity.
In experimental models of liver injury, a role for TNF‐α cell
death has been elucidated. Following partial hepatectomy,
massive hepatocyte cell death occurs after completion of cell
cycle progression due to sustained TNF‐α signaling in mice
lacking tissue inhibitor of metalloproteinase 3 (Timp3), a
model characterized by abnormal chronically elevated TNF‐α
activity [26]. In TNFR1‐deficient ethanol‐fed mice, hepato-
cyte apoptosis, serum alanine aminotransferase levels (ALT),
and inflammatory foci were decreased as compared to wild‐type
ethanol‐fed mice; TNFR2‐deficient mice developed liver
injury and apoptosis comparable to those in wild‐type controls
[27]. In ischemia reperfusion injury mice lacking TNFR1 and
treated with pentoxyfylline, a pharmacologic TNF‐α inhibitor,
liver injury and apoptosis are significantly reduced [28]. Liver
samples from patients with alcoholic steatohepatitis or non‐
alcoholic steatohepatitis demonstrate enhanced TNFR1
expression [29]. Serum levels of TNFR1 in patients with alco-
holic hepatitis are predictive of three‐month survival [30].
Thus, the TNF‐α cascade is activated in patients with many
liver diseases, including fulminant hepatic failure, alcoholic
steatohepatitis, non‐alcoholic steatohepatitis, chronic hepatitis
C, and chronic hepatitis B [29, 31, 32]; it is indeed a hallmark
of inflammatory changes in these conditions and likely con-
tributes to hepatocyte apoptosis in vivo. Our understanding of
why the TNFR1‐initiated NFκB cell survival pathways fail in
these diseases remains rudimentary.
 17:  Hepatocyte Apoptosis: Mechanisms and Relevance in Liver Diseases 199

Fas affect HGF binding to its receptor Met. Pretreatment of cells

Fas (also known as Apo‐1, CD95) is ubiquitously expressed in
liver cell types. Hepatocytes are exquisitely sensitive to Fas‐
induced apoptosis, and exogenously administered Fas agonistic
antibody results in fulminant hepatic failure in mice [33]. Fas
signaling usually results in hepatocyte apoptosis, although there
are reports of Fas‐induced proliferation of T cells and fibro-
blasts, chemokine secretion from macrophages, and Fas‐medi-
ated acceleration of liver regeneration after partial hepatectomy
in mice [34]. Fas–Fas ligand (FasL) binding leads to receptor
oligomerization, bringing together the intracellular DD, recruit-
ment of FADD, and procaspase 8 or 10 at the DISC (Figure 17.4).
This leads to activation and autoproteolytic activation of procas-
pase 8 or 10, generation of tBid, activation of Bax and Bak,
mitochondrial permeabilization with eventual activation of cas-
pase 3 and 7. Fas can be activated by soluble or circulating as
well as membrane‐bound FasL. FasL is expressed by cells of the
immune system, such as cytotoxic T lymphocytes (CTLs) and
natural killer (NK) cells [35]. The liver is enriched in both these
cell populations, therefore under constant “Fas attack.” However
Fas‐induced signaling is regulated at many levels. Cell surface
expression of Fas, levels of FasL, and cFLIP inhibition of cas-
pase 8 activation at the DISC are potential regulatory sites. Of
interest in hepatocytes is the sequestration of Fas by the hepato-
cyte growth factor receptor (HGF) Met [14]. Met–Fas com-
plexes prevent binding of FasL to Fas; however, Fas does not
with HGF releases Fas from this complex, and enhances FasL
binding and toxicity at lower concentrations of FasL. High con-
centrations of FasL are maximally toxic even in the absence of
HGF. Thus, the Met–Fas complex fine tunes and regulates the
biologic availability of Fas in hepatocytes. In embryonic hepat-
ocytes, Met prevents Fas‐induced cFLIP degradation, thus pre-
venting apoptosis.
In adult mice, genetic deficiency of Fas leads to hepatic
hyperplasia, in addition to enlargement of lymph nodes and
spleen [36]. The induction of fulminant hepatic failure in mice
by exogenous administration of Fas agonistic antibody is further
regulated by the Bcl‐2 family of proteins. It can be abrogated by
overexpression of Bcl‐2 and enhanced by genetic inhibition of
Bcl‐xL [37]. Genetic inhibition of Fas itself or Bid mitigates
liver injury by Fas agonists [38]. Neutralization of Fas reduces
warm ischemia/reperfusion‐related liver injury [39].
Circulating levels of serum Fas are elevated in patients with
fulminant hepatic failure [40]. Levels of serum Fas vary by eti-
ology, and the highest levels occur in patients with drug‐induced
liver injury. Fas expression and apoptosis are enhanced in liver
samples from patients with chronic hepatitis C [41]. Circulating
levels of soluble Fas correlate with histologic activity, and along
with levels of caspase 3 activity, are predictive of response to
therapy [42]. Similarly, in patients with chronic hepatitis B
hepatocyte Fas levels and circulating levels of soluble Fas are
elevated [41, 43]. Fas expression is enhanced in liver samples

Figure 17.4  Fas and TRAIL receptor signaling: Fas and TRAIL receptors are activated by ligand binding, which leads to receptor oligomeriza-
tion, bringing together their conserved death domains (DD). The adaptor protein Fas‐associated death domain (FADD) binds to the trimerized
intracellular DD and via its death effector domain (DED) leads to activation of procaspase 8. Active caspase 8 leads to proteolytic cleavage of Bid
to tBid and downstream mitochondrial permeabilization via activation of Bax and Bak. Mitochondrial permeabilization leads to release of the
contents of the intermembrane space including cytochrome c, smac/DIABLO, Apaf‐1, and endonuclease G, culminating in the activation of caspase
3/7 and cleavage of cellular proteins.
200 THE LIVER:  THE INTRINSIC PATHWAY
from patients with non‐alcoholic fatty liver disease [7]. In
experimental models of dietary and genetic fatty liver, steatotic
livers are sensitized to exogenous Fas administration. Indeed, in
patients with non‐alcoholic fatty liver disease, the inhibition of
Fas by Met is diminished, providing another mechanism to
explain the enhanced sensitivity to Fas‐induced hepatocyte
apoptosis. Furthermore, free fatty acid treatment can increase
Fas expression in vitro in cell culture models of hepatocyte stea-
tosis, sensitizing cells to Fas‐induced apoptosis. In the bile duct‐
ligated mouse model of cholestatic liver injury, hepatocyte
apoptosis is mediated by Fas, and Fas‐induced apoptosis pro-
motes hepatic fibrosis [44]. Toxic bile acids promote cell sur-
face expression of Fas, and can lead to ligand‐independent Fas
oligomerization and induction of hepatocyte apoptosis [45, 46].
In bile salt‐mediated ligand‐independent hepatocyte apoptosis,
Fas phosphorylation is required for its translocation to the cell
surface; this can occur in a Yes kinase/epidermal growth factor
receptor‐dependent and JNK‐dependent manner [47].
Tumor necrosis factor‐related apoptosis‐
inducing ligand
The role of tumor necrosis factor‐related apoptosis‐inducing
ligand (TRAIL, also known as Apo‐2 ligand) and its receptors in
liver disease is an area with remarkable recent advances. TRAIL
binds with several receptors [48]. TRAIL receptor 1 (TRAIL‐
R1/death receptor (DR) 4) and TRAIL receptor 2 (TRAIL‐R2/
DR5/killer/TRICK2) are complete receptors and can induce
apoptosis via caspase activation, similar to Fas [49]. This occurs
via the adaptor protein FADD, recruitment of procaspase 8 and
10 to the TRAIL receptor DISC, in a cFLIP‐regulated manner
(Figure  17.4). TRAIL receptor 3 (TRAIL‐R3/Apo‐3/TRAMP/
WSL‐1/LARD, decoy receptor 1 (DcR1)) and TRAIL receptor
4 (TRAIL‐R4, DR6, decoy receptor 2 (DcR2)) are incomplete
cell surface receptors and cannot stimulate apoptotic signaling.
Normal human hepatocytes, in situ and in vivo, are considered
resistant to TRAIL‐induced apoptosis, though there are occa-
sional reports of in vitro TRAIL‐induced hepatocyte apoptosis
[50]. This resistance to cell death may be secondary to cFLIP‐
induced inhibition of caspase 8 activation at the DISC or cell
surface expression/availability of TRAIL‐R1 or TRAIL‐R2.
However, diseased hepatocytes are sensitized to TRAIL‐induced
apoptosis [51]. TRAIL also sensitizes to Fas‐induced hepato-
cyte apoptosis by activating JNK and the proapoptotic BH3‐
only protein Bim.
TRAIL‐induced hepatocyte apoptosis has been demonstrated
in cholestatic, viral, and metabolic liver diseases. Toxic bile
acids transcriptionally regulate hepatocyte cell surface TRAIL‐
R2 expression in Fas‐deficient cells and inactivate cFLIP by
phosphorylation, thus dually sensitizing cells to TRAIL‐induced
apoptosis [52]. In the bile duct‐ligated mouse model of choles-
tasis, hepatocyte TRAIL‐R2 expression is enhanced and hepato-
cytes are sensitized to exogenously administered TRAIL [53].
By corollary, liver injury and hepatocyte apoptosis are signifi-
cantly reduced in TRAIL‐deficient mice following bile duct
ligation. Steatosis is also associated with increased hepatocyte
expression of TRAIL‐R2 and TRAIL‐R1, which imparts sensi-
tivity to TRAIL toxicity, and TRAIL or TRAIL receptor dele-
tion in mice improves dietary obesity‐associated fatty liver
disease due to a reduction in hepatocyte apoptosis [54, 55]. Free
fatty acids, which are elevated in the metabolic syndrome, tran-
scriptionally enhance TRAIL‐R2 expression in cell culture and
render steatotic cells sensitive to TRAIL toxicity [51]. In acute
hepatitis B‐induced liver failure in humans and experimental
adenoviral acute hepatitis in mice, TRAIL‐R2 expression is
enhanced, as is sensitivity to TRAIL. This occurs independently
of Kupffer cells and NK cells, suggesting a hepatocyte‐gener-
ated paracrine loop for elimination of virally infected cells [56].
Circulating soluble TRAIL levels are elevated in patients with
chronic viral hepatitis B. Hepatitis B X antigen increases
TRAIL‐R1 expression in cell culture experiments, conferring
sensitivity to TRAIL. In liver samples from patients with chronic
hepatitis C, TRAIL‐R1 and TRAIL‐R2 expression and TRAIL‐
induced apoptosis were enhanced [50]. Hepatitis C virus core
protein also selectively modulates cellular responsiveness to
TRAIL by promoting TRAIL‐induced Bid cleavage [57].
THE INTRINSIC PATHWAY
Intracellular stress leads to the activation of the intrinsic path-
way of apoptosis. Stress can be perceived and transduced by any
membrane‐defined organelle in the cell. For example, lys-
osomes can mediate steatotic liver cell death, as can the endo-
plasmic reticulum (ER). DNA damage can lead to genotoxic
stress and steatosis can activate JNK, also a mediator of the
intrinsic pathway of apoptosis. These processes converge on
mitochondria and are transduced by the Bcl‐2 family of pro-
teins, and so are usually referred to as the Bcl‐2‐regulated or
mitochondrial pathway of apoptosis. The Bcl‐2 family consists
of proapoptotic and antiapoptotic proteins, which are classified
into groups on the basis of the number of shared Bcl‐2 homol-
ogy (BH) domains. The proapoptotic proteins are structurally
divided based on the number of shared BH domains into multi-
domain (Bak and Bax, display BH1, 2, and 3 domains) and
BH3‐only proteins (Bid, Noxa, Puma, Bim, Bmf, Bik, Hrk, and
Bad). The antiapoptotic proteins include Bcl‐2, Bcl‐xL, Bcl‐w,
A1, Mcl‐1, and Boo, and share four BH domains with the excep-
tion of Mcl‐1, which shares three BH domains with the rest of
the antiapoptotic Bcl‐2 family members. The liver expresses
Bcl‐xL and Mcl‐1; Bcl‐2 is not expressed by hepatocytes. Bax
and Bak are both abundantly expressed by hepatocytes. The
antiapoptotic members of this family are located on the cyto-
plasmic aspect of membrane‐bound organelles, primarily the
mitochondria, though also on other organelles such as the ER.
They protect cells from death by preventing spontaneous oli-
gomerization of Bax and Bak, and may be necessary for survival
of certain cell types. Given this constitutive role in preventing
apoptosis, predictably, hepatocyte‐specific Bcl‐xL‐ or Mcl‐1‐
knockout mouse models demonstrate an increase in spontane-
ous hepatocyte apoptosis, liver injury, and fibrosis [58, 59]. Bax
and Bak are required for mitochondrial permeabilization, while
Bax is located in the cytosol and translocates to mitochondria
upon activation; Bak is a resident mitochondrial outer mem-
brane protein. The activation of Bax and Bak is regulated by
interactions between the antiapoptotic Bcl‐2 proteins and the
BH3‐only proapoptotic proteins. Several models have been
 17:  Hepatocyte Apoptosis: Mechanisms and Relevance in Liver Diseases 201
proposed to explain the biochemical activation of Bax or Bak by
proapoptotic BH3‐only proteins. Using Bim as an example,
upon activation, Bim is released from the dynein motor com-
plex, and can directly engage and activate Bax and Bak.
Alternatively, Bim can bind and negate the inhibitory effect of
Bcl‐2 or Bcl‐xL, releasing Bax and Bak from inhibition by these
proteins (the derepression model).
Although the large number of BH3‐only proteins imparts
redundancy, its primary effect is to impart stimulus specificity.
For example, free fatty acids activate Bim and Puma [60]; Puma
and Noxa are target genes of the tumor suppressor p53 [61].
Cytosolic Bid must be cleaved by active caspase 8 or 10 to
generate truncated Bid (tBid), which translocates to the mito-
chondria to activate Bak or Bax. Bim is sequestered by the
microtubule‐associated dynein motor complex in the cytosol,
from which it is dissociated by proapoptotic stimuli. For exam-
ple, JNK‐mediated Bim phosphorylation promotes its mito-
chondrial translocation. Bim is also regulated transcriptionally,
and known to be increased in ER stress‐induced apoptosis.
Cytosolic Bad is bound to the death inhibitory protein 14‐3‐3.
Activating and inhibitory phosphorylation events have been
described for both Bim and Bad.
Mitochondria
In addition to the metabolic functions of mitochondria, hepato-
cytes require mitochondria to die. The mitochondrial intermem-
brane space sequesters a number of proapoptotic proteins
including cytochrome c, SMAC/DIABLO (second mitochon-
drial activator of caspase/direct IAP‐binding protein with low
pI), HtrA2/Omi, AIF (apoptosis‐inducing factor), and endonu-
clease G [2, 15]. Current models suggest that active Bax or Bak
form pores in the outer mitochondrial membrane leading to
mitochondrial outer membrane permeabilization (MOMP) and
release of these mediators into the cytosol [62]. In the absence
of Bak and Bax, cells are resistant to apoptotic stimuli; thus,
Bak and Bax are critical effectors of MOMP. Though previously
thought to be a rapid and complete phenomenon wherein all
mitochondria within a cell would experience MOMP within
minutes, recent work has challenged this paradigm by describ-
ing partial MOMP in cells undergoing apoptotic stress. Since
partial MOMP does not result in apoptosis, this type of cellular
stress has been named sublethal stress. Inactive Bak and Bax
support mitochondrial fusion, whereas active Bak and Bax shift
mitochondrial dynamics toward fission. MOMP can also occur
secondary to permeability transition (PT) of the inner mitochon-
drial membrane (IMM) via the PT pore (PTP). The identity of
the PTP has received much attention and several candidate pro-
teins have been evaluated, including adenine nucleotide trans-
porter and the voltage‐dependent anion channel. Current
evidence suggests that the PTP is formed from FOF1 ATP syn-
thase dimers in the IMM [63], though these is some evidence to
suggest that spastic paraplegia 7 could also form the PTP.
Opening of the permeability transition pore leads to rapid fluxes
of ions and water, dissipation of the mitochondrial inner trans-
membrane potential, swelling of the mitochondria, and outer
mitochondrial membrane rupture leading to the release of the
contents into the intermembrane space. MOMP releases inter-
membrane contents into the cytosol and commits the cell to
apoptosis. SMAC inactivates post‐mitochondrial inhibitors of
apoptosis proteins (IAP). Cytosolic cytochrome c, apoptotic
protease‐activating factor 1 (Apaf) and ATP form a complex
called the apoptosome, leading to activation of procaspase 9 and
effector caspases 3 and 7 [64]. These effector caspases cleave
over 500 substrates resulting in cellular demolition.
Cytokeratin 18 is a structural protein expressed in most epi-
thelial cells that is cleaved by caspase 3 at aspartate positions
238 and 396. The fragment generated by this cleavage, cytoker-
atin 18–aspartate 396 (CK18‐asp396) forms a neo‐epitope that
is recognized by the M30 antibody. This neoepitope can be
detected in apoptotic tissues as well as serum by a commercially
available ELISA. Indeed circulating levels of CK18‐asp396 are
elevated in patients with liver injury and can correlate with
outcome [4]. Thus this biomarker presents a noninvasive, sim-
ple, and mechanistic tool to monitor progress and response to
therapy in liver injury.
Lysosomes
Lysosomes are intracellular organelles with acid intravesicular
pH that contain lysosomal proteases, known as cathepsins [65].
Cathepsin B and D, two of 11 known human cathepsins, are
stable and active at neutral pH. Methodical dissection of path-
ways that mediate intracellular death signals demonstrates that
lysosomes can be involved in the intrinsic pathway of cell death.
Typically, lysosomal permeabilization, when it mediates apop-
tosis, is selective and partial and is observed upstream of mito-
chondrial permeabilization. Cathepsin B‐induced mitochondrial
permeabilization can occur via caspase 2 (in mice) and via pro-
teolytic cleavage of Bid similar to death receptor‐induced acti-
vation of Bid [66, 67]. Indeed Bid also links death receptors to
lysosomal permeabilization; providing crosstalk between death
receptors and their engagement of the lysosomal and mitochon-
drial pathways [66]. Bax activation by intracellular stress can
also result in lysosomal permeabilization [68]. Cathepsin D lev-
els were elevated in serum from patients with fulminant hepatic
failure as well as chronic hepatitis [69, 70]. Cathepsin B‐defi-
cient mice are resistant to TNF‐α‐induced hepatocyte apoptosis
[23]. In models of cellular steatosis, cathepsin B inhibition pre-
vents mitochondrial permeabilization and apoptosis. In cathep-
sin B‐deficient mice liver apoptosis, injury, and fibrosis are
diminished following bile duct ligation [71]; liver apoptosis and
injury are abrogated in ischemia reperfusion injury as well [72].
Endoplasmic reticulum
The ER has an inbuilt mechanism to cope with excess or altered
unfolded proteins that serves to correct the inciting imbalance.
This process is termed the unfolded protein response (UPR).
The UPR can also be activated by stimuli that affect the function
of the ER, such as calcium depletion, glycosylation inhibition
(tunicamycin), ultraviolet radiation, and insulin resistance. The
ER stress response consists of a series of compensatory pro-
cesses to correct both the excess and the stress of the unfolded
proteins. Global translation is attenuated to reduce the func-
tional protein load of the ER. There is also selective translation
of UPR target genes aimed at protecting the ER [73, 74]. The
transducers of ER stress are membrane proteins that have an ER
202 THE LIVER:  THE INTRINSIC PATHWAY
luminal domain and a cytosolic domain. Inositol‐requiring pro-
tein 1 alpha (IRE1α) and protein kinase RNA‐like ER kinase
(PERK) auto‐transphosphorylate when released from the ER
chaperone BiP/Grp78. IRE1α possesses endoribonucleolytic
activity leading to excision of an intron within X‐box binding
protein 1 (XBP1) mRNA to generate spliced XBP1 (sXBP1), a
transcription factor that activates a subset of UPR target genes.
IRE1α also recruits TRAF2, leading to JNK activation. PERK
phosphorylates and inactivates the eukaryotic translation initia-
tion factor 2α (eIF2α), resulting in global translation attenuation
with selective translation of activating transcription factor 4
(ATF4), which leads to transcription of C/EBP‐homologous
protein (CHOP), and the ER chaperone BiP/Grp78. Activating
transcription factor 6 (ATF6) is cleaved within the ER mem-
brane, generating an ATF6 fragment that translocates to the
nucleus and activates a subset of UPR target genes. It is not
known if ATF6 also regulates apoptotic signaling.
ER stress also activates a negative feedback regulatory loop
that terminates the UPR; however in the setting of sustained ER
stress, proapoptotic signaling occurs [75]. Bax and Bak both
bind to the cytoplasmic domain of IRE1α, and in cells lacking
Bax and Bak, IRE1 stress‐generated JNK activation and XBP1
splicing are reduced [76], thus linking the core apoptotic
machinery to ER stress response. Bax and Bak localize on the
ER membrane, in addition to mitochondrial membranes. In cells
lacking both Bax and Bak, the ER is depleted of calcium and
unable to respond to certain death stimuli [77]. The proapop-
totic transcription factor CHOP can increase Bim expression
transcriptionally and by inhibiting its proteasomal degradation,
leading to Bim‐dependent ER stress‐induced apoptosis [78].
CHOP can also upregulate TRAIL‐R2 expression, sensitizing
cancer cells to TRAIL‐induced apoptosis [79].
The involvement of the ER stress‐induced apoptotic pathway
in liver diseases is an area of emerging research. In the bile duct‐
ligated mouse model of cholestasis, an early and transient
induction of CHOP expression is observed [80]. Mice deficient
in CHOP are protected from hepatocyte apoptosis, liver injury,
and liver fibrosis. In cell culture, the toxic bile acid glycocheno-
deoxycholic acid also induces ER stress and CHOP expression
in isolated rat hepatocytes [81]. In transgenic mice expressing
hepatitis C viral core and E2 proteins, hepatocyte apoptosis is
associated with CHOP expression [82]. Cycloheximide, an
inhibitor of protein synthesis, induces ER stress, induction of
CHOP expression and apoptotic hepatocyte cell death in rat liv-
ers [83]. In non‐alcoholic fatty liver disease, markers of ER
stress were variably activated [84]. Toxic saturated fatty acids
also induce ER stress and apoptosis in liver cell lines [85]. In a
mouse model of alcohol‐induced liver injury, CHOP‐deficient
mice are protected from hepatocyte apoptosis, though able to
mount an ER stress response [86].
c‐Jun N‐terminal kinase
Given the role of JNK in multiple models of cell death, it war-
rants a separate discussion as a final common cell death media-
tor. JNK1 and 2 are ubiquitously expressed, including liver,
whereas JNK3 is not expressed in the liver [87]. JNK activation
occurs downstream of kinase cascades that can be activated by
multiple stimuli including TNF‐α, IRE1α, reactive oxygen
species, free fatty acids, and bile acids [88, 89]. JNK involve-
ment in apoptosis is temporally regulated and stimulus specific
[90]. The same inciting stimulus (e.g. TNF‐α) can induce bipha-
sic JNK activation mediated by distinct intracellular pathways.
Transient and early JNK activation promotes survival; and sus-
tained and late activation of JNK promotes apoptosis [91]. In
the case of TNF‐α, production of reactive oxygen species medi-
ates the delayed and sustained activation of JNK. Other stimuli,
such as toxic free fatty acids, result in early and sustained JNK
activation, culminating in apoptotic signaling [92]. JNK‐stimu-
lated proapoptotic signaling converges on mitochondria via the
activation of Bax and Bak. In the absence of Bax and Bak, JNK‐
induced cell death is mitigated. Furthermore, mitochondrial per-
meabilization and release of cytochrome c are abolished in cells
derived from mice lacking Jnk1 and 2 genes, in response to
stimuli that cause intracellular stress [90]. JNK‐mediated
phosphorylation of pro‐ and antiapoptotic proteins upstream of
mitochondria also regulates apoptotic sensitivity. JNK can
phosphorylate and activate the BH3‐only proteins; for example,
Bim phosphorylation releases it from binding to the dynein
motor complex and promotes apoptosis [93]. Sustained JNK
activation promotes caspase 8 formation at the DISC by activa-
tion of the E3 ubiquitin ligase Itch, which ubiquinates and
degrades cFLIP, promoting liver cell death [94]. JNK can phos-
phorylate the antiapoptotic proteins Bcl‐2, Bxl‐xL, and Mcl‐1,
and the proapoptotic proteins Bmf and Bad.
JNK1 and JNK2 can both mediate liver injury in a stimulus‐
specific manner. In a murine model of steatohepatitis induced
by methionine‐ and choline‐deficient diet, JNK1 plays a pre-
dominant role. In high‐fat diet‐induced obesity and genetic
obesity in mice, JNK was activated; this was found to be pre-
dominantly JNK1, though JNK2 plays a role that is unmasked
in the absence of JNK1 [95]. In free fatty acid‐based cellular
models of hepatocyte steatosis, JNK2 is the predominant iso-
form that mediates apoptosis [92]. Oleic acid, a minimally
toxic free fatty acid, also sensitizes steatotic hepatocytes to
TRAIL‐induced apoptosis by JNK‐dependent transcriptional
upregulation of the death receptor TRAIL‐R2 [51]. This mech-
anism is shared by toxic bile acids, which also sensitize hepato-
cytes to TRAIL‐induced apoptosis by transcriptionally
activating TRAIL‐R2 expression in a JNK‐dependent manner
[53, 96]. Liver injury induced by ischemia reperfusion is also
mediated by JNK, and pharmacologic inhibition of JNK in
donor livers improved graft survival and decreased apoptosis
after orthotopic liver transplantation [97]. In acetaminophen‐
induced acute liver injury JNK activation was robust and sus-
tained, and led to Bax translocation to mitochondria and poor
animal survival. Pharmacologic inhibition of JNK decreased
liver injury and hepatocyte cell death and improved survival;
utilizing genetically deficient models of Jnk1 or Jnk2 it was
demonstrated that both mediate liver injury, though JNK2 was
predominant. JNK activation was observed in hepatocytes in
human liver samples from patients with acetaminophen‐
induced acute liver failure. JNK inhibition was more effective
in decreasing hepatocyte cell death than N‐acetylcysteine in a
murine model of acetaminophen‐induced liver injury. In a
murine model of TNF‐α‐induced liver injury utilizing galac-
tosamine and lipopolysaccharide, JNK2 mediated caspase 8
activation and mitochondrial permeabilization.
 17:  Hepatocyte Apoptosis: Mechanisms and Relevance in Liver Diseases 203
THE CONSEQUENCES OF HEPATOCYTE
APOPTOSIS
Apoptosis, inflammation, and injury are in some ways insepara-
ble, and it is difficult sometimes to dissect the primary event.
However, based on the inciting stimulus, apoptosis or inflam-
matory signaling may be the primary event, each stimulating the
other. The liver has a large population of Kupffer cells, NK
cells, and NK T cells. These cells are a ready source of TNF‐α
and other cytokines that mediate inflammation, such as Fas,
TRAIL, and TNF‐α that mediate hepatocyte apoptosis and
transforming growth factor beta (TGF‐β) that activates stellate
cells. Apoptotic hepatocytes can be engulfed by Kupffer cells,
leading to generation of cytokines [6]. In keeping with this the
pharmacologic inhibition of apoptosis prevents Kupffer cell
activation. Also, in the bile duct‐ligated mouse, Kupffer cell
depletion decreases hepatocyte apoptosis, liver injury, and liver
inflammation. In addition, stressed hepatocytes increase expression
of NKG2D ligands, thus inviting NK‐ and NKT cell‐mediated
destruction.
Fibrosis is the hallmark of ongoing liver injury. Hepatic stel-
late cells mediate hepatic fibrosis. In the normal liver, stellate
cells maintain a quiescent phenotype. On activation, they
undergo a metamorphosis to become myofibroblasts, secreting
collagen which leads to liver fibrosis. Stellate cells in vitro can
engulf apoptotic hepatocytes, leading to their activation, and
increased expression of TGF‐β, alpha smooth muscle actin, and
collagen alpha 1 [98]. Similarly, in vivo hepatocyte apoptosis is
a fibrogenic stimulus. Several experimental studies have dem-
onstrated that the inhibition of hepatocyte apoptosis abrogates
liver fibrosis [5, 71, 99]. By corollary, apoptosis of activated
stellate cells should decrease liver fibrosis and dissociate ongo-
ing hepatocyte apoptosis from the ensuing fibrogenic response.
Indeed, activated stellate cells are sensitized to apoptotic signaling.
This can be achieved by inhibition of NFκB, TRAIL‐mediated
stellate cell apoptosis, and NK cell‐mediated stellate cell apop-
tosis. Indeed, the resolution phase of fibrosis requires apoptosis
of activated hepatic stellate cells.
Lastly, the clinical applications of apoptosis are discussed in
the conclusion of this chapter. The cytokeratin 18‐derived M30
neoantigen reflects epithelial cell apoptosis, is abundant in
hepatocytes, can easily be measured in serum by a commercially
available ELISA, and correlates with hepatocyte apoptosis in
diverse liver diseases. In a study with a small number of patients
with chronic hepatitis C, pretreatment M30 levels were predic-
tive of response to therapy, inferring from this that patients with
an apoptotic response to virally infected hepatocytes are more
likely to have a treatment response. In another study with chronic
hepatitis C patients with normal transaminases, serum M30 lev-
els correlated with fibrosis. In patients with non‐alcoholic fatty
liver disease, serum M30 levels offer reliable discrimination of
patients with steatohepatitis from simple steatosis, and increas-
ing levels are predictive of a higher likelihood of inflammation.
Caspase inhibitors have demonstrated efficacy in preventing
hepatocyte apoptosis and injury in experimental models of liver
injury [99, 100]. In patients with chronic hepatitis C, orally
administered caspase inhibitor was found to be safe, and lowered
transaminases. The caspase inhibitor Emricasan (IDUN‐6556) is
currently in clinical trials for non‐alcoholic steatohepatitis.
In conclusion, hepatocyte apoptosis is a key mediator of liver
injury and inflammation in most forms of liver disease. Multiple
apoptotic pathways are activated by a given injurious stimulus
in a vulnerable hepatocyte. The predominant signaling pathway
that results in mitochondrial dysfunction in a given cell is diffi-
cult to discern; however, multiple pathways could potentially
cooperate or oppose each other, to eventually result in mito-
chondrial permeabilization. Once mitochondrial permeabiliza-
tion occurs, the hepatocyte is committed to cell death. Evidence
of hepatocyte apoptosis can be demonstrated by serum markers
and early studies demonstrate prognostic significance of apop-
tosis markers. Lastly, therapeutic manipulation of apoptosis is
of benefit in preventing liver injury and fibrosis.
ACKNOWLEDGMENTS
Support for this work is provided by NIH grant DK 41876
(GJG), DK 111378 (HM), and the Mayo Foundation.
REFERENCES
1. Malhi, H., Gores, G.J., and Lemasters, J.J. Apoptosis and necrosis in the
liver: a tale of two deaths? Hepatology, 2006;43(2 Suppl 1):S31–44.
2. Green, D.R. and Kroemer, G. The pathophysiology of mitochondrial cell
death. Science, 2004;305(5684):626–9.
3. Xiang, J., Chao, D.T., and Korsmeyer, S.J. BAX‐induced cell death may not
require interleukin 1 beta‐converting enzyme‐like proteases. Proc Natl Acad
Sci U S A, 1996;93(25):14559–63.
4. Rutherford, A.E., Hynan, L.S., Borges, C.B. et al. Serum apoptosis markers
in acute liver failure: a pilot study. Clin Gastroenterol Hepatol, 2007;
5(12):1477–83.
5. Canbay, A., Higuchi, H., Bronk, S.F. et al. Fas enhances fibrogenesis in the
bile duct ligated mouse: a link between apoptosis and fibrosis. Gastroenterology,
2002;123(4):1323–30.
6. Canbay, A., Friedman, S., and Gores, G.J. Apoptosis: the nexus of liver
injury and fibrosis. Hepatology, 2004;39(2):273–8.
7. Feldstein, A.E., Canbay, A., Angulo, P. et al. Hepatocyte apoptosis and fas
expression are prominent features of human nonalcoholic steatohepatitis.
Gastroenterology, 2003;125(2):437–43.
8. Natori, S., Rust, C., Stadheim, L.M. et al. Hepatocyte apoptosis is a patho-
logic feature of human alcoholic hepatitis. J Hepatol, 2001;34(2):248–53.
9. Kohli, V., Selzner, M., Madden, J.F., Bentley, R.C., and Clavien, P.A.
Endothelial cell and hepatocyte deaths occur by apoptosis after ischemia‐
reperfusion injury in the rat liver. Transplantation, 1999;67(8):1099–105.
10. Guicciardi, M.E. and Gores, G.J. Life and death by death receptors. FASEB
J. 2009;23(6):1625–37.
11. Muzio, M., Stockwell, B.R., Stennicke, H.R., Salvesen, G.S., and Dixit,
V.M. An induced proximity model for caspase‐8 activation. J Biol Chem.
1998;273(5):2926–30.
12. Scaffidi, C., Fulda, S., Srinivasan, A. et al. Two CD95 (APO‐1/Fas) signaling
pathways. EMBO J, 1998;17(6):1675–87.
13. Yin, X.M. Bid, a critical mediator for apoptosis induced by the activation of
Fas/TNF‐R1 death receptors in hepatocytes. J Mol Med (Berl), 2000
;78(4):203–11.
14. Wang, X., DeFrances, M.C., Dai, Y. et  al. A mechanism of cell survival:
sequestration of Fas by the HGF receptor Met. Mol Cell, 2002;9(2):411–21.
15. Danial, N.N. and Korsmeyer, S.J. Cell death: critical control points. Cell,
2004;116(2):205–19.
16. Yamada, Y., Webber, E.M., Kirillova, I., Peschon, J.J., and Fausto, N.
Analysis of liver regeneration in mice lacking type 1 or type 2 tumor necrosis
factor receptor: requirement for type 1 but not type 2 receptor. Hepatology,
1998;28(4):959–70.
204 THE LIVER:  REFERENCES
17. Tartaglia, L.A., Ayres, T.M., Wong, G.H., and Goeddel, D.V. A novel domain
within the 55 kd TNF receptor signals cell death. Cell, 1993;74(5):845–53.
18. Micheau, O. and Tschopp, J. Induction of TNF receptor I‐mediated apopto-
sis via two sequential signaling complexes. Cell, 2003;114(2):181–90.
19. Grootjans, S., Vanden Berghe, T., and Vandenabeele, P. Initiation and execu-
tion mechanisms of necroptosis: an overview. Cell Death Differ,
2017;24(7):1184–95.
20. Cai, Z., Jitkaew, S., Zhao, J. et al. Plasma membrane translocation of trimer-
ized MLKL protein is required for TNF‐induced necroptosis. Nat Cell Biol,
2014;16(1):55–65.
21. Sass, G., Shembade, N.D., Haimerl, F. et  al. TNF pretreatment interferes
with mitochondrial apoptosis in the mouse liver by A20‐mediated down‐
regulation of Bax. J Immunol, 2007;179(10):7042–9.
22. Guicciardi, M.E., Deussing, J., Miyoshi, H. et al. Cathepsin B contributes to
TNF‐alpha‐mediated hepatocyte apoptosis by promoting mitochondrial
release of cytochrome c. J Clin Invest, 2000;106(9):1127–37.
23. Guicciardi, M.E., Miyoshi, H., Bronk, S.F., and Gores, G.J. Cathepsin B
knockout mice are resistant to tumor necrosis factor‐alpha‐mediated hepato-
cyte apoptosis and liver injury: implications for therapeutic applications. Am
J Pathol, 2001;159(6):2045–54.
24. Varfolomeev, E., Blankenship, J.W., Wayson, S.M. et  al. IAP antagonists
induce autoubiquitination of c‐IAPs, NF‐kappaB activation, and TNFalpha‐
dependent apoptosis. Cell, 2007;131(4):669–81.
25. Kim, Y.S., Morgan, M.J., Choksi, S., and Liu, Z.G. TNF‐induced activation
of the Nox1 NADPH oxidase and its role in the induction of necrotic cell
death. Mol Cell, 2007;26(5):675–87.
26. Mohammed, F.F., Smookler, D.S., Taylor, S.E. et al. Abnormal TNF activity
in Timp3‐/‐ mice leads to chronic hepatic inflammation and failure of liver
regeneration. Nat Genet, 2004;36(9):969–77.
27. Yin, M., Wheeler, M.D., Kono, H. et al. Essential role of tumor necrosis fac-
tor alpha in alcohol‐induced liver injury in mice. Gastroenterology,
1999;117(4):942–52.
28. Rudiger, H.A. and Clavien, P.A. Tumor necrosis factor alpha, but not Fas,
mediates hepatocellular apoptosis in the murine ischemic liver.
Gastroenterology, 2002;122(1):202–10.
29. Ribeiro, P.S., Cortez‐Pinto, H., Sola, S. et al. Hepatocyte apoptosis, expres-
sion of death receptors, and activation of NF‐kappaB in the liver of nonalco-
holic and alcoholic steatohepatitis patients. Am J Gastroenterol,
2004;99(9):1708–17.
30. Spahr, L., Giostra, E., Frossard, J.L. et al. Soluble TNF‐R1, but not tumor
necrosis factor alpha, predicts the 3‐month mortality in patients with alco-
holic hepatitis. J Hepatol, 2004;41(2):229–34.
31. Torre, F., Rossol, S., Pelli, N. et al. Kinetics of soluble tumour necrosis factor
(TNF)‐alpha receptors and cytokines in the early phase of treatment for chronic
hepatitis C: comparison between interferon (IFN)‐alpha alone, IFN‐alpha plus
amantadine or plus ribavirin. Clin Exp Immunol, 2004; 136(3):507–12.
32. Fang, J.W., Shen, W.W., Meager, A., and Lau, J.Y. Activation of the tumor
necrosis factor‐alpha system in the liver in chronic hepatitis B virus infec-
tion. Am J Gastroenterol, 1996;91(4):748–53.
33. Ogasawara, J., Watanabe‐Fukunaga, R., Adachi, M. et al. Lethal effect of the
anti‐Fas antibody in mice. Nature, 1993;364(6440):806–9.
34. Desbarats, J. and Newell, M.K. Fas engagement accelerates liver regenera-
tion after partial hepatectomy. Nat Med, 2000;6(8):920–3.
35. Berke, G. The CTl’s kiss of death. Cell, 1995;81(1):9–12.
36. Adachi, M., Suematsu, S., Kondo, T. et al. Targeted mutation in the Fas gene
causes hyperplasia in peripheral lymphoid organs and liver. Nat Genet,
1995;11(3):294–300.
37. Lacronique, V., Mignon, A., Fabre, M. et  al. Bcl‐2 protects from lethal
hepatic apoptosis induced by an anti‐Fas antibody in mice. Nat Med,
1996;2(1):80–6.
38. Zhang, H., Cook, J., Nickel, J. et al. Reduction of liver Fas expression by an
antisense oligonucleotide protects mice from fulminant hepatitis. Nat
Biotechnol, 2000;18(8):862–7.
39. Al‐Saeedi, M., Steinebrunner, N., Kudsi, H. et al. Neutralization of CD95
ligand protects the liver against ischemia‐reperfusion injury and prevents
acute liver failure. Cell Death Dis, 2018;9(2):132.
40. Ryo, K., Kamogawa, Y., Ikeda, I. et al. Significance of Fas antigen‐mediated
apoptosis in human fulminant hepatic failure. Am J Gastroenterol,
2000;95(8):2047–55.
41. Kiyici, M., Gurel, S., Budak, F. et  al. Fas antigen (CD95) expression and
apoptosis in hepatocytes of patients with chronic viral hepatitis. Eur J
Gastroenterol Hepatol, 2003;15(10):1079–84.
42. Toyoda, M., Kakizaki, S., Horiguchi, N. et al. Role of serum soluble Fas/
soluble Fas ligand and TNF‐alpha on response to interferon‐alpha therapy in
chronic hepatitis C. Liver, 2000;20(4):305–11.
43. Song, le H., Binh, V.Q., Duy, D.N. et al. Variations in the serum concentra-
tions of soluble Fas and soluble Fas ligand in Vietnamese patients infected
with hepatitis B virus. J Med Virol, 2004;73(2):244–9.
44. Miyoshi, H., Rust, C., Roberts, P.J., Burgart, L.J., and Gores, G.J. Hepatocyte
apoptosis after bile duct ligation in the mouse involves Fas. Gastroenterology,
1999;117(3):669–77.
45. Sodeman, T., Bronk, S.F., Roberts, P.J., Miyoshi, H., and Gores, G.J. Bile
salts mediate hepatocyte apoptosis by increasing cell surface trafficking of
Fas. Am J Physiol Gastrointest Liver Physiol, 2000;278(6):G992–9.
46. Faubion, W.A., Guicciardi, M.E., Miyoshi, H. et al. Toxic bile salts induce
rodent hepatocyte apoptosis via direct activation of Fas. J Clin Invest,
1999;103(1):137–45.
47. Reinehr, R., Becker, S., Wettstein, M., and Haussinger, D. Involvement of the
Src family kinase yes in bile salt‐induced apoptosis. Gastroenterology,
2004;127(5):1540–57.
48. Kimberley, F.C. and Screaton, G.R. Following a TRAIL: update on a ligand
and its five receptors. Cell Res, 2004;14(5):359–72.
49. Schneider, P., Thome, M., Burns, K. et al. TRAIL receptors 1 (DR4) and 2
(DR5) signal FADD‐dependent apoptosis and activate NF‐kappaB.
Immunity, 1997;7(6):831–6.
50. Volkmann, X., Fischer, U., Bahr, M.J. et  al. Increased hepatotoxicity of
tumor necrosis factor‐related apoptosis‐inducing ligand in diseased human
liver. Hepatology, 2007;46(5):1498–508.
51. Malhi, H., Barreyro, F.J., Isomoto, H., Bronk, S.F., and Gores, G.J. Free fatty
acids sensitise hepatocytes to TRAIL mediated cytotoxicity. Gut,
2007;56(8):1124–31.
52. Higuchi, H., Yoon, J.H., Grambihler, A. et  al. Bile acids stimulate cFLIP
phosphorylation enhancing TRAIL‐mediated apoptosis. J Biol Chem,
2003;278(1):454–61.
53. Higuchi, H., Bronk, S.F., Taniai, M., Canbay, A., and Gores, G.J. Cholestasis
increases tumor necrosis factor‐related apoptotis‐inducing ligand (TRAIL)‐
R2/DR5 expression and sensitizes the liver to TRAIL‐mediated cytotoxicity.
J Pharmacol Exp Ther, 2002;303(2):461–7.
54. Idrissova, L., Malhi, H., Werneburg, N.W. et al. TRAIL receptor deletion in
mice suppresses the inflammation of nutrient excess. J Hepatol,
2015;62(5):1156–63.
55. Hirsova, P., Weng, P., Salim, W. et al. TRAIL deletion prevents liver inflam-
mation but not adipose tissue inflammation during murine diet‐induced obe-
sity. Hepatol Commun, 2017;1(7):648–62.
56. Mundt, B., Kuhnel, F., Zender, L. et al. Involvement of TRAIL and its recep-
tors in viral hepatitis. FASEB J, 2003;17(1):94–6.
57. Chou, A.H., Tsai, H.F., Wu, Y.Y. et  al. Hepatitis C virus core protein
modulates TRAIL‐mediated apoptosis by enhancing Bid cleavage and acti-
vation of mitochondria apoptosis signaling pathway. J Immunol, 2005;
174(4):2160–6.
58. Takehara, T., Tatsumi, T., Suzuki, T. et al. Hepatocyte‐specific disruption of
Bcl‐xL leads to continuous hepatocyte apoptosis and liver fibrotic responses.
Gastroenterology, 2004;127(4):1189–97.
59. Weber, A., Boger, R., Vick, B. et  al. Hepatocyte‐specific deletion of the
antiapoptotic protein myeloid cell leukemia‐1 triggers proliferation and
hepatocarcinogenesis in mice. Hepatology, 2010;51(4):1226–36.
60. Barreyro, F.J., Kobayashi, S., Bronk, S.F. et al. Transcriptional regulation of
Bim by Fox03A mediates hepatocyte lipoapoptosis. J Biol Chem,
2007;282(37):27141–54.
61. Yu, J. and Zhang, L. The transcriptional targets of p53 in apoptosis control.
Biochem Biophys Res Commun, 2005;331(3):851–8.
62. Kalkavan, H. and Green, D.R. MOMP, cell suicide as a BCL‐2 family busi-
ness. Cell Death Differ, 2018;25(1):46–55.
63. Giorgio, V., von Stockum, S., Antoniel, M. et al. Dimers of mitochondrial
ATP synthase form the permeability transition pore. Proc Natl Acad Sci U S
A, 2013;110(15):5887–92.
64. Riedl, S.J. and Salvesen, G.S. The apoptosome: signalling platform of cell
death. Nat Rev Mol Cell Biol, 2007;8(5):405–13.
65. Guicciardi, M.E., Leist, M., and Gores, G.J. Lysosomes in cell death.
Oncogene, 2004;23(16):2881–90.
66. Guicciardi, M.E., Bronk, S.F., Werneburg, N.W., Yin, X.M., and Gores, G.J.
Bid is upstream of lysosome‐mediated caspase 2 activation in tumor necro-
sis factor alpha‐induced hepatocyte apoptosis. Gastroenterology, 2005;
129(1):269–84.
 17:  Hepatocyte Apoptosis: Mechanisms and Relevance in Liver Diseases 205
67. Stoka, V., Turk, B., Schendel, S.L. et  al. Lysosomal protease pathways to
apoptosis. Cleavage of bid, not pro‐caspases, is the most likely route. J Biol
Chem, 2001;276(5):3149–57.
68. Feldstein, A.E., Werneburg, N.W., Li, Z., Bronk, S.F., and Gores, G.J.
Bax inhibition protects against free fatty acid‐induced lysosomal perme-
abilization. Am J Physiol Gastrointest Liver Physiol, 2006;290(6):
G1339–46.
69. Gove, C.D., Wardle, E.N., and Williams, R. Circulating lysosomal enzymes
and acute hepatic necrosis. J Clin Pathol, 1981;34(1):13–16.
70. Kyaw, A., Aung, T., Htut, T., Myint, H., and Tin, K.M. Lysosomal enzyme
activities in normals and in patients with chronic liver diseases. Clin Chim
Acta, 1983;131(3):317–23.
71. Canbay, A., Guicciardi, M.E., Higuchi, H. et  al. Cathepsin B inactivation
attenuates hepatic injury and fibrosis during cholestasis. J Clin Invest,
2003;112(2):152–9.
72. Baskin‐Bey, E.S., Canbay, A., Bronk, S.F. et  al. Cathepsin B inactivation
attenuates hepatocyte apoptosis and liver damage in steatotic livers after cold
ischemia‐warm reperfusion injury. Am J Physiol Gastrointest Liver Physiol,
2005;288(2):G396–402.
73. Ron, D. and Walter, P. Signal integration in the endoplasmic reticulum
unfolded protein response. Nat Rev Mol Cell Biol, 2007;8(7):519–29.
74. Malhi, H. and Kaufman, R.J. Endoplasmic reticulum stress in liver disease.
J Hepatol, 2011;54(4):795–809.
75. Lin, J.H., Li, H., Yasumura, D. et al. IRE1 signaling affects cell fate during
the unfolded protein response. Science, 2007;318(5852):944–9.
76. Hetz, C., Bernasconi, P., Fisher, J. et al. Proapoptotic BAX and BAK modu-
late the unfolded protein response by a direct interaction with IRE1alpha.
Science, 2006;312(5773):572–6.
77. Scorrano, L., Oakes, S.A., Opferman, J.T. et al. BAX and BAK regulation of
endoplasmic reticulum Ca2+: a control point for apoptosis. Science,
2003;300(5616):135–9.
78. Puthalakath, H., O’Reilly, L.A., Gunn, P. et al. ER stress triggers apoptosis
by activating BH3‐only protein Bim. Cell, 2007;129(7):1337–49.
79. He, Q., Luo, X., Jin, W. et  al. Celecoxib and a novel COX‐2 inhibitor
ON09310 upregulate death receptor 5 expression via GADD153/CHOP.
Oncogene, 2008;27(18):2656–60.
80. Tamaki, N., Hatano, E., Taura, K. et al. CHOP deficiency attenuates choles-
tasis‐induced liver fibrosis by reduction of hepatocyte injury. Am J Physiol
Gastrointest Liver Physiol, 2008;294(2):G498–505.
81. Tsuchiya, S., Tsuji, M., Morio, Y., and Oguchi, K. Involvement of endoplas-
mic reticulum in glycochenodeoxycholic acid‐induced apoptosis in rat
hepatocytes. Toxicol Lett, 2006;166(2):140–9.
82. Tumurbaatar, B., Sun, Y., Chan, T., and Sun, J. Cre‐estrogen receptor‐medi-
ated hepatitis C virus structural protein expression in mice. J Virol Methods,
2007;146(1–2):5–13.
83. Ito, K., Kiyosawa, N., Kumagai, K. et al. Molecular mechanism investiga-
tion of cycloheximide‐induced hepatocyte apoptosis in rat livers by morpho-
logical and microarray analysis. Toxicology, 2006;219(1–3):175–86.
84. Puri, P., Mirshahi, F., Cheung, O. et al. Activation and dysregulation of the
unfolded protein response in nonalcoholic fatty liver disease.
Gastroenterology, 2008;134(2):568–76.
85. Wei, Y., Wang, D., Topczewski, F., and Pagliassotti, M.J. Saturated fatty
acids induce endoplasmic reticulum stress and apoptosis independently of
ceramide in liver cells. Am J Physiol Endocrinol Metab, 2006;
291(2):E275–81.
86. Ji, C., Mehrian‐Shai, R., Chan, C., Hsu, Y.H., and Kaplowitz, N. Role of
CHOP in hepatic apoptosis in the murine model of intragastric ethanol
feeding. Alcohol Clin Exp Res, 2005;29(8):1496–503.
87. Czaja, M.J. The future of GI and liver research: editorial perspectives. III.
JNK/AP‐1 regulation of hepatocyte death. Am J Physiol Gastrointest Liver
Physiol, 2003;284(6):G875–9.
88. Urano, F., Wang, X., Bertolotti, A. et al. Coupling of stress in the ER to
activation of JNK protein kinases by transmembrane protein kinase IRE1.
Science, 2000;287(5453):664–6.
89. Ueda, S., Masutani, H., Nakamura, H. et al. Redox control of cell death.
Antioxid Redox Signal, 2002;4(3):405–14.
90. Tournier, C., Hess, P., Yang, D.D. et  al. Requirement of JNK for stress‐
induced activation of the cytochrome c‐mediated death pathway. Science,
2000;288(5467):870–4.
91. Ventura, J.J., Hubner, A., Zhang, C. et al. Chemical genetic analysis of the
time course of signal transduction by JNK. Mol Cell, 2006;21(5):701–10.
92. Malhi, H., Bronk, S.F., Werneburg, N.W., and Gores, G.J. Free fatty acids
induce JNK‐dependent hepatocyte lipoapoptosis. J Biol Chem,
2006;281(17):12093–101.
93. Lei, K. and Davis, R.J. JNK phosphorylation of Bim‐related members of
the Bc12 family induces Bax‐dependent apoptosis. Proc Natl Acad Sci U S
A, 2003;100(5):2432–7.
94. Chang, L., Kamata, H., Solinas, G. et al. The E3 ubiquitin ligase itch cou-
ples JNK activation to TNFalpha‐induced cell death by inducing c‐FLIP(L)
turnover. Cell, 2006;124(3):601–13.
95. Hirosumi, J., Tuncman, G., Chang, L. et al. A central role for JNK in obe-
sity and insulin resistance. Nature, 2002;420(6913):333–6.
96. Higuchi, H., Bronk, S.F., Takikawa, Y. et al. The bile acid glycochenode-
oxycholate induces trail‐receptor 2/DR5 expression and apoptosis. J Biol
Chem, 2001;276(42):38610–18.
97. Uehara, T., Xi Peng, X., Bennett, B. et al. c‐Jun N‐terminal kinase mediates
hepatic injury after rat liver transplantation. Transplantation, 2004;
78(3):324–32.
98. Canbay, A., Taimr, P., Torok, N. et  al. Apoptotic body engulfment by a
human stellate cell line is profibrogenic. Lab Invest, 2003;83(5):655–63.
99. Canbay, A., Feldstein, A., Baskin‐Bey, E., Bronk, S.F., and Gores, G.J. The
caspase inhibitor IDN‐6556 attenuates hepatic injury and fibrosis in the bile
duct ligated mouse. J Pharmacol Exp Ther, 2004;308(3):1191–6.
100. Natori, S., Higuchi, H., Contreras, P., and Gores, G.J. The caspase inhibitor
IDN‐6556 prevents caspase activation and apoptosis in sinusoidal endothelial
cells during liver preservation injury. Liver Transpl, 2003;9(3):278–84.
SECTION B:
THE HEPATOCYTE
 Copper Metabolism
18 and the Liver
Cynthia Abou Zeid, Ling Yi, and Stephen G. Kaler
Section on Translational Neuroscience, Molecular Medicine Branch, Intramural Research Program, National
Institutes of Health, Bethesda, MD, USA
OVERVIEW
Among its several roles in the human body, the liver is a piv-
otal organ involved in metabolizing and excreting various
toxic substances into the bile. One such function is elimination
of trace metals, such as copper, that are necessary for several
physiological processes but are also toxic if present in excess.
Copper is a micronutrient acquired from the diet and plays a
critical role in human growth and development. It is a divalent
metal ion that can be found in both the cuprous (Cu+) and
cupric (Cu2+) states. Its capacity to readily gain and donate
electrons renders it important in various metabolic pathways.
Copper is used as a cofactor for enzymes involved in the mito-
chondrial electron transport chain, detoxification of reactive
oxygen species, iron transport, connective tissue metabolism,
melanin pigment production, synthesis of amidated neuropep-
tides, and catecholamine metabolism [1–5]. However, the
chemical properties of copper also expose to the risk of oxida-
tive damage when homeostatic mechanisms are impaired.
Consequently, tissue copper levels are tightly regulated by sev-
eral transporters and chaperone proteins involved in systemic
copper metabolism [3, 6]. The main route of systemic copper
regulation is considered to be biliary excretion. Emerging data
support regulation at the intestinal absorption level, but these
mechanisms are not completely elucidated [7–9]. In all cases,
the liver is still counted as the major organ of copper metabo-
lism, since it functions not only in the elimination of excessive
copper, but also in its distribution to several destinations fol-
lowing intestinal uptake.
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
THE LIVER: A CENTRAL ORGAN
OF COPPER METABOLISM
Dietary copper absorption primarily occurs in the duodenum.
The process begins with uptake of reduced copper via the apical
membrane of the enterocyte, through the copper transporter
Ctr1 [10, 11]. DMT1 is a less‐specific divalent metal transporter
that also imports iron, manganese, and nickel, and can play an
additional role in copper uptake [12]. Once inside the entero-
cyte, copper is shuttled by its chaperone ATOX1 to the copper‐
transporting ATPase ATP7A. The latter is responsible for copper
export via the enterocyte’s basolateral pole into the portal circu-
lation (Figure  18.1). ATP7A has ubiquitous distribution and
important roles in several cell types [13]. In hepatocytes, ATP7A
has higher expression in the neonatal period with a reduction
over time [14]. In the liver, the predominant copper‐transporting
ATPase is ATP7B, a protein closely related to ATP7A with simi-
larities in its structure and function [15]. The general role of
these pumps depends on copper levels in the cell. In physiologi-
cal copper states, ATP7A and ATP7B localize to the trans‐Golgi
network and are responsible for conveying copper into the
secretory pathway, for metalation of copper‐dependent enzymes.
In case of elevated intracellular levels of copper, these transport-
ers traffic to the plasma membrane to export copper to the extra-
cellular milieu. The latter mechanism is how ATP7A works in
intestinal absorption of copper, by sending the metal from the
enterocyte into the portal venous circulation [16, 17].
The copper pool in blood is then bound to serum proteins such
as albumin and α2‐macroglobulin [18] and directed to the liver.
210 THE LIVER:  COPPER, THE LIVER, AND DISEASE

Figure 18.1  Copper metabolism in enterocytes and hepatocytes. (a) In enterocytes, copper (Cu+) uptake is mediated by CTR1, possibly in concert
with the metalloreductases STEAP1, STEAP2, and dCYTB. The roles of CTR2 (not shown) and DMT1 in this process are less certain. Within
enterocytes, GSH and MT function in copper sequestration and storage. The chaperones CCS, ATOX1, Cox17, Cox11, and Sco1 ferry copper to
specific proteins or organelles. With an increase in copper levels, ATP7A traffics to the basolateral surface and pumps copper into the blood. (b) In
hepatocytes, ATOX1 provides copper to ATP7B for metalation of ceruloplasmin (CP), and traffics to the apical membrane to pump copper into the
bile, the body’s major mechanism for copper removal. CCS has been proposed to deliver copper to XIAP, which may interact with COMMD1, a
protein mutated in hepatic copper toxicosis of Bedlington terriers and which may modulate ATP7B activity. This putative pathway is denoted by
dashed lines. Modified from [3] with permission of Springer Nature.

Copper is imported through the basolateral membrane of hepato-
cytes by Ctr1 and DMT1 (Figure 18.1) [3]. For release into the
bloodstream, copper needs to be bound to proteins, the main spe-
cies in blood being ceruloplasmin, a secreted glycoprotein that
acts as a ferroxidase and functions in systemic iron metabolism
[19]. It is synthetized in the liver, and is initially devoid of copper
(apoceruloplasmin) [20]. In the trans‐Golgi network of hepato-
cytes, copper is loaded onto apoceruloplasmin by ATP7B and
subsequently secreted into the blood circulation. When liver cop-
per levels increase, ATP7B trafficks to the apical membrane of
hepatocytes, from where it excretes excess copper into the biliary
canaliculi. Through biliary excretion, the liver is the major organ
responsible for copper level regulation by elimination of the
metal. Renal excretion of copper also occurs, and becomes clini-
cally significant only in the presence of pathological mecha-
nisms of copper overload from metabolic disorders [3], or in
conditions resulting in biliary obstruction [21–24].
The liver is therefore the central organ in copper metabolism,
which explains how dysfunction in one can impact the other.
Disorders of copper metabolism manifest in the liver with clini-
cal, laboratory, and pathological signs of hepatic dysfunction, as
discussed below.
COPPER, THE LIVER, AND DISEASE
Copper metabolism disorders involving
the liver
Wilson disease
Wilson disease was first described in 1912 by Samuel Alexander
Kinnier Wilson as a familial disorder associating liver cir-
rhosis with neurological manifestations, also referred to as
hepato‐lenticular degeneration [25]. It was not until three
decades later that elevated copper levels and accumulation in
the brain and liver were identified, which led to the classifica-
tion of this disease as a disorder of copper metabolism [26].
This biochemical phenotyping facilitated discovery of the
putative role of ATP7B in Wilson disease, supported by the
similarities with ATP7A [27].
Wilson disease is an autosomal recessive disorder caused by
a dysfunction of the copper transporter ATP7B, and more than
600 disease‐causing mutations have been identified to date [27,
28]. The distribution and frequency of specific mutations vary
depending on the geographic region and population studied
[29–32]. The types of amino acid changes most frequently
found are missense mutations, small insertions or deletions, and
splice site junction mutations [33]. Pathological variations in
the sequence of this copper transporter cause copper overload
and injury in the hepatocytes. Excessive copper eventually spills
into the blood circulation and accumulates in several extrahe-
patic tissues, namely the brain, kidneys, and cornea, which
helps explain the signs and symptoms of this disease [34].
Wilson disease classically presents in individuals ranging in
age from 3 to 50 years, with a variable combination of neuro-
logical, psychiatric, and hepatic manifestations [35, 36]. Clinical
signs and symptoms vary greatly among affected individuals,
even within a single family [30]. Liver disease classically occurs
in Wilson disease individuals who are symptomatic and diag-
nosed at an early age from childhood through early adulthood.
Neurological and psychiatric manifestations are more frequent
in older individuals, where hepatic involvement may be less
pronounced [37]. In addition, very late onset (after age 70)
and very early onset (nine months of age) of Wilson disease
have been reported [38, 39]. Hepatic presentations can be both
acute or chronic. Acute Wilsonian liver disease manifests as
a rapidly occurring jaundice from a hepatitis‐like injury or a
 18:  Copper Metabolism and the Liver 211
Coombs‐negative hemolytic disease and progresses to fulmi-
nant liver failure. The illness may also present as chronic liver
disease, with or without cirrhosis. Especially since there are
treatments available, Wilson disease should be considered in the
differential diagnosis of any unexplained chronic liver disease
or new‐onset neuropsychiatric symptoms.
Psychiatric manifestations of Wilson disease include mood
and personality disorders and sometimes cognitive deteriora-
tion. Neurological presentations are dominated by extrapyrami-
dal symptoms including tremor, chorea, and choreoathetosis,
problems in coordination and fine motor control, as well as
rigidity and gait disturbance. From the ophthalmological stand-
point, copper accumulation in the cornea may form a brown
pathognomonic circle called the Kayser–Fleischer ring visible
on slit lamp examination [30] (Figure  18.2). If clearly docu-
mented, this finding can confirm the diagnosis in a patient sus-
pected of having Wilson disease.
Confirmatory laboratory tests in Wilson disease include low
levels of serum copper and ceruloplasmin that result from the
inability of mutated ATP7B to load copper onto apoceruloplas-
min, and increased 24‐hour urinary copper excretion. If liver
biopsy is obtained during diagnostic workup of unexplained
liver disease, hepatic copper levels are usually markedly ele-
vated. A scoring system for the diagnosis of Wilson disease
based on both clinical and laboratory criteria is available to
guide clinicians faced with a possible case of this disorder [40].
Definitive confirmation is obtained by molecular analysis of the
ATP7B gene.
Treatment of Wilson disease should begin as soon as the
diagnosis is made to prevent evolution to fulminant liver dis-
ease, cirrhosis, or irreversible neuropsychiatric damage. Therapy
in symptomatic patients is based on the administration of copper
chelators such as d‐penicillamine or trientine to prevent its
accumulation in organs. Zinc salts can also be used as means to
inhibit copper absorption in the intestines. Oral zinc induces
Figure 18.2  Kayser–Fleischer ring in the cornea of an adult patient
with Wilson disease, resulting from copper deposition in Descemet’s
membrane. Image courtesy of T.U. Hoogenraad, M.D., Department of
Neurology, University Hospital, Utrecht, The Netherlands. Reproduced
from Kaler, Wilson disease, in Cecil’s Textbook of Medicine, 23rd edn.
(eds. L. Goldman and D. Ausiello), Saunders, Philadelphia, 2008, ch.
230, pp. 1593–5 with permission of Elsevier.
expression of metallothionein, an intracellular metal chelator
that binds copper with an even higher affinity than zinc, and
impairs its absorption. Pharmacotherapy using copper chelators
is ideally continued as a lifelong treatment. However, side
effects such as dermatological changes, hypersensitivity
­reactions, autoimmune disease, bone marrow suppression, and
nephrotoxicity can be an obstacle for treatment adherence.
Resistant cases that do not respond to pharmacological therapy
and life‐threatening cases of Wilson disease may be treated with
liver transplantation. A liver transplant that replaces the major
dysfunctional organ is usually curative [41, 42].
A number of new therapies for Wilson disease are currently
in preclinical stage studies. Targets are inhibition of ATP7B
mutant retention and degradation in the endoplasmic reticulum,
promotion of copper excretion [43, 44], and introduction of
engineered hepatic cells with the hope of repopulating the liver
[45]. Viral gene therapy for Wilson disease aims to introduce
working copies of ATP7B by delivering the normal ATP7B
complementary DNA (cDNA) to hepatocytes. This is accom-
plished by incorporating the normal gene’s cDNA in a nonpath-
ogenic virus that targets the liver [46–48]. Gene therapy showed
promise in a mouse model with an advanced stage of Wilson
disease; liver enzymes and copper excretion were normalized
[49]. This approach has potential for future application in clini-
cal trials.
Huppke–Brendel syndrome
Huppke–Brendel syndrome is a recently described autosomal
recessive copper metabolism disorder with a biochemical
phenotype similar to that of Wilson disease, but with different
etiology and clinical manifestations. Patients with Huppke–Brendel
syndrome are infants and children with low serum levels of cop-
per and ceruloplasmin associated with syndromic features
including congenital cataracts, hearing loss, and severe develop-
mental delay [50]. On brain imaging, cerebellar atrophy, wid-
ened subarachnoid spaces, and hypomyelination are evident.
This disorder is lethal in early childhood, with death reported
between 22 months and 6 years of age in patients from the origi-
nal cohort. Initially, several known copper transporters and pro-
teins were investigated for their potential role in the pathogenesis
of this syndrome, with no success [51]. Homozygosity mapping
and deep sequencing later identified the genetic basis of the dis-
order, mutations in SLC33A1, a gene that encodes the acetyl‐
CoA transporter AT‐1 [52]. It is thought that AT‐1 is needed to
acetylate glycoproteins and gangliosides in the endoplasmic
reticulum and/or Golgi apparatus [53]. Copper metabolism
imbalance in Huppke–Brendel patients could be secondary to
the effects of a defective AT‐1 on trafficking of copper pumps
ATP7A and ATP7B, since both proteins normally undergo acet-
ylation of specific lysine residues (L. Yi and S.G. Kaler, unpub-
lished data).
MEDNIK syndrome
MEDNIK is an acronym used to describe an autosomal recessive
neurocutaneous syndrome that was first defined in French‐
Canadian families sharing common ancestors [54]. MEDNIK
stands for mental retardation, enteropathy, deafness, neuropathy,
ichthyosis, and keratodermia. In addition to these clinical signs
212 THE LIVER:  CONCLUSIONS
and symptoms, dysfunction in copper metabolism was later
identified in one of the index cases reported with this disease, and
subsequently confirmed in the original cohort [55]. Patients with
MEDNIK present a mixed biochemical phenotype with elements
of both copper deficiency and copper excess. Developmental
delays and brain atrophy found in MEDNIK subjects are reminis-
cent of Menkes disease, a neurodegenerative disorder secondary
to mutations of ATP7A [3]. At the same time, MEDNIK patients
show features of copper overload and liver disease with hypocer-
uloplasminemia, accumulation of copper in the liver, and intrahe-
patic cholestasis [55]. These changes are similar to those seen in
Wilson disease caused by ATP7B mutations.
The molecular basis of MEDNIK syndrome is mutation in the
AP1S1 gene that encodes the sigma 1A (σ1A) subunit of adaptor
protein complex 1 (AP‐1), a complex that normally mediates
intracellular trafficking of certain transmembrane proteins
between the trans‐Golgi network, endosomes, and plasma mem-
brane of cells. The ATP7A and ATP7B copper ATPases are both
transmembrane proteins with similar structures, and their func-
tion depends on their ability to translocate to different cellular
compartments in response to changing cellular copper levels. In
studies of ATP7A trafficking, the AP‐1 complex was implicated
as a key player in this intracellular itinerary by interaction with a
di‐leucine motif near the C‐terminus [56, 57]. This conclusion
was extrapolated to ATP7B, given the similarities between these
two ATPases, including the di‐leucine motif to which AP‐1 binds.
With a defective σ1A subunit of the AP‐1 complex, ATP7A/B
appear unable to remain docked in the trans‐Golgi network under
normal or low copper states [57]. This defect in copper ATPase
localization and trafficking may represent the link between
AP1S1 mutations and the copper metabolism disturbances found
in patients with MEDNIK. It seems conceivable that ATP7B
function in the liver could be impacted more than ATP7A in other
tissues, since two other isoforms of the σ1 subunit, σ1B and σ1C,
exist and may substitute for σ1A, depending on their expression
levels (L. Yi and S.G. Kaler, unpublished data). The cutaneous
and other clinical features of MEDNIK presumably relate to
impaired function of other transmembrane protein cargos.
Zinc acetate therapy has been suggested and used to reduce
copper overload in the liver of MEDNIK patients [55]. Further
understanding of copper ATPase trafficking in MEDNIK subjects
could shed light on the precise pathophysiology of the disorder,
and allow the development of more precisely targeted therapies.
Primary disorders of the liver and copper
imbalance
Since the liver is the main regulatory organ of copper metabolism,
the consequences of an imbalance in the metal’s levels are
expected in hepatocytes, as previously described in diseases such
as Wilson disease and MEDNIK. Conversely, evidence of copper
level dysregulation is also present in primary hepatic diseases.
Non‐alcoholic fatty liver disease
Non‐alcoholic fatty liver disease (NAFLD) is the accumulation
of fat in the liver with structural disorganization, in the absence
of excessive alcohol consumption. The severity of steatosis and
histological changes varies, ranging from benign steatosis to
significant fat deposits accompanied by hepatocellular inflam-
mation and necrosis (non‐alcoholic steatohepatitis or NASH).
The latter can potentially progress to fibrosis and cirrhosis, end‐
stage liver disease, and hepatocellular carcinoma [58]. NAFLD
is closely related to the metabolic syndrome, which encom-
passes abdominal obesity, hypertension, dyslipidemia, and
impaired glucose tolerance or diabetes [59]. The pathophysiol-
ogy of this disorder is complex and incompletely understood,
with a possible role of copper in the process. Interestingly, a
study of 124 patients with NAFLD showed reduced hepatic cop-
per concentrations compared to controls and other liver diseases.
This correlation was more pronounced in patients with advanced
hepatic steatosis, non‐alcoholic steatohepatitis, and additional
features of the metabolic syndrome [60]. Copper deficiency has
previously been linked to atherogenic dyslipidemia, and a recent
study of ATP7B showed additional data connecting copper and
lipid metabolism at the enterocyte level [7, 61, 62]. Copper‐defi-
cient diets in rats induced hepatic steatosis and insulin resistance,
which further supports the role of low copper in the pathogenesis
of NAFLD [60–62]. Additionally, beta‐oxidation occurs in per-
oxisomes and mitochondria, and the process is downregulated in
states of copper deficiency [63]. Copper is needed for normal
mitochondrial function, since it serves as a critical cofactor for
cytochrome c oxidase, the last enzyme complex in the mitochon-
drial electron transport chain. Copper‐deficient mice show mito-
chondrial dysfunction [64], and similar alterations can be seen in
humans with NAFLD. Finally, copper is also a cofactor of super-
oxide dismutase 1 (SOD1), an important enzyme for controlling
oxidative stress, which has been linked to the hepatic damage in
NAFLD and NASH [4, 60].
Hepatocellular carcinoma
The role of copper in cancer development and progression has
been widely studied, with several reports showing high serum
and tumor copper levels in multiple cancer types [65–67]. High
copper levels promote angiogenesis and oxidative stress genera-
tion, processes which create a suitable environment for tumor
growth, and which could be counteracted by lowering copper
levels [67–70]. These realizations led to the investigation of
copper chelation as a potential anticancer therapy [71, 72], with
applications in hepatocellular carcinoma (HCC).
HCC is one of the cancers in which studies have documented
high tumor copper levels [73–75]. Serum copper and cerulo-
plasmin levels were also found to be elevated in patients with
HCC, and higher levels correlated with worse prognosis [76].
Interestingly, other types of hepatic neoplasms, including cholan-
giocarcinoma and metastatic liver disease, did not show significant
copper level elevation, which helps distinguish them from HCC
[77, 78]. The therapeutic strategy of lowering copper levels to treat
cancer tested in vivo with HCC yielded promising results showing
suppression of angiogenesis and tumor growth [72, 79]. Additional
study of this therapeutic approach for other cancers is warranted.
CONCLUSIONS
The liver is the main regulator of copper levels in the blood,
through its role in the distribution and excretion of this trace
 18:  Copper Metabolism and the Liver 213
metal. Hepatocytes process copper by directing it to various path-
ways, depending on the body’s needs and on intracellular levels.
Copper is shuttled by the transporter ATP7B into the secretory
compartment of hepatic cells, where it is incorporated into the
serum glycoprotein ceruloplasmin. In conditions of excessive
intracellular copper, the metal is excreted into the bile from the
canalicular (apical) pole of hepatocytes. Defects in ATP7B func-
tion or trafficking lead to copper accumulation in the liver with
consequent toxicity and hepatocellular injury. These mechanisms
explain the liver manifestations seen in Wilson disease, Huppke–
Brendel syndrome, and MEDNIK syndrome. Copper dyshomeo-
stasis is also found in HCC and NAFLD, with implications for
lipid metabolism, but the mechanisms underlying these condi-
tions are incompletely understood. Further insights into copper
metabolism at the cellular level, as well as carefully designed
clinical trials involving novel copper chelators or viral gene ther-
apy are needed to further advance understanding of these disease
mechanisms and assess potential therapeutic remedies.
REFERENCES
1. Baker, Z.N., Cobine, P.A., and Leary, S.C. The mitochondrion: a central
architect of copper homeostasis. Metallomics, 2017;15;9(11):1501–12.
2. Bousquet‐Moore, D., Mains, R.E., and Eipper, B.A. PAM and copper  –  a
gene/nutrient interaction critical to nervous system function. J Neurosci Res,
2010;88(12):2535–45.
3. Kaler, S.G. ATP7A‐related copper transport diseases‐emerging concepts and
future trends. Nat Rev Neurol, 2011;7(1):15–29.
4. McCord, J.M. and Fridovich, I. Superoxide dismutase an enzymic function
for erythrocuprein (hemocuprein). J Biol Chem, 1969;244(22):6049–55.
5. Rucker, R.B., Kosonen, T., Clegg, M.S. et  al. Copper, lysyl oxidase, and
extracellular matrix protein cross‐linking. Am J Clin Nutr, 1998;67(5
Suppl):996S–1002S.
6. Lutsenko, S. Human copper homeostasis: a network of interconnected path-
ways. Curr Opin Chem Biol, 2010;14(2):211–17.
7. Pierson, H., Muchenditsi, A., Kim B.‐E. et al. The function of ATPase copper
transporter ATP7B in intestine. Gastroenterology, 2018;154(1):168–180.e5.
8. Zerounian, N.R., Redekosky, C., Malpe, R., and Linder, M.C. Regulation of
copper absorption by copper availability in the Caco‐2 cell intestinal model.
Am J Physiol Gastrointest Liver Physiol, 2003;284(5):G739–747.
9. Berghe, V.D., Ve, P., and Klomp, L.W. New developments in the regulation of
intestinal copper absorption. Nutr Rev, 2009;67(11):658–72.
10. Nose, Y., Kim, B.‐E., and Thiele, D.J. Ctr1 drives intestinal copper absorp-
tion and is essential for growth, iron metabolism, and neonatal cardiac func-
tion. Cell Metab, 2006;4(3):235–44.
11. Lee, J., Peña, M.M.O., Nose, Y., and Thiele, D.J. Biochemical characteriza-
tion of the human copper transporter Ctr1. J Biol Chem, 2002;277(6):
4380–7.
12. Garrick, M.D., Singleton, S.T., Vargas, F. et al. DMT1: which metals does it
transport? Biol Res, 2006;39(1):79–85.
13. Lutsenko, S., Barnes, N.L., Bartee, M.Y., and Dmitriev, O.Y. Function and
regulation of human copper‐transporting ATPases. Physiol Rev, 2007;87(3):
1011–46.
14. Kim, B.‐E., Turski, M.L., Nose, Y. et al. Cardiac copper deficiency activates
a systemic signaling mechanism that communicates with the copper acquisi-
tion and storage organs. Cell Metab, 2010;11(5):353–63.
15. Linz, R. and Lutsenko, S. Copper‐transporting ATPases ATP7A and ATP7B:
cousins, not twins. J Bioenerg Biomembr, 2007;39(5–6):403–7.
16. Monty, J.‐F., Llanos, R.M., Mercer, J.F.B., and Kramer, D.R. Copper expo-
sure induces trafficking of the menkes protein in intestinal epithelium of
ATP7A transgenic mice. J Nutr, 2005;135(12):2762–6.
17. Ravia, J.J., Stephen, R.M., Ghishan, F.K., and Collins, J.F. Menkes copper
ATPase (Atp7a) is a novel metal‐responsive gene in rat duodenum, and
immunoreactive protein is present on brush‐border and basolateral mem-
brane domains. J Biol Chem, 2005;280(43):36221–7.
18. Liu, N., Lo, L.S., Askary, S.H. et al. Transcuprein is a macroglobulin regu-
lated by copper and iron availability. J Nutr Biochem, 2007;18(9):597–608.
19. Hellman, N.E. and Gitlin, J.D. Ceruloplasmin metabolism and function.
Annu Rev Nutr, 2002;22:439–58.
20. Hellman, N.E., Kono, S., Mancini, G.M. et  al. Mechanisms of Copper
Incorporation into Human Ceruloplasmin. J Biol Chem, 2002;277(48):
46632–8.
21. Kuo, Y.M., Gitschier, J., and Packman, S. Developmental expression of the
mouse mottled and toxic milk genes suggests distinct functions for the
Menkes and Wilson disease copper transporters. Hum Mol Genet,
1997;6(7):1043–9.
22. Pyatskowit, J.W. and Prohaska, J.R. Copper deficient rats and mice both
develop anemia but only rats have lower plasma and brain iron levels. Comp
Biochem Physiol C Toxicol Pharmacol, 2008;147(3):316–23.
23. Schilsky, M.L. and Thiele, D.J. Copper metabolism and the liver, in The
Liver (eds. I.M. Arias et  al.), Wiley‐Blackwell, Chichester, 2009, pp.
221–33.
24. Smallwood, R.A., Williams, H.A., Rosenoer, V.M., and Sherlock, S. Liver‐
copper levels in liver disease: studies using neutron activation analysis.
Lancet, 1968;292(7582):1310–13.
25. Wilson, S.A.K. Progressive lenticular degeneration: a familial nervous disease
associated with cirrhosis of the liver. Brain, 1912;34(4):295–507.
26. Cumings, J.N. The copper and iron content of brain and liver in the normal
and in hepato‐lenticular degeneration. Brain, 1948;71(4):410–15.
27. Bull, P.C., Thomas, G.R., Rommens, J.M., Forbes, J.R., and Cox, D.W. The
Wilson disease gene is a putative copper transporting P‐type ATPase similar
to the Menkes gene. Nat Genet, 1993;5(4):327–37.
28. Bennett, J. and Hahn, S.H. Clinical molecular diagnosis of Wilson disease.
Semin Liver Dis, 2011;31(3):233–8.
29. Ferenci, P. Regional distribution of mutations of the ATP7B gene in patients
with Wilson disease: impact on genetic testing. Hum Genet,
2006;120(2):151–9.
30. Weiss, K.H. Wilson disease. GeneReviews [internet], 2016. https://www.
ncbi.nlm.nih.gov/books/NBK1512/ (Accessed Oct 26, 2018).
31. Gomes, A. and Dedoussis, G.V. Geographic distribution of ATP7B mutations
in Wilson disease. Ann Hum Biol, 2016;43(1):1–8.
32. Chang, I.J. and Hahn, S.H. The genetics of Wilson disease. Handb Clin
Neurol, 2017;142:19–34.
33. Bandmann, O., Weiss, K.H., and Kaler, S.G. Wilson disease and other neuro-
logical copper disorders. Lancet Neurol, 2015;14(1):103–13.
34. Roberts, E.A. and Schilsky, M.L. A practice guideline on Wilson disease.
Hepatology, 2003;37(6):1475–92.
35. Ferenci, P., Członkowska, A., Merle, U. et  al. Late‐onset Wilson disease.
Gastroenterology, 2007;132(4):1294–8.
36. Kaler, S.G. Inborn errors of copper metabolism, in Handbook of Clinical
Neurology (eds. O. Dulac, M. Lassonde, and H.B. Sarnat), Elsevier, New
York, 2013, pp. 1745–54.
37. Machado, A., Chien, H.F., Deguti, M.M. et al. Neurological manifestations
in Wilson disease: report of 119 cases. Mov Disord, 2006;21(12):2192–6.
38. Ala, A., Borjigin, J., Rochwarger, A., Schilsky, M. Wilson disease in septua-
genarian siblings: raising the bar for diagnosis. Hepatology, 2005;41(3):
668–70.
39. Kim, J.W., Kim, J.H., Seo, J.K. et al. Genetically confirmed Wilson disease
in a 9‐month old boy with elevations of aminotransferases. World J Hepatol,
2013;5(3):156–9.
40. Ferenci, P., Caca, K., Loudianos, G. et al. Diagnosis and phenotypic classifi-
cation of Wilson disease. Liver Int, 2003;23(3):139–42.
41. Roberts, E.A. and Schilsky, M.L. Diagnosis and treatment of Wilson disease:
an update. Hepatology, 2008;47(6):2089–111.
42. Schilsky, M.L. Liver transplantation for Wilson disease. Ann N Y Acad Sci,
2014;1315:45–9.
43. Kaler, S.G. Microbial peptide de‐coppers mitochondria: implications for
Wilson disease. J Clin Invest, 2016;126(7):2412–4.
44. Lichtmannegger, J., Leitzinger, C., Wimmer, R. et  al. Methanobactin
reverses acute liver failure in a rat model of Wilson disease. J Clin Invest,
2016;126(7):2721–35.
45. Ranucci, G., Polishchuck, R., and Iorio, R. Wilson disease: prospective
developments towards new therapies. World J Gastroenterol, 2017;23(30):
5451–6.
46. Meng, Y., Miyoshi, I., Hirabayashi, M. et al. Restoration of copper metabolism
and rescue of hepatic abnormalities in LEC rats, an animal model of Wilson
214 THE LIVER:  REFERENCES
disease, by expression of human ATP7B gene. Biochim Biophys Acta,
2004;1690(3):208–19.
47. Roybal, J.L., Endo, M., Radu, A. et al. Early gestational gene transfer with
targeted ATP7B expression in the liver improves phenotype in a murine
model of Wilson disease. Gene Ther, 2012;19(11):1085–94.
48. Murillo, O., Luqui, D.M., Gazquez, C. et al. Long‐term metabolic correction
of Wilson disease in a murine model by gene therapy. J Hepatol,
2016;64(2):419–26.
49. Murillo, O., Moreno, D., Gazquez, C. et al. Improvement of gene therapy for
Wilson disease. Mol Ther, 2016;24:S64.
50. Huppke, P., Brendel, C., Kalscheuer, V. et al. Mutations in SLC33A1 cause a
lethal autosomal‐recessive disorder with congenital cataracts, hearing loss, and
low serum copper and ceruloplasmin. Am J Hum Genet, 2012;90(1):61–8.
51. Horváth, R., Freisinger, P., Rubio, R. et  al. Congenital cataract, muscular
hypotonia, developmental delay and sensorineural hearing loss associated
with a defect in copper metabolism. J Inherit Metab Dis, 2005;
28(4):479–92.
52. Kanamori, A., Nakayama, J., Fukuda, M.N. et  al. Expression cloning and
characterization of a cDNA encoding a novel membrane protein required for
the formation of O‐acetylated ganglioside: a putative acetyl‐CoA transporter.
Proc Natl Acad Sci U S A, 1997;94(7):2897–902.
53. Hirabayashi, Y., Nomura, K.H., and Nomura, K. The acetyl‐CoA transporter
family SLC33. Mol Aspects Med, 2013;34(2–3):586–9.
54. Montpetit, A., Côté, S., Brustein, E. et al. Disruption of AP1S1, causing a
novel neurocutaneous syndrome, perturbs development of the skin and spi-
nal cord. PLoS Genet, 2008;4(12):e1000296.
55. Martinelli, D., Travaglini, L., Drouin, C.A. et  al. MEDNIK syndrome: a
novel defect of copper metabolism treatable by zinc acetate therapy. Brain,
2013;136(Pt 3):872–81.
56. Holloway, Z.G., Velayos‐Baeza, A., Howell, G.J. et  al. Trafficking of the
Menkes copper transporter ATP7A is regulated by clathrin‐, AP‐2‐, AP‐1‐,
and Rab22‐dependent steps. Mol Biol Cell, 2013 Jun 1;24(11):1735–48.
57. Yi, L. and Kaler, S.G. Direct interactions of adaptor protein complexes 1 and
2 with the copper transporter ATP7A mediate its anterograde and retrograde
trafficking. Hum Mol Genet, 2015;24(9):2411–25.
58. Adams, L.A., Angulo, P., and Lindor, K.D. Nonalcoholic fatty liver disease.
CMAJ, 2005;172(7):899–905.
59. Greenfield, V., Cheung, O., and Sanyal, A.J. Recent advances in nonalcholic
fatty liver disease. Curr Opin Gastroenterol, 2008;24(3):320–7.
60. Aigner, E., Strasser, M., Haufe, H. et al. A role for low hepatic copper con-
centrations in nonalcoholic fatty liver disease. Am J Gastroenterol,
2010;105(9):1978–85.
61. al‐Othman, A.A., Rosenstein, F., Lei, K.Y. Copper deficiency alters plasma
pool size, percent composition and concentration of lipoprotein components
in rats. J Nutr, 1992;122(6):1199–204.
62. al‐Othman, A.A., Rosenstein, F., and Lei, K.Y. Copper deficiency increases
in vivo hepatic synthesis of fatty acids, triacylglycerols, and phospholipids in
rats. Proc Soc Exp Biol Med, 1993;204(1):97–103.
63. Tosco, A., Fontanella, B., Danise, R. et al. Molecular bases of copper and
iron deficiency‐associated dyslipidemia: a microarray analysis of the rat
intestinal transcriptome. Genes Nutr, 2010;5(1):1–8.
64. Ibdah, J.A., Perlegas, P., Zhao, Y. et al. Mice heterozygous for a defect in
mitochondrial trifunctional protein develop hepatic steatosis and insulin
resistance. Gastroenterology, 2005;128(5):1381–90.
65. Apelgot, S., Coppey, J., Fromentin, A. et al. Altered distribution of copper
(64Cu) in tumor‐bearing mice and rats. Anticancer Res, 1986;6(2):159–64.
66. Daniel, K.G., Harbach, R.H., Guida, W.C., and Dou, Q.P. Copper storage
diseases: Menkes, Wilsons, and cancer. Front Biosci, 2004;9:2652–62.
67. Gupte, A. and Mumper, R.J. Elevated copper and oxidative stress in cancer
cells as a target for cancer treatment. Cancer Treat Rev, 2009;35(1):32–46.
68. Brewer, G.J. Copper lowering therapy with tetrathiomolybdate as an antian-
giogenic strategy in cancer. Curr Cancer Drug Targets, 2005;5(3):195–202.
69. Lowndes, S.A. and Harris, A.L. The role of copper in tumour angiogenesis.
J Mammary Gland Biol Neoplasia, 2005;10(4):299–310.
70. Sproull, M., Brechbiel, M., and Camphausen, K. Antiangiogenic therapy
through copper chelation. Expert Opin Ther Targets, 2003;7(3):405–9.
71. Goodman, V.L., Brewer, G.J., and Merajver, S.D. Copper deficiency as an
anti‐cancer strategy. Endocr Relat Cancer, 2004;11(2):255–63.
72. Yoshii, J., Yoshiji, H., Kuriyama, S. et al. The copper‐chelating agent, trien-
tine, suppresses tumor development and angiogenesis in the murine hepato-
cellular carcinoma cells. Int J Cancer, 2001;94(6):768–73.
73. Vecchio, F.M., Federico, F., and Dina, M.A. Copper and hepatocellular car-
cinoma. Digestion, 1986;35(2):109–14.
74. Ebara, M., Fukuda, H., Hatano, R. et al. Relationship between copper, zinc
and metallothionein in hepatocellular carcinoma and its surrounding liver
parenchyma. J Hepatol, 2000;33(3):415–22.
75. Ebara, M., Fukuda, H., Hatano, R. et al. Metal contents in the liver of patients
with chronic liver disease caused by hepatitis C virus. Reference to hepato-
cellular carcinoma. Oncology, 2003;65(4):323–30.
76. Casaril, M., Capra, F., Marchiori, L. et al. Serum copper and ceruloplasmin
in early and in advanced hepatocellular carcinoma: diagnostic and prognos-
tic relevance. Tumori, 1989;75(5):498–502.
77. Haratake, J., Horie, A., Takeda, S. et al. Tissue copper content in primary and
metastatic liver cancers. Acta Pathol Jpn, 1987;37(2):231–8.
78. Gurusamy, K. and Davidson, B.R. Trace element concentration in metastatic
liver disease: a systematic review. J Trace Elem Med Biol,
2007;21(3):169–77.
79. Yoshiji, H., Kuriyama, S., Yoshii, J. et al. The copper‐chelating agent, trientine,
attenuates liver enzyme‐altered preneoplastic lesions in rats by angiogenesis
suppression. Oncol Rep, 2003;10(5):1369–73.
 The Central Role of the Liver
19 in Iron Storage and Regulation
of Systemic Iron Homeostasis
Tracey A. Rouault1, Victor R. Gordeuk2, and Gregory J. Anderson3
1
Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of
Health, Bethesda, MD, USA
2
University of Illinois at Chicago, Chicago, IL, USA
3
QIMR Berghofer Medical Research Institute, Brisbane, Queensland, Australia

THE LIVER AS A MAJOR IRON
REPOSITORY
In healthy adult humans, the liver is estimated to store from
0.07 to 0.4 g of iron. A combined analysis of previous studies
using quantitative phlebotomy indicated that total body
median storage iron was approximately 0.8 g among 39 nor-
mal men predominantly in the third and fourth decades of life,
and 0.4 g among 20 normal women of similar age [1–5].
Based on the assumption that one‐third of iron stores are nor-
mally in the liver, this would translate to a normal median
hepatic iron content of 0.27 g for men and 0.13 g for women.
Iron is sequestered within the iron storage protein ferritin, a
large 24‐subunit protein composed of ferritin L (light) and H
(heavy) chains (Figure 19.1). The ferritin subunits combine to
form a spherical protein shell that contains numerous chan-
nels. On the inside of the spherical protein, iron is oxidized
from the ferrous (Fe2+) form to the ferric (Fe3+) form by the
ferroxidase activity of the ferritin H chain, and insoluble ferric
iron deposits grow from their initial deposition sites on side‐
chains of ferritin L subunits [6, 7]. As ferritin iron uptake and
oxidation continue, thousands of ferric iron atoms accumulate
within the ferritin sphere, which has the capacity to store up to
4500 iron atoms. As the iron contained within ferritin is not
very bioavailable, ferritin sequesters iron and reduces the
availability of cytosolic iron to interact with oxygen and gen-
erate harmful reactive oxygen species. Ferritin also serves as a
source of iron for the metabolic needs of the cell. Iron can be
released from intact ferritin that has been mono‐ubiquitinated
and will be ultimately degraded by the proteasome [8].
Alternatively, ferritin multimers can undergo degradation in
lysosomes to release iron. This process is guided by interac-
tion of ferritin with NCOA4, which binds to ferritin and
­delivers it to the lysosome (Figure 19.1) [9].
SERUM TRANSFERRIN IS THE MAJOR
SOURCE OF IRON FOR TISSUES
Transferrin (TF) is a high‐affinity binding protein for ferric iron
that circulates in the plasma with a concentration of 9–28 nmol
L−1. It is synthesized mainly by hepatocytes and is secreted into
the circulation. The plasma usually contains an excess of TF,
such that less than one‐third of the iron‐binding sites on TF are
occupied by iron. TF contains two lobes, each of which contains
a ferric (Fe3+) iron‐binding site [11]. Cells in tissues throughout
the body can obtain iron from circulating TF by increasing
expression of transferrin receptors (TFRs). TFR1 is the major
receptor for TF and binds diferric TF with high affinity (107–109
M−1) at physiological pH. Monoferric and apoTF are bound with
much lower affinity (106 and <105 M−1 respectively) [12]. Upon
binding diferric TF, a TF–TFR complex forms and then inter-
nalizes within endosomes. As endosomes undergo acidification,
iron is released from the TFR–TF complex. The released iron is
still in the oxidized ferric form (Fe3+) and must undergo reduc-
tion by a reductase of the STEAP family of proteins [13] prior
to being transported from within the endosome to the cytosol by
a divalent metal‐ion transporter (DMT) known as DMT1 [14].
When iron enters the cytosol, it can be incorporated into many

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
216 THE LIVER:  NON‐TRANSFERRIN‐BOUND IRON UPTAKE AND HEPATIC IRON OVERLOAD
Figure 19.1  Ferritin is a 24 subunit protein in which subunits of H or
L chains coassemble to create a protein spherical structure with a cen-
tral cavity. In this depiction, generated from coordinates of a reported
ferritin crystal structure [10], three separate subunits on the anterior of
the spherical protein are in dark gray, to show the arrangement of indi-
vidual subunits. Pores in the structure allow entry of iron, which accu-
mulates as a ferric precipitate within the hollow core. Each ferritin
multimer can sequester up to 4500 iron atoms.
enzymes and proteins that require iron for function. Iron can
also be transported to subcellular locations such as mitochon-
dria, which use an iron importer known as mitoferrin to facili-
tate mitochondrial iron uptake [15].
THE ROLE OF TFR2, A SECOND
TRANSFERRIN RECEPTOR HIGHLY
EXPRESSED BY HEPATOCYTES
Whereas in most tissues TFR1 is the first component of the
major cellular iron uptake pathway, the liver expresses large
amounts of a second, homologous TFR, TFR2 [16]. In fact,
TFR2 appears to be considerably more abundant in the liver than
TFR1 [17]. Despite their homology, recent data suggest that
there are differences in the way the two TFRs bind TF [18]. The
finding that TFR2 has a 25‐fold lower affinity for diferric TF
than TFR1 [16, 19] led to the supposition that TFR2 was not
important for iron uptake. However, the concentration of diferric
TF in plasma is estimated at about 8–12 μM [20], a concentration
that would likely be sufficient to saturate both TFRs, suggesting
that both might play a role in hepatic iron uptake. Despite this,
studies under defined conditions in hepatoma cells demonstrate
that the TFR1‐mediated uptake of TF‐bound iron is by far the
dominant iron uptake pathway [21]. In addition, the livers of
patients with mutations in TFR2 or mice lacking TFR2 readily
load with iron, indicating that TFR2 is dispensible for hepatic
iron upake [22]. As will be described in more detail below, the
major role of TFR2 in the liver is as a regulator of hepcidin, and
hence it plays a critical role in body iron homeostasis.
TFR2 EXPRESSION IS REGULATED BY
DIFFERENT MECHANISMS THAN TFR1
Aside from the fact that TFR2 is mainly expressed in hepatocytes,
whereas TFR1 is expressed highly in erythroid progenitors and
to varying degrees in most other tissues, the major difference
between the two receptors is a difference in the mechanisms by
which their expression is regulated. TFR1 is regulated at the
transcriptional level by hypoxia [23, 24], but it is also potently
regulated by a post‐transcriptional regulatory system that con-
sists of iron‐regulatory proteins (IRPs), which bind to RNA
stem‐loops known as iron responsive elements (IREs) in tran-
scripts such as ferritin and TFR1 (Figure 19.2) [25]. In cells that
have cytosolic iron deficiency, IRPs bind to IREs, and IRP bind-
ing prevents translation of transcripts such as those encoding the
ferritin H and L chains, which have IREs located within their 5′
UTRs (Figure 19.3a). Conversely, IRP binding to IREs in the 3′
UTR of TFR1 stabilizes the transcript, leading to an increase in
mRNA levels, and a concomitant increase in TFR1 protein lev-
els (Figure 19.3b). In most cells that are iron‐loaded, IRPs do
not bind to IREs, and ferritin synthesis increases, whereas TFR1
expression decreases. Thus, iron loading in most cells leads to
an increase in iron sequestration and to a decrease in TFR1‐
mediated iron uptake (reviewed in [25, 26]).
Hepatocytes differ from other cells because they express high
levels of TFR2, and TFR2 mRNA levels remain high with iron
loading [27]. In fact, binding of diferric TF leads to increased
TFR2 stability by changing TFR2 trafficking [28]. One reason
why hepatocytes are an important iron repository may thus be
because TFR2‐dependent iron uptake continues even when
expression of TFR1 is diminished by iron overload in other cells
throughout the body. The propensity to maintain iron uptake
through non‐IRP‐regulated TFR expression may also be valua-
ble for the hepatocyte in its role as a systemic iron regulator, as
discussed later. Although TFR2 is implicated in a signaling
pathway that responds to iron status (see later), the liver also has
a strong capacity to take up non‐transferrin‐bound iron (NTBI).
NON‐TRANSFERRIN‐BOUND IRON
UPTAKE AND HEPATIC IRON OVERLOAD
In addition to the uptake of TF‐bound iron, the liver can efficiently
take up iron that is not sequestered by TF (i.e. NTBI). NTBI
becomes detectable in the plasma mainly when TF iron‐binding
sites become fully saturated, as occurs in various diseases of
systemic iron overload. The plasma half‐life of NTBI is less
than 30 seconds, whereas TF‐bound iron has a half‐life of
approximately 50 minutes [29, 30]. Thus, the isolated perfused
liver can remove 60–75% of NTBI from plasma on first pass,
but less than 2% of TF‐bound iron [29, 31, 32]. In addition, if
plasma TF is first saturated with iron, the clearance of a subse-
quent dose of radioactive iron, given either orally or intrave-
nously, is very rapid, with the liver being the major site of iron
deposition [30]. Perhaps the best demonstration of the capacity
of the liver to acquire NTBI comes from studies of congenital
hypotransferrinemia, where little TF is synthesized in either a
 19:  The Central Role of the Liver in Iron Storage and Regulation of Systemic Iron Homeostasis 217
A consensus motif for the IRE
GU can also be A
A G
C X U/C/A
N-N′
N-N′
N-N′ 5 base pair stem
N-N′
N-N′
Unpaired C
nucleoside − N-N′
usually cytosine N-N′ Lower stem
N-N′ (at least 3 base pairs)
5′-N-N′-3′
Consensus IRE
Functional IREs in 5′UTRs
Ferritin
Mitochondrial aconitase
Succinate dehydrogenase D. melanogaster
eALASynthase
Ferroportin-basolateral iron exporter
HIF2 alpha
Figure 19.2  Iron responsive elements (IREs) are RNA stem‐loops found in the mRNAs encoding numerous proteins important in iron homeosta-
sis. IREs consist of a base‐paired stem, an unpaired cytosine that separates the upper and lower stems, and a six‐membered loop with the sequence
CAGUGX, in which X can be any residue except G and there is a base pair between A2 and G5 of the loop. Reproduced from [25] with permission
of Elsevier.
mouse model [33] or human patients [34]. In this condition, the
liver and some other organs develop massive iron loading at a
very young age. Such studies show that the role of TF is not
simply to deliver iron to tissues, but to allow the regulated deliv-
ery of the metal. TF has such a high binding affinity for iron
under normal physiological conditions (affinity constant 1023
M−1) [35] that the amount of circulating NTBI is extremely
small and is essentially undetectable. Where NTBI does become
important is in various pathological states, particularly iron‐
loading disorders in which the capacity of plasma TF to bind
iron is exceeded. In HFE‐associated hemochromatosis, for
example, NTBI can reach 10–15 μM [36]. It has proved difficult
to precisely define the chemical nature of NTBI, but most of it
may be chelated by small organic acids such as citrate and
amino acids, and some can also be bound by proteins such as
albumin [37].
Until recently, the uptake pathway of NTBI had been a
­significant unresolved problem in iron homeostasis. Early stud-
ies suggested that NTBI uptake may be mediated by DMT1 [38,
39] but when DMT1 is disrupted, the liver is still able to accu-
mulate NTBI, indicating that other uptake pathways were
involved [40, 41]. Another candidate is ZRT‐ and IRT‐like pro-
tein 14 (ZIP14) or SLC39A14 [42]. This plasma membrane
transporter is expressed in the liver and was shown to mediate
NTBI uptake in hepatoma cells [43]. Subsequent studies with
ZIP14‐knockout mice have shown that ZIP14 provides the main
pathway for the uptake of NTBI by the liver and pancreas [43].
Indeed, the same study showed that deletion of ZIP14 in mice
can completely prevent hepatic iron loading in mouse models of
hemochromatosis. Interestingly, the reduction of ferric iron
appears to be required before it can be utilized by hepatic ZIP14
G
C
G
C
5′
3′
Functional IREs in 3′UTRs
Transferrin receptor
Divalent metal transporter
(non-consensus)
CDC14A
and this activity can be supplied by the prion protein Pr(C) [44].
Early studies on the kinetics of NTBI uptake indicated that the
process was increased under iron loading conditions and it was
suggested that this plays a protective role by removing poten-
tially toxic “free” iron from the circulation [45]. Consistent with
this finding is the more recent observation that ZIP14 levels are
upregulated by iron loading [46].
THE RELEASE OF IRON FROM HEPATIC
CELLS
Although a great deal is known about the uptake of iron, the
mechanisms by which the liver can release stored iron in times
of need are less well understood. The capacity of the liver to
relinquish stored iron is particularly well illustrated during
phlebotomy therapy for primary iron‐loading diseases such as
HFE‐associated hemochromatosis [47]. In such diseases, the
liver may contain 20 g or more of stored iron, but all of this can
be mobilized for hemoglobin synthesis following phlebotomy.
While in HFE‐associated hemochromatosis most of the iron
storage in the liver occurs in hepatocytes (see later), in transfu-
sional iron overload the resident macrophages (the Kupffer
cells) also can store much of the iron, and iron can be mobilized
from Kupffer cells as well as hepatocytes by phlebotomy. In
addition, Kupffer cells play an important role in the recycling of
hemoglobin‐derived iron from senescent red blood cells, and
this iron must also be released into the circulation [48].
The protein responsible for iron release from most body cells
is ferroportin (FPN) [49–51], a highly conserved ferrous iron
218 THE LIVER:  THE RELEASE OF IRON FROM HEPATIC CELLS

(a)
Ferritin mRNA One IRE in 5′UTR

5′ 3′
AUG

−Fe +Fe
IRP

40s AAAAAA

IRE occupied by IRP,

inhibiting translation
initiation

IRE unoccupied, allowing

60s polysome formation and
increased ferritin synthesis
(b)
TFR mRNA Five IREs in 3′UTR

AAAAAA
Protein 3′
5′
coding

−Fe

IRP +Fe

Endonuclease
AAAAAA cleavage site
Protein
coding AAAAAA
Protein
coding
One or more IREs occupied
by IRP, protecting mRNA
from rate-determining IRE unoccupied, rendering
step mRNA degradation mRNA susceptible to an
endonuclease

Figure 19.3  (a) Iron‐regulatory proteins (IRP) 1 and 2 repress translation of transcripts that contain an iron responsive element (IRE) in the 5′
UTR of the transcript by preventing binding of translation factors and ribosomes. (b) Binding of IRPs to IREs in the 3′ UTR protects the transferrin
receptor (TFR1) transcript from degradation, and likely also protects other transcripts that contain a 3′ IRE by a similar mechanism.

export protein that contains 9 or 10 transmembrane domains.
While FPN is most highly expressed in cell types that efflux
large amounts of iron, such as intestinal enterocytes and mac-
rophages [52], it also performs an iron efflux role in all cells
studied to date, including hepatocytes [53]. FPN mRNA con-
tains an IRE in its 5′ UTR, like the ferritin genes, and its expres-
sion in the liver increases with iron loading, suggesting that it is
under iron‐dependent translational control through the IRE/IRP
system [51]. Since FPN levels increase with iron loading, it is
feasible that this increase represents a protective mechanism to
help the liver limit its iron accumulation.
Efficient export of iron from tissues requires an iron oxidase
in addition to FPN. For most tissues the oxidase activity is
supplied by circulating ceruloplasmin (CP) [54]. CP is the major
copper‐containing protein in the plasma, but it is a protein of
iron metabolism rather than copper metabolism. In the absence
of CP (e.g. in rare cases of aceruloplasminemia), the plasma
iron level is reduced and iron accumulates in the tissues, whereas
copper homeostasis is unaffected [55, 56]. Recent data have
indicated that the membrane‐bound CP homolog hephaestin
(HEPH) can also contribute to hepatic iron efflux [57]. How CP
contributes to iron efflux is unclear, but there is some evidence
that a glycophosphatidylinositol (GPI)‐linked form of CP on
glioma cells can stabilize FPN on the cell surface by oxidizing
transported ferrous iron [58], and HEPH has been reported to
stabilize FPN in hippocampal neurons [59]. More plasma
 19:  The Central Role of the Liver in Iron Storage and Regulation of Systemic Iron Homeostasis 219
membrane FPN means increased cellular iron release. Whether
this is a general mechanism of CP/HEPH action that also applies
to the liver remains to be determined. The liver is the principal
site of synthesis of CP and thus it plays an important role sys-
temically in facilitating iron efflux.
HEPATOCYTES SYNTHESIZE
AND SECRETE THE REGULATORY
PEPTIDE HORMONE HEPCIDIN
In systemic iron regulation, secretion of the peptide hormone
hepcidin has an important role in determining how much iron is
released from macrophages, and likely other body cells, and
how much iron is absorbed through the duodenum [60].
Hepcidin is related to other peptide hormones known as
defensins that function in defense against bacterial infections. It
is synthesized as an 84 amino acid precursor, which undergoes
cleavage and is released into the circulation as an active 25
amino acid peptide hormone [61]. Forms of hepcidin that are
truncated at the N‐terminus and contain 22 or 20 amino acids
are also present in serum, but these forms are not active [61].
(Figure 19.4 shows the hepcidin sequence and the positions of
disulfide bonds.)
As a circulating peptide, hepcidin can interact with tissues
and cells throughout the body. Hepcidin binds to the iron
exporter FPN on target cell membranes, where it forms a com-
plex that internalizes and undergoes ubiquitination and lysoso-
mal degradation [62, 63]. This in turn reduces the flow of iron
into the circulation. It is not yet clear if this mechanism of action
can explain all of the effects of hepcidin (see later).
The transcription of hepcidin is low when there is systemic
iron depletion, and high when there is systemic iron overload.
Levels are also reduced when erythropoiesis is stimulated, and
increased under inflammatory conditions. The mechanisms by
which hepatocytes transduce information about systemic iron
status into transcriptional activity of the hepcidin promoter have
been extensively investigated [65], but the regulatory system
remains incompletely understood.
In general, peptide hormones such as hepcidin effect changes
by binding to specific receptors in target cell populations.
Binding of a ligand to a receptor often activates a process of
signal transduction, a cascade of intracellular events that
changes the transcription of various genes in the target cell pop-
ulation. Hepcidin signaling does not conform to this general
pathway, because hepcidin binds to the iron exporter FPN rather
than to a member of a previously recognized receptor family.
Upon binding hepcidin, FPN is internalized and subsequently
1 5 10 15 20 25
DTHFPICIFCCGCCHRSKCGMCCKT
Figure 19.4  Sequence of the peptide hormone hepcidin. The peptide
contains eight cysteines and is believed to contain four disulfide bonds.
Arcs indicate the proposed cysteine pairs that form the disulfide bonds
that lead to formation of a hairpin in the hepcidin structure. Reproduced
from [64] with permission of Elsevier.
degraded in lysosomes [61–63], which results in reduced export
of iron because FPN is the only known iron exporter [66, 67].
This is the proposed mechanism by which hepcidin reduces iron
efflux from macrophages, duodenal enterocytes, hepatocytes,
and other target cells.
Although hepcidin‐dependent degradation of FPN is accepted
as a major mechanism by which intestinal iron uptake is
decreased by hepcidin overexpression, other levels of regulation
also exist. For instance, expression of the mRNAs encoding
FPN, DMT1, and DCYTB is increased in enterocytes under
iron‐deficient conditions [68] and this cannot be readily
explained through the direct actions of hepcidin. These changes
may be secondary to reduced iron levels in the enterocytes,
which could influence DMT1 levels via the IRE/IRP system and
the expression of FPN and DCYTB through the hypoxia‐induc-
ible factor HIF-2 [69]. This could also explain why the effects of
hepcidin on macrophages occur before the duodenal mucosa is
affected [70]. Changes in the subcellular localization of FPN
and DMT1 can also occur in enterocytes with changes in iron
levels, with higher levels on the plasma membrane under iron‐
deficient conditions, adding further to the complexity [71, 72].
TRANSCRIPTIONAL REGULATION
OF HEPCIDIN EXPRESSION IS THE MAJOR
DETERMINANT OF SYSTEMIC IRON
HOMEOSTASIS
Hepcidin is the central regulator of systemic iron homeostasis as
it coordinates release of iron to serum from macrophages, intes-
tinal enterocytes, and other body cells. Whereas it was once
believed that sensing of systemic iron homeostasis occurred in
the cells of the duodenal mucosa, numerous studies now indi-
cate that the most important site of sensing is within hepatocytes
[65, 73], with signaling information converging at the hepcidin
promoter. Over the last 15 years a great deal has been learned
about the mechanisms regulating the expression of the HAMP
gene, which encodes hepcidin. The BMP–SMAD signaling
pathway has proved to be essential for both basal and regulated
transcription of HAMP [74], but a number of other pathways are
also involved. These explain how hepcidin expression responds
to changes in iron levels, inflammation, erythropoiesis, and sev-
eral other stimuli. These are summarized in Figure 19.5. A great
deal of our knowledge of hepcidin regulation has come from the
analysis of human disorders of iron homeostasis, and particu-
larly the various forms of iron‐loading disease known as heredi-
tary hemochromatosis. These conditions are briefly summarized
here, but are considered in more detail is subsequent sections.
The essential role of the BMP–SMAD pathway in hepci-
din regulation was identified in family studies of patients
who develop severe juvenile‐onset hemochromatosis due to
mutations in the hemojuvelin gene (HJV) [75]. HJV is a
member of the bone morphogenetic protein receptor (BMPR)
family and the signaling pathways of these receptors have
been studied extensively in other physiological settings. In
general, BMP receptors function by activating SMAD pro-
teins, a family of transcriptional activators/repressors that
220 THE LIVER:  TRANSCRIPTIONAL REGULATION OF HEPCIDIN EXPRESSION IS THE MAJOR DETERMINANT

Figure 19.5  The hepcidin signaling pathway in hepatocytes in response to iron and inflammation. Changes in the expression of hepcidin in
response to altered body iron requirements occur through at least two pathways. One is through the iron‐dependent regulation of BMP6 and BMP2
in hepatic nonparenchymal cells, which subsequently signal through the BMPR/SMAD pathway on hepatocytes to alter HAMP transcription. The
other is through the influence of diferric TF, which likely modulates the interactions between HFE and TFRs 1 and 2. How this in turn alters HAMP
expression is not known, but it is likely to involve effects on BMP/SMAD signaling. Interleukin 6 (IL‐6) signals through a separate surface receptor.
Transcription factors activated by this pathway activate hepcidin transcription by binding to a STAT3‐binding site. Ultimately, hepcidin (HAMP)
transcription increases in hepatocytes that sense high systemic iron or inflammatory signals.

function in numerous pathways to connect receptor‐binding
events to changes in nuclear transcription of target genes
[74]. Notably, animals that lack hepatocyte SMAD4, a pro-
tein that combines with other members of the SMAD family
to regulate transcription of target genes, develop significant
iron overload associated with a profound reduction in hepci-
din expression [76]. Multiple studies have now shown that a
functioning BMP–SMAD pathway is essential for HAMP
transcription, but the pathway is also important for the iron‐
dependent regulation of hepcidin. The levels of two BMPR
ligands, BMP2 and BMP6, have been shown to be increased
when body iron levels rise and these in turn can signal
through the BMP–SMAD pathway to increase HAMP tran-
scription [77–79].
Another link between the BMP–SMAD pathway and hep-
cidin expression is provided by the transmembrane serine
protease TMPRSS6. TMPRSS6 acts to suppress BMP–
SMAD signaling, so when TMPRSS6 is defective in either
mice or humans [80, 81] this suppression is relieved and hep-
cidin levels are constitutively high. This in turn leads to sys-
temic iron deficiency due to reduced dietary iron absorption,
and in humans the condition is known as iron‐refractory iron‐
deficiency anemia [81]. It was originally thought that the
main substrate of TMPRSS6 was HJV [82, 83], but more
recent data have suggested that that the system is far more
complicated and that the enzyme can cleave multiple compo-
nents of the hepcidin regulatory pathway in addition to HJV,
including multiple BMPRs (ALK2, ALK3, ACTRIIA,
BMPR2) as well as HFE and TFR2 [84]. Consistent with
this, TMPRSS6 has been shown to suppress hepcidin expres-
sion independently of HJV [84].
The first gene identified as a cause of hereditary hemochro-
matosis when mutated, HFE [85], is an upstream regulator
of  hepcidin expression. The mechanisms by which HFE
­influences HAMP expression are incompletely understood,
although early studies suggested that it could modulate B­ MP–
SMAD signaling [86, 87]. Patients with HFE‐related
hemochromatosis have reduced hepcidin expression and con-
sequently increased dietary iron absorption, hence the iron
loading [88]. Much rarer than mutations in HFE are those in
the TFR2 gene, but they also lead to reduced hepcidin expres-
sion and hemochromatosis [89], and TFR2 also appears to
affect BMP–SMAD signaling [90]. How HFE and TFR2 reg-
ulate hepcidin expression is only partially understood. HFE is
able to interact with TFR1 and HFE and TF can compete for
binding to TFR1 [91, 92]. An intact HFE/TFR1 complex
appears to be able to reduce hepcidin expression [93], so it
has been proposed that when diferric TF levels are high, the
 19:  The Central Role of the Liver in Iron Storage and Regulation of Systemic Iron Homeostasis 221
complex dissociates and hepcidin levels are increased [65].
Other studies have shown that TFR2 binds HFE when both
proteins are overexpressed [94], and HFE binding appears to
increase the affinity of TFR2 for diferric TF [95]. Thus,
TFR2‐dependent TF binding may play an important role in
hepatocyte iron sensing and systemic iron homeostasis.
Precisely how HFE and TFR2 might work together to alter
hepcidin expression is unclear. One possibility is that under
low iron conditions HFE binds to TFR1, but when the level of
diferric TF rises, it displaces HFE from TFR1 and HFE then
binds to TFR2 and alters hepcidin transcription, possibly by
modulating BMP/SMAD signaling [94, 96]. Interestingly,
when both HFE and TFR2 are knocked out together in mice,
the phenotype is more severe than either of the single knock-
outs, suggesting the two proteins might have at least some
independent effects [97]. Recent data suggest that HFE can
also stabilize some BMP receptors [98], adding a further layer
of complexity to this regulatory system.
Plasma levels of diferric TF have been strongly implicated
in the hepcidin signaling pathway [99]. The degree of satura-
tion of plasma TF saturation integrates the removal and
return of iron to the circulation from many tissue sources into
one common value. Thus, the TF saturation may be the prin-
cipal clinically measurable indicator of iron status that con-
veys information about bodily iron status to multiple tissues
[65, 100–102]. Such a value could feed into the HFE/TFR2‐
mediated pathway of hepcidin regulation, but could also
affect the expression of BMPs 2 and 6. It is likely that multi-
ple pathways are involved. Recently, levels of diferric TF
have also been implicated in mediating (at least in part) the
suppression of HAMP expression in response to stimulated
erythropoiesis [103].
In addition to responding to signaling through the BMP–
SMAD pathway, hepcidin expression is strongly induced by
inflammation (Figure 19.5) [104]. This response of hepcidin
is considered to be a major determinant (though likely not the
only one) that underlies the anemia associated with chronic
inflammation [105, 106]. LPS, as well as the proinflammatory
cytokines interleukin 6 (IL‐6), IL‐1, IL‐22, and interferon‐α
are all able to activate hepcidin expression through a pathway
that involves JAK–STAT signaling and a binding site for the
transcription factor STAT3 in the HAMP promoter [100, 107–
109]. However, the stimulation of HAMP expression in
response to inflammation also requires a functioning BMP/
SMAD pathway as IL‐6 cannot enhance hepcidin expression
in Smad4‐knockout mice [76] or in the absence of the type I
BMPR ALK3 [110].
A final point should be made on the cellular location of the
hepcidin regulatory machinery. Most work has focused on the
hepatocyte as this in the principal site of hepcidin synthesis.
HFE, TFR2, and HJV are also more strongly expressed in hepat-
ocytes than in other types of hepatic cells (such as Kupffer cells,
endothelial cells and stellate cells). However, another key regu-
latory component, BMP6, is synthesized predominantly by
endothelial cells, and to a lesser extent by Kupffer cells and
hepatic stellate cells [111, 112]. This indicates that there is
important intrahepatic crosstalk between different cell types in
the liver, adding another layer of complexity to the hepcidin
regulatory network.
MUTATIONS IN GENES INVOLVED
IN HEPATOCYTE SENSING OF SYSTEMIC
IRON LEVELS CAUSE
HEMOCHROMATOSIS
“Hemochromatosis” is the term for a group of inherited diseases
characterized initially by increased serum iron levels, and later
by parenchymal iron overload in the liver, heart, and endocrine
organs [113]. Macrophages and the duodenal mucosa play
­particularly important roles in the disturbed iron homeostasis
associated with hemochromatosis. Macrophages are normally a
major repository of iron in the body, because they metabolize
senescent red cells and store any iron derived from heme catab-
olism that is not released to plasma transferrin. Whereas the
bone marrow macrophages of non‐hemochromatosis patients
are usually rich in stored iron, as assessed by Prussian blue iron
stain, the macrophages of hemochromatosis patients have
reduced ability to retain iron retrieved from erythrophagocyto-
sis, and consequently hepatic and splenic macrophages contain
little stainable iron. Moreover, despite high serum iron levels,
duodenal iron absorption remains relatively high in hemochro-
matosis patients. As the disease progresses, increased serum
iron levels cause parenchymal iron overload in the liver, heart,
and endocrine organs.
There are now five different genes whose products are recog-
nized to cause forms of genetic hemochromatosis. These are
HFE, TFR2, HJV, FPN (in part), and HAMP, and it is likely that
the list of disease genes will grow as research continues
(Figure 19.6). All of these disorders are characterized by abnor-
mally reduced hepcidin production in conjunction with elevated
plasma transferrin saturation and increased parenchymal rather
than macrophage iron stores. Fortunately, impaired function of
these genes does not adversely affect hematopoiesis. As a result,
iron balance can be restored by regular phlebotomy procedures
to remove red cells rich in hemoglobin iron; these red cells are
readily replaced by increased erythropoiesis.
THE DISCOVERY OF HFE, THE FIRST
IDENTIFIED CAUSE OF GENETIC
HEMOCHROMATOSIS
Although early studies demonstrated that the commonest form of
genetic hemochromatosis was caused by a mutation on chromo-
some 6 based on linkage to the HLA region [114], the affected
gene, HFE, was not identified until 1996, in part because the dis-
ease gene resided in a region of the chromosome that was not
prone to the informative recombination events that facilitate
positional cloning [85]. Initially, a single mutation in the coding
region of HFE that results in substitution of a tyrosine for a
cysteine at codon 282 (C282Y) was recognized, and a PCR test
for the C282Y mutation confirmed that it was widespread in pop-
ulations of Northern European descent [115]. HFE did not belong
to a known gene family, and initial insight into its potential func-
tion came from the observation that it could bind to TFR1 [116]
and interfere with the binding of diferric transferrin [116a].
222 THE LIVER:  TFR2 AND HEMOCHROMATOSIS

Figure 19.6  A model for systemic iron homeostasis. Serum signals reflecting iron stores, erythropoiesis, and inflammation are transduced by
receptors, and information converges on the hepcidin promoter, which determines how much hepcidin is synthesized. Hepcidin binds to ferroportin
on the membrane surface of macrophages, enterocytes, and other target cells, leading to reduced ferroportin expression, and thereby enhancing iron
flow into the plasma. Concomitantly, more iron is sequestered in macrophages and other body cells and duodenal iron absorption is reduced.

However, an animal model indicated that HFE does not exert its
effects on phenotype directly through binding to TFR1 [93], and
subsequent studies showed that HFE functions predominantly as
a regulator of hepcidin expression [47]. As noted above, it may do
this in conjunction with TFR2, which is highly expressed in the
liver where systemic iron sensing occurs [28, 95].
The presence of homozygosity for the C282Y mutation in
HFE predisposes to iron loading, but does not mandate it. Many
C282Y homozygotes have little evidence of iron overload and
few symptoms [117]. One large study concluded that overt
hemochromatosis occurred in 28% of male homozygotes for the
C282Y mutation, but only 1% of female homozygotes [118].
Both environmental and genetic factors can play a role in ame-
liorating the iron‐loading phenotype. Environmental factors,
such as low dietary iron intake, physiological and pathological
blood loss, and blood donation, are likely to be most important
in limiting iron accumulation. However, there are also genetic
modifiers that can influence whether HFE mutations will cause
disease. Examples of these include mutations in the HAMP
[119], DCYTB [120], and GNPAT [121] genes.
Numerous tests have been used to diagnose hemochromato-
sis, including TF saturation, serum ferritin level, liver biopsy
with evidence of parenchymal iron overload and cirrhosis, and
genetic testing for the C282Y mutation. However, a single
serum ferritin determination, with a follow‐up mutation test,
may be the most cost‐effective approach to discovering patients
with HFE‐related hemochromatosis, and serial serum ferritin
measurements can be used to monitor the iron load [117, 122].
TFR2 AND HEMOCHROMATOSIS
Initially discovered as a second TFR with 45% homology to
TFR1, TFR2 differs in that it is mainly expressed in the liver,
with some expression in the brain, heart, and spleen, but little
expression in other tissues [16]. Soon after its initial discovery,
mutations in TFR2 were found to cause hemochromatosis simi-
lar to HFE hemochromatosis in both the histological pattern of
iron overload (parenchymal loading with sparing of macrophages)
and in disease severity. TFR2‐associated hemochromatosis is
more severe than HFE‐related disease, but not as severe as
juvenile hemochromatosis. The disease usually presents in
adults [123, 124], although it sometimes presents in juveniles
[125]. Recapitulation of the essential features of the disease in a
mouse model confirmed that TFR2 was another cause of
classical hemochromatosis [126]. Interestingly, TFR2 levels
increase with binding of TF, due to decreased degradation of
TFR2 in lysosomes, and binding of TF results in more recycling
of TFR2 [28].
 19:  The Central Role of the Liver in Iron Storage and Regulation of Systemic Iron Homeostasis 223
CAUSES OF JUVENILE
HEMOCHROMATOSIS
Among hemochromatosis patients, the severity of disease can
vary substantially. For many years, investigators recognized a
distinct type of hemochromatosis in which severe iron overload
and organ damage occurred before patients reached the age of
30 (i.e. juvenile hemochromatosis). This unusually severe clini-
cal presentation correlates with mutations in genes that encode
proteins that are central to body iron homeostasis: the HAMP
gene that encodes hepcidin and the HJV gene that encodes
hemojuvelin.
Hepcidin mutations cause severe early‐onset
hemochromatosis
As described earlier, hepcidin is now acknowledged to be the
most important regulator of systemic iron homeostasis through
its capacity to regulate iron release from macrophages, duode-
nal enterocytes, and other cells, and it is not surprising that
mutations in hepcidin itself cause a very severe iron‐loading
phenotype [127].
Hemojuvelin mutations cause severe
early‐onset hemochromatosis
Using positional cloning approaches to investigate families
with juvenile hemochromatosis, a gene not previously
­recognized as important in iron homeostasis was discovered
to harbor mutations that cause disease [75, 128]. The gene
product was given the name “hemojuvelin” because of its
causal role in juvenile hemochromatosis, in which patients
typically develop hypogonadotrophic hypogonadism, hepatic
fibrosis or cirrhosis, and cardiomyopathy before age 30. By
sequence homology, HJV belongs in the family of repulsive
guidance molecules (RGMs), proteins that bind to pairs of
type I and type II serine/threonine kinase receptors that are
important in transducing signals from proteins in the
transforming growth factor β (TGF‐β) superfamily. Upon
­
ligand binding, these type I and type II receptor pairs activate
the SMAD proteins, which translocate to the nucleus to mod-
ulate transcription of various target genes after they have
undergone receptor‐dependent phosphorylation. BMPs are
signaling molecules in the TGF‐β superfamily that interact
with a specific set of type I and type II receptor pairs known
as BMP receptors (reviewed in [129]). For some receptors in
the BMP signaling pathway, RGMs bind to the receptor–
ligand complex and potentiate signaling [130]. It appears that
HJV, which is required for the increase of hepcidin transcrip-
tion that occurs in response to iron loading [131], is a co‐
receptor that facilitates BMP signaling in hepatocytes [132].
HJV is linked to membranes through a GPI membrane link-
age, and it specifically facilitates signaling of BMP2 and
BMP4 ligands [133]. Mouse models in which the hemojuvelin
gene is deleted (Hjv−/−) develop parenchymal iron overload typ-
ical of hemochromatosis, but they develop relatively little
organ damage, unlike humans [131, 134], an observation that
remains unexplained.
MUTATIONS OF THE IRON EXPORTER
FERROPORTIN CAUSE A GROUP
OF DISTINCT IRON OVERLOAD
SYNDROMES
Mutations in the gene encoding the iron export protein FPN
(SLC40A1) can also lead to iron loading [135]. Patients with
mutations in SLC40A1 fall into two classes, with distinct pheno-
types depending on the particular mutation they possess [136].
One type of mutation is autosomal dominant and the resulting
syndrome is referred to by some experts as ferroportin disease.
This type of mutation leads to reduced levels of FPN on the
plasma membrane or reduced iron transport capacity. Reduced
iron export means that less iron enters the plasma and that iron
accumulates in tissue macrophages, as internal iron recycling is
disrupted. Thus the characteristic of this type of ferroportin-
linked disease is reduced plasma iron in the face of reticuloen-
dothelial cell iron accumulation. In the liver, iron accumulation
is seen predominantly in Kupffer cells, which is distinct from
the periportal hepatocyte‐dominated distribution observed in
the early stages of HFE‐related hemochromatosis. The reduced
plasma iron can also mean that a mild anemia tends to develop,
especially with phlebotomy therapy. This in turn can stimulate
iron absorption and increase the body iron load. In ferroportin
disease, iron absorption is less efficient due to functional FPN
deficiency, but there still appears to be a net accretion of iron by
the body. The second type of ferroportin mutation is autosomal
recessive and is referred to by some experts as ferroportin‐asso-
ciated hemochromatosis. The condition is characterized by mis-
sense mutations in SLC401 that generate a FPN protein that has
reduced responsiveness to hepcidin. In this situation, the iron
loading is quite similar to that associated with disruption of
HFE, TFR2, hemojuvelin, and hepcidin, with elevated transfer-
rin saturation and periportal deposition of iron in the hepato-
cytes of the liver. The failure of FPN to respond appropriately to
hepcidin means that the body’s normal feedback mechanism for
limiting iron loading is dysfunctional.
SUPPRESSION OF HEPCIDIN EXPRESSION
BY ERYTHROFERRONE IS IMPORTANT
IN IRON‐LOADING ANEMIAS
Patients with anemias characterized by ineffective erythropoiesis,
such as thalassemia major and intermedia syndromes, develop
systemic iron overload in the absence of blood transfusions due
to enhanced intestinal iron absorption [137]. The magnitude of
iron overload is independent of the degree of anemia [138] and
the iron loading is predominantly parenchymal, similar to
hereditary hemochromatosis due to mutations in HFE, TFR2,
HJV, and HAMP. These observations suggested that there is a
common pathophysiologic pathway for the iron loading in ane-
mias characterized by ineffective erythropoiesis and the iron
loading in hereditary hemochromatosis [139]. Indeed, there is
now ample evidence that deficiency of hepcidin with respect to
the body’s iron burden underlies the iron loading seen in both
hereditary hemochromatosis and anemias characterized by
224 THE LIVER:  ACKNOWLEDGMENTS
ineffective erythropoiesis [140]. Investigators recently discov-
ered that erythroferrone, a factor secreted by erythroid progeni-
tors in the bone marrow, is an important mediator of the
suppression of hepatic hepcidin expression that is observed with
increased erythropoiesis, especially ineffective erythropoiesis
[141, 142]. As noted earlier, BMP signaling through phospho-
rylation of SMAD proteins is a central pathway to regulate hep-
cidin transcription [132, 143]. Erythropoietin robustly induces
bone marrow erythroferrone mRNA, and erythroferrone appears
to act through hepatocyte SMAD1 and SMAD5 signaling to
suppress hepcidin production [144] but precisely how erythro-
ferrone interacts with hepatocytes to modulate SMAD signaling
is unknown. Other recent evidence indicates that a reduction in
the level of diferric TF can contribute to the reduction in HAMP
expression associated with stimulated erythropoiesis [103].
ELEVATED HEPCIDIN EXPRESSION
EXPLAINS THE ANEMIA OF CHRONIC
DISEASE
The anemia of chronic disease, better termed the anemia of
chronic inflammation, is characterized by decreased production
of erythrocytes by the bone marrow without underlying nutrient
or hormonal deficiency and with no intrinsic abnormality of the
bone marrow. This hypoproliferative anemia is marked by the
presence of one of several chronic underlying inflammatory dis-
orders, which can be of infectious, malignant, inflammatory, or
traumatic origin. Typically the anemia is mild or moderate in
character, the mean corpuscular volume is mildly reduced, the
serum concentrations of iron and transferrin are decreased, and
the serum ferritin concentration is in the upper end of the normal
range or frankly elevated. Duodenal iron absorption is reduced
and storage of iron in macrophages is increased. The changes in
iron metabolism in the anemia of chronic inflammation are likely
attributable to increased secretion of hepcidin by hepatocytes.
Several experiments indicate that exogenous delivery of hep-
cidin or overexpression of hepcidin can cause chronic anemia.
The injection of hepcidin in mice rapidly leads to a prolonged
hypoferremia [145]. Mice that overexpress hepcidin under the
control of a liver‐specific promoter depend on parenteral iron
injections after birth, suggesting that hepcidin is a negative reg-
ulator of iron transport in the small intestine and that it induces
iron retention in macrophages that normally recycle iron from
senescent erythrocytes [146]. Chronic overexpression of human
hepcidin in mice leads to hypoferremia, increased hepatic iron,
and anemia [147]. Patients with large hepatic adenomas that
produce inappropriately high levels of hepcidin mRNA have
iron‐refractory anemia similar to that observed in anemia of
chronic disease, and this anemia resolves spontaneously after
adenoma resection or liver transplantation [148].
In addition to hepatocytes, monocytes also produce low lev-
els of hepcidin and such autocrine production may contribute to
macrophage iron loading and the anemia of chronic inflamma-
tion [149]. Changes in iron metabolism probably do not fully
explain the anemia of chronic disease. Additional factors may
include suppression of erythroid precursors by inflammatory
cytokines and some limitation to the production of or response
to erythropoietin.
THE LIVER IS IMPORTANT IN CLEARING
REACTIVE HEME FROM CIRCULATION
Hepatocytes are the main site of synthesis of hemopexin, an
abundant serum protein that binds free heme with high affinity
[150]. The structure of hemopexin enables it to sequester heme
in a nonreactive form [151] and heme–hemopexin complexes
are then metabolized by cells that have hemopexin receptors,
including hepatocytes, macrophages, neurons, and syncytio-
trophoblasts [152]. Increased hepatic uptake of hemopexin trig-
gers an increase in hepatic heme oxygenase‐1 expression, which
enables hepatocytes to fully metabolize the newly absorbed
heme. The hemopexin system is a major defense against tissue
injury by free heme, which is important in the pathogenesis of
ischemia–reperfusion and traumatic injuries. Mice engineered
to lack hemopexin are more prone to developing renal failure
after hemolysis than wild‐type mice, indicating that hemopexin
is important for protecting animals from the toxic effects of free
heme in the circulation [153].
Haptoglobin is a tetrameric protein expressed by the liver that
combines with free plasma hemoglobin and protects tissues,
particularly the kidney, from hemoglobin‐dependent damage
[154]. Hemoglobin–haptoglobin complexes are cleared from
the circulation by liver and spleen macrophages, which express
the hemoglobin–haptoglobin receptor CD163 [155]. The liver
protects the other organs against damage from hemolysis and
trauma by synthesizing and secreting hemopexin and haptoglobin,
and also by taking up heme–hemopexin and hemoglobin–
haptoglobin complexes, and catabolizing the heme it acquires
through these pathways using heme oxygenase.
HEPATIC IRON OVERLOAD AND
THE RELATIONSHIP OF IRON
OVERLOAD TO ALCOHOLIC
LIVER DISEASE
A third or more of alcoholics develop increased amounts of
hepatic iron compared to controls [156, 157]. In rural Africa,
there is a strong association among the consumption of a tradi-
tional fermented beverage with high ionic iron concentration,
increased body iron stores, and liver toxicity and cirrhosis
[158–162]. In experimental rats, dietary iron supplementation
exacerbates alcohol‐induced hepatocyte damage and promotes
liver fibrogenesis [163]. Similarly, the presence of elevated
iron stores accentuates the hepatic toxicity of alcohol in the
setting of HFE hemochromatosis in Caucasians [164]. The
increased iron absorption associated with alcohol ingestion
may be related to suppression of hepcidin production by
hepatocytes [165–167].
ACKNOWLEDGMENTS
We thank Helge Uhrigshardt for creating the ferritin structural
representation.
 19:  The Central Role of the Liver in Iron Storage and Regulation of Systemic Iron Homeostasis 225

REFERENCES 28. Johnson, M.B., Chen, J., Murchison, N., Green, F.A., and Enns, C.A.
Transferrin receptor 2: evidence for ligand‐induced stabilization and redirec-
tion to a recycling pathway. Mol Biol Cell, 2007;18(3):743–54.
1. Pritchard, J.A, and Mason, R.A. Iron stores of normal adults and replenish-
29. Zimelman, A.P., Zimmerman, H.J., McLean, R., and Weintraub, L.R. Effect
ment with oral iron therapy. JAMA, 1964;190:897–901.
of iron saturation of transferrin on hepatic iron uptake: an in vitro study.
2. Haskins, D., Stevens, A.R., Jr, Finch, S., and Finch, C.A. Iron metabolism;
Gastroenterology, 1977;72(1):129–31.
iron stores in man as measured by phlebotomy. J Clin Invest, 1952;
30. Craven, C.M., Alexander, J., Eldridge, M. et al. Tissue distribution and clear-
31(6):543–7.
ance kinetics of non‐transferrin‐bound iron in the hypotransferrinemic
3. Walters, G.O., Miller, F.M., and Worwood, M. Serum ferritin concentration
mouse: a rodent model for hemochromatosis. Proc Natl Acad Sci U S A,
and iron stores in normal subjects. J Clin Pathol, 1973;26(10):770–2.
1987;84(10):3457–61.
4. Olsson, K.S. Iron stores in normal men and male blood donors. As measured
31. Brissot, P., Wright, T.L., Ma, W.L., and Weisiger, R.A. Efficient clearance of
by desferrioxamine and quantitative phlebotomy. Acta Med Scand, 1972;
non‐transferrin‐bound iron by rat liver. Implications for hepatic iron loading
192(5):401–7.
in iron overload states. J Clin Invest, 1985;76(4):1463–70.
5. Balcerzak, S.P., Westerman, M.P., Heinle, E.W., and Taylor, F.H. Measurement
32. Wright, T.L., Brissot, P., Ma, W.L., and Weisiger, R.A. Characterization of
of iron stores using deferoxamine. Ann Intern Med, 1968;68(3):518–25.
non‐transferrin‐bound iron clearance by rat liver. J Biol Chem,
6. Harrison, P.M. and Arosio, P. The ferritins: molecular properties, iron storage
1986;261(23):10909–14.
function and cellular regulation. Biochim Biophys Acta, 1996;1275(3):
33. Bernstein, S.E. Hereditary hypotransferrinemia with hemosiderosis, a
161–203.
murine disorder resembling human atransferrinemia. J Lab Clin Med,
7. Theil, E.C. Ferritins. Annu Rev Biochem, 1987;56:289–315.
1987;110(6):690–705.
8. De Domenico, I., Vaughn, M.B., Li, L. et al. Ferroportin‐mediated mobiliza-
34. Hayashi, A., Wada, Y., Suzuki, T., and Shimizu, A. Studies on familial
tion of ferritin iron precedes ferritin degradation by the proteasome. EMBO
hypotransferrinemia: unique clinical course and molecular pathology. Am J
J, 2006;25(22):5396–404.
Hum Genet, 1993;53(1):201–13.
9. Mancias, J.D., Pontano Vaites, L., Nissim, S. et  al. Ferritinophagy via
35. Aisen, P., Leibman, A., and Zweier, J. Stoichiometric and site characteristics
NCOA4 is required for erythropoiesis and is regulated by iron dependent
of the binding of iron to human transferrin. J Biol Chem,
HERC2‐mediated proteolysis. Elife, 2015;4.
1978;253(6):1930–1937.
10. Ha, Y., Shi, D., Small, G.W., Theil, E.C., and Allewell, N.M. Crystal struc-
36. Breuer, W., Ronson, A., Slotki, I.N. et al. The assessment of serum nontrans-
ture of bullfrog M ferritin at 2.8 A resolution: analysis of subunit interactions
ferrin‐bound iron in chelation therapy and iron supplementation. Blood,
and the binuclear metal center. J Biol Inorg Chem, 1999;4(3):243–56.
2000;95(9):2975–82.
11. Wally, J. and Buchanan, S.K. A structural comparison of human serum trans-
37. Hider, R.C. Nature of nontransferrin‐bound iron. Eur J Clin Invest,
ferrin and human lactoferrin. Biometals, 2007;20(3–4):249–62.
2002;32(Suppl 1):50–4.
12. Chua, A.C., Graham, R.M., Trinder, D., and Olynyk, J.K. The regulation of
38. Baker, E., Baker, S.M., and Morgan, E.H. Characterisation of non‐transfer-
cellular iron metabolism. Crit Rev Clin Lab Sci, 2007;44(5–6):413–59.
rin‐bound iron (ferric citrate) uptake by rat hepatocytes in culture. Biochim
13. Ohgami, R.S., Campagna, D.R., McDonald, A., and Fleming, M.D. The
Biophys Acta, 1998;1380(1):21–30.
Steap proteins are metalloreductases. Blood, 2006;108(4):1388–94.
39. Trinder, D., Oates, P.S., Thomas, C., Sadleir, J., and Morgan, E.H.
14. Mims, M.P. and Prchal, J.T. Divalent metal transporter 1. Hematology,
Localisation of divalent metal transporter 1 (DMT1) to the microvillus mem-
2005;10(4):339–45.
brane of rat duodenal enterocytes in iron deficiency, but to hepatocytes in
15. Shaw, G.C., Cope, J.J., Li, L. et al. Mitoferrin is essential for erythroid iron
iron overload. Gut, 2000;46(2):270–6.
assimilation. Nature, 2006;440(7080):96–100.
40. Mims, M.P., Guan, Y., Pospisilova, D. et al. Identification of a human muta-
16. Kawabata, H., Yang, R., Hirama, T. et al. Molecular cloning of transferrin
tion of DMT1 in a patient with microcytic anemia and iron overload. Blood,
receptor 2. A new member of the transferrin receptor‐like family. J Biol
2005;105(3):1337–42.
Chem, 1999;274(30):20826–32.
41. Thompson, K., Molina, R.M., Brain, J.D., and Wessling‐Resnick, M.
17. Chloupkova, M., Zhang, A.S., and Enns, C.A. Stoichiometries of transferrin
Belgrade rats display liver iron loading. J Nutr, 2006;136(12):3010–14.
receptors 1 and 2 in human liver. Blood Cells Mol Dis, 2010;44(1): 28–33.
42. Aydemir, T.B. and Cousins, R.J. The multiple faces of the metal transporter
18. Kleven, M.D., Jue, S., and Enns, C.A. Transferrin receptors TfR1 and TfR2 bind
ZIP14 (SLC39A14). J Nutr, 2018;148(2):174–84.
transferrin through differing mechanisms. Biochemistry, 2018;57(9):1552–9.
43. Jenkitkasemwong, S., Wang, C.Y., Coffey, R. et al. SLC39A14 is required
19. West, A.P., Jr., Bennett, M.J., Sellers, V.M. et al. Comparison of the interactions
for the development of hepatocellular iron overload in murine models of
of transferrin receptor and transferrin receptor 2 with transferrin and the heredi-
hereditary hemochromatosis. Cell Metab, 2015;22(1):138–50.
tary hemochromatosis protein HFE. J Biol Chem, 2000; 275(49):38135–8.
44. Tripathi, A.K., Haldar, S., Qian, J. et al. Prion protein functions as a ferrire-
20. Huebers, H.A. and Finch, C.A. The physiology of transferrin and transferrin
ductase partner for ZIP14 and DMT1. Free Radic Biol Med,
receptors. Physiol Rev, 1987;67(2):520–82.
2015;84:322–30.
21. Herbison, C.E., Thorstensen, K., Chua, A.C. et  al. The role of transferrin
45. Randell, E.W., Parkes, J.G., Olivieri, N.F., and Templeton, D.M. Uptake of
receptor 1 and 2 in transferrin‐bound iron uptake in human hepatoma cells.
non‐transferrin‐bound iron by both reductive and nonreductive processes is
Am J Physiol Cell Physiol, 2009;297(6):C1567–75.
modulated by intracellular iron. J Biol Chem, 1994;269(23):16046–53.
22. Chen, J. and Enns, C.A. Hereditary hemochromatosis and transferrin recep-
46. Zhao, N., Zhang, A.S., Worthen, C., Knutson, M.D., and Enns, C.A. An iron‐
tor 2. Biochim Biophys Acta, 2012;1820(3):256–63.
regulated and glycosylation‐dependent proteasomal degradation pathway for
23. Tacchini, L., Bianchi, L., Bernelli‐Zazzera, A., and Cairo, G. Transferrin
the plasma membrane metal transporter ZIP14. Proc Natl Acad Sci U S A,
receptor induction by hypoxia. HIF‐1‐mediated transcriptional activation
2014;111(25):9175–80.
and cell‐specific post‐transcriptional regulation. J Biol Chem,
47. Brissot, P., Pietrangelo, A., Adams, P.C. et al. Haemochromatosis. Nat Rev
1999;274(34):24142–6.
Dis Primers, 2018;4:18016.
24. Lok, C.N. and Ponka, P. Identification of a hypoxia response element in the
48. Naito, M., Hasegawa, G., Ebe, Y., and Yamamoto, T. Differentiation and
transferrin receptor gene. J Biol Chem, 1999;274(34):24147–52.
function of Kupffer cells. Med Electron Microsc, 2004;37(1):16–28.
25. Addess, K.J., Basilion, J.P., Klausner, R.D., Rouault, T.A., and Pardi, A.
49. McKie, A.T., Marciani, P., Rolfs, A. et al. A novel duodenal iron‐regulated
Structure and dynamics of the iron responsive element RNA: implications
transporter, IREG1, implicated in the basolateral transfer of iron to the circu-
for binding of the RNA by iron regulatory binding proteins. J Mol Biol,
lation. Mol Cell, 2000;5(2):299–309.
1997;274(1):72–83.
50. Donovan, A., Brownlie, A., Zhou, Y. et  al. Positional cloning of zebrafish
26. Ghosh, M.C., Zhang, D.L., and Rouault, T.A. Iron misregulation and neuro-
ferroportin1 identifies a conserved vertebrate iron exporter. Nature,
degenerative disease in mouse models that lack iron regulatory proteins.
2000;403(6771):776–81.
Neurobiol Dis, 2015;81:66–75.
51. Abboud, S. and Haile, D.J. A novel mammalian iron‐regulated protein involved
27. Fleming, R.E., Migas, M.C., Holden, C.C. et al. Transferrin receptor 2: con-
in intracellular iron metabolism. J Biol Chem, 2000;275(26):19906–12.
tinued expression in mouse liver in the face of iron overload and in hereditary
52. Knutson, M. and Wessling‐Resnick, M. Iron metabolism in the reticuloen-
hemochromatosis. Proc Natl Acad Sci U S A, 2000;97(5):2214–19.
dothelial system. Crit Rev Biochem Mol Biol, 2003;38(1):61–88.
226 THE LIVER:  REFERENCES
53. Zhang, A.S., Xiong, S., Tsukamoto, H., and Enns, C.A. Localization of iron
metabolism‐related mRNAs in rat liver indicate that HFE is expressed pre-
dominantly in hepatocytes. Blood, 2004;103(4):1509–14.
54. Hellman, N.E. and Gitlin, J.D. Ceruloplasmin metabolism and function.
Annu Rev Nutr, 2002;22:439–58.
55. Harris, Z.L., Durley, A.P., Man, T.K., and Gitlin, J.D. Targeted gene disrup-
tion reveals an essential role for ceruloplasmin in cellular iron efflux. Proc
Natl Acad Sci U S A, 1999;96(19):10812–17.
56. Xu, X., Pin, S., Gathinji, M., Fuchs, R., and Harris, Z.L. Aceruloplasminemia:
an inherited neurodegenerative disease with impairment of iron homeostasis.
Ann N Y Acad Sci, 2004;1012:299–305.
57. Fuqua, B.K., Lu, Y., Darshan, D. et al. Deletion of the multicopper ferroxi-
dases hephaestin and ceruloplasmin disrupts systemic iron distribution with-
out affecting intestinal iron absorption. Cell Mol Gastroenterol Hepatol,
2018. https://doi.org/10.1016/j.jcmgh.2018.06.006
58. De Domenico, I., Ward, D.M., di Patti, M.C. et  al. Ferroxidase activity is
required for the stability of cell surface ferroportin in cells expressing GPI‐
ceruloplasmin. EMBO J, 2007;26(12):2823–31.
59. Ji, C., Steimle, B.L., Bailey, D.K., and Kosman, D.J. The ferroxidase
hephaestin but not amyloid precursor protein is required for ferroportin‐sup-
ported iron efflux in primary hippocampal neurons. Cell Mol NeuroBiol,
2018;38(4):941–54.
60. Drakesmith, H., Nemeth, E., and Ganz, T. Ironing out ferroportin. Cell
Metab, 2015;22(5):777–87.
61. Nemeth, E., Preza, G.C., Jung, C.L. et  al. The N‐terminus of hepcidin is
essential for its interaction with ferroportin: structure‐function study. Blood,
2006;107(1):328–33.
62. Qiao, B., Sugianto, P., Fung, E. et al. Hepcidin‐induced endocytosis of fer-
roportin is dependent on ferroportin ubiquitination. Cell Metab,
2012;15(6):918–24.
63. Preza, G.C., Pinon, R., Ganz, T., and Nemeth, E. Cellular catabolism of the
iron‐regulatory peptide hormone hepcidin. PLoS One, 2013;8(3):e58934.
64. Clark, R.J., Tan, C.C., Preza, G.C. et al. Understanding the structure/activity
relationships of the iron regulatory peptide hepcidin. Chem Biol,
2011;18:336–43.
65. Frazer, D.M. and Anderson, G.J. The orchestration of body iron intake: how
and where do enterocytes receive their cues? Blood Cells Mol Dis,
2003;30(3):288–97.
66. Nemeth, E., Tuttle, M.S., Powelson, J. et al. Hepcidin regulates cellular iron
efflux by binding to ferroportin and inducing its internalization. Science,
2004;306(5704):2090–3.
67. Delaby, C., Pilard, N., Goncalves, A.S., Beaumont, C., and Canonne‐
Hergaux, F. Presence of the iron exporter ferroportin at the plasma mem-
brane of macrophages is enhanced by iron loading and down‐regulated by
hepcidin. Blood, 2005;106(12):3979–84.
68. Gulec, S., Anderson, G.J., and Collins, J.F. Mechanistic and regulatory
aspects of intestinal iron absorption. Am J Physiol Gastrointest Liver
Physiol, 2014;307(4):G397–409.
69. Mastrogiannaki, M., Matak, P., Keith, B. et  al. HIF‐2alpha, but not
HIF‐1alpha, promotes iron absorption in mice. J Clin Invest,
2009;119(5):1159–66.
70. Chaston, T., Chung, B., Mascarenhas, M. et  al. Evidence for differential
effects of hepcidin in macrophages and intestinal epithelial cells. Gut,
2008;57(3):374–82.
71. Oates, P.S. The role of hepcidin and ferroportin in iron absorption. Histol
Histopathol, 2007;22(7):791–804.
72. Laftah, A.H., Ramesh, B., Simpson, R.J. et al. Effect of hepcidin on intesti-
nal iron absorption in mice. Blood, 2004;103(10):3940–4.
73. Vujic Spasic, M., Kiss, J., Herrmann, T. et  al. Hfe acts in hepatocytes to
prevent hemochromatosis. Cell Metab, 2008;7(2):173–8.
74. Parrow, N.L. and Fleming, R.E. Bone morphogenetic proteins as regulators
of iron metabolism. Annu Rev Nutr, 2014;34:77–94.
75. Papanikolaou, G., Samuels, M.E., Ludwig, E.H. et al. Mutations in HFE2
cause iron overload in chromosome 1q‐linked juvenile hemochromatosis.
Nat Genet, 2004;36(1):77–82.
76. Wang, R.H., Li, C., Xu, X. et  al. A role of SMAD4 in iron metabolism
through the positive regulation of hepcidin expression. Cell Metab,
2005;2(6):399–409.
77. Andriopoulos, B., Jr., Corradini, E., Xia, Y. et al. BMP6 is a key endogenous
regulator of hepcidin expression and iron metabolism. Nat Genet,
2009;41(4):482–7.
78. Meynard, D., Kautz, L., Darnaud, V. et al. Lack of the bone morphogenetic
protein BMP6 induces massive iron overload. Nat Genet, 2009;
41(4):478–81.
79. Canali, S., Wang, C.Y., Zumbrennen‐Bullough, K.B., Bayer, A., and Babitt,
J.L. Bone morphogenetic protein 2 controls iron homeostasis in mice inde-
pendent of Bmp6. Am J Hematol, 2017;92(11):1204–13.
80. Du, X., She, E., Gelbart, T. et al. The serine protease TMPRSS6 is required
to sense iron deficiency. Science, 2008;320(5879):1088–92.
81. Finberg, K.E., Heeney, M.M., Campagna, D.R. et  al. Mutations in
TMPRSS6 cause iron‐refractory iron deficiency anemia (IRIDA). Nat
Genet, 2008;40(5):569–71.
82. Maxson, J.E., Chen, J., Enns, C.A., and Zhang, A.S. Matriptase‐2‐ and propro-
tein convertase‐cleaved forms of hemojuvelin have different roles in the down‐
regulation of hepcidin expression. J Biol Chem, 2010;285(50):39021–8.
83. Silvestri, L., Pagani, A., Nai, A. et  al. The serine protease matriptase‐2
(TMPRSS6) inhibits hepcidin activation by cleaving membrane hemojuve-
lin. Cell Metab, 2008;8(6):502–11.
84. Wahedi, M., Wortham, A.M., Kleven, M.D. et al. Matriptase‐2 suppresses
hepcidin expression by cleaving multiple components of the hepcidin
induction pathway. J Biol Chem, 2017;292(44):18354–71.
85. Feder, J.N., Gnirke, A., Thomas, W. et al. A novel MHC class I‐like gene is
mutated in patients with hereditary haemochromatosis. Nat Genet,
1996;13(4):399–408.
86. Kautz, L., Meynard, D., Besson‐Fournier, C. et al. BMP/Smad signaling is
not enhanced in Hfe‐deficient mice despite increased Bmp6 expression.
Blood, 2009;114(12):2515–20.
87. Ryan, J.D., Ryan, E., Fabre, A., Lawless, M.W., and Crowe, J. Defective
bone morphogenic protein signaling underlies hepcidin deficiency in HFE
hereditary hemochromatosis. Hepatology, 2010;52(4):1266–73.
88. Bridle, K.R., Frazer, D.M., Wilkins, S.J. et al. Disrupted hepcidin regula-
tion in HFE‐associated haemochromatosis and the liver as a regulator of
body iron homoeostasis. Lancet, 2003;361(9358):669–73.
89. Nemeth, E., Roetto, A., Garozzo, G., Ganz, T., and Camaschella, C.
Hepcidin is decreased in TFR2 hemochromatosis. Blood, 2005;105(4):
1803–6.
90. Corradini, E., Rozier, M., Meynard, D. et al. Iron regulation of hepcidin
despite attenuated Smad1,5,8 signaling in mice without transferrin receptor
2 or Hfe. Gastroenterology, 2011;141(5):1907–14.
91. West, A.P., Jr., Giannetti, A.M., Herr, A.B. et al. Mutational analysis of the
transferrin receptor reveals overlapping HFE and transferrin binding sites.
J Mol Biol, 2001;313(2):385–97.
92. Giannetti, A.M. and Bjorkman, P.J. HFE and transferrin directly compete
for transferrin receptor in solution and at the cell surface. J Biol Chem,
2004;279(24):25866–75.
93. Schmidt, P.J., Toran, P.T., Giannetti, A.M., Bjorkman, P.J., and Andrews,
N.C. The transferrin receptor modulates Hfe‐dependent regulation of hep-
cidin expression. Cell Metab, 2008;7(3):205–14.
94. Goswami, T. and Andrews, N.C. Hereditary hemochromatosis protein,
HFE, interaction with transferrin receptor 2 suggests a molecular mecha-
nism for mammalian iron sensing. J Biol Chem, 2006;281(39):28494–8.
95. Waheed, A., Britton, R.S., Grubb, J.H., Sly, W.S., and Fleming, R.E. HFE
association with transferrin receptor 2 increases cellular uptake of transfer-
rin‐bound iron. Arch Biochem Biophys, 2008;474(1):193–7.
96. Zhao, N., Zhang, A.S., and Enns, C.A. Iron regulation by hepcidin. J Clin
Invest, 2013;123(6):2337–43.
97. Wallace, D.F., Summerville, L., Crampton, E.M. et al. Combined deletion
of Hfe and transferrin receptor 2 in mice leads to marked dysregulation of
hepcidin and iron overload. Hepatology, 2009;50(6):1992–2000.
98. Wu, X.G., Wang, Y., Wu, Q. et al. HFE interacts with the BMP type I recep-
tor ALK3 to regulate hepcidin expression. Blood, 2014;124(8):1335–43.
99. Wilkins, S.J., Frazer, D.M., Millard, K.N., McLaren, G.D., and Anderson,
G.J. Iron metabolism in the hemoglobin‐deficit mouse: correlation of difer-
ric transferrin with hepcidin expression. Blood, 2006;107(4):1659–64.
100. Lee, P., Peng, H., Gelbart, T., Wang, L., and Beutler, E. Regulation of hep-
cidin transcription by interleukin‐1 and interleukin‐6. Proc Natl Acad Sci U
S A, 2005;102(6):1906–10.
101. Darshan, D. and Anderson, G.J. Liver‐gut axis in the regulation of iron
homeostasis. World J Gastroenterol, 2007;13(35):4737–45.
102. Anderson, G.J., Darshan, D., Wilkins, S.J., and Frazer, D.M. Regulation of
systemic iron homeostasis: how the body responds to changes in iron
demand. Biometals, 2007;20(3–4):665–74.
 19:  The Central Role of the Liver in Iron Storage and Regulation of Systemic Iron Homeostasis 227
103. Mirciov, C.S.G., Wilkins, S.J., Hung, G.C.C. et al. Circulating iron levels
influence the regulation of hepcidin following stimulated erythropoiesis.
Haematologica, 2018;103(10):1616–26.
104. Nicolas, G., Chauvet, C., Viatte, L. et al. The gene encoding the iron regu-
latory peptide hepcidin is regulated by anemia, hypoxia, and inflamma-
tion. J Clin Invest, 2002;110(7):1037–44.
105. Ganz, T. and Nemeth, E. Iron homeostasis in host defence and inflamma-
tion. Nat Rev Immunol, 2015;15(8):500–10.
106. Kim, A., Fung, E., Parikh, S.G. et al. A mouse model of anemia of inflam-
mation: complex pathogenesis with partial dependence on hepcidin.
Blood, 2014;123(8):1129–36.
107. Wrighting, D.M. and Andrews, N.C. Interleukin‐6 induces hepcidin
expression through STAT3. Blood, 2006;108(9):3204–9.
108. Verga Falzacappa, M.V., Vujic Spasic, M., Kessler, R. et al. STAT3 medi-
ates hepatic hepcidin expression and its inflammatory stimulation. Blood,
2007;109(1):353–8.
109. Schmidt, P.J. Regulation of iron metabolism by hepcidin under conditions
of inflammation. J Biol Chem, 2015;290(31):18975–83.
110. Mayeur, C., Lohmeyer, L.K., Leyton, P. et al. The type I BMP receptor
Alk3 is required for the induction of hepatic hepcidin gene expression by
interleukin‐6. Blood, 2014;123(14):2261–8.
111. Enns, C.A., Ahmed, R., Wang, J. et al. Increased iron loading induces
Bmp6 expression in the non‐parenchymal cells of the liver independent of
the BMP‐signaling pathway. PLoS One. 2013;8(4):e60534.
112. Canali, S., Zumbrennen‐Bullough, K.B., Core, A.B. et  al. Endothelial
cells produce bone morphogenetic protein 6 required for iron homeostasis
in mice. Blood, 2017;129(4):405–14.
113. Pietrangelo, A. Hereditary hemochromatosis. Biochim Biophys Acta,
2006;1763(7):700–10.
114. Simon, M., Bourel, M., Fauchet, R., and Genetet, B. Association of HLA‐
A3 and HLA‐B14 antigens with idiopathic haemochromatosis. Gut,
1976;17(5):332–4.
115. Merryweather‐Clarke, A.T., Pointon, J.J., Shearman, J.D., and Robson,
K.J. Global prevalence of putative haemochromatosis mutations. J Med
Genet, 1997;34(4):275–8.
116. Lebron, J.A., Bennett, M.J., Vaughn, D.E. et al. Crystal structure of the
hemochromatosis protein HFE and characterization of its interaction with
transferrin receptor. Cell, 1998;93(1):111–23.
116a. Lebrón, J.A., West, A.P., and Bjorkman, P.J. The hemochromatosis protein
HFE competes with transferrin for binding to the transferrin receptor.
J Mol Biol, 1999;294:239–45.
117. Waalen, J. and Beutler, E. Hereditary hemochromatosis: screening and
management. Curr Hematol Rep, 2006;5(1):34–40.
118. Allen, K.J., Gurrin, L.C., Constantine, C.C. et al. Iron‐overload‐related dis-
ease in HFE hereditary hemochromatosis. N Engl J Med,
2008;358(3):221–30.
119. Merryweather‐Clarke, A.T., Cadet, E., Bomford, A. et al. Digenic inherit-
ance of mutations in HAMP and HFE results in different types of haemo-
chromatosis. Hum Mol Genet, 2003;12(17):2241–7.
120. Constantine, C.C., Anderson, G.J., Vulpe, C.D. et al. A novel association
between a SNP in CYBRD1 and serum ferritin levels in a cohort study of
HFE hereditary haemochromatosis. Br J Haematol, 2009;147(1):140–9.
121. McLaren, C.E., Emond, M.J., Subramaniam, V.N. et al. Exome sequencing in
HFE C282Y homozygous men with extreme phenotypes identifies a GNPAT
variant associated with severe iron overload. Hepatology, 2015;62(2):429–39.
122. Waalen, J., Felitti, V.J., Gelbart, T., and Beutler, E. Screening for hemo-
chromatosis by measuring ferritin levels: a more effective approach.
Blood, 2008;111(7):3373–6.
123. Camaschella, C., Roetto, A., Cali, A. et al. The gene TFR2 is mutated in a
new type of haemochromatosis mapping to 7q22. Nat Genet, 2000;25(1):
14–15.
124. Roetto, A., Totaro, A., Piperno, A. et al. New mutations inactivating trans-
ferrin receptor 2 in hemochromatosis type 3. Blood, 2001;97(9):2555–60.
125. Loreal, O., Ropert, M., Mosser, A. et al. [Pathophysiology and genetics of clas-
sic HFE (type 1) hemochromatosis]. Presse Med, 2007;36(9 Pt 2):1271–7.
126. Fleming, R.E., Ahmann, J.R., Migas, M.C. et al. Targeted mutagenesis of
the murine transferrin receptor‐2 gene produces hemochromatosis. Proc
Natl Acad Sci U S A, 2002;99(16):10653–8.
127. Roetto, A., Papanikolaou, G., Politou, M. et al. Mutant antimicrobial pep-
tide hepcidin is associated with severe juvenile hemochromatosis. Nat
Genet, 2003;33(1):21–2.
128. Lanzara, C., Roetto, A., Daraio, F. et  al. Spectrum of hemojuvelin gene
mutations in 1q‐linked juvenile hemochromatosis. Blood, 2004;103(11):
4317–21.
129. Herpin, A. and Cunningham, C. Cross‐talk between the bone morphoge-
netic protein pathway and other major signaling pathways results in tightly
regulated cell‐specific outcomes. FEBS J, 2007;274(12):2977–85.
130. Halbrooks, P.J., Ding, R., Wozney, J.M., and Bain, G. Role of RGM core-
ceptors in bone morphogenetic protein signaling. J Mol Signal, 2007;2:4.
131. Niederkofler, V., Salie, R., and Arber, S. Hemojuvelin is essential for die-
tary iron sensing, and its mutation leads to severe iron overload. J Clin
Invest, 2005;115(8):2180–6.
132. Babitt, J.L., Huang, F.W., Wrighting, D.M. et al. Bone morphogenetic pro-
tein signaling by hemojuvelin regulates hepcidin expression. Nat Genet,
2006;38(5):531–9.
133. Lin, L., Valore, E.V., Nemeth, E. et al. Iron transferrin regulates hepcidin
synthesis in primary hepatocyte culture through hemojuvelin and BMP2/4.
Blood, 2007;110(6):2182–9.
134. Huang, F.W., Pinkus, J.L., Pinkus, G.S., Fleming, M.D., and Andrews, N.C. A
mouse model of juvenile hemochromatosis. J Clin Invest, 2005;115(8):
2187–91.
135. Pietrangelo, A. The ferroportin disease. Blood Cells Mol Dis,
2004;32(1):131–8.
136. Pietrangelo, A. Ferroportin disease: pathogenesis, diagnosis and treatment.
Haematologica, 2017;102(12):1972–84.
137. Pootrakul, P., Kitcharoen, K., Yansukon, P. et  al. The effect of erythroid
hyperplasia on iron balance. Blood, 1988;71(4):1124–9.
138. Cazzola, M., Barosi, G., Bergamaschi, G. et al. Iron loading in congenital
dyserythropoietic anaemias and congenital sideroblastic anaemias. Br J
Haematol, 1983;54(4):649–54.
139. Gordeuk, V.R., McLaren, G.D., and Samowitz, W. Etiologies, conse-
quences, and treatment of iron overload. Crit Rev Clin Lab Sci,
1994;31(2):89–133.
140. Camaschella, C. Iron and hepcidin: a story of recycling and balance.
Hematology Am Soc Hematol Educ Program, 2013;2013:1–8.
141. Pak, M., Lopez, M.A., Gabayan, V., Ganz, T., and Rivera, S. Suppression
of  hepcidin during anemia requires erythropoietic activity. Blood,
2006;108(12):3730–5.
142. Kautz, L., Jung, G., Valore, E.V. et al. Identification of erythroferrone as an
erythroid regulator of iron metabolism. Nat Genet, 2014;46(7):678–84.
143. Kautz, L., Meynard, D., Monnier, A. et al. Iron regulates phosphorylation
of Smad1/5/8 and gene expression of Bmp6, Smad7, Id1, and Atoh8 in the
mouse liver. Blood, 2008;112(4):1503–9.
144. Wang, C.Y., Core, A.B., Canali, S. et  al. Smad1/5 is required for
erythropoietin‐mediated suppression of hepcidin in mice. Blood,
­
2017;130(1):73–83.
145. Rivera, S., Nemeth, E., Gabayan, V. et al. Synthetic hepcidin causes rapid
dose‐dependent hypoferremia and is concentrated in ferroportin‐containing
organs. Blood, 2005;106(6):2196–9.
146. Nicolas, G., Bennoun, M., Porteu, A. et al. Severe iron deficiency anemia
in transgenic mice expressing liver hepcidin. Proc Natl Acad Sci U S A,
2002;99(7):4596–601.
147. Rivera, S., Liu, L., Nemeth, E. et al. Hepcidin excess induces the sequestra-
tion of iron and exacerbates tumor‐associated anemia. Blood, 2005;105(4):
1797–802.
148. Weinstein, D.A., Roy, C.N., Fleming, M.D. et al. Inappropriate expression
of hepcidin is associated with iron refractory anemia: implications for the
anemia of chronic disease. Blood, 2002;100(10):3776–81.
149. Theurl, I., Theurl, M., Seifert, M. et al. Autocrine formation of hepcidin
induces iron retention in human monocytes. Blood, 2008;111(4):2392–9.
150. Eskew, J.D., Vanacore, R.M., Sung, L., Morales, P.J., and Smith, A. Cellular
protection mechanisms against extracellular heme. heme‐hemopexin, but
not free heme, activates the N‐terminal c‐jun kinase. J Biol Chem,
1999;274(2):638–48.
151. Baker, H.M., Anderson, B.F., and Baker, E.N. Dealing with iron: common
structural principles in proteins that transport iron and heme. Proc Natl
Acad Sci U S A, 2003;100(7):3579–83.
152. Hvidberg, V., Maniecki, M.B., Jacobsen, C. et al. Identification of the receptor
scavenging hemopexin‐heme complexes. Blood, 2005; 106(7):2572–9.
153. Tolosano, E., Hirsch, E., Patrucco, E. et al. Defective recovery and severe
renal damage after acute hemolysis in hemopexin‐deficient mice. Blood,
1999;94(11):3906–14.
228 THE LIVER:  REFERENCES
154. Fagoonee, S., Gburek, J., Hirsch, E. et  al. Plasma protein haptoglobin
modulates renal iron loading. Am J Pathol, 2005;166(4):973–83.
155. Schaer, D.J., Alayash, A.I., and Buehler, P.W. Gating the radical hemo-
globin to macrophages: the anti‐inflammatory role of CD163, a scavenger
receptor. Antioxid Redox Signal, 2007;9(7):991–9.
156. Fletcher, L.M., Halliday, J.W., and Powell, L.W. Interrelationships of alcohol
and iron in liver disease with particular reference to the iron‐binding proteins,
ferritin and transferrin. J Gastroenterol Hepatol, 1999;14(3):202–14.
157. Suzuki, Y., Saito, H., Suzuki, M. et al. Up‐regulation of transferrin receptor
expression in hepatocytes by habitual alcohol drinking is implicated in
hepatic iron overload in alcoholic liver disease. Alcohol Clin Exp Res,
2002;26(8 Suppl):26S–31S.
158. Gordeuk, V.R., Boyd, R.D., and Brittenham, G.M. Dietary iron overload
persists in rural sub‐Saharan Africa. Lancet, 1986;1(8493):1310–13.
159. Gordeuk, V., Mukiibi, J., Hasstedt, S.J. et  al. Iron overload in Africa.
Interaction between a gene and dietary iron content. N Engl J Med,
1992;326(2):95–100.
160. Friedman, B.M., Baynes, R.D., Bothwell, T.H. et al. Dietary iron overload
in southern African rural blacks. S Afr Med J, 1990;78(6):301–5.
161. Moyo, V.M., Gangaidzo, I.T., Gomo, Z.A. et al. Traditional beer consump-
tion and the iron status of spouse pairs from a rural community in
Zimbabwe. Blood, 1997;89(6):2159–66.
162. Moyo, V.M., Mandishona, E., Hasstedt, S.J. et  al. Evidence of
genetic transmission in African iron overload. Blood, 1998;91(3):
1076–82.
163. Tsukamoto, H., Horne, W., Kamimura, S. et al. Experimental liver cirrhosis
induced by alcohol and iron. J Clin Invest, 1995;96(1):620–30.
164. Fletcher, L.M. and Powell, L.W. Hemochromatosis and alcoholic liver dis-
ease. Alcohol, 2003;30(2):131–6.
165. Kohgo, Y., Ohtake, T., Ikuta, K. et al. Iron accumulation in alcoholic liver
diseases. Alcohol Clin Exp Res, 2005;29(11 Suppl):189S–93S.
166. Flanagan, J.M., Peng, H., and Beutler, E. Effects of alcohol consumption on
iron metabolism in mice with hemochromatosis mutations. Alcohol Clin
Exp Res, 2007;31(1):138–43.
167. Ohtake, T., Saito, H., Hosoki, Y. et al. Hepcidin is down‐regulated in alco-
hol loading. Alcohol Clin Exp Res, 2007;31(1 Suppl):S2–8.
 Disorders of Bilirubin
20 Metabolism
Namita Roy Chowdhury, Yanfeng Li, and Jayanta Roy Chowdhury
Departments of Medicine and Genetics and Marion Bessin Liver Research Center, Albert Einstein College of
Medicine, Bronx, New York, NY, USA
INTRODUCTION
Bilirubin is the end product of degradation of the heme moiety
of hemoproteins. Hemoglobin, derived from senescent erythro-
cytes, is the major source of bilirubin. Significant fractions are
also derived from other hemoproteins of liver and other organs.
Historically, hyperbilirubinemia has attracted the attention of
clinicians as a marker of liver dysfunction. Subsequently, the
studies of bilirubin chemistry, synthesis, transport, metabolism,
distribution, and excretion have provided important insights
into the transport, metabolism, and excretion of biologically
important organic anions, particularly those with limited aque-
ous solubility.
Bilirubin is cytoprotective at relatively low concentrations,
but is potentially toxic at higher levels. It is normally rendered
harmless by tight binding to albumin in the circulation, efficient
extraction and storage by hepatocytes, enzyme‐catalyzed detox-
ification, and subsequent active biliary excretion. Patients with
very high levels of unconjugated hyperbilirubinemia are at risk
for bilirubin encephalopathy (kernicterus). Kernicterus is found
in some cases of severe neonatal jaundice and in inherited disor-
ders associated with severe unconjugated hyperbilirubinemia.
This chapter will provide a brief summary of bilirubin metabo-
lism and its inherited disorders, with emphasis on some signifi-
cant recent developments.
BILIRUBIN METABOLISM
Bilirubin is derived from the breakdown of hemoglobin released
from senescent erythrocytes and other hemoproteins, particu-
larly cytochromes and several enzymes. Of the 250–400 mg of
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
bilirubin produced daily in normal adults, ~80% is derived from
the hemoglobin, while the remainder comes from more rapidly
turning‐over hemoproteins, predominantly in the liver.
Microsomal heme oxygenase (HO) cleaves heme at the α‐
methene bridge by a reaction requiring oxygen and a reducing
agent, such as NADPH. HO activity is rate limiting in bilirubin
production and its inhihition by nonmetabolized “dead‐end”
inhibitors, such as tin‐protoporphyrin or tin‐mesoporphyrin,
reduces serum bilirubin levels in neonates [1, 2]. The reaction
opens the tetrapyrrole ring of heme, giving rise to a linear
tetrapyrrole, biliverdin, and releasing 1 mol each of carbon
monoxide (CO) and iron. Normally, oxidation of the α‐chain
carbon of heme accounts for the great majority of the body’s
endogenous CO production, only a small fraction being contrib-
uted by intestinal bacteria [3]. Therefore, measurement of CO
production by breath analysis can be used to quantify bilirubin
production and heme breakdown, which, at a steady state,
equals heme synthesis.
Of the two HO isoforms, HO‐1, expressed from the HMOX1
gene, is ubiquitous and inducible by protoheme IX and various
stress signals. The kelchlike ECH‐associated protein 1 (KEAP1)–
nuclear factor E2‐related factor 2 (Nrf2) pathway regulates HO‐1
expression. At resting state, KEAP1 degrades NRF2. Oxidative/
electrophilic modification of KEAP1 or its sequestration in
autophagosomes permits stabilization of NRF2 and its transloca-
tion to the nucleus, where it activates target genes, including
HO‐1 [4].
HO‐1 and its products, biliverdin and bilirubin, provide tissue
protection against reactive oxygen radicals, and CO released by
the reaction regulates the vascular tone in the liver and in other
organs, such as the heart, under conditions of stress. HO‐1
induction in the gastrointestinal tract protects against ischemia–
reperfusion injury, lipopolysaccharide (LPS)‐associated sepsis,
230 THE LIVER:  POTENTIAL BENEFICIAL EFFECTS OF BILIRUBIN
as well as tissue injury caused by indometacin, trinitrobenzene
sulfonic acid, and dextran sulfate [5]. Inherited HO‐1 deficiency
is associated with extensive endothelial injury, causing con-
sumption coagulopathy and microangiopathic hemolytic ane-
mia. In addition, these patients have growth retardation and
hyperlipidemia [6]. The cytoprotective function of HO‐1 induc-
tion is also important in the renal vascular endothelium and
tubular epithelium. Some effects of HO‐1 induction, such as
induction of subsets of regulatory T‐lymphocytes may be unre-
lated to its products, CO and bilirubin [7].
The released iron is reutilized, but can be deleterious through
mitochondrial non‐transferrin iron sequestration that can con-
tribute to bioenergy failure, such as in Alzheimer disease. HO‐2
is expressed constitutively, mainly in the brain, where its upreg-
ulation in response to anoxia may have a protective function
during ischemic attacks [8].
In most vertebrates, including early vertebrates, such as
early teleost and elasmobranch fish [9], biliverdin is reduced to
bilirubin by the action of two biliverdin reductases, BVRA and
BVRB. BVRB, which is expressed earlier in embryogenesis,
catalyzes the reduction of biliverdin IXβ, the product of fetal
heme‐IXβ, whereas BVRA mediates the reduction of biliverdin
IXα, the product of adult heme‐IXα [10]. The gene encoding
BVRA is located on chromosome 7 (pter>q22) [11], whereas
the BVRB‐encoding gene is on chromosome 19. The catalytic
site for biliverdin reduction is located within the N‐terminal
domain of both isozymes [12]. However, the C‐terminal domain
of BVRA contains a basic‐leucine‐zipper (bZiP) domain to
function as a transcription factor, which does not exist in
BVRB. The C‐terminal domain of BVRA also contains both a
nuclear localization sequence (NLS) and a nuclear export
sequence (NES), enabling BVRA to bind to DNA sequences
such as the antioxidant response element (ARE) and hypoxia
response elements (HRE), thereby recruiting Nrf2. Nrf2
induces HO‐1, which protects against oxidative injury.
Additionally, BVRA is a member of the insulin receptor sub-
strate family with serine/threonine/tyrosine kinase activity and
through its C‐terminal region interacts with both major arms of
the insulin signaling pathway, namely the phosphatidylinositol
3‐kinase (PI3‐kinase)/Akt pathway and the IRK/IRS/PI3‐
kinase/MAPK pathway [10].
The evolutionary conservation of the energy‐consuming pro-
cess of generation and excretion of the nonpolar bilirubin sug-
gests that the stronger antioxidant activity of bilirubin may be
particularly important during the neonatal period, when concen-
trations of other intracellularly available antioxidants are low in
body fluids. Another potential advantage of bilirubin formation
may be that, being more nonpolar, bilirubin is more efficiently
extracted by the placenta in intrauterine life, although the mech-
anism of bilirubin extraction by the placenta from the fetal cir-
culation has not been elucidated fully. Cord blood bilirubin
concentrations of fetuses born to mothers who have unconju-
gated hyperbilirubinemia due to bilirubin glucuronidation
deficiency are similar to the maternal serum bilirubin levels,
suggesting that the placenta does not pose a barrier to equilibra-
tion of maternal and fetal serum unconjugated bilirubin [13].
As heme oxygenase acts specifically at the α‐bridge of heme,
physiologically generated biliverdin and bilirubin are desig-
nated as biliverdin IXα and bilirubin IXα, respectively. The two
dipyrrolic halves are joined by a central methane bridge. Each
half has a propionic acid side‐chain. X‐ray diffraction crystal-
lography and nuclear magnetic resonance studies have revealed
that the propionic acid side‐chains of bilirubin are internally
hydrogen‐bonded to the pyrrolic and lactam sites on the oppo-
site half of the molecule [14]. The internal hydrogen bonding
engages all polar groups and “buries” the central methane
bridge that joins the two dipyrrolic halves of the molecule.
Physiologically, the hydrogen bonds can be disrupted by
enzyme‐catalyzed conjugation of one or both propionic acid
side‐chains, forming bilirubin mono‐ and diglucuronides,
respectively.
In the van den Bergh reaction [15] that is commonly employed
in clinical analysis of serum bilirubin levels, the diazo reagents,
which attack the central methane bridge, act rapidly on the non‐
hydrogen‐bonded conjugated bilirubin (“direct”‐reacting biliru-
bin), whereas the unconjugated fraction reacts rapidly only
when the hydrogen bonds are disrupted by a chemical accelera-
tor (“total” bilirubin). The hydrogen bonds can also be disrupted
transiently by configurational isomerization of bilirubin, which
occurs upon exposure to light. These bilirubin photoproducts
can be excreted into bile without conjugation, which explains
the efficacy of phototherapy in reducing serum unconjugated
bilirubin levels.
POTENTIAL BENEFICIAL EFFECTS
OF BILIRUBIN
Although clinicians are generally concerned with serum
bilirubin levels as an indicator of liver function and disease,
and the toxicity of bilirubin, within a near‐physiological
range of plasma concentrations, the antioxidative action of
bilirubin may provide beneficial effects. Serum bilirubin
levels were found to be inversely related to the risk of obesity
and metabolic syndrome [16] and ischemic coronary artery
disease in middle‐aged men [17]. An inverse relationship
between serum bilirubin levels and cancer mortality was
reported [18]. The study of a large number of subjects in the
United States revealed that the odds ratio for history of
colorectal cancer was reduced by 0.295 in men and 0.186 in
women per 1 mg dL−1 increment in serum bilirubin concentra-
tions [19]. Subjects with mildly elevated plasma bilirubin lev-
els were found to have lower levels of abdominal obesity and
reduced risk of metabolic syndrome. Consistent with this,
obese individuals with elevated insulin and visceral adiposity
had lower plasma bilirubin [20].
It should be noted that these convincing statistical associa-
tions do not, by themselves, establish a causative role of
bilirubin in reducing the incidence of a number of common
diseases in the population. Mechanistically, HO‐1 induction
by cobalt‐protoporphyrin administration in leptin‐deficient
ob/ob mice led to the recruitment of FGF21, PPARα, and
Glut1, thereby reducing hepatic heme, body weight gain,
plasma glucose, fatty acid synthase, and hepatic steatosis
[21]. Many of these effects could be reproduced by bilirubin
administration in mice with diet‐induced obesity or leptin
receptor deficiency (db/db) [22].
 20:  Disorders of Bilirubin Metabolism 231

HEPATIC METABOLISM because of acquired or inherited liver diseases, a significant

AND ELIMINATION OF BILIRUBIN
Bilirubin circulates in plasma bound to albumin. Albumin bind-
ing keeps unconjugated bilirubin in solution and prevents its dif-
fusion into tissues and all its toxic effects. Unconjugated
bilirubin binds more tightly to albumin than conjugated biliru-
bin. Therefore, in the absence of proteinuria, unconjugated bili-
rubin does not undergo glomerular filtration significantly. As
albumin is normally present in approximately threefold molar
excess to bilirubin, there is a significant reserve bilirubin bind-
ing capacity, which acts as a buffer for fluctuations of serum
bilirubin levels. However, a number of metabolites and drugs
affect albumin binding of bilirubin and thereby risk of neurotox-
icity. Therefore, measurement of unbound plasma bilirubin and
the reserve bilirubin binding capacity could provide a more
accurate estimate of the risk of bilirubin‐induced neurological
damage (BIND). This is particularly important in premature
infants, in whom the threshold total plasma bilirubin concentra-
tions used for instituting phototherapy and/or exchange transfu-
sion in full‐term infants may be misleading. Directly measured
free (unbound) serum bilirubin levels (Bf) are more sensitive
and specific predictors of BIND, as evidenced by incidence of
hearing defect than total serum bilirubin or bilirubin/albumin
ratio [23]. Peroxidase treatment, gel chromatography, electro-
phoretic analysis, and direct fluorometry have been used for Bf
determination [24], However, these approaches are not in clini-
cal use, except for Bf determination by peroxidase treatment,
which is used routinely in Japan [25].
Because of the tight binding of unconjugated bilirubin to
albumin, bilirubin is not excreted in the urine in the absence of
proteinuria. Conjugated bilirubin binds less tightly to albumin.
Therefore, when conjugated bilirubin accumulates in plasma
amount of conjugated bilirubin is excreted in urine. In the pres-
ence of prolonged accumulation of conjugated bilirubin in
plasma, a fraction of the conjugated bilirubin becomes cova-
lently bound to albumin. The covalently bound fraction, termed
delta‐bilirubin, is not excreted in bile or urine, and may persist
in plasma even after biliary obstruction or intrahepatic cholesta-
sis is relieved.
In the liver, bilirubin dissociates from albumin and is internal-
ized by hepatocytes via facilitated diffusion. Although the trans-
port protein SLC21A6 (OATP‐2) was implicated in bilirubin
internalization, this could not be confirmed in subsequent studies
[26]. Within the hepatocyte, binding of bilirubin to glutathione‐
S‐transferases (GSTs) inhibits its efflux, thereby increasing
the net uptake. Microsomal uridinediphosphoglucuronate glucu-
ronosyltransferase type 1 (UGT1A1) catalyzes the transfer of glu-
curonic acid from UDP‐glucuronate to bilirubin, forming
mono‐ and diglucuronides. Glucuronidation makes bilirubin
water soluble, reduces its toxicity, and promotes its secretion into
bile. Glucuronide conjugation is critical in biliary excretion of
bilirubin. Significant reduction of hepatic bilirubin glucuronidat-
ing activity results in the accumulation of unconjugated bilirubin
in plasma (see later). Finally, the bilirubin glucuronides are
transported into the bile canaliculi by an energy‐consuming pro-
cess mediated by ABCC2 (also termed MRP2) (Figure 20.1).
Among the many isoforms of uridinediphosphoglucuro-
nate glucuronosyltransferase (UGT), UGT1A1 is the only
one that contributes significantly to bilirubin glucuronida-
tion. UGT1A1 [27] also accepts estradiol and several drugs
as substrates. UGT1A1 is expressed from an unusually
organized gene that expresses eight UGT1 isoforms from
eight different promoters, each next to a unique exon encod-
ing the N‐terminal half of a specific isoform. The transcript

Bilirubin
Albumin + glucuronides
Albumin-Bilirubin Bilirubin
Sinusoidal 1
surface 6 7
2
3
Contiguous GST’s Bilirubin ABCC3 OATP01B1 OATP01B3
membrane UDPGA
4 UGT1A1
UDP Bilirubin
5 glucuronides
Bilirubin glucuronides ABCC2
(MRP2)
ABCC2 Canalicular
(MRP2) surface

Figure 20.1  Summary of hepatic metabolism of bilirubin. Bilirubin is strongly bound to albumin in the circulation (1). At the sinusoidal surface
of the hepatocyte, this complex dissociates, and bilirubin enters hepatocytes by fascilitated diffusion (2). This process is non‐adenosine triphosphate
(ATP)‐dependent and bidirectional. Within the hepatocyte, bilirubin binds to a group of cytosolic proteins, mainly to glutathione‐S‐transferases
(GSTs) (3). GST binding inhibits the efflux of bilirubin from the cell, thereby increasing the net uptake. A specific form of uridine diphosphoglu-
curonate glucuronosyltransferase, UGT1A1, located in the endoplasmic reticulum, catalyzes the transfer of the glucuronic acid moiety from UDP‐
glucuronic acid (UDPGA) to bilirubin, forming bilirubin glucuronides (diglucuronide and monoglucuronide) (4). Glucuronidation is necessary for
efficient excretion of bilirubin in bile. Canalicular excretion of bilirubin and other organic anions (except most bile acids) is primarily an energy‐
dependent process, mediated by the ATP‐utilizing transporter ABCC2, also known as multidrug‐resistance‐related proteins (MRP2) (5). Excess
bilirubin glucuronides are pumped back into the plasma by ABCC3 located at the sinusoidal membrane (6) and undergo reuptake by hepatocytes
located downstream to the portal blood flow via sinusoidal surface organic anion transport proteins, OATP01B1 and OATP01B3 (7).
232 THE LIVER:  DISORDERS OF BILIRUBIN METABOLISM
initiated from each promoter is spliced to four common
region exons (exons 2–5) that encode the identical C‐termi-
nal half of all UGT1A isoforms [28]. The presence of a dif-
ferent promoter for each isoform permits its independent
regulation. Mutations in any of the five exons (1A1–5)
encoding UGT1A1 can cause complete or partial deficiency
of bilirubin glucuronidation, resulting in Crigler–Najjar
syndrome type 1 (CN1) or Crigler–Najjar syndrome type 2
(CN2), which are characterized by hyperbilirubinemia with
potential cerebral injury [29, 30]. Gilbert syndrome, a milder
and much more common type of inherited hyperbilirubine-
mia, results from a variant TATA element within the
UGT1A1 promoter, which leads to reduced synthesis of
structurally normal UGT1A1 protein [31].
Finally, the bilirubin glucuronides are transported into the
bile canaliculi against a steep concentration gradient by the
ATP‐hydrolyzing canalicular pump ABCC2 (also termed
MRP2), which is rate limiting in bilirubin throughput. Thus,
efficient uptake and conjugation of bilirubin by the periportal
(zone 1) hepatocytes that are initially exposed to bilirubin in
portal blood may saturate the storage capacity of these cells.
A portion of the bilirubin glucuronides produced in these
cells is secreted into the sinusoidal plasma by the multispe-
cific sinusoidal export pump MRP3 (ABCC3). Conjugated
bilirubin flowing toward the central vein undergoes reuptake
via the organic anion‐transporting polypeptides OATP1B1
(SLC01B1) and OATP1B3 (SLC01B3) located in hepatocyte
sinusoidal membranes. This process recruits additional hepat-
ocytes, thereby increasing the bilirubin‐excreting capacity of
the liver [32].
Bilirubin is degraded by intestinal bacteria into a series of
urobilinogen and related products. Most of the urobilinogen
reabsorbed from the intestine is excreted in bile, but a small
fraction is excreted in urine. Absence of urobilinogen in stool
and urine indicates complete obstruction of the bile duct. In
liver disease and states of increased bilirubin production, uri-
nary urobilinogen excretion is increased. Urobilinogen is color-
less; its oxidation product, urobilin, contributes to the color of
normal urine and stool.
Normally, the capacity of the liver for bilirubin uptake,
conjugation, and excretion is closely balanced, so that reduc-
tion of any of these processes can limit the rate of bilirubin
throughput by the liver. On the other hand, increase of the
bilirubin handling capacity of the liver, for example, in
response to increased bilirubin load, requires that all these
pathways are coordinately upregulated. Several nuclear
receptor proteins, such as CAR and PXR, may regulate such
coordinated regulation [33–35].
DISORDERS OF BILIRUBIN
METABOLISM
At higher concentrations, unconjugated bilirubin is toxic to
many cells and organelles. Since albumin binding abrogates the
toxic effect of bilirubin, the harmful effects occur when uncon-
jugated bilirubin is present in a molar excess over albumin. Such
conditions are usually limited to neonates with a high level of
unconjugated hyperbilirubinemia and subjects with inherited
disorders of bilirubin conjugation.
Neonatal hyperbilirubinemia
Normally, newborns have higher serum bilirubin levels than adults.
Eighty percent of all term infants exhibit clinical jaundice during
the first 5 days of life. For clinical management, serum bilirubin
levels in newborns need to be interpreted according to the infant’s
age in hours. Bilirubin levels increase during the first few days of
life, typically peaking at the age of 96 hours, when the 50th percen-
tile in healthy newborns in Western countries ranges from 8 to 9 mg
dL−1 and the 95th percentile is 15–17.5 mg dL−1 [36]. These levels
are considered to be harmless. After this, serum bilirubin levels
decline to less than 1 mg dL−1 in 7–10 days. Exaggeration of this
physiological jaundice increases the risk of BIND.
Neonatal hyperbilirubinemia results from a combination of
increased bilirubin production and lower hepatic bilirubin excre-
tory capacity. Exacerbation of these factors, with or without
additional complicating disorders, increases BIND risk. In
addition to prematurity, common risk factors for severe hyperbili-
rubinemia in babies born at 35 weeks or more gestation are exclusive
breastfeeding (particularly with excessive weight loss), clinical
jaundice noted within the first 24 hours, hemolytic diseases (e.g.
glucose 6‐phosphate dehydrogenase (G6PD) deficiency), and
cephalohematoma or significant bruising during birth [37].
Contribution of genetic factors is suggested by increased
incidence of hyperbilirubinemia in infants of East Asian ethnicity
and history of neonatal jaundice in a previous sibling [38].
Mechanisms of neonatal jaundice are briefly considered here.
Increased bilirubin production
Bilirubin production, as measured by CO excretion in breath, is
increased during the newborn period [39]. The excess bilirubin is
derived from shortened erythrocyte half‐life and also from non-
erythroid sources [40]. Mother–fetus Rh incompatibility has
become infrequent since the availability of anti‐Rh immunoglobu-
lins [41], but ABO blood group incompatibility continues to be a
common cause of exaggerated neonatal hyperbilirubinemia. Other
common hemolytic disorders include sickle cell disease, glucose
6‐phosphate dehydrogenase deficiency, hereditary spherocytosis,
and toxic or allergic drug reactions. Ineffective erythropoiesis, as in
thalassemia, vitamin B12 deficiency, and congenital dyserythro-
poietic anemias can also result in excessive bilirubin production.
Low hepatic bilirubin uptake
Low uptake rate during the first few days of life may result from
delayed closure of the ductus venosus and low levels of cyto-
solic GSTs [42], which increase net bilirubin uptake by reduc-
ing its efflux.
Reduced bilirubin glucuronidation
Both full‐term and premature infants are born with approxi-
mately 1% of adult hepatic UGT1A1 activity, which increases
to adult levels by 14 weeks, regardless of the gestational age at
birth [43]. Inhibition of UGT1A1 exacerbates and prolongs
­neonatal jaundice.
 20:  Disorders of Bilirubin Metabolism 233
Maternal milk jaundice
In general, breastfed infants have higher serum bilirubin levels
than formula‐fed babies [44]. Mild maternal milk jaundice may
abate despite continuation of breastfeeding and even in more
severe cases, resolves promptly upon discontinuation of breast-
feeding, If breastfeeding is continued, the hyperbilirubinemia
may persist for weeks and, in some cases, may increase to 15–24
mg dL−1 by the age of 10–19 days. Although maternal milk jaun-
dice is usually benign [45], kernicterus can occur in rare
cases  [46]. The mechanism of maternal milk jaundice is not
clear. Polyunsaturated free fatty acids produced by lipolytic
enzymes in some maternal milk samples may be responsible for
the inhibition of bilirubin glucuronidation, as suggested by
increased inhibitory effect of maternal milk upon storage and
marked reduction of the inhibitory effect by heating the milk to
56°C [47].
Maternal serum jaundice
Lucey and associates [48] described a syndrome manifested by
moderate to severe unconjugated hyperbilirubinemia (8.9–65
mg dL−1) within the first 4 days of life, which is thought to be
caused by an unidentified inhibitor of UGT1A1 present in
maternal serum. Jaundice may persist several weeks and is
occasionally associated with kernicterus.
Delayed maturation of canalicular bilirubin excretion
Maturation of canalicular excretion mechanism may lag behind
the maturation of uptake and conjugation, so that in the late
newborn period, canalicular excretion becomes rate limiting in
hepatic bilirubin throughput. In these cases, conjugated biliru-
bin may accumulate in serum [49].
Increased intestinal reabsorption
Deconjugation of bilirubin by intestinal β‐glucuronidase
releases unconjugated bilirubin, which is not further degraded
because of the absence of an established intestinal microbiota in
the newborn. This results in increased bilirubin absorption [50],
which may be enhanced by feeding maternal milk.
DISORDERS ASSOCIATED
WITH UNCONJUGATED
HYPERBILIRUBINEMIA
Predominantly unconjugated hyperbilirubia can result from
increased bilirubin production due to hemolytic disorders or
ineffective erythropoiesis. In the presence of normal liver func-
tion, increased bilirubin production does not lead to plasma lev-
els of bilirubin exceeding 5 mg dL−1.
Three inherited disorders associated with UGT1A1 defi-
ciency and consequent reduction of bilirubin glucuronidation
have been described. A near‐complete deficiency of UGT1A1
activity results in Crigler–Najjar syndrome type 1 (CN1); severe
but incomplete deficiency of UGT1A1 activity is termed
Crigler–Najjar syndrome type 2 (CN2), also known as Arias
syndrome; and a mild reduction of UGT1A1 activity results in a
common benign disorder called Gilbert syndrome, which is
characterized by a mild, fluctuating, and often intermittent
unconjugated hyperbilirubinemia.
Crigler–Najjar syndrome type 1
This rare syndrome with autosomal recessive inheritence was
described in 1952 by Crigler and Najjar in six infants from three
unrelated families [51] and was later found to result from lack of
UGT1A1 activity [52]. All patients had lifelong severe non-
hemolytic unconjugated hyperbilirubinemia resulting in bilirubin‐
induced encephalopathy and death within 15 months in five of
the six originally reported cases. The remaining patient devel-
oped kernicterus for the first time at the age of 15 years, and
died six months after that [52]. A related patient remained
without brain damage until 18 years of age, but then developed
kernicterus and died at the age of 24 [53]. As the original fami-
lies described by Crigler and Najjar had a high degree of
consanguinity, several other recessively inherited disorders,
­
including Morquio syndrome, homocystinuria, metachromatic
leukodystrophy, and bird‐headed dwarfism existed in these fam-
ilies. However, in other families CN1 was not associated with
additional inherited disorders. Several hundred CN1 patients
were described subsequently in all races.
After Arias reported a milder variant of this condition (see
below under Crigler–Najjar syndrome type 2), the original
potentially lethal Crigler–Najjar syndrome was designated
Crigler–Najjar syndrome type 1 (CN1), whereas the milder var-
iant of the disease was termed Crigler–Najjar syndrome type 2
(CN2). Jaundice is often the only clinical finding, although
some patients may have residual neurologic abnormalities, from
previous episodes of bilirubin encephalopathy. With routine use
of phototherapy and plasmapheresis during acute bilirubin
encephalopathy, many patients with CN1 now survive beyond
childhood, but many survivors develop kernicterus around
puberty or in early adult life [54]. Orthotopic or auxilliary liver
transplantation results in normalization of serum bilirubin.
Because of the relatively high concentration of unconjugated
bilirubin excreted in bile as a result of phototherapy, pigment
gallstones are common.
Laboratory tests
Serum bilirubin levels usually range from 20 to 25 mg dL−1, but
may reach 50 mg dL−1 [51, 52]. Serum bilirubin is all unconju-
gated and tightly bound to albumin, therefore bilirubinuria is
absent. The bile contains only small amounts of unconjugated
bilirubin [55]. Although fecal urobilinogen excretion is reduced,
the stool color remains normal [51]. Normal bile canalicular
transport is evidenced by normal plasma clearance of bromosul-
fophthalein and indocyanin green, as well as visualization of the
biliary tree by cholecystographic agents [51].
Liver histology
Historically, liver histology has been reported as normal other
than bilirubin plugs in bile canaliculi and bile ducts [51, 55].
However, a recent systematic analysis of 22 CN1 cases under-
going liver transplantation at a single center showed liver fibro-
sis of various degrees in 41% of the explanted livers [56]. The
234 THE LIVER:  DISORDERS ASSOCIATED WITH UNCONJUGATED HYPERBILIRUBINEMIA
liver fibrosis was not associated with portal hypertension and
there was no significant correlation with gallstones. The trans-
plant recipients with liver fibrosis were older in age, suggesting
an accrual of the risk of fibrosis with age.
Abnormalities of hepatic UGTs
Hepatic UGT activity toward bilirubin is virtually absent in all
patients with CN1. In addition, many of these patients have
reduced glucuronidation of phenolic substrates [57], which can
be explained by the location of mutation in UGT1A1 exons (see
later). By immunological analysis, expression of UGT isoforms
may vary in the livers of different CN1 patients [58].
Molecular bases of CN1 and CN2
Description of the layout of the UGT1A locus in 1992 by Ritter
et al. [59] and Bosma et al. [60] explained the structural charac-
teristics and different substrate specificities of various isoforms
of the UGT1A family. Each isoform has a unique N‐terminal
domain but they all have identical C‐terminal domains.
Processed UGT1 mRNAs consist of five exons. The UGT1A
locus comprises four exons encoding the common C‐terminal
domain of all UGT1A isoforms that contains the binding site of
the glucuronic acid donor substrate uridine diphosphoglucu-
ronic acid and the single transmembrane region of the protein.
Twelve unique exon 1 sequences are located 5′ to these common
region exons, only one of which is used in a given UGT1A iso-
form, encoding its unique N‐terminal domain that imparts agly-
cone substrate specificity to specific isoforms. Each unique
exon 1 has a separate proximal upstream promoter, so that
depending on which promoter is used for transcription, RNA
transcripts of different lengths are produced. In each transcript
only the exon 1 located at the 5′ end of the transcript is spliced
to the common region exon 2, and all other exon 1’s are treated
as intervening sequences and deleted. This results in the
translation of nine UGT1 isoforms, of which only UGT1A1
contributes significantly to bilirubin glucuronidation [61].
Consequently, genetic lesions within any of the five exons com-
prising the UGT1A1 gene may lead to complete or near‐com-
plete loss of hepatic bilirubin glucuronidation.
Such genetic lesions may consist of point mutations, dele-
tions, or insertions within the coding region or introns at splice
donor or acceptor sites [29, 54]. The genetic lesions may result
in mutation of a single critical amino acid or deletion of seg-
ments of the enzyme. The various genetic lesions described in
the literature have been reviewed [29, 62]. In cases where the
genetic lesions are present in exons 2–5, all isoforms expressed
from the UGT1A locus are affected, whereas mutation located
in the unique first exon of UGT1A1 affects the gluronidation of
only the substrates of this isoform.
As CN1 or CN2 can be caused by any of a large number of
mutations, deletions, or insertions, no particular mutation is
very common in any ethnic group or community. An exception
to this is seen where there is a strong founder effect and a high
level of consanguinity, such as in the Amish–Mennonite com-
munity, in which there is a high incidence of CN1 and all people
with CN1 carry a specific nonsense mutation in exon 1 of
UGT1A1 [63]. In nearly all cases molecular diagnosis of CN1
and CN2 can be made by sequence determination of products of
polymerase chain reaction (PCR) amplification of the five
UGT1A1 exons and their flanking splice sites. The same strat-
egy can be used for prenatal diagnosis by analyzing DNA
extracted from amniotic cells or chorionic villus samples [64].
As in other recessively inherited inherited disorders, in nearly
all CN1 and CN2 cases, one mutant allele is derived from each
heterozygous parent, giving rise to homozygous or compound
homozygous states. However, a person with CN1 with unipa-
rental isodisomy has been reported, in whom both mutant alleles
were inherited from the father [57]. The mother’s UGT1A1
genotype was normal. This case highlights the desirability of
analyzing the genotype of both parents to determine the mode of
inheritance of CN syndromes.
Animal models of CN1
A mutant strain of Wistar rats exhibiting lifelong nonhemolytic
unconjugated hyperbilirubinemia inherited as an autosomal
recessive characteristic was reported by Gunn in 1938 [65].
Subsequently, the cause of jaundice was found to be deficiency
of UGT‐mediated glucuronidation of bilirubin. Jaundiced Gunn
rats are homozygous for the deletion of a single guanosine resi-
due in the common region exon 4, causing a frameshift and a
premature termination codon that results in the expression of a
truncated protein lacking 150 amino acid residues at the C‐
terminus. This results in the loss of activity of all UGT isoforms
expressed from the UGT1 locus, but UGT isoforms expressed
from other loci (e.g. UGT2) are not affected [66]. Experiments
using Gunn rats have provided important information on biliru-
bin toxicity and have helped in developing novel cell and gene‐
based therapies of CN1 [67–69].
A mouse knockout model of CN1 was created by disrupting
exon 4 of the mouse UGT1 locus [70]. The knockout mice have
higher bilirubin levels than Gunn rats, and have a high spontane-
ous mortality rate unless treated with intensive phototherapy.
The UGT1‐KO mice have enabled the pathophysiological study
of BIND and development of novel therapeutic strategies.
Treatment of CN1
Management of CN1 centers on maintaining serum bilirubin
concentrations below neurotoxic levels. Partial or whole liver
transplantation cures the disease, but commits the pateint to pro-
longed immunosuppressive therapy. Therapies based on hepato-
cyte transplantation and gene therapy are still considered
experimental. A brief discussion of these treatment modalities
follows.
Phototherapy
Phototherapy is routinely used to reduce the level of plasma
unconjugated bilirubin [55]. Banks of fluorescent lamps, or
more recently light‐emitting diode (LED) lamps, are used with
devices for shielding the eyes. LED “light blankets,” or “light
jackets,” have been also devised. Light converts bilirubin IXα‐
ZZ to its photoisomers (see the section on Bilirubin chemistry),
which are excreted in bile and partly degraded. During the neo-
natal period, use of phototherapy is based on age‐related serum
bilirubin concentrations: 15 mg dL−1 (260 μM) at age 24–48
hours, 18 mg dL−1 (310 μM) at 49–72 hours, and 20 mg dL−1
 20:  Disorders of Bilirubin Metabolism 235
(340 μM) above 72 hours [24]. If serum bilirubin remains above
these levels and is not reduced by at least 1–2 mg dL−1 within
4–6 hours despite intensive phototherapy, plasmapheresis is
considered. The efficiency of phototherapy is reduced beyond
the age of 3 or 4 years, because of the thickening of skin, pig-
mentation, and reduction of surface area relative to body mass,
requiring readjustment of photherapy intensity and duration.
Plasmapheresis
During neurologic emergencies, serum bilirubin concentration
can be acutely reduced by plasmapheresis [55]. If serum biliru-
bin levels exceed the target levels described in the previous
paragraph by 5 mg dL−1, despite intensive phototherapy, plas-
mapheresis is added to continued intensive photherapy, because
after removal of albumin‐bound bilirubin from blood, bilirubin
is mobilized from tissue stores to the plasma, leading to second-
ary increase of bilirubin levels.
Orthotopic liver transplantation
At present, the transplantation of whole liver or a segment of the
liver is the only available curative therapy for CN1 [71].
Although this procedure commits the patient to prolonged
immunosuppressive therapy, liver transplantation has dramati-
cally improved the outlook for CN1 patients.
Experimental methods for reducing serum bilirubin levels
Inhibiting heme oxygenase activity
Non‐iron metalloporphyrins are dead‐end inhibitors of microso-
mal heme oxygenase [72]. Tin‐protoporphyrin injection sup-
presses neonatal hyperbilirubinemia in rhesus monkeys [73].
Injection of tin‐mesoporphyrin (0.5 μmol kg−1, three times a
week for 13–23 weeks) in two 17‐year‐old male CN1 patients
reduced serum bilirubin concentrations modestly. However,
the  duration and safety of this treatment of CN1 are not yet
established.
Hepatocyte transplantation
As UGT1A1 activity is present in excess in normal liver, partial
replacement of the enzyme activity in livers of CN1 patients
should reduce serum bilirubin to nontoxic levels. After exten-
sive validation in Gunn rats [68], isolated allogeneic human
hepatocytes were transplanted into the liver of an adolescent
CN1 patient through a catheter placed percutaneously into the
portal vein [74]. Transplantation of 7.5 × 109 hepatocytes
reduced bilirubin levels by about 50% and enabled reduction of
the duration of phototherapy [74]. However, although bilirubin
glucuronides were detectable in bile for up to two and a half
years, serum bilirubin level gradually increased to pretrans-
plantation levels. The patient received an auxiliary liver trans-
plantation, which rapidly reduced the serum bilirubin levels
to normal (J. Roy Chowdhury, personal communication).
Experience in this case and a number of other cases of hepato-
cyte transplantation for CN1, as well as for a number of other
inherited liver diseases indicated that the number of adult
hepatocytes that can be transplanted at a single procedure is
insufficient for fully curing inherited liver‐based metabolic
disorders [75]. Because of this and the increasing shortage of
good‐quality donor livers for hepatocyte isolation, novel strate-
gies are being explored to induce preferential proliferation of
transplanted normal hepatocytes over the mutant host cells. As
adult hepatocytes retain a remarkable capacity to proliferate
and the liver to body weight ratio is regulated tightly by physi-
ological mechanisms, transplanted hepatocytes must compete
with host liver hepatocytes for preferential proliferation.
Controlled regional irradiation of the liver, in combination with
a variety of mitotic stimuli can enhance both initial hepatocyte
engraftment in Gunn rats and subsequent proliferation of the
donor cells, leading to normalization of serum bilirubin levels
[76]. Following initial evaluation in non‐human primates [77],
a clinical trial has been initiated to evaluate preparative hepatic
irradiation for hepatocyte transplantation in patients with inher-
ited metabolic disorders of the liver [78]. Hepatocyte‐like cells
(iHep) generated by differentiating human‐induced pluripotent
stem cells derived by reprogramming somatic cells (e.g. skin
fibroblasts, bone marrow cells, peripheral blood mononuclear
cells, or epithelial cells shed in the urine) can be transplanted
into the livers of experimental animals. Transplantation of
human iHep cells into immunosuppressed Gunn rats subjected
to preparative X‐irradiation resulted in significant reduction of
serum bilirubin levels [69].
Gene therapy
Gene therapy approaches aimed at reconstitution of the missing
UHT1A1 activity include: (i) ex vivo gene therapy, which con-
sists of transduction of primary hepatocytes using viral vectors,
followed by transplantation into the liver; (ii) systemic adminis-
tration of viral or nonviral vectors to transduce hepatocytes in
vivo with transcription units expressing UGT1A1; and (iii) tar-
geted gene editing through homologous recombination for cor-
rection of a specific mutation in the UGT1A1 gene, or for
insertion of a UGT1A1 transcription unit at a genomic “safe
haven” site of choice, or for targeted insertion of a UGT1A1
open reading frame downstream to a highly expressed gene (e.g.
albumin) to take advantage of the strong endogenous promoter.
After validation in preclinical experiments, a clinical trial has
been initiated in CN1 patients using recombinant adeno‐associated
viral vectors.
CN2 (Arias syndrome)
In this milder variant of Crigler–Najjar syndrome described by
Arias in 1962 [79], serum bilirubin, which is mostly unconju-
gated, usually ranges from 8 to 18 mg dL−1. During intercurrent
illness, general anesthesia, or prolonged fasting, serum bilirubin
can increase to 40 mg dL−1 [80]. Kernicterus is unusual, but has
been reported during episodes of exacerbated hyperbilirubine-
mia [79–81]. As in CN1, there is no evidence of hemolysis or
other liver dysfunction. CN2 can be clinically differentiated
from CN1 by at least 25% decrease of serum bilirubin after
induction of the residual UGT1A1 activity by administration of
drugs, such as phenobarbital. Also in contrast to CN1, bile con-
tains a significant amount of bilirubin glucuronides. In CN2, as
in Gilbert syndrome (see later), bilirubin monoglucuronide
exceeds 30% of total conjugated bilirubin (normal, ~10%),
236 THE LIVER:  DISORDERS ASSOCIATED WITH UNCONJUGATED HYPERBILIRUBINEMIA
reflecting a reduced hepatic UGT1A1 activity. Liver histology
is normal, and UGT1A1 activity is usually reduced to 10%
normal [82].
Genetic lesions
Molecular genetic studies are consistent with autosomal reces-
sive inheritance [83]. In CN2, UGT1A1 coding region mutations
always result in single amino acid transitions that significantly
reduce the UGT1A1 activity, without completely abolishing it.
In some cases, the mutation has been shown to increase the Km
for bilirubin [54]. The various lesions causing CN2 have been
reviewed [29].
Gilbert syndrome
Gilbert and Lereboullet described this common disorder, associ-
ated with a mild, fluctuating unconjugated hyperbilirubinemia
in 1901 [84]. More than a century later, Bosma and associates
discovered that this syndrome is caused by a polymorphism in
the proximal promoter region that reduces the expression of
UGT1A1 [85]. Gilbert syndrome is usually diagnosed in young
adults during blood tests for routine check up, screening, or
investigation of unrelated illnesses. In many cases, hyperbiliru-
binemia is intermittent and is usually less than 3 mg dL−1.
Bilirubin levels increase during intercurrent illness, stress, fast-
ing, or menstruation [86]. Mild icterus is the only positive clini-
cal finding, and predominantly unconjugated hyperbilirubinemia
is the only abnormality found on routine blood tests. Some
patients report fatigue and abdominal discomfort, which may be
manifestations of anxiety or superimposed illnesses. Oral chol-
ecystrography visualizes the gallbladder normally, but there
may be a higher incidence of gallstones. Liver biopsy is not nec-
essary for diagnosis, but when performed, shows normal histol-
ogy, except for some nonspecific lipofuscin accumulation in the
centrilobular zone.
Incidence
Gilbert syndrome is one of the most common inherited disorder
in humans, the reported incidence ranging from 3% to 7% of
most populations [87]. As males have higher average serum bili-
rubin levels than females, Gilbert syndrome is diagnosed more
frequency in males [87]. Gilbert syndrome is often recognized
around puberty, which may be related to increased red cell mass
and consequent increased bilirubin production, as well as inhi-
bition of bilirubin glucuronidation by endogenous steroid
hormones.
Diagnosis
Gilbert syndrome is diagnosed in individuals with mild uncon-
jugated hyperbilirubinemia in the absence of normal liver serol-
ogy and without evidence of hemolysis. Hepatic UGT1A1
activity is consistently low (around 30% of normal) in Gilbert
syndrome [88]. Although hemolysis is not a feature of Gilbert
syndrome, coexistent hemolytic disorders, such as glucose 6‐
phosphate dehydrogenase deficiency can make jaundice clini-
cally obvious, thereby bringing the patient to the attention of a
physician. Sequence determination of the TATAA element
within the proximal promoter region of UGT1A1 provides a
simple means of diagnosis (see section on Genetic basis of
Gilbert syndrome). If confirmation of diagnosis is required,
chromatographic analysis of bile for determination of the biliru-
bin monoglucuronide to diglucuronide ratio can be used. As in
CN2, bile in Gilbert syndrome contains an increased proportion
(over 10%) of bilirubin monoglucuronide [82], reflecting
reduced hepatic UGT1A1 activity in these syndromes.
Reduction of caloric intake to 400 kcal day−1 for 2 days or nico-
tinic acid administration [89] increase serum bilirubin levels in
Gilbert syndrome as well as in normal subjects, therefore these
tests do not provide definitive diagnosis. Needle biopsy of the
liver is not recommended for the diagnosis.
Genetic basis of Gilbert syndrome
Gilbert syndrome is associated with a varient TATAA box in the
promoter upstream to exon 1 of UGT1A1. The normal TATAA
element has the sequence A[TA]6TAA, whereas subjects with
Gilbert syndrome are homozygous for insertion of two nucleo-
tides, making its sequence T[TA]7TAA, which reduces the
expression of UGT1A1 to approximately 30% of normal [85].
The Gilbert‐type TATAA element has been designated as
UGT1A1*28 [90]. Approximately 9% of most populations are
homozygous for UGT1A1*28 and about 42% are heterozygous
carriers. However, all subjects who are homozygous for the
UGT1A1*28 allele do not exhibit the clinical phenotype,
indicating that other variables, particularly the rate of bilirubin
production may be necessary for manifestation of hyperbiliru-
binemia. Thus, although the inheritance is autosomal (chromo-
some 2q37), jaundice is uncomon in women, probably because
of lower daily bilirubin production. The gene frequency for
UGT1A1*28 may be lower in Japan. Some UGT1A1 coding
region mutations may cause mild hyperbilirubinemia, compati-
ble with the clinical diagnosis of Gilbert syndrome [91, 92].
These mutations have been reported only in populations of East
Asian origin. Some of these mutations were rported to be domi-
nant negative, suggesting that they reduce the activity of a nor-
mal allele [93].
Some heterozygous carriers of structural mutations (CN1 or
CN2) may also carry the Gilbert‐type promoter on the structur-
ally normal allele, which reduces the expression of the only
structurally normal allele. Such combinations may give rise to
intermediate levels of hyperbilirubinemia [94, 95], explaining
the long‐standing observation that intermediate levels of hyper-
bilirubinemia are common among relatives of patients with
CN1 or CN2. Because of the rarity of CN1 and CN2 mutations
and the very high frequency of the UGT1A1*28 allele, this type
of combination is a more common cause of intermediate levels
of hyperbilirubinemia than the presence of a coding region
mutation on both UGT1A1 alleles.
Health implications of Gilbert syndrome
Gilbert syndrome is innocuous, but its recognition is considered
important mainly for reassuring the patient and the physician
that no underlying liver disease is responsible for the mild
hyperbilirubinemia. In fact, mild elevation of serum bilirubin
may have health benefits (see section on Potential beneficial
effects of bilirubin). However, reduced UGT1A1 activity can
 20:  Disorders of Bilirubin Metabolism 237
affect detoxification of certain drugs [96]. Gilbert syndrome is
associated with a high incidence of diarrhea in patients treated
with the anticancer drug irinotecan [97]. Oxidative paracetamol
metabolism is associated with drug toxicity. There is conflicting
evidence in literature for increased oxidative metabolism of par-
acetamol (acetaminophen) and its decreased glucuronidation in
subjects with the UGT1A1*28 allele [98, 99]. Nilotinib and
probably other tyrosine kinase inhibitors used in the treatment
of leukemia are not metabolized by UGT1A1, but may inhibit
the enzyme activity, thereby enhancing hyperbilirubinemia in
patients with Gilbert syndrome, as well as CN2 [100].
Animal model
The Bolivian population of squirrel monkeys (Saimiri sciureus)
has higher serum unconjugated bilirubin concentrations and a
greater hyperbilirubinemic response to fasting than does a
closely related Brazilian population [101, 102]. Plasma clear-
ance of intravenously administered bilirubin is lower in the
Bolivian squirrel monkey population. As in subjects with Gilbert
syndrome, hepatic UGT activity toward bilirubin, and bilirubin
diglucuronide:monoglucuronide ratio in bile are lower in the
Bolivian population. Fasting hyperbilirubinemia is rapidly
reversed by oral or intravenous administration of carbohydrates,
but not lipids [101].
DISORDERS PREDOMINANTLY
ASSOCIATED WITH CONJUGATED
HYPERBILIRUBINEMIA
Normally, by chromatographic measurement, approximately
4% of bilirubin in plasma is conjugated. The proportion of con-
jugated bilirubin increases in plasma when bilirubin glucuron-
ides formed inside hepatocytes are transferred back to plasma as
a result of biliary obstruction, inflammatory or ischemic injury
of the liver, inherited disorders of bile canalicular excretion, or
failure of reuptake of bilirubin glucuronides transported out of
the hepatocytes into sinusoidal blood. In addition, there are sev-
eral disorders, collectively termed progressive familial intrahe-
patic cholestasis, that are caused by inherited abnormalities of
bile canalicular transport proteins or a tight junction protein. In
another genetic condition, paucity of bile ducts leads to conju-
gated hyperbilirubinemia. A brief description of these disorders
follows.
Dubin–Johnson syndrome
Dubin and Johnson [103] and Sprinz and Nelson [104] described
a syndrome characterized by chronic conjugated hyperbiliru-
binemia and grossly pigmented, but otherwise histologically
normal liver. Mild icterus is the only consistent physical finding,
but rarely, hepatosplenomegaly has been observed [105, 106].
Patients are usually asymptomatic, but an occasional patient
complains of weakness and vague abdominal pain. Serum bile
acid levels are nearly normal [105], and pruritus is absent.
Serum bilirubin levels are increased during intercurrent illness,
intake of oral contraceptives, and pregnancy [105]. Diagnosis is
usually made after puberty and, in some cases, during pregnancy
or contraceptive use [105, 107].
Laboratory tests
Liver function tests, including serum bile acid levels, are normal
[107]. Serum bilirubin levels usually range from 2 to 5 mg dL−1,
but can rarely reach 20–25 mg dL−1. Over 50% of total serum
bilirubin is direct‐reacting, and bilirubin is excreted in urine.
Because the hepatocellular canalicular excretion is abnormal for
many organic anions, except bile acids, oral cholecystography,
even using a “double dose” of the contrast material, fails to visu-
alize the gallbladder. Gallbladder may be visualized 4 to 6 hours
after intravenous administration of meglumine iodipamide
(Biligrafin) [108]. Macroscopically, the liver is black, and light
microscopy reveals a dense pigment. After intravenous infusion
of 3H‐epinephrine into mutant Corriedale sheep (an animal
model for Dubin–Johnson syndrome), radioactivity is incorpo-
rated into the dark brown pigment [109], which is not authentic
melanin, but may consist of polymers of epinephrine metabo-
lites [110]. Following liver regeneration after hepatocellular
apoptosis, such as after acute viral hepatitis, the pigment is
cleared from the liver and reaccumulates slowly after recovery.
Organic anion transport
Organic anions other than bile acids are transported into the bile
canaliculus from the hepatocyte against a concentration gradi-
ent by an ATP‐dependent energy‐consuming process, mediated
by a protein that had been originally termed the canalicular mul-
tispecific organic anion transporter (cMOAT), and is now
termed MRP2 or ABCC2 [111–114]. These anions include bili-
rubin glucuronides, the leukotriene LTC4, oxidized and reduced
glutathione, and numerous glucuronide and glutathione conju-
gates. In contrast, with some exceptions, bile acids are secreted
normally. MRP2 is one of the ABC transporters [112]. Direct
evidence for its involvement in canalicular transport came from
the discovery of a frameshift mutation in the gene encoding
Mrp2 in the TR− rat [115]. Studies in TR− mice indicated that
some sulfated or glucuronidated bile acids require Mrp2 for
biliary excretion [116]. Consistent with this, in a multicenter
study in Japan, neonates with Dubin–Johnson syndrome had
significantly increased serum total bile acid levels [117].
Despite the lack of MRP2 function, the serum bilirubin is
only mildly elevated in Dubin–Johnson syndrome, suggesting
the presence of additional mechanisms of biliary excretion of
bilirubin conjugates. In addition, accumulation of organic ani-
ons within the hepatocytes due to MRP2 deficiency leads to
upregulation of the expression of MRP1 and MRP3 in the baso-
lateral surface of the hepatocytes. These and possibly additional
ATP‐consuming pumps may actively export both unconjugated
and conjugated bilirubin from the hepatocyte to plasma via the
space of Disse [118].
After intravenous injection of the organic anion bromosul-
fophthalein (BSP), plasma BSP concentration decreases at near‐
normal rate for 45 minutes, indicating normal uptake across the
sinusoidal surface of hepatocytes and normal storage capacity
of hepatocytes. However, in 90% of patients, plasma BSP con-
centration exhibits a secondary increase, so that plasma BSP
concentration at 90 minutes exceeds that at 45 minutes. This
238 THE LIVER:  DISORDERS PREDOMINANTLY ASSOCIATED WITH CONJUGATED HYPERBILIRUBINEMIA
secondary rise is due to reflux of glutathione‐conjugated BSP
from hepatocytes into the circulation [119]. A similar secondary
rise occurs after intravenous administration of bilirubin.
However, such secondary rise of plasma BSP can also occur in
some other hepatobiliary disorders [120], therefore, this phe-
nomenon is not pathognomonic of Dubin–Johnson syndrome.
Urinary coproporphyrin excretion
Coproporphyrins excreted in urine consists of two isomers,
coproporphyrin I and coporophyrin III. Normally, in adults
approximately 75% of the urinary coproporphyrin is isomer III,
which is the precursor of heme. Total urinary coproporphyrin
excretion is normal in Dubin–Johnson syndrome, but over 80%
is isomer I [121]. Neonates have a higher proportion of urinary
coproporphyrin I than adults, but the proportion is not as high as
in Dubin–Johnson syndrome. The relationship of the abnormal
pattern of urinary porphyrin excretion to the organic anion
transport defect is not fully understood. When correlated with
history and physical examination, the urinary coproporphyrin
excretion pattern is diagnostic of Dubin–Johnson syndrome.
Genetic basis and inheritance
Dubin–Johnson syndrome is inherited as an autosomal recessive
trait, and has been reported in all races and both sexes. It is rare
in all races, except in Iranian, Iraqi, and Moroccan Jews living
in Israel, who have an incidence of 1 in 1300 [106], in whom it
is associated with clotting factor VII deficiency [122]. Dubin–
Johnson syndrome may be relatively common in some areas of
Japan, where consanguinity is frequent.
Dubin–Johnson syndrome is caused by insertions, deletions,
and nonsense mutations of the MRP2 (ABCC2) gene, or abnor-
mal splicing of the RNA transcript that causes loss of MRP2
expression, or missense mutations of the gene that interfere with
normal localization of MRP2 in the bile canalicular plasma
membrane of hepatocytes [115, 117, 123, 124]. Human MRP2
gene has been localized to chromosome 10q23–q24 [125].
Single amino acid transitions most commonly involve the cru-
cial ATP‐binding region. Some mutations may lead to impaired
glycosylation of MRP2, leading to premature proteasome‐
dependent degradation [126].
Animal models
In mutant Corriedale sheep, biliary excretion of conjugated biliru-
bin, glutathione‐conjugated BSP, iopanoic acid, and indocyanine
green is decreased, whereas the transport of taurocholate [127]
and unconjugaed BSP [128] is normal. The secretion of organic
cations, such as procaine amide ethobromide, is unaffected.
Serum bilirubin is mildly increased, with 60% of the bilirubin
being conjugated. Other than dark brown pigmentation, liver his-
tology is normal [109]. As in human Dubin–Johnson syndrome,
total urinary coproporphyrin excretion is normal, but the propor-
tion of coproporphyrin I is increased. Thus, the mutant Corriedale
sheep is a model of human Dubin–Johnson syndrome.
As in humans with Dubin–Johnson syndrome, Eisai hyper-
bilirubinemic (EHBR) and TR− rats lack Mrp2 and exhibit
reduced biliary excretion of conjugated bilirubin, leukotriene
LTC4 and glutathione, as glucuronic acid and glutathione
conjugates [129] and many other organic anions [130],
Coproporphyrin I is the major porphyrin isomer excreted in
urine [131]. The liver is not normally pigmented, but intracel-
lular pigments are deposited upon feeding a diet enriched in
tryptophan, tyrosine, and phenylalanine. Impaired excretion of
anionic metabolites of these amino acids may result in their
retention, oxidation, polymerization, and subsequent lysosomal
accumulation in hepatocytes [132]. Experiments in these rat
models showed that the pathway for secretion of bilirubin con-
jugates and many other organic anions into the bile canaliculus
differed from that for secretion of most bile acids. Bile acids
with free 3‐OH groups are transported by the bile salt export
pump (BSEP), but sulfated or glucuronide‐conjugated bile acids
are transported by Mrp2 [133].
The golden lion tamarin monkey (Leontopithecus rosalia),
which manifests elevated serum conjugated bilirubin levels, is a
non‐human primate model of Dubin–Johnson syndrome [134].
Rotor syndrome
Rotor, Manahan, and Florentin reported two families with sev-
eral patients with lifelong predominantly conjugated hyperbili-
rubinemia without evidence of hemolysis [135]. Other routine
blood biochemistries and hematologic tests were normal and, in
contrast to Dubin–Johnson syndrome, there was no pigmenta-
tion in the liver. Liver histology was normal. Rotor syndrome is
harmless [135]. Although rare, it has been described in several
races.
Organic anion excretion
The organic anion excretion defect in Rotor syndrome is differ-
ent from that in Dubin–Johnson syndrome. As in many acquired
liver diseases, over 25% of injected BSP is retained in serum at
45 minutes, and there is no secondary rise of plasma BSP level
[136]. Plasma clearance of intravenously administrated uncon-
jugated bilirubin and indocyanine green is also delayed. In con-
trast to the findings in Dubin–Johnson syndrome, bile canalicular
export pumps are normal in Rotor syndrome, therefore, the gall-
bladder is visualized by oral cholecystography [137, 137A].
Urinary coproporphyrin excretion
Total urinary coproporphyrin is increased two‐ to fivefold over
normal in Rotor syndrome, of which approximately 65% is iso-
mer I [138]. These findings are similar to those seen in many
other hepatobiliary disorders, and distinguish Rotor syndrome
from Dubin–Johnson syndrome. However, two brothers with
clinical features of Rotor syndrome were reported to have over
80% of urinary coproporphyrins as isomer I [139], raising doubts
about the usefulness of using urinary coproporphyrin analysis in
distinguishing the two syndromes. More recently, the discovery
of the genetic basis of Rotor syndrome has clarified the defect in
organic anion handling by hepatocytes in this disorder (see later).
Molecular basis of Rotor syndrome
Van de Steeg et  al. found a tight linkage between simultaneous
mutations in the SLC01B1 and SLC01B3 genes and Rotor syn-
drome in six families [140]. These mutations result in deficiency
 20:  Disorders of Bilirubin Metabolism 239
of the organic anion transporting polypeptides OATP1B1 and
OATP1B3, respectively. Using mice deficient in the multispecific
sinusoidal export pump Abcc3, as well as Oatp1a/1b, van de Steeg
et  al. demonstrated that Abcc3 secretes bilirubin glucuronides
from hepatocytes to blood, whereas Oatp1a/1b mediate their reup-
take. Because of the efficient uptake and conjugation of bilirubin
by zone 1 hepatocytes from blood entering the hepatic sinusoids,
these bilirubin glucuronides formed in these cells may exceed
their canalicular excretory capacity. A portion of bilirubin glucuro-
nides formed in the zone 1 hepatocytes is transported back into the
sinusoidal blood and is subsequently taken up by hepatocytes
located downstream to the blood flow (toward Zzone 3) via
Oatp1b1/3. By recruiting additional hepatocytes, this mechanism
increases the capacity of liver to handle a bilirubin load.
In humans, OATP1B1 and OATP1B3 belong to the 11‐member
OATP family that have 12 membrane‐spanning domains and
mediate the uptake of a wide variety of compounds. They are
almost exclusively localized in the sinusoidal domain of hepato-
cyte plasma membranes. OATP1B1 and OATP1B3 transport dif-
ferent, but partly overlapping substrates. In addition to bilirubin
glucururonides, substrates for both transporters include drugs such
as hydroxymethylglutaryl (HMG)‐CoA reductase inhibitors
(statins), angiotensin II receptor blockers (sartans), angiotensin‐
converting enzyme (ACE) inhibitors, and antidiabetes (glinides)
as substrates [141]. In addition to organic anions, OATP1B1 and
OATP1B3 accept neutral compunds, such as digoxin, ouabain,
lopinavir, as well as zwitter ionic drugs such as fexofenadine.
OATP1B1 preferentially recognizes estrone‐3‐sulfate , whereas
cholecystokinin octapeptide CCK‐8, telmisartan, paclitaxel, and
docetaxel are specifically transported via OATP1B3. OATPs are
structurally unrelated to ABCC2 (MRP2), which is located in the
bile canalicular domain of hepatocyte plasma membranes, but
they share many of their substrates, including bilirubin glucuron-
ides, which coordinates hepatocellular uptake with and canalicular
excretion. For certain substrates, OATP1B1 and OATP1B3 cannot
compensate for the lack of each other. For example, simvastatin‐
associated myopathy is strongly linked to a single nucleotide poly-
morphism of SLC01B1 [142]. On the other hand, for the reuptake
of bilirubin glucuronides, OATP1B1 and OATP1B3 can compen-
sate for the lack of one another, whereby simultaneous mutation of
both is required for the manifestation of Rotor syndrome.
Animal model
A subgroup of Southdown sheep exhibits mild conjugated
hyperbilirubinemia and photosensitivity [143]. Southdown
sheep with these clinical features were found to have a missense
(glycine to arginine) mutation in Slc01b3 [144]. This glycine
residue is conserved in seven other mammalian species and is
thought to be functionally critical.
INHERITED CHOLESTASIS SYNDROMES
Progressive familial intrahepatic cholestasis
Three life‐threatening disorders, collectively termed progressive
familial intrahepatic cholestasis (PFIC1, PFIC2, PFIC3, and
PFIC4), can cause cholestasis of varying severities by affecting
the secretion of bile acids or other components of bile across the
bile canaliculi of hepatocytes [145]. Benign recurrent intrahe-
patic cholestasis (BRIC) is a disorder that is genetically related
to PFIC1 or PFIC2. The PFIC syndromes usually present during
infancy or childhood, and often lead to growth failure and pro-
gressive liver disease. Several additional genetic cholestatic
disorders have been described during the last few years. PFIC
syndromes and the other inherited cholestatic diseases have been
reviewed recently [146] and are discussed in brief here.
Progressive familial intrahepatic cholestasis type 1
PFIC1 was originally described in an Amish–Mennonite family
and was named Byler disease, after the family name of the first
patient [147]. PFIC1 is associated with severe life‐threatening
cholestasis. It is caused by mutations in the P‐type ATPase gene
FIC1 (also termed ATP8B1), which is located on chromosome
18q21 [148]. FIC1‐mediated ATP hydrolysis is coupled with the
translocation of acidic phospholipids. How FIC1 mutation
causes cholestasis is not fully understood [149]. FIC1 defi-
ciency may interfere with the nuclear translocation of the
nuclear receptor FXR [150], thereby downregulating the expres-
sion of the bile salt export protein (BSEP) (see section on
Progressive familial intrahepatic cholestasis type 2).
Benign recurrent intrahepatic cholestasis
BRIC was first described in 1959 [151]. It presents in adoles-
cence or early adulthood with recurrent episodes of cholestasis,
characterized by conjugated hyperbilirubinemia, together with
malaise, anorexia, pruritus, weight loss, and malabsorption.
During the attacks, which last weeks to months, laboratory tests
reveal biochemical evidence of cholestasis without severe hepa-
tocellular injury [152–154]. These attacks are followed by a
complete clinical, biochemical, and histological remission. The
clinical features and duration of the episodes in a given patient
resemble those in previous attacks. Liver biopsy shows non-
inflammatory intrahepatic cholestasis without fibrosis, not-
withstanding the number and severity of the attacks. During
remission, light or electron microscopy of the liver tissue returns
to normal [155]. Like PFIC1, the inheritance shows an autoso-
mal recessive pattern. Interestingly, this relatively benign disor-
der is also caused by certain missense mutations in the FIC1
gene, other mutations of which cause the much more severe
­disorder PFIC1.
There is no specific treatment for BRIC. Some cases are asso-
ciated with mutations of ABCB11, which is associated with
PFIC2. Because of this, some authors classify BRIC into BRIC‐I
and BRIC‐II.
Progressive familial intrahepatic cholestasis type 2
This disorder, which resembles Byler disease clinically, occurs in
mainly in Middle Eastern and European races. It is caused by
defects of the ABCB11 gene that encodes the bile salt export
pump (BSEP), previously termed sister P‐glycoprotein. BSEP is
a canalicular ATP‐dependent transporter of bile acids from the
hepatocytes into the bile [156]. The ABCB11 gene is located on
chromosome 2q24 [157]. A large number of different point muta-
tions in ABCB11 have been described in patients with PFIC2.
240 THE LIVER:  ACKNOWLEDGMENT
Although both PFIC1 and PFIC2 are associated with life‐
threatening cholestasis, serum γ‐glutamyl transpeptidase (GGT)
levels are normal or nearly so in both disorders, which differen-
tiates them from PFIC3 [145, 157].
BSEP is expressed specifically in the liver and, as expected,
liver transplantation ameliorates all manifestations of PFIC2.
However, in several cases, PFIC‐like symptoms recur after liver
transplantation consequent to development of immunoglobulin
G‐type antibodies against BSEP. Although antibodies develop
against both the N‐terminal and the C‐terminal regions of the
protein, only sera recognizing the first extracellular loop (ECL1)
inhibit transepithelial taurocholate transport [158].
Progressive familial intrahepatic cholestasis type 3
PFIC3 involves mutations in ABCB4, an ATP‐utilizing bile can-
alicular pump (also known as the multidrug resistance protein‐3
P‐glycoprotein (PGY3, MDR3). ABCB4 transports phosphati-
dylcholine from the inner lipid leaflet of the bile canaliculus to
the outer leaflet, thereby replenishing the phospholipid in the
outer leaflet, which is continuously removed by contact with
bile acids. Failure of phosphatidylcholine translocation, the
canalicular membrane, and small bile ducts are chronically
injured by bile salts [158, 159]. In contrast to PFIC1 and PFIC2,
in which the canalicular membrane is protected by reduced
secretion of bile salts in the canalicular bile, serum GGT activity
is increased in PFIC3 [145]. Human ABCB4 deficiency can pre-
sent with a variety of hepatobiliary disorders, including small‐
duct primary sclerosing cholangitis [160] and cholesterol
gallstones [161]. Interestingly, asymptomatic female heterozy-
gous carriers of nonsense or missense mutations of the ABCB4
gene have been reported to manifest familial intrahepatic chol-
estasis of pregnancy [162, 163], the cause of which was
unknown heretofore.
Liver transplantation is the only treatment option for PFIC3
available at this time. Transplantation of normal hepatocytes in
a knockout mouse model of PFIC3 has resulted in spontaneous
massive liver repopulation with the transplanted cells, resulting
in long‐term amelioration of the phospholipid transport defect
[164, 165].
Progressive familial intrahepatic cholestasis type 4
As discussed earlier, PFIC syndromes can be classified clinically
by serum GGT levels that reflect bile salt‐induced injury of bile
canaliculi or smallest bile ductules. In PFIC1 and PFIC2 serum
GGT is normal, whereas in PFIC3 it is elevated. In 2014,
Sambrota et al. [166] reported a multi‐institution collaborative
study of 33 children from 29 families who had severe chronic
cholestatic liver disease, with low GGT for the degree of choles-
tasis. However, these children did not have mutations of either
ABCB11 and ATP8B1, which cause PFIC1 and PFIC2, respec-
tively. In these 29 families, 18 parents were consanguineous.
Genetic analysis by targeted resequencing (TRS), whole‐exome
sequencing (WES), or both revealed protein‐truncating muta-
tions in the tight junction protein 2 gene (TJP2), which resulted
in failure of incorporation of TJP2 protein in desmosomes delim-
iting the bile canaliculi. PFIC resulting from TJP2 mutations has
been designated PFIC4. Two cases of hepatocellular carcinoma
in infants with this disorder have been reported [167].
OTHER INHERITED CHOLESTATIC
DISORDERS
Mutation of CIRH1A, the gene encoding a mitochondrial scaf-
fold protein, has been reported to manifest as cholangiopathy of
North American Indian childhood cirrhosis. Clinically, this
disorder resembles extrahepatic biliary atresia, but there is no
evidence of biliary tract obstruction [168].
GRACILE syndrome
Genetic lesions of another mitochondrial scaffold gene,
BCS1L, presents potentially lethal intrahepatic cholestasis
associated with fetal growth retardation, amino aciduria, iron
overload, and lactic acidosis (GRACILE) [169]. Interestingly,
neither North American Indian childhood cirrhosis nor
GRACILE exhibit the usual features of mitochondriopathy,
such as central nervous system (CNS) lesions or hepatocellu-
lar steatosis.
Alagille syndrome
Alagille syndrome affects the development of multiple organ sys-
tems, including liver, heart, kidneys, eyes, vertebrae (sagittal,
cleft), and CNS [170]. Paucity of bile ducts is found in 89% of
cases [171]. Most patients exhibit a characteristic facies. Although
the CNS lesions can be the result of malnutrition, even after care-
ful nutritional balance, intracranial bleeding is the most common
CNS complication [171]. Alagille syndrome is inherited as an
autosomal dominant characteristic with markedly variable pene-
trance. Lesions of the Jagged1 (JAG1) gene, on chromosome
20p12 [172] are responsible for this disease. Jagged‐1 is a ligand
of Notch, and is important in the Notch signaling pathway, which
is critical in organ development. Coding region mutations of
JAG1 have been identified in about 70% of the patients with
Alagille syndrome. Surprisingly, 50–70% of these mutations are
de novo, and not found in the parents. About 21–50% of patients
with Alagille syndrome who manifest hepatic symptoms in
infancy eventually require liver transplantation.
Villin disease
Villin is a tissue‐specific actin‐modifying protein, critical in
the bundling, nucleation, capping, and severing of actin fila-
ments [173]. In vertebrates, villin is expressed at the apical
surface of epithelial cells and is a major structural component
of the brush border cytoskeleton. Several children with fea-
tures of cholestasis, who later developed cirrhosis of the liver
requiring liver transplantation, were found to have lesions of
the VILLIN1 gene [174].
ACKNOWLEDGMENT
This work was supported in part by NIH grants RO1‐DK‐092469,
RO1‐DK 100490 and P30‐DK‐41296.
 20:  Disorders of Bilirubin Metabolism 241

REFERENCES 25. Nakamura, H., Yonetani, M., Uetani, Y., Funato, M., and Lee, Y.
Determination of serum unbound bilirubin for prediction of kernicterus in
low birthweight infants. Acta Paediatr Jpn, 1992;34: 642–7.
1. Kappas, A. and Drummond, G.S. Direct comparison of Sn‐mesoporphyrin,
26. Wang, P., Kim, R.B., Roy‐Chowdhury, J., and Wolkoff, A.W. The human
an inhibitor of bilirubin production, and phototherapy in controlling hyper-
organic anion transport protein SLc21A6 is not sufficient for bilirubin trans-
bilirubinemia in term and near‐term newborns. Pediatrics, 1995;95:468.
port. J Biol Chem, 2003;278:20695.
2. Valaes, T., Petmezaki, S., Henschke, C., Drummond, G.S., and Kappas, A.
27. Bosma, P.J., Seppen, J., Goldhoorn, B. et al. Bilirubin UDP‐glucuronosyl-
Control of jaundice in preterm newborns by an inhibitor of bilirubin produc-
transferase 1 is the only relevant bilirubin glucuronidating isoform in
tion: studies with tin‐mesoporphyrin. Pediatrics, 1994;93:1.
man. J Biol Chem, 1994;269:17960.
3. Westlake, D.W.S., Roxburgh, J.M., and Talbot, G. Microbial production of
28. Ritter, J.K., Chen, F., Sheen, Y.Y. et  al. A novel complex locus UGT1
carbon monoxide from flavonoids. Nature, 1961;189:510.
encodes human bilirubin, phenol, and other UDP‐glucuronosyltransferase
4. Itoh, K., Chiba, T., Takahashi, S. et al. An Nrf2/small Maf heterodimer medi-
isozymes with identical carboxyl termini. J Biol Chem, 1992;267:3257.
ates the induction of phase II detoxifying enzyme genes through antioxidant
29. Kadakol, A., Ghosh, S.S., Sappal, B.S. et  al. Genetic lesions of bilirubin
response elements. Biochem Biophys Res Commun, 1997;236(2):313–22.
uridinediphosphoglucuronate glucuronosyltransferase causing Crigler‐
5. Naito, Y., Takagi, T., Uchiyama, K., and Yoshikawa, T. Heme oxygenase‐1: a
Najjar and Gilbert’s syndromes: correlation of genotype to phenotype. Hum
novel therapeutic target for gastrointestinal diseases. J Clin Biochem Nutr,
Mutat, 2000;16:297.
2011;48:126.
30. Gantla, S., Bakker, C.T.M., Deocharan, B. et  al. Splice site mutations: a
6. Yachie, A., Niida, Y., Wada, T. et  al. Oxidative stress causes enhanced
novel genetic mechanism of Crigler‐Najjar syndrome type 1. Am J Hum
endothelial cell injury in human heme oxygenase‐1 deficiency. J Clin Invest,
Genet, 1998;62:585.
1999;103:129–35.
31. Bosma, P.J., Roy Chowdhury, J., Bakker, C. et al. A sequence abnormality in
7. Mougiakakos, D., Jitschin, R., Johansson, C.C. et al. The impact of inflam-
the promoter region results in reduced expression of bilirubin‐UDP‐glucu-
matory licensing on heme oxygenase‐1‐mediated induction of regulatory T
ronosyltransferase‐1 in Gilbert syndrome. N Engl J Med, 1995;333:1171.
cells by human mesenchymal stem cells. Blood, 2011;117:4826–35.
32. van de Steeg, E., Stranecky, V., Hartmannova, H. et al. Complete OATP1B1
8. Basuroy, S., Bhattacharya, S., Tcheranova, D. et al. HO‐2 provides endogenous
and OATP1B3 deficiency causes human Rotor syndrome by interrupting
protection against oxidative stress and apoptosis caused by TNF‐α in cerebral
conjugated bilirubin reuptake into the liver. J Clin Invest, 2012;122:519.
vascular endothelial cells. Am J Physiol Cell Physiol, 2006;291:C897–C908.
33. Huang, W., Zhang, J., Chua, S.S. et al. Induction of bilirubin clearance by the
9. Chowdhury, R., Roy Chowdhury, N., and Arias, I.M. Bilirubin conjugation
constitutive androstane receptor. Proc Natl Acad Sci U S A, 2003;100:4156.
in the spiny dogfish Squalus acanthias, the small skate Raja erinacea and the
34. Xie, W., Yeuh, M.F., Radominska‐Pandya, A. et al. Control of steroid, heme,
winter flounder Pseudopleuronectes americanus. Comp Biochem Physiol,
and carcinogen metabolism by nuclear pregnane X receptor and constitutive
1980;668:523.
androstane receptor. Proc Natl Acad Sci U S A, 2003;100:4150.
10. O’Brien, L., Hosick, P.A., John, K., Stec, D.E., and Hinds, T.D., Jr. Biliverdin
35. Roy‐Chowdhury, J., Locker, J., and Roy‐Chowdhury, N. Nuclear receptors
reductase isozymes in metabolism. Trends Endocrinol Metab, 2015;26:212–20.
orchestrate detoxification pathways. Dev Cell, 2003;4:607.
11. Parkar, M., Jeremiah, S.J., Povey, S. et al. Confirmation of the assignment of
36. Maisels, M.J. Managing the jaundiced newborn: a persistent challenge. Can
human biliverdin reductase to chromosome 7. Ann Human Genet, 1984;48:
Med Assoc J, 2015;187:335–43.
57–60.
37. American Academy of Pediatrics Subcommittee on Hyperbilirubinemia.
12. Fu, G., Liu, H., and Doerksen, R.J. Molecular modeling to provide insight
Management of hyperbilirubinemia in the newborn infant 35 or more weeks
into the substrate binding and catalytic mechanism of human biliverdin‐
of gestation. Pediatrics, 2004;114:297–316. Erratum October 01, 2004.
IXalpha reductase. J Phys Chem B, 2012;116:9580–94.
38. Bhutani, V.K., Stark, A.R., Lazzeroni, L.C. et al. Initial clinical testing evalua-
13. Hannam, S., Moriaty, P., O’Reilly, H. et al. Normal neurological outcome in
tion and risk assessment for universal screening for Hyperbilirubinemia
two infants treated with exchange transfusions born to mothers with Crigler‐
Screening Group. Predischarge screening for severe neonatal hyperbilirubine-
Najjar Type 1 disorder. Eur J Pediatr, 2009;168:427–9.
mia identifies infants who need phototherapy. J Pediatr, 2013;162:477–82.
14. Bonnet, R.J., Davis, E., and Hursthouse, M.B. Structure of bilirubin. Nature,
39. Maisels, M.J., Pathak, A., Nelson, N.M. et  al. Endogenous production of
1976;262:326.
carbon monoxide in normal and erythroblastic newborn infants. J Clin
15. Van den Bergh, A.A.H. and Muller, P. Uber eine direkte und eine indirekte
Invest, 1971;50:1.
Diazoreaktion auf Bilirubin. Biochem Z, 1916;77:90.
40. Vest, M., Strebel, L., and Hauensiein, D. The extent of “shunt” bilirubin and
16. Choi, S.H., Yun, K.E., and Choi, H.J. Relationships between serum total
erythrocyte survival in the newborn infant measured by the administration of
bilirubin levels and metabolic syndrome in Korean adults. Nutr Metab
(15N) glycine. Biochem J, 1965;95:11c.
Cardiovasc Dis, 2013;23:31.
41. Clarke, C.A., Donohoe, W.T.A., Finn, R. et al. Prevention of Rh hemolytic
17. Breimer, L.H., Wannamethee, G., Ebrahim, S., and Shaper, A.G. Serum bili-
disease: Final results of the “high risk” clinical trial. A combined study from
rubin and risk of ischemic heart disease in middle‐aged British men. Clin
centers in England and Baltimore. Br Med J, 1971;2:607.
Chem, 1995;41:1504.
42. Levi, A.J., Gatmaitan, Z., and Arias, I.M. Deficiency of hepatic organic
18. Temme, E.H.N., Zhang, J., Schouten, E.G. et al. Serum bilirubin and 10‐year
anion‐binding protein, impaired organic anion uptake by liver and “physio-
mortality risk in a Belgian population. Cancer Causes Control, 2001;12:887.
logic” jaundice in newborn monkeys. N Engl J Med, 1970;283:1136.
19. Zucker, S.D., Horn, P.S., and Sherman, K.E. Serum bilirubin levels in the US
43. Kawade, N. and Onishi, S. The prenatal and postnatal development of
population: gender effect and inverse correlation with colorectal cancer.
UDP‐glucuronosyltransferase activity towards bilirubin and the effect of
Hepatology, 2004;40:827.
premature birth on this activity in human liver. Biochem J, 1981;196:257.
20. Torgerson, J.S., Lindroos, A.K., Sjostrom, C.D. et  al. Are elevated ami-
44. Arthur, L.J., Bevan, B.R., and Holton, J.B. Neonatal hyperbilirubinemia and
notransferases and decreased bilirubin additional characteristics of the meta-
breast feeding. Dev Med Child Neurol, 1966;8:279.
bolic syndrome? Obesity Res, 1997;5:105–14.
45. Arias, I.M., Gartner, L.M., Seifter, S., and Furman, M. Prolonged neonatal
21. Hinds, T.D. Jr. Sodhi, K., Meadows, C. et al. Increased HO‐1 levels amelio-
unconjugated hyperbilirubinemia associated with breast feeding and a ster-
rate fatty liver development through a reduction of heme and recruitment of
oid, pregnane‐3α, 20β‐diol, in maternal milk that inhibits glucuronide forma-
FGF21. Obesity, 2014;22:705–12.
tion in vitro. J Clin Invest, 1964;42:2037.
22. Dong, H., Huang, H., Yun, X. et al. Bilirubin increases insulin sensitivity in
46. Maisels, M.J. and Newman, T.B. Kernicterus in otherwise healthy breast‐fed
leptin‐receptor deficient and diet‐induced obese mice through suppression of
term newborns. Pediatrics, 1995;96:730.
ER stress and chronic inflammation. Endocrinology, 2014;155:818–28.
47. Foliot, A., Ploussard, J.P., Housett, E. et al. Breast milk jaundice: In vitro
23. Amin, S.B., Saluja, S., Saili, A. et al. Chronic auditory toxicity in late pre-
inhibition of rat liver bilirubin‐uridine diphosphate glucuronosyltransferase
term and term infants with significant hyperbilirubinemia. Pediatrics,
activity and Z protein‐bromosulfophthalein binding by human breast milk.
2017;(4):pii:e20164009.
Pediatr Res, 1976;10:594.
24. Ahlfors, C.E., Vreman, H.J., Wong, R.J. et  al. Effects of sample dilution,
48. Lucey, J.F. and Driscol, J.J. Physiological jaundice re‐examined, in
peroxidase concentration, and chloride ion on the measurement of unbound
Kernicterus (ed. A. Sass‐Kortsak), University of Toronto Press, Toronto,
bilirubin in premature newborns. Clin Biochem, 2007;40:261–7.
1961.
242 THE LIVER:  REFERENCES
49. Nies, A.T., Gatmaitan, Z., and Arias, I.M. ATP‐dependent phosphatidylcho-
line translocation in rat liver canalicular plasma membrane vesicles. J Lipid
Res, 1996;37:1125.
50. Poland, R.L. and Odell, G.B. Physiologic jaundice: the enterohepatic circu-
lation of bilirubin. N Engl J Med, 1971;284:1
51. Crigler, J.F. and Najjar, V.A. Congenital familial non‐hemolytic jaundice
with kernicterus. Pediatrics, 1952;10:169.
52. Childs, B., Sidbury, J.B., and Migeon, C.J. Glucuronic acid conjugation by
patients with familial non‐hemolytic jaundice and their relatives. Pediatrics,
1959;23:903.
53. Berk, P.D., Martin, F., Blaschke, T.F. et al. Unconjugated hyperbilirubine-
mia: physiological evaluation and experimental approaches to therapy. Ann
Intern Med, 1975;82:552.
54. Seppen, J., Bosma, P.J., Goldhoorn, B.G. et  al. Discrimination between
Crigler‐Najjar type I and II by expression of mutant bilirubin uridine diphos-
phate‐glucuronosyltransferase. J Clin Invest, 1994;94:2385.
55. Wolkoff, A.W., Roy Chowdhury, J., Gartner, L.A. et al. Crigler‐Najjar syn-
drome (type I) in an adult male. Gastroenterology, 1979;76:3380.
56. Mitchell, E., Ranganathan, S., McKiernan, P. et  al. Hepatic parenchymal
injury in Crigler‐Najjar type I. J Pediatr Gastroenterol Nutr, 2018;66:
588–94.
57. Petit, F.M., Gajdos, V., Parisot, F. et al. Parental isodisomy for chromosome
2 as the cause of Crigler‐Najjar syndrome type I syndrome. Eur J Hum
Genet, 2005;13:278–82.
58. Van Es, H.H.G., Goldhoorn, B., Paul‐Abrahamse, M. et al. Immunochemical
characterization of UDP‐glucuronosyltransferase in four patients with the
Crigler‐Najjar type I syndrome. J Clin Invest, 1990;85:1199.
59. Ritter, J.K., Crawford, J.M., and Owens, I.S. Cloning of two human liver
bilirubin‐UDP‐glucuronosyltransferase cDNAs with expression in COS‐1
cells. J Biol Chem, 1991;266:1043.
60. Bosma, P.J., Roy Chowdhury, N., Goldhoorn, B.G. et al. Sequence of exons
and the flanking regions of human bilirubin‐UDP‐glucuronosyltransferase
gene complex and identification of a genetic mutation in a patient with
Crigler‐Najjar syndrome, type I. Hepatology, 1992;15:941–7.
61. Bosma, P.J., Seppen, J., Goldhoorn, B. et al. Bilirubin UDP‐glucuronosyl-
transferase 1 is the only relevant bilirubin glucuronidating isoform in man.
J Biol Chem, 1994;269:17960.
62. Servedio, V., d’Apolito, M., Maiorano, N. et al. Spectrum of UGT1A1 muta-
tions in Crigler‐Najjar (CN) syndrome patients: identification of twelve
novel alleles and genotype‐phenotype correlation. Hum Mutat, 2005;25:325.
63. Strauss, K.A., Robinson, D.L., Vreman, H.J. et al. Management of hyperbili-
rubinemia and prevention of kernicterus in 20 patients with Crigler‐Najjar
disease. Eur J Pediatr, 2006;165:306–19.
64. Kadakol, A., Deocharan, B., Mukhopadhyay, L. et al. Rapid prenatal diagno-
sis of Crigler‐Najjar syndrome type 1 by genetic analysis of chorionic villus
samples. Hepatology, 1998;28:316A.
65. Gunn, C.H. Hereditary acholuric jaundice in a new mutant strain of rats.
J Hered, 1938;29:137.
66. Roy Chowdhury, N., Kondapalli, R., and Roy Chowdhury, J. The Gunn rat:
An animal model for inherited deficiency of bilirubin glucuronidation, in
Animal Models in Liver Research (eds. C.E. Cornelius, R.R. Marshak, and
E.C. Melby), Academic Press, New York, 1993, p. 150.
67. Roy Chowdhury, N., Kondapalli, R., and Roy Chowdhury, J. The Gunn rat:
an animal model for inherited deficiency of bilirubin glucuronidation, in
Animal Models in Liver Research (eds. C.E. Cornelius, R.R. Marshak, and
E.C. Melby), Academic Press, New York, 1993, pp. 150–75.
68. Polgar, Z., Li, Y., Li Wang, X. et al. Gunn rats as a surrogate model for evalu-
ation of hepatocyte transplantation‐based therapies of Crigler‐Najjar syn-
drome type 1. Methods Mol Biol, 2017;1506:131–47.
69. Chen, Y., Li, Y., Wang, X. et al. Amelioration of hyperbilirubinemia in Gunn
rats after transplantation of human induced pluripotent stem cell‐derived
hepatocytes. Stem Cell Reports, 2015;5:22–30.
70. Nguyen, N., Bonzo, J.A., Chen, S. et al. Disruption of the Ugt1 locus in mice
resembles human Crigler‐Najjar type I disease. J Biol Chem, 2008;
283:7901–11.
71. Mazariegos, G., Shneider, B., Burton, B. et al. Liver transplantation for pedi-
atric metabolic disease. Mol Genet Metab, 2014;111:418–27.
72. Kappas, A., Drummond, G.S., Manola, T., Petmezaki, S., and Valaes T Sn‐
protoporphyrin use in the management of hyperbilirubinemia in term new-
borns with direct Coombs‐positive ABO incompatibility. Pediatrics,
1988;81(4):485–97.
73. Galbraith, R.A., Drummond, G.S., and Kappas A. Suppression of bilirubin
production in the Crigler‐Najjar, type I syndrome: studies with heme oxyge-
nase inhibitor tin‐mesoporphyrin. Pediatrics, 1992;89:175.
74. Fox, I.J., Roy Chowdhury, J., Kaufman, S.S. et  al. Treatment of Crigler‐
Najjar syndrome type I with hepatocyte transplantation. N Engl J Med,
1998;333:1422.
75. Iansante, V., Mitry, R.R., Filippi, C., Fitzpatrick, E., and Dhawan, A. Human
hepatocyte transplantation for liver disease: current status and future per-
spectives. Pediatr Res, 2018;83:232–40.
76. Zhou, H., Dong, X., Kabarriti, R. et al. Single liver lobe repopulation with
wildtype hepatocytes using regional hepatic irradiation cures jaundice in
Gunn rats. PLoS One, 2012;7:e46775.
77. Yannam, G.R., Han, B., Setoyama, K. et al. A nonhuman primate model of
human radiation‐induced venocclusive liver disease and hepatocyte injury.
Int J Radiat Oncol Biol Phys, 2014;88:404–11.
78. Soltys, K.A., Setoyama, K., Tafaleng, E.N. et  al. Host conditioning and
rejection monitoring in hepatocyte transplantation in humans. J Hepatol,
2016;66:987–1000.
79. Arias, I.M. Chronic unconjugated hyperbilirubinemia without overt signs of
hemolysis in adolescents and adults. J Clin Invest, 1962;41:2233.
80. Gollan, J.L., Huang, S.M., Billing, B., and Sherlock, S. Prolonged survival in
three brothers with severe type II Crigler‐Najjar syndrome: ultrastructural
and metabolic studies. Gastroenterology, 1975;68:1543.
81. Arias, I.M., Gartner, L.M., Cohen, M. et al. Chronic nonhemolytic unconju-
gated hyperbilirubinemia with glucuronosyltransferase deficiency: clinical,
biochemical, pharmacologic, and genetic evidence for heterogeneity. Am J
Med 1969;47:395.
82. Fevery, J., Blanckaert, N., Heirwegh, K.P.M. et al. Unconjugated bilirubin
and an increased proportion of bilirubin monoconjugates in the bile of
patients with Gilbert’s syndrome and Crigler‐Najjar syndrome. J Clin Invest,
1977;60:970.
83. Bosma, P.J., Golhoorn, B., Oude Elferink, R.P. et al. A mutation in bilirubin
uridine 5′‐diphosphate glucuronosyltransferase isoforms 1 causing Crigler‐
Najjar syndrome type II. Gastroenterology, 1993;105:216.
84. Gilbert, A. and Lereboullet, P. La cholamae simple familiale. Semin Med,
1901;21:241.
85. Bosma, P.J., Roy Chowdhury, J., Bakker, C. et al. A sequence abnormality in
the promoter region results in reduced expression of bilirubin‐UDP‐glucu-
ronosyltransferase‐1 in Gilbert syndrome. N Engl J Med, 1995;333:1171.
86. Thompson, R.P.H. (1981) Genetic transmission of Gilbert’s syndrome, in
Familial Hyperbilirubinemia (ed. L. Okolicsanyl), Wiley, New York, 1981,
p. 91.
87. Powell, L.W., Hemingway, E., Billing, B.H., and Sherlock, S. Idiopathic
unconjugated hyperbilirubinemia (Gilbert’s syndrome): a study of 42 fami-
lies. N Engl J Med, 1967;277:1108.
88. Arias, I.M. and London, I.M. Bilirubin glucuronide formation in vitro:
Demonstration of a defect in Gilbert’s disease. Science, 1957;126:563.
89. Davidson, A.R., Rojas‐Beuno, A., Thompson, R.P.H., and Williams, R.
Reduced caloric intake and nicotinic acid provocation tests in diagnosis of
Gilbert’s syndrome. Br Med J, 1975;2:480.
90. Bosma, P.J., Roy Chowdhury, J., Bakker, C. et al. A sequence abnormality in
the promoter region results in reduced expression of bilirubin‐UDP‐glucu-
ronosyltransferase‐1 in Gilbert syndrome. N Engl J Med, 1995;333:1171.
91. Maruo, Y., Sato, H., Yamano, T. et al. Gilbert’s syndrome caused by homozy-
gous missense mutation (Tyr486Asp) of bilirubin‐UDP glucuronosyl trans-
ferase. J Pediatr, 1998;132:1045.
92. Soeda, Y., Yamamoto, K., Adachi, Y. et al. Predicted homozygous mis‐sense
mutation in Gilbert’s syndrome. Lancet, 1995;346:1494.
93. Koiwai, O., Nishizawa, M., Hasada, K. et al. Gilbert’s syndrome is caused by
a heterozygous missense mutation in the gene for bilirubin UDP‐glucurono-
syltransferase. Hum Mol Genet, 1995;4:1183.
94. Kadakol, A., Sappal, B.S., Ghosh, S.S. et  al. Interaction of coding region
mutations and the Gilbert‐type promoter abnormality of the UGT1A1 gene
causes moderate degrees of unconjugated hyperbilirubinemia and may lead
to neonatal kernicterus. J Med Genet, 2001;38:244.
95. Labrune, P., Myara, A., Chalas, J. et al. Association of a homozygous (TA)8
promoter polymorphism and a N400D mutation of UGT1A1 in a child with
Crigler‐Najjar type II syndrome. Hum Mutat, 2002;20:399.
96. Wells, P.G., Mackenzie, P.I., Roy‐Chowdhury, J. et al. Glucuronidation and
the UDP‐glucuronosyltransferases in health and disease. Drug Metab
Dispos, 2004;32:281.
 20:  Disorders of Bilirubin Metabolism 243
97. Iyer, L., King, C.D., Whitington, P.F. et al. Genetic predisposition to the
metabolism of irinotecan (CPT‐11): Role of uridine diphosphate glucu-
ronosyltransferase isoform 1A1 in the glucuronidation of its active metabo-
lite (SN‐38) in human liver microsomes. J Clin Invest, 1998;101:847.
98. de Morais, S.M., Uetrecht, J.P., and Wells, P.G. Decreased glucuronidation
and increased bioactivation of acetaminophen in Gilbert’s syndrome.
Gastroenterology, 1992;102:577.
99. Ullrich, D., Sieg, A., Blume, R. et al. Normal pathways for glucuronidation,
sulphation and oxidation of paracetamol in Gilbert’s syndrome. Eur J Clin
Invest, 1987;17:237.
100. Abumiya, M., Takahashi, N., Niioka, T. et al. Influence of UGT1A1 6, 27,
and 28 polymorphisms on nilotinib‐induced hyperbilirubinemia in Japanese
patients with chronic myeloid leukemia. Drug Metab Pharmacokinet,
2014;29:449–54.
101. Portman, O.W., Alexander, M., Roy Chowdhury, J. et al. Effects of nutri-
tion on hyperbilirubinemia in Bolivian squirrel monkeys. Hepatology,
1984;4:454.
102. Portman, O.W., Roy Chowdhury, J., Roy Chowdhury, N. et  al. A non‐
human primate model for Gilbert’s syndrome. Hepatology, 1984;4:175.
103. Dubin, I.N. and Johnson, F.B. (1954) Chronic idiopathic jaundice with uni-
dentified pigment in liver cells: a new clinicopathologic entity with a report
of 12 cases. Medicine (Baltimore) 1954;33:155.
104. Sprinz, H. and Nelson, R.S. Persistent nonhemolytic hyperbilirubinemia
associated with lipochrome‐like pigment in liver cells: report of four cases.
Ann Intern Med, 1954;41:952.
105. Dubin, I.N. Chronic idiopathic jaundice: a review of fifty cases. Am J Med,
1958;23:268.
106. Shani, M., Seligshon, U., Gilon, E. et  al. Dubin‐Johnson syndrome in
Israel: clinical, laboratory, and genetic aspects of 101 cases. West J Med,
1970;39:549.
107. Come, S.E., Shohet, S.B., and Robinson, S.H. Surface remodeling vs.
whole‐cell hemolysis of reticulocytes produced with erythroid stimulation
or iron deficiency anemia. Blood, 1974;44:817.
108. Morita, M. and Kihava, T. Intravenous cholecystography and metabolism
of meglumine iodipamide (biligrafin) in Dubin‐Johnson syndrome.
Radiology, 1971;95:57.
109. Arias, I.M., Bernstein, L., Roffler, R., and Ben Ezzer, J. Black liver diseases
in Corriedale sheep: metabolism of tritiated epinephrine and incorporation
of isotope into the hepatic pigment in vivo. J Clin Invest, 1965;44:1026.
110. Swartz, H.M., Sarna, T., and Varma, R.R. On the nature and excretion of the
hepatic pigment in the Dubin‐Johnson syndrome. Gastroenterology, 1979;76:958.
111. Ishikawa, T., Muller, M., Klunemann, C. et  al. ATP‐dependent primary
active transport of cysteinyl leukotrienes across liver canalicular mem-
branes. J Biol Chem, 1990;265:19279.
112. Jedlitschky, G., Leier, I., Buchholz, U. et al. ATP‐dependent transport of
glutathione S‐conjugates by the multidrug resistance‐associated protein.
Cancer Res, 1994;54:4833.
113. Kobayashi, K., Sogame, Y., Hara, H., and Hayashi, K. Mechanism of glu-
tathione S‐conjugate transport in canalicular and basolateral rat liver
plasma membranes. J Biol Chem, 1990;265:7737.
114. Nishida, T., Hardenbrook, C., Gatmaitan, Z., and Arias, I.M. ATP‐depend-
ent organic anion transport system in normal and TR− rat liver canalicular
membranes. Am J Physiol, 1992;262:G629.
115. Paulusma, C.C., Bosma, P.J., Zaman, G.J.R. et al. Congenital jaundice in
rats with a mutation in a multidrug resistance‐associated protein gene.
Science, 1996;271:1126.
116. Takikawa, H., Sano, N., Narita, T. et al. Biliary excretion of bile acid con-
jugates in a hyperbilirubinemic mutant sprague‐dawley rat. Hepatology,
1991;14:352–60.
117. Togawa, T., Mizuochi, T., Sugiura, T. et al. Clinical, pathologic, and genetic
features of neonatal Dubin‐Johnson syndrome: a multicenter study in
Japan. J Pediatr, 2018;196:161–7.
118. Konig, J., Rost, D., Cui, Y., and Keppler, D. Characterization of the human
multidrug resistance protein isoform MRP3 localized to the basolateral
hepatocyte membrane. Hepatology, 1999;29:1156.
119. Charbonnier, A. and Brisbois, P. Etude chromatographique de la BSP au
cours de l’epreuve clinique d’epuration plasmatique de ce colorant. Rev
Intern Hepatol, 1960;10:1163.
120. Rodes, J., Zubizarreta, A., and Bruguera, M. Metabolism of bromsulphoph-
thalein in Dubin‐Johnson syndrome: Diagnostic value of the paradoxical
increase in plasma levels of BSP. Am J Dig Dis, 1972;17:545.
121. Kaplowitz, N., Javitt, N., and Kappas, A. Coproporphyrin I and III excre-
tion in bile and urine. J Clin Invest, 1972;51:2895.
122. Seligsohn, U., Shani, M., Ramot, B. et al. Dubin‐Johnson syndrome in
Israel. II. Association with factor‐VII deficiency. Q J Med, 1970;39:569.
123. Kamisako, T., Leier, I., Cui, Y. et al. Transport of monoglucuronosyl and
bisglucuronosyl bilirubin by recombinant human and rat multidrug resist-
ance protein 2. Hepatology, 1999;30:485.
124. Kartenbeck, J., Leuschner, U., Mayer, R., and Keppler, D. Absence of the
canalicular isoform of the MRP gene‐encoded conjugate export pump
from the hepatocytes in Dubin‐Johnson syndrome. Hepatology, 1996;
23:1061.
125. Allikmets, R., Gerrard, B., Hutchinson, A., and Dean, M. Characterization
of the human ABC superfamily: isolation and mapping of 21 new genes
using the Expressed Sequence Tags database. Hum Mol Gen, 1996;5:1649.
126. Keitel, V., Karenbeck, J., Nies, A.T. et al. Impaired protein maturation of the
conjugate export pump multidrug resistance protein 2 as a consequence of a
deletion mutation in Dubin‐Johnson syndrome. Hepatology, 2000;32:1317.
127. Cornelius, C.E., Arias, I.M., and Osburn, B.I. Hepatic pigmentation with
photosensitivity: a syndrome in Corriedale sheep resembling Dubin‐
Johnson syndrome in man. J Am Vet Med Assoc, 1965;146:709.
128. Barnhart, J.L., Gronwall, R.R., and Combes, B. Biliary excretion of
­sulfobromophthalein compounds in normal and mutant Corriedale sheep:
evidence for a disproportionate transport defect for conjugated sulfo-
bromophthalein. Hepatology, 1981;1:441.
129. Bohan, A. and Boyer, J.L. Mechanisms of hepatic transport of drugs:
Implications for cholestatic drug reactions. Semin Liver Dis, 2002;22:123.
130. Jansen, P.L.M., van Klinken, J.W., van Gelder, M. et al. Preserved organic
anion transport in mutant TR‐ rats with a hepatobiliary secretion defect.
Am J Physiol, 1993;265:G445.
131. Jansen, P.L.M., Peters, W.H.M., and Lamers, W.H. Hereditary chronic
conjugated hyperbilirubinemia in mutant rats caused by defective hepatic
anion transport. Hepatology, 1985;5:573.
132. Kitamura, T., Alroy, J., Gatmaitan, Z. et al. Defective biliary excretion of
epinephrine metabolites in mutant (TR‐) rats: Relation to the pathogene-
sis of black liver in the Dubin‐Johnson syndrome and Correidale sheep
with an analogous excretory defect. Hepatology, 1992;15:1154.
133. Takikawa, H., Nishikawa, K., Sano, N. et  al. Mechanisms of biliary
excretion of lithocholate‐3‐sulfate in Eisai hyperbilirubinemic rats
(EHBR). Dig Dis Sci, 1995;40:1792.
134. Schulman, F.Y., Montali, R.J., Bush, M. et al. Dubin‐Johnson‐like syn-
drome in Golden Lion tamarins (Leontopithecus rosalia rosalia). Vet
Pathol, 1993;30:491.
135. Rotor, A.B., Manahan, L., and Florentin, A. Familial nonhemolytic jaun-
dice with direct van den Bergh reaction. Acta Med Philos, 1948;5:37.
136. Wolpert, E., Pascasio, F.M., Wolkoff, A.W., and Arias, I.M. Abnormal
sulfobromophthalein metabolism in Rotor’s syndrome and obligate het-
erozygotes. N Engl J Med, 1977;296:1099.
137. Schiff, L., Billing, B.H., and Oikawa Y (1959) Familial nonhemolytic
jaundice with conjugated bilirubin in the serum; a case study. N Engl J
Med, 1959;260(26):1315–18.
137A. Porush, J.G., Delman, A.J., and Feuer, M.M. Chronic idiopathic jaundice
with normal liver histology. Arch Intern Med, 1962;109:102.
138. Wolkoff, A.W., Wolpert, E., Pascasio, F.N., and Arias, I.M. Rotor’s syn-
drome: A distinct inheritable pathophysiologic entity. Am J Med, 1976;60:173
139. Rapacini, G.L., Topi, G.C., Anti, M. et al. Porphyrins in Rotor syndrome:
A study on an Italian family. Hepatogastroenterology, 1986;33:11.
140. van de Steeg, E., Stranecky’, V., Hartmannova, H. et al. Complete OATP1B1
and OATP1B3 deficiency causes human Rotor syndrome by interrupting
conjugated bilirubin reuptake into the liver. J Clin Invest, 2012;122:519–28.
141. Maeda, K. Organic anion transporting polypeptide (OATP)1B1 and
OATP1B3 as important regulators of the pharmacokinetics of substrate
drugs. Biol Pharm Bull, 2015;38:155–68.
142. Link, E., Parish, S., Armitage, J. et al. SEARCH Collaborative Group.
SLC01B1 variants and statin‐induced myopathy: a genomewide study. N
Engl J Med, 2008;359:789–99.
143. Cornelius, C.E. and Gronwall, R.R. Congenital photosensitivity and
hyperbilirubinemia in Southdown sheep in the United States. Am J Vet
Res, 1968;29:291–5.
144. Posbergh, C.J.,.Kalla, S.E., Sutter, N.B., Tennant, B.C., and Huson, H.J.
Mutation responsible for congenital photosensitivity and hyperbilirubine-
mia in Southdown sheep. Am J Vet Res, 2018;79:538–45.
244 THE LIVER:  REFERENCES
145. Jansen, P.L. and Muller, M.M. Progressive familial intrahepatic cholestasis
types 1, 2, and 3. Gut, 1998;42:766.
146. Sticova, E., Jirsa, M., and Pawlowska, J. New insights in genetic cholestasis:
from molecular mechanisms to clinical implications. Can J Gastroenterol
Hepatol, 2018, Article ID 2313675. https://doi.org/10.1155/2018/2313675
147. Clayton, R.J., Iber, F.L., Ruebner, B.H. et al. Byler disease: Fatal familial
intrahepatic cholestasis in an Amish kindred. Am J Dis Child, 1969;117:112.
148. Bull, L.N., Van Eijk, M.J., Pawlikowaska, L. et al. A gene encoding a P‐
type ATPase mutated in two forms of hereditary cholestasis. Nat Genet,
1998;18:219.
149. Klomp, L.W., Vargas, J.C., van Mil, S.W. et al. Characterization of muta-
tions in ATP8B1 associated with hereditary cholestasis. Hepatology,
2004;40:27.
150. Chen, F., Ananthanarayanan, M., Emre, S. et al. Progressive familial intra-
hepatic cholestasis, type 1, is associated with decreased farnesoid X receptor
activity. Gastroenterology, 2004;126:756.
151. Summerskill, W.H.J. and Walshe, J.M. Benign recurrent intrahepatic
obstructive jaundice. Lancet 1959;2:686.
152. Tygstrup, N. and Jensen, B. Intermittent intrahepatic cholestasis of
unknown etiology in five young males from the Faroe Islands. Acta Med
Scand, 1969;185:523.
153. Summerfield, J.A., Scott, J., Berman, M. et al. Benign recurrent intrahepatic
cholestasis: Studies of bilirubin kinetics, bile acids, and cholangiography.
Gut, 1980;21:154.
154. Summerskill, W.H.J. The syndrome of benign recurrent cholestasis. Am J
Med, 1965;38:298.
155. Biempica, L., Gutstein, S., and Arias, I.M. Morphological and biochemi-
cal studies of benign recurrent cholestasis. Gastroenterology, 1967;
52:521.
156. Thompson, R. and Strautnieks, S. BSEP: Function and role in progressive
familial intrahepatic cholestasis. Semin Liver Dis, 2001;21:545.
157. Strautnieks, S.S., Kagalwalla, A.F., Tanner, M.S. et al. Identification of a
locus for progressive familial intrahepatic cholestasis (PFIC2) on chromo-
some 2q24. Am J Hum Genet, 1997;61:630.
158. Stindt, J., Kluge, S., Dröge, C. et al. Bile salt export pump‐reactive antibod-
ies form a polyclonal, multi‐inhibitory response in antibody‐induced bile
salt export pump deficiency. Hepatology, 2016;63:524–37.
159. Nies, A.T., Gatmaitan, Z., and Arias, I.M. ATP‐dependent phosphatidylcho-
line translocation in rat liver canalicular plasma membrane vesicles. J Lipid
Res, 1996;37:1125.
160. Strautnieks, S.S., Knisley, A.S., Gerred, S. et al. Mutations in MDR3 in adult
onset cholangiopathy, in Bile Acids: From Genomics to Disease and Therapy.
Proceedings of Falk Symposium 129 (eds. D. Keppler, U. Leuschner, G.
Paumgartner, and A. Stiehl), Kluwer, Dordrecht, 2003, p. 184.
161. Rosmorduc, O., Hermelin, B., and Poupon, R. MDR3 gene defect in adults
with symptomatic intrahepatic and gallbladder cholesterol cholelithiasis.
Gastroenterology, 2001;120:1459.
162. Dixon, P.H., Weerasekera, N., Linton, K.J. et al. Heterozygous MDR3 mis-
sense mutation associated with intrahepatic cholestasis of pregnancy: evi-
dence for a defect in protein trafficking. Hum Mol Genet, 2000;9:1209.
163. Jacquemin, E., Cresteil, D., Manouvrier, S. et al. Heterozygous non‐sense
mutation associated with intrahepatic cholestasis of pregnancy. Lancet,
1999;353:210.
164. de Vree, J., Jacquemin, E., Strum, E. et al. Mutations in the MDR3 gene
cause progressive familial intrahepatic cholestasis. Proc Acad Natl Sci U S
A, 1998;25:282.
165. de Vree, J.M.L., Ottenhoff Smith, A.J., Aten, J. et al. Rapid correction of
Mdr2 deficiency by transplantation of Mdr3 transgenic hepatocytes.
Hepatology, 1998;28:387A.
166. Sambrotta, M., Strautnieks, S., Papouli, E. et al. Mutations in TJP2 cause
progressive cholestatic liver disease. Nat Genet, 2014;46:326–8.
167. Zhou, S., Hertel, P.M., Finegold, M.J. et al. Hepatocellular carcinoma associ-
ated with tight‐junction protein 2 deficiency. Hepatology, 2015;62:1914–16.
168. Drouin, E., Russo, P., Tuchweber, B. et al. North American Indian cirrhosis
in children: a review of 30 cases. J Pediatr Gastroenterol Nutr, 2000;31:395.
169. Rapola, J., Heikkila, P., and Fellman, V. (2002) Pathology of lethal fetal
growth retardation syndrome with aminoaciduria, iron overload and lactric
acidosis (GRACILE). Pediatr Pathol Mol Med, 2002;21:183.
170. Emerick, E.M., Rand, E.B., Goldmuntz, E. et al. Features of Alagille syn-
drome in 92 patients: frequency and relation to prognosis. Hepatology,
1999;29:822.
171. Piccoli, D.A. and Spinner, N.A. Alagille syndrome and the Jagged1 gene.
Semin Liver Dis, 2001;21:525.
172. Oda, T., Elkaholoun, A.G., Meltzer, P.S., and Chandrashekharappa, S.C.
Identification and cloning of the human homolog (Jag1) of the rat jagged1
gene from the Alagille syndrome critical region at 20p12. Genomics,
1997;43:376.
173. Friederich, E., Vancompernolle, K., Louvard, D., and Vandekerckhove, J.
Villin function in the organization of the actin cytoskeleton. Correlation of
in vivo effects to its biochemical activities in vitro. J Biol Chem,
1999;274:26751–60.
174. Phillips, M.J., Azuma, T., Meredith, S.L. et al. Abnormalities in villin gene
expression and canalicular microvillus structure in progressive cholestatic
liver disease if childhood. Lancet, 2003;362:1090.
 Hepatic Lipid Droplets
21 in Liver Function
and Disease
Douglas G. Mashek1,2, Wenqi Cui1, Linshan Shang1, and Charles P. Najt1
1
Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota,
Minneapolis, MN, USA
2
Department of Medicine, Division of Diabetes, Endocrinology and Metabolism, University of Minnesota,
Minneapolis, MN, USA
INTRODUCTION
Lipid droplet (LD) accumulation is the defining characteristic of
non‐alcoholic and alcoholic fatty liver disease. Historically con­
sidered to be inert and simply a sign of disease, LDs are increas­
ingly recognized as etiological factors in numerous liver diseases
as well as having important non‐pathological roles in cell signal­
ing and function. These dynamic properties of LDs are highly
regulated by the hundreds of proteins that coat the LD surface
and control lipid trafficking and flux. In this chapter, we will
highlight the major pathways of lipid metabolism that influence
LD accumulation, explore key proteins that regulate LD turnover
and that link LDs to cellular dysfunction and liver disease.
LIPID UPTAKE AND SYNTHESIS
Fatty acids (FAs) serve as the building blocks for the biosynthesis
of many complex lipids including triacylglycerols (TAGs),
phospholipids, and cholesterol esters. The liver derives these
FAs from three main sources  –  direct uptake from the blood,
uptake and degradation of lipoprotein remnants, and de novo
lipogenesis (DNL). Under fasting conditions, nearly all of the
FAs entering the liver are derived from direct uptake as a result
of enhanced adipose tissue lipolysis [1]. In the fed state, there is
a decreased reliance from adipose‐derived FAs that coincides
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
with an increase in the contribution of DNL and remnant uptake,
both of which are highly influenced by diet composition.
Interestingly, the contribution of FAs from DNL is markedly
increased in subjects with non‐alcoholic fatty liver disease
(NAFLD) and can account for up to nearly 30% of hepatic FAs
[2]. The increased flux through the DNL pathway likely contrib­
utes to NAFLD development and its comorbidities. Of special
note is the role of dietary sucrose or its monosaccharide con­
stituent fructose in DNL flux. Numerous studies have linked
sucrose and/or fructose to NAFLD etiology, which results from
high rate of flux of fructose carbons to the liver and its unique
metabolism relative to glucose (see [3] for review).
Intracellular FAs must first be activated to their respective
acyl‐CoAs prior to entrance in downstream metabolic pathways.
This occurs in an ATP‐dependent reaction catalyzed by a family
of long‐chain acyl‐CoA synthetases (ACSL). In addition, the
family of fatty acid transport proteins (FATP) also possess acyl‐
CoA synthetase activity and contribute to FA uptake. Importantly,
both the ACSL and FATP families consist of multiple isoforms
with different intracellular localizations and substrate specificity
[4]. As examples, several isoforms including ACSL3 and 5 and
FATP2 and 5 promote TAG synthesis in the liver [5–8], highlighting
important roles in trafficking FAs to anabolic pathways. ACSL1
activates and channels FAs to the mitochondria for oxidation in
heart and adipose tissue [9, 10], but has minimal effects on FA
oxidation in the liver [11]. The isoform responsible for this initial
step in hepatic FA β‐oxidation has yet to be elucidated.
246 THE LIVER:  LIPOLYSIS
TRIACYLGLYCEROL BIOSYNTHESIS
AND LIPID DROPLET BIOGENESIS
The primary route for disposal of FAs is their esterification to
TAG and its subsequent storage in LDs. The coordination of
numerous acyltransferase enzymes and the phosphatase enzyme
lipin are required for the synthesis of TAG [12]. Each of the
enzymes in this pathway has numerous isoforms with different
substrate specificity, which influence TAG composition.
Examples are the initial (glycerol‐3‐phosphate acyltransferase,
GPAT) and terminal (diacylglycerol acyltransferase, DGAT)
enzymes in the classical Kennedy pathway of TAG synthesis.
GPAT1 shows substrate specificity for saturated FAs, which
accounts for the high percentage of saturated FAs in the sn‐1
position of TAG [13]. Additionally, DGAT1 channels exoge­
nous FAs into TAG, whereas DGAT2 is more selective to FAs
derived from DNL [14–16]. Thus, the relative activity levels of
these various isoforms dictate the acyl composition of TAG. It
should also be noted that more recent work has identified mona­
cylglycerol acyltransferases (MGAT), historically thought to
exist solely in the intestine, to be present in the liver and have
increased expression in models of hepatic steatosis [17].
Although the limited studies on the role of these enzymes are
not conclusive [18, 19], the contribution of the monacylglycerol
synthetic pathway to hepatic TAG synthesis will likely be a
focus of future studies.
LDs are thought to develop in the endoplasmic reticulum
(ER), where neutral lipids accumulate within the leaflets of the
membrane bilayer. As TAG begins to accumulate, it forms a
neutral lipid core surrounded by a phospholipid monolayer
embedded with proteins. This core and associated phospholipid
monolayer eventually bud from the ER, thus forming a cytosolic
LD (see [20, 21] for excellent reviews on LD biogenesis). The
nascent LDs formed at the ER are small in diameter and are
coated with DGAT1 [22]. For LDs to increase in size, they can
either undergo fusion with other LDs or synthesize TAG at the
surface of LDs. Cell death‐inducing DFF45‐like effector C
(CIDEC) is a key protein involved in the fusion of LDs [23, 24].
In response to FAs, CIDEC translocates from the ER to LDs
where it initiates fusion and the formation of large LDs [25].
The LD monolayer membrane contains numerous enzymes
involved in TAG synthesis. For example, several ACSLs as well
as GPAT4 and DGAT2 are present on the LD surface to promote
TAG synthesis and expansion of preexisting LDs [22].
As the LD grows with TAG, the phospholipid monolayer
must also expand. As such, the synthesis of phosphatidylcholine
(PC), the most abundant LD phospholipid, is a critical compo­
nent regulating LD growth and dynamics [26]. Deposition of
TAG in the LD core increases the size of the droplet, thereby
diluting the relative amount of PC in the phospholipid mon­
olayer. As diacylglycerol (DAG) and phosphatidic acid become
enriched at the LD surface, the rate‐limiting enzyme in PC syn­
thesis, CTP:phosphor‐choline cytidylyltransferase α (CCTα),
translocates from the nucleus and binds to the LD monolayer.
Once there, CCTα becomes activated to facilitate CDP‐choline
synthesis. Interestingly, other enzymes critical for PC synthesis
are not localized on LDs, suggesting a unique role of CCTα in
controlling LD PC metabolism [26].
LIPOLYSIS
Catabolism of TAG stored within LDs is a major determinant of
LD size and number. Numerous proteins directly or indirectly
influence this process and, as a result, impact the development
or resolution of hepatic steatosis. Although named after the tis­
sue where it has its highest expression, adipose triglyceride
lipase (ATGL) is also involved in hepatic TAG turnover. First
discovered in 2004, ATGL is now widely recognized as the pri­
mary cytosolic TAG hydrolase in numerous tissues [27].
Ablation of ATGL in the liver results in TAG accumulation and
reduced FA oxidation, while ATGL overexpression alleviates
hepatic steatosis; ATGL does not influence very low density
lipoprotein (VLDL) secretion [28–30]. The increased FA oxida­
tion is driven via enhanced PGC‐1α/PPAR‐α signaling. We have
shown that ATGL drives this transcriptional network through
the protein deacetylase sirtuin 1 (SIRT1), which is required for
ATGL‐mediated induction of oxidative metabolism [31]. Thus,
ATGL appears to be an important signaling node that links LD
catabolism to cell signaling as a means to coordinate down­
stream transcriptional networks that control metabolism. The
next enzyme in the classical lipolysis pathway is hormone‐­
sensitive lipase (HSL), however, the contribution of hepatocyte
HSL to liver TAG catabolism is not known; HSL has been
shown to catalyze cholesterol ester hydrolysis in hepatocytes
[32]. Monoacylglycerol lipase (MAGL), which catalyzes the
last step in hepatic TAG breakdown, has not been studied exclu­
sively in the liver. However, studies in mice lacking MAGL or
utilizing MAGL inhibitors reveal that MAGL plays a key role in
hepatic injury via decreasing endocannabinoids and, thereby,
promoting inflammation [33].
Given the importance of ATGL in catalyzing the initial step
in cytosolic TAG hydrolysis, perhaps it is not surprising that
ATGL activity is highly regulated. A host of proteins that
directly interact with ATGL to influence its activity have been
identified. CGI‐58 is widely accepted as the primary coactiva­
tor of ATGL [34]. Although ATGL activity is an important
mechanism through which CGI‐58 elicits its effects, it clearly
has functions beyond ATGL. For example, CGI‐58 ablation
results in an age‐dependent and robust (8‐ to 52‐fold) increase
in liver TAG, whereas ATGL ablation only increases hepatic
TAG by ~3‐fold [29, 35]. Moreover, CGI‐58 ablation regu­
lates hepatic TAG hydrolysis independent of ATGL, suggest­
ing, at least in the liver, a more complicated role of these
proteins in LD turnover [36].
The perilipin (PLIN) family were the first LD proteins to be
characterized. They consist of five family members, with vary­
ing structural homology, that act to antagonize ATGL‐catalyzed
lipolysis among other functions [37]. PLIN2 is the isoform with
the highest expression in the liver and is induced in response to
FA loading in cells and NAFLD in both humans and rodents
[38]. Liver‐specific ablation of PLIN2 prevents hepatic steatosis
and inflammation [39, 40]. Similar to PLIN2, ablation of PLIN3
in the liver also reduces steatosis, but does not influence inflam­
matory markers [41]. PLIN5 is highly expressed in oxidative
tissues including the liver and its expression is robustly upregu­
lated in response to fasting [42]. Liver‐specific PLIN5 ablation
reduces steatosis, whereas its overexpression promotes steatosis,
which is likely due to its anti‐lipolytic effects [43, 44].
 21:  Hepatic Lipid Droplets in Liver Function and Disease 247
In addition to the perilipin proteins discussed above, a host of
other proteins have been shown to directly interact with and
antagonize ATGL. G0/G1 switch gene 2 (G0S2) is a potent
inhibitor of ATGL that influences hepatic energy metabolism.
G0S2 overexpression in the liver promotes steatosis and reduces
FA oxidation, while ablation of hepatic G0S2 does the opposite
[45–47]. Other interacting proteins including pigment epithelial
derived factor [48], hypoxia‐inducible protein 2 [49], and
CIDEC [50] inhibit ATGL activity directly while UBXD8 tar­
gets it for degradation [51].
Very low density lipoprotein secretion
Approximately 70% of hepatic FAs that are destined for VLDL–
TAG first transit through the cytosolic LD pool before lipopro­
tein assembly [52, 53]. Members of the carboxylesterase (Ces)
family of enzymes contribute to turnover of cytosolic LDs and
their trafficking to LDs destined for VLDL secretion. In particu­
lar, ablation of hepatic Ces1d (also known as triglyceride hydro­
lase, TGH), which colocalizes to ER luminal LDs, reduces
VLDL secretion and increases the size, but decreases the num­
ber, of cytosolic LDs [54, 55]. It appears that Ces1d reduces the
transfer of lipids to existing cytosolic LDs, which may be due to
transfer of ER lipids to VLDL for packaging. While other CES
proteins have been studied, their role specifically in liver and
VLDL secretion has not been characterized. In addition to the
carboxylesterases, the LD protein CIDEB plays an essential role
in VLDL lipidation. Specifically, ablation of CIDEB reduces
VLDL–TAG secretion via an apolipoprotein B (ApoB)‐
dependent mechanism [56]. In contrast, ancient ubiquitous protein
1, which partially colocalizes to LDs, antagonizes ApoB‐medi­
ated VLDL lipidation as a means to suppress VLDL production
[57]. These studies also highlight the importance of ApoB as a
nexus between LDs and VLDL lipidation. Indeed, ApoB colocal­
izes to LDs on extensions of the ER, which is likely critical in
cytosolic transfer of lipids to the developing VLDL particle [58].
LIPOPHAGY
Autophagy is a highly conserved and well‐characterized mecha­
nism through which cellular organelles, protein aggregates, and
macronutrients are degraded during times of nutrient insuffi­
ciency. Although an extensive body of literature characterizing
autophagy of numerous organelles exists, autophagic degrada­
tion of LDs, termed lipophagy, was relatively more recently
identified [59] (see [60] for an extensive review). In this pro­
cess, multiple autophagic arms contribute to LD degradation.
Macrolipophagy is the classical process through which
autophagosomes bud off part of the LD prior to fusing with
lysosomes to form autolysosomes, which can then degrade
the lipid cargo. Microlipophagy describes the direct interaction
and transfer of lipids between LDs and lysosomes. Finally,
chaperone‐mediated autophagy is characterized by degradation
of select proteins, which also may indirectly influence lipophagy.
For example, degradation of specific LD proteins such as
PLIN2 via chaperone‐mediated autophagy allows for the subse­
quent degradation of LDs [61]. Although macroautophagy and
microautophagy are precisely regulated and considered as
pro‐survival mechanisms in response to oxidative stress and
nutrient limitation, chaperone‐mediated autophagy serves a
housekeeping role in the clearance and recycling of misfolded
protein aggregates to maintain cellular integrity. Through the
process of macro‐ and micro‐lipophagy autophagosomes and
lysosomes, respectively, target LDs for degradation. Once inter­
nalized, the lipids are hydrolyzed by lysosomal acid lipase, the
only known lysosomal lipase with activity towards neutral lipids
such as TAG and cholesterol ester. The FAs produced from the
hydrolysis of TAG and cholesterol ester are then available for
β‐oxidation or other downstream pathways.
Since its initial discovery, a rapidly growing body of litera­
ture has characterized autophagy/lipophagy in the liver. To date,
most studies have evaluated lipophagy in the broader context of
global autophagy induction or inhibition. Lipophagy is broadly
regulated similar to autophagy; nutrients, insulin, and mTOR
antagonize autophagy/lipophagy, whereas depletion of nutrients
or increase in activity of low energy sensors (SIRT1, AMPK,
etc.) promotes autophagy/lipophagy. At the transcriptional
level, autophagy is regulated by a host of transcription factors
coinciding with their known roles in nutrient/energy sensing
[60]. Consistent with nutrient regulation, many small molecules
(caffeine, ginsenoside, and quercetin to name a few) reduce
hepatic steatosis through alterations in lipophagy [62–64].
Figure 21.1 provides an overview of the above major pathways
involving LDs and key protein mediators of each pathway.
Lipolysis of TAG via cytosolic lipases and lipophagy are the
two pathways thought to contribute to hepatic LD degradation.
We recently explored the interrelationship between the two
pathways. These studies identified ATGL to act via SIRT1 sign­
aling as an upstream driver of autophagy/lipophagy [65].
Moreover, autophagy/lipophagy is required for ATGL to drive
LD degradation. Consistent with these studies, ablation of
PLIN2 also promotes LD degradation in a lipophagy‐dependent
manner [66]. Thus, these studies point towards a dominant role
of lipophagy in hepatic LD degradation. As our understanding
of the lipophagic process and its significance to hepatic LD
turnover evolves, this pathway will be undoubtedly targeted
through small molecules, as mentioned above, or transcript/pro­
tein targeting as a means to alleviate hepatic steatosis.
LIPID DROPLETS AND LIVER DISEASE
Non‐alcoholic fatty liver disease
Lipid storage in LDs is a hallmark of NAFLD. Alterations in the
balance of LD anabolism and catabolism are integral to the
development and resolution of NAFLD. Thus, dozens of studies
show that reducing substrates for TAG synthesis, inhibiting
enzymes involved in TAG synthesis, or increasing TAG hydrol­
ysis can prevent or alleviate hepatic steatosis. Indeed, ongoing
clinical trials are testing these pathways, especially those
involved in TAG synthesis.
PNPLA3 is an LD protein that belongs to the same family of
enzymes as ATGL (i.e. PNPLA2). A nucleotide substitution in
PNPLA3 (rs738409, I148M) is widely recognized as the single
248 THE LIVER:  LIPID DROPLETS AND LIVER DISEASE
Figure 21.1  Key lipid droplet (LD) protein mediators of specific lipid metabolic pathways influencing LDs. Proteins that antagonize specific
pathways are highlighted in red.
largest genetic predictor of NAFLD [67]. The prevalence of this
mutation is ~15–25% in most populations but increases to ~50%
in Hispanics [67]. PNPLA3 is highly upregulated in response to
high carbohydrate diets and the I148M single‐nucleotide poly­
morphism (SNP) is predictive of steatosis, especially when
­carriers consume a high carbohydrate/sugar diet [68]. Similarly,
the PNPLA3 variant is robustly increased on LDs in mice fed
high sucrose diets [69]. The mechanism(s) through which
PNPLA3 conveys its phenotype is currently under intense
debate. Although ample studies show that the I148M variant
attenuates LD catabolism [70], other studies suggest a lipogenic
role of I148M through its lysophosphatidic acid acyltransferase
activity [71]. Regardless of the mechanism of action, the pres­
ence of the I148M variant also progresses NAFLD to liver dam­
age and injury in numerous models of liver disease [72] as
discussed in more detail later.
In addition to PNPLA3, several other LD proteins are altered
in NAFLD and appear to be involved in disease etiology. The
first discovered and perhaps most studied perilipin isoform is
PLIN1. In adipose tissue, PLIN1 binds CGI‐58 under basal con­
ditions, but phosphorylation in response to lipolytic stimuli dis­
rupts this interaction allowing CGI‐58 to partner with ATGL to
promote lipolysis [73]. Although PLIN1 is normally expressed
at very low or undetectable levels in the liver, it is upregulated in
livers of humans with NAFLD [38, 74]. It remains unknown if
this increase is merely a consequence of increased LDs or if
PLIN1 plays an etiological role in NAFLD. PLIN2 and PLIN3
are also increased in human NAFLD [38]. Consistent with their
anti‐lipolytic roles, liver‐specific PLIN2 or PLIN3 ablation pre­
vents hepatic steatosis [39–41].
The surface of the LD is coated with hundreds of proteins
whose presence and/or abundance are dynamic. Although LD
proteomics has been conducted in numerous cell types, several
studies have characterized the hepatic LD proteome and how
it changes in response to fasting/refeeding [75] or NAFLD
[76–78]. From these studies, we can glean that LDs interact
with numerous cellular organelles to coordinate metabolism as
evidenced by the abundant presence of numerous organelle‐spe­
cific proteins in the LD proteome. Moreover, these studies have
expedited the discovery of novel proteins with pathological
roles in liver disease. One such example is 17β‐hydroxysteroid
dehydrogenase‐13 (17β‐HSD13), which was identified and
characterized to be increased in a LD proteomics screen of liver
biopsies from NAFLD patients relative to non‐steatotic controls
[77]. Overexpression of 17β‐HSD13 in the liver of mice
increases steatosis confirming a direct effect on etiology of the
disease [77]. Although the underlying mechanism of action is
still under investigation, 17β‐HSD13 overexpression in hepato­
cytes increases lipogenesis, suggesting the increased FA supply
and subsequent TAG synthesis may be a driving factor underly­
ing its ability to promote steatosis [77]. Since 17β‐HSD13 is
thought to be involved in estrogen and androgen metabolism, it
has been proposed that it may act to alter local pools of hor­
mones, leading to downstream changes in lipogenesis [79].
Genome‐wide association studies also identified a polymor­
phism (rs641738 C>T) that contains the membrane‐bound
O‐acyltransferase domain‐containing 7 gene (MBOAT7, or
LPIAT1) and transmembrane channel‐like 4 gene (TMC4) to be
associated with increased hepatic fat content mediated by
changes in the phosphatidylinositol acyl‐chain remodeling [80,
81]. MBOAT7 is localized on intracellular membranes such as
ER, mitochondria, and LDs and functions as a lysophosphoino­
sitol acyltransferase [80]. Another prominent NAFLD‐associated
polymorphism (rs58542926, E167K) was identified in trans­
membrane 6 superfamily member 2 (TM6SF2) [82]. Despite
being localized in the ER and the ER–Golgi intermediate com­
partment, but not on LDs, TM6SF2 regulates cholesterol metab­
olism and incorporation of polyunsaturated FAs into hepatic
TAG [83, 84]. TM6SF2 deficiency causes LD accumulation
due, at least in part, to reduced TAG secretion [84].
 21:  Hepatic Lipid Droplets in Liver Function and Disease 249
Alcoholic fatty liver disease
Relative to NAFLD, our understanding of LDs in alcoholic fatty
liver disease (AFLD) is less developed. The I148M polymor­
phism of PNPLA3 is positively associated with AFLD in
genome‐wide association studies [85, 86] consistent with its
broad role in promoting liver disease. In addition, TM6SF2
(rs58542926) and MBOAT7 (rs641738) genes are also identi­
fied as risk loci for alcohol‐related cirrhosis [87], suggesting the
importance of lipid metabolism in the pathogenesis of AFLD.
The expression level of PLIN2 increases to coincide with expan­
sion of the LD pool following ethanol exposure [88], suggesting
reduced lipolysis may be one factor contributing to increased
LDs. Lipophagy and expression of RAG GTPase Rab7, which is
required for lipophagy, are greatly inhibited by ethanol, further
supporting reduced LD catabolism as a contributing factor to
AFLD [83]. Moreover, ethanol exposure inhibits β‐adrenergic
induced phosphorylation of HSL and the recruitment of ATGL
to the LD surface, leading to decreased lipolysis [89]. More
recently, a novel role of ceramide synthase 6 (CerS6) has been
uncovered. Previous studies have shown that increased cera­
mide synthesis and accumulation are linked to AFLD [90]. The
expression of CerS6, which is involved in ceramide synthesis, is
increased in response to alcohol [91], and inhibition of ceramide
synthesis attenuates alcohol‐mediated steatosis and liver dys­
function [92, 93]. Moreover, CerS6 drives PLIN2 expression
and knockdown of PLIN2 attenuates ceramide production fol­
lowing ethanol exposure [91]. Finally, CerS6 localizes to LDs
and interacts with ACSL5, suggesting that ceramides are pro­
duced at the LD surface [91, 94]. Although excess efflux of
ethanol carbons to DNL has long been thought to be the driving
force in AFLD, these more recent yet limited studies suggest a
more complex etiology.
Non‐alcoholic steatohepatitis
The transition from simple steatosis to non‐alcoholic steato­
hepatitis (NASH) is a pivotal point in disease pathology.
Inflammation and reactive oxygen species are key components
in advancing progression to NASH. Since LDs directly interact
with mitochondria and can generate inflammatory eicosanoids
at the LD surface [95], perhaps it is not surprising that numerous
LD proteins are implicated in NAFLD progression. The I148M
mutation of PNPLA3, in addition to promoting steatosis, is also
a major driver of progression from steatosis to NASH in both
adult and pediatric populations [96, 97]. Ablation of CGI‐58
causes NASH in mice [35], while ATGL ablation has no effect
on liver fibrosis [29], suggesting lipolysis‐independent effects
of CGI‐58 on NAFLD progression. Liver‐specific knockout of
PLIN2 alleviates NASH pathologies in a methionine choline‐
deficient (MCD) model of NASH [40]. Although its role in
NASH etiology is not known, PLIN1 is highly expressed in
adults and children with NASH [98]. In contrast to proteins
involved in TAG breakdown, the DGAT enzymes appear to have
opposing effects on NASH. Treatment of mice with antisense
oligonucleotides (ASOs) targeting DGAT1 attenuates fibrosis,
induced by the MCD diet, and reduces stellate cell activation
without influencing hepatic steatosis [99]. In contrast, DGAT2
ablation in the same model reduces steatosis, but increases
Figure 21.2  Lipid droplet proteins that link progression of steatosis to
non‐alcoholic steatohepatitis (NASH), hepatitis C virus (HCV), hepato­
cellular carcinoma (HCC), and insulin/glucose homeostasis.
hepatic inflammation, lipid peroxidation, and fibrosis [100].
Thus, this terminal step in the TAG synthetic pathway appears
to play an important role in coupling/uncoupling steatosis from
the liver dysfunction that precedes NASH.
Serine/threonine protein kinase 25 (STK25) is one of the few
kinases that have been characterized to reside on LDs in hepato­
cytes [101]. Overexpression of STK25 promotes steatosis, liver
inflammation, and fibrosis in mice [84]. Moreover, STK25 pro­
tein abundance correlates with steatosis in humans [101].
Although the proteins downstream of STK25 remain to be elu­
cidated, these initial studies suggest that STK25 could play a
major role in hepatic LD biology and disease progression
through its phosphorylation of LD proteins.
Another important aspect of LDs during NASH development
is their lipid composition. Specifically, Ioannou et al. observed
the presence of cholesterol crystals in LDs subjects with NASH
but not simple steatosis [102]. Increasing dietary cholesterol
intake in rodents also increases LD cholesterol crystals and pro­
motes inflammation and fibrosis [103, 104]. These studies are
consistent with human studies that show increased dietary cho­
lesterol to be an independent risk factor for NASH and cirrhosis
[105]. Activated Kupffer cells surround dead hepatocytes laden
with cholesterol crystals, suggesting that cell death and signal­
ing to immune cells may be directly involved in explaining how
cholesterol crystals promote inflammation [102]. Additionally,
PC levels are reduced in subjects with NAFLD and NASH,
which suggests that alterations in the LD monolayer could also
influence disease progression [106]. Figure 21.2 highlights the
key LD proteins involved in the development of NASH and
other complications downstream of steatosis.
Insulin resistance and glucose dysregulation
Ectopic LD accumulation is correlated with insulin resistance
in numerous tissues including the liver. The consequences of
hepatic insulin resistance include excess hepatic glucose pro­
duction and VLDL secretion, which contribute to the hyper­
glycemia and hypertriglyceridemia, respectively, commonly
present in prediabetes and type 2 diabetes. Although numerous
mechanisms underlying the development of insulin resistance
exist, alterations in lipid metabolism are a common theme.
250 THE LIVER:  LIPID DROPLETS AND LIVER DISEASE
Accumulation of intermediates in lipid metabolic pathways are
thought to be the key drivers of hepatic insulin resistance rather
than TAG itself. Intermediates in the lipid biosynthesis pathway
including phosphatidic acid, DAG, and ceramides may directly
antagonize insulin signaling. Reactive oxygen species may
result from the oversupply of FAs to mitochondria and/or mito­
chondrial dysfunction, leading to insulin resistance. Finally,
production of inflammatory signaling molecules within hepato­
cytes or hepatic immune cells may also interfere with normal
hepatocyte function and insulin signaling.
LD dynamics appear to play a major role in the above pro­
cesses that promote insulin resistance. PLIN proteins have sig­
nificant but differing effects on hepatic insulin sensitivity. For
example, ablation of PLIN2 or PLIN3 in the liver improves
insulin sensitivity, coinciding with reduced steatosis [39–41].
An SNP within the PLIN2 gene impairs VLDL assembly and
correlates with type 2 diabetes [107]. In contrast, overexpres­
sion of PLIN5 in the liver increases steatosis, but improves insu­
lin sensitivity [43]. These effects are consistent with studies on
PLIN5 in other tissues, including muscle and heart, where LD
accumulation is dissociated from insulin resistance [108, 109].
Although the exact role and function of the perilipin proteins
remains an active area of research, much of their effects are
thought to occur via direct interactions with other LD proteins
to influence TAG turnover.
One mechanism through which perilipin proteins may work
is their interaction with lipases, such as ATGL, to inhibit lipol­
ysis under basal conditions [37]. ATGL overexpression in the
liver improves hepatic insulin signaling and reduces DAGs
and ceramides in response to high fat feeding without altering
glucose tolerance [30]. Knockdown of liver ATGL in high fat
fed mice does not affect insulin signaling, but reduces gluco­
neogenesis and improves glucose tolerance without changing
DAG levels [110]. ATGL is now recognized to form a regula­
tory loop with FoxO1 in which ATGL drives FoxO1 activity
and FoxO1 drives ATGL expression reciprocally [111]. This
regulatory circuitry promotes lipolysis and FA oxidation,
which support hepatic gluconeogenesis [111]. Knockdown of
the ATGL coactivator CGI‐58 increases liver TAG, DAG, and
ceramides but improves glucose tolerance and insulin resist­
ance [112]. The inhibitor of ATGL, G0S2, also influences
insulin resistance. G0S2 overexpression improves glucose tol­
erance in mice with increased hepatic glucose uptake [45],
whereas G0S2 ablation in liver reduces steatosis, but improves
whole‐body insulin sensitivity and glucose tolerance despite
increased hepatic gluconeogenesis [46]. Ablation of hypoxia‐
inducible factor 2, an ATGL inhibitor, also promotes TAG
catabolism and FA oxidation while improving glucose toler­
ance [113]. Taken as a whole, inhibiting lipolysis causes
hepatocytes to preferentially take up and burn glucose, which
may explain the generally improved glucose tolerance in mod­
els of reduced hepatic lipolysis. In contrast, promoting lipoly­
sis causes less steatosis as a result of enhanced FA oxidation,
but also increases gluconeogenesis with variable response to
systemic glucose homeostasis depending on the protein
manipulated to regulate lipolysis. Similarly, DGATs also influ­
ence insulin sensitivity. DGAT2 overexpression increases
TAG, DAG, and ceramides but maintains insulin sensitivity
[114]. Thus, while complicated, these studies suggest that
accumulation of TAG and other signaling lipids in the liver can
be dissociated from insulin resistance or glucose intolerance,
suggesting that there are likely many mechanisms that can link
NAFLD to insulin/glucose homeostasis.
In contrast to its effects on liver diseases, the PNPLA3 I148M
polymorphism appears to uncouple NAFLD from its commonly
associated metabolic complications. I148M carriers do not have
elevated serum TAG, which is common in subjects with NAFLD
[67]. Most studies have reported I148M carriers do not have
increased risk for insulin resistance or type 2 diabetes despite
the presence of NAFLD [115–117]. The latter study concluded
that individuals with the I148M SNP, compared to noncarrier
steatotic controls, have increased saturated and monosaturated
hepatic lipids and reduced polyunsaturated lipid species, which
may explain why the lipids are less toxic [89, 90]. Similar to
PNPLA3 I148M, the TM6SF2 E167K variant also does not
increase type 2 diabetes risk [118]. It will be of interest for
future studies to further clarify the role of the PNPLA3 and
TM6SF2 polymorphisms in NAFLD‐related complications and
identify other dietary, genetic, or environmental modifiers to
this relationship.
Hepatitis C virus
The majority of hepatitis C virus (HCV)‐infected patients
have steatosis, which is even more pronounced in those with
genotype 3A [119]. The increase in LD accumulation is likely
due to the fact that the HCV life cycle is tightly intertwined
with LD metabolism. Several HCV proteins bind directly to
LDs and the blockade of de novo LD formation via DGAT1
inhibition prevents the translocation of the proteins from the
ER to LDs and blocks viral replication [120, 121]. Similarly,
the LD proteins PLIN3 and Rab18 are also required for direct
interaction with HCV proteins and viral replication [122,
123]. HCV itself can promote hepatic steatosis and it appears
to do so through the direct inhibition of LD catabolism. HCV
infection reduces TAG hydrolysis, in part through indirect
inhibition of ATGL‐mediated TAG hydrolysis [124]. Similarly,
HCV suppressed the expression of the putative triglyceride
lipase arylacetamide deacetylase (AADAC) to reduce
lipolysis in the early stages of infection [125]. Although LDs
are required for viral replication, VLDL secretion is essential
for packaging and export of newly formed virus. Along
these lines, ablation of AADAC, which also suppresses VLDL
secretion, prevents virus production, suggesting that lipolysis
may be required for late‐stage packaging and export of
the virus [125]. Consistent with this logic, the ATGL coacti­
vator CGI‐58 is also essential for viral replication [126].
These effects may be due to increased CGI‐58‐mediated
VLDL production. Given its role in VLDL secretion, perhaps
it is not surprising that CIDEB is also required for HCV
release [127]. Nonetheless, CIDEB also directly interacts
with the core protein of HCV and is essential for entry and
replication of the virus [127], suggesting a prominent role for
CIDEB in the entire HCV life cycle. Although these data
clearly highlight an essential role of LD proteins and LD
metabolism in regulating the HCV life cycle, much remains to
be elucidated regarding the molecular underpinnings of this
relationship.
 21:  Hepatic Lipid Droplets in Liver Function and Disease 251
Hepatocellular carcinoma
Although LD accumulation is a common observation in hepatic
neoplastic samples, the specific role of LDs and LD proteins in
development of hepatocellular carcinoma (HCC) has not been
extensively studied. Given its high penetrance, the I148M poly­
morphism in PNPLA3 also is likely a major driver of HCC since
it is positively associated with the development of HCC [128,
129]. These effects are expected based on the robust effects of
the I148M SNP on the development of NASH, a known risk
factor for HCC [130]. In addition to PNPLA3 (rs738409),
TM6SF2 (rs58542926) is also a risk factor for the development
of HCC in alcohol‐related cirrhosis [131]. Similar to its pres­
ence in human NAFLD, PLIN1 expression is also increased in
neoplastic liver cells although it is not clear if this merely
reflects more LDs or if it has a pathological role [132]. Finally,
MAGL has been shown to promote HCC via increased inflam­
mation and has been used as a prognostic indicator of HCC
[133, 134]. Clearly, much remains to be understood regarding
the role of LDs in HCC and related hepatic malignancies.
Not all fatty liver (or LDs) is created equal
As our understanding of NAFLD progresses, there is a growing
appreciation that NAFLD is a very heterogeneous disease. As
discussed earlier, numerous factors including diet, genetics, pre­
disposing diseases (obesity, diabetes, etc.) and HCV can lead to
NAFLD through different etiological paths. As a result, mani­
festation of the disease itself, as well as its complications, can
vary widely among individuals. These differences also likely
contribute to the difficulties in identifying efficacious treatment
regiments for NAFLD. Clearly, an important focus of research
and diagnostic medicine moving forward should be towards
characterizing these distinct forms of NAFLD and targeting
therapies specific for each case.
Analogous to NAFLD, there is also immense variability in
LDs both across the liver and within cells. LD proteins are com­
monly observed to have uneven distribution across LDs within
an individual cell. Similarly, the lipid composition of LDs varies
greatly; Raman spectroscopy of NAFLD reveals substantial het­
erogeneity of LD composition [135]. Moreover, the liver is
highly zonated with periportal and perivenous hepatocytes hav­
ing unique and very different gene signatures and metabolic
functions [136]. Beyond differential expression of PLIN pro­
teins [132], almost nothing is known regarding the differences
in LDs among these different populations of cells. It would be
logical to assume that LD metabolism and signaling differ
greatly among and within hepatocytes, which is an area that
begs for future investigation.
Perspective
LDs are the distinguishing trait of NAFLD and play important
roles in both the development and progression of NAFLD, a
disease without efficacious and approved treatments. Under­
standing how alterations in the LD proteome and lipidome
change to influence disease development and severity and define
unique subtypes of NAFLD will be a critical area of research
moving forward. New therapeutic strategies including ASOs
and CRISPR/Cas9 technologies hold great potential for the
liver‐specific targeting of proteins to reverse or attenuate
­d isease pathology. Undoubtedly, the LD will likely be a
key target to help alleviate the burden of NAFLD and its
comorbidities.
REFERENCES
1. Barrows, B.R. and Parks, E.J. Contributions of different fatty acid sources to
very low‐density lipoprotein‐triacylglycerol in the fasted and fed states.
J Clin Endocrinol Metab, 2006;91(4):1446–52.
2. Donnelly, K.L., Margosian, M.R., Sheth, S.S., Lusis, A.J., and Parks, E.J.
Increased lipogenesis and fatty acid reesterification contribute to hepatic
triacylglycerol stores in hyperlipidemic Txnip‐/‐ mice. J Nutr, 2004;
134(6):1475–80.
3. Hannou, S.A., Haslam, D.E., McKeown, N.M., and Herman, M.A. Fructose
metabolism and metabolic disease. J Clin Invest, 2018;128(2):545–55.
4. Grevengoed, T.J., Klett, E.L., and Coleman, R.A. Acyl‐CoA metabolism and
partitioning. Annu Rev Nutr, 2014;34:1–30.
5. Bu, S.Y. and Mashek, D.G. Hepatic long‐chain acyl‐CoA synthetase 5 medi­
ates fatty acid channeling between anabolic and catabolic pathways. J Lipid
Res, 2010;51(11):3270–80.
6. Bu, S.Y., Mashek, M.T., and Mashek, D.G. Suppression of long chain acyl‐
CoA synthetase 3 decreases hepatic de novo fatty acid synthesis through
decreased transcriptional activity. J Biol Chem, 2009;284(44):30474–83.
7. Falcon, A., Doege, H., Fluitt, A. et al. FATP2 is a hepatic fatty acid trans­
porter and peroxisomal very long‐chain acyl‐CoA synthetase. AJP
Endocrinol Metab, 2010;299:E384–93.
8. Doege, H., Baillie, R.A., Ortegon, A.M. et al. Targeted deletion of FATP5
reveals multiple functions in liver metabolism: alterations in hepatic lipid
homeostasis. Gastroenterology, 2006;130:1245–58.
9. Ellis, J.M., Mentock, S.M., Depetrillo, M.A. et al. Mouse cardiac acyl coen­
zyme a synthetase 1 deficiency impairs fatty acid oxidation and induces car­
diac hypertrophy. Mol Cell Biol, 2011;31(6):1252–62.
10. Ellis, J.M., Li, L.O., Wu, P.‐C. et al. Adipose acyl‐CoA synthetase‐1 directs
fatty acids toward beta‐oxidation and is required for cold thermogenesis.
Cell Metab, 2010;12(1):53–64.
11. Li, L.O., Ellis, J.M., Paich, H.A. et al. Liver‐specific loss of long chain acyl‐
CoA synthetase‐1 decreases triacylglycerol synthesis and beta‐oxidation and
alters phospholipid fatty acid composition. J Biol Chem, 2009;284(41):
27816–26.
12. Coleman, R.A. and Mashek, D.G. Mammalian triacylglycerol metabolism:
synthesis, lipolysis, and signaling. Chem Rev, 2011;111(10):6359–86.
13. Vancura, A. and Haldar, D. Purification and characterization of glycerophos­
phate acyltransferase from rat liver mitochondria. J Biol Chem, 1994;
269(44):27209–15.
14. Wurie, H.R., Buckett, L., and Zammit, V.A. Diacylglycerol acyltransferase 2
acts upstream of diacylglycerol acyltransferase 1 and utilizes nascent diglyc­
erides and de novo synthesized fatty acids in HepG2 cells. FEBS J,
2012;279(17):3033–47.
15. Qi, J., Lang, W., Geisler, J.G. et al. The use of stable isotope‐labeled glycerol
and oleic acid to differentiate the hepatic functions of DGAT1 and ‐2. J Lipid
Res, 2012;53(6):1106–16.
16. Villanueva, C.J., Monetti, M., Shih, M. et  al. Specific role for acyl
CoA:diacylglycerol acyltransferase 1 (Dgat1) in hepatic steatosis due to
exogenous fatty acids. Hepatology, 2009;50(2):434–42.
17. Hall, A.M., Kou, K., Chen, Z. et al. Evidence for regulated monoacylglyc­
erol acyltransferase expression and activity in human liver. J Lipid Res,
2012;53(5):990–9.
18. Soufi, N., Hall, A.M., Chen, Z. et al. Inhibiting monoacylglycerol acyltrans­
ferase 1 ameliorates hepatic metabolic abnormalities but not inflammation
and injury in mice. J Biol Chem, 2014;289(43):30177–88.
19. Hall, A.M., Soufi, N., Chambers, K.T. et al. Abrogating monoacylglycerol
acyltransferase activity in liver improves glucose tolerance and hepatic insu­
lin signaling in obese mice. Diabetes, 2014;63(7):2284–96.
20. Wilfling, F., Haas, J.T., and Walther, T.C., and Farese, R.V. Jr. Lipid droplet
biogenesis. Curr Opin Cell Biol, 2014;29:39–45.
252 THE LIVER:  REFERENCES
21. Pol, A., Gross, S.P., and Parton, R.G. Review: biogenesis of the multifunc­
tional lipid droplet: lipids, proteins, and sites. J Cell Biol, 2014;204(5):
635–46.
22. Wilfling, F., Wang, H., Haas, J.T. et al. Triacylglycerol synthesis enzymes
mediate lipid droplet growth by relocalizing from the ER to lipid droplets.
Dev Cell, 2013;24(4):384–99.
23. Xu, W., Wu, L., Yu, M. et al. Differential Roles of Cell Death‐inducing DNA
fragmentation factor‐α‐like effector (CIDE) proteins in promoting lipid
droplet fusion and growth in subpopulations of hepatocytes. J Biol Chem,
2016;291(9):4282–93.
24. Gao, G., Chen, F.‐J., Zhou, L. et  al. Control of lipid droplet fusion and
growth by CIDE family proteins. Biochim Biophys Acta Mol Cell Biol
Lipids, 2017;1862(10, Part B):1197–204.
25. Li, H., Chen, A., Shu, L. et  al. Translocation of CIDEC in hepatocytes
depends on fatty acids. Genes Cells, 2014;19(11):793–802.
26. Krahmer, N., Guo, Y., Wilfling, F. et al. Phosphatidylcholine synthesis for
lipid droplet expansion is mediated by localized activation of
CTP:phosphocholine cytidylyltransferase. Cell Metab, 2011;14(4):504–15.
27. Zimmermann, R., Strauss, J.G., Haemmerle, G. et  al. Fat mobilization in
adipose tissue is promoted by adipose triglyceride lipase. Science,
2004;306(5700):1383–6.
28. Ong, K.T., Mashek, M.T., Bu, S.Y., Greenberg, A.S., and Mashek, D.G.
Adipose triglyceride lipase is a major hepatic lipase that regulates triacylg­
lycerol turnover and fatty acid signaling and partitioning. Hepatology,
2011;53(1):116–26.
29. Wu, J.W., Wang, S.P., Alvarez, F. et al. Deficiency of liver adipose triglycer­
ide lipase in mice causes progressive hepatic steatosis. Hepatology,
2011;54(1):122–32.
30. Turpin, S.M., Hoy, A.J., Brown, R.D. et al. Adipose triacylglycerol lipase is
a major regulator of hepatic lipid metabolism but not insulin sensitivity in
mice. Diabetologia, 2011;54(1):146–56.
31. Khan, S.A., Sathyanarayan, A., Mashek, M.T. et al. ATGL‐catalyzed lipoly­
sis regulates SIRT1 to control PGC‐1α/PPAR‐α signaling. Diabetes,
2015;64(2):418–26.
32. Sekiya, M., Osuga, J.‐I., Yahagi, N. et  al. Hormone‐sensitive lipase is
involved in hepatic cholesteryl ester hydrolysis. J Lipid Res, 2008;49(8):
1829–38.
33. Cao, Z., Mulvihill, M.M., Mukhopadhyay, P. et al. Monoacylglycerol lipase
controls endocannabinoid and eicosanoid signaling and hepatic injury in
mice. Gastroenterology, 2013;144(4):808–17.e15.
34. Lass, A., Zimmermann, R., Haemmerle, G. et al. Adipose triglyceride lipase‐
mediated lipolysis of cellular fat stores is activated by CGI‐58 and defective
in Chanarin‐Dorfman syndrome. Cell Metab, 2006;3(5):309–19.
35. Guo, F., Ma, Y., Kadegowda, A.K.G. et al. Deficiency of liver Comparative
Gene Identification‐58 causes steatohepatitis and fibrosis in mice. J Lipid
Res, 2013;54(8):2109–20.
36. Lord, C.C., Ferguson, D., Thomas, G. et al. Regulation of hepatic triacylg­
lycerol metabolism by CGI‐58 does not require ATGL co‐activation. Cell
Rep, 2016;16(4):939–49.
37. Kimmel, A.R. and Sztalryd, C. The perilipins: major cytosolic lipid droplet‐
associated proteins and their roles in cellular lipid storage, mobilization, and
systemic homeostasis. Annu Rev Nutr, 2016;36:471–509.
38. Straub, B.K., Stoeffel, P., Heid, H., Zimbelmann, R., and Schirmacher, P.
Differential pattern of lipid droplet‐associated proteins and de novo per­
ilipin expression in hepatocyte steatogenesis. Hepatology, 2008;47(6):
1936–46.
39. Imai, Y., Boyle, S., Varela, G.M. et al. Effects of perilipin 2 antisense oligo­
nucleotide treatment on hepatic lipid metabolism and gene expression.
Physiol Genomics, 2012;44(22):1125–31.
40. Najt, C.P., Senthivinayagam, S., Aljazi, M.B. et  al. Liver‐specific loss of
perilipin 2 alleviates diet‐induced hepatic steatosis, inflammation, and fibro­
sis. Am J Physiol Gastrointest Liver Physiol, 2016;310(9):G726–38.
41. Carr, R.M., Patel, R.T., Rao, V. et al. Reduction of TIP47 improves hepatic
steatosis and glucose homeostasis in mice. Am J Physiol Regul Integr Comp
Physiol, 2012;302(8):R996–1003.
42. Wolins, N.E., Quaynor, B.K., Skinner, J.R. et al. OXPAT/PAT‐1 is a PPAR‐
induced lipid droplet protein that promotes fatty acid utilization. Diabetes,
2006;55(12):3418–28.
43. Trevino, M.B., Mazur‐Hart, D., Machida, Y. et al. Liver perilipin 5 expres­
sion worsens hepatosteatosis but not insulin resistance in high fat‐fed mice.
Mol Endocrinol, 2015;29(10):1414–25.
44. Wang, C., Zhao, Y., Gao, X. et al. Perilipin 5 improves hepatic lipotoxicity
by inhibiting lipolysis. Hepatology, 2015;61(3):870–82.
45. Wang, Y., Zhang, Y., Qian, H. et al. The g0/g1 switch gene 2 is an important
regulator of hepatic triglyceride metabolism. PloS One, 2013;8(8):e72315.
46. Zhang, X., Xie, X., Heckmann, B.L. et al. Targeted disruption of G0/G1 switch
gene 2 enhances adipose lipolysis, alters hepatic energy balance, and alleviates
high‐fat diet‐induced liver steatosis. Diabetes, 2014;63(3):934–46.
47. Sugaya, Y. and Satoh, H. Liver‐specific G0 /G1 switch gene 2 (G0s2) expres­
sion promotes hepatic insulin resistance by exacerbating hepatic steatosis in
male Wistar rats. J Diabetes, 2017;9(8):754–63.
48. Chung, C., Doll, J.A., Gattu, A.K. et  al. Anti‐angiogenic pigment epithe­
lium‐derived factor regulates hepatocyte triglyceride content through adi­
pose triglyceride lipase (ATGL). J Hepatol, 2008;48(3):471–8.
49. Das, K.M.P., Wechselberger, L., Liziczai, M. et al. Hypoxia‐inducible lipid
droplet‐associated protein inhibits adipose triglyceride lipase. J Lipid Res,
2018;59(3):531–41.
50. Grahn, T.H.M., Kaur, R., Yin, J. et al. Fat‐specific protein 27 (FSP27) inter­
acts with adipose triglyceride lipase (ATGL) to regulate lipolysis and insulin
sensitivity in human adipocytes. J Biol Chem, 2014;289(17):12029–39.
51. Olzmann, J.A., Richter, C.M., and Kopito, R.R. Spatial regulation of UBXD8
and p97/VCP controls ATGL‐mediated lipid droplet turnover. Proc Natl
Acad Sci U S A, 2013;110(4):1345–50.
52. Yang, L.Y., Kuksis, A., Myher, J.J., and Steiner, G. Origin of triacylglycerol
moiety of plasma very low density lipoproteins in the rat: structural studies.
J Lipid Res, 1995;36(1):125–36.
53. Yang, L.Y., Kuksis, A., Myher, J.J., and Steiner, G. Contribution of de novo
fatty acid synthesis to very low density lipoprotein triacylglycerols: evidence
from mass isotopomer distribution analysis of fatty acids synthesized from
[2H6]ethanol. J Lipid Res, 1996;37(2):262–74.
54. Wang, H., Wei, E., Quiroga, A.D. et  al. Altered lipid droplet dynamics in
hepatocytes lacking triacylglycerol hydrolase expression. Mol Biol Cell,
2010;21(12):1991–2000.
55. Lian, J., Wei, E., Wang, S.P. et  al. Liver specific inactivation of carboxy­
lesterase 3/triacylglycerol hydrolase decreases blood lipids without causing
severe steatosis in mice. Hepatology, 2012;56(6):2154–62.
56. Ye, J., Li, J.Z., Liu, Y. et al. Cideb, an ER‐ and lipid droplet‐associated pro­
tein, mediates VLDL lipidation and maturation by interacting with apolipo­
protein B. Cell Metab, 2009;9(2):177–90.
57. Zhang, J., Zamani, M., Thiele, C. et al. AUP1 (ancient ubiquitous protein 1)
is a key determinant of hepatic very‐low‐density lipoprotein assembly and
secretion. Arterioscler Thromb Vasc Biol, 2017;37(4):633–42.
58. Ohsaki, Y., Cheng, J., Suzuki, M., Fujita, A., and Fujimoto, T. Lipid droplets
are arrested in the ER membrane by tight binding of lipidated apolipoprotein
B‐100. J Cell Sci, 2008;121(14):2415–22.
59. Singh, R., Kaushik, S., Wang, Y. et al. Autophagy regulates lipid metabolism.
Nature, 2009;458(7242):1131–5.
60. Schulze, R.J., Sathyanarayan, A., and Mashek, D.G. Breaking fat: the regula­
tion and mechanisms of lipophagy. Biochim Biophys Acta, 2017;1862(10 Pt
B):1178–87.
61. Kaushik, S. and Cuervo, A.M. AMPK‐dependent phosphorylation of lipid
droplet protein PLIN2 triggers its degradation by CMA. Autophagy,
2016;12(2):432–8.
62. Sinha, R.A., Farah, B.L., Singh, B.K. et al. Caffeine stimulates hepatic lipid
metabolism by the autophagy‐lysosomal pathway in mice. Hepatology,
2014;59(4):1366–80.
63. Huang, Q., Wang, T., Yang, L., and Wang, H.‐Y. Ginsenoside Rb2 alleviates
hepatic lipid accumulation by restoring autophagy via induction of Sirt1 and
activation of AMPK. Int J Mol Sci, 2017;18(5).
64. Zhu, X., Xiong, T., Liu, P. et al. Quercetin ameliorates HFD‐induced NAFLD
by promoting hepatic VLDL assembly and lipophagy via the IRE1a/XBP1s
pathway. Food Chem Toxicol, 2018;114:52–60.
65. Sathyanarayan, A., Mashek, M.T., and Mashek, D.G. ATGL promotes
autophagy/lipophagy via SIRT1 to control hepatic lipid droplet catabolism.
Cell Rep, 2017;19(1):1–9.
66. Tsai, T.‐H., Chen, E., Li, L. et  al. The constitutive lipid droplet protein
PLIN2 regulates autophagy in liver. Autophagy, 2017;13(7):1130–44.
67. Romeo, S., Kozlitina, J., Xing, C. et al. Genetic variation in PNPLA3 confers
susceptibility to nonalcoholic fatty liver disease. Nat Genet, 2008;40(12):
1461–5.
68. Davis, J.N., Lê, K.‐A., Walker, R.W. et  al. Increased hepatic fat in over­
weight Hispanic youth influenced by interaction between genetic variation in
 21:  Hepatic Lipid Droplets in Liver Function and Disease 253
PNPLA3 and high dietary carbohydrate and sugar consumption. Am J Clin
Nutr, 2010;92(6):1522–7.
69. Smagris, E., BasuRay, S., Li, J. et al. Pnpla3I148M knockin mice accumu­
late PNPLA3 on lipid droplets and develop hepatic steatosis. Hepatology,
2015;61(1):108–11.
70. He, S., McPhaul, C., Li, J.Z. et al. A sequence variation (I148M) in PNPLA3
associated with nonalcoholic fatty liver disease disrupts triglyceride hydroly­
sis. J Biol Chem, 2010;285(9):6706–15.
71. Kumari, M., Schoiswohl, G., Chitraju, C. et al. Adiponutrin functions as a
nutritionally regulated lysophosphatidic acid acyltransferase. Cell Metab,
2012;15(5):691–702.
72. Bruschi, F.V., Tardelli, M., Claudel, T., and Trauner, M. PNPLA3 expression
and its impact on the liver: current perspectives. Hepatic Med Evid Res,
2017;9:55–66.
73. Subramanian, V., Rothenberg, A., Gomez, C. et al. Perilipin A mediates the
reversible binding of CGI‐58 to lipid droplets in 3T3‐L1 adipocytes. J Biol
Chem, 2004;279(40):42062–71.
74. Pawella, L.M., Hashani, M., Eiteneuer, E. et al. Perilipin discerns chronic
from acute hepatocellular steatosis. J Hepatol, 2014;60(3):633–42.
75. Kramer, D.A., Quiroga, A.D., Lian, J., Fahlman, R.P., and Lehner, R. Fasting
and refeeding induces changes in the mouse hepatic lipid droplet proteome.
J Proteomics, 2018;181:213–24.
76. Khan, S.A., Wollaston‐Hayden, E.E., Markowski, T.W., Higgins, L., and
Mashek, D.G. Quantitative analysis of the murine lipid droplet‐associated
proteome during diet‐induced hepatic steatosis. J Lipid Res, 2015;56(12):
2260–72.
77. Su, W., Wang, Y., Jia, X. et al. Comparative proteomic study reveals 17β‐
HSD13 as a pathogenic protein in nonalcoholic fatty liver disease. Proc Natl
Acad Sci U S A, 2014;111(31):11437–42.
78. Liu, M., Ge, R., Liu, W. et  al. Differential proteomics profiling identifies
LDPs and biological functions in high‐fat diet‐induced fatty livers. J Lipid
Res, 2017;58(4):681–94.
79. Zhang, X., Wang, Y., and Liu, P. Omic studies reveal the pathogenic lipid
droplet proteins in non‐alcoholic fatty liver disease. Protein Cell,
2017;8(1):4–13.
80. Mancina, R.M., Dongiovanni, P., Petta, S. et al. The MBOAT7‐TMC4 vari­
ant rs641738 increases risk of nonalcoholic fatty liver disease in individuals
of European descent. Gastroenterology, 2016;150(5):1219–30.e6.
81. Luukkonen, P.K., Zhou, Y., Hyötyläinen, T. et  al. The MBOAT7 variant
rs641738 alters hepatic phosphatidylinositols and increases severity of non‐
alcoholic fatty liver disease in humans. J Hepatol, 2016;65(6):1263–5.
82. Kozlitina, J., Smagris, E., Stender, S. et al. Exome‐wide association study
identifies a TM6SF2 variant that confers susceptibility to nonalcoholic fatty
liver disease. Nat Genet, 2014;46(4):352–6.
83. Mahdessian, H., Taxiarchis, A., Popov, S. et al. TM6SF2 is a regulator of
liver fat metabolism influencing triglyceride secretion and hepatic lipid
droplet content. Proc Natl Acad Sci U S A, 2014;111(24):8913–18.
84. Luukkonen, P.K., Zhou, Y., Nidhina Haridas, P.A. et  al. Impaired hepatic
lipid synthesis from polyunsaturated fatty acids in TM6SF2 E167K variant
carriers with NAFLD. J Hepatol, 2017;67(1):128–36.
85. Tian, C., Stokowski, R.P., Kershenobich, D., Ballinger, D.G., and Hinds,
D.A. Variant in PNPLA3 is associated with alcoholic liver disease. Nat
Genet, 2010;42(1):21–3.
86. Chamorro, A.‐J., Torres, J.‐L., Mirón‐Canelo, J.‐A. et al. Systematic review
with meta‐analysis: the I148M variant of patatin‐like phospholipase domain‐
containing 3 gene (PNPLA3) is significantly associated with alcoholic liver
cirrhosis. Aliment Pharmacol Ther, 2014;40(6):571–81.
87. Buch, S., Stickel, F., Trépo, E. et al. A genome‐wide association study con­
firms PNPLA3 and identifies TM6SF2 and MBOAT7 as risk loci for alcohol‐
related cirrhosis. Nat Genet, 2015;47(12):1443–8.
88. Mak, K.M., Ren, C., Ponomarenko, A., Cao, Q., and Lieber, C.S. Adipose
differentiation‐related protein is a reliable lipid droplet marker in alcoholic
fatty liver of rats. Alcohol Clin Exp Res, 2008;32(4):683–9.
89. Schott, M.B., Rasineni, K., Weller, S.G. et  al. β‐Adrenergic induction of
lipolysis in hepatocytes is inhibited by ethanol exposure. J Biol Chem,
2017;292(28):11815–28.
90. Carr, R.M., Dhir, R., Yin, X., Agarwal, B., and Ahima, R.S. Temporal effects
of ethanol consumption on energy homeostasis, hepatic steatosis, and insulin
sensitivity in mice. Alcohol Clin Exp Res, 2013;37(7):1091–9.
91. Williams, B., Correnti, J., Oranu, A. et al. A novel role for ceramide synthase
6 in mouse and human alcoholic steatosis. FASEB J, 2018;32(1):130–42.
92. Lizarazo, D., Zabala, V., Tong, M., Longato, L., and de la Monte, S.M.
Ceramide inhibitor myriocin restores insulin/insulin growth factor signal­
ing for liver remodeling in experimental alcohol‐related steatohepatitis.
J Gastroenterol Hepatol, 2013;28(10):1660–8.
93. Tong, M., Longato, L., Ramirez, T. et al. Therapeutic reversal of chronic
alcohol‐related steatohepatitis with the ceramide inhibitor myriocin. Int J Exp
Pathol, 2014;95(1):49–63.
94. Senkal, C.E., Salama, M.F., Snider, A.J. et al. Ceramide is metabolized to
acylceramide and stored in lipid droplets. Cell Metab, 2017;25(3):686–97.
95. Bozza, P.T., Bakker‐Abreu, I., Navarro‐Xavier, R.A., and Bandeira‐Melo,
C. Lipid body function in eicosanoid synthesis: an update. Prostaglandins
Leukot Essent Fatty Acids, 2011;85(5):205–13.
96. Singal, A.G., Manjunath, H., Yopp, A.C. et al. The effect of PNPLA3 on
fibrosis progression and development of hepatocellular carcinoma: a meta‐
analysis. Am J Gastroenterol, 2014;109(3):325–34.
97. Valenti, L., Alisi, A., Galmozzi, E. et al. I148M patatin‐like phospholipase
domain‐containing 3 gene variant and severity of pediatric nonalcoholic
fatty liver disease. Hepatology, 2010;52(4):1274–80.
98. Carr, R.M., Dhir, R., Mahadev, K. et  al. Perilipin staining distinguishes
between steatosis and non‐alcoholic steatohepatitis in adults and children.
Clin Gastroenterol, 2017;15(1):145–7.
99. Yamaguchi, K., Yang, L., McCall, S. et al. Diacylglycerol acyltranferase 1
anti‐sense oligonucleotides reduce hepatic fibrosis in mice with nonalco­
holic steatohepatitis. Hepatology, 2008;47(2):625–35.
100. Yamaguchi, K., Yang, L., McCall, S. et al. Inhibiting triglyceride synthesis
improves hepatic steatosis but exacerbates liver damage and fibrosis in
obese mice with nonalcoholic steatohepatitis. Hepatology, 2007;45(6):
1366–74.
101. Amrutkar, M., Cansby, E., Nuñez‐Durán, E. et al. Protein kinase STK25
regulates hepatic lipid partitioning and progression of liver steatosis and
NASH. FASEB J, 2015;29(4):1564–76.
102. Ioannou, G.N., Haigh, W.G., Thorning, D., and Savard, C. Hepatic choles­
terol crystals and crown‐like structures distinguish NASH from simple
steatosis. J Lipid Res, 2013;54(5):1326–34.
103. Ioannou, G.N., Subramanian, S., Chait, A. et al. Cholesterol crystallization
within hepatocyte lipid droplets and its role in murine NASH. J Lipid Res,
2017;58(6):1067–79.
104. Savard, C., Tartaglione, E.V., Kuver, R. et  al. Synergistic interaction of
dietary cholesterol and dietary fat in inducing experimental steatohepatitis.
Hepatology, 2013;57(1):81–92.
105. Yasutake, K., Nakamuta, M., Shima, Y. et al. Nutritional investigation of
non‐obese patients with non‐alcoholic fatty liver disease: the significance
of dietary cholesterol. Scand J Gastroenterol, 2009;44(4):471–7.
106. Puri, P., Baillie, R.A., Wiest, M.M. et al. A lipidomic analysis of nonalco­
holic fatty liver disease. Hepatology, 2007;46(4):1081–90.
107. Sentinelli, F., Capoccia, D., Incani, M. et al. The perilipin 2 (PLIN2) gene
Ser251Pro missense mutation is associated with reduced insulin secretion
and increased insulin sensitivity in Italian obese subjects. Diabetes Metab
Res Rev, 2016;32(6):550–6.
108. Bosma, M., Sparks, L.M., Hooiveld, G.J. et al. Overexpression of PLIN5 in
skeletal muscle promotes oxidative gene expression and intramyocellular
lipid content without compromising insulin sensitivity. Biochim Biophys
Acta, 2013;1831(4):844–52.
109. Kuramoto, K., Okamura, T., Yamaguchi, T. et al. Perilipin 5, a lipid droplet‐
binding protein, protects heart from oxidative burden by sequestering fatty
acid from excessive oxidation. J Biol Chem, 2012;287(28):23852–63.
110. Ong, K.T., Mashek, M.T., Bu, S.Y., and Mashek, D.G. Hepatic ATGL
knockdown uncouples glucose intolerance from liver TAG accumulation.
FASEB J, 2013;27(1):313–21.
111. Zhang, W., Bu, S.Y., Mashek, M.T. et al. Integrated regulation of hepatic
lipid and glucose metabolism by adipose triacylglycerol lipase and FoxO
proteins. Cell Rep, 2016;15(2):349–59.
112. Brown, J.M., Betters, J.L., Lord, C. et  al. CGI‐58 knockdown in mice
causes hepatic steatosis but prevents diet‐induced obesity and glucose intol­
erance. J Lipid Res, 2010;51(11):3306–15.
113. DiStefano, M.T., Danai, L.V., Roth Flach, R.J. et al. The lipid droplet pro­
tein hypoxia‐inducible gene 2 promotes hepatic triglyceride deposition by
inhibiting lipolysis. J Biol Chem, 2015;290(24):15175–84.
114. Monetti, M., Levin, M.C., Watt, M.J. et al. Dissociation of hepatic steatosis
and insulin resistance in mice overexpressing DGAT in the liver. Cell
Metab, 2007;6(1):69–78.
254 THE LIVER:  REFERENCES
115. Luukkonen, P.K., Zhou, Y., Sädevirta, S. et al. Ceramides dissociate steato­
sis and insulin resistance in the human liver in non‐alcoholic fatty liver
disease. J Hepatol, 2016;64(5):1167–75.
116. Speliotes, E.K., Butler, J.L., Palmer, C.D. et al. PNPLA3 variants specifi­
cally confer increased risk for histologic nonalcoholic fatty liver disease but
not metabolic disease. Hepatology, 2010;52(3):904–12.
117. Sookoian, S. and Pirola, C.J. Meta‐analysis of the influence of I148M vari­
ant of patatin‐like phospholipase domain containing 3 gene (PNPLA3) on
the susceptibility and histological severity of nonalcoholic fatty liver
disease. Hepatology, 2011;53(6):1883–94.
118. Pirola, C.J. and Sookoian, S. The dual and opposite role of the TM6SF2‐
rs58542926 variant in protecting against cardiovascular disease and confer­
ring risk for nonalcoholic fatty liver: a meta‐analysis. Hepatology,
2015;62(6):1742–56.
119. Adinolfi, L.E., Gambardella, M., Andreana, A. et al. Steatosis accelerates
the progression of liver damage of chronic hepatitis C patients and corre­
lates with specific HCV genotype and visceral obesity. Hepatology,
2001;33(6):1358–64.
120. Herker, E., Harris, C., Hernandez, C. et al. Efficient hepatitis C virus parti­
cle formation requires diacylglycerol acyltransferase‐1. Nat Med,
2010;16(11):1295–8.
121. Camus, G., Herker, E., Modi, A.A. et  al. Diacylglycerol acyltransferase‐1
localizes hepatitis C virus NS5A protein to lipid droplets and enhances NS5A
interaction with the viral capsid core. J Biol Chem, 2013; 288(14):9915–23.
122. Vogt, D.A., Camus, G., Herker, E. et  al. Lipid droplet‐binding protein
TIP47 regulates hepatitis C Virus RNA replication through interaction with
the viral NS5A protein. PLoS Pathog, 2013;9(4):e1003302.
123. Salloum, S., Wang, H., Ferguson, C., Parton, R.G., and Tai, A.W. Rab18
binds to hepatitis C virus NS5A and promotes interaction between sites of
viral replication and lipid droplets. PLoS Pathog, 2013;9(8):e1003513.
124. Camus, G., Schweiger, M., Herker, E. et al. The hepatitis C virus core protein
inhibits adipose triglyceride lipase (ATGL)‐mediated lipid mobilization and
enhances the ATGL interaction with comparative gene identification 58
(CGI‐58) and lipid droplets. J Biol Chem, 2014; 289(52):35770–80.
125. Nourbakhsh, M., Douglas, D.N., Pu, C.H. et al. Arylacetamide deacetylase:
a novel host factor with important roles in the lipolysis of cellular triacylg­
lycerol stores, VLDL assembly and HCV production. J Hepatol, 2013;
59(2):336–43.
126. Vieyres, G., Welsch, K., Gerold, G. et al. ABHD5/CGI‐58, the Chanarin‐
Dorfman syndrome protein, mobilises lipid stores for hepatitis C virus
production. PLoS Pathog, 2016;12(4):e1005568.
127. Cai, H., Yao, W., Li, L. et  al. Cell‐death‐inducing DFFA‐like effector B
contributes to the assembly of hepatitis C virus (HCV) particles and inter­
acts with HCV NS5A. Sci Rep, 2016;6:27778.
128. Liu, Y.‐L., Patman, G.L., Leathart, J.B.S. et al. Carriage of the PNPLA3
rs738409 C >G polymorphism confers an increased risk of non‐alcoholic
fatty liver disease associated hepatocellular carcinoma. J Hepatol,
2014;61(1):75–81.
129. Krawczyk, M., Stokes, C.S., Romeo, S., and Lammert, F. HCC and liver
disease risks in homozygous PNPLA3 pI148M carriers approach mono­
genic inheritance. J Hepatol, 2015;62(4):980–1.
130. Cholankeril, G., Patel, R., Khurana, S., and Satapathy, S.K. Hepatocellular
carcinoma in non‐alcoholic steatohepatitis: current knowledge and implica­
tions for management. World J Hepatol, 2017;9(11):533–43.
131. Stickel, F., Buch, S., Nischalke, H.D. et al. Genetic variants in PNPLA3
and TM6SF2 predispose to the development of hepatocellular carcinoma in
individuals with alcohol‐related cirrhosis. Am J Gastroenterol, 2018;
113(10):1475–83.
132. Straub, B.K., Herpel, E., Singer, S. et al. Lipid droplet‐associated PAT‐
proteins show frequent and differential expression in neoplastic steatogen­
esis. Mod Pathol, 2010;23(3):480–92.
133. Zhu, W., Zhao, Y., Zhou, J. et al. Monoacylglycerol lipase promotes progression
of hepatocellular carcinoma via NF‐κB‐mediated epithelial‐mesenchymal
transition. J Hematol Oncol, 2016;9(1):127.
134. Zhang, J., Liu, Z., Lian, Z. et al. Monoacylglycerol lipase: a novel potential
therapeutic target and prognostic indicator for hepatocellular carcinoma.
Sci Rep, 2016;6:35784.
135. Kochan, K., Maslak, E., Krafft, C. et al. Raman spectroscopy analysis of
lipid droplets content, distribution and saturation level in non‐alcoholic
fatty liver disease in mice. J Biophotonics, 2015;8(7):597–609.
136. Guzmán, M. and Castro, J. Zonation of fatty acid metabolism in rat liver.
Biochem J, 1989;264(1):107–13.
 Lipoprotein Metabolism
22 and Cholesterol Balance
Mariana Acuña‐Aravena and David E. Cohen
Division of Gastroenterology and Hepatology, Joan & Sanford I. Weill Department of Medicine, Weill Cornell
Medical College, New York, NY, USA
INTRODUCTION
Lipids are insoluble or sparingly soluble molecules that are
essential for membrane biogenesis and maintenance of mem-
brane integrity. They also serve as energy sources, hormone
precursors, and signaling molecules. In order to facilitate trans-
port through the relatively aqueous blood, nonpolar lipids,
such as cholesteryl esters or triglycerides, are packaged within
lipoproteins.
Increased concentrations of certain lipoproteins in the circula-
tion are associated strongly with atherosclerosis. Much of the
prevalence of cardiovascular disease, the leading cause of death
in the United States and most Western countries, can be attributed
to elevated plasma concentrations of cholesterol‐rich low‐density
lipoprotein (LDL) particles, as well as lipoproteins that are rich
in triglycerides. Epidemiologically, decreased concentrations of
high‐density lipoprotein (HDL) cholesterol also predispose to
atherosclerotic disease. This chapter highlights the biochemistry
and ­physiology of cholesterol and lipoproteins. Because abun-
dant clinical outcomes data have proven that morbidity and mor-
tality from cardiovascular disease can be reduced by the use of
lipid‐­lowering drugs, mechanisms of pharmacologic interven-
tions that can ameliorate hyperlipidemia will be discussed.
BIOCHEMISTRY AND PHYSIOLOGY
OF CHOLESTEROL AND LIPOPROTEIN
METABOLISM
Lipoproteins are macromolecular aggregates that transport
triglycerides and cholesterol through the blood. Circulating
­ apoA‐I, apoC‐II, and apoE. The following discussion presents
lipoproteins can be differentiated on the basis of density, size,
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
and protein content (Table 22.1). As a general rule, larger, less
dense lipoproteins have a greater percentage composition of
lipids; chylomicrons are the largest and least dense lipoprotein
subclass, whereas HDL are the smallest lipoproteins, containing
the lowest lipid content and the highest proportion of protein.
Structurally, lipoproteins are microscopic spherical particles
ranging from 7 to 100 nm in diameter. Lipoprotein particles
consist of a monolayer of polar, amphipathic lipids surrounding
a hydrophobic core. Each lipoprotein particle also contains one
or more types of apolipoprotein (Table 22.1). The polar lipids
that comprise the surface coat are unesterified cholesterol and
phospholipid molecules, arranged in a monolayer. The hydro-
phobic core of a lipoprotein contains the cholesteryl esters
(cholesterol molecules linked by an ester bond to a fatty acid)
and triglycerides (three fatty acids esterified to a glycerol mol-
ecule). The apolipoproteins (also referred to as apoproteins) are
amphipathic proteins that intercalate into the lipid membrane of
lipoproteins. In addition to stabilizing the structure of lipopro-
teins, apolipoproteins engage in biological functions. They may
act as receptor ligands for lipoprotein particles or may activate
enzymatic activities in the plasma. As will be discussed, the
apolipoprotein composition determines the metabolic fate of
the lipoprotein.
From a metabolic perspective, lipoprotein particles can be
divided into lipoproteins that participate in the delivery of tri-
glyceride molecules to muscle and fat tissue (the apolipoprotein
B (apoB)‐containing lipoproteins: chylomicrons and very‐low‐
density lipoprotein, VLDL) and lipoproteins that are involved
primarily in cholesterol transport (HDL and the remnants of
apoB‐containing lipoproteins). HDL also serves as a reservoir
for exchangeable apolipoproteins in the plasma, including
each lipoprotein class in the context of its function.
256 THE LIVER:  BIOCHEMISTRY AND PHYSIOLOGY OF CHOLESTEROL AND LIPOPROTEIN METABOLISM
Table 22.1  Characteristics of plasma lipoproteins
CM
Density (g mL ) −1
<0.95
Diameter (nm) 75–1200
Total lipid (% wt) 98
Composition, % dry weight
Protein 2 10
Triglycerides 83
Unesterified cholesterol + 8 22
cholesterol ester
Phospholipids (% wt lipid) 7 18
Electrophoretic mobilitya None
Major apolipoproteins B48, A‐I, A‐IV, E,
C‐I, C‐II, C‐III
CM, chylomicron; VLDL, very‐low‐density lipoprotein; IDL, intermediate‐density lipoprotein; LDL, low‐density lipoprotein; HDL, high‐density lipoprotein.
a
 Electrophoretic mobility of lipoprotein particles is designated relative to migration of plasma α‐ and β‐globulins.
Adapted from [103] with permission of Elsevier.
Metabolism of apoB‐containing lipoproteins
The primary function of apoB‐containing lipoproteins is to deliver
fatty acids in the form of triglycerides to muscle tissue to be used
for ATP biogenesis and to adipose tissue for long‐term storage.
Chylomicrons are formed in the intestine and transport dietary
triglycerides, whereas VLDL particles are formed by the liver and
transport triglycerides that are synthesized endogenously. For
conceptual purposes, the metabolic lifespan of apoB‐containing
lipoproteins can be divided into three phases: assembly, intravas-
cular metabolism, and receptor‐mediated clearance. This is also a
convenient categorization because pharmacologic agents are
available that influence these different phases.
Assembly of apoB‐containing lipoproteins
The cellular mechanisms by which chylomicrons and VLDL are
assembled are quite similar. Regulation of the assembly process
depends upon the availability of apoB and triglycerides, as well as
the activity of microsomal triglyceride transfer protein (MTP) [1].
The gene that encodes apoB is transcribed principally in the
intestine and the liver. Apart from this tissue‐specific expres-
sion, there is little transcriptional regulation of the apoB gene.
By contrast, a key regulatory event that differentiates chylomi-
cron from VLDL metabolism is editing of apoB mRNA. Within
enterocytes but not hepatocytes, a protein named apoB‐editing
complex 1 (apobec‐1) is expressed. This protein constitutes the
catalytic subunit of the apoB‐editing complex, which deami-
nates a cytosine at position 6666 of the apoB mRNA molecule
[2]. This converts the cytosine to a uridine base. As a result, the
codon containing this nucleotide is converted from a glutamine
to a premature stop. When translated, the intestinal form apoB48
is 48% as long as the full‐length protein that is expressed in the
liver and referred to as apoB100. As a consequence, chylomi-
crons, the apoB‐containing lipoprotein produced by the intes-
tine, contain apoB48. By contrast, VLDL particles produced by
the liver contain apoB100.
Figure  22.1a illustrates the cellular mechanisms for the
assembly and secretion of apoB‐containing lipoproteins. As
the  apoB protein is translated by ribosomes, it crosses into
the ­endoplasmic reticulum (ER) [3, 4]. Within the ER, triglycer-
ide molecules are added co‐translationally to the elongating
apoB protein (i.e. apoB is lipidated) by the action of MTP [3].
VLDL IDL LDL HDL
0.95–1.006 1.006–1.019 1.019–1.063 1.063–1.210
30–80 25–35 18–25 5–12
90 82 75 40
18 25 33
50 31 9 8
29 45 30
22 21 29
Pre‐β β β α or Pre‐β
B100, E, C‐I, B100, E, C‐I, B100 A‐I, A‐II, C‐I,
C‐II, C‐III C‐II, C‐III C‐II, C‐III, E
Once apoB has been fully translated, the nascent lipoprotein is
enlarged in the Golgi apparatus, during which MTP adds addi-
tional triglycerides to the core of the particle [4]. MTP also
enhances rates of cholesterol esterification by removing choles-
teryl esters from their site of synthesis in the endoplasmic retic-
ulum and transferring them into nascent apoB lipoproteins [5].
This assembly process produces the lipoprotein particles, each
of which contains a single molecule of apoB.
The diet is the main source of triglycerides in chylomicrons
(Figure 22.1b), so their assembly, secretion, and metabolism are
collectively referred to as the exogenous pathway of lipoprotein
metabolism. By contrast, cholesteryl esters in chylomicrons are
derived mainly from biliary cholesterol (approximately 75%),
with the remainder contributed by dietary sources. During
digestion, cholesteryl esters and triglycerides in food are hydro-
lyzed by lipases secreted from the pancreas to form unesterified
cholesterol, free fatty acids, and monoglycerides [3]. Bile salts,
phospholipids, and cholesterol are secreted by the liver into bile
and stored in the gallbladder during fasting as micelles and vesi-
cles, which are macromolecular lipid aggregates that form due
to the detergent properties of bile salt molecules. The stimulus
of eating a meal promotes emptying of gallbladder bile into the
small intestine, where micelles and vesicles solubilize the
digested lipids. Lipid absorption into enterocytes of the duode-
num and jejunum is facilitated mainly by micelles [6]. Long‐
chain fatty acids and monoglycerides are taken up separately
into the enterocyte by carrier‐mediated transport and then re‐
esterified to form triglycerides by the enzyme diacylglycerol
acyltransferase (DGAT) [3]. By contrast, medium‐chain fatty
acids are absorbed directly into the portal blood and metabo-
lized by the liver. The steps of cholesterol absorption are mecha-
nistically defined [7]. Dietary and biliary cholesterol from
micelles enter the enterocyte via a protein channel named
Niemann-Pick C1-like 1 protein (NPC1L1). Some of this cho-
lesterol is immediately pumped back into the intestinal lumen
by the ATP‐dependent action of a heterodimeric protein, ATP‐
binding cassette (ABC) G5/ABCG8 (ABCG5/G8) expressed
in enterocytes and hepatocytes [8]. The fraction of cholesterol
that remains is esterified to a long‐chain fatty acid by acyl‐
CoA:cholesterol acyltransferase (ACAT). Once triglycerides
and cholesteryl esters are packaged together with apoB48,
apoA‐IV is added to stabilize the surface, allowing additional
 22:  Lipoprotein Metabolism and Cholesterol Balance 257

(a)
Enterocyte or hepatocyte

Cytosol
Presence of triglycerides

Lipidation
Ribosome CM

MTP MTP
VLDL
ApoB

Absence of triglycerides

Degradation

ApoB Endoplasmic reticulum

(b)
Cholesterol absorption Triglyceride absorption
LYMPH ENTEROCYTE INTESTINAL LYMPH ENTEROCYTE INTESTINAL
LUMEN LUMEN

Cholesterol 2 Fatty acid

+
NPC1L1 Monoglyceride
ACAT

DGAT
Cholesteryl
ABCG5/G8 Exogenous
ester
triglyceride
Bile salt

(c)
Chylomicron formation
LYMPH ENTEROCYTE INTESTINAL
LUMEN
Exogenous
CM triglycerides
apoB48

Cholesteryl
ester

Figure 22.1  Assembly and secretion of apolipoprotein B‐containing lipoproteins. (a) Chylomicron (CM) and VLDL particles are assembled and
secreted by similar mechanisms in the enterocyte and hepatocyte, respectively. The apoB protein (i.e. apoB48 or apoB100) is translated by ribo-
somes and enters the lumen of the endoplasmic reticulum. If triglycerides are available, the apoB protein is lipidated by the action of microsomal
triglyceride transfer protein (MTP) in two distinct steps, accumulating triglyceride as well as cholesteryl ester molecules. The resulting CM or
VLDL particle is secreted by exocytosis into the lymphatics by enterocytes or into the plasma by hepatocytes. In the absence of triglycerides, the
apoB protein is degraded. Modified from [101] with permission of John Wiley & Sons. (b) Exogenous triglycerides and cholesterol absorption are
simultaneously absorbed from the intestinal lumen by different mechanisms. Cholesterol is taken up from micelles across a regulatory channel
named NPC1L1. A fraction of the cholesterol is pumped back into the lumen by ABCG5/G8, a heterodimeric ATP‐dependent plasma membrane
protein. The remainder of the cholesterol is converted to cholesteryl esters by ACAT. Triglycerides are taken up as fatty acids and monoglycerides,
which are re‐esterified by DGAT. Modified from [102] with permission of Wolters Kluwer. (c) Exogenous triglycerides plus cholesteryl esters are
assembled into CM within the enterocyte.
258 THE LIVER:  BIOCHEMISTRY AND PHYSIOLOGY OF CHOLESTEROL AND LIPOPROTEIN METABOLISM
core lipidation, and subsequently apoA‐I is attached [9]. The
assembled chylomicron particle is then exocytosed into the lym-
phatics for transport into the circulation via the thoracic duct
(Figure  22.1c). The plasma concentration of triglyceride‐rich
chylomicrons varies in proportion to dietary fat intake.
VLDL particles comprise triglycerides that are assembled by
the liver using plasma fatty acids derived from adipose tissue or
synthesized by de novo lipogenesis. For this reason, the assem-
bly, secretion, and metabolism of VLDL particles are often
referred to as the endogenous pathway of lipoprotein metabo-
lism. Hepatocytes synthesize triglycerides in response to
increased free fatty acid flux to the liver. This typically occurs in
response to fasting, thereby insuring a continuous supply of
fatty acids for delivery to muscle in the absence of triglycerides
from the diet. Dietary saturated fats as well as carbohydrates
also stimulate the ­synthesis of triglycerides within the liver [10,
11]. By cellular mechanisms that are similar to those that pro-
duce chylomicrons, MTP in hepatocytes lipidates apoB100 to
form nascent VLDL particles. Under the continued influence of
MTP, the nascent VLDL particles coalesce with larger triglycer-
ide droplets and are secreted directly into the circulation. VLDL
particles may also acquire apoE, apoC‐I, apoC‐II, and apoC‐III
within the hepatocyte prior to secretion. However, these apoli-
poproteins may also be transferred to VLDL from HDL during
their circulation through the bloodstream.
The translation of apoB48 in the intestine and apoB100 in the
liver is constitutive. This permits the immediate production of
chylomicrons and VLDL particles when triglyceride molecules
are available. In the absence of triglycerides, such as in entero-
cytes during fasting, apoB is degraded by a variety of cellular
mechanisms [12].
Recent evidence from whole‐exome sequencing, exome chip,
and genome‐wide association studies (GWAS) have identified
new genes that regulate apoB‐containing lipoprotein metabo-
lism [13]. One of these is SORT1, which encodes the lysosomal
trafficking protein sortilin‐1. A single nucleotide polymorphism
(SNP) which increases the hepatic expression of sortilin‐1 is
associated with reduced plasma levels of LDL cholesterol. This
is partially explained because sortilin‐1 facilitates the post‐
transcriptional degradation of apoB by a lysosomal‐dependent
pathway, leading to decreased VLDL secretion [14]. Another
SNP found in the gene encoding transmembrane 6 superfamily
member 2 (TM6SF2), which reduces the protein level, is associ-
ated with low plasma triglyceride and cholesterol concentra-
tions, but also non‐alcoholic fatty liver disease (NAFLD). This
is because TM6SF2 decreases VLDL concentrations in the
plasma, thereby increasing hepatic triglyceride contents [15, 16].
Additional studies have revealed that the lowered TM6SF2 pro-
tein levels impair apoB lipidation and VLDL assembly, but not
its secretion [17]. Similar to TM6SF2, a loss‐of‐function SNP in
patatin‐like phospholipase domain‐containing 3 (PNPLA3) is
also associated with NAFLD, and impairs VLDL secretion [18].
Intravascular metabolism of apoB‐containing lipoproteins
Within the circulation, chylomicrons and VLDL particles must
be activated in order to target triglyceride delivery to muscle and
fat tissues (Figure 22.2a). This requires the addition of an opti-
mal complement of apoC‐II molecules, which occurs at least in
part by aqueous transfer of apoC‐II from HDL particles [19].
Because there is an inherent delay in the transfer of apoC‐II to
chylomicrons and VLDL particles, this allows time for
­widespread circulation of triglyceride‐rich particles throughout
the body.
Lipoprotein lipase (LPL) is a lipolytic enzyme that is primar-
ily synthesized and secreted by a variety of parenchymal cells,
including myocytes, adipocytes, skeletal muscle cells, mac-
rophages, and mammary gland cells [20, 21]. This glycoprotein
is transported from the interstitial space to endothelial cell sur-
face of the luminal capillary in muscle and fat tissues by the
glycosylphosphatidylinositol (GPI)‐linked protein (GPIHBP1)
[22]. GPIHBP1 also serves to anchor LPL to the endothelial cell
plasma membrane. Once chylomicrons and VLDL particles
acquire apoC‐II, they can bind to LPL, which hydrolyzes tri-
glycerides from the core of the lipoprotein (Figure 22.2b). LPL‐
mediated lipolysis liberates free fatty acids and glycerol, which
are then taken up by the neighboring parenchymal cells. Because
the expression level and intrinsic activity of LPL in muscle
­versus fat tissue is regulated according to the fed/fasting state,
this allows the body to direct the delivery of fatty acids prefer-
entially to muscle during fasting and to fat post prandially [21].
(a)
VLDL CM
LIVER apoB100 apoB48 INTESTINE
apoE apoC-II
α-HDL apoAI
(b) Capillary in
muscle or adipose tissue
VLDL CM
apoB48
apoB100
Fatty Fatty
acids Lipoprotein acids
lipase
Endothelium
Figure 22.2  Intravascular metabolism of apoB‐containing lipoproteins.
(a) Following secretion, chylomicron (CM) and VLDL particles are
­activated for lipolysis when they encounter HDL particles in the plasma
and acquire the exchangeable apolipoprotein apoC‐II. (b) When CM and
VLDL particles circulate into capillaries of muscle or fat tissue, apoC‐II
promotes binding of the particle to lipoprotein lipase, which is bound to
the surface of endothelial cells. Lipoprotein lipase mediates hydrolysis of
triglycerides, but not cholesteryl esters, from the core of the lipoprotein
particle. The resulting fatty acids are taken up into muscle or fat tissue.
Modified from [101] with permission of John Wiley & Sons.
 22:  Lipoprotein Metabolism and Cholesterol Balance 259
The rate of lipolysis of chylomicrons and VLDL triglycerides is
also controlled by apoC‐III, which is an inhibitor of LPL a­ ctivity
[23]. Antagonism of LPL activity by apoC‐III may be an
additional mechanism promoting widespread distribution of
­
­triglyceride‐rich particles in the circulation.
Receptor‐mediated clearance of apoB‐containing
lipoproteins
As LPL continues to hydrolyze triglycerides from chylomi-
crons and VLDL, the particles become progressively depleted
of ­triglycerides and relatively enriched with cholesterol. Once
approximately 50% of triglycerides have been removed, the
particles lose affinity for LPL and dissociate. The exchangeable
apolipoproteins apoA‐I and apoC‐II (as well as apoC‐I and
apoC‐III) are then transferred to HDL in exchange for apoE
(Figure 22.3a) [24], which serves as a high‐affinity ligand for
receptor‐mediated clearance [25]. Upon acquiring apoE, the
particles are termed chylomicron or VLDL remnants [26]. As
explained below, intermediate‐density lipoproteins (IDLs) are
also remnant particles.
Remnants of chylomicrons and VLDL particles, as well as
some IDL particles, are taken up by the liver in a three‐step
process (Figure  22.3b) [26]. The first step is sequestration
of  the particles within the space of Disse between the fenes-
trated endothelium of the liver sinusoids and the sinusoidal
(­basolateral) plasma membrane of hepatocytes. Sequestration
requires that the remnant particles become small enough dur-
ing the process of lipolysis to fit in between the endothelial
cells. Once in the space of Disse, remnants are bound and
sequestered by large heparan sulfate proteoglycans (HSPGs).
The next step is remodeling within the space of Disse by the
action of hepatic lipase, a lipolytic enzyme that is similar to
LPL, but is expressed by hepatocytes. Hepatic lipase appears to
optimize the triglyceride content of remnant particles, so they
are efficiently cleared by receptor‐mediated mechanisms. The
final phase of remnant clearance is receptor‐mediated uptake
[27]. This is accomplished by one of four pathways. At the
sinusoidal hepatocyte plasma membrane, remnant particles
may be bound and taken up by the LDL receptor, the LDL
receptor‐related protein (LRP), or by HSPGs. A separate path-
way is mediated by the combined activities of LRP and HSPGs.
These efficient and redundant mechanisms allow for efficient
particle clearance, so that the half‐life of remnants in the
plasma is approximately 30 minutes.
Formation and clearance of LDL particles
Whereas apoB48‐containing chylomicron remnants are com-
pletely cleared from the plasma, the presence of apoB100 alters
the metabolism of VLDL remnants so that only approximately
50% are cleared by the pathways for remnant particles. The dif-
ference begins when the other 50% are metabolized to a greater
extent by LPL, becoming smaller and relatively deficient in tri-
glycerides and enriched in cholesteryl esters. When converted to
remnants following exchange of apolipoproteins with HDL,
these more dense particles become IDL. Because IDL particles
contain apoE, some of them may be cleared into the liver by
remnant receptor pathways (Figure  22.3b) [28]. However, the
(a)
VLDL
VLDL CM
CM
remnant remnant
apoB100 apoB48
apoB100 apoB48
apoCII
apoE
α-HDL apoAI
(b) Sinusoidal endothelium CM or VLDL remnant
HSPG
Space
Sequestration of
Disse
Hepatic lipase
Lipolysis
Uptake LDLr LRP
HSPG
Hepatocyte
Figure 22.3  Formation and hepatic uptake of remnant particles. (a)
Upon completion of hydrolysis, chylomicron (CM) and VLDL particles
lose affinity for lipoprotein lipase. When an HDL particle is encoun-
tered, apoC‐II is transferred back to HDL particles in exchange for
apoE. The resulting particles are CM and VLDL remnants and interme-
diate‐density lipoproteins (IDLs), which are VLDL particles that inter-
act for prolonged periods with LPL to become more dense. (b) The
activity of lipoprotein lipase results in remnant lipoprotein particles that
are small enough in size to enter the space of Disse. Remnant lipopro-
teins are sequestered in the space of Disse by binding to high molecular
weight heparin sulfate proteoglycan (HSPG) molecules. This is fol-
lowed by binding of hepatic lipase, which promotes lipolysis of some
residual triglycerides in the core of the remnant lipoproteins and the
release of fatty acids (as indicated by the dashed arrow). Uptake of rem-
nant lipoprotein particles into hepatocytes is mediated by the LDL
receptor (LDLr), the LDL‐related protein (LRP), a complex formed
between LRP and HSPG, or HSPG alone. Modified from [101] with
permission of John Wiley & Sons.
remainder are converted to LDL by hepatic lipase, which further
hydrolyzes triglycerides in the core of IDL. The further reduc-
tion in size of the particles results in the transfer of apoE to
HDL. As a result, LDLs are distinct, cholesteryl ester‐enriched
lipoproteins with apoB100 as their only apolipoprotein [29].
The LDL receptor is the primary receptor capable of clearing
significant amounts of LDL from the plasma [30]. It is concen-
trated on the surface of hepatocytes, macrophages, lympho-
cytes, adrenocortical cells, gonadal cells, and smooth muscle
cells. Due to the lack of apoE, the LDL particles are relatively
weak ligands for the LDL receptor [30]. As a result, the half‐life
of LDL in the circulation is markedly prolonged (2–4 days).
This explains why LDL cholesterol accounts for approximately
65–75% of total plasma cholesterol.
Interaction of apoB100 with the LDL receptor facilitates
receptor‐mediated endocytosis of LDL particles and subsequent
260 THE LIVER:  BIOCHEMISTRY AND PHYSIOLOGY OF CHOLESTEROL AND LIPOPROTEIN METABOLISM
vesicle fusion with a lysosome [30]. The LDL receptor is recy-
cled to the cell surface, while the LDL particle is hydrolyzed
within the lysosome by lysosomal acid lipase to release unest-
erified cholesterol and free fatty acids. LDL‐derived cholesterol
is released from lysosomes through the action of two proteins,
the membrane‐bound Niemann–Pick C1 (NPC1) and the solu-
ble (NPC2) [31]. Recent studies suggest that cholesterol is
transported first to replenish the plasma membrane in order to
maintain optimal levels, and the excess is then trafficked to the
ER, where it impacts three major homeostatic pathways [32].
First, intracellular cholesterol inhibits 3‐hydroxy‐3‐methylglu-
taryl‐coenzyme A reductase (HMG‐CoA reductase), the enzyme
that catalyzes the rate‐limiting step in de novo cholesterol
­synthesis. Second, cholesterol activates ACAT to increase ester-
ification and storage of cholesterol in the cell. Third, LDL
receptor expression is downregulated, reducing further uptake
of cholesterol into the cells. The majority of LDL receptors
(approximately 70%) are expressed on the surface of hepato-
cytes. As a result, the liver is primarily responsible for the
removal of LDL particles from the circulation. Reductions in
intracellular cholesterol levels, such as occur during therapy
with statin drugs, lead to LDL receptor upregulation. This is due
to the proteolytic processing and activity of sterol regulatory
element‐binding protein 2 (SREBP‐2) [33].
Sortilin‐1 also promotes the uptake of LDL by LDL receptor‐
dependent and ‐independent pathways. In the independent path-
way, sortilin‐1 serves as a cell surface receptor for LDL, facilitating
its cellular uptake and lysosomal degradation [14]. Another
important regulator of LDL receptor clearance is the proprotein
convertase subtilisin/kexin type 9 (PCSK9). PCSK9 is a serine
protease that binds LDL receptor and induces its degradation in
lysosomes via a mechanism that requires neither ubiquitination
nor autophagy [34]. Loss‐of‐function mutations in PCSK9 result
in marked decreased plasma LDL cholesterol concentrations.
A variant of LDL is produced when the glycoprotein apo(a) is
covalently attached to apoB100 by a disulfide bond, generating
lipoprotein(a) or Lp(a) [35]. Apo(a) contains from 3 to more
than 50 plasminogen‐like domains, which give rise to a hetero-
geneous population of Lp(a) particles. The physiological func-
tions and catabolism of Lp(a) remain under study [36].
High plasma concentrations of LDL and Lp(a) particles
­constitute risk factors for the development of atherosclerotic
cardiovascular disease [37]. LDL particles that are not taken up
by LDL receptor may penetrate the intima of blood vessels and
bind to proteoglycans. There, they are subject to oxidation,
which results in lipid peroxidation and may create reactive alde-
hyde intermediates that fragment apoB100. The modified LDL
is internalized by scavenger receptors (e.g. SR‐A), which are
expressed predominantly by mononuclear phagocytic cells [38].
Unlike the LDL receptor, scavenger receptors are not downreg-
ulated when the phagocytic immune cells begin to accumulate
cholesterol. As a result, the continued accumulation of oxidized
LDL in macrophages can lead to foam cell formation (choles-
terol‐rich macrophages). Oxidized LDL also causes upregula-
tion of cytokine production, impairs endothelial function, and
increases expression of endothelial adhesion molecules. All of
these effects increase the local inflammatory response and
­promote atherosclerosis. Foam cells are a major constituent of
atherosclerotic lesions, and excessive foam cell death can
destabilize atherosclerotic plaques, in part due to the liberation
of matrix metalloproteinases which may precipitate acute
ischemic cardiovascular events in the form of myocardial infarc-
tions and strokes. Moreover, Lp(a) exhibits prothrombotic/anti‐
fibrinolytic effects, which can accelerate atherosclerosis and
lead to the intimal deposition of Lp(a) cholesterol [39].
HDL metabolism and reverse cholesterol
transport
Virtually all cells in the body are capable of synthesizing all of
the cholesterol they require. However, only the liver has the
capacity to eliminate cholesterol by secretion into bile in its unes-
terified form or following its conversion to bile salts. In addition
to serving as a reservoir for exchangeable apolipoproteins for the
metabolism of apoB‐containing lipoproteins, HDLs are lipopro-
teins that play a key role in cholesterol homeostasis by removing
excess cholesterol from cells and transporting it through plasma
to the liver. This process is often referred to as reverse cholesterol
transport (RCT) (Figure 22.4a) [40]. The process of RCT protects
against the development of atherosclerosis by removing excess
cholesterol from plaque macrophages [41]. Inflammatory and
oxidative damage also play pivotal roles in atherogenesis, and
certain HDL particle populations have been shown to be pro­
tective by preventing oxidation of LDL particles and inhibiting
endothelial expression of adhesion molecules [42]. Moreover,
epidemiological studies have shown that low plasma levels of
HDL cholesterol are associated with an increased risk in cardio-
vascular disease. Therefore, high plasma concentrations of HDL
cholesterol in plasma are considered atheroprotective [43].
HDL particles are heterogeneous in structure and biological
activity, and have been classified by physical properties into a
variety of subclasses [44]. These particles contain 14 different
apolipoproteins, the major ones being apoA‐I and apoA‐II [45].
ApoA‐I is the main structural determinant of HDL, and partici-
pates in the formation, as well as interaction of the particle with
its receptor, scavenger receptor class B type I (SR‐BI) [40, 46].
ApoA‐I is also important for the antioxidant properties of HDL
[47]. ApoA‐II is more hydrophobic than apoA‐I, which helps
to confer the lipid surface curvature in HDL maturation and to
stabilize the particle [48].
HDL formation
HDL formation occurs mainly in the liver, although a percent-
age is contributed by the small intestine [43]. Unlike other lipo-
proteins, HDL particles are not assembled in cells; instead they
are formed in the extracellular space and mature in the plasma
as they are remodeled. The earliest events occur when lipid‐poor
apoA‐I is secreted by the liver or intestine [49], or dissociates
from lipoprotein particles in the plasma [50]. These amphip-
athic apoA‐I molecules interact with ABCA1, which is imbed-
ded in the sinusoidal membrane of hepatocytes or basolateral
membrane of enterocytes [51, 52]. ABCA1 incorporates a small
amount of membrane phospholipids and unesterified choles-
terol into the apoA‐I molecule [52]. The resulting small disk‐
shaped HDL particle consists mainly of phospholipid and
apoA‐I. The particle that is formed is referred to as nascent or
pre‐β‐HDL, due to its characteristic migration on agarose gels.
 22:  Lipoprotein Metabolism and Cholesterol Balance 261

(a)
LIVER BLOOD TISSUES

Cholesterol LCAT
α-HDL
pre-β-HDL
Phospholipid
apoA-I
ABCA1

Cholesterol
SR-BI
LCAT
PLTP
Hepatic CETP
lipase

(b)

CM or VLDL
Albumin
remnant

Lyso-PC
PLTP
LCAT

Cell
CETP

α -HDL

Figure 22.4  Reverse cholesterol transport. (a) The process of reverse cholesterol transport begins when apoA‐I is secreted from the liver. ApoA‐I
in plasma interacts with ATP‐binding cassette protein AI (ABCA1), which incorporates a small amount of phospholipid and unesterified cholesterol
from hepatocyte plasma membranes to form a discoidal‐shaped pre‐β‐HDL particle. Due to the activity of lecithin:cholesterol acyltransferase
(LCAT) in plasma, pre‐β‐HDL particles mature to form spherical α‐HDL. Spherical α‐migrating HDL particles function to accept excess unesterified
cholesterol from the plasma membranes of cells in a wide variety of tissues. The unesterified cholesterol is transferred from the cell to nearby HDL
particles by diffusion through the plasma. As shown in (b), LCAT and phospholipid transfer protein (PLTP) increase the capacity of HDL to accept
unesterified cholesterol molecules from cells by allowing for expansion of the core and the surface coat of the particle. Cholesteryl ester transfer
protein (CETP) exchanges some of the cholesteryl ester molecules from the core of HDL for triglycerides from the core of remnant particles. HDL
particles interact with scavenger receptor, class B type I (SR‐BI), which mediates selective hepatic uptake of cholesteryl esters, but not apoA‐I. This
process is facilitated when hepatic lipase hydrolyzes triglycerides from the core of the particle. The remaining apoA‐I molecules may begin the cycle
of reverse cholesterol transport again. In (a), solid arrows indicate metabolic events in HDL metabolism, whereas dashed arrows denote transfer of
molecules. (b) LCAT, PLTP, and CETP promote the removal of excess cholesterol from the plasma membranes of cells. LCAT removes a fatty acid
from a phosphatidylcholine molecule in the surface coat of α‐HDL (or pre‐β‐HDL) and esterifies an unesterified cholesterol molecule on the surface
of the particle. The resulting lysophosphatidylcholine (lyso‐PC) becomes bound to albumin in the plasma, whereas the cholesteryl ester migrates
spontaneously into the core of the lipoprotein particle. The unesterified cholesterol molecules that are consumed by LCAT are replaced by unesteri-
fied cholesterol from cells. HDL phospholipids that are consumed by LCAT action are replaced with excess phospholipids from remnant particles
by the activity of PLTP. As described in (a), CETP increases the efficiency of cholesterol movement to the liver by exchanging cholesteryl ester
molecules from the core of α‐HDL for triglycerides from the core of remnant particles. In (b), solid arrows denote protein‐mediated lipid transfer,
whereas dashed arrows indicate that lipids move by diffusion through the plasma. Modified from [101] with permission of John Wiley & Sons.

Intravascular maturation of HDL

occurs due to the activity of two distinct circulating proteins
Because disk‐shaped pre‐β‐HDL particles are relatively ineffi- (Figure 22.4b). Lecithin:cholesterol acyltransferase (LCAT) binds
cient at removing excess cholesterol from cell membranes, they preferentially to disk‐shaped HDL particles and converts choles-
must mature into spherical particles within the plasma. This terol molecules within the particle to cholesteryl esters [53].
262 THE LIVER:  BIOCHEMISTRY AND PHYSIOLOGY OF CHOLESTEROL AND LIPOPROTEIN METABOLISM
This is accomplished by transesterification of a fatty acid from a
phosphatidylcholine molecule on the surface of the HDL to the
hydroxyl group of a cholesterol molecule. This two‐step reaction
also creates a lysophosphatidylcholine molecule, which departs
from the particle upon binding to serum albumin [53]. Because
they are highly insoluble, cholesteryl esters produced by LCAT
spontaneously migrate into the core of the HDL particle. The
development of a hydrophobic core converts the pre‐β‐HDL to a
spherical HDL particle, which exhibits α migration on agarose
gels [45].
The second important protein that contributes to HDL matura-
tion in the plasma is phospholipid transfer protein (PLTP), which
is mainly expressed in adipose tissue, lung and liver [54]. PLTP
transfers phospholipids from the surface coat of apoB‐contain-
ing remnant particles to the surface coat of HDL by penetrating
into the surface of both lipoproteins and forming a ternary com-
plex [55]. During LPL‐mediated lipolysis of apoB‐containing
lipoproteins, the particles become smaller as triglycerides are
removed from the core. This leaves a relative excess of phospho-
lipids on the surface of the particle. Because phospholipids are
insoluble and cannot otherwise dissociate from a particle, PLTP
removes excess phospholipids and thereby maintains the appro-
priate surface concentration for the shrinking core. By transfer-
ring phospholipids to the surface of HDL, PLTP also replaces the
molecules that are consumed by the LCAT reaction. This allows
the core of HDL to continue to enlarge. Based on studies on mice
lacking PLTP, it has been demonstrated that the majority of HDL
phospholipids are transferred from the surface coat of remnant
particles [54].
HDL‐mediated cholesterol efflux from cells
Cellular cholesterol efflux is the mechanism by which excess
insoluble cholesterol molecules are removed from cells. This
occurs when unesterified cholesterol is transferred from the
plasma membrane of cells to an HDL particle. The mechanism
of cholesterol efflux varies depending upon the cell type and the
type of HDL particle [56]. Lipid‐poor pre‐β‐HDL particles can
promote cholesterol efflux by interacting with ABCA1. In addi-
tion to beginning the process of HDL formation by the liver, this
is also an active mechanism for removing excess cholesterol
from cells that serves to protect macrophages within the suben-
dothelial space from cholesterol‐induced cytotoxicity. Besides
ABCA1‐mediated cholesterol efflux, spherical HDL particles
very efficiently stimulate cholesterol export by three other path-
ways [56]. First, a passive mechanism includes simple diffusion
via the aqueous phase in the absence of binding to a specific cell
surface protein. Although cholesterol has very low monomeric
solubility, it can dissociate in appreciable amounts and travel
short distances through the plasma to acceptor particles that
are enriched with phospholipids on their surfaces. The second
is facilitated diffusion mediated by SR‐BI, whereby HDL parti-
cles interact with SR‐BI on the plasma membrane. promoting
­cholesterol efflux. Finally, an active efflux of cholesterol to
spherical HDL particles is mediated by ABCG1, which is
expressed in macrophages. Quantitatively, efflux to spherical
HDL particles accounts for most of the removal of excess cho-
lesterol from cells. This capacity of HDL to remove cellular
cholesterol is enhanced through the activities of LCAT and
PLTP, which prevent the surface coat of the particle from
becoming saturated with unesterified cholesterol from cells.
Delivery of HDL cholesterol to the liver
When mature HDL particles circulate to the liver, they interact
with SR‐BI, the principal HDL receptor [57], which is expressed
on the sinusoidal plasma membranes of hepatocytes. Although
it can mediate cholesterol efflux from cells with excess choles-
terol, SR‐BI in the liver promotes selective uptake of lipids, a
process whereby the cholesterol and cholesteryl esters of HDL
particles are taken up into the cell, in the absence of uptake of
apolipoproteins [46]. During SR‐BI‐mediated selective lipid
uptake, apoA‐I is liberated to participate in pre‐β‐HDL forma-
tion. As a result, the “lifespan” of an HDL particle is 2–5 days,
suggesting that each apoA‐I molecule can participate in many
cycles of RCT. Among the tissues that express high levels of
SR‐BI are the adrenal glands and gonads, presumably reflecting
their requirement for cholesterol in order to support steroido-
genesis [58].
Delivery of cholesterol from extrahepatic tissues to the liver
is optimized by two additional proteins, cholesteryl ester trans-
fer protein (CETP) and hepatic lipase. CETP is a plasma protein
that transfers cholesteryl esters from mature spherical HDL par-
ticles to the cores of remnant lipoproteins in exchange for a tri-
glyceride molecule, which is inserted into the core of the HDL
particle (Figure  22.4b) [59]. This process allows the body to
utilize remnant particles that have completed their function of
triglyceride transport for purposes of transporting cholesterol to
the liver. Removal of cholesteryl ester molecules from HDL
appears to serve two functions. First, it further increases the
capacity of HDL to take on additional cholesterol molecules
from cells. Second, it makes the process of selective uptake by
SR‐BI more efficient [59]. This is because hydrolysis of triglyc-
erides by hepatic lipase on the hepatocyte surface facilitates the
activity of SR‐BI (Figure 22.4a).
There are conflicting data concerning the net effects of CETP
on lipid metabolism. Studies in humans have suggested that
CETP may improve cholesterol efflux by increasing the capac-
ity of HDL particles to take up cholesterol from macrophages,
preventing lipid accumulation in these cells [60]. On the other
hand, the observation that deleterious gene mutations in CETP
gene had the effect of increasing plasma HDL cholesterol and
apoA‐I concentrations in humans with no evidence for prema-
ture atherosclerosis suggests that CETP inhibition could repre-
sent a promising therapeutic approach for cardiovascular disease
[43]. This notion was supported by the selective beneficial
effects of certain hypolipidemic drugs observed in in rodents,
which are naturally deficient in CETP, but not in humans [61].
Although CETP inhibitors were advanced to phase 3 clinical
trials, an apparent lack of benefit along with concerns around
toxicity prevented any candidate drugs from coming to the
market.
Biliary lipid secretion
Once cholesterol is delivered to the liver, it is eliminated by
­biliary secretion. A key essential step occurs  when a fraction
of  the cholesterol is converted to bile salts. The enzymatic
 22:  Lipoprotein Metabolism and Cholesterol Balance 263
convertion into bile salt molecules catalyzed by CYP7A1 is
referred to as the classical pathway; an alternative pathway is
initiated by the sterol 27‐hydroxylase (CYP27A) followed by
oxysterol 7α‐hydroxylase (CYP7B1) [62]. In humans the alter-
native pathway contributes only a minor proportion of bile salts
under normal circumstances, whereas in rodents the classical
and alternative pathways contribute about equally to bile acid
synthesis [63]. Bile salts, unlike cholesterol, are highly soluble
in water. Moreover, bile salts are biological detergents that pro-
mote the formation of micelles. These macromolecular aggre-
gates are rich in phospholipids, which are derived from
hepatocyte membranes and solubilize cholesterol in bile for
transport from the liver to the small ­intestine, essentially func-
tioning as counterparts to HDL particles in plasma [64].
Bile formation begins when ABCB11, an ATP‐driven canali-
cular membrane transporter, functions to pump bile salts into
bile [65]. Bile salts within bile then activate two other ATP‐
dependent transporters on the canalicular membrane of the
hepatocyte: ABCB4 and the heterodimer ABCG5/G8 [66].
These transporters promote the secretion of phospholipid and
cholesterol molecules, respectively, which together with bile
salts form mixed micelles. Biliary lipids are stored in the gall-
bladder during fasting. The stimulus of a fatty meal leads to
gallbladder contraction, which propels its contents into the
small intestine. As described above, bile facilitates the digestion
and absorption of fats, in addition to promoting the elimination
of endogenous cholesterol. Hepatic hypersecretion of choles-
terol into bile and gallbladder hypomotility, among other genetic
and environmental factors, promote the development of choles-
terol gallstone disease (cholelithiasis) [67].
Although biliary secretion into the intestinal lumen was clas-
sically considered as the only significant route for cholesterol
elimination, more recent evidence has revealed that transintesti-
nal cholesterol excretion (TICE) accounts for up to 35% [68, 69].
TICE reflects the net effect of four cholesterol fluxes within the
enterocyte: cholesterol absorption from the intestine by the
action of NPC1L1 on the apical plasma membrane, cholesterol
excretion into the intestinal lumen mainly via ABCG5/8, also on
the apical membrane, the uptake of plasma lipoproteins at the
basolateral membrane, and the basolateral excretion of choles-
terol in chylomicrons into the lymphatics [70].
Cholesterol balance
Because cholesterol is converted by the liver to bile salts and is
secreted unmodified into bile, overall cholesterol balance
depends upon the disposition of both types of molecules. Rather
than being lost in the feces after participating in cholesterol
transport and fat digestion, most bile salt molecules are recy-
cled when they are taken up by high‐affinity transport proteins
in the distal ileum. Bile salts enter the portal circulation and are
transported back to the liver, where they are cleared by hepato-
cytes from the blood with high first‐pass efficiency. Bile salts
are then re‐secreted into bile. This process of recycling bile salts
between the liver and intestine is referred to as the enterohepatic
circulation [71].
The enterohepatic circulation is highly efficient, allowing
<5% of secreted bile salts to be lost in the feces. However,
because bile salts are secreted in such large amounts, the small
fractional loss of bile salts amounts to about 0.4 g per day [71].
Considering that cholesterol is the substrate for bile salt synthe-
sis, fecal bile salts represent a source of cholesterol loss from
the body. Sensitive nuclear hormone receptors within the liver
are capable of detecting the rate of loss of bile salts into the
feces [63]. These receptors tightly regulate transcription of bile
salt synthetic genes. As a result, the liver synthesizes precisely
an amount of bile salts that is sufficient to replace what is lost in
the feces. The gut microbiota also regulates bile acids synthesis,
bile acid pool size and composition, and enterohepatic circula-
tion of bile acids [72].
On a daily basis, an average of 24 g of bile salts are secreted
[71], together with 11 g of phospholipids [73] and approxi-
mately 0.9 g of cholesterol. TICE accounts for about 0.7 g of
cholesterol, and the average American diet contributes approxi-
mately 0.4 g per day to intestinal cholesterol. Therefore, dietary
cholesterol represents only a minor fraction (20%) compared to
the endogenous (i.e. biliary) cholesterol that passes through the
intestine [74]. The extent to which intestinal cholesterol is
absorbed appears to be genetically regulated. Each individual
absorbs a fixed percentage of intestinal cholesterol. In the popu-
lation, percentages range from as low as 20% to more than 80%
[75, 76]. For example, when an average individual absorbs
50%, this will amount to half of the 2.0 g (i.e. 0.9 g of biliary
cholesterol plus 0.7 g TICE cholesterol plus 0.4 g of dietary
cholesterol). The other half (1.0 g) is lost in the feces. This,
combined with a loss of 0.4 g per day of cholesterol in the form
of fecal bile salts, yields a total cholesterol loss from the body
of 1.4 g per day. Taking into account intestinal absorption of
dietary cholesterol and reabsorption of biliary cholesterol, total
body cholesterol synthesis is 1.0 g (i.e. cholesterol synthesis =
fecal loss of cholesterol + bile salts − dietary cholesterol intake)
[74]. This amounts to more than double what is consumed in the
average diet.
PHARMACOLOGIC CLASSES
AND AGENTS
The decision to treat dyslipidemia is largely dependent upon the
calculated cardiovascular risk. A number of clinical algorithms
exist for determining initiation of therapy. Goals for lipid lower-
ing were established in the 2001 National Cholesterol Education
Program Adult Treatment Panel III (ATP III) guidelines [77],
which were updated in 2004 based upon the results of several
additional large, randomized clinical trials [78]. These guidelines
provide target LDL levels based on 10‐year risk of death from
cardiovascular disease. More recent guidelines for the manage-
ment of dyslipidemia from the American Heart Association are
no longer based on target LDL levels. Importantly, the guidelines
stress that the first intervention should be therapeutic lifestyle
changes, including reduction of dietary saturated fat and choles-
terol intake, weight reduction, increased physical activity, and
possibly stress reduction.
Whereas successful dietary therapy can reduce total choles-
terol by 5–25%, depending upon adherence and the metabolic
basis for elevated cholesterol concentrations. If this approach is
unsuccessful or insufficient to normalize lipid levels, drug
264 THE LIVER:  PHARMACOLOGIC CLASSES AND AGENTS
therapy is generally recommended. Eight classes of agents,
including pharmacologic drugs and dietary supplements, are
available for modification of lipid metabolism, and more are
under study. Three of these classes (i.e. inhibitors of cholesterol
synthesis, bile salt sequestrants, and cholesterol absorption
inhibitors) have relatively well‐defined effects on lipid metabo-
lism. Whereas the overall effects of the other five classes are
clear, their molecular mechanisms of action are diverse and still
subjects of active investigation. The following is a condensed
summary of the mechanisms of these agents.
Inhibitors of cholesterol synthesis
Cholesterol synthesis inhibitors, commonly known as statins,
competitively inhibit the activity of HMG‐CoA reductase, the
rate‐limiting enzyme in cholesterol synthesis [30]. This acti-
vates a cellular signaling cascade culminating in the activation
of SREBP‐2, which is a transcription factor that upregulates
expression of the gene encoding the LDL receptor. Increased
LDL receptor expression causes increased uptake of plasma
LDL and, consequently, decreases plasma LDL cholesterol con-
centrations [79].
Inhibitors of bile salt absorption
The bile salt sequestrants are cationic polymers that bind non-
covalently to negatively charged bile salt molecules in the small
intestine. The resin–bile salt complex cannot be reabsorbed in
the distal ileum and is excreted into the stool. Decreased bile
salt reabsorption by the ileum partially interrupts enterohepatic
bile salt circulation, causing hepatocytes to upregulate CYP7A1,
the rate‐limiting enzyme in bile salt synthesis from cholesterol
[80]. The increase in bile salt synthesis decreases hepatocyte
cholesterol concentration, leading to increased expression of
the LDL receptor and enhanced LDL clearance from the circu-
lation [30]. The effectiveness of bile salt sequestrants in clear-
ing LDL from plasma is partially offset by concurrent
upregulation of hepatic cholesterol and triglyceride synthesis,
which stimulates the production of VLDL particles by the liver
[80]. As a result, bile salt sequestrants may also raise triglycer-
ide levels and should be used with caution in patients with
hypertriglyceridemia.
Inhibitors of cholesterol absorption
Cholesterol absorption inhibitors reduce cholesterol absorption
by the small intestine. This includes dietary cholesterol, but
more important they reduce the reabsorption of biliary choles-
terol, which comprises the majority of intestinal cholesterol
[74]. Like statins and bile salt‐binding resins, cholesterol
absorption inhibitors reduce LDL cholesterol by increasing
clearance via the LDL receptor. In doing so, they enhance the
catabolism of VLDL and LDL apoB100, without affecting the
production rate of these lipoproteins [80].
There are two available cholesterol absorption inhibitors:
plant sterols and ezetimibe. Plant sterols and stanols are
­naturally present in vegetables and fruits. They are similar in
molecular structure to cholesterol, but are substantially more
hydrophobic [81]. As a result, plant sterols and stanols
displace cholesterol from micelles, increasing the loss of
cholesterol from the feces [82]. The plant sterols and stanols
are themselves poorly absorbed. Based upon their mechanism
of action, gram quantities of plant sterols and stanols are
required to reduce plasma LDL cholesterol concentrations by
approximately 15%. Because an average diet contains 200–
400 mg of plant sterols and stanols, the molecules must be
enriched in dietary supplements (approximately 2 g) to be
effective [83].
Ezetimibe at very low concentrations reduces intestinal choles-
terol absorption by about 50%, without reducing the absorption
of triglycerides or fat‐soluble vitamins. It decreases cholesterol
movement from micelles into the enterocyte by inhibiting uptake
through the brush border protein NPC1L1 [84] (Figure  22.1b).
TICE is markedly increased by inhibition of NPC1L1 using
ezetimibe [69], suggesting that modulation of TICE by other
mechanisms could provide an additional strategy for reducing the
risk of atherosclerotic cardiovascular disease [68].
A reduction in cholesterol absorption that is achieved by
either plant sterols and stanols or ezetimibe would be expected
to decrease the cholesterol content of chylomicrons and there-
fore the movement of cholesterol from the intestine to the liver.
This is because chylomicrons retain the intestinal cholesterol
that was originally incorporated into the particles after they are
metabolized to remnants by the liver. Within the liver, choles-
terol derived from chylomicron remnants contributes to the
cholesterol that is packaged into VLDL particles. Therefore,
inhibiting cholesterol absorption reduces cholesterol incorpo-
ration into VLDL particles, and decreases LDL cholesterol
concentrations in the plasma. Reduced hepatic cholesterol con-
tents lead to upregulation of the LDL receptor, which also con-
tributes to the mechanism of LDL lowering by cholesterol
absorption inhibitors [80].
Fibrates
Fibrates bind and activate peroxisome proliferator‐activated
receptor α (PPARα), a nuclear receptor expressed in hepato-
cytes, skeletal muscle, macrophages, and the heart [85, 86].
Upon fibrate binding, PPARα heterodimerizes with the retinoid
X receptor (RXR). This heterodimer binds to peroxisome
proliferator response elements (PPREs) present in the pro-
­
moter region of specific genes and activates transcription of the
target genes [85].
Activation of PPARα by fibrates results in numerous changes
in lipid metabolism that act to decrease plasma triglyceride
­levels and increase plasma HDL [87]. The decrease in plasma
triglyceride levels is caused by increased muscle cell expression
of LPL, decreased hepatic expression of apoC‐III and increased
hepatic oxidation of fatty acids. The increased expression of
LPL in muscle results in increased uptake of triglyceride‐rich
lipoproteins, with a resultant decrease in plasma triglyceride
levels. Because apoC‐III normally functions to inhibit interac-
tion of triglyceride‐rich lipoproteins with their receptors, the
decrease in hepatic production of apoC‐III may potentiate the
increased LPL activity [23].
The mechanisms by which fibrates raise plasma HDL levels
are complex [88]. Whereas PPARα decreases hepatocyte
­production of apoA‐I in the mouse, the opposite appears be the
 22:  Lipoprotein Metabolism and Cholesterol Balance 265
case in humans. This contributes directly to increased plasma
HDL. Upregulation of SR‐BI and ABCA1 in macrophages pre-
sumably promotes cholesterol efflux from these cells in vivo.
Hepatocytes also reduce expression of SR‐BI in response to
PPARα, which provides another mechanism for increased HDL
levels in the plasma.
Fibrates also lower LDL levels modestly. The lower LDL lev-
els result from a PPARα‐mediated shift in hepatocyte metabo-
lism toward fatty acid oxidation. PPARα increases the expression
of numerous enzymes involved in fatty acid transport and oxida-
tion [85]. This increases fatty acid catabolism, leading to
decreased triglyceride synthesis and VLDL production. PPARα
activation also results in LDL particles of larger size, which
appear to be taken up more efficiently by LDL receptor. Many
of these effects of PPARα on lipid metabolism are still the sub-
ject of basic and clinical investigation to develop of more selec-
tive PPARα agonists that are capable of targeting selective
aspects of lipid metabolism [89].
Niacin
Niacin (nicotinic acid, vitamin B3) is a water‐soluble vitamin.
At physiologic concentrations, it is a substrate in the synthesis
of nicotinamide adenine dinucleotide (NAD) and nicotinamide
adenine dinucleotide phosphate (NADP), which are important
cofactors in intermediary metabolism.
The pharmacologic use of niacin necessitates large doses
(1500–3000 mg day−1) and is independent of the conversion of
nicotinic acid to NAD or NADP. Niacin increases HDL cho-
lesterol and decreases plasma LDL cholesterol and triglycer-
ide concentrations, as well as Lp(a) levels [90]. The effects of
niacin are partially mediated by a G protein‐coupled receptor
on adipocytes, which decreases the activity of adipose tissue
hormone‐sensitive lipase. Decreased hormone‐sensitive lipase
reduces peripheral tissue triglyceride catabolism, and there-
fore decreases the flux of free fatty acids to the liver. This,
together with local effects of niacin in liver, decreases the rate
of hepatic triglyceride synthesis and VLDL production. Niacin
also decreases the fractional catabolic rate of apoA‐I [90]. The
increased plasma apoA‐I increases plasma HDL concentra-
tions and presumably augments reverse cholesterol transport.
Omega‐3 fatty acids
The omega‐3 fatty acids eicosopantaenoic acid (EPA) and doco-
sahexaenoic acid (DHA), also referred to as fish oil, are effec-
tive at reducing plasma triglycerides [91]. Although the
molecular mechanisms are incompletely understood, the effects
are reduced hepatic triglyceride biosynthesis and increased fatty
acid oxidation in the liver [92].
Proprotein convertase subtilisin/kexin type 9
inhibitors
PCSK9, which increases LDL cholesterol concentrations by pro-
moting the degradation of the LDL receptor, was mapped as the
third locus associated with autosomal dominant hypercholes-
terolemia together with LDL receptor and apoB, suggesting that
gain‐of‐function mutations produce increase levels of LDL cho-
lesterol in the plasma [93]. Monoclonal antibodies, genetic
knockdown, and gene repair techniques have been used to inhibit
PCSK9 [94]. The commercially available monoclonal antibodies
alirocumab and evolocumab reduce LDL cholesterol levels from
30 to 70%, even in severe familial hypercholesterolemia and sta-
tin‐resistant patients [95], and may produce other pleiotropic
anti‐atherogenic effects independent of LDL lowering [96].
Inhibitors of VLDL secretion
Therapies against dyslipidemia include the reduction of LDL
cholesterol and triglyceride levels using agents that reduce
VLDL secretion from the liver into the plasma. Mipomersen is
an antisense oligonucleotide to apoB, whereas lomitapide is an
MTP inhibitor [97]. Both are effective at lowering plasma LDL
concentration, even in patients with homozygous familial
hypercholesterolemia, who lack both alleles of the LDL
­receptor. However, a mechanism‐based side‐effect is hepatic
steatosis, and steatorrhea in the case of lomitapide.
Agents in development
Bempedoic acid, currently in phase 3 clinical trials, is an inhib-
itor of adenosine triphosphate citrate lyase (ACLY), a cytosolic
enzyme upstream of HMG‐CoA reductase, which catalyzes the
reaction that produces acetyl‐CoA from citrate, inhibiting the
­synthesis of cholesterol and ultimately leading to upregulation
of LDL receptor [98]. Other agents that increase LPL activity
and thereby reduce plasma triglycerides concentrations are
currently under study, including antisense oligonucleotides to
acpoC-III, and gemcabene, which is a small molecule that
reduces apoC‐III expression, also inhibits lipid synthesis, as
well as enhancing VLDL clearance [97, 99]. Other approaches
to increasing LPL activity are antisense oligonucleotides and
monoclonal antibodies against angiopoietin‐like protein 3,
which is secreted by the liver and inhibits LPL [100].
CONCLUSIONS
The efficacy of available lipid‐lowering drugs that reduce LDL
represents an important advance in reducing cardiovascular
disease mortality. Future advances will build upon a rich
­
­knowledge base of lipoprotein metabolism in order to identify
new biochemical targets for the management of plasma lipids,
as well as related diseases such as type 2 diabetes and non‐­
alcoholic fatty liver disease.
ACKNOWLEDGMENTS
This work was supported in part by research grants DK056626,
DK048873, and DK103046 from the National Institutes of
Health (US Public Health Service) to DEC. MAA is the
­recipient of a 2017 NASH Fatty Liver Postdoctoral Research
Fellowship Award from the American Liver Foundation.
266 THE LIVER:  REFERENCES

REFERENCES 27. Foley, E.M., Gordts, P., Stanford, K.I. et  al. Hepatic remnant lipoprotein
clearance by heparan sulfate proteoglycans and low‐density lipoprotein
receptors depend on dietary conditions in mice. Arterioscler Thromb Vasc
1. Hussain, M.M., Rava, P., Walsh, M., Rana, M., and Iqbal, J. Multiple
Biol, 2013;33:2065–74.
­functions of microsomal triglyceride transfer protein. Nutr Metab (Lond),
28. Packard, C.J. and Shepherd, J. Lipoprotein heterogeneity and apolipoprotein
2012;9:14.
B metabolism. Arterioscler Thromb Vasc Biol, 1997;17:3542–56.
2. Smith, H.C. RNA binding to APOBEC deaminases: not simply a substrate
29. Ramasamy, I. Recent advances in physiological lipoprotein metabolism.
for C to U editing. RNA Biol, 2017;14:1153–65.
Clin Chem Lab Med, 2014;52:1695–727.
3. Hussain, M.M. Intestinal lipid absorption and lipoprotein formation. Curr
30. Brown, M.S. and Goldstein, J.L. A receptor‐mediated pathway for choles-
Opin Lipidol, 2014;25:200–6.
terol homeostasis. Science, 1986;232:34–47.
4. Sirwi, A. and Hussain, M.M. Lipid transfer proteins in the assembly of apoB‐
31. Kwon, H.J., Abi‐Mosleh, L., Wang, M.L. et  al. Structure of N‐terminal
containing lipoproteins. J Lipid Res, 2018;59:1094–102.
domain of NPC1 reveals distinct subdomains for binding and transfer of cho-
5. Iqbal, J., Rudel, L.L., and Hussain, M.M. Microsomal triglyceride transfer
lesterol. Cell, 2009;137:1213–24.
protein enhances cellular cholesteryl esterification by relieving product
32. Infante, R.E. and Radhakrishnan, A. Continuous transport of a small fraction
­inhibition. J Biol Chem, 2008;283:19967–80.
of plasma membrane cholesterol to endoplasmic reticulum regulates total
6. Carey, M.C., Small, D.M., and Bliss, C.M. Lipid digestion and absorption.
cellular cholesterol. Elife, 2017;6.
Annu Rev Physiol, 1983;45:651–77.
33. Horton, J.D., Goldstein, J.L., and Brown, M.S. SREBPs: activators of the
7. Wang, D.Q. and Cohen, D.E. Absorption and excretion of cholesterol and
complete program of cholesterol and fatty acid synthesis in the liver. J Clin
other sterols, in Clinical Lipidology: Companion to Braunwald’s Heart
Invest, 2002;109:1125–31.
Disease (ed. C. Ballantyne), Elsevier, Philadelphia, PA, 2014, pp. 25–42.
34. Wang, Y., Huang, Y., Hobbs, H.H., and Cohen, J.C. Molecular characteriza-
8. Yu, X.H., Qian, K., Jiang, N. et al. ABCG5/ABCG8 in cholesterol excretion
tion of proprotein convertase subtilisin/kexin type 9‐mediated degradation of
and atherosclerosis. Clin Chim Acta, 2014;428:82–8.
the LDLR. J Lipid Res, 2012;53:1932–43.
9. Hesse, D., Jaschke, A., Chung, B., and Schurmann, A. Trans‐Golgi proteins
35. Brown, M.S. and Goldstein, J.L. Plasma lipoproteins: teaching old dogmas
participate in the control of lipid droplet and chylomicron formation. Biosci
new tricks. Nature, 1987;330:113–14.
Rep, 2013;33:1–9.
36. Banach, M. Lipoprotein (a)‐We know so much yet still have much to learn.
10. Xu, X., So, J.S., Park, J.G., and Lee, A.H. Transcriptional control of
J Am Heart Assoc, 2016;5.
hepatic  lipid metabolism by SREBP and ChREBP. Semin Liver Dis,
37. Emdin, C.A., Khera, A.V., Natarajan, P. et al. Phenotypic characterization of
2013;33:301–11.
genetically lowered human lipoprotein(a) levels. J Am Coll Cardiol,
11. Wang, Y., Viscarra, J., Kim, S.J., and Sul, H.S. Transcriptional regulation of
2016;68:2761–72.
hepatic lipogenesis. Nat Rev Mol Cell Biol, 2015;16:678–89.
38. Boullier, A., Bird, D.A., Chang, M.K. et  al. Scavenger receptors, oxidized
12. Fisher, E.A. The degradation of apolipoprotein B100: multiple opportunities
LDL, and atherosclerosis. Ann N Y Acad Sci, 2001;947:214–22; discussion
to regulate VLDL triglyceride production by different proteolytic pathways.
22–3.
Biochim Biophys Acta, 2012;1821:778–81.
39. Nordestgaard, B.G., Chapman, M.J., Ray, K. et  al. Lipoprotein(a) as a
13. Christoffersen, M. and Tybjaerg‐Hansen, A. Novel genes in LDL metabo-
­cardiovascular risk factor: current status. Eur Heart J, 2010;31:2844–53.
lism – a comprehensive overview. Curr Opin Lipidol, 2015;26:179–87.
40. Santos‐Gallego, C.G., Badimon, J.J., and Rosenson, R.S. Beginning to
14. Strong, A., Patel, K., and Rader, D.J. Sortilin and lipoprotein metabolism:
understand high‐density lipoproteins. Endocrinol Metab Clin North Am,
making sense out of complexity. Curr Opin Lipidol, 2014;25:350–7.
2014;43:913–47.
15. Holmen, O.L., Zhang, H., Fan, Y. et al. Systematic evaluation of coding
41. Rye, K.A., Bursill, C.A., Lambert, G., Tabet, F., and Barter, P.J. The metabo-
­variation identifies a candidate causal variant in TM6SF2 influencing
lism and anti‐atherogenic properties of HDL. J Lipid Res, 2009;50(Suppl):
total cholesterol and myocardial infarction risk. Nat Genet, 2014;46:
S195–200.
345–51.
42. Tabet, F. and Rye, K.A. High‐density lipoproteins, inflammation and oxida-
16. Kozlitina, J., Smagris, E., Stender, S. et al. Exome‐wide association study
tive stress. Clin Sci (Lond), 2009;116:87–98.
identifies a TM6SF2 variant that confers susceptibility to nonalcoholic fatty
43. Kardassis, D., Mosialou, I., Kanaki, M., Tiniakou, I., and Thymiakou, E.
liver disease. Nat Genet, 2014;46:352–6.
Metabolism of HDL and its regulation. Curr Med Chem, 2014;21:2864–80.
17. Smagris, E., Gilyard, S., BasuRay, S., Cohen, J.C., and Hobbs, H.H.
44. Rosenson, R.S., Brewer, H.B., Jr., Chapman, M.J. et al. HDL measures, par-
Inactivation of Tm6sf2, a gene defective in fatty liver disease, impairs lipida-
ticle heterogeneity, proposed nomenclature, and relation to atherosclerotic
tion but not cecretion of very low density lipoproteins. J Biol Chem,
cardiovascular events. Clin Chem, 2011;57:392–410.
2016;291:10659–76.
45. Zannis, V.I., Fotakis, P., Koukos, G. et al. HDL biogenesis, remodeling, and
18. Pirazzi, C., Adiels, M., Burza, M.A. et  al. Patatin‐like phospholipase
catabolism. Handb Exp Pharmacol, 2015;224:53–111.
domain‐containing 3 (PNPLA3) I148M (rs738409) affects hepatic VLDL
46. Gillard, B.K., Bassett, G.R., Gotto, A.M., Jr, Rosales, C., and Pownall, H.J.
secretion in humans and in vitro. J Hepatol, 2012;57:1276–82.
Scavenger receptor B1 (SR‐B1) profoundly excludes high density lipopro-
19. Wolska, A., Dunbar, R.L., Freeman, L.A. et  al. Apolipoprotein C‐II: New
tein (HDL) apolipoprotein AII as it nibbles HDL‐cholesteryl ester. J Biol
findings related to genetics, biochemistry, and role in triglyceride metabo-
Chem, 2017;292:8864–73.
lism. Atherosclerosis, 2017;267:49–60.
47. Tabet, F., Remaley, A.T., Segaliny, A.I. et  al. The 5A apolipoprotein A‐I
20. Goldberg, I.J. 2017 George Lyman Duff Memorial Lecture: Fat in the blood,
mimetic peptide displays antiinflammatory and antioxidant properties in
fat in the artery, fat in the heart: triglyceride in physiology and disease.
vivo and in vitro. Arterioscler Thromb Vasc Biol, 2010;30:246–52.
Arterioscler Thromb Vasc Biol, 2018;38:700–6.
48. Gao, X., Yuan, S., Jayaraman, S., and Gursky, O. Role of apolipoprotein A‐II
21. He, P.P., Jiang, T., OuYang, X.P. et al. Lipoprotein lipase: biosynthesis, regu-
in the structure and remodeling of human high‐density lipoprotein (HDL):
latory factors, and its role in atherosclerosis and other diseases. Clin Chim
protein conformational ensemble on HDL. Biochemistry, 2012;51:4633–41.
Acta, 2018;480:126–37.
49. Miles, R.R., Perry, W., Haas, J.V. et al. Genome‐wide screen for modulation
22. Fong, L.G., Young, S.G., Beigneux, A.P. et al. GPIHBP1 and plasma triglyc-
of hepatic apolipoprotein A‐I (ApoA‐I) secretion. J Biol Chem, 2013;288:
eride metabolism. Trends Endocrinol Metab, 2016;27:455–69.
6386–96.
23. Rocha, N.A., East, C., Zhang, J., and McCullough, P.A. ApoCIII as a cardio-
50. Gursky, O. Structural stability and functional remodeling of high‐density
vascular risk factor and modulation by the novel lipid‐lowering agent volane-
lipoproteins. FEBS Lett, 2015;589:2627–39.
sorsen. Curr Atheroscler Rep, 2017;19:62.
51. Segrest, J.P., Jones, M.K., Catte, A. et al. Surface density‐induced pleating of
24. Cooper, A.D. Hepatic uptake of chylomicron remnants. J Lipid Res,
a lipid monolayer drives nascent high‐density lipoprotein assembly.
1997;38:2173–92.
Structure, 2015;23:1214–26.
25. Phillips, M.C. Apolipoprotein E isoforms and lipoprotein metabolism.
52. Wang, S. and Smith, J.D. ABCA1 and nascent HDL biogenesis. Biofactors,
IUBMB Life, 2014;66:616–23.
2014;40:547–54.
26. Mahley, R.W. and Ji, Z.S. Remnant lipoprotein metabolism: key pathways
53. Ossoli, A., Simonelli, S., Vitali, C., Franceschini, G., and Calabresi, L. Role
involving cell‐surface heparan sulfate proteoglycans and apolipoprotein E.
of LCAT in atherosclerosis. J Atheroscler Thromb, 2016;23:119–27.
J Lipid Res, 1999;40:1–16.
 22:  Lipoprotein Metabolism and Cholesterol Balance 267
54. Jiang, X.C. Phospholipid transfer protein: its impact on lipoprotein homeo-
stasis and atherosclerosis. J Lipid Res, 2018;59:764–71.
55. Zhang, M., Zhai, X., Li, J. et al. Structural basis of the lipid transfer mecha-
nism of phospholipid transfer protein (PLTP). Biochim Biophys Acta,
2018;1863:1082–94.
56. Phillips, M.C. Molecular mechanisms of cellular cholesterol efflux. J Biol
Chem, 2014;289:24020–9.
57. Acton, S., Rigotti, A., Landschulz, K.T. et al. Identification of scavenger recep-
tor SR‐BI as a high density lipoprotein receptor. Science, 1996;271:518–20.
58. Shen, W.J., Hu, J., Hu, Z., Kraemer, F.B., and Azhar, S. Scavenger receptor
class B type I (SR‐BI): a versatile receptor with multiple functions and
actions. Metabolism, 2014;63:875–86.
59. Lund-Katz, S. and Phillips, M.C. High density lipoprotein structure-function
and role in reverse cholesterol transport. Subcell Biochem, 2010;51:183–227.
60. Villard, E.F., El Khoury, P., Duchene, E. et  al. Elevated CETP activity
improves plasma cholesterol efflux capacity from human macrophages in
women. Arterioscler Thromb Vasc Biol, 2012;32:2341–9.
61. Hogarth, C.A., Roy, A., and Ebert, D.L. Genomic evidence for the absence
of a functional cholesteryl ester transfer protein gene in mice and rats. Comp
Biochem Physiol B Biochem Mol Biol, 2003;135:219–29.
62. Chiang, J.Y. Bile acid metabolism and signaling. Compr Physiol, 2013;3:
1191–212.
63. Chiang, J.Y. Recent advances in understanding bile acid homeostasis.
F1000Res, 2017;6:2029.
64. Wang, D.Q., Cohen, D.E., and Carey, M.C. Biliary lipids and cholesterol
gallstone disease. J Lipid Res, 2009;509Suppl):S406–11.
65. Reshetnyak, V.I. Physiological and molecular biochemical mechanisms of
bile formation. World J Gastroenterol, 2013;19:7341–60.
66. Wang, H.H., Garruti, G., Liu, M., Portincasa, P., and Wang, D.Q. Cholesterol
and lipoprotein metabolism and atherosclerosis: Recent advances in reverse
cholesterol transport. Ann Hepatol, 2017;16:s27–s42.
67. Lammert, F., Gurusamy, K., Ko, C.W. et al. Gallstones. Nat Rev Dis Primers,
2016;2:16024.
68. Cohen, D.E. Of TICE in men. Cell Metab, 2016;24:773–4.
69. Jakulj, L., van Dijk, T.H., de Boer, J.F. et  al. Transintestinal cholesterol
­transport is active in mice and humans and controls ezetimibe‐induced fecal
neutral sterol excretion. Cell Metab, 2016;24:783–94.
70. Reeskamp, L.F., Meessen, E.C.E., and Groen, A.K. Transintestinal choles-
terol excretion in humans. Curr Opin Lipidol, 2018;29:10–17.
71. Carey, M.C. and Duane, W.C. Enterohepatic circulation, in The Liver:
Biology and Pathobiology (eds. I.M. Arias et al.), Raven Press, New York,
1994, pp. 719–67.
72. Wahlstrom, A., Sayin, S.I., Marschall, H.U., and Backhed, F. Intestinal
crosstalk between bile acids and microbiota and its impact on host metabo-
lism. Cell Metab, 2016;24:41–50.
73. Cohen, D.E. Hepatocellular transport and secretion of biliary phospholipids.
Semin Liver Dis, 1996;16:191–200.
74. Cohen, D.E. Balancing cholesterol synthesis and absorption in the gastroin-
testinal tract. J Clin Lipidol, 2008;2:S1–3.
75. Bosner, M.S., Lange, L.G., Stenson, W.F., Ostlund, R.E., Jr. Percent
­cholesterol absorption in normal women and men quantified with dual stable
isotopic tracers and negative ion mass spectrometry. J Lipid Res, 1999;40:
302–8.
76. Sehayek, E., Nath, C., Heinemann, T. et  al. U‐shape relationship between
change in dietary cholesterol absorption and plasma lipoprotein responsive-
ness and evidence for extreme interindividual variation in dietary cholesterol
absorption in humans. J Lipid Res, 1998;39:2415–22.
77. Expert Panel on Detection, Evaluation, and Treatment of High Blood
Cholesterol in Adults. Executive summary of the third report of the national
cholesterol education program (NCEP) expert panel on detection, evalua-
tion, and treatment of high blood cholesterol in adults (Adult Treatment
Panel III). JAMA, 2001;285:2486–97.
78. Grundy, S.M., Cleeman, J.I., Merz, C.N. et al. Implications of recent clinical
trials for the national cholesterol education program Adult Treatment
Panel III guidelines. J Am Coll Cardiol, 2004;44:720–32.
79. Young, S.G. and Fong, L.G. Lowering plasma cholesterol by raising LDL
receptors – revisited. N Engl J Med, 2012;366:1154–5.
80. Couture, P. and Lamarche, B. Ezetimibe and bile acid sequestrants: impact
on lipoprotein metabolism and beyond. Curr Opin Lipidol, 2013;24:
227–32.
81. Armstrong, M.J. and Carey, M.C. Thermodynamic and molecular determi-
nants of sterol solubilities in bile salt micelles. J Lipid Res, 1987;28:1144–55.
82. Taha, D.A., Wasan, E.K., Wasan, K.M., and Gershkovich, P. Lipid‐lowering
activity of natural and semi‐synthetic sterols and stanols. J Pharm Pharm
Sci, 2015;18:344–67.
83. Gylling, H., Plat, J., Turley, S. et al. Plant sterols and plant stanols in the
management of dyslipidaemia and prevention of cardiovascular disease.
Atherosclerosis, 2014;232:346–60.
84. Pirillo, A., Catapano, A.L., and Norata, G.D. Niemann‐Pick C1‐like 1
(NPC1L1) Inhibition and cardiovascular diseases. Curr Med Chem,
2016;23:983–99.
85. Rakhshandehroo, M., Knoch, B., Muller, M., and Kersten, S. Peroxisome
proliferator‐activated receptor alpha target genes. PPAR Res, 2010;2010.
86. Tyagi, S., Gupta, P., Saini, A.S., Kaushal, C., and Sharma, S. The peroxisome
­proliferator-activated receptor: A family of nuclear receptors role in various
diseases. J Adv Pharm Technol Res, 2011;2(4):236–40.
87. Shah, A., Rader, D.J., and Millar, J.S. The effect of PPAR‐alpha agonism on
apolipoprotein metabolism in humans. Atherosclerosis, 2010;210:35–40.
88. Botta, M., Audano, M., Sahebkar, A. et al. PPAR agonists and metabolic
syndrome: an established role? Int J Mol Sci, 2018;19.
89. Ferri, N., Corsini, A., Sirtori, C., and Ruscica, M. PPAR‐alpha agonists are
still on the rise: an update on clinical and experimental findings. Expert
Opin Investig Drugs, 2017;26:593–602.
90. Song, W.L. and FitzGerald, G.A. Niacin, an old drug with a new twist.
J Lipid Res, 2013;54:2586–94.
91. Del Gobbo, L.C., Imamura, F., Aslibekyan, S. et al. Omega‐3 polyunsatu-
rated fatty acid biomarkers and coronary heart disease: pooling project of
19 cohort studies. JAMA Intern Med, 2016;176:1155–66.
92. Ooi, E.M., Ng, T.W., Watts, G.F., and Barrett, P.H. Dietary fatty acids and
lipoprotein metabolism: new insights and updates. Curr Opin Lipidol,
2013;24:192–7.
93. Abifadel, M., Varret, M., Rabes, J.P. et al. Mutations in PCSK9 cause auto-
somal dominant hypercholesterolemia. Nat Genet, 2003;34:154–6.
94. Banach, M., Aronow, W.S., Serban, M.C. et al. Lipids, blood pressure and
kidney update 2015. Lipids Health Dis, 2015;14:167.
95. Moriarty, P.M., Jacobson, T.A., Bruckert, E. et al. Efficacy and safety of
alirocumab, a monoclonal antibody to PCSK9, in statin‐intolerant patients:
design and rationale of ODYSSEY ALTERNATIVE, a randomized phase 3
trial. J Clin Lipidol, 2014;8:554–61.
96. Karagiannis, A.D., Liu, M., Toth, P.P. et al. Pleiotropic anti‐atherosclerotic
effects of PCSK9 inhibitorsfrom molecular biology to clinical translation.
Curr Atheroscler Rep, 2018;20:20.
97. Ajufo, E. and Rader, D.J. New therapeutic approaches for familial hyper-
cholesterolemia. Annu Rev Med, 2018;69:113–31.
98. Saeed, A. and Ballantyne, C.M. Bempedoic acid (ETC‐1002): a current
review. Cardiol Clin, 2018;36:257–64.
99. Gaudet, D., Alexander, V.J., Baker, B.F. et al. Antisense inhibition of apoli-
poprotein C‐III in patients with hypertriglyceridemia. N Engl J Med,
2015;373:438–47.
100. Graham, M.J., Lee, R.G., Brandt, T.A. et al. Cardiovascular and metabolic
effects of ANGPTL3 antisense oligonucleotides. N Engl J Med, 2017;377:
222–32.
101. Scappa, M.C., Kanno, K., and Cohen, D.E. Lipoprotein metabolism, in
The Textbook of Hepatology: From Basic Science to Clinical Practice, 3rd
edn (eds. J. Rod’es, J.P. Benhamou, M.A. Rizzetto, J. Reichen and A. Blei),
Blackwell, Oxford, 2007.
102. Cohen, D.E. and Armstrong, E.J. Pharmacology of cholesterol and
lipoprotein metabolism, in Principles of Pharmacology: The Patho­
­
physiologic Basis of Drug Therapy, 2nd edn (eds. D.E. Golan, A.H.
Tashjian, E.J. Armstrong et  al.), Lippincott, Williams and Wilkins,
Philadelphia, 2007.
103. Jonas, A. Lipoprotein structure, in Biochemistry of Lipids, Lipoproteins
and  Membranes, 4th edn (eds. D.E. Vance and J.E. Vance), Elsevier,
Amsterdam, 2002, pp. 483–504.
SECTION C:
TRANSPORTERS,
BILE ACIDS, AND
CHOLESTASIS
 Bile Acid Metabolism
23 in Health and Disease:
An Update
Tiangang Li1 and John Y.L. Chiang2
1
Department of Pharmacology, Toxicology and Therapeutics, University of Kansas Medical Center, Kansas
City, KS, USA
2
Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA

INTRODUCTION chapter, we will summarize the basic knowledge of bile acid

One of the major functions of the liver is to generate bile. Bile
consists of ~95% water plus organic solutes, inorganic electro-
lytes, and proteins. Bile acids are synthesized from cholesterol
in the hepatocytes and are the major organic solutes in bile. Bile
acid synthesis accounts for a major fraction of hepatic choles-
terol turnover, and hepatic secretion of bile acids is a major driv-
ing force to generate bile flow. Following conjugation with
glycine or taurine, bile acids are efficiently reabsorbed in small
intestine and circulated back to the liver in a process called the
enterohepatic circulation of bile acids. Many physiological
functions of bile acids depend on their physiochemical proper-
ties as amphipathic detergent molecules, while other functions
are related to their signaling properties. In bile, bile acids pre-
vent cholesterol precipitation by forming mixed micelles with
phospholipids. In the small intestine, bile acids facilitate the
digestion of dietary fat by the pancreatic lipase and colipase,
which forms mixed micelles with fatty acids and monoacylg-
lycerols. These compounds are absorbed into enterocytes,
re‐esterified to form triglycerides and then assembled into
­chylomicrons that are transported via the lymphatic system and
delivered into circulation. In liver and intestine, bile acids serve
as endogenous ligands for nuclear receptors and cell surface
G  protein‐coupled bile acid receptors that regulate metabolic
homeostasis and immune responses.
Disrupted bile flow due to biliary obstruction, genetic defects
of bile acid transporters, or acquired autoimmune destruction of
bile ducts causes cholestasis in humans. Pharmacological
approaches targeting bile acid metabolism, transport, and sign-
aling have been developed to treat hepatobiliary diseases. In this
chemistry and biology, new mechanistic understanding of the
regulation of bile acid homeostasis, human diseases related to
altered bile acid metabolism, and the recent advances in bile
acid‐based therapies.
BILE ACID SYNTHESIS, FUNCTION,
AND REGULATION
Bile acid biosynthesis
Conversion of cholesterol to bile acids is a multistep reaction
that involves enzymes located in the endoplasmic reticulum,
mitochondria, cytosol, and peroxisomes of hepatocytes [1]. As
shown in Figure  23.1a, the conversion of cholesterol to bile
acids (cholic acid as an example) involves hydroxylation,
­oxidation of the 3‐hydroxy group with isomerization of the
­double bond from C5 to C4, stereo‐specific reduction of the
double bond at C4, and the oxidative cleavage of the side‐chain
to convert the isooctane side‐chain of cholesterol to an isopen-
tanoic acid of C24 bile acid molecules. In mammals, most bile
acids have a 5β‐hydrogen group and a cis‐configuration along
the plane of the steroid nucleus. The space‐filling model
(Figure 23.1a, right) of cholesterol shows that the carbon skel-
eton of cholesterol (black) is in a planar structure with one
hydroxy group at 3β‐position and a double bond at C5–C6 posi-
tion. In cholic acid, the A and B rings are kinked along the
C5–C10 bond and all three α‐hydroxyl groups and a ‐COOH
group are positioned on one side of the steroid nucleus

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
272 THE LIVER:  BILE ACID SYNTHESIS, FUNCTION, AND REGULATION

(a) 21
22

20 23
18 26
12 24
17 25
19 16 27
11 13 D
C
15
2 1 9 14
A
10
B 8 Cholesterol
3β HO 3 5
7
OH Hydrophobic face
4 6

12α OH

COOH
Cholic acid

3α HO OH 7α
OH COOH
Hydrophilic face

(b)
Major bile acid species in humans and mice.

Bile acids C-3 C-7 C-12 C-6

Chenodeoxycholic acid α-OH α-OH H H
Cholic acid α-OH α-OH α-OH H
Deoxycholic acid α-OH H α-OH H
Lithocholic acid α-OH H H H
α-muricholic acid α-OH α-OH H β-OH
β-muricholic acid α-OH β-OH H β-OH
ω-muricholic acid α-OH β-OH H α-OH
Ursodeoxycholic acid α-OH β-OH H H

Figure 23.1  Bile acid structure. (a) Structures of cholesterol and cholic acid are illustrated as examples. Bile acids are amphipathic molecules.
The 3α‐, 7α‐, and 12α‐HO groups of cholic acid face to one side of the steroid rings to form a hydrophilic face and the carbon skeleton forms a
hydrophobic face on the other side. Saturation of the C5–6 double bond creates a kink at A/B rings. (b) Major bile acids in humans and rodents have
hydroxy groups specifically located at C‐3, C‐7, C‐12, and C‐6 positions.

opposing the other hydrophobic side, which renders bile acids
amphipathic molecules with physiological detergent proper-
ties. Figure 23.1b shows major bile acids in humans and rodents
with one‐ to three‐hydroxy groups located stereo‐specifically at
C3, C7, C12, and C6 positions. Chenodeoxycholic acid (CDCA,
3α,7α‐dihydroxy bile acid) and cholic acid (CA, 3α,7α,12α‐tri-
hydroxy bile acid) are major primary bile acids in humans,
while cholic acid and α‐muricholic acid (α‐MCA, 3α,6β,7α‐tri-
hydroxy bile  acid) and β‐muricholic acid (β‐MCA, 3α,6β,7β‐
trihydroxy bile acid)  are primary bile acids in rodents. The
7α‐hydroxy group in CA and CDCA are removed by gut bacte-
rial 7α‐dehydroxylase to form secondary bile acids, deoxy-
cholic acid (DCA, 3α,12α‐dihydroxy bile acid) and lithocholic
acids (LCA, 3α‐monohydroxy bile acid) respectively. The gut
microbes also convert α‐ and β‐muricholic acids to ω‐muricholic
acid (ω‐MCA, 3α,6α,7β‐trihydroxy bile acid) by epimerization
for fecal excretion. Ursodeoxycholic acid (UDCA, 3α,7β‐
dihydroxy bile acid) is a minor primary in rodents and secondary
bile acid in humans formed by isomerization of the 7α‐hydroxy
group in CDCA to 7β‐position. Conjugated bile acids are ion-
ized at physiological pH. After synthesis, the side‐chain of bile
acids is efficiently linked to the amino acids glycine (G) or tau-
rine (T) to form N‐acyl‐amidates, a process that is referred to as
“conjugation” or “amidation.” The side‐chain of the glycine
conjugates and taurine conjugates have a pKa value of ~4.0 and
~2.0, respectively. Conjugated bile acids are soluble, mem-
brane‐impermeable, and resistant to precipitation at physiolog-
ical pH. The amide bond is resistant to cleavage by the
pancreatic carboxypeptidases, and conjugated bile acids are
highly stable in the intestine lumen.
In humans, the bile acid pool consists of two primary bile
acids, CA and CDCA, and secondary bile acids including DCA
and a very low concentration of LCA. Primary bile acids are the
direct end‐products of bile acid biosynthesis in hepatocytes,
while secondary bile acids are formed from primary bile acids
in the small and large intestine by bacterial enzymes. The human
bile acid pool contains about 80% CDCA and CA, about 20%
DCA, trace amounts of LCA and UDCA, and other secondary
bile acids in negligible amounts. In humans, CDCA and CA
generally present in roughly equal amounts but this varies by
individual. TCA and Tα‐MCA and Tβ‐MCA are the predomi-
nant bile acids in mice.
Bile acid biosynthesis in hepatocytes
In hepatocytes, two major bile acid biosynthesis pathways,
namely the classic pathway and the alternative pathway, synthe-
size primary bile acids (Figure 23.2a). Under normal physiology,
 23:  BILE ACID METABOLISM IN HEALTH AND DISEASE 273

bile acid synthesis is more active in the pericentral hepatocytes
and low in the periportal hepatocytes. This is because higher
sinusoidal bile acid uptake into the periportal hepatocytes inhib-
its bile acid synthesis. The relative contributions of the classic
pathway and the alternative pathway to hepatic bile acid produc-
tion vary in different species. In humans, the classic bile acid
synthesis pathway is the predominant pathway and accounts for
about 90% of total hepatic bile acid production, while the alter-
native pathway accounts for less than 10% of total hepatic bile
acid production. In rodents, the classic and alternative pathways
contribute about equally to the synthesis of primary bile acids.
The classic bile acid synthesis pathway is initiated by the rate‐
limiting enzyme cholesterol 7α‐hydroxylase (CYP7A1), a
cytochrome p450 enzyme located in the endoplasmic reticulum.
CYP7A1 catalyzes the stereo‐specific hydroxylation at the
C‐7α ­position of cholesterol to produce 7α‐hydroxycholesterol
[2]. 3β‐Hydroxy‐Δ5‐C27‐steroid dehydrogenase (HSD3B7) con-
verts 7α‐hydroxycholesterol to 7α‐hydroxy‐4‐cholesten‐3‐one
(C4), which has been used as a surrogate serum marker for the
rate of hepatic bile acid synthesis in humans [3, 4]. If the C‐12
position on C4 is left unmodified, CDCA will be produced.
Alternatively, microsomal sterol 12α‐hydroxylase (CYP8B1)
hydroxylates the C‐12 position on C4, leading to the production
of CA. Therefore, C4 is a common precursor of CA and CDCA,
and CYP8B1 activity is one of the factors that affect the
CA:CDCA ratio in the bile acid pool. In this series of enzymatic
reactions catalyzed by Δ4–3‐oxosteroid‐5β‐reductase (aldo‐keto
reductase, AKR1D1) and 3α‐hydroxysteroid dehydrogenase
(AKR1C4), the 3β‐hydroxyl group of the steroid nucleus is
converted to a 3α‐hydroxyl group, the Δ5,6 double bond on
­
the steroid nucleus is saturated, and the A ring and B ring of the
steroid nucleus assume a cis‐configuration. After steroid ring
modifications, the mitochondrial enzyme sterol 27‐hydroxylase
(CYP27A1) catalyzes the steroid side‐chain hydroxylation and
oxidation of C27 (or C26) of the sterol intermediates to produce
3α,7α,12α‐trihydroxycholestanoic acid (THCA), leading to cholic
acid synthesis, and 3α,7α‐dihydroxycholestanic acid (DHCA),
leading to CDCA synthesis. Peroxisomal very‐long‐chain acyl‐
CoA synthase (VLACS, SLC27A5) then converts THCA and
DHCA to acyl‐CoA thioesters, which are transported into
­peroxisomes via bile acid‐acyl transporter ABCD3 [5]. In the
peroxisomes, α‐methylacyl‐CoA racemase (AMACR), enoyl‐
CoA hydratase/3‐hydroxy acyl‐CoA dehydrogenase (EHHADH,
bifunctional enzyme) and acyl‐CoA oxidase 2 (ACOX2) catalyze a
series of racemizations, hydrations, and dehydrogenations (β‐
oxidation reaction), and then thiolase/sterol carrier protein X
(SCPx) cleaves a propionyl‐CoA from THCA and DHCA to pro-
duce C24‐CoA thioesters, cholyl‐CoA, and chenodeoxycholyl‐
CA, respectively. Cholyl‐CoA and chenodeoxycholyl‐CoA
are subsequently conjugated to taurine or glycine by bile acid-
CoA:amino acid N‐acyltransferase (BAAT) [6, 7]. Unconjugated
bile acids returning to hepatocytes via p­ ortal circulation can be
converted to their CoA thioesters by  microsomal bile acid-
CoA synthase (BACS) and then r­econjugated to taurine or
­glycine by BAAT.
The alternative pathway is initiated by mitochondrial
CYP27A1, which hydroxylates and oxidizes the cholesterol
side‐chain to produce 27‐hydroxycholesterol and then 3β‐
hydroxy‐5‐cholestenoic acid prior to enzymatic modifications
of the steroid nucleus. In the alternative pathway, hydroxylation
of the C‐7 position is catalyzed by the oxysterol 7α‐hydroxylase
(CYP7B1). CYP7B1 is expressed in liver for bile acid synthesis,
steroidogenic tissues for steroid synthesis, and macrophages
for oxysterol synthesis. In the brain, sterol 24‐hydroxylase
(CYP46A1) converts cholesterol to 24‐hydroxycholesterol,
which is hydroxylated at the 7α‐position by liver sterol 7α‐
hydroxylase (CYP39A1). In the liver, sterol 25‐hydroxylase
(non‐CYP enzyme) can convert cholesterol to 25‐hydroxycho-
lesterol, a major oxysterol in circulation. Oxysterols generated
in extrahepatic tissues can be transported to the liver to be con-
verted to bile acids. The enzyme cascade involved in bile acid
synthesis from oxysterols has not been identified.
The classic versus alternative bile acid synthesis pathways
The classic pathway is also called “the neutral pathway” because
the steroid nucleus modifications occur before the side‐chain
oxidation and thus most of the intermediates in this pathway do
not have a carboxylic acid group. In contrast, the alternative
pathway is referred to as “the acidic pathway” because bile acid
synthesis in this pathway is initiated with cholesterol side‐chain
oxidation to a C27‐carboxylic acid group. It is believed that
hepatocytes are the only cells that express the entire set of
enzymes required for de novo bile acid synthesis. The classic
bile acid synthesis pathway is tightly regulated, whereas the
alternative pathway is constitutively active and is not regulated
by bile acid feedback. In adult liver, the classic bile acid synthe-
sis pathway is the predominant pathway for bile acid synthesis,
whereas in the newborn infant the alternative pathway produces
bile acids before CYP7A1 is expressed at weaning. There is a
misconception that the alternative pathway only produces
CDCA. In mice deficient of Cyp7a1, Cyp8b1 expression is
increased and Cyp7b1 in the alternative pathway is doubled to
produce a smaller bile acid pool, but gallbladder bile still con-
tains a substantial amount of TCA (~30% compared to 45%
TCA in wild‐type mice) [8]. Mutations of the CYP7A1 gene in
a human patient only caused mild hypercholesterolemia and
premature gallstone disease, indicating that the alternative bile
acid synthesis pathway is activated to produce bile acids [9]. It
has been reported that in inborn errors of bile acid metabolism,
C27‐acidic bile acid intermediates are accumulated, indicating
the alternative bile acid synthesis pathway may be stimulated.
Therefore, bile acid synthesis can be switched from the classic
pathway to the alternative pathways to synthesize both CA and
CDCA. After synthesis, bile acids are conjugated to the amino
acids glycine or taurine at C24‐COOH group immediately in
hepatocytes to form amidated bile acids (Figure 23.2b).
Species differences in bile acid synthesis
Bile acid synthesis pathways are remarkably similar between
humans and other mammals. In mice and rats, the majority of
CDCA is converted to α‐MCA and β‐MCA (Figure 23.1b). Both
α‐MCA and β‐MCA have a hydroxyl group at the 6β‐position,
which increases solubility. A recent study reported that Cyp2c‐
cluster null mice lacked both α‐MCA and β‐MCA and had high
CDCA and UDCA content [10]. Further analysis showed that
recombinant Cyp2c70 produced α‐MCA and β‐MCA using
CDCA and UDCA, respectively, as a substrate (Figure 23.2c).
 (a) (b)
Alternative pathway HO

Sterol 24-hydroxylase 24
12
(CYP46A1) (Brain) 12 C-NH-CH2-COO–
24
Classic pathway
O
3 7 Sterol 25-hydroxylase
HO 6
3 7
Cholesterol OH
Cholesterol 7α-hydroxylase Sterol 27-hydroxylase HO
6
(CYP7A1) (CYP27A1)
COOH Sterol 7α-hydroxylase Glyco-cholic acid
12 12
24 24 (CYP39A1)

Liver
3 7 3 7
Macrophage 24
HO OH HO 6
6
C-NH-CH2-CH2-SO3–
7α-hydroxycholesterol 3β-hydroxy-5-cholestanoic acid
Oxysterol 7α-hydroxylase O
3β-Hydroxy Δ5-C27 steroid
dehydrogenase (CYP7B1)
COOH
(HSD3B7) 12 12 Oxysterols 3 7
24 24 OH
HO
6
3 7
OH
3 7 Tauro-chenodeoxycholic acid
O 6 HO OH
6
7α-hydroxy-4-cholesten-3-one (C4) (c)
Sterol 12α-hydroxylase ? Hyocholic acid T/G-CA
T/G-CDCA
(CYP8B1) AKR1D1 (3α, 6α, 7α)
Aldo-keto reductases AKR1C4
Mitochondria
(AKR1D1, AKR1C4) 6α-hydroxylase Bile salt hydrolase
Sterol 27-hydroxylase (BSH)
(CYP27A1) α-MCA 6β-hydroxylase CDCA CA
3α, 7α, 12α-trihydroxycholestanoic acid 3α, 7α-dihydroxycholestanoic acid (3α, 6β, 7α)
(3α, 7α) (3α, 7α, 12α)
(THCA) (DHCA) (Cyp2c70)
Peroxisome Very long-chain acyl-CoA synthase CoA
(VLCAS), ABCD3, acyl-CoA oxidase, 7α/β-hydroxysteroid
CoA dehydrogenase/ 7α-dehydroxylase
Epimerase
bifunctional enzyme, thiolase/SCP2 Epimerase
HO Propionyl-CoA Cyp2c70
24 7β-hydroxylase DCA
24 β-MCA UDCA LCA
COOH (CoA) COOH (CoA) (3α, 12α)
12 (3α, 6β, 7β) (3α, 7β) (3α)
7β-dehydroxylase
3 7
3 7 Epimerase
HO OH 6α-hydroxylase 6β-hydroxylase
HO OH 6
6
Cholic acid (CA)(-CoA) Chenodeoxycholic acid (CDCA)(-CoA)
ω-MCA
Bile acid-CoA synthase (BACS) Hyodeoxycholic
(3α, 6α, 7β) Murideoxycholic
Bile acid-CoA amino N-acyltransferase acid (3α, 6α) acid (3α, 6β)
(BAAT)
T/G-CA T/G-CDCA

Figure 23.2  Bile acid synthesis pathways. (a) The classic and alternative bile acid synthesis pathways are shown. Enzymes involved in bile acid synthesis are described in the text. (b) Structure
of glycine‐ and taurine‐conjugated bile acids. Glycocholic acid and taurochenodeoxycholic acid are shown as examples. (c) Bile acid biotransformation by gut bacteria. Bacterial enzymes involved
in hydroxylation and dihydroxylation produce various secondary bile acids. Species differences in bile acid transformation are described in the text.
 23:  BILE ACID METABOLISM IN HEALTH AND DISEASE 275
CDCA is isomerized to UDCA in liver and gut bacteria, which
may be the predominant substrate for Cyp2c70 to synthesize
β‐MCA in rodent liver.
Another aspect of the species‐dependent difference in bile
acid pool composition is the preferential use of glycine (G) or
taurine (T) for bile acid conjugation. The bile acid pool in many
mammalian species such as mice, rats, and dogs contains pre-
dominantly taurine conjugates, while the bile acid pool in
humans, hamsters, and rabbits contains about two‐thirds or
more glycine conjugates, with the rest being taurine conjugates
[11, 12]. Taurine‐conjugated bile acids are less toxic than their
corresponding glycine‐conjugated and unconjugated bile acids.
From an evolutionary point of view, the use of glycine for bile
acid conjugation appears to be a more recent event. The spe-
cies‐dependent preference for using glycine or taurine for bile
acid conjugation is thought to be determined by BAAT substrate
specificity and possibly the availability of taurine and glycine in
the peroxisomes [6, 13, 14].
Bacterial modification of bile acids: synthesis
of secondary bile acids
When bile is released into the small intestine, a portion of bile
acids can further undergo complex modifications by bacterial
enzymes in the ileum and large intestine [15]. Bacterial bile salt
hydrolases (BSH) deconjugate T/G‐conjugated bile acids to
form unconjugated bile acids (Figure 23.2c). In the colon, bac-
terial 7α‐dehydroxylases convert CA to DCA, and CDCA to
LCA [15–17]. DCA is the predominant bile acid in the colon.
Unconjugated primary and secondary bile acids are hydropho-
bic and are passively absorbed in the ileum and large intestine.
DCA is delivered to the liver where it is reconjugated and joins
the circulating bile acids [18]. Bacterial 3α,7α,12α‐hydroxyster-
oid dehydrogenases (HSDHs) epimerize the α‐hydroxyl groups
of bile acids to carbonyl groups to form 3‐oxo, 7‐oxo‐, and 12‐
oxo‐bile acids, respectively. Then the carbonyl groups in oxo‐
bile acids are converted to β epimers: iso‐bile acids and epi‐bile
acids by 3β‐, 7β‐, and 12β‐HSDHs. The 7α‐hydroxy‐bile acids
have higher bactericidal activity than oxo‐bile acids and β‐
epimers. In humans and mice, bacterial 7β‐HSDH epimerizes
the 7α‐hydroxyl group of CDCA (3α,7α) to 7β, forming UDCA.
The 7β‐epimerization converts highly hydrophobic CDCA to
hydrophilic and nontoxic UDCA.
LCA is the most hydrophobic and toxic bile acid and is pre-
sent in trace amounts in the human bile acid pool. In human
hepatocytes, LCA is mainly detoxified by sulfonation at the C‐3
position by a family of sulfotransferases (SULT2A1, SULT2B8,
etc.). Sulfated LCA is poorly reabsorbed in the small intestine
and is excreted into feces. In mice, LCA is sulfated at the C‐7
position but sulfonation is not a major detoxification pathway.
Mouse gut bacteria detoxify bile acids mainly by hydroxylation
of mono‐ and di‐hydroxyl bile acids to polyhydroxylated bile
acids for reabsorption or secretion.
In pig, CDCA is 6α‐hydroxylated to hyocholic acid
(3α,6α,7α) and LCA is converted to the hyodeoxycholic
acid (3α,6α). Only rodents can convert LCA back to UDCA by
7β‐hydroxylase and to murideoxycholic acid (3α,6β) by bacte-
rial 6α‐ and 6β‐hydroxylases (Figure 23.2c). Epimerization of
β‐MCA to ω‐MCA increased the solubility of MCA, leading to
fecal excretion or circulating in portal blood. The fecal bile
acids consist of predominantly unconjugated bile acids. Some
secondary bile acids are reabsorbed in the intestine and trans-
ported to the liver and bile by portal circulation. Secondary bile
acids circulated to hepatocytes can be detoxified by sulfonation
for renal secretion, reconjugated to glycine or taurine, or glu-
curonidated for secretion into bile.
ENTEROHEPATIC CIRCULATION
OF BILE ACIDS
Overview of bile acid circulation
and homeostasis
Bile acids are actively transported across the hepatocyte cana-
licular membrane, which plays an important role in promoting
water and other biliary lipid secretion and provides a major
osmotic force for bile salt‐dependent bile flow. Under normal
conditions, bile flows freely in intrahepatic bile ducts due to the
relatively low pressure in the biliary tree. Under fasting condi-
tions, little bile enters the duodenum due to the high tone of the
sphincter of Oddi, which therefore promotes bile entrance into the
gallbladder via the cystic duct. In the gallbladder, water and elec-
trolytes are reabsorbed and bile is concentrated ~5‐ to 10‐fold.
After food intake, dietary fat and amino acids stimulate the neu-
roendocrine cells of the duodenum to secrete a peptide hormone
called cholecystokinin (CCK), which stimulates gallbladder con-
traction and sphincter of Oddi relaxation to release concentrated
bile into the duodenum [19]. CCK also stimulates pancreatic aci-
nar cells to simultaneously secrete digestive juice into duodenum.
In the small intestine lumen, bile acids form mixed micelles with
dietary lipids and cholesterol to facilitate digestion by pancreatic
lipases and colipases. Bile acids are efficiently reabsorbed in the
ileum and transported back to the liver via portal circulation. In
humans, the bile acid pool circulates a few times a day between
the liver and the intestine (Figure 23.3). Because bile acids are
reabsorbed in the small intestine at ~95% efficiency, the bile acid
pool is largely conserved. In humans, the bile acid pool contains
about 2–4 g of bile acids. Roughly 0.2–0.6 g of bile acids is syn-
thesized daily by the hepatocytes to replace the bile acids that
are lost in the feces. The enterohepatic circula­tion of bile
acids is driven by a network of bile acid efflux and uptake trans-
porters expressed in hepatocytes, cholangiocytes, and enterocytes
briefly described here (Figure 23.3).
Biliary secretion of bile acids
Hepatocytes are polarized cells. The apical membrane of two
adjacent hepatocytes form a bile canaliculi surrounded by tight
junctions. A network of bile canalicular collects bile generated
by hepatocytes and drains bile into small bile ducts that are
formed by biliary epithelial cholangiocytes at the portal triad.
Cholangiocytes can further modify bile in the biliary tree by
regulated secretion (e.g. water, HCO3−) and reabsorption (e.g. bile
acids, glucose, and other solutes). Bile acids, phospholipids, and
276 THE LIVER:  ENTEROHEPATIC CIRCULATION OF BILE ACIDS
ABCG5/G8
Cholesterol
CYP7A1
Bile
BSEP
Bile acids
OSTα/β
MRP3/4
NTCP OATPs
Systemic
circulation
Portal
blood
Hepatocyte
Cholangiocyte
Enterocyte
Figure 23.3  Enterohepatic circulation of bile acids. The major transporters involved in bile formation and enterohepatic circulation of bile acids
are illustrated in hepatocytes, cholangiocytes, and enterocytes.
cholesterol are the major organic solutes in bile. Once secreted
by hepatocytes into bile, they form mixed micelles to increase
cholesterol solubility. Genetic defects in either biliary bile acid
secretion or phospholipid secretion are known to result in
cholesterol monohydrate precipitation and gallstone disease
in pediatric patients. In addition, mixed micelle formation
decreases the biliary concentration of monohydroxy bile acids,
which may damage the biliary tract upon chronic exposure at
high concentrations. Canalicular bile acid secretion into the bile
against the concentration gradient is the rate‐limiting step in bile
formation [20]. The ATP‐binding cassette (ABC) transporter
bile salt export pump (BSEP, ABCB11) is the major canalicular
bile acid efflux transporter in hepatocytes [21]. Hydrolysis of
ATP provides the driving force for active transport of bile acids
against a high bile acid concentration gradient in bile. BSEP
expression is hepatocyte‐specific, and BSEP has high substrate
specificity for conjugated bile acids. Multidrug resistance‐asso-
ciated protein 2 (MRP2, ABCC2) mediates the secretion of
bilirubin glucuronide, glutathione, and some sulfated and conju-
gated bile acids into bile. The ABCG5 and ABCG8 transporters
form a heterodimer to mediate free cholesterol secretion into
bile [22]. Genetic variations of ABCG5 and ABCG8 are associ-
ated with gallstone disease in humans because biliary choles-
terol hypersecretion is the major risk factor for gallstone
formation. ABCG5 and ABCG8 are also highly expressed in the
MDR3
Phospholipid
Bile
MRP2
Bilirubin glucuronide
Sulfated bile acids
Cholehepatic shunt
ASBT
OSTα/β
Bacteria-mediated
Bile acids biotransformation
of bile acids in gut
OSTα/β ASBT
small intestine where the heterodimer mediates the secretion of
cholesterol and other plant sterols back into the lumen. This pro-
cess is significant in that it prevents accumulation of toxic die-
tary sitosterols in the blood and tissues, and it also limits
intestinal cholesterol absorption efficiency at about 50%.
Multidrug‐resistant 3 (MDR3, ABCB4) mediates the canalicu-
lar secretion of phosphatidylcholine, which is the major phos-
pholipid in bile [23]. Detailed description of bile acid
transporters can be found in Chapters 26 and 27.
Intestinal reabsorption of bile acids
Intestinal bile acids are reabsorbed via passive diffusion of
unconjugated bile acids and active transport of conjugated bile
acids, with the latter being the quantitatively major mechanism
(Figure 23.3). Bile acid reabsorption mainly occurs at the termi-
nal ileum where apical sodium‐dependent bile salt transporter
(ASBT, SLC10A2) is highly expressed [24]. Humans and mice
lacking functional ASBT present with bile acid malabsorption
[25, 26]. Once absorbed into enterocytes, bile acids bind to the
bile acid‐binding protein (I‐BABP), which transports bile acids
to the basolateral membrane for secretion by the organic solute
transporters OSTα (SLC51A) and OSTβ (SLC51B) [27]. OSTα
and OSTβ form a functional heterodimer to efflux bile acids
across the basolateral membrane into the portal circulation [28].
 23:  BILE ACID METABOLISM IN HEALTH AND DISEASE 277
Mice lacking a functional OSTα/OSTβ heterodimer had signifi-
cantly decreased intestinal bile acid absorption, lower plasma
bile acid concentration, and a small bile acid pool [29]. ASBT
and OSTα/β are also expressed at the apical side and the baso-
lateral side, respectively, of cholangiocytes and renal proximal
tubular cells [28]. In the biliary tract, ASBT and OSTα/β func-
tion sequentially to transport a small flux of bile acids across the
biliary epithelium into the peribiliary plexus. From here, bile
acids are further transported to hepatic sinusoidal cells and are
reabsorbed by hepatocytes.
If unconjugated dihydroxy bile acids are secreted into the bil-
iary ductules, they may be absorbed passively. This process is
called cholehepatic shunting, which generates a hypercholeresis
effect by recycling biliary bile acids for hepatic apical resecre-
tion [30]. In the kidney proximal tubule cells, ASBT and OSTα/β
coordinate the efflux of bile acids from kidney back to the
systemic circulation. Inhibition of ASBT is expected to increase
urinary bile acid excretion [31].
Hepatic uptake of bile acids
Hepatic basolateral uptake of bile acids is more active in the peri-
portal hepatocytes rather than the pericentral hepatocytes. The
first‐pass extraction rate is about 60–90%, depending on the bile
acid with little spillover of portal bile acids into the systemic cir-
culation [32]. Na+‐taurocholate co‐transporting polypeptide
(NTCP, SLC10A1) is the major basolateral conjugated
bile acid uptake transporter in hepatocytes [33–35]. Coupling bile
acid transport to the Na+ gradient provides energy for active trans-
port of serum bile acids across the sinusoidal membrane to hepat-
ocytes. The human isoforms of organic anion transporters
(OATP), OATP1A2, OATP1B1, and OATP1B3, mediate Na+‐
independent basolateral transport of conjugated and unconju-
gated bile acids and other organic anions, such as steroids and
drugs [18, 36]. So far there has been only one reported case of
human NTCP loss‐of‐function mutation that resulted in a hyper-
cholanemic phenotype with significantly elevated circulating bile
acids of ~1500 μM (normally less than 10 μM) without clinical
signs of cholestasis [37]. Experiments conducted in mice showed
that the plasma clearance of taurocholate after intravenous admin-
istration was delayed but not abolished in mice lacking NTCP
[38]. Only a small subset of the Ntcp knockout mice showed
marked elevation of plasma bile acid concentration, while the
majority of the Ntcp knockout mice had normal plasma bile acid
concentration [38]. A more recent study showed that pharmaco-
logical inhibition of NTCP in mice lacking OATP1A/1B achieved
close to complete inhibition of plasma taurocholate clearance and
markedly elevated plasma bile acids [39]. These findings imply
that in mice both NTCP and OATPs mediate quantitative basolat-
eral bile acid uptake into hepatocytes. In humans, it is possible
that NTCP is the predominant basolateral bile acid uptake trans-
porter and its loss‐of‐function cannot be compensated to a
­significant degree by the presence of OATPs.
Plasma bile acids
The presence of bile acids in the systemic circulation is mainly
due to incomplete extraction of portal bile acids by the liver.
Under normal physiological conditions, bile acid concentration
in the systemic blood is very low. It is uncertain if such low
concentration of bile acids in the systemic blood can elicit sig-
nificant physiological functions in peripheral organs that express
bile acid receptors. In addition, bile acid composition in the
systemic blood may differ significantly from that of bile due to
the differential extraction rate of various bile acid species via
active transport or passive diffusion along the sinusoid. In chol-
estasis, basolateral efflux of bile acids is increased and causes
significantly elevated bile acid concentration in the systemic
circulation [28, 40, 41]. Several transporters including OSTα/β,
MRP3, and MRP4 mediate basolateral efflux of bile acids from
hepatocytes [18, 28, 42]. Plasma bile acids increased in choles-
tasis are sulfated and eliminated mainly via renal excretion.
Species differences in serum and bile acid concentration and
composition have been reported [43].
REGULATION OF BILE ACID SYNTHESIS,
TRANSPORT, AND HOMEOSTASIS
Bile acid pool size is usually kept at a relatively constant level
because daily bile acid synthesis roughly equals the daily
amount of bile acids excreted in feces. Such a balance is main-
tained by bile acid sensing mechanisms that regulate hepatic
bile acid synthesis and intestinal bile acid reabsorption.
When  bile acid concentration increases in the enterohepatic
system, bile acids inhibit the transcription of genes in bile acid
synthesis and intestinal bile acid transport, leading to decreased
hepatic output and increased fecal loss. When bile acid con-
centration decreases, these two pathways are upregulated. Bile
acid‐activated farnesoid X receptor (FXR) plays a key role in
mediating bile acid effects in the enterohepatic system. FXR
belongs to the nuclear receptor superfamily, which consists of
a group of ligand‐activated transcription factors [44]. Upon
ligand binding, FXR binds to its target gene promoters and
activates gene transcription. Both conjugated and unconju-
gated CDCA and CA are endogenous ligands for FXR, but CA
(EC50 = 586 μM) is a much weaker FXR agonist than CDCA
(EC50 = 17 μM). The hydrophilic bile acids T‐UDCA and T‐
MCAs do not activate FXR but are FXR antagonists [45]. FXR
is expressed in hepatocytes and enterocytes that are routinely
exposed to high concentrations of bile acids. The conjugated
and nonconjugated secondary bile acids LCA and DCA acti-
vate membrane G protein‐coupled bile acid receptor 1
(Gpbar‐1) [46], aka Takeda G receptor 5 (TGR5) [47]. The
roles of FXR and TGR5 in liver physiology are reviewed in
Chapters 24 and 25. The following sections will only focus on
the role of bile acid receptors in the regulation of bile acid
synthesis and transport.
Regulation of bile acid synthesis
The early evidence supporting the presence of bile acid feed-
back inhibition of bile acid synthesis came from experimental
observations that hepatic CYP7A1 enzyme activity markedly
decreased when rats were fed a diet containing bile acids, and
that hepatic CYP7A1 enzyme activity was induced when intestine
bile acid absorption was disrupted by bile acid sequestrants.
278 THE LIVER:  REGULATION OF BILE ACID SYNTHESIS, TRANSPORT, AND HOMEOSTASIS

Hepatic CYP7A1 function is mainly controlled at the transcrip-
tional level and evidence suggesting direct regulation of
CYP7A1 protein stability or enzyme activity is scarce. The
CYP7A1 gene promoter has been well characterized. It contains
one or two bile acid response elements (BARE‐1 and BARE‐2).
Human, mouse, and rat CYP7A1 gene proximal promoters con-
tain a BARE‐1 that binds two nuclear receptors, hepatocyte
nuclear factor 4α (HNF4α) and liver‐related homolog 1
(LRH‐1). The putative endogenous ligands for HNF4α and
LRH‐1 are phospholipids, and they are believed to be constitu-
tively active in hepatocytes. These two nuclear receptors are
largely responsible for the basal expression of hepatic CYP7A1,
because mutagenesis that abolishes their binding results in sig-
nificant reduction of CYP7A1 promoter activity. In addition,
bile acids mainly inhibit bile acid synthesis by repressing
HNF4α and LRH‐1 transactivation of CYP7A1 via BARE‐2.
Two major mechanisms are believed to mediate bile acid
inhibition of CYP7A1 (and also CYP8B1) (Figure 23.4). One is
mediated by FXR in hepatocytes and the other is initiated by
activation of FXR in the small intestine. When bile acid concen-
tration increases within hepatocytes, such as in cholestatic liver
injury, bile acids may activate FXR to induce nuclear receptor
small heterodimer partner (SHP). SHP is an atypical nuclear
receptor without DNA binding activity and often acts as a
co‐repressor to inhibit other transcriptional factors via protein–
protein interactions. In this case, SHP inhibits the trans‐
activating activity of HNF4α and LRH‐1, leading to the
inhibition of CYP7A1 gene transcription [48, 49]. However,
intrahepatic TCA and TCDCA concentrations may not be high
enough to activate FXR under physiological conditions in mice
and humans. It was reported in 1995 that intraduodenal infu-
sion, but not intravenous infusion, of TCA inhibited CYP7A1
expression in rats with bile fistula, suggesting that intestinal
factor(s) might be required to mediate bile acid feedback inhibi-
tion of bile acid synthesis [50]. The intestine is exposed to high
concentrations of bile acids. Not only is trans‐enterocyte bile
acid flux highly regulated, the intestine stores a significantly
larger portion of the bile acid pool (~70–80%) compared to the
gallbladder and the liver [51]. Bile acid activation of intestine
FXR induces the transcription of fibroblast growth factor 15
(FGF15), which is subsequently released into the portal circula-
tion and acts as an endocrine signaling molecule to inhibit

Cholesterol

HNF4α
ERK CYP7A1
LRH-1

FGFR4 SHP

FXR FXR

Bile acids BSEP

NTCP

FGF15/19 Bile acids

Portal blood

OSTα/β
TCDCA
ASBT
IBABP
FXR

FGF15/19

Insulin
sensitivity TGR5 FXR
GLP-1

L cells

Figure 23.4  Mechanisms of bile acid feedback regulation of bile acid synthesis and homeostasis. Bile acid‐activated FXR regulates bile acid
synthesis and bile acid transport in hepatocytes and enterocytes. FXR and TGR5 are coexpressed in the intestinal L cells.
 23:  BILE ACID METABOLISM IN HEALTH AND DISEASE 279
CYP7A1 gene transcription in hepatocytes [52]. FGF15 binds
and activates membrane FGF receptor 4 (FGFR4), which is the
predominant FGF receptor expressed in hepatocytes. FGFR4
forms a complex with β‐Klotho to activate ERK1/2 to repress
CYP7A1 gene expression. The downstream target of FGF15
signaling is still not fully clear, but FGF15‐activated ERK1/2
may phosphorylate HNF4α to inhibit its binding to the CYP7A1
gene [53]. FGF15 is also required for gallbladder refilling [54],
and is an insulin‐independent postprandial regulator of hepatic
protein and glucose metabolism [55].
FGF19 is the human ortholog of FGF15 and shares ~51%
amino acid sequence identity with mouse FGF15. FGF19 tran-
scription is also induced by FXR [53, 56]. Functionally, FGF19
activates FGFR4 and ERK1/2 and strongly represses CYP7A1
gene transcription in human hepatocytes [53, 56]. Human pri-
mary hepatocytes express FGF19, while mouse hepatocytes do
not express FGF15 [53, 57]. Human patients with extrahepatic
obstructive cholestasis had increased liver FGF19 mRNA and
plasma FGF19 concentration [57], while mouse intestine FGF15
expression significantly decreased in a bile duct ligation mouse
model of obstructive cholestasis [52]. These findings suggest an
autocrine model by which hepatocyte FGF19 inhibits CYP7A1
in response to intrahepatic bile acid concentration.
The physiological role of FXR‐mediated regulation of hepatic
bile acid synthesis is best demonstrated in genetically modified
mice with disrupted FXR, SHP, or FGF15 [58–62]. Another line
of information came from mice with disrupted enterohepatic
bile acid transport. In mice lacking OSTα, hepatic CYP7A1 was
increased, despite reduced bile acids returning to the liver and a
significantly smaller bile acid pool size [29, 63]. These observa-
tions indicate that an enlarged bile acid pool in the enterocytes
can exert a dominant effect on inhibition of hepatic bile acid
synthesis and the intestine‐to‐liver axis via FXR/FGF15 signal-
ing plays a critical role in bile acid feedback regulation.
Several studies have shown that CYP7A1 mRNA was strongly
repressed in hepatocytes and in mice within 2–6 hours post
FGF15 or FGFG19 administration [58, 64, 65]. Such rapid
CYP7A1 downregulation is consistent with the reported short
CYP7A1 mRNA half‐life of about 30 minutes in cultured cells
[66]. In a recent study, FXR activation was shown to induce
RNA‐binding protein ZFP36L1 to decrease the stability of
CYP7A1 mRNA [67]. Liver CYP7A1 mRNA was markedly
decreased in bile duct‐ligated mice [68, 69]. It is expected that
intestine FGF15 production is significantly decreased under
obstructive cholestasis. Both hepatic FXR activation and proin-
flammatory cytokines can cause CYP7A1 repression in obstruc-
tive cholestasis [70]. These findings demonstrate that significant
redundancy exists regarding bile acid feedback inhibition of bile
acid synthesis in both physiological and pathological conditions.
Bile acid synthesis is the major pathway for cholesterol
catabolism in hepatocytes. In mice and rats, but not in humans,
cholesterol accumulation induces CYP7A1 via activation of the
nuclear receptor liver X receptor (LXR) [71, 72]. LXR in hepat-
ocytes also induces the transcription of ABCG5/G8 and ABCA1
and ABCG1, which mediate cholesterol efflux, alleviating intra-
hepatic cholesterol accumulation [73–75]. LXRα binds BARE‐1
in the mouse and rat Cyp7a1 gene promoter, but does not bind
BARE‐1 in the human CYP7A1 gene promoter due to sequence
differences [76, 77].
Regulation of bile acid transport
Biliary bile acid secretion and ileal bile acid uptake are two key
mechanisms driving the enterohepatic circulation of bile acids.
In hepatocytes, FXR induces BSEP transcription to promote
apical bile acid secretion [78]. On the basolateral side, FXR
activation inhibits NTCP expression [79]. Similarly, in entero-
cytes, FXR activation induces intestinal bile acid binding pro-
tein (IBABP), OSTα, and OSTβ and inhibits ASBT [80–82].
FXR induces BSEP, I‐BABP, OSTα, and OSTβ gene transcrip-
tion via direct binding to their gene promoters. FXR inhibition
of ASBT may be mediated by SHP. Therefore, FXR‐mediated
regulation of bile acid transporter expression does not promote
trans‐hepatocyte or trans‐enterocyte bile acid flux. Instead, the
main purpose seems to be decreasing intracellular bile acid con-
centration. This FXR‐regulated bile acid transport may serve as
an adaptive response to protect against bile acid accumulation in
hepatocytes during cholestasis. Pharmacological inhibition of
NTCP has hepato‐protective effects in mouse models of choles-
tasis [83]. Isoforms of the basolateral transporters OSTα/β and
MRP 3/4 are induced during cholestasis to efflux bile acids into
the systemic circulation [28, 40, 41].
Gut microbiota and bile acid metabolism
High‐fat diets, circadian disruption, time of feeding/fasting,
drugs, alcohol, and hormones all can shape the gut microbiota
to alter bile acid metabolism, homeostasis and host metabo-
lism [84]. Biotransformation of primary bile acids to second-
ary bile acids occurs in the small intestine and the colon. The
small intestine has a low bacterial population (a gradient of
103–108 g−1 wet weight in feces) but is the major site for bile
acid reabsorption, and nutrient digestion and absorption. The
colon has a high bacterial population (1011 g−1 wet weight).
DCA is a potent antimicrobial agent that controls bacterial
overgrowth and maintains intestinal barrier function. BSH and
7α‐HSDH activities determine the extent of secondary bile
acid synthesis, the bile acid species present, and the hydropho-
bicity of bile acids in enterohepatic circulation. Antibiotic‐
treated mice and germ‐free mice have increased bile acid
synthesis and pool size (~30%) compared to conventionally
raised mice [45]. Thus, the gut microbiota determines the total
bile acid pool size in the liver, gallbladder, bile, and intestine.
Gut bacteria utilize bile acids, polysaccharides, cellulose, and
starch to produce short‐chain fatty acids such as acetate,
butyrate, and propionate, and amino acids for energy metabo-
lism and growth. In the small intestine and colon, BSH activity
is  high in the Gram‐positive bacteria genera Clostridium,
Enterococcus, Bifidobacterium, and Lactobacillus, and Gram‐
negative genus Bacteroides. These anaerobic bacteria contain
high HSDH activity and are involved in multiple steps of 7α‐
dehydroxylation during secondary bile acid synthesis by gut
bacteria. The bile acid‐inducible operon (bai) in Clostridium
species has been elucidated [15]. The baiE gene encodes bile
acid 7α‐HSDH, and the baiI gene may encode 7β‐HSDH.
Antibiotic‐resistant Clostridium difficile (C. dif) outbreaks in
hospitals are a growing concern for infection and patient mor-
tality. Studies show that 7α‐HSDH in Clostridium scindens
may protect from C. dif. infection [85].
280 THE LIVER:  BILE ACID METABOLIC DISEASE AND THERAPY
Recent studies of human and mouse gut microbiomes by 16S
ribosomal RNA sequencing have identified gut microbe‐associ-
ated diseases, such as non‐alcoholic fatty liver disease (NAFLD),
type 2 diabetes, and hepatocellular carcinoma (HCC). In
human feces, 90% of the bacteria belong to two phyla: Firmicutes
and Bacteroidetes. The relative abundance of Firmicutes and
Bacteroidetes is associated with obesity in mouse models
and  human volunteers [86]. A higher ratio of Firmicutes to
Bacteroidetes enables the gut microbiota to extract energy more
efficiently from high‐fat diets, thus increasing adiposity. Cholic
acid feeding is known to increase the ratio of Firmicutes to
Bacteroidetes in mice. An animal‐based diet promotes dysbiosis
by increasing the abundance of bile‐tolerant bacteria and by
decreasing Firmicutes, which metabolizes plant polysaccharides
[87]. A high‐fat diet significantly induced TCA, which expanded
Bacteroidetes and Bilophila wadsworthia populations to promote
colitis in IL10−/− mice [88]. On the other hand, the antioxidant
tempol reduced Lactobacillus and decreased BSH activity to
increase Tβ‐MCA, which antagonized intestinal FXR and resulted
in increased bile acid synthesis and improved diet‐induced obe-
sity and diabetes [89]. Interestingly, a recent study reports that
bile acids induce uncoupling protein 1 (UCP‐1) in brown adipose
tissue to stimulate energy metabolism in mice housed at thermo-
neutrality, and reducing ambient temperature increases brown
adipose tissue thermogenesis to reduce weight in diet‐induced
obese mice [90]. It was found that hepatic cholesterol was
increased during cold‐induced thermogenesis to stimulate bile
acid synthesis by inducing Cyp7b1 of the alternative bile acid
synthesis pathway. Increased bile acid synthesis via the alterna-
tive pathway shaped the gut microbiota to stimulate adaptive ther-
mogenesis in brown adipose tissues in mice [91]. It is not clear
why the alternative pathway was preferentially stimulated in
cold‐induced thermogenesis. Stimulating the alternative bile acid
synthesis pathway alters bile acid composition by reducing tauro-
cholic acid synthesis and increasing CDCA, which may be con-
verted to LCA by gut microbiota to stimulate TGR5‐mediated
glucagon‐like peptide‐1 (GLP‐1) secretion to promote adipocyte
browning [92].
Circadian and nutrient interaction
in the control of bile acid homeostasis
Circadian metabolic homeostasis is maintained by the mamma-
lian biological clock located in the hypothalamic suprachias-
matic nucleus (SCN). The central clock in the SCN is entrained
by the daily environmental light/dark cycle and directs the phys-
iological timing of peripheral clocks located in nearly all organs
and tissues. In the liver, lipid, glucose, cholesterol, and bile acid
metabolisms are under circadian control. Disruption in circa-
dian timing (e.g. sleep deprivation, jet leg, shift work) uncou-
ples the central and peripheral clocks to cause dysbiosis and can
alter metabolic homeostasis and contribute to the pathogenesis
of cardiovascular disease, metabolic syndromes, gastrointesti-
nal disorders, non‐alcoholic steatohepatitis (NASH), and HCC
[93–98]. The Cyp7a1 gene expression and bile acid synthesis
exhibit a distinct circadian rhythm, peaking in day in humans
and night in rodents [99, 100]. This rhythm is altered by high‐fat
diets, alcohol consumption, and sleep disruption. Feeding
­rapidly stimulated Cyp7a1 gene expression but suppressed
Cyp8b1 gene expression, whereas fasting suppressed Cyp7a1
and stimulated Cyp8b1 expression [101]. The gut microbiome
also exhibits a circadian rhythm in composition, which is damp-
ened by HFD and circadian disruption, causing dysbiosis and
impairment of  barrier function (leaky gut), and is associated
with the pathogenesis of inflammatory bowel diseases, diabetes,
and obesity [102]. The feeding and fasting cycle modulates cir-
cadian rhythmicity and bile acid metabolism. Interestingly,
intermittent ­fasting shapes the gut microbiome by increasing the
ratio of gut Firmicutes to Bacteroidetes, which in turn increases
acetate and lactate. It also promotes white adipose tissue brown-
ing and decreases obesity in mice [103].
BILE ACID METABOLIC DISEASE
AND THERAPY
Bile acid synthesis defects
Analysis of serum and urine bile acid intermediates has been
used to identify 13 human genes in inborn errors of bile acid
synthesis [104]. Deficiency of primary bile acid synthesis
causes malabsorption of fat, steroids, and fat‐soluble vitamins,
leading to steatorrhea and growth retardation. Reduced bile acid
synthesis and feedback inhibition causes stimulation of CYP7A1
and CYP8B1 expression and accumulation of bile acid interme-
diates at the points of metabolic block [104]. Deficiency of
CYP7A1 does not cause severe metabolic disorder in humans.
Homozygous CYP7A1 mutation in humans resulted in decreased
bile acids levels, hypercholesterolemia and premature gallstone
disease, supporting the key role of bile acid synthesis in choles-
terol metabolism [9]. Individuals who were heterozygous for
the CYP7A1 mutation also showed a hyperlipidemic phenotype.
CYP7A1 polymorphisms have also been associated with risk of
gallstones, hyperlipidemia, and cardiovascular events [105–
107]. In contrast, mutation of the CYP7B1 gene in an infant
caused neonatal cholestasis and fibrosis [108, 109], and pro-
gressive spastic paraplegia [110]. In CYP7B1‐deficient patients,
3β‐monohydroxy‐Δ5 bile acids are accumulated, which are
toxic and cause cholestatic liver injury, giant cell hepatitis,
severe fibrosis, and cirrhosis. Numerous CYP27A1 mutations
identified in humans result in the rare lipid storage disorder
­cerebrotendinous xanthomatosis (CTX) [111], which is charac-
terized by accumulation of cholesterol and cholestanol in xan-
thomas and in the brain. CTX patients have xanthomatosis,
premature atherosclerosis and progressive neurological disorders.
Deficiency of CYP27A1 prevented both classic and alternative
bile acid synthesis, and 5β‐cholestane‐3α,7α,12α‐triol is converted
to bile alcohol and cholestenol, which accumulate in tissues to
form xanthomas. CTX can be effectively managed by CDCA
therapy, which inhibits CYP7A1 to alleviate accumulation of
7α‐hydroxylated cholesterol metabolites. Several mutations in
3β‐hydroxy‐Δ5 ‐C27‐steroid oxidoreductase (HSD3B7) and
Δ4–3‐oxidoreductase (AKR1D1) have been identified. Mutations
of peroxisomal very‐long‐chain acyl‐CoA synthase (SLC27A5),
bile acid‐acyl transporter (ABCD3), α‐methylacyl‐CoA race-
mase (AMACR), acyl‐CoA oxidase 2 (ACOX2), D‐bifunctional
protein (HSD17B4), bile acyl‐CoA:amino acid N‐acyltransferase
 23:  BILE ACID METABOLISM IN HEALTH AND DISEASE 281
(BAAT), and sterol carrier protein X (SCPx) cause Zellweger‐
related peroxisome diseases. Zellweger disorders are caused by
deficiency of PEX genes involved in peroxisome biogenesis and
abnormal C27‐bile acid intermediates and C29‐dicarboxylic
acid are accumulated in serum.
Cholestatic liver diseases
Cholestasis is a chronic liver condition resulting from obstructed
hepatic bile flow, leading to accumulation of bile acids in the
liver and increased bile acids in the systemic circulation.
Chronic cholestasis leads to fibrosis, cirrhosis, liver failure, and
higher risk of hepatocellular or cholangiocellular carcinomas.
Genetic mutations of bile acid transporter genes and autoim-
mune destruction of small bile ductules result in intrahepatic
cholestasis, while obstruction of extrahepatic bile ducts by com-
mon bile duct stones or tumors of the bile duct or pancreas can
cause extrahepatic cholestasis [112, 113].
Congenital cholestasis is usually early‐onset and is associated
with jaundice, pruritus, and growth failure. Progressive familial
intrahepatic cholestasis (PFIC) and benign recurrent intrahepatic
cholestasis (BRIC) are autosomal recessive diseases associated
with genetic mutations in ATP8B1 (type 1, PFIC1), BSEP (type
2, PFIC2), and MDR3 (type 3, PFIC3) [112]. The ATP8B1 gene
encodes a p‐type cation transporter and phospholipid flippase
that transports phosphatidylserine from the outer to the inner
leaflet to maintain an asymmetric membrane with higher phos-
phatidylcholine on the outer layer of the membrane. The current
hypothesis is that ATP8B1 loss of function alters membrane
structure, resulting in impaired bile acid transporter function. In
PFIC2, lack of functional BSEP at the apical hepatocyte mem-
brane causes hepatic bile acid accumulation, giant cell hepatitis,
and hepatocellular necrosis [114]. PFIC2 is also associated with
higher incidence of gallstones. PFIC3 is associated with muta-
tions of the apical phospholipid transporter MDR3. The resulting
biliary injury is due to chronic exposure of cholangiocytes to
high concentrations of non‐micellar bile acids. Decreased phos-
pholipid secretion also results in unstable micelles that favor
cholesterol crystallization and small bile duct obstruction. High
serum levels of γ‐glutamyl transpeptidase (GGT), a marker of
ductular damage, is a characteristic feature of PFIC3 and distin-
guishes it from PFIC1 and PFIC2. Mutation of the TJP2 gene,
which encodes tight junction protein 2 localized to hepatic tight
junctions, causes progressive cholestasis in humans and has been
referred to as PFIC4 [115]. More recently, children with neonatal
cholestasis associated with FXR mutation have also been
reported [116]. These patients had undetectable hepatic expres-
sion of BSEP, normal or near‐normal serum GGT, and rapidly
progressed to end‐stage liver disease.
Among the acquired forms of cholestasis, intrahepatic choles-
tasis of pregnancy (ICP) is a common pregnancy‐related liver
disease that affects ~1% of all pregnant women. ICP is primarily
diagnosed in the third trimester and pruritus is a major initial
clinical presentation of ICP. ICP is reversible and rapidly resolves
after delivery. Environmental, hormonal, and genetic factors are
thought to be involved in the pathogenesis. Genetic variations of
the BSEP gene may be associated with ICP [117, 118].
Primary biliary cholangitis (PBC), previously known as pri-
mary biliary cirrhosis, is another form of acquired chronic
cholestasis associated with autoimmune destruction of the small
bile ducts causing portal infiltration and fibrosis. About 95% of
the PBC patients are middle‐aged females. Primary sclerosing
cholangitis (PSC) is associated with injury and fibrosis of both
intrahepatic and extrahepatic bile ducts, resulting in biliary
strictures and obstructed bile flow. PSC is a male‐dominant dis-
ease with a male:female ratio of ~2:1.
Bile acids as therapeutic agents
Non‐alcoholic fatty liver diseases
Diabetes and fatty liver are inflammatory diseases associated
with dyslipidemia and hepatic insulin resistance. NAFLD is the
most common chronic liver disease with prevalence in ~30% of
the US population, and is a risk factor for type 2 diabetes, obe-
sity, and cardiovascular disease [119–121]. NASH is a progres-
sive form of NAFLD that may lead to cirrhosis and liver cancer.
The molecular mechanism of progression from simple steatosis
to fibrosis in NASH is not completely understood, and there is no
effective drug therapy for NASH. Recent studies indicate NASH
patients have increased circulating conjugated primary bile acids
(especially G/TCA and TCDCA), an increased ratio of conju-
gated CA to CDCA, and decreased secondary bile acids [122].
The cause of altered serum bile acid composition in NASH is not
understood. It is likely that the gut microbiota plays a critical role
in the control of bile acid biotransformation, bile acid pool size,
and bile acid feedback regulation in NASH and type 2 diabetes.
Recently, bile acid‐based therapies have been developed to treat
cholestasis and NASH. Activation of FXR and TGR5 signaling
protects against inflammation of liver and intestine.
Ursodeoxycholic acid
Ursodeoxycholic acid (UDCA) (trade name, ursodiol) is a 7β‐
epimer of CDCA and a hydrophilic bile acid present in small
quantities in human bile. Pharmacological doses of UDCA do
not produce a toxic effect in humans. UDCA has been used for
cholesterol gallstone dissolution [123]. UDCA is more effective
towards smaller stones and complete dissolution may take
~6–24 months. UDCA can be beneficial in the prevention of
cholelithiasis in certain high‐risk conditions such as during
pregnancy or following bariatric surgery. Currently, UDCA is
the first‐line therapy for PBC [124]. It significantly improves
liver function tests and delays the need for liver transplantation
in PBC patients. UDCA can also provide beneficial effects in
ICP and PFIC3. UDCA can increase bile acid pool hydrophilic-
ity and decrease bile acid toxicity [125]. It stimulates bile flow
and promotes biliary HCO3− secretion [126, 127]. and it also
exhibits anti‐inflammatory and anti‐apoptosis effects [127,
128]. However, about 40% of PBC patients do not adequately
respond to UDCA therapy [129].
Many clinical trials have been conducted to test UDCA ther-
apy for treating PSC, but its use in PSC has not been recom-
mended in clinical practice guidelines due to the lack of
consistently perceived benefits [130, 131]. Nor‐ursodeoxy-
cholic acid (norUDCA) is a side‐chain‐shortened C23 homolog
of UDCA [132, 133]. In enterohepatic circulation, norUDCA
does not undergo significant amidation but is partially glucuro-
nidated. NorUDCA undergoes cholehepatic shunting and
282 THE LIVER:  REFERENCES
promotes HCO3− secretion [127] and exhibits potent anti‐­
cholestasis and anti‐inflammatory effects in experimental
­models [134]. A recent randomized controlled phase 2 trial
showed that norUDCA significantly and dose‐dependently
reduced alkaline phosphatase in PSC patients in a 12‐week
treatment [135], which warrants further clinical investigation. The
use of norUDCA appeared to be safe in PSC patients, though
norUDCA did not improve pruritus.
Farnesoid X receptor agonists
Several potent FXR agonists have been developed, among which
is obeticholic acid (OCA), a 6α‐ethyl CDCA derivative that selec-
tively activates FXR with an EC50 of ~100 nM [136]. Activation
of FXR decreases bile acid synthesis, increases intracellular bile
acid accumulation, and alleviates inflammation, which underlie
the beneficial effects of OCA in experimental cholestasis [136]
and human PBC [137–139]. OCA has recently been approved by
the US Food and Drug Administration (FDA) to treat PBC
patients who do not respond to or cannot tolerate UDCA. In phase
2 trials, OCA also improved NASH scores and is under further
testing for the treatment of fatty liver disease [140].
Bile acid sequestrants
Bile acid sequestrants are a class of drugs that bind negatively
charged bile acids in the intestine and prevent bile acid reab-
sorption. This action leads to stimulated hepatic bile acid syn-
thesis, increased hepatic LDL uptake, and decreased plasma
LDL cholesterol. Cholestyramine and colestipol are classic bile
acid sequestrants used to lower plasma cholesterol in humans,
but their use is less common given the availability of other lipid‐
lowering drugs. Bile acid sequestrants are used to treat pruritus
in cholestasis. Colesevelam [141], a second‐generation bile acid
sequestrant, improves glycemic control in type 2 diabetes. The
mechanism of action of bile acid sequestrants is not completely
understood, but may be attributed to increased GLP‐1 secretion
[142] and reduced hepatic glucose production [143].
Bariatric surgery
Bariatric surgeries for weight reduction result in increased serum
conjugated bile acids, which is positively correlated to improved
insulin resistance in obese patients shortly after bariatric surgery
[144, 145]. The early increase of serum bile acids and insulin
sensitivity after Roux‐en‐Y gastric bypass may be linked to
increased secondary bile acids and GLP‐1 [146]. Vertical sleeve
gastrectomy improved insulin sensitivity in high‐fat diet‐fed
wild‐type mice but not in Fxr−/− mice or Tgr5−/− mice [147, 148],
suggesting that FXR and TGR5 may be involved. However, the
underlying molecular mechanism of bile acid receptor signaling
in improving diabetes after bariatric surgery is not clear.
CONCLUSION
This chapter summarizes the physiochemical properties of bile
acids and updated the mechanisms of bile acid synthesis, trans-
port, and regulation of homeostasis. The major physiological
functions of bile acids in the digestive system depend on both
the detergent and signaling properties of bile acids. Nuclear
receptors, the gut microbiota, nutrients, and circadian rhythms
control bile acid metabolism and homeostasis. New findings
in  these areas demonstrated the high complexity of bile acid
metabolism, regulation, and function in physiology and diseases.
Inborn errors of bile acid metabolism are rare but are usually
associated with severe hepatic complications. Significant pro-
gress has been made in obtaining new knowledge of bile acid
biology and in translating these basic research findings to clinical
applications. Current studies support the concept of targeting
different steps in the enterohepatic circulation of bile acids to
achieve therapeutic benefits in various liver and metabolic diseases.
Further developments in bile acid‐based therapies are anticipated
in the years to come.
ACKNOWLEDGMENTS
This work was supported in part by NIH grants, DK44442 and
DK58379 to JYLC, and 1R01DK102487‐01 to TL. We thank
Dr. Alan Hofmann for his critical review of the manuscript.
REFERENCES
1. Russell, D.W. The enzymes, regulation, and genetics of bile acid synthesis.
Annu Rev Biochem, 2003;72:1370174.
2. Myant, N.B. and Mitropoulos, K.A. Cholesterol 7 alpha‐hydroxylase. J Lipid
Res, 1977;18(2):135–53.
3. Axelson, M., Mork, B., and Sjovall, J. Occurrence of 3 beta‐hydroxy‐5‐
cholestenoic acid, 3 beta,7 alpha‐dihydroxy‐5‐cholestenoic acid, and 7
alpha‐hydroxy‐3‐oxo‐4‐cholestenoic acid as normal constituents in human
blood. J Lipid Res, 1988;29(5):629–41.
4. Matsuzawa, N., Takamura, T., Kurita, S. et al. Lipid‐induced oxidative stress
causes steatohepatitis in mice fed an atherogenic diet. Hepatology,
2007;46(5):1392–403.
5. Mihalik, S.J., Steinberg, S.J., Pei, Z. et al. Participation of two members of
the very long‐chain acyl‐CoA synthetase family in bile acid synthesis and
recycling. J Biol Chem, 2002;277(27):24771–9.
6. Falany, C.N., Johnson, M.R., Barnes, S., and Diasio, R.B. Glycine and tau-
rine conjugation of bile acids by a single enzyme. Molecular cloning and
expression of human liver bile acid CoA:amino acid N‐acyltransferase. J
Biol Chem, 1994;269(30):19375–9.
7. Pellicoro, A., van den Heuvel, F.A., Geuken, M. et al. Human and rat bile
acid‐CoA:amino acid N‐acyltransferase are liver‐specific peroxisomal
enzymes: implications for intracellular bile salt transport. Hepatology, 2007;
45(2):340–8.
8. Ferrell, J.M., Boehme, S., Li, F., and Chiang, J.Y. Cholesterol 7α‐hydroxy-
lase‐deficient mice are protected from high fat/high cholesterol diet‐induced
metabolic disorders. J Lipid Res, 2016;57:1144–54.
9. Pullinger, C.R., Eng, C., Salen, G. et al. Human cholesterol 7alpha‐hydroxy-
lase (CYP7A1) deficiency has a hypercholesterolemic phenotype. J Clin
Invest, 2002;110(1):109–17.
10. Takahashi, S., Fukami, T., Masuo, Y. et al. Cyp2c70 is responsible for the
species difference in bile acid metabolism between mice and humans. J Lipid
Res, 2016;57(12):2130–7.
11. Hofmann, A.F., Hagey, L.R., and Krasowski, M.D. Bile salts of vertebrates:
structural variation and possible evolutionary significance. J Lipid Res,
2010;51(2):226–46.
12. Moschetta, A., Xu, F., Hagey, L.R. et al. A phylogenetic survey of biliary
lipids in vertebrates. J Lipid Res, 2005;46(10):2221–32.
13. Falany, C.N., Fortinberry, H., Leiter, E.H., and Barnes, S. Cloning, expres-
sion, and chromosomal localization of mouse liver bile acid CoA:amino acid
N‐acyltransferase. J Lipid Res, 1997;38(6):1139–48.
 23:  BILE ACID METABOLISM IN HEALTH AND DISEASE 283
14. Hardison, W.G. Hepatic taurine concentration and dietary taurine as regula-
tors of bile acid conjugation with taurine. Gastroenterology, 1978;
75(1):71–5.
15. Ridlon, J.M., Kang, D.J., and Hylemon, P.B. Bile salt biotransformations by
human intestinal bacteria. J Lipid Res, 2006;47(2):241–59.
16. Bjorkhem, I., Einarsson, K., Melone, P., and Hylemon, P. Mechanism of
intestinal formation of deoxycholic acid from cholic acid in humans: evidence
for a 3‐oxo‐delta 4‐steroid intermediate. J Lipid Res, 1989;30(7):1033–9.
17. Hylemon, P.B., Melone, P.D., Franklund, C.V., Lund, E., and Bjorkhem, I.
Mechanism of intestinal 7 alpha‐dehydroxylation of cholic acid: evidence
that allo‐deoxycholic acid is an inducible side‐product. J Lipid Res, 1991;
32(1):89–96.
18. Trauner, M. and Boyer, J.L. Bile salt transporters: molecular characterization,
function, and regulation. Physiol Rev, 2003;83(2):633–71.
19. Otsuki, M. Pathophysiological role of cholecystokinin in humans. J
Gastroenterol Hepatology, 2000;15(Suppl):D71–83.
20. Boyer, J.L. Bile formation and secretion. Compr Physiol, 2013;3(3):
1035–78.
21. Childs, S., Yeh, R.L., Georges, E., and Ling, V. Identification of a sister gene
to P‐glycoprotein. Cancer Res, 1995;55(10):2029–34.
22. Berge, K.E., Tian, H., Graf, G.A. et al. Accumulation of dietary cholesterol
in sitosterolemia caused by mutations in adjacent ABC transporters. Science,
2000;290(5497):1771–5.
23. Smit, J.J., Schinkel, A.H., Oude Elferink, R.P. et al. Homozygous disruption
of the murine mdr2 P‐glycoprotein gene leads to a complete absence of
phospholipid from bile and to liver disease. Cell, 1993;75(3):451–62.
24. Shneider, B.L., Dawson, P.A., Christie, D.M. et al. Cloning and molecular
characterization of the ontogeny of a rat ileal sodium‐dependent bile acid
transporter. J Clin Invest, 1995;95(2):745–54.
25. Dawson, P.A., Haywood, J., Craddock, A.L. et al. Targeted deletion of the
ileal bile acid transporter eliminates enterohepatic cycling of bile acids in
mice. J Biol Chem, 2003;278(36):33920–7.
26. Oelkers, P., Kirby, L.C., Heubi, J.E., and Dawson, P.A. Primary bile acid
malabsorption caused by mutations in the ileal sodium‐dependent bile acid
transporter gene (SLC10A2). J Clin Invest, 1997;99(8):1880–7.
27. Gong, Y.Z., Everett, E.T., Schwartz, D.A., Norris, J.S., and Wilson, F.A.
Molecular cloning, tissue distribution, and expression of a 14‐kDa bile acid‐
binding protein from rat ileal cytosol. Proc Natl Acad Sci U S A, 1994;
91(11):4741–5.
28. Ballatori, N., Christian, W.V., Lee, J.Y. et al. OSTalpha‐OSTbeta: a major
basolateral bile acid and steroid transporter in human intestinal, renal, and
biliary epithelia. Hepatology, 2005;42(6):1270–9.
29. Rao, A., Haywood, J., Craddock, A.L. et al. The organic solute transporter
alpha‐beta, Ostalpha‐Ostbeta, is essential for intestinal bile acid transport
and homeostasis. Proc Natl Acad Sci U S A, 2008;105(10):3891–6.
30. Xia, X., Francis, H., Glaser, S., Alpini, G., and LeSage, G. Bile acid interac-
tions with cholangiocytes. World J Gastroenterol, 2006;12(22):3553–63.
31. Soroka, C.J., Mennone, A., Hagey, L.R., Ballatori, N., and Boyer, J.L. Mouse
organic solute transporter alpha deficiency enhances renal excretion of bile
acids and attenuates cholestasis. Hepatology, 2010;51(1):181–90.
32. Meier, P.J. Molecular mechanisms of hepatic bile salt transport from sinusoi-
dal blood into bile. Am J Physiol, 1995;269(6 Pt 1):G801–12.
33. Hagenbuch, B., Stieger, B., Foguet, M., Lubbert, H., and Meier, P.J.
Functional expression cloning and characterization of the hepatocyte Na+/
bile acid cotransport system. Proc Natl Acad Sci U S A, 1991;88(23):
10629–33.
34. Hagenbuch, B. and Meier, P.J. Molecular cloning, chromosomal localization,
and functional characterization of a human liver Na+/bile acid cotransporter.
J Clin Invest, 1994;93(3):1326–31.
35. Meier, P.J. and Stieger, B. Bile salt transporters. Annu Rev Physiol,
2002;64:635–61.
36. Suga, T., Yamaguchi, H., Sato, T. et al. Preference of conjugated bile acids
over unconjugated bile acids as substrates for OATP1B1 and OATP1B3.
PLoS One, 2017;12(1):e0169719.
37. Vaz, F.M., Paulusma, C.C., Huidekoper, H. et  al. Sodium taurocholate
cotransporting polypeptide (SLC10A1) deficiency: conjugated hyper-
cholanemia without a clear clinical phenotype. Hepatology, 2015;
61(1):260–7.
38. Slijepcevic, D., Kaufman, C., Wichers, C.G. et al. Impaired uptake of conjugated
bile acids and Hepatitis B Virus preS1‐binding in Na ‐taurocholate cotrans-
porting polypeptide knockout mice. Hepatology, 2015;62(1):207–19.
39. Slijepcevic, D., Roscam Abbing, R.L.P., Katafuchi, T. et al. Hepatic uptake
of conjugated bile acids is mediated by both sodium taurocholate cotrans-
porting polypeptide and organic anion transporting polypeptides and modulated
by intestinal sensing of plasma bile acid levels in mice. Hepatology, 2017;
66(5):1631–43.
40. Boyer, J.L., Trauner, M., Mennone, A. et al. Upregulation of a basolateral
FXR‐dependent bile acid efflux transporter OSTalpha‐OSTbeta in cholesta-
sis in humans and rodents. Am J Physiol Gastrointest Liver Physiol, 2006;
290(6):G1124–30.
41. Cui, Y.J., Aleksunes, L.M., Tanaka, Y., Goedken, M.J., and Klaassen, C.D.
Compensatory induction of liver efflux transporters in response to ANIT‐
induced liver injury is impaired in FXR‐null mice. Toxicol Sci, 2009;
110(1):47–60.
42. Kullak‐Ublick, G.A., Stieger, B., Hagenbuch, B., and Meier, P.J. Hepatic
transport of bile salts. Semin Liver Dis, 2000;20(3):273–92.
43. Thakare, R., Alamoudi, J.A., Gautam, N., Rodrigues, A.D., and Alnouti, Y.
Species differences in bile acids I. Plasma and urine bile acid composition. J
Appl Toxicol, 2018;38(10):1323–35.
44. Mangelsdorf, D.J., Thummel, C., Beato, M. et  al. The nuclear receptor
superfamily: the second decade. Cell, 1995;83(6):835–9.
45. Sayin, S.I., Wahlstrom, A., Felin, J. et al. Gut microbiota regulates bile acid
metabolism by reducing the levels of tauro‐beta‐muricholic acid, a naturally
occurring FXR antagonist. Cell Metab, 2013;17(2):225–35.
46. Maruyama, T., Miyamoto, Y., Nakamura, T. et  al. Identification of mem-
brane‐type receptor for bile acids (M‐BAR). Biochem Biophys Res Commun,
2002;298(5):714–19.
47. Kawamata, Y., Fujii, R., Hosoya, M. et  al. A G protein‐coupled receptor
responsive to bile acids. J Biol Chem, 2003;278:9435–40.
48. Goodwin, B., Jones, S.A., Price, R.R. et  al. A regulatory cascade of the
nuclear receptors FXR, SHP‐1, and LRH‐1 represses bile acid biosynthesis.
Mol Cell, 2000;6(3):517–26.
49. Lu, T.T., Makishima, M., Repa, J.J. et al. Molecular basis for feedback regu-
lation of bile acid synthesis by nuclear receptors. Mol Cell, 2000;6(3):
507–15.
50. Pandak, W.M., Heuman, D.M., Hylemon, P.B., Chiang, J.Y., and Vlahcevic,
Z.R. Failure of intravenous infusion of taurocholate to down‐regulate choles-
terol 7 alpha‐hydroxylase in rats with biliary fistulas. Gastroenterology,
1995;108(2):533–44.
51. Li, T., Francl, J.M., Boehme, S. et al. Glucose and insulin induction of bile
acid synthesis: mechanisms and implication in diabetes and obesity. J Biol
Chem, 2012;287(3):1861–73.
52. Inagaki, T., Choi, M., Moschetta, A. et al. Fibroblast growth factor 15 func-
tions as an enterohepatic signal to regulate bile acid homeostasis. Cell
Metab, 2005;2(4):217–25.
53. Song, K.H., Li, T., Owsley, E., Strom, S., and Chiang, J.Y. Bile acids activate
fibroblast growth factor 19 signaling in human hepatocytes to inhibit choles-
terol 7alpha‐hydroxylase gene expression. Hepatology, 2009;49(1):
297–305.
54. Choi, M., Moschetta, A., Bookout, A.L. et al. Identification of a hormonal
basis for gallbladder filling. Nat Med, 2006;12(11):1253–5.
55. Kir, S., Beddow, S.A., Samuel, V.T. et al. FGF19 as a postprandial, insulin‐
independent activator of hepatic protein and glycogen synthesis. Science,
2011;331(6024):1621–4.
56. Holt, J.A., Luo, G., Billin, A.N. et al. Definition of a novel growth factor‐
dependent signal cascade for the suppression of bile acid biosynthesis.
Genes Dev, 2003;17(13):1581–91.
57. Schaap, F.G., van der Gaag, N.A., Gouma, D.J., and Jansen, P.L. High expres-
sion of the bile salt‐homeostatic hormone fibroblast growth factor 19 in the liver
of patients with extrahepatic cholestasis. Hepatology, 2009;49(4):1228–35.
58. Kong, B., Wang, L., Chiang, J.Y. et  al. Mechanism of tissue‐specific
farnesoid X receptor in suppressing the expression of genes in bile‐acid syn-
thesis in mice. Hepatology, 2012;56(3):1034–43.
59. Yu, C., Wang, F., Kan, M. et al. Elevated cholesterol metabolism and bile
acid synthesis in mice lacking membrane tyrosine kinase receptor FGFR4. J
Biol Chem, 2000;275(20):15482–9.
60. Li, S., Hsu, D.D., Li, B. et al. Cytoplasmic tyrosine phosphatase Shp2 coor-
dinates hepatic regulation of bile acid and FGF15/19 signaling to repress bile
acid synthesis. Cell Metab, 2014;20(2):320–32.
61. Lin, B.C., Wang, M., Blackmore, C., and Desnoyers, L.R. Liver‐specific
activities of FGF19 require Klotho beta. J Biol Chem, 2007;282(37):
27277–84.
284 THE LIVER:  REFERENCES
62. Ito, S., Fujimori, T., Furuya, A. et al. Impaired negative feedback suppres-
sion of bile acid synthesis in mice lacking betaKlotho. J Clin Invest,
2005;115(8):2202–8.
63. Lan, T., Rao, A., Haywood, J., Kock, N.D., and Dawson, P.A. Mouse organic
solute transporter alpha deficiency alters FGF15 expression and bile acid
metabolism. J Hepatol, 2012;57(2):359–65.
64. Kong, B. and Guo, G.L. Soluble expression of disulfide bond containing
proteins FGF15 and FGF19 in the cytoplasm of Escherichia coli. PLoS One,
2014;9(1):e85890.
65. Potthoff, M.J., Boney‐Montoya, J., Choi, M. et  al. FGF15/19 regulates
hepatic glucose metabolism by inhibiting the CREB‐PGC‐1alpha pathway.
Cell Metab, 2011;13(6):729–38.
66. Baker, D.M., Wang, S.L., Bell, D.J., Drevon, C.A., and Davis, R.A. One or
more labile proteins regulate the stability of chimeric mRNAs containing the
3’‐untranslated region of cholesterol‐7alpha ‐hydroxylase mRNA. J Biol
Chem, 2000;275(26):19985–91.
67. Tarling, E.J., Clifford, B.L., Cheng, J. et al. RNA‐binding protein ZFP36L1
maintains posttranscriptional regulation of bile acid metabolism. J Clin
Invest, 2017;127(10):3741–54.
68. Wang, M., Tan, Y., Costa, R.H., and Holterman, A.X. In vivo regulation of
murine CYP7A1 by HNF‐6: a novel mechanism for diminished CYP7A1
expression in biliary obstruction. Hepatology, 2004;40(3):600–8.
69. Li, J., Woolbright, B.L., Zhao, W. et al. Sortilin 1 loss‐of‐function protects
against cholestatic liver injury by attenuating hepatic bile acid accumulation
in bile duct ligated mice. Toxicol Sci, 2018;161(1):34–47.
70. Miyake, J.H., Wang, S.L., and Davis, R.A. Bile acid induction of cytokine
expression by macrophages correlates with repression of hepatic cholesterol
7alpha‐hydroxylase. J Biol Chem, 2000;275(29):21805–8.
71. Peet, D.J., Turley, S.D., Ma, W. et al. Cholesterol and bile acid metabolism
are impaired in mice lacking the nuclear oxysterol receptor LXR alpha. Cell,
1998;93(5):693–704.
72. Lehmann, J.M., Kliewer, S.A., Moore, L.B. et al. Activation of the nuclear
receptor LXR by oxysterols defines a new hormone response pathway. J Biol
Chem, 1997;272(6):3137–40.
73. Repa, J.J., Berge, K.E., Pomajzl, C. et al. Regulation of ATP‐binding cassette
sterol transporters ABCG5 and ABCG8 by the liver X receptors alpha and
beta. J Biol Chem, 2002;277(21):18793–800.
74. Repa, J.J., Turley, S.D., Lobaccaro, J.A. et al. Regulation of Absorption and
ABC1‐mediated efflux of cholesterol by RXR heterodimers. Science,
2000;289(5484):1524–9.
75. Venkateswaran, A., Laffitte, B.A., Joseph, S.B. et al. Control of cellular cho-
lesterol efflux by the nuclear oxysterol receptor LXR alpha. Proc Natl Acad
Sci U S A, 2000;97(22):12097–102.
76. Chiang, J.Y., Kimmel, R., and Stroup, D. Regulation of cholesterol 7alpha‐
hydroxylase gene (CYP7A1) transcription by the liver orphan receptor
(LXRalpha). Gene, 2001;262(1–2):257–65.
77. Agellon, L.B., Drover, V.A., Cheema, S.K., Gbaguidi, G.F., and Walsh, A.
Dietary cholesterol fails to stimulate the human cholesterol 7alpha‐hydroxylase
gene (CYP7A1) in transgenic mice. J Biol Chem, 2002;277(23):20131–4.
78. Ananthanarayanan, M., Balasubramanian, N., Makishima, M., Mangelsdorf,
D.J., and Suchy, F.J. Human bile salt export pump promoter is transactivated
by the farnesoid X receptor/bile acid receptor. J Biol Chem,
2001;276(31):28857–65.
79. Denson, L.A., Sturm, E., Echevarria, W. et al. The orphan nuclear receptor,
shp, mediates bile acid‐induced inhibition of the rat bile acid transporter,
ntcp. Gastroenterology, 2001;121(1):140–7.
80. Lee, H., Zhang, Y., Lee, F.Y. et al. FXR regulates organic solute transporters
alpha and beta in the adrenal gland, kidney, and intestine. J Lipid Res,
2006;47(1):201–14.
81. Grober, J., Zaghini, I., Fujii, H. et al. Identification of a bile acid‐responsive
element in the human ileal bile acid‐binding protein gene. Involvement of the
farnesoid x receptor/9‐cis‐retinoic acid receptor heterodimer. J Biol Chem,
1999;274(42):29749–54.
82. Neimark, E., Chen, F., Li, X., and Shneider, B.L. Bile acid‐induced negative
feedback regulation of the human ileal bile acid transporter. Hepatology,
2004;40(1):149–56.
83. Slijepcevic, D., Roscam Abbing, R.L.P., Fuchs, C.D. et al. Na+‐taurocholate
cotransporting polypeptide inhibition has hepatoprotective effects in choles-
tasis in mice. Hepatology, 2018; 68(3):1057–69.
84. Wahlstrom, A., Sayin, S.I., Marschall, H.U., and Backhed, F. Intestinal
crosstalk between bile acids and microbiota and its impact on host metabo-
lism. Cell Metab, 2016;24(1):41–50.
85. Studer, N., Desharnais, L., Beutler, M. et al. Functional Intestinal Bile Acid
7alpha‐Dehydroxylation by Clostridium scindens associated with protec-
tion from Clostridium difficile infection in a gnotobiotic mouse model.
Front Cell Infect Microbiol, 2016;6:191.
86. Ley, R.E., Turnbaugh, P.J., Klein, S., and Gordon, J.I. Microbial ecology:
human gut microbes associated with obesity. Nature, 2006;
444(7122):1022–3.
87. David, L.A., Maurice, C.F., Carmody, R.N. et al. Diet rapidly and repro-
ducibly alters the human gut microbiome. Nature, 2014(505):559–63.
88. Devkota, S., Wang, Y., Musch, M.W. et al. Dietary‐fat‐induced taurocholic
acid promotes pathobiont expansion and colitis in I110‐/‐ mice. Nature,
2012;487(7405):104–8.
89. Li, F., Jiang, C., Krausz, K.W. et al. Microbiome remodelling leads to inhi-
bition of intestinal farnesoid X receptor signalling and decreased obesity.
Nat Commun, 2013;4:2384.
90. Zietak, M., Kovatcheva‐Datchary, P., Markiewicz, L.H. et  al. Altered
microbiota contributes to reduced diet‐induced obesity upon cold exposure.
Cell Metab, 2016;23(6):1216–23.
91. Worthmann, A., John, C., Ruhlemann, M.C. et al. Cold‐induced conversion
of cholesterol to bile acids in mice shapes the gut microbiome and promotes
adaptive thermogenesis. Nat Med, 2017;23(7):839–49.
92. Pathak, P., Cen, X., Nichols, R.G. et al. Intestine farnesoid X receptor ago-
nist and the gut microbiota activate G‐protein bile acid receptor‐1 signaling
to improve metabolism. Hepatology, 2018;68(4):1574–88.
93. Turek, F.W., Joshu, C., Kohsaka, A. et al. Obesity and metabolic syndrome
in circadian Clock mutant mice. Science, 2005;308(5724):1043–5.
94. Morris, C.J., Purvis, T.E., Hu, K., and Scheer, F.A. Circadian misalignment
increases cardiovascular disease risk factors in humans. Proc Natl Acad Sci
U S A, 2016;113(10):E1402–11.
95. Bass, J. and Takahashi, J.S. Circadian integration of metabolism and ener-
getics. Science, 2010;330(6009):1349–54.
96. Panda, S., Antoch, M.P., Miller, B.H. et al. Coordinated transcription of key
pathways in the mouse by the circadian clock. Cell, 2002;109(3):307–20.
97. Kettner, N.M., Voicu, H., Finegold, M.J. et  al. Circadian homeostasis of
liver metabolism suppresses hepatocarcinogenesis. Cancer Cell,
2016;30(6):909–24.
98. Maury, E., Hong, H.K., and Bass, J. Circadian disruption in the pathogen-
esis of metabolic syndrome. Diabetes Metab, 2014;40(5):338–46.
99. Galman, C., Angelin, B., and Rudling, M. Bile acid synthesis in humans has
a rapid diurnal variation that is asynchronous with cholesterol synthesis.
Gastroenterology, 2005;129(5):1445–53.
100. Lundasen, T., Galman, C., Angelin, B., and Rudling, M. Circulating intes-
tinal fibroblast growth factor 19 has a pronounced diurnal variation and
modulates hepatic bile acid synthesis in man. J Intern Med,
2006;260(6):530–6.
101. Pathak, P., Li, T., and Chiang, J.Y. Retinoic acid‐related orphan receptor
alpha regulates diurnal rhythm and fasting induction of sterol 12alpha‐
hydroxylase in bile acid synthesis. J Biol Chem, 2013;288(52):37154–65.
102. Hatori, M., Vollmers, C., Zarrinpar, A. et al. Time‐restricted feeding with-
out reducing caloric intake prevents metabolic diseases in mice fed a high‐
fat diet. Cell Metab, 2012;15(6):848–60.
103. Li, G., Xie, C., Lu, S. et  al. Intermittent fasting promotes white adipose
browning and decreases obesity by shaping the gut microbiota. Cell Metab,
2017;26(4):672–85 e4.
104. Vaz, F.M. and Ferdinandusse, S. Bile acid analysis in human disorders of
bile acid biosynthesis. Mol Aspects Med, 2017;56:10–24.
105. Hofman, M.K., Princen, H.M., Zwinderman, A.H., and Jukema, J.W.
Genetic variation in the rate‐limiting enzyme in cholesterol catabolism
(cholesterol 7alpha‐hydroxylase) influences the progression of atheroscle-
rosis and risk of new clinical events. Clin Sci, 2005;108(6):539–45.
106. Kajinami, K., Brousseau, M.E., Ordovas, J.M., and Schaefer, E.J.
Interactions between common genetic polymorphisms in ABCG5/G8 and
CYP7A1 on LDL cholesterol‐lowering response to atorvastatin.
Atherosclerosis, 2004;175(2):287–93.
107. Jiang, Z.Y., Han, T.Q., Suo, G.J. et  al. Polymorphisms at cholesterol
7alpha‐hydroxylase, apolipoproteins B and E and low density lipoprotein
receptor genes in patients with gallbladder stone disease. World J
Gastroenterol, 2004;10(10):1508–12.
108. Setchell, K.D.R., Schwarz, M., O’Connell, N.C. et al. Identification of a
new inborn error in bile acid synthesis: mutation of the oxysterol 7a‐
hydroxylase gene causes severe neonatal liver disease. J Clin Invest,
1998;102(9):1690–703.
 23:  BILE ACID METABOLISM IN HEALTH AND DISEASE 285
109. Ueki, I., Kimura, A., Nishiyori, A. et al. Neonatal cholestatic liver disease
in an Asian patient with a homozygous mutation in the oxysterol 7alpha‐
hydroxylase gene. J Pediatr Gastroenterol Nutr, 2008;46(4):465–9.
110. Tsaousidou, M.K., Ouahchi, K., Warner, T.T. et  al. Sequence alterations
within CYP7B1 implicate defective cholesterol homeostasis in motor‐neu-
ron degeneration. Am J Hum Genet, 2008;82(2):510–15.
111. Leitersdorf, E., Safadi, R., Meiner, V. et al. Cerebrotendinous xanthomato-
sis in the Israeli Druze: molecular genetics and phenotypic characteristics.
Am J Hum Genet, 1994;55(5):907–15.
112. Srivastava, A. Progressive familial intrahepatic cholestasis. J Clin Exp
Hepatol, 2014;4(1):25–36.
113. Zollner, G. and Trauner, M. Mechanisms of cholestasis. Clin Liver Dis,
2008;12(1):1–26, vii.
114. Strautnieks, S.S., Kagalwalla, A.F., Tanner, M.S. et al. Identification of a
locus for progressive familial intrahepatic cholestasis PFIC2 on chromo-
some 2q24. Am J Hum Genet, 1997;61(3):630–3.
115. Carlton, V.E., Harris, B.Z., Puffenberger, E.G. et al. Complex inheritance of
familial hypercholanemia with associated mutations in TJP2 and BAAT.
Nat Genet, 2003;34(1):91–6.
116. Gomez‐Ospina, N., Potter, C.J., Xiao, R. et al. Mutations in the nuclear bile
acid receptor FXR cause progressive familial intrahepatic cholestasis. Nat
Commun, 2016;7:10713.
117. Pauli‐Magnus, C., Lang, T., Meier, Y. et al. Sequence analysis of bile salt
export pump (ABCB11) and multidrug resistance p‐glycoprotein 3
(ABCB4, MDR3) in patients with intrahepatic cholestasis of pregnancy.
Pharmacogenetics, 2004;14(2):91–102.
118. Noe, J., Kullak‐Ublick, G.A., Jochum, W. et al. Impaired expression and
function of the bile salt export pump due to three novel ABCB11 mutations
in intrahepatic cholestasis. J Hepatol, 2005;43(3):536–43.
119. Farrell, G.C. and Larter, C.Z. Nonalcoholic fatty liver disease: from steato-
sis to cirrhosis. Hepatology, 2006;43(2 Suppl 1):S99–S112.
120. Cohen, J.C., Horton, J.D., and Hobbs, H.H. Human fatty liver disease: old
questions and new insights. Science, 2011;332(6037):1519–23.
121. Brunt, E.M. Pathology of nonalcoholic fatty liver disease. Nat Rev
Gastroenterol Hepatol, 2010;7(4):195–203.
122. Puri, P., Daita, K., Joyce, A. et al. The presence and severity of nonalco-
holic steatohepatitis is associated with specific changes in circulating bile
acids. Hepatology, 2017.
123. Lioudaki, E., Ganotakis, E.S., and Mikhailidis, D.P. Lipid lowering drugs
and gallstones: a therapeutic option? Curr Pharm Des, 2011;17(33):
3622–31.
124. Dyson, J.K., Hirschfield, G.M., Adams, D.H. et al. Novel therapeutic tar-
gets in primary biliary cirrhosis. Nat Rev Gastroenterol Hepatol, 2015;
12(3):147–58.
125. Dilger, K., Hohenester, S., Winkler‐Budenhofer, U. et al. Effect of ursode-
oxycholic acid on bile acid profiles and intestinal detoxification machinery
in primary biliary cirrhosis and health. J Hepatol, 2012;57(1):133–40.
126. Prieto, J., Garcia, N., Marti‐Climent, J.M. et  al. Assessment of biliary
bicarbonate secretion in humans by positron emission tomography.
Gastroenterology, 1999;117(1):167–72.
127. Beuers, U. Drug insight: mechanisms and sites of action of ursodeoxy-
cholic acid in cholestasis. Nat Clin Pract Gastroenterol Hepatol,
2006;3(6):318–28.
128. Poupon, R. Ursodeoxycholic acid and bile‐acid mimetics as therapeu-
tic agents for cholestatic liver diseases: an overview of their mecha-
nisms of action. Clin Res Hepatol Gastroenterol, 2012;36(Suppl
1):S3–12.
129. Pares, A., Caballeria, L., and Rodes, J. Excellent long‐term survival in
patients with primary biliary cirrhosis and biochemical response to ursode-
oxycholic acid. Gastroenterology, 2006;130(3):715–20.
130. European Association for the Study of the Liver. EASL Clinical Practice
Guidelines: management of cholestatic liver diseases. J Hepatol,
2009;51(2):237–67.
131. Chapman, R., Fevery, J., Kalloo, A. et al. Diagnosis and management of
primary sclerosing cholangitis. Hepatology, 2010;51(2):660–78.
132. Yoon, Y.B., Hagey, L.R., Hofmann, A.F. et al. Effect of side‐chain shorten-
ing on the physiologic properties of bile acids: hepatic transport and effect
on biliary secretion of 23‐nor‐ursodeoxycholate in rodents. Gastroenterology,
1986;90(4):837–52.
133. Yeh, H.Z., Schteingart, C.D., Hagey, L.R. et al. Effect of side chain length
on biotransformation, hepatic transport, and choleretic properties of cheno-
deoxycholyl homologues in the rodent: studies with dinorchenodeoxy-
cholic acid, norchenodeoxycholic acid, and chenodeoxycholic acid.
Hepatology, 1997;26(2):374–85.
134. Halilbasic, E., Fiorotto, R., Fickert, P. et al. Side chain structure determines
unique physiologic and therapeutic properties of norursodeoxycholic acid
in Mdr2‐/‐ mice. Hepatology, 2009;49(6):1972–81.
135. Fickert, P., Hirschfield, G.M., Denk, G. et  al. norUrsodeoxycholic acid
improves cholestasis in primary sclerosing cholangitis. J Hepatol,
2017;67(3):549–58.
136. Pellicciari, R., Fiorucci, S., Camaioni, E. et al. 6alpha‐ethyl‐chenodeoxy-
cholic acid (6‐ECDCA), a potent and selective FXR agonist endowed with
anticholestatic activity. J Med Chem, 2002;45(17):3569–72.
137. Hirschfield, G.M., Mason, A., Luketic, V. et al. Efficacy of obeticholic acid
in patients with primary biliary cirrhosis and inadequate response to urso-
deoxycholic acid. Gastroenterology, 2015;148(4):751–61 e8.
138. Kowdley, K.V., Luketic, V., Chapman, R. et al. A randomized trial of obet-
icholic acid monotherapy in patients with primary biliary cholangitis.
Hepatology, 2018;67(5):1890–902.
139. Nevens, F., Andreone, P., Mazzella, G. et al. A placebo‐controlled trial of
obeticholic acid in primary biliary cholangitis. N Engl J Med,
2016;375(7):631–43.
140. Neuschwander‐Tetri, B.A., Loomba, R., Sanyal, A.J. et  al. Farnesoid X
nuclear receptor ligand obeticholic acid for non‐cirrhotic, non‐alcoholic
steatohepatitis (FLINT): a multicentre, randomised, placebo‐controlled
trial. Lancet, 2015;385(9972):956–65.
141. Fonseca, V.A., Handelsman, Y., and Staels, B. Colesevelam lowers glucose
and lipid levels in type 2 diabetes: the clinical evidence. Diabetes Obes
Metab, 2010;12(5):384–92.
142. Shang, Q., Saumoy, M., Holst, J.J., Salen, G., and Xu, G. Colesevelam
improves insulin resistance in a diet‐induced obesity (F‐DIO) rat model by
increasing the release of GLP‐1. Am J Physiol Gastrointest Liver Physiol,
2010;298(3):G419–24.
143. Potthoff, M.J., Potts, A., He, T. et al. Colesevelam suppresses hepatic gly-
cogenolysis by TGR5‐mediated induction of GLP‐1 action in DIO mice.
Am J Physiol Gastrointest Liver Physiol, 2013;304(4):G371–80.
144. Patti, M.E., Houten, S.M., Bianco, A.C. et al. Serum bile acids are higher in
humans with prior gastric bypass: potential contribution to improved glu-
cose and lipid metabolism. Obesity (Silver Spring), 2009;17(9):1671–7.
145. Simonen, M., Dali‐Youcef, N., Kaminska, D. et al. Conjugated bile acids
associate with altered rates of glucose and lipid oxidation after Roux‐en‐Y
gastric bypass. Obes Surg, 2012;22(9):1473–80.
146. Albaugh, V.L., Flynn, C.R., Cai, S. et al. Early increases in bile acids post
Roux‐en‐Y gastric bypass are driven by insulin‐sensitizing, secondary bile
acids. J Clin Endocrinol Metab, 2015;100(9):E1225–33.
147. Ryan, K.K., Tremaroli, V., Clemmensen, C. et al. FXR is a molecular target
for the effects of vertical sleeve gastrectomy. Nature, 2014;509:183–8.
148. McGavigan, A.K., Garibay, D., Henseler, Z.M. et al. TGR5 contributes to
glucoregulatory improvements after vertical sleeve gastrectomy in mice.
Gut, 2015;66(2):226–34.
 TGR5 (GPBAR1)
24 in the Liver
Verena Keitel1, Christoph G.W. Gertzen2, Lina Spomer1, Holger Gohlke2,
and Dieter Häussinger1
1
Clinic for Gastroenterology, Hepatology and Infectious Diseases, University Hospital Düsseldorf, Medical
Faculty at Heinrich‐Heine‐University, Düsseldorf, Germany
2
Institute of Pharmaceutical and Medicinal Chemistry, Heinrich‐Heine University Düsseldorf, Düsseldorf,
Germany

INTRODUCTION intestine [15–20]. Thus, FXR acts as important metabolic regulator

Bile acids are synthesized in hepatocytes from cholesterol, lead-
ing to the formation of cholic acid (CA) and chenodeoxycholic
acid (CDCA) in humans, while in mice CDCA is mostly con-
verted to α‐muricholic acid (αMCA) and β‐muricholic acid
(βMCA) [1, 2]. These primary bile acids are subsequently con-
jugated at carbon 24 to either taurine or glycine and secreted
into bile, forming mixed micelles with phospholipids and cho-
lesterol, which are released into the small intestine following
food intake, where they facilitate absorption of dietary lipids as
well as cholesterol excretion [1–3]. In the intestine, bile acids
are deconjugated by microbial bile salt hydrolase (BSH) to form
unconjugated bile acids, which subsequently can be converted
into the secondary bile acids deoxycholic acid (DCA) and litho-
cholic acid (LCA) through 7α‐dehydroxylation [4, 5]. Bile acids
are actively reabsorbed in the terminal ileum and transported
back to the liver via the portal venous blood [6]. In the liver, bile
acids are taken up into hepatocytes, reconjugated with taurine or
glycine and again secreted across the canalicular hepatocyte
membrane into bile [6]. This enterohepatic circulation of bile
acids takes place 6–10 times per day in humans [3, 6]. Only a
small amount of secondary bile acids is excreted with the feces
and hence replaced by de novo synthesis of bile acids from cho-
lesterol in the liver [2].
Over the past decades, bile acids have emerged as important
signaling molecules that activate different classes of receptors,
allowing for a bile acid‐ and cell type‐specific response and
explaining the pleiotropic effects of bile acids in the organism
[7–11]. Bile acids were first identified as ligands for the nuclear
receptor (NR) farnesoid X receptor (FXR, NR1H4) in 1999
[12–14]. FXR transcriptionally regulates key genes involved in
bile acid, glucose, and lipid metabolism in both liver and
and the FXR ligand obeticholic acid (6α‐ethyl‐CDCA), which
has already been approved for the treatment of primary biliary
cholangitis (PBC), significantly improved histological features
of non‐alcoholic steatohepatitis (NASH) in a phase 2 trial [21,
22]. Further NRs activated by bile acids are the pregnane X
receptor (PXR, NR1I2) and the vitamin D receptor (VDR,
NR1I1) [23–25]. Several G protein‐coupled receptors (GPCRs)
are also responsive to bile acids, including different types of
muscarinic (acetylcholine) receptors (e.g. M2 and M3 recep-
tors) [26–29], formyl‐peptide receptors (FPR) [11, 30, 31], the
sphingosine‐1‐phosphate receptor 2 (S1PR2) [32–37] and the
Takeda G protein‐coupled receptor 5 (TGR5) [38–40]. TGR5 is
also known as G protein‐coupled bile acid receptor 1 (GPBAR1)
or membrane‐type bile acid receptor (M‐BAR) and was the first
GPCR identified as bile acid receptor in 2002/2003 [38, 40].
TGR5 EXPRESSION AND LOCALIZATION
IN RODENT AND HUMAN LIVER
The human TGR5 gene encompasses two exons and is located in
the chromosomal region 2q35 [41]. Exon 2 contains the entire
coding region of 993 base pairs, which translates to 330 amino
acids [40, 41]. In contrast, rat and murine Tgr5 genes are present
on chromosome 9q33 and chromosome 1qC3, respectively, and
contain coding regions of 990 base pairs each, resulting in pro-
teins of 329 amino acids [42]. The protein sequences of human,
bovine, rabbit, rat, and mouse TGR5 are highly conserved, with
amino acid identities of 82–91% [38, 40].
TGR5 mRNA is expressed almost ubiquitously in rodent and
human tissues [10, 39–40, 43, 44]. High mRNA levels were

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
 24:  TGR5 (GPBAR1) in the Liver 287
found in organs involved in the enterohepatic circulation and
excretion of bile acids such as liver, gallbladder, small and large
intestine, and kidney [38, 40, 42, 44–46]. Moreover, TGR5
mRNA was also present in heart, lung, spleen, CD14‐positive
monocytes, placenta, female and male reproductive organs,
adrenal glands, brain, and adipose tissue [38, 40, 42, 44–46].
Expression levels in whole liver tissue were lowest in mice,
higher in rats and highest in humans [39].
On the protein level, TGR5 was detected in different non‐
parenchymal cells of human and rodent livers, such as liver
sinusoidal endothelial cells (LSECs), liver‐resident mac-
rophages (Kupffer cells), biliary epithelial cells of small and
large intrahepatic bile ducts (cholangiocytes, BECs), BECs of
the extrahepatic bile duct, and activated hepatic stellate cells
(HSCs) (Figure  24.1) [10, 39, 47–52]. TGR5 immunofluores-
cence staining intensity was minimal or absent in hepatocytes
and quiescent HSCs (Figure 24.1) [10, 39, 48, 53]. In the gall-
bladder, TGR5 was localized in the epithelium and in smooth
muscle cells of the muscularis propria [54, 55].
On the subcellular level, TGR5 was predominantly found in
the plasma membrane [38, 40]. In polarized cells, such as gall-
bladder epithelial cells and cholangiocytes, TGR5 was localized
in the apical membrane and on the primary cilia, a sensory orga-
nelle extending from the apical plasma membrane into the bile
duct lumen [47, 49, 51, 54, 56–58]. Interestingly, TGR5 has also
been detected in the nuclear membrane of cholangiocytes; how-
ever, it is unknown whether the receptor can be activated intra-
cellularly [56].
NATURAL AND SYNTHETIC
TGR5 LIGANDS
Endogenous ligands of TGR5
A wide spectrum of hydrophobic endogenous ligands can acti-
vate TGR5 [59]. Those ligands comprise all known bile acids
and many hydrophobic neurosteroids, such as pregnanolone,
allopregnanolone, pregnanediol, and estradiol [45, 59, 60]. In
contrast to other bile acid receptors, TGR5 is rather promiscuous
as it recognizes bile acids regardless of their substitution pattern
and conjugation state [59]. However, the potency of bile acids
increases with the hydrophobicity of their cholane scaffold and
varies with the state and type of conjugation. As such, taurolitho-
cholic acid (TLCA; Figure 24.2), a secondary bile acid with only
a hydroxyl group in position 3 of the cholane scaffold and with
taurine conjugation, is the most potent natural agonist of TGR5
(EC50 = 0.29 μM, Figure 24.2); its unconjugated derivative, litho-
cholic acid (LCA), has a potency of EC50 = 0.58 μM (Figure 24.2)
[59]. Addition of hydroxyl groups in position 12 in the secondary
bile acid deoxycholic acid (DCA) or position 7 in the primary
bile acid CDCA increases the EC50 4‐fold and 23‐fold compared
to TLCA, respectively [59] (Figure 24.2). TGR5 also shows an
epimeric selectivity for bile acids with an α‐hydroxyl group in
position 7 of the cholane scaffold, as seen in the 5‐fold higher
efficacy of CDCA than ursodeoxycholic acid (UDCA)
(Figure  24.2) [59]. The effects of structural modifications on
agonistic potency are explained by a binding mode model.
TGR5‐binding mode model
Bile acids have been predicted to bind to the orthosteric site in
the GPCR TGR5 [61]. There, bile acids bridge the transmem-
brane helices (TM) 3 and 6 to activate the receptor, as is common
for GPCR agonists [62], while addressing additional TMs to
increase their efficacy [61] (Figure 24.3a,b). Taurine conjugation
increases the size of a bile acid compared to unconjugated bile
acids, which allows the bridging of residues R79 (extracellular
loop EL1) and Y240 (TM6) in TGR5 [61]. The salt‐bridge inter-
action between the negatively charged sulfonic acid moiety and
the positively charged R79 likely increases the affinity of those
bile acids towards TGR5. The localization of the sulfonic acid
moiety of taurine‐conjugated bile acids high up in the b­ inding
pocket of TGR5 may also explain why TGR5 can be activated by
cholestyramine‐bound bile acids [63]; it may still be possible for
bile acids to reach TM6 with their cholane ­scaffold even when
bound to cholestyramine with their acidic moiety. All bile acids
employ the 3‐hydroxyl groups of their cholane scaffold to form a
hydrogen bond to Y240, and this interaction is further stabilized
by a hydrogen bond to E169 (TM5; Figure  24.3a,b) [61]. The
hydrogen bond interaction with Y240 is vital for the activation of
the receptor because a Y240F variant, lacking the hydroxyl
group in the phenylalanine ring, is unresponsive to bile acids
[61]. Agonistic neurosteroids such as pregnanediol also utilize
their hydroxyl or carbonyl groups to interact with Y240 in TGR5
[61]. Lacking acidic groups, they mainly form additional hydro-
phobic contacts with Y89 in TM3 to bind to and activate TGR5
at a reasonable EC50 (e.g. pregnanediol EC50 = 0.58 μM [59]
(Figure 24.2)), allowing them to activate TGR5 in the brain. The
epimeric selectivity has been explained by a hydrogen bond for-
mation of CDCA’s 7α‐hydroxyl group to Y89 in TM3 of TGR5.
In contrast, due to the β‐configuration, UDCA cannot form such
a hydrogen bond with its 7‐hydroxyl group.
Synthetic TGR5 agonists and antagonists
To date, a range of TGR5 agonists with a nonsteroidal core are
known [64–68] (Figure 24.2). Here, an acidic or amide moiety is
linked to a system of three to four variably interconnected aro-
matic and aliphatic rings. The ring furthest from the acid or
amide moiety always contains a hetero‐atom. Although the
binding mode of nonsteroidal TGR5 agonists is unknown, it is pos-
sible that the hetero‐atom is necessary to form a hydrogen bond
to Y240 (TM6), which is crucial for the activation of TGR5 [69].
Intestinal specificity can be achieved either by addition of a qua-
ternary ammonia moiety to the synthetic TGR5 agonists (26a,
Figure 24.2) or by linking two agonists via polyethylene glycol
(15c, Figure 24.2). Ligands with quaternary ammonia moieties
can usually not be absorbed due to the permanent charge, while
the interconnected agonists are likely too spacious to pass the
cell membrane in the intestine and enter the bloodstream.
So far, despite the broad spectrum of ligands recognized by
TGR5, only one antagonist has been discovered (SBI‐115,
Figure  24.2) [70]. It is composed of two interconnected aro-
matic rings and an ethyl sulfo substituent and is slightly smaller
than the known agonists. Due to the lack of an acid moiety and
a steroidal core, the binding mode of this antagonist cannot
easily be inferred from the agonist binding mode.
288 THE LIVER:  TGR5‐DEPENDENT SIGNALING AND REGULATION OF TGR5 EXPRESSION AND LOCALIZATION

Figure 24.1  Localization of TGR5 in different liver cells. The receptor has been detected in sinusoidal endothelial cells (LSEC), in liver‐resident
macrophages (Kupffer cells, KCs), in biliary epithelial cells (BEC) and in activated hepatic stellate cells (HSCs). In contrast, little or no immunofluo-
rescence staining could be observed in hepatic parenchymal cells (HPCs, hepatocytes) and quiescent HSC. A–D Staining for TGR5 in rat liver is shown
in red. CD163, Reca‐1, CK‐19, α‐SMA, and GFAP served as marker proteins for KCs, LSECs, BECs, activated HSCs, and quiescent HSCs, respec-
tively. Panels (a–c) show livers from control animals. Panel (d) depicts a rat liver from an animal treated with carbon tetrachloride. Inset in (a): TGR5
staining is increased in KC after three days of common bile duct ligation. Inset in (d): Quiescent HSCs (GFAP, in green) showed no staining for TGR5
(red), which was present in LSECs (stained in blue). Panels (e–h) show immunofluorescence staining of TGR5 in human liver (e–g) and human gall-
bladder (h). CD163 served as marker protein for KCs, while CK‐7 was used to visualize BECs. MRP2 and Na/K‐ATPase were used to stain the apical
and basolateral membrane of gallbladder epithelial cells, respectively (h). Panels (i–m) show staining of TGR5 in BECs. TGR5 is localized in the pri-
mary cilium of cultured murine cholangiocytes (i, j) and of a human cholangiocarcinoma cell line (TFK‐1, (k)). Phalloidin was used to stain F‐actin fila-
ments (i) while acetylated α‐tubulin (α‐tub) served as marker protein for primary cilia (j, k). Panels (l) and (m) depict reduced fluorescence intensity
for TGR5 in BECs of a PSC liver (l) and an increase in fluorescence staining of TGR5 in tumor cells of human cholangiocarcinoma (m). Bars = 10 μm.

TGR5‐DEPENDENT SIGNALING can also be activated [10, 52, 57, 58]. In the esophageal
AND REGULATION OF TGR5 adenocarcinoma cell line FLO, a coupling of TGR5 with
EXPRESSION AND LOCALIZATION Gα q and Gα i3 upon activation by taurodeoxycholic acid
(TDCA) has been shown [71]. In ciliated and nonciliated
cholangiocytes, different effects of TGR5 agonists on prolif-
TGR5 downstream signaling eration were observed, which have been linked to TGR5
Bile acid binding to TGR5 leads to Gαs‐mediated activation coupling to either Gαs or Gαi [56]. TGR5 thus shows a func-
of the adenylyl cyclase/cyclic AMP (cAMP) signaling path- tional selectivity that may be related to the occurrence of
way [38]. In addition, other cell‐specific signaling pathways different active conformations and to subcellular
 24:  TGR5 (GPBAR1) in the Liver 289

Figure 24.2  Important TGR5 ligands. The table shows the substitution pattern for selected natural bile acid agonists with their EC50 values [59].
In addition, a neurosteroid agonist, pregnanediol, an antagonist, SBI‐115, two intestinal agonists, 26a and 15c, and three synthetic agonists, 23g,
18, and INT‐777, are shown [64–68, 70].

localization [56, 72]. Yet, in transfected HEK293 cells or
colonocytes, after activation by endogenous agonists, TGR5
interacts neither with GPCR kinases 2, 5, or 6 nor with β‐
arrestins 1 or 2 and, accordingly, shows no trafficking to
endosomes and no desensitization even after repeated stimu-
lation [73].
Ligand binding to TGR5 not only triggers an elevation of
intracellular cAMP but also of intracellular calcium levels, pro-
motes activation of different ion channels, regulates gene
expression, and induces mitochondrial fission [38, 45, 48, 50,
55, 74–77]. Moreover, stimulation of TGR5 by bile acids acti-
vates a broad range of kinase signaling pathways, including pro-
tein kinase A (PKA), protein kinase B (AKT), epidermal growth
factor receptor (EGFR), mammalian target of rapamycin
(mTOR), Rho kinase, and extracellular signal‐regulated kinase
(ERK) pathways [38, 52, 55, 56, 76, 78].
Regulation of TGR5 through di‐ and 
oligomerization
It is now well established that GPCRs can form homo‐ and het-
ero‐oligomers, and oligomer formation can affect a broad range
of biological functions [79]. For TGR5, a combined strategy
using cell biology, multiparameter fluorescence image spectros-
copy (MFIS) for quantitative fluorescence resonance energy
transfer (FRET) analysis, and integrative modeling was applied
to obtain structural information on dimerization and oligomeri-
zation of TGR5, fused to fluorescent proteins, in live cells [80].
The correlation of FRET parameters with increasing FRET
acceptor‐to‐donor ratio showed that TGR5 wild‐type forms
higher-order oligomers, a process disrupted in a TGR5 Y111A
variant; residue 111 is located in TM3 within the highly con-
served D/ERY motif. From the ratio dependence it was
290 THE LIVER:  FUNCTION OF TGR5 IN DIFFERENT LIVER CELLS
concluded that higher-order oligomers of TGR5 wild‐type are
formed from dimers interacting with other dimers. In contrast,
dimers are the largest unit suggested for the TGR5 Y111A vari-
ant. Higher-order oligomers likely have a linear arrangement
with interaction sites involving TM1 and helix 8 in the cytoplasm
as well as TM5 (Figure 24.3c). The latter interaction was sug-
gested to be disrupted by the Y111A mutation. The importance
of helix 8 for TGR5 dimerization, together with reports that
GPCR dimerization in the endoplasmic reticulum is obligatory
for their membrane trafficking, might explain why a membrane‐
proximal, C‐terminal helix is structurally required for plasma
membrane localization and function of TGR5 [69]. Finally, time‐
series MFIS–FRET analysis before, at the time point of, and
after stimulation of TGR5 with taurocholate (TCA, Figure 24.2),
a bile acid less cytotoxic than TLCA in live cells, indicated that
TCA, and subsequent G-protein coupling, does not influence the
oligomerization state of TGR5 wild‐type and Y111A [80], an
observation supported by the literature [81].
Regulation of TGR5 expression and function
Mechanisms regulating TGR5 expression, localization, and
function are mostly elusive to date. Reduced TGR5 mRNA and
protein levels were observed in cultured rat astrocytes after stim-
ulation with ammonia (NH4Cl; 0.5–5 mM over 72 hours) [45].
Moreover, TGR5 mRNA expression was significantly lower in
cortical brain tissue obtained from patients with hepatic enceph-
alopathy (HE) as compared to samples from patients without HE
[45]. Incubation of cultured rat astrocytes or human macrophages
with the TGR5 ligands 5β‐pregnan‐3α‐ol‐20‐one (pregnanolone)
(10 μM, 72 hours), 5α‐pregnan‐3α‐ol‐20‐one (allopregnanolone)
and 5α‐pregnan‐3β‐ol‐20‐one (isopregnanolone) (25 μM, 12 or
72 hours), respectively, resulted in a significant downregulation
of TGR5 mRNA levels as compared to vehicle‐treated cells [45,
46]. This suggests that continuous ligand stimulation and the
resulting downregulation of TGR5 mRNA expression represents
an important mechanism of receptor desensitization, especially
since TGR5 is not retrieved from the plasma membrane in
response to repetitive ligand activation [45, 46, 73].
In contrast, an upregulation of TGR5 mRNA levels has been
described in the frontal brain cortex of mice following azoxym-
ethane‐induced liver failure [82]. Moreover, an increase in
TGR5 immunofluorescence staining was observed in Kupffer
cells in rat liver following common bile duct ligation (CBDL)
for 72 hours [48]. The molecular mechanisms regulating TGR5
mRNA expression are yet unknown.
FUNCTION OF TGR5 IN DIFFERENT
LIVER CELLS
TGR5 in liver sinusoidal endothelial cells
Stimulation of TGR5 in cultivated liver sinusoidal endothelial
cells (LSECs) from rat liver resulted in activation of adeny-
lyl cyclase, elevation of intracellular cAMP, subsequent
PKA‐mediated serine phosphorylation of endothelial nitric oxide
synthase (eNOS), and increased generation and release of nitric
oxide (NO) [10, 39, 50]. Furthermore, TGR5 activation triggered
an AKT‐dependent serine phosphorylation of cystathionine‐γ‐
lyase (CSE), an enzyme essential for the generation of the vaso-
dilatory molecule hydrogen sulfide (H2S) [39, 83, 84]. On the
genomic level, ligand binding to TGR5 promoted upregulation
of eNOS and CSE mRNA levels, while it suppressed expression
of endothelin 1 (ET‐1) (Figure  24.4) [50, 83–85]. Therefore,
TGR5 activation within LSECs favors the generation and
release of vasodilatory molecules (NO, H2S) while it inhibits the
expression of ET‐1, thus reducing HSC contractility as well as
portal pressure (Figure 24.4) [39, 83].
TGR5 in hepatic stellate cells
TGR5 mRNA and protein expression were negligible in freshly
isolated, quiescent rat HSCs [53]. Immunofluorescent staining
for TGR5 was negative in glial fibrillary acidic protein (GFAP)‐
positive quiescent HSC in rodent livers [48, 53]. However,
TGR5 mRNA and protein levels increased over days in cul-
tured, myofibroblast‐like activated HSCs and in damaged livers
in vivo (Figures 24.1 and 24.4) [39, 48, 53]. Cyclic AMP was
shown to desensitize the ETA receptor towards ET‐1 in activated
HSCs through receptor internalization [86]. Therefore, TGR5
stimulation in activated HSCs may counteract contraction of
these cells and ameliorate portal hypertension.
TGR5 in Kupffer cells and macrophages
TGR5 ligand activation exerts strong anti‐inflammatory effects in
Kupffer cells as well as bone marrow‐derived macrophages
(BMDMs), where activation of the receptor inhibits the expres-
sion and secretion of proinflammatory cytokines [48, 87, 88].
Activation of TGR5 through elevation of cAMP inhibits phos-
phorylation of IκBα and thus prevents translocation of the nuclear
factor (NF)‐κB p65 into the nucleus [39, 89]. Furthermore, via
PKA‐mediated phosphorylation of nucleotide‐binding domain
(NB) and leucine‐rich repeat (LRR)‐containing receptor protein 3
(NLRP3), TGR5 suppresses caspase 1‐dependent maturation of
proinflammatory cytokines, such as interleukin 1β (IL‐1β) and
IL‐18 [39, 88]. Ligand activation of TGR5 also inhibits chemokine
expression via a PKB (AKT)–mTOR signaling pathway, result-
ing in the increased expression of the CCAAT/enhancer binding
protein β (C/EBPβ) isoform liver inhibitory protein (LIP) [39, 76,
90]. Other anti‐inflammatory mechanisms triggered by TGR5
activation comprise reduced macrophage migration and phagocy-
tosis as well as preferential differentiation of monocytes into an
anti‐inflammatory, regulatory phenotype (Figure  24.5) [76, 89,
91–94]. Similar to FXR, TGR5 promotes a variety of anti‐inflam-
matory effects [39].
TGR5 in the biliary epithelium
and the gallbladder
TGR5 is expressed in cholangiocytes of the small and large
intrahepatic ducts as well as the extrahepatic bile duct, where
 24:  TGR5 (GPBAR1) in the Liver 291

(a) (c)

(b)

Figure 24.3  Interactions of TGR5 with TLC and TGR5 oligomer formation. (a) Binding mode model of TLC (orange sticks and transparent
surface) in TGR5 (gray, cartoon) [61]. Important interacting residues are shown as sticks. TLC bridges TM3 and TM6 by interacting with Y240 and
Y89, which are important for TGR5 activation. E169 and R79 are important for the efficacy of TLC. (b) Two‐dimensional diagram showing interac-
tions of TLC within TGR5. Hydrogen bonding and salt‐bridge interactions are shown as dashed lines, hydrophobic interactions are shown in green.
(c) Schematic of TGR5 oligomerization [80]. TGR5 forms dimers involving TM1 and helix 8 (1–8). Then, oligomerization can occur from dimers
either via the TM4–TM5 interface (4–5) or the TM5–TM6 interface (5–6).
292 THE LIVER:  FUNCTION OF TGR5 IN DIFFERENT LIVER CELLS
the receptor is localized in the apical plasma membrane and the
primary cilium [39, 43, 47–49, 51, 52, 56–58]. Ligand binding
to TGR5 activates adenylyl cyclase, triggers cAMP elevation,
and subsequent stimulation of the cystic fibrosis transmembrane
conductance regulator (CFTR, ABCC7) resulting in increased
chloride secretion (Figure  24.6) [39, 49, 54, 57]. Chloride is
subsequently exchanged against bicarbonate via the anion
exchanger 2 (AE2, SLC4A2), leading to bicarbonate‐rich
choleresis and formation of a protective bicarbonate layer
known as the bicarbonate umbrella [39, 47, 49, 54, 67, 95–97].
Cyclic AMP not only enhances transport activity but also trig-
gers the insertion of CFTR and AE2 into the apical plasma
membrane from intracellular vesicles, resulting in increased bil-
iary secretion [47, 49, 54, 58]. Absence of TGR5 renders chol-
angiocytes more susceptible towards bile acid‐induced toxicity
[43, 47, 52, 96]. Activation of TGR5 also triggers increased
expression and phosphorylation of the junctional adhesion
Figure 24.4  Function of TGR5 in liver sinusoidal endothelial cells
(LSECs) and activated hepatic stellate cells (HSCs). Ligand binding to
TGR5 stimulates adenylyl cyclase and the generation of cyclic AMP
(cAMP), which in turn triggers activation of further downstream targets
comprising cAMP response element binding protein (CREB) and pro-
tein kinase A (PKA). CREB promotes increased expression of endothe-
lial nitric oxide synthase (eNOS) and cystathionine‐γ‐lyase (CSE),
resulting in increased generation of NO and hydrogen sulfide (H2S).
This is further enhanced through PKA‐ and AKT‐mediated activation
of eNOS and CSE. Additionally, TGR5 suppresses expression of
endothelin‐1 (ET‐1). In activated HSCs, TGR5 inhibits ET‐1 signaling
and, thus, contraction through cAMP‐dependent internalization of the
endothelin receptor. In summary, TGR5 activation promotes secretion
of vasodilators and inhibits release of vasoconstrictors, leading to
reduced portal pressure. Modified from [39] with permission of Georg
Thieme Verlag KG Stuttgart.
molecule A (JAM‐A), thereby reducing paracellular permeabil-
ity and protecting liver tissue from bile leakage (Figure  24.6)
[98]. TGR5 can exert opposing effects on cholangiocyte prolif-
eration dependent on the receptor’s subcellular localization at
time of activation [39, 47]. Ligand binding to TGR5 on the pri-
mary cilia of simian virus 40‐transformed human cholangio-
cytes (H69 cells) leads to coupling to an inhibitory Gαi protein,
reduces intracellular cAMP concentrations and impairs cell pro-
liferation [39, 47, 56]. Stimulation of TGR5 in non‐ciliated H69
cells promotes cell proliferation through elevation of cAMP [47,
56]. In contrast, incubation of murine cholangiocytes with bile
acids or synthetic TGR5 agonists results in generation of reac-
tive oxygen species (ROS), stimulation of Src kinase and matrix
metalloproteinases, transactivation of the EGFR, phosphoryla-
tion of ERK1/2, and significantly increased cell proliferation
[39, 43, 47, 52]. This process is independent of adenylyl cyclase
activation [52]. In contrast, TGR5 attenuates CD95 ligand‐
induced apoptosis via stimulation of adenylyl cyclase, elevation
of cAMP and PKA activation (Figure 24.6) [47, 52].
In the gallbladder epithelium, TGR5 activation also enhances
CFTR‐mediated chloride secretion, while in gallbladder smooth
muscle cells TGR5 stimulation through a cAMP–PKA pathway
results in the opening of ATP‐sensitive potassium channels
(KATP) and subsequent muscle cell relaxation [54, 55]. Studies
with full‐thickness gallbladder tissue preparations and exposure
of the epithelial cell layer to the bile acids also resulted in
smooth muscle cell relaxation [55]. Thus, absorption of bile
acids through the epithelium and subsequent TGR5 activation
Figure 24.5  TGR5 exerts anti‐inflammatory effects in Kupffer cells
and macrophages. TGR5‐dependent elevation of cAMP impairs IκB
kinase activity and phosphorylation of IκB, thereby stabilizing IκB as
well as NF‐κB p65 within the cytoplasm. This results in decreased tran-
scriptional activity of NF‐κB and, thus, reduced expression of proin-
flammatory cytokines. TGR5 activation impairs caspase 1‐dependent
cleavage of proinflammatory cytokines through PKA‐dependent phos-
phorylation of NLRP3. Moreover, TGR5 activation reduces chemokine
expression and secretion via an AKT–mTOR–LIP signaling pathway.
In summary, TGR5 stimulation attenuates expression of proinflamma-
tory cytokines and chemokines and stabilizes or increases the transcrip-
tion of anti‐inflammatory molecules. Modified from [39] with
permission of Georg Thieme Verlag KG Stuttgart.
 24:  TGR5 (GPBAR1) in the Liver 293
Figure 24.6  Function of TGR5 in biliary epithelial cells. Ligand bind-
ing to TGR5 triggers activation of a stimulatory G protein and adenylyl
cyclase (AC). Through elevation of intracellular cyclic AMP (cAMP)
TGR5 promotes chloride and subsequent bicarbonate excretion via
CFTR and the anion exchanger 2 (AE2), resulting in the formation of the
bicarbonate umbrella. Furthermore, cAMP‐dependent activation of pro-
tein kinase A (PKA) leads to serine/threonine phosphorylation of the
CD95 receptor and inhibition of apoptosis. Stimulation of TGR5 induces
expression and phosphorylation of the junctional adhesion molecule A
(JAM‐A) and regulates tight junction (TJ) integrity and paracellular per-
meability. Ligand binding to TGR5 can increase the levels of reactive
oxygen species (ROS) independent of AC activation, resulting in Src
kinase activation, matrix metalloproteinase (MMP)‐dependent shedding
of the epidermal growth factor (EGF), transactivation of the EGF receptor
(EGFR), followed by phosphorylation of MAP kinase ERK1/2 and
increased cell proliferation, while activation of TGR5 in primary cilia
inhibited cell proliferation. Modified from [47, 49, 52].
on gallbladder smooth muscle cells is a rapid and local mecha-
nism for gallbladder filling [55, 66]. Moreover, bile acids medi-
ate gallbladder filling not only through TGR5 but also through
the ileal FXR downstream effector fibroblast growth factor 19
(FGF19, which is FGF15 in mice) [55, 95, 99].
INSIGHTS FROM TGR5‐KNOCKOUT
MICE AND APPLICATION OF TGR5
LIGANDS TO RODENTS
Targeted deletion of TGR5 in mice is not associated with an
obvious phenotype or development of spontaneous liver disease
[42, 44]. However, TGR5‐knockout mice have a smaller bile
acid pool size with reduced levels of the FXR‐antagonistic bile
acid TβMCA and an increase in the fraction of TCA and TDCA,
which may be attributed to lower expression levels of Cyp7b1
and suppression of bile acid synthesis via the alternative path-
way [43, 95, 100, 101]. This further supports an overlap of
TGR5 and FXR functions, since reduced concentrations of
TβMCA may disinhibit FXR signaling in the absence of TGR5
[43, 101].
Role of TGR5 in gallbladder and biliary
diseases: lessons from mouse models
As described above, TGR5‐knockout mice not only have
reduced bile flow but also a very small gallbladder volume,
which is explained by impaired smooth muscle cell relaxation in
absence of TGR5 [43, 55, 95]. Oral administration of TGR5
ligands to wild‐type mice (6α‐ethyl‐23(S)‐methyl‐cholic acid
(INT‐777, EMCA, EC50 = 0.82 μM), or the highly potent small
molecules compound 18 (EC50 = 25 nM) or compound 23g
(EC50 = 0.72 nM)) (Figure 24.2) induced biliary bile flow and
promoted gallbladder filling with an increase in gallbladder size
to up to 231% (compound 23g, Figure 24.2) [43, 66, 67, 95].
The latter compound (23g), which is a 4‐phenoxy‐nicotinamide
derivative and thus a non‐bile acid TGR5 ligand, accumulated in
plasma, bile, and gallbladder tissue in mice, underscoring a
direct, local effect of this compound on TGR5 in gallbladder
smooth muscle cells [67]. Intestinal‐specific TGR5 ligands with
low systemic availability (such as compound 15c, Figure 24.2)
failed to increase gallbladder volume significantly [65]. This
suggests that TGR5 ligands need to accumulate in gallbladder
tissue either through absorption from the gallbladder lumen or
through accumulation in plasma in order to facilitate gallbladder
filling. Whether systemic application of a TGR5 ligand, which
is not secreted into bile, is sufficient to trigger gallbladder
smooth muscle cell relaxation needs to be determined.
Interestingly, mice with targeted deletion of TGR5 are pro-
tected from cholesterol gallstone formation when administered
a lithogenic diet [42, 43]. Absence of TGR5 from gallbladder
smooth muscle cells may prevent bile acid‐dependent reduction
in contractility, while overexpression of TGR5, as observed in
human tissue from patients with gallstone disease, may promote
gallbladder hypomotility, a prerequisite for gallstone formation
[54, 55]. Whether this mechanism plays a role in human gall-
stone disease needs further investigation. However, monitoring
of gallbladder function is warranted if TGR5 agonists are evalu-
ated for clinical applications [39].
Targeted deletion of TGR5 renders mice more susceptible
towards different types of cholestatic liver injury [39, 43, 47, 52].
TGR5‐knockout mice display increased liver damage as deter-
mined by elevated liver enzymes or by histology in response to
cholic acid feeding (CA, 0.5% for 7 days or 1% for 5 days) or
common bile duct ligation (CBDL) for up to 7 days [39, 43, 47,
52, 63]. Gene deletion of TGR5 impaired cholangiocyte and
hepatocyte proliferation in response to cholestasis in vivo [47, 52,
63]. In isolated, cultured murine cholangiocytes, TGR5 was
essential for bile acid‐induced proliferation, which was mediated
by a TGR5–ROS–Src–MMP–EGFR–ERK1/2 signaling cascade
(Figure  24.6) [52]. In contrast, the mechanisms responsible for
reduced hepatocyte proliferation in TGR5‐knockout mice are as
yet unknown [43]. However, impaired hepatocyte proliferation
was also detected after partial hepatectomy (PHx) in TGR5‐
knockout mice and may be attributed in part to higher hepatic bile
acid concentrations as well as a more hydrophobic bile acid pool
composition in the absence of TGR5 [43, 63]. Liver injury after
PHx could be attenuated in TGR5-knockout through bile acid
depletion with cholestyramine (2%), underscoring the hypothesis
that elevated bile acid levels as well as altered composition of the
bile acid pool underlie the observed phenotype [43, 63].
294 THE LIVER:  HUMAN LIVER DISEASE
Role of TGR5 in inflammatory and metabolic
liver disease: lessons from mouse models
TGR5 is an important negative regulator of inflammation both
systemically but also locally in the liver as described earlier.
TGR5‐deficient mice were more susceptible towards intraperi-
toneal injection of lipopolysaccharide resulting in elevated
serum levels of alanine (ALT) and aspartate (AST) aminotrans-
ferases and inflammatory cytokines, such as IL‐1β, compared to
their wild‐type littermates [39, 43, 87]. Increased inflammatory
infiltrates as well as enhanced hepatocyte apoptosis were pre-
sent in the livers of TGR5‐knockout mice as compared to wild‐
type animals [39, 43, 87]. Simultaneous intraperitoneal injection
of both LPS and TLCA or LPS and the synthetic TGR5 ligand
INT‐777 attenuated LPS‐mediated elevation of IL‐1β and IL‐18,
an effect which was abrogated in mice deficient for TGR5 or the
NLRP3 inflammasome [39, 88].
Activation of TGR5 has been shown to improve various
aspects of the metabolic syndrome in different animal models
[11, 43, 65–67, 74, 89, 102]. Therefore, TGR5 has emerged as
an attractive drug target for metabolic diseases including obe-
sity, insulin resistance, atherosclerosis, and non‐alcoholic fatty
liver disease (NAFLD) leading to the development of different
synthetic TGR5 agonists [11, 43, 65–68, 90, 103, 104].
Steatohepatitis is characterized by inflammation and infiltra-
tion of Ly‐6Chigh monocytes, and the differentiation of mac-
rophages towards an inflammatory phenotype is a typical
finding in progressive disease [39, 105]. Application of a dual
FXR and TGR5 agonist (6α‐ethyl‐24‐nor‐5β‐cholane‐3α,7α,23‐
trio‐23‐sulfate sodium salt, INT‐767) to obese db/db mice for
six weeks reduced levels of Ly‐6Chigh and increased levels of
Ly‐6Clow monocytes, resulting in a significant improvement of
hepatic inflammation, steatosis, and hepatocyte ballooning as
well as profibrotic gene expression [39, 106]. Stimulation of
TGR5 with the synthetic agonist INT‐777 in mice on a high‐fat
diet reduced fat mass and obesity and also triggered intestinal
secretion of glucagon‐like peptide 1 (GLP‐1) thereby enhancing
glucose tolerance [43, 74, 75]. The reduction of body weight
and the GLP‐1‐mediated improvement of glucose tolerance
were also observed with other systemically available non‐bile
acid TGR5 agonists [66, 67, 103]. Furthermore, INT‐777
administration attenuated hepatic steatosis through reduction of
hepatic fatty acid and triglyceride concentrations [43, 74].
Inhibition of hepatic and adipose tissue inflammation improved
insulin sensitivity, enhanced energy expenditure in brown adi-
pose tissue and skeletal muscle, as well as increased lipolysis,
mitochondrial biogenesis, and mitochondrial fission. Induction
of beiging and thermogenesis in white adipose tissue may thus
underlie the beneficial effects of TGR5 agonist treatment on
obesity and steatohepatitis [74–76, 90, 102].
Interestingly, FXR and TGR5 ligands exert overlapping and
complementary beneficial effects on various aspects of the meta-
bolic syndrome including NAFLD. While FXR agonists regulate
lipid and glucose homeostasis in the liver directly as well as indi-
rectly through intestinal induction and secretion of FGF15/19,
TGR5 agonists promote anti‐inflammatory effects in the liver and
adipose tissue, induce energy expenditure as well as lipolysis and
thermogenesis in skeletal muscle and white adipose tissue, respec-
tively, and trigger GLP‐1 secretion from the intestine [75, 107].
Role of TGR5 in progressive liver disease:
lessons from rodent models
Portal hypertension represents a common complication of
advanced liver fibrosis and liver cirrhosis and is associated with
anatomical and biochemical changes in both LSECs and HSCs
[39, 108, 109]. Loss of fenestration, increased deposition of
extracellular matrix in the space of Disse, reduced generation
and secretion of vasodilatory mediators as well as increased
expression and release of profibrotic molecules, such as trans-
forming growth factor β (TGF‐β) and platelet‐derived growth
factor‐C (PDGF‐C), are characteristics of LSEC dysfunction in
portal hypertension [39, 109]. HSC activation and transforma-
tion into myofibroblast‐like cells is typically associated with
loss of vitamin A granules, production of extracellular matrix
and increased expression of alpha‐smooth muscle actin (α‐
SMA) and also of TGR5 [39, 53, 109]. TGR5 modulates portal
pressure and hepatic microcirculation under physiological con-
ditions [39, 50]. TGR5 activation by a synthetic agonist BAR501
in mice with hepatic fibrosis induced by carbon tetrachloride
administration over nine weeks improved portal hypertension
significantly [43, 85]. Reduced expression of ET‐1 and increased
expression and activity of CSE, which is essential for the gen-
eration of the vasodilatory molecule hydrogen sulfide, underlie
this beneficial effect [39, 85]. Furthermore, cAMP‐mediated
internalization of the ETA receptor leads to reduced responsive-
ness of activated HSCs towards ET‐1 and may also contribute to
the reduction in portal pressure [86]. Despite the improvement
in portal hypertension, BAR501 did not reverse the increased
expression levels of profibrogenic genes, such as TGF-β1,
collagen1α1, and α‐SMA [85]. Whether TGR5 agonists can
improve liver fibrosis or even cirrhosis in other models of liver
damage is unknown.
HUMAN LIVER DISEASE
Role of TGR5 in biliary diseases
Genome‐wide association studies suggested an association of
the TGR5 gene locus with ulcerative colitis and primary scleros-
ing cholangitis (PSC) [110, 111]. Resequencing of TGR5 in
patients with PSC and healthy blood donors failed to support a
role for TGR5 coding sequence (CDS) variants in the disease
pathogenesis [41, 47, 49, 112]. Nonsynonymous variants in the
TGR5 CDS were rare and variants on both alleles were not
­identified in any of the patients or controls [41, 47, 49, 112].
However, a common polymorphism rs11554825 was identified
within the noncoding first exon of TGR5 and in almost com-
plete linkage disequilibrium with a second polymorphism
rs3731859 within the 5′ untranslated region (UTR), which may
affect TGR5 promotor activity and, thus, TGR5 mRNA expres-
sion levels [41, 47, 49]. TGR5 mRNA levels were lowest in
­individuals homozygous for the G‐allele at rs3731859, which
was linked to the presence of the C‐allele at rs11554825 [41].
Furthermore, presence of the risk allele at rs11554825 (C‐allele)
was significantly associated to PSC across four cohorts (odds
ratio 1.14) [39, 41]. Reduced TGR5 mRNA levels impair
 24:  TGR5 (GPBAR1) in the Liver 295
TGR5‐mediated cytoprotective effects and render cholangio-
cytes more susceptible towards bile acid‐induced cell damage,
thereby contributing to the development or progression of bil-
iary diseases such as PSC and PBC [39, 47, 52]. Interestingly,
cholangiocytes in livers of patients with PSC showed a dimin-
ished immunofluorescence staining intensity as compared to
control livers independent of the common TGR5 rs11554825
polymorphism (Figure 24.1  l) [39, 47, 49]. Furthermore, a simi-
lar reduction of TGR5 fluorescence staining was observed in
cholangiocytes in livers of Mdr2 (Abcb4)‐knockout mice, which
serve as a model for PSC [39, 43, 47, 49, 113]. The mechanisms
and timing underlying TGR5 downregulation in PSC are cur-
rently under investigation.
In contrast to sclerosing cholangitis, an overexpression of
TGR5 mRNA has been observed in intrahepatic cholangiocarci-
noma (iCCA) as well as in extrahepatic, perihilar CCA com-
pared to tissue from the non‐malignant resection margins [47,
52, 114]. On the cellular level, an approximately threefold
increase of TGR5 immunofluorescence staining intensity was
measured in iCCA cells compared to cholangiocytes from the
non‐malignant surrounding tissue (Figure  24.1m), while the
staining intensity of cytokeratin 19 was unchanged [39, 47, 52].
Incubation of CCA‐derived cell lines (EGI‐1 and TFK‐1) with
different TGR5 agonists (TLCA or synthetic ligands) triggered
transactivation of the EGFR as well as ERK1/2 phosphoryla-
tion, leading to increased cell proliferation, which was attenu-
ated by siRNA knockdown of TGR5 [39, 47, 52]. Stimulation of
TGR5 also increased migration and invasiveness of CCA cell
lines, which again was abolished by targeted deletion of TGR5
[47, 52, 114].
An increased expression of TGR5 has also been detected in
cystic cholangiocytes in polycystic liver disease (PLD) [39, 43,
47, 70, 113]. Activation of TGR5 in cholangiocytes triggered
cell proliferation and cyst growth through elevation of cAMP, a
process which could be abolished in vitro by application of a
novel TGR5 inhibitor (m‐tolyl 5‐chloro‐2‐(ethylsulfonyl)‐
pyrimidine‐4‐carboxylate; SBI‐115, Figure 24.2) or attenuated
in vivo by targeted deletion of TGR5 [39, 70]. Thus, in PLD and
CCA inhibition of TGR5 signaling may offer novel therapeutic
options in these difficult to treat disorders [39, 43, 47, 70].
Role of TGR5 in extrahepatic complications
of chronic liver disease: hepatic
encephalopathy
TGR5 is expressed in virtually all cell types in the central nerv-
ous system and can even be recovered in synapses [45, 115].
Here, TGR5 acts as a neurosteroid receptor and its stimulation
triggers activation of adenylate cyclase, leads to elevation of
intracellular calcium levels, and induces the formation of ROS
[45]. In hepatic encephalopathy, which is seen as the clinical
manifestation of a low‐grade cerebral edema with oxidative
stress [116], TGR5 is strongly downregulated, which may
reflect a compensatory response with regard to oxidative stress
[45]. In microglial cells, stimulation of TGR5 with pregna-
nolone inhibits expression of proinflammatory cytokines and
thus exerts anti‐inflammatory and neuroprotective effects in
hepatic encephalopathy [82, 117].
SUMMARY AND PERSPECTIVE
Bile acid effects in liver are mediated through different bile
acid receptors, including both NRs and GPCRs. TGR5
(GPBAR1) is localized in different nonparenchymal liver cells,
where the receptor modulates liver microcirculation, inflam-
matory response, and biliary function. In humans, a role for
TGR5 has been suggested in biliary diseases such as PBC and
PSC, where reduced levels of TGR5 may accelerate disease
progression. Furthermore, an overexpression of the receptor
has been reported in malignant and cystic cholangiocytes in
human cholangiocarcinoma and polycystic liver disease,
respectively. In both diseases, activation of the receptor pro-
motes disease progression and, hence, disruption of TGR5
signaling may prove beneficial in these conditions. Mice with
targeted deletion of TGR5 develop more severe liver damage
with increased parenchymal and biliary injury, reduced adap-
tive hepatocyte and cholangiocyte proliferation as well as
increased immune activation in response to bile acid feeding,
partial hepatectomy, common bile duct ligation, or treatment
with lipopolysaccharide. Furthermore, TGR5 activation
improves many aspects of the metabolic syndrome, including
steatohepatitis, glucose homeostasis, adipose tissue inflamma-
tion, atherosclerosis, and kidney injury. Administration of
TGR5 agonists also ameliorates obesity through increased
energy expenditure in brown adipose tissue and skeletal muscle
as well as through enhanced lipolysis and mitochondrial bio-
genesis in white adipose tissue. Despite these beneficial effects,
TGR5 ligands promote gallbladder filling via gallbladder
smooth muscle relaxation, which may lead to abdominal dis-
comfort, gallstone disease, and its potential complications, all
of which may limit the use of systemically available TGR5
agonists. Intestinal‐specific TGR5 ligands promote GLP‐1
secretion, improve glucose homeostasis, and are devoid of the
unwanted side‐effects on gallbladder function; however, many
of the beneficial effects of targeting systemic TGR5 will not be
induced by these molecules. Further studies are needed to
investigate the potential of targeting intestinal TGR5 and its
impact on gut liver axis as well as on the microbiome host
interactions.
ACKNOWLEDGMENTS
Our studies are supported by Deutsche Forschungsgemeinschaft
(DFG) through SFB 974 and KFO 217.
REFERENCES
1. Li, T. and Chiang, J.Y. Bile acid signaling in metabolic disease and drug ther-
apy. Pharmacol Rev, 2014;66(4):948–83.
2. Russell, D.W. The enzymes, regulation, and genetics of bile acid synthesis.
Annu Rev Biochem, 2003;72:137–74.
3. Hofmann, A.F. Bile acids: the good, the bad, and the ugly. News Physiol Sci,
1999;14:24–9.
4. Joyce, S.A. and Gahan, C.G. Bile acid modifications at the microbe‐host
interface: potential for nutraceutical and pharmaceutical interventions in host
health. Annu Rev Food Sci Technol, 2016;7:313–33.
296 THE LIVER:  REFERENCES
5. Wahlstrom, A., Kovatcheva‐Datchary, P., Stahlman, M., Backhed, F., and
Marschall, H.U. Crosstalk between bile acids and gut microbiota and its
impact on farnesoid X receptor signalling. Dig Dis, 2017;35(3):246–50.
6. Kullak‐Ublick, G., Stieger, B., and Meier, P.J. Enterohepatic bile salt trans-
porters in normal physiology and liver disease. Gastroenterology,
2004;126(1):322–42.
7. Häussinger, D., Reinehr, R., and Keitel, V. Bile acid signaling in the liver and
the biliary tree, in Hepatobiliary Transport in Health and Disease (eds. D.
Häussinger, V. Keitel, and R Kubitz), DeGruyter Publishing, Berlin, 2012,
pp. 85–102.
8. Copple, B.L. and Li, T. Pharmacology of bile acid receptors: evolution of
bile acids from simple detergents to complex signaling molecules. Pharmacol
Res, 2016;104:9–21.
9. Godoy, P., Hewitt, N.J., Albrecht, U. et al. Recent advances in 2D and 3D in
vitro systems using primary hepatocytes, alternative hepatocyte sources and
non‐parenchymal liver cells and their use in investigating mechanisms of hepa-
totoxicity, cell signaling and ADME. Arch Toxicol, 2013;87(8): 1315–530.
10. Keitel, V. and Häussinger, D. Perspective: TGR5 (Gpbar‐1) in liver physiol-
ogy and disease. Clin Res Hepatol Gastroenterol, 2012;36(5):412–19.
11. Pols, T.W., Noriega, L.G., Nomura, M., Auwerx, J., and Schoonjans, K. The
bile acid membrane receptor TGR5 as an emerging target in metabolism and
inflammation. J Hepatol, 2011;54(6):1263–72.
12. Makishima, M., Okamoto, A.Y., Repa, J.J. et al. Identification of a nuclear
receptor for bile acids. Science, 1999;284(5418):1362–5.
13. Parks, D.J., Blanchard, S.G., Bledsoe, R.K. et al. Bile acids: natural ligands
for an orphan nuclear receptor. Science, 1999;284(5418):1365–8.
14. Wang, H., Chen, J., Hollister, K., Sowers, L.C., and Forman, B.M.
Endogenous bile acids are ligands for the nuclear receptor FXR/BAR. Mol
Cell, 1999;3(5):543–53.
15. Sinal, C.J., Tohkin, M., Miyata, M. et al. Targeted disruption of the nuclear
receptor FXR/BAR impairs bile acid and lipid homeostasis. Cell,
2000;102(6):731–44.
16. Watanabe, M., Houten, S.M., Wang, L. et al. Bile acids lower triglyceride
levels via a pathway involving FXR, SHP, and SREBP‐1c. J Clin Invest,
2004;113(10):1408–18.
17. Kalaany, N.Y. and Mangelsdorf, D.J., LXRS and FXR: the yin and yang of
cholesterol and fat metabolism. Annu Rev Physiol, 2006;68:159–91.
18. Modica, S., Gadaleta, R.M., and Moschetta, A. Deciphering the nuclear bile
acid receptor FXR paradigm. Nucl Recept Signal, 2010;8:e005.
19. Inagaki, T., Choi, M., Moschetta, A. et al. Fibroblast growth factor 15 func-
tions as an enterohepatic signal to regulate bile acid homeostasis. Cell
Metab, 2005;2(4):217–25.
20. Kir, S., Beddow, S.A., Samuel, V.T. et al. FGF19 as a postprandial, insulin‐
independent activator of hepatic protein and glycogen synthesis. Science,
2011;331(6024):1621–4.
21. Neuschwander‐Tetri, B.A., Loomba, R., Sanyal, A.J. et  al. Farnesoid X
nuclear receptor ligand obeticholic acid for non‐cirrhotic, non‐alcoholic
steatohepatitis (FLINT): a multicentre, randomised, placebo‐controlled trial.
Lancet, 2015;385(9972):956–65.
22. Nevens, F., Andreone, P., Mazzella, G. et al. A placebo‐controlled trial of obet-
icholic acid in primary biliary cholangitis. N Engl J Med, 2016;375(7):631–43.
23. Makishima, M., Lu, T.T., Xie, W. et al. Vitamin D receptor as an intestinal
bile acid sensor. Science, 2002;296(5571):1313–16.
24. Staudinger, J.L., Goodwin, B., Jones, S.A. et al. The nuclear receptor PXR is
a lithocholic acid sensor that protects against liver toxicity. Proc Natl Acad
Sci U S A, 2001;98(6):3369–74.
25. Xie, W., Radominska‐Pandya, A., Shi, Y. et al. An essential role for nuclear
receptors SXR/PXR in detoxification of cholestatic bile acids. Proc Natl
Acad Sci U S A, 2001;98(6):3375–80.
26. Cheng, K., Chen, Y., Zimniak, P. et al. Functional interaction of lithocholic
acid conjugates with M3 muscarinic receptors on a human colon cancer cell
line. Biochim Biophys Acta, 2002;1588(1):48–55.
27. Raufman, J.P., Chen, Y., Cheng, K. et al. Selective interaction of bile acids
with muscarinic receptors: a case of molecular mimicry. Eur J Pharmacol,
2002;457(2–3):77–84.
28. Raufman, J.P., Chen, Y., Zimniak, P., and Cheng, K. Deoxycholic acid con-
jugates are muscarinic cholinergic receptor antagonists. Pharmacology,
2002;65(4):215–21.
29. Sheikh Abdul Kadir, S.H., Miragoli, M., Abu‐Hayyeh, S. et  al. Bile acid‐
induced arrhythmia is mediated by muscarinic M2 receptors in neonatal rat
cardiomyocytes. PloS One, 2010;5(3):e9689.
30. Chen, X., Yang, D., Shen, W. et  al. Characterization of chenodeoxycholic
acid as an endogenous antagonist of the G‐coupled formyl peptide receptors.
Inflamm Res, 2000;49(12):744–55.
31. Ferrari, C., Macchiarulo, A., Costantino, G., and Pellicciari, R.
Pharmacophore model for bile acids recognition by the FPR receptor. J
Comput Aided Mol Des, 2006;20(5):295–303.
32. Chiang, J.Y. Sphingosine‐1‐phosphate receptor 2: a novel bile acid receptor
and regulator of hepatic lipid metabolism? Hepatology,
2015;61(4):1118–20.
33. Kwong, E., Li, Y., Hylemon, P.B., and Zhou, H. Bile acids and sphingo-
sine‐1‐phosphate receptor 2 in hepatic lipid metabolism. Acta Pharm Sin B,
2015;5(2):151–7.
34. Liu, R., Li, X., Qiang, X. et  al. Taurocholate induces cyclooxygenase‐2
expression via the sphingosine 1‐phosphate receptor 2 in a human cholangio-
carcinoma cell line. J Biol Chem, 2015;290(52):30988–1002.
35. Liu, R., Zhao, R., Zhou, X. et al. Conjugated bile acids promote cholangio-
carcinoma cell invasive growth through activation of sphingosine 1‐phos-
phate receptor 2. Hepatology, 2014;60(3):908–18.
36. Nagahashi, M., Takabe, K., Liu, R. et al. Conjugated bile acid‐activated S1P
receptor 2 is a key regulator of sphingosine kinase 2 and hepatic gene expres-
sion. Hepatology, 2015;61(4):1216–26.
37. Studer, E., Zhou, X., Zhao, R. et al. Conjugated bile acids activate the sphin-
gosine‐1‐phosphate receptor 2 in primary rodent hepatocytes. Hepatology,
2012;55(1):267–76.
38. Kawamata, Y., Fujii, R., Hosoya, M. et  al. A G protein‐coupled receptor
responsive to bile acids. J Biol Chem, 2003;278(11):9435–40.
39. Keitel, V. and Häussinger, D. Role of TGR5 (GPBAR1) in liver disease.
Semin Liver Dis, 2018;38(4):333–9.
40. Maruyama, T., Miyamoto, Y., Nakamura, T. et  al. Identification of mem-
brane‐type receptor for bile acids (M‐BAR). Biochem Biophys Res Commun,
2002;298(5):714–19.
41. Hov, J.R., Keitel, V., Laerdahl, J.K. et al. Mutational characterization of the
bile acid receptor TGR5 in primary sclerosing cholangitis. PloS One
2010;5(8):e12403.
42. Vassileva, G., Golovko, A., Markowitz, L. et al. Targeted deletion of Gpbar1
protects mice from cholesterol gallstone formation. Biochem J,
2006;398(3):423–30.
43. Reich, M., Klindt, C., Deutschmann, K. et al. Role of the G protein‐coupled
bile acid receptor TGR5 in liver damage. Dig Dis, 2017;35(3):235–40.
44. Maruyama, T., Tanaka, K., Suzuki, J. et al. Targeted disruption of G protein‐
coupled bile acid receptor 1 (Gpbar1/M‐Bar) in mice. J Endocrinol,
2006;191(1):197–205.
45. Keitel, V., Görg, B., Bidmon, H.J. et  al. The bile acid receptor TGR5
(Gpbar‐1) acts as a neurosteroid receptor in brain. Glia, 2010;58(15):
1794–805.
46. Keitel, V., Spomer, L., Marin, J.J. et al. Effect of maternal cholestasis on TGR5
expression in human and rat placenta at term. Placenta, 2013;34(9):810–16.
47. Deutschmann, K., Reich, M., Klindt, C. et al. Bile acid receptors in the biliary
tree: TGR5 in physiology and disease. Biochim Biophys Acta, 2018;1864(4 Pt
B):1319–25.
48. Keitel, V., Donner, M., Winandy, S., Kubitz, R., and Häussinger, D.,
Expression and function of the bile acid receptor TGR5 in Kupffer cells.
Biochem Biophys Res Commun, 2008;372(1):78–84.
49. Keitel, V., Reich, M., and Häussinger, D. TGR5: pathogenetic role and/or
therapeutic target in fibrosing cholangitis? Clin Rev Allergy
Immunol,2015;48(2–3):218–25.
50. Keitel, V., Reinehr, R., Gatsios, P. et  al. The G‐protein coupled bile salt
receptor TGR5 is expressed in liver sinusoidal endothelial cells. Hepatology,
2007;45(3):695–704.
51. Keitel, V., Ullmer, C., and Häussinger, D., The membrane‐bound bile acid
receptor TGR5 (Gpbar‐1) is localized in the primary cilium of cholangio-
cytes. Biol Chem, 2010;391(7):785–9.
52. Reich, M., Deutschmann, K., Sommerfeld, A. et al. TGR5 is essential for
bile acid‐dependent cholangiocyte proliferation in vivo and in vitro. Gut
2016;65(3):487–501.
53. Sawitza, I., Kordes, C., Götze, S., Herebian, D., and Häussinger, D., Bile
acids induce hepatic differentiation of mesenchymal stem cells. Sci Rep,
2015;5:13320.
54. Keitel, V., Cupisti, K., Ullmer, C. et  al. The membrane‐bound bile acid
receptor TGR5 is localized in the epithelium of human gallbladders.
Hepatology, 2009;50(3):861–70.
 24:  TGR5 (GPBAR1) in the Liver 297
55. Lavoie, B., Balemba, O.B., Godfrey, C. et al. Hydrophobic bile salts inhibit
gallbladder smooth muscle function via stimulation of GPBAR1 receptors
and activation of KATP channels. J Physiol, 2010;588(Pt 17):3295–305.
56. Masyuk, A.I., Huang, B.Q., Radtke, B.N. et al. Ciliary subcellular localiza-
tion of TGR5 determines the cholangiocyte functional response to bile acid
signaling. Am J Physiol Gastrointest Liver Physiol, 2013;304(11):
G1013–24.
57. Keitel, V. and Häussinger, D. TGR5 in the biliary tree. Dig Dis,
2011;29(1):45–7.
58. Keitel, V. and Häussinger, D. TGR5 in cholangiocytes. Curr Opin
Gastroenterol, 2013;29(3):299–304.
59. Sato, H., Macchiarulo, A., Thomas, C. et al. Novel potent and selective bile
acid derivatives as TGR5 agonists: biological screening, structure‐activity
relationships, and molecular modeling studies. J Med Chem,
2008;51(6):1831–41.
60. Martin, R.E., Bissantz, C., Gavelle, O. et al. 2‐Phenoxy‐nicotinamides are
potent agonists at the bile acid receptor GPBAR1 (TGR5). ChemMedChem,
2013;8(4):569–76.
61. Gertzen, C.G., Spomer, L., Smits, S.H. et  al. Mutational mapping of the
transmembrane binding site of the G‐protein coupled receptor TGR5 and
binding mode prediction of TGR5 agonists. Eur J Med Chem,
2015;104:57–72.
62. Katritch, V. and Abagyan, R. GPCR agonist binding revealed by modeling
and crystallography. Trends Pharmacol Sci, 2011;32(11):637–43.
63. Pean, N., Doignon, I., Garcin, I. et al. The receptor TGR5 protects the liver
from bile acid overload during liver regeneration in mice. Hepatology,
2013;58(4):1451–60.
64. Cao, H., Chen, Z.X., Wang, K. et  al. Intestinally‐targeted TGR5 agonists
equipped with quaternary ammonium have an improved hypoglycemic effect
and reduced gallbladder filling effect. Sci Rep, 2016;6:28676.
65. Duan, H., Ning, M., Zou, Q. et  al. Discovery of intestinal targeted TGR5
agonists for the treatment of type 2 diabetes. J Med Chem,
2015;58(8):3315–28.
66. Briere, D.A., Ruan, X., Cheng, C.C. et al. Novel small molecule agonist of
TGR5 possesses anti‐diabetic effects but causes gallbladder filling in mice.
PloS One, 2015;10(8):e0136873.
67. Duan, H., Ning, M., Chen, X. et al. Design, synthesis, and antidiabetic activ-
ity of 4‐phenoxynicotinamide and 4‐phenoxypyrimidine‐5‐carboxamide
derivatives as potent and orally efficacious TGR5 agonists. J Med Chem,
2012;55(23):10475–89.
68. Pellicciari, R., Gioiello, A., Macchiarulo, A. et  al. Discovery of 6alpha‐
ethyl‐23(S)‐methylcholic acid (S‐EMCA, INT‐777) as a potent and selective
agonist for the TGR5 receptor, a novel target for diabesity. J Med Chem,
2009;52(24):7958–61.
69. Spomer, L., Gertzen, C.G., Schmitz, B. et al. A membrane‐proximal, C‐ter-
minal alpha‐helix is required for plasma membrane localization and function
of the G Protein‐coupled receptor (GPCR) TGR5. J Biol Chem,
2014;289(6):3689–702.
70. Masyuk, T.V., Masyuk, A.I., Lorenzo Pisarello, M. et al. TGR5 contributes
to hepatic cystogenesis in rodents with polycystic liver diseases through
cyclic adenosine monophosphate/Galphas signaling. Hepatology,
2017;66(4):1197–218.
71. Hong, J., Behar, J., Wands, J. et al. Role of a novel bile acid receptor TGR5
in the development of oesophageal adenocarcinoma. Gut, 2010;
59(2):170–80.
72. Wacker, D., Wang, C., Katritch, V. et al. Structural features for functional
selectivity at serotonin receptors. Science, 2013;340(6132):615–19.
73. Jensen, D.D., Godfrey, C.B., Niklas, C. et al. The bile acid receptor TGR5 does
not interact with beta‐arrestins or traffic to endosomes but transmits sustained
signals from plasma membrane rafts. J Biol Chem, 2013;288(32):22942–60.
74. Thomas, C., Gioiello, A., Noriega, L. et al. TGR5‐mediated bile acid sensing
controls glucose homeostasis. Cell Metabol, 2009;10(3):167–77.
75. Velazquez‐Villegas, L.A., Perino, A., Lemos, V. et al. TGR5 signalling pro-
motes mitochondrial fission and beige remodelling of white adipose tissue.
Nat Commun, 2018;9(1):245.
76. Perino, A., Pols, T.W., Nomura, M. et al. TGR5 reduces macrophage migra-
tion through mTOR‐induced C/EBPbeta differential translation. J Clin
Invest, 2014;124(12):5424–36.
77. Lieu, T., Jayaweera, G., Zhao, P. et al. The bile acid receptor TGR5 activates
the TRPA1 channel to induce itch in mice. Gastroenterology, 2014;
147(6):1417–28.
78. Rajagopal, S., Kumar, D.P., Mahavadi, S. et al. Activation of G protein‐
coupled bile acid receptor, TGR5, induces smooth muscle relaxation via
both Epac‐ and PKA‐mediated inhibition of RhoA/Rho kinase pathway.
Am J Physiol Gastrointest Liver Physiol, 2013;304(5):G527–35.
79. Hiller, C., Kuhhorn, J., and Gmeiner, P., Class A G‐protein‐coupled receptor
(GPCR) dimers and bivalent ligands. J Med Chem, 2013;56(17):6542–59.
80. Greife, A., Felekyan, S., Ma, Q. et al. Structural assemblies of the di‐ and
oligomeric G‐protein coupled receptor TGR5 in live cells: an MFIS‐FRET
and integrative modelling study. Sci Rep, 2016;6:36792.
81. Patowary, S., Alvarez‐Curto, E., Xu, T.R. et al. The muscarinic M3 acetyl-
choline receptor exists as two differently sized complexes at the plasma
membrane. Biochem J, 2013;452(2):303–12.
82. McMillin, M., Frampton, G., Tobin, R. et al. TGR5 signaling reduces
neuroinflammation during hepatic encephalopathy. J Neurochem,
2015;135(3):565–76.
83. Fiorucci, S., Zampella, A., Cirino, G., Bucci, M., and Distrutti, E., Decoding
the vasoregulatory activities of bile acid‐activated receptors in systemic and
portal circulation: role of gaseous mediators. Am J Physiol Heart Circ
Physiol, 2017;312(1):H21–H32.
84. Renga, B., Bucci, M., Cipriani, S. et  al. Cystathionine gamma‐lyase, a
H2S‐generating enzyme, is a GPBAR1‐regulated gene and contributes to
vasodilation caused by secondary bile acids. Am J Physiol Heart Circ
Physiol, 2015;309(1):H114–26.
85. Renga, B., Cipriani, S., Carino, A. et al. Reversal of endothelial dysfunction by
GPBAR1 agonism in portal hypertension involves a AKT/FOXOA1 dependent
regulation of H2S generation and endothelin‐1. PloS One, 2015;10(11):e0141082.
86. Reinehr, R., Fischer, R., and Häussinger, D. Regulation of endothelin‐A
receptor sensitivity by cyclic adenosine monophosphate in rat hepatic stel-
late cells. Hepatology, 2002;36(4 Pt 1):861–73.
87. Wang, Y.D., Chen, W.D., Yu, D., Forman, B.M., and Huang, W. The G‐pro-
tein‐coupled bile acid receptor, Gpbar1 (TGR5), negatively regulates
hepatic inflammatory response through antagonizing nuclear factor kappa
light‐chain enhancer of activated B cells (NF‐kappaB) in mice. Hepatology,
2011;54(4):1421–32.
88. Guo, C., Xie, S., Chi, Z. et al. Bile acids control inflammation and meta-
bolic disorder through inhibition of NLRP3 inflammasome. Immunity,
2016;45(4):802–16.
89. Pols, T.W., Nomura, M., Harach, T. et al. TGR5 activation inhibits athero-
sclerosis by reducing macrophage inflammation and lipid loading. Cell
Metab, 2011;14(6):747–57.
90. Perino, A. and Schoonjans, K., TGR5 and immunometabolism: insights from
physiology and pharmacology. Trends Pharmacol Sci, 2015;36(12):847–57.
91. Haselow, K., Bode, J.G., Wammers, M. et al. Bile acids PKA‐dependently
induce a switch of the IL‐10/IL‐12 ratio and reduce proinflammatory capa-
bility of human macrophages. J Leukoc Biol, 2013;94(6):1253–64.
92. Hogenauer, K., Arista, L., Schmiedeberg, N. et al. G‐Protein‐coupled bile
acid receptor 1 (GPBAR1, TGR5) agonists reduce the production of proin-
flammatory cytokines and stabilize the alternative macrophage phenotype.
J Med Chem, 2014;57(24):10343–54.
93. Ichikawa, R., Takayama, T., Yoneno, K. et al. Bile acids induce monocyte
differentiation toward interleukin‐12 hypo‐producing dendritic cells via a
TGR5‐dependent pathway. Immunology, 2012;136(2):153–62.
94. Wammers, M., Schupp, A.K., Bode, J.G. et  al. Reprogramming of pro‐
inflammatory human macrophages to an anti‐inflammatory phenotype by
bile acids. Sci Rep, 2018;8(1):255.
95. Li, T., Holmstrom, S.R., Kir, S. et al. The G protein‐coupled bile acid receptor,
TGR5, stimulates gallbladder filling. Mol Endocrinol, 2011;25(6):1066–71.
96. Beuers, U., Hohenester, S., de Buy Wenniger, L.J. et al. The biliary HCO(3)
(‐) umbrella: a unifying hypothesis on pathogenetic and therapeutic aspects
of fibrosing cholangiopathies. Hepatology, 2010;52(4):1489–96.
97. Hohenester, S., Wenniger, L.M., Paulusma, C.C. et  al. A biliary HC03‐
umbrella constitutes a protective mechanism against bile acid‐induced
injury in human cholangiocytes. Hepatology, 2012;55(1):173–83.
98. Merlen, G., Ursic‐Bedoya, J., Kahale, N. et al. The bile acid receptor TGR5
regulates paracellular permeability and protects the liver through an impact
on the tight junction protein JAM‐A. J Hepatol, 2017;66(1):S44–S45.
99. Choi, M., Moschetta, A., Bookout, A.L. et al. Identification of a hormonal
basis for gallbladder filling. Nat Med, 2006;12(11):1253–5.
100. Donepudi, A.C., Boehme, S., Li, F., and Chiang, J. Y. G‐protein‐coupled
bile acid receptor plays a key role in bile acid metabolism and fasting‐
induced hepatic steatosis in mice. Hepatology, 2017;65(3):813–27.
298 THE LIVER:  REFERENCES
101. Häussinger, D. and Keitel, V. Dual role of the bile acid receptor Takeda G‐
protein‐coupled receptor 5 for hepatic lipid metabolism in feast and famine.
Hepatology, 2017;65(3):767–70.
102. Watanabe, M., Houten, S.M., Mataki, C. et  al. Bile acids induce energy
expenditure by promoting intracellular thyroid hormone activation. Nature,
2006;439(7075):484–9.
103. Ullmer, C., Alvarez Sanchez, R., Sprecher, U. et al. Systemic bile acid sens-
ing by G protein‐coupled bile acid receptor 1 (GPBAR1) promotes PYY
and GLP‐1 release. Br J Pharmacol, 2013;169(3):671–84.
104. Phillips, D.P., Gao, W., Yang, Y. et  al. Discovery of
trifluoromethyl(pyrimidin‐2‐yl)azetidine‐2‐carboxamides as potent,
orally bioavailable TGR5 (GPBAR1) agonists: structure‐activity relation-
ships, lead optimization, and chronic in vivo efficacy. J Med Chem, 2014;
57(8):3263–82.
105. Tacke, F. Targeting hepatic macrophages to treat liver diseases. J Hepatol,
2017;66(6):1300–12.
106. McMahan, R.H., Wang, X.X., Cheng, L.L. et al. Bile acid receptor activa-
tion modulates hepatic monocyte activity and improves nonalcoholic fatty
liver disease. J Biol Chem, 2013;288(17):11761–70.
107. Chavez‐Talavera, O., Tailleux, A., Lefebvre, P., and Staels, B., Bile acid
control of metabolism and inflammation in obesity, type 2 diabetes, dyslipi-
demia, and nonalcoholic fatty liver disease. Gastroenterology, 2017;
152(7):1679–94 e3.
108. Iwakiri, Y. and Groszmann, R.J. Vascular endothelial dysfunction in cirrhosis.
J.Hepatol, 2007;46(5):927–34.
109. Iwakiri, Y., Shah, V., and Rockey, D.C. Vascular pathobiology in chronic
liver disease and cirrhosis – current status and future directions. J Hepatol,
2014;61(4):912–24.
110. Franke, A., Balschun, T., Karlsen, T.H. et  al. Sequence variants in IL10,
ARPC2 and multiple other loci contribute to ulcerative colitis susceptibil-
ity. Nat Genet, 2008;40(11):1319–23.
111. Karlsen, T.H., Franke, A., Melum, E. et al. Genome‐wide association anal-
ysis in primary sclerosing cholangitis. Gastroenterology, 2010;
138(3):1102–11.
112. Hov, J.R., Keitel, V., Schrumpf, E., Häussinger, D., and Karlsen, T.H.
TGR5 sequence variation in primary sclerosing cholangitis. Dig Dis,
2011;29(1):78–84.
113. Masyuk, T.V., Masyuk, A.I., and LaRusso, N.F. TGR5 in the cholangiocili-
opathies. Dig Dis, 2015;33(3):420–5.
114. Erice, O., Labiano, I., Arbelaiz, A. et  al. Differential effects of FXR or
TGR5 activation in cholangiocarcinoma progression. Biochim Biophys
Acta, 2018;1864(4 Pt B):1335–44.
115. Doignon, I., Julien, B., Serriere‐Lanneau, V. et al. Immediate neuroendocrine
signaling after partial hepatectomy through acute portal hyperpressure and
cholestasis. J Hepatol, 2011;54(3):481–8.
116. Häussinger, D. and Schliess, F. Pathogenetic mechanisms of hepatic
encephalopathy. Gut, 2008;57(8):1156–65.
117. Karababa, A., Groos‐Sahr, K., Albrecht, U. et al. Ammonia attenuates LPS‐
induced upregulation of pro‐inflammatory cytokine mRNA in co‐cultured
astrocytes and microglia. Neurochem Res, 2017;42(3):737–49.
 Bile Acids as Signaling
25 Molecules
Thierry Claudel and Michael Trauner
Hans Popper Laboratory of Molecular Hepatology, Division of Gastroenterology and Hepatology,
Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
INTRODUCTION
Primary bile acids are synthesized by enzymes belonging to the
cytochrome P450 family (CYP). In humans the “classical path-
way,” with CYP7A1 as limiting step, produces the primary bile
acids cholic acid (CA) and chenodeoxycholic acid (CDCA),
while the “alternative pathway,” involving CYP27A1, produces
CDCA [1, 2] (Table  25.1). In rodents, CDCA is metabolized
into α‐muricholic acid (αMCA) by Cyp2c70 and epimerized in
the intestine into β‐muricholic acid (βMCA) [3] (Table  25.1).
Primary bile acids are conjugated with taurine in rodents (TCA
and TβMCA) and glycine in humans (GCA and GCDCA).
Intestinal bacteria deconjugate these primary into secondary
bile acids; βMCA can be then epimerized at C‐6 into ωMCA
(Table  25.1); while CA, CDCA, and ωMCA can be 7α‐dehy-
droxylated into deoxycholic acid (DCA), lithocholic acid
(LCA), and hyodeoxycholic acid (HDCA), respectively
(Table  25.1) [4, 5]. CDCA is also converted into ursodeoxy-
cholic acid (UDCA) by the hydroxysteroid dehydrogenase and
subsequently UDCA can be 7α‐dehydroxylated into LCA or
oxidized in the liver by Cyp2c70 into βMCA (Table 25.1). DCA,
LCA, UDCA, HDCA, and ωMCA are secondary bile acids and
display either higher hydrophobicity and toxicity (LCA, DCA)
or hydrophilicity (UDCA, HDCA) than their primary bile acids
precursors. Therefore bile acid composition depends on micro-
biota composition and, conversely, bile acids also contribute to
shape the microbiome [6].
Since bile acids are amphipathic molecules which can dam-
age cell membranes, induce apoptosis, and have potential pro-
inflammatory actions at higher concentrations [7], various sensor
mechanisms are necessary to coordinate metabolic and trans-
port processes limiting the bile acid load and exposure, thus pro-
tecting the cellular integrity in the different organs handling bile
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
acids by controlling a network of transporters and enzymes.
These sensors are nuclear, intracellular, and membrane proteins
that coordinate intra‐ and inter‐organ signaling and communica-
tions by regulating bile acid homeostasis and a wide range of
other metabolic and immunological pathways (Figure 25.1) [1].
BILE ACID‐ACTIVATED NUCLEAR
RECEPTORS: FXR, PXR, CAR, VDR,
AND GR
Nuclear receptors (NRs) are transcription factors that, upon
ligand binding or post‐translational modifications and cofactor
recruitment, modulate RNA polymerase II activity and gene
expression. Several of these NRs are activated by bile acids.
Farnesoid X receptor
The farnesoid X receptor (FXR, NR1H4) [8, 9] was originally
described as farnesol‐activated receptor [8], but several years
later its key function as the main bile acid‐activated NR was
identified [10–12]. Human FXR is maximally activated by
CDCA (Table 25.1, Figure 25.1), followed by DCA, CA, LCA
and weakly UDCA, whereas mouse Fxr is marginally respon-
sive to CDCA [13] which is only an intermediate in the synthe-
sis of MCA in rodents (see earlier and Table  25.1). FXR can
accommodate both glycine/taurine conjugated‐ and free bile
acids in its ligand‐binding domain (LBD) [14]. In addition to
bile acids, fatty acids such as arachidonic acid, docosahexaenoic
acid, and linoleic acid were found to activate FXR [15, 16].
Moreover, various semi‐synthetic bile acids such as the obeticholic
acid (OCA)/6‐ethyl‐CDCA [17] and isoxazole derivatives such
300 THE LIVER:  BILE ACID‐ACTIVATED NUCLEAR RECEPTORS: FXR, PXR, CAR, VDR, AND GR

Table 25.1  Bile acid nomenclature, properties, and biological activities

Bile acid Full name Origin Biological activity
CA Cholic acid Liver (Cyp7a1) FXR agonist
CDCA Chenodeoxycholic acid Liver (Cyp27a1 and Cyp7a1) Most potent FXR agonist in human but not in mouse;
FPR receptor antagonist
αMCA α‐Muricholic acid Liver: oxidation of CDCA by Cyp2c70in rodents only FXR antagonist
βMCA β‐Muricholic acid Liver: oxidation of UDCA by Cyp2c70 in rodents only FXR antagonist
Intestine: epimerization of αMCA
LCA Lithocholic acid Intestine: 7α‐dehydroxylation of CDCA by Clostridium and FXR, PXR, VDR, TGR5 agonist
Eubacterium and/or 7α‐dehydroxylation of UDCA
UDCA Ursodeoxycholic acid Intestine: hydroxysteroid dehydrogenization of CDCA FXR partial agonist
GRα agonist
HDCA Hyodeoxycholic acid Intestine: 7α‐dehydroxylation of ωMCA by Clostridium and Lowers FXR antagonist ωMCA concentration
Eubacterium
DCA Deoxycholic acid Intestine: 7α‐dehydroxylation of CA by Clostridium and FXR, TGR5, S1PR2, CHMR3 agonist; FPR receptor
Eubacterium antagonist
ωMCA ω‐Muricholic acid Intestine: epimerization in C‐6 of βMCA by Bacteroides, FXR antagonist?
Escherichia, Clostridium, Eubacterium

as GW4064 [18] act as FXR ligands. These were used as discov-
ery tools in preclinical studies (GW4064) [19, 20] or developed
further for treatment of primary biliary cholangitis [21] and
non‐alcoholic fatty liver (OCA) [22]. In addition to bile acid‐
derived FXR ligands, nonsteroidal FXR agonists were also
developed, which may differ in their pharmacokinetic and phar-
macodynamic properties (see also Chapter 30) [23]. FXR can
form heterodimers with the retinoid X receptor alpha (RXRα;
NR2B1), binding to FXR response elements (FXRE) and stimu-
lating gene transcription [8, 9, 24], but FXR also binds as a
monomer repressing [25] or increasing gene expression [26].
Based on genome binding analysis, FXREs are located in the
vicinity of genes and clusters involved in lipid, fatty acid, ster-
oid metabolism, glycolysis, and transporters shared with another
NR, the liver receptor homolog 1 (LRH‐1; NR5A2) [27].
Whole‐body [28] or tissue‐specific [29] Fxr‐deficient mice
revealed the central role of FXR signaling in lipid and bile acid
homeostasis. At least four isoforms of FXR are expressed in
human and mouse due to alternative first exon use and a stretch
of amino acids inserted in the hinge domain between the DNA‐
binding domain (DBD) and the LBD [30]. Hereditary defects of
FXR result in a severe form of progressive familial intrahepatic
cholestasis (PFIC) [31], while polymorphisms predispose to
intrahepatic cholestasis of pregnancy (ICP) [32] or influence
fasting glycemia and free fatty acid levels [33]. A second FXR
locus, called Fxrβ (NR1H5), identified in rodents and shown to
encode a functional protein responsive to lanosterol, but not bile
acid, is only a pseudogene in primates [34].
Pregnane X receptor/constitutive
androstane receptor
In addition to FXR, other NRs such as the pregnane X receptor
(PXR; NRI2) [35, 36] and the constitutive androstane receptor
(CAR; NR1I3) [37] were shown to respond to bile acids
(Figure 25.1 and Table 25.1). Fxr‐deficient mice have a massive
induction of enzymes involved in bile acids transport and detox-
ification, which are target genes of PXR and CAR [38], allow-
ing these receptors and their downstream targets to partly
compensate for loss of FXR. Moreover, combined deficiency of
PXR and CAR increased the sensitivity to bile acid toxicity in
mice [39].
Initial studies showed that PXR was activated by 21‐carbon
steroids called pregnanes, resulting in the name pregnane X
receptor [40]. PXR can form heterodimers with RXR and stimu-
lates expression of CYP3A4, a key enzyme for drug detoxifica-
tion and drug/drug interactions [41–43]. PXR is mainly
expressed in liver and intestine [40–43] and is a promiscuous
receptor activated by nearly 60% of the drugs existing due to its
abnormally large LBD, accommodating a wide range of ligands
[44] including bile acids such as LCA.
CAR is highly expressed in intestine and liver and is consti-
tutively active even in absence of ligands, forming heterodimers
with RXR [45].
Despite species differences in their LBDs, PXR and CAR
retained the capacity to respond to bile acids and to coordinate
bile acid detoxification. Hence PXR activation lowers Cyp7a1
(see later and Figure 25.2) without affecting Cyp27, Cyp8b1, or
Cyp7b1 in mice [46]. Moreover, LCA (Table  25.1) and the
metabolite 3‐keto‐LCA binds to the LBD and activates PXR at
concentrations found in cholestasis [46]. PXR in turn represses
Cyp7a1 and induces the bile acid transporter Oatp2, facilitating
LCA uptake into the hepatocytes where it is subsequently
detoxified via coordinately induced Cyp3a11 [46]. CAR is pre-
sent in cytosol and unable to affect gene transcription until its
translocation to the nucleus after phosphorylation, where it
upregulates Cyp2b genes [47]. CAR activation in mouse liver
results in induction of Fxr, Cyp8b1, Cy27a1, Cyp39a1, and
Ugt1a1, which altogether not only reduces the total amount of
bile acid but also makes the bile acid pool composition less
toxic [48]. Taken together, CAR and PXR act as a second line of
defense against toxic bile acids as additional backup to FXR
signaling.
Vitamin D receptor
Vitamin D receptor (VDR; NR1I1) forms heterodimers with
RXR and is primarily activated by vitamin D3 and as such regu-
lates calcium homeostasis as well as inflammation [49].
Moreover, VDR was identified as an additional bile acid‐respon-
sive NR in intestine, where it is activated by LCA (Table 25.1
and Figure 25.1) [50] and induces CYP3A4, metabolizing LCA
and counteracting its pro‐carcinogenic properties  [50]. VDR
activation after vitamin D3 gavage reduced intrahepatic and
 25:  Bile Acids as Signaling Molecules 301

N‐Formyl CDCA
peptides DCA LCA DCA DCA
Pro-EGF EGF

TGR5 S1pR2 MMP EGFR

FPR receptor CHMR3
PLCβ Cytosol
Gα cAMP RAS

ERK AKT PLC ERK

PKC PKC

PLA2 Ca2+
COX influx
5-LO
ROS

LTB4
FXR PXR CREB SPHK2
VDR p-FXR Nucleus

PPARγ p-SHP

Chemotaxis BA/lipid/glucose NO production BA/lipid

homeostasis Anti-inflammatory metabolism

Figure 25.1  Signaling pathways activated by bile acids. CDCA, LCA, and DCA are the main ligands of the nuclear receptor FXR as key regulator
of bile acid homeostasis, lipid, and glucose metabolism (center). LCA activates the nuclear receptors PXR and VDR, regulating bile acid detoxifica-
tion and inflammation (center). LCA and DCA are also ligands for the membrane‐bound G‐coupled receptor TGR5, which stimulates the produc-
tion of cyclic AMP (cAMP) activating the transcription factor cAMP response element‐binding protein (CREB), which in turn regulates nitric oxide
(NO) levels and suppresses inflammation (center). Bile acids such as DCA can activate either the sphingosine 1 phosphate receptor 2 (S1PR2) or
the muscarinic receptor 3 (CHMR3) or the epidermal growth factor receptor (EGFR), which all reside in the cell membrane (right part of the
image). S1PR2 and CHMR3 activate the phospholipase C beta (PLCβ), which activates membrane metalloproteinase (MMP). MMP releases EGF
from the membrane‐bound precursor pro‐EGF, allowing EGF then to bind to EGFR, in turn activating the small GTPase RAS resulting in activation
of protein kinase C zeta (PKCζ), phosphorylating FXR and SHP. S1PR2 can also directly activate ERK signaling, leading to increased activity of
the sphingosine kinase 2 (SPHK2), catalyzing the phosphorylation of sphingosine and controlling angiogenesis, tumorigenesis, and lipid metabo-
lism. N‐Formyl peptides are bacterial components activating the G‐coupled receptor formyl peptide receptor (FPR), CDCA blocks ligand binding
to this receptor (left part of the image). FPR stimulates phospholipase C (PLC) and protein kinase C (PKC) signaling, which together with AKT
leads to calcium influx and reactive oxygen species (ROS) formation. Calcium influx also initiates a negative feedback desensitizing FPR/FPRL1.
FPR/FPRL1 also activates extracellular signal‐regulated kinase (ERK), which increases the activity of the phospholipase A2 (PLA2), arachidonate
5‐lipoxygenase (5‐LO) and cyclooxygenase (COX) which produce leukotriene B4 (LTB4), a potent ligand activating the nuclear receptor PPARγ.

circulating bile acid levels by promoting their urinary excretion
in mice [51]. Since LCA is a secondary bile acid obtained after
bacterial modification (Table 25.1), the impact of VDR on intes-
tinal microbiota and bile acid metabolism will be of future
interest.
Glucocorticoid receptor
Glucocorticoids activate the glucocorticoid receptor (GR;
NR3C1) [52]. Two isoforms, GRα and GRβ, exist by exon 9
splicing: GRα is a classical NR activated by glucocorticoids,
while GRβ acts as a dominant negative receptor [53].
Importantly. GRα is also activated by UDCA but not other bile
acids (Table  25.1) [54] but the binding of UDCA to GRα is
weaker than other glucocorticoids, such as dexamethasone [55].
Activation of GRα may contribute to the immunosuppressive
properties of UDCA, such as reducing the hepatic expression of
class I and II major histocompatibility complex in PBC [56, 57].
GRα may also be required to mediate other immunomodulatory
effects of UDCA, such as suppressing interleukin 18 and inter-
feron gamma production by Kupffer cells and NKT cells,
respectively [58]. Novel UDCA derivatives have been
developed which may act as new GR modulators, retaining
anti‐inflammatory properties without disturbing glucose
metabolism [59].
Fibroblast growth factor 19
FXR induces the expression of Fgf15 (mice) [60]/FGF19
(humans) [61] in the terminal ileum. In contrast to mouse Fgf15,
FGF19 is also expressed in hepatocytes [61]. Intestinal FGF19/
Fgf15 reaches the liver via the portal vein and binds to the tyros-
ine kinase receptor FGFR4, or alternatively to FGFR1c in mac-
rophages, adipocytes, and brain (Figure  25.3) [62]. The
interactions between FGF19/Fgf15 and FGFR4 or FGFR1c are
only possible via the membrane‐bound protein βKlotho, which
allows a tissue‐specific fine tuning of the ligand–receptor inter-
action since its expression is more restricted than that of FGFR4
or FGFR1c [62]. PXR and VDR can also activate the intestinal
expression of FGF19/Fgf15 [62]. FGF19/Fgf15 reduces bile
acid synthesis (by repressing CYP7A1), gluconeogenesis, and
lipogenesis and increases hepatic glycogen and protein synthe-
sis independently of insulin signaling (see Figures  25.2 and
25.3) [63]. In addition, FGF19/Fgf15 controls gallbladder
302 THE LIVER:  BILE ACID‐ACTIVATED NUCLEAR RECEPTORS: FXR, PXR, CAR, VDR, AND GR

3 Alternative export BA

MRP4 MRP3 OSTα/β Phase III

CAR CAR PXR VDR FXR

OATP1A1
2
CAR UGTs
BA OATP1B2 G/S-BA Phase II
SULTs
OA CAR MRP2 OA, Bili
OATP1B3 FXR CYPs ox-BA Phase I

Canalicular export
PXR
OATP1A4 PXR BA
Uptake

CYP7A1 PPARα MDR2/3 PL

BA HNF4α FGF19 FXR BSEP BA

BA NTCP HNF4α SHP FXR ABCG5

Chol
ABCG8
BA
1

4 Bile synthesis and detoxification Hepatocyte

Figure 25.2  Nuclear receptors and transcriptional regulation of hepatocellular bile acid transport and metabolism. (1) Bile acid (BA) uptake:
HNF4α regulates basolateral Na+‐dependent (NTCP) and is repressed directly by bile acids or indirectly via FXR/SHP. FXR upregulates human
OATP1B3, while PXR increases OATP1A4 expression. (2) Canalicular export of bile acids into the bile: all canalicular transporters involved in bile
formation (BSEP for BAs, MDR3/Mdr2 for phospholipids (PL), MRP2 for BAs, conjugated bilirubin, and GSH, ABCG5/8 for cholesterol (Chol))
are upregulated by FXR. CAR and PXR positively regulate MRP2 and PPARα upregulates MDR3/Mdr2. FXR also upregulates PPARα expression
in humans. (3) Alternative export into blood to allow renal excretion: CAR positively regulates MRP3 and MRP4, while PXR and VDR upregulate
MRP3. FXR induces OSTα/β. (4) Bile acid synthesis and detoxification: CYP7A1 is the rate‐limiting step in BA synthesis. FXR represses CYP7A1
expression via the induction of the hepatic and intestinal FGF19 in humans, Fgf15 in mouse which is produced only in the terminal ileum (not
shown). FXR also represses CYP7A1 expression via the induction of SHP which in turn decreases HNF4α activity. HNF4α activity can also be
directly impaired by BAs. Bile acids can be detoxified during phase I by oxidation catalyzed by enzymes from the cytochrome P450 family (CYPs).
After oxidation, ox‐bile acids can be conjugated by sulfation catalyzed by the sulfotransferases (SULTs) or glucuronidated by the UDP‐glucurono-
syltransferases (UGTs). FXR/CAR/PXR upregulate CYPs, SULTs, and UGTs expression. After conjugation, glucuronidated bile acids (G‐BA) and
sulfated‐bile acids (S‐BA) can be excreted either into the bile via MRP2 or into the bloodstream via the alternative transporters MRP3, MRP4, and
OSTα/β. Arrows: upregulation or metabolite flux; blocked arrows: transcriptional repression.

filling (via FGFR4), lowers food intake and increases insulin
sensitivity in the brain (via FGFR1c), metabolically activates
white adipose tissue (WAT) adipocytes (via FGFR1c), and
decreases macrophage uptake of oxidized low‐density lipopro-
teins (LDL) (and thereby foam cell formation) by reducing
CD36 expression (via FGFR1c) (Figure  25.3) [62]. Thus
FGF19/Fgf15 acts as a bile acid‐induced key regulator of inter-
mediary metabolism.
Bile acid signaling actions mediated
by nuclear receptors
Bile acid synthesis
CYP7A1 is the rate‐limiting enzyme in bile acid synthesis and is
transcriptionally regulated via a negative feedback by bile acids
[64]. Bile acids binding to FXR induces expression of the NR
small heterodimer partner (SHP; NR0B2), which interacts with
LRH‐1 to suppress CYP7A1 transcription (Figure 25.2) [65, 66].
Lrh‐1 and the NR hepatocyte nuclear factor 4 alpha (Hnf4α;
NR2A1) drive Cyp7a1 baseline expression by binding to specific
promoter sequences  [67] and by promoting histone modifica-
tions reversed by Shp [68]. However, the rather weak bile acid
phenotype of Shp‐deficient mice [69, 70] combined with the
marked decrease of Cyp8b1 but preserved expression of Cyp7a1
in Lrh‐1‐deficient mice [71, 72], suggested that intestinal factors
might also contribute to Cyp7a1 regulation [29].
Indeed, experiments with liver and ileum‐specific Fxr‐defi-
cient mice suggest that the ileal route of Cyp7a1 repression via
the FGF19/Fgf15 pathway dominates over hepatic‐negative
feedback pathways, indicating that a functional gut–liver sign-
aling is a prerequisite for bile acid homeostasis [29]
(Figure 25.3). FGF19/Fgf15 after binding to FGFR4/βKlotho
activates the tyrosine kinase function of Fgfr4, which represses
Cyp7a1 through c‐Jun terminal kinase (Jnk) and Shp‐depend-
ent pathways [60]. In addition to Fxr/Shp and FXR/FGF19,
bile acids can directly decrease Hnf4α gene expression [73]
and impair Hnf4α activation of the Cyp7a1 promoter by block-
ing the recruitment of its coactivators the peroxisome prolif-
erator activated receptor gamma coactivator (Pgc‐1α) or the
cAMP response element‐binding protein (CBP) [74].
Moreover, PXR (e.g. activated by hydrophobic bile acids)
 25:  Bile Acids as Signaling Molecules 303

FGF15

FXR βKlotho GLP1-R

TGR5 FGFR1c
FXR FGFR4
Glycogen Neuro- Glucose Insulin
synthesis transmitters metabolism sensitivity
TG FA metabolic &
clearance oxidation rate satiety
&
TG HDL Protein feeding

synthesis synthesis synthesis

Brain

FGF15 FGF15
βKlotho βKlotho/FGFR1c TGR5/FGFR1c
FGF15
FGFR4

Liver WAT Macrophage

Less TG uptake CD36
Leptin Low intake oxLDL
Small adipocytes Foam cells
FXR VDR FGF15
VDR
Endothelium
FXR
TGR5
eNOS TGR5
AKT
GLP-1 TGR5
BAT Muscle
PXR FXR βKlotho Intestine
VDR FGFR1c/FGFR4 DIO2 energy
expenditure
SMC Insulin sensitivity
DIO2 energy
Cyps Cell migration Role for FGF19? expenditure
induction &
detoxification

Figure 25.3  FXR and TGR5 control lipid and glucose homeostasis. FXR activation in the liver promotes triglycerides (TG) clearance (green),
reduces TG synthesis (red). Concomitantly, FXR activation increases glycogen and protein synthesis (green) thus helping to control glycemia. FXR
and VDR activation in the intestine release FGF19 in humans (Fgf15 in mice) which via the portal vein circulates to the liver where it binds to its
receptor FGFR4 and βKlotho to repress bile acid synthesis. TGR5 activation in L‐cells produces the glucagon‐like peptide 1 (GLP‐1), which regu-
lates glucose sensitivity and insulin resistance. FXR activation in endothelium increases smooth muscle cell (SMC) motility in vitro and the
endothelial nitric oxide synthase expression (eNOS). FGFR4, βKlotho, and FGFR1c are expressed in smooth muscle cells but the FGF15/19 role
is not yet clear. FGF15/19 signals in white adipose tissue (WAT) where it decreases TG uptake and adipocyte size and increases leptin expression.
TGR5 activation in brown adipose tissue (BAT) increases energy expenditure via the induction of the expression of the type II iodothyronine deio-
dinase (DIO2). In muscles TGR5 activation increases AKT signaling and increases DIO2 expression, energy expenditure, and insulin sensitivity. In
macrophages TGR5 activation and FGF15/19 signaling via FGFR1c/βKlotho lowers CD36 expression, reducing oxidized LDL (ox‐LDL) intake
and thus foam cells formation. In the brain FXR and TGR5 regulates neurotransmitters. GLP‐1 resulting from TGR5 activation in the gut increases
insulin sensitivity and FGF15/19 signaling increases the metabolic rate and improves glucose metabolism.

impairs HNF4α binding to the CYP7A1 promoter by reducing
the interaction of PGC‐1α with HNF4α [75] (Figure  25.2).
FXR induces the expression of peroxisome proliferator acti-
vated receptor alpha (PPARα; NR1C1) in humans [76] and
PPARα in turn reduces CYP7A1 transcription via reduced
HNF4α binding, effects which might contribute to the risk of
gallstone formation after treatment with PPARα activators
such as fibrates but could also contribute to their anti‐choles-
tatic effects [1, 77] (Figure 25.2).
Another enzyme, CYP8B1, controls the relative hydropho-
bicity of the bile acid pool through CDCA synthesis and
Cyp8b1‐deficient mice lack the negative feedback on Cyp7a1
[78] due to increased levels of FXR antagonists such as αMCA,
βMCA, and UDCA (Table 25.1) [79, 80]. Lrh‐1‐deficient mice
have reduced Cyp8b1 expression and low CA levels in their bile
acid pool composition, therefore identifying Lrh‐1 as the central
regulator of Cyp8b1 expression [71, 72], also indirectly control-
ling Cyp7a1 expression by generating FXR antagonists.
Regulation of phase I and II bile acid metabolism
by bile acid‐activated nuclear receptors
Bile acid hydroxylation (phase I) and conjugation (phase II)
makes them more hydrophilic and facilitates their excretion into
urine as an additional/alternative route for bile acid elimination
under cholestatic conditions [81]. Human CYP3A4 and the
rodent Cyp3a11 are key cytochrome P450 enzymes for bile acid
oxidation (see earlier). CYP3A4 expression is induced by FXR,
PXR, VDR, and CAR, providing  –  together with sulfation or
glucuronidation – a mechanism to limit hepatocellular damage
by accumulating bile acids (Figure 25.2).
304 THE LIVER:  BILE ACID‐ACTIVATED NUCLEAR RECEPTORS: FXR, PXR, CAR, VDR, AND GR
SULT2A1, the dehydroepiandrosterone sulfotransferase, cat-
alyzes sulfo‐conjugation of bile acids (an important bile acid
detoxification pathway during cholestasis in humans) and is
induced by FXR, PXR, VDR, and CAR [1]. Bile acid glucuro-
nidation is catalyzed by the UDP‐glucuronosyltransferases
UGT2B4 and UGT2B7 [26, 82]. Bile acids induce human
UGT2B4 via activation of FXR [26], but UGT2B4 is also acti-
vated by PPARα agonists [82]. Since FXR upregulates PPARα
and its target genes expression [76], bile acids could induce
UGT2B4 expression directly via FXR and indirectly via FXR‐
mediated induction of PPARα.
Regulation of bile acid transport by nuclear receptors
Basolateral hepatocellular bile acid uptake
Hepatic bile acid uptake is mediated by a high‐affinity Na+‐
dependent transporter NTCP (SLC10A1) and a family of multi‐
specific organic anion transporters (OATPs; SLC21A),
mediating the quantitatively less important Na+‐independent
uptake of conjugated or free bile acids mediated by facilitated
exchange with intracellular anions [83] (Figure 25.2).
Regulation of NTCP by bile acids differs considerably among
humans, mice, and rats [84]. Negative feedback inhibition of
Ntcp is mediated via Fxr‐Shp‐dependent or ‐independent mech-
anisms to limit bile acid uptake [85, 86]. Induction of Shp by
bile acid‐activated FXR inhibits the activity of Hnf4α [87]
(Figure  25.2). In humans, the NTCP promoter lacks HNF4α
response elements [84] and SHP acts mainly by suppressing
GRα‐mediated activation of NTCP [88]. Similar to NTCP,
repression of the sodium‐independent bile acid uptake system in
humans, OATP1B1, is also mediated by FXR, involving SHP,
HNF4α, and HNF1α [89].
Canalicular bile acid excretion
Canalicular excretion of bile acids is the rate‐limiting step in
bile formation. Bile acids such as CA, CDCA, and UDCA and
their tauro/glycoconjugate are excreted into the bile canaliculus
via the bile salt export pump BSEP (ABCB11) (Figure  25.2).
Human, rat, and mouse BSEP promoters are transcriptionally
activated by FXR [1, 90] and Bsep baseline expression is
reduced in Fxr‐knockout mice [28] and humans with FXR loss
of function [31]. A role for VDR in BSEP repression via direct
VDR–FXR interaction in vitro has also been postulated [91].
Bile acids with two negative charges such as sulfated tauro‐
or glyco‐LCA are transported by multidrug resistance‐associ-
ated protein MRP2 (ABCC2) [92]. MRP2 is regulated by several
NRs, reflecting its diverse substrate spectrum. FXR binds to
FXREs in the promoter, shared with CAR and PXR [93]
(Figure 25.2).
Additional hepatobiliary transport systems in the canalicular
membrane include a phospholipid floppase (MDR3/Mdr2 in
rodents) for phosphatidylcholine translocation, the cholesterol
two half‐transporters ABCG5/8 for sitosterol and cholesterol
export, a P‐type ATPase (Fic1/FIC1; ATP8B1) whose substrate/
function is still unclear and an Cl−/HCO3− anion exchanger 2
(SLC4A2/AE2), all of them involved in bile formation [1, 81].
FXR enhances human MDR3 transcription, while PPARα stim-
ulates the expression of rodent Mdr2 and MDR3; since FXR
induces PPARα in human, FXR could also activate MDR3 indi-
rectly via PPARα (Figure 25.2) [1]. FXR also induces ABCG5/
G8 expression, resulting in increased cholesterol excretion in
bile [1].
Canalicular bile acid efflux is thus mainly mediated via bile
acids requiring FXR/CAR/PXR/VDR as a key transcription fac-
tor coordinating a balanced bile composition.
Alternative basolateral bile acid export
Basolateral bile acid excretion back into portal blood represents
an alternative route for bile acid elimination during cholestasis
(when canalicular excretion is blocked). Alternative basolateral
bile acid export is mediated by MRP3 and MRP4 and the
organic solute transporters OSTα and OSTβ [89] (Figure 25.2).
These export systems are expressed at low levels under normal
conditions but can be significantly upregulated in cholestasis
[89]. Since MRP3, MRP4, and OSTα/OSTβ are able to trans-
port sulfated and glucuronidated bile acids, the induction of
these transporters may facilitate renal excretion of these bile
acids in patients with chronic cholestasis [89].
Induction of both Mrp3 and Mrp4 is independent of Fxr in
rodent models of cholestasis [94, 95] (Figure  25.2). PXR and
VDR are able to induce human and rodent MRP3 expression
[96, 97], while CAR ligands induced both human and rodent
MRP3 and MRP4 [81]. The bile acid transporter OSTα/β is
induced by FXR and baseline Ostα/Ostβ expression is reduced
in Fxr‐knockout mice [1]. Taken together, similar to canalicular
transporters, multiple NRs (FXR, PXR, VDR, and CAR) coor-
dinate basolateral bile acid efflux.
Bile acid‐activated nuclear receptor signaling
in lipid/glucose metabolism, endothelial
function, and inflammation
High‐density lipoprotein metabolism
High‐density lipoproteins (HDLs) are heterogeneous lipid parti-
cles containing phospholipids, cholesterol, sphingosine 1 phos-
phate (S1P), and apolipoprotein A‐I as their major structural
proteins. HDLs transport cholesterol from the periphery back to
the liver where it can be excreted as free cholesterol or, after
conversion into bile acids, into bile to be eliminated via feces in
a process called “reverse cholesterol transport” (Figure  25.1).
Resins (such as cholestyramine), bind to bile acids in the intes-
tine, resulting in bile acid malabsorption and increasing HDL
cholesterol levels, while dietary bile acid supplementation low-
ers HDL cholesterol [98]. This can be explained by repression
of apoA‐I gene expression by FXR, resulting in reduced hepatic
HDL synthesis (Figure 25.3) [25]. FXR activation also increases
cholesteryl ester transfer protein (CETP) mass and activity [99]
as well as phospholipid transfer protein (PLTP) in vitro [98].
PLTP transfers phospholipids from triglycerides to HDL,
whereas CETP catalyzes exchange of triglycerides and choles-
terol esters between HDL and apoB lipoproteins and CETP
activity correlates with cardiovascular disease risk. This,
together with lowering HDL cholesterol and increasing LDL
cholesterol (see later) (Figure  25.3) may impact on long‐term
cardiovascular safety of FXR agonists [98].
 25:  Bile Acids as Signaling Molecules 305
Triglyceride and low‐density lipoprotein metabolism
Bile acid supplementation to patients lowered serum triglycer-
ides [98], while Fxr‐deficient mice are hypertriglyceridemic due
to an increase of the lipoprotein lipase (LPL) activity inhibitor
apoC‐III and lowering of the LPL activator apoC‐II [28, 100].
FXR (via induction of SHP) also reduces de novo lipogenesis by
repressing Srebp1c [101] and reduces very low‐density lipopro-
tein (VLDL) synthesis via repression of HNF4α transcriptional
activity on the promoter of the microsomal triglyceride transfer
protein (MTTP) [102]. Conversely FXR increases the expres-
sion of the VLDL receptor (VLDLR) [103] and PPARα as well
as β‐oxidation [76] (Figure  25.3). Lipoprotein(a) (Lp(a)) is a
lipoprotein formed via the covalent binding of apoA with
apoB‐100 in LDL and high plasma Lp(a) levels increased ath-
erogenicity [104]. FXR directly lowers Lp(a) gene expression
and indirectly via FGF19‐mediated mechanisms [105, 106].
FXR activation in humanized mouse with chimeric livers also
increased LDL cholesterol levels due to SREBP2 repression,
which in turn reduces LDL receptor expression in hepatocytes
and LDL uptake by the liver [107]. In conclusion, bile acids are
hypotriglyceridemic through lowering the production and
increasing clearance of triglycerides, but conversely increase
LDL cholesterol (Figure 25.3).
Glucose metabolism
FXR deletion in mouse resulted in insulin resistance coupled
with glucose intolerance; whereas treatment with FXR agonist
lowered blood glucose by repressing hepatic gluconeogenesis
and enhancing glycogen synthesis and storage (Figure  25.3)
[108], partially via FGF19 [63,109]. Moreover, absence of FXR
increased plasma free fatty acids levels and glucose production
in liver, thus impairing peripheral glucose disposal in mice
[110]. Finally, FXR enhances glucose‐stimulated insulin secre-
tion in pancreatic beta‐cells [111]. Altogether, FXR activation
counteracts insulin resistance and glucose intolerance, which
may be of interest for the treatment of type 2 diabetes mellitus
and related/associated disorders (Figure 25.3).
Signaling in endothelium and smooth muscle cells
Vascular smooth muscle cells express FXR [112], where it con-
trols PLTP expression, apoptosis, and attenuates inflammation
by inducing SHP expression which in turn interferes with NFκB
signaling [113] (Figure  25.3). In addition, FXR activation
reduces platelet activation/aggregation in response to thrombin
and collagen [114]. FGFR4 and FGFR1c are also expressed in
smooth muscle cells in human arteries [115], together with
βKlotho, where it mediates FGF21 signaling [116], while the
relevance of FGF19 signaling in vascular wall is still unclear
(Figure 25.3). FXR also regulates vascular tension by control-
ling endothelial nitric oxide synthase eNOS (Figure 25.3) [117].
Since chronic liver disease leads to portal hypertension and cir-
rhosis, FXR activation in sinusoidal endothelial cells reduces
intrahepatic vascular resistance and thereby portal pressure by
increasing eNOS and reducing endothelin 1 [118]. VDR is also
expressed in endothelial cells (outside the liver) and its disrup-
tion promotes a pro-inflammatory phenotype with induction of
interleukin 6 (IL‐6) and VCAM‐1, resulting in enhanced rolling
and plaque accumulation favoring the development of
atherosclerosis [119]. Like FXR, VDR also induces eNOS in
endothelial cells [120] (Figure 25.3).
Inflammation
FXR, CAR, PXR, and VDR are all negative acute phase genes
(i.e. their expression and signaling are reduced during the acute
phase response) [121–123]. FXR stimulation by bile acids (e.g.
accumulating during sepsis‐induced cholestasis) is anti‐inflam-
matory by antagonizing nuclear factor‐κB (NF‐κB) signaling
[124]. In addition, FXR antagonizes JNK signaling during
inflammation by inducing superoxide dismutase 3 (SOD3)
[125]. Conversely, binding of NF‐κB to well‐conserved binding
sites in the promoters of FXR target genes resulted in repression
of BSEP, ABCG5/G8, and MRP2, while promoters of OSTα
and OSTβ were unexpectedly activated rather than repressed.
This may suggest that activation of NF‐κB could help to main-
tain bile acid excretion via OSTα/β into the blood during inflam-
mation/cholestasis [126]. Activation of FXR in the small
intestine and colon attenuates interleukin responses (IL‐6,
IL‐1β), controls microbiota, maintains intestinal integrity, and
overall ameliorates inflammation [127, 128]. VDR and FXR
upregulate cathelicidin expression in liver, biliary ducts, and
intestine to control bacterial growth [129]. Gut microbiota also
regulates FXR activation in intestine by controlling the release
of the FXR antagonist βMCA (see earlier) [79, 130].
G PROTEIN‐COUPLED‐RECEPTOR
FOR BILE ACIDS TAKEDA G PROTEIN‐
COUPLED RECEPTOR 5
Discovery and properties
In addition to NRs, bile acids also activate G protein‐coupled
receptors (GPCRs) such as TGR5 (also see Chapter  25). Two
groups cloned a membrane receptor which was shown to be
activated by bile acid and named BG37 or TGR5 [131, 132].
TGR5 is activated mainly by LCA or DCA, with a divergent
potency from FXR [131, 132] (Figure 25.1 and Table 25.1) and
by bile acid concentrations achieved in post‐prandial state, dur-
ing cholestasis and portosystemic shunting or therapeutic bile
acid therapy. Notably, TGR5 is expressed in tissues where FXR
is absent, such as lung, spleen, leukocytes, white (WAT) and
brown adipose tissues (BAT) [131, 132], with its highest expres-
sion in colon and gallbladder [133] (for more details see
Chapter 24).
Metabolism and energy homeostasis
Mice fed with high‐fat diet containing bile acids were protected
from obesity – despite preserved food intake – which was attrib-
uted to increased energy expenditure [134]. TGR5 activation by
bile acid induced cAMP production, resulting in upregulation of
deiodinase 2 (D2) expression, which converts inactive thyroid
hormone T4 into the active T3 form. T3, by activating the NR
thyroid hormone receptor alpha (TRα1, NR1A1), induced
306 THE LIVER:  SPHINGOSINE 1 PHOSPHATE RECEPTOR 2, CHOLINERGIC RECEPTOR MUSCARINIC 3
uncoupling protein 1 (UCP1) expression, thus increasing energy
expenditure in BAT and skeletal muscle [134] (Figure 25.3), an
effect also principally found in humans [135]. TGR5 also stimu-
lates production of glucagon‐like peptide 1 (GLP‐1) in intesti-
nal enteroendocrine cells and therefore contributes to control of
glucose homeostasis [136]. The use of bile acid‐binding resins
(cholestyramine, colestipol, colesevelam), resulting in activa-
tion of intestinal TGR5 by shifting bile acid signaling to the
colon by inhibition of ileal bile acid absorption, improved the
control of glycemia in diabetic patients and hypercholester-
olemia [98] and reduced hepatic glycogenolysis and glucose
production in mice [137] (Figure 25.3). Bile acid‐binding resins
have beneficial metabolic effects by depleting hepatic choles-
terol via stimulation of bile acid synthesis/CYP7A1 activity and
improving glucose control by GLP‐1 induction via TGR5.
Therefore activating TGR5 might be an attractive approach to
treat metabolic conditions such as diabetes, fatty liver, or
dyslipidemia.
Anti‐inflammatory effects of TGR5
Tgr5‐deficient mice are more susceptible to liver inflamma-
tion induced by LPS [138]. TGR5 inhibits inflammatory
cytokine production by c‐FOS/p65 in macrophages [139],
while pharmacological stimulation of TGR5 reduced myeloid
cell activation and autoimmune encephalitis [140] and pre-
vented atherosclerosis by lowering macrophage plaque con-
tent and inflammatory cytokine production [141, 142]. In
addition, TGR5 repressed CD36 and SR‐A expression in mac-
rophages, thus reducing uptake of oxidized LDL [141]
(Figure  25.3). FXR/TGR5 coactivation also enhanced the
intrahepatic content of alternatively activated macrophages
and anti‐inflammatory monocytes in mice with NAFLD
[143]. Such anti‐inflammatory effects could contribute to the
beneficial effects of pharmacological TGR5 stimulation in
cholestatic and metabolic disorders (see Chapter  24 for fur-
ther details).
Role of TGR5 in the biliary tree
In human cholangiocytes, TGR5 colocalizes with the chloride
channel cystic fibrosis transmembrane conductance regulator
(CFTR) and the apical sodium‐dependent bile salt uptake
transporter (ASBT) in the apical membrane of gallbladder epi-
thelial cells, where it mediates chloride secretion via CFTR
[144]. TGR5 is also localized to primary cilium of cholangio-
cytes, where it regulates ductular bile formation by increasing
bile acid reabsorption and water secretion [145]. Targeting
TGR5 in inflammatory biliary diseases might thus be tempting
but TGR5 has recently been implicated in the pathogenesis of
pruritus [146].
Moreover, TGR5 also promotes cell proliferation in
response to bile acid in adenocarcinoma [147] and induces
cardiac hypertrophy in cardiomyocytes [148], TGR5 is also a
neurosteroid receptor in brain involved in the generation of
reactive oxygen species (ROS) [149]. Therefore the quest for
safe TGR5 modulators with therapeutic applications for
cholestatic and metabolic disorders may be challenging (also
see Chapter 24).
FORMYL PEPTIDE RECEPTORS
Formyl peptide receptors (FPRs) are G‐coupled receptors
activated by bacterial chemotactic formylated peptides and play
a key role in sensing of bacteria and mediating leukocyte recruit-
ment to infection sites. N‐Formyl‐methionyl‐leucylphenylala-
nine (fMLP) is one of the most powerful ligands activating the
high‐affinity FPR and formyl peptide receptor like 1 (FPRL1)
(Figure 25.1) [150]. Cholestatic patients showed an impairment
of FPR and FPRL1 activity [151], since bile acids such as
CDCA and CA act as FPR and FPRL1 antagonists, while
UDCA has only a rather weak inhibitory effect [151]. Bile acids
prevent binding of FPR and FPRL1 with fMLP by steric hin-
drance and glycoconjugated CDCA was identified as the most
efficient competitor [151, 152]. Bile acid antagonism of FPR/
FPRL1 could thus contribute to immunosuppressive effects of
bile acids [153, 154]. During the acute phase response up to
25% of the proteins produced in the liver are accounted for by
serum amyloid A (SAA), which also binds and activates FPRL1.
FPRL1 induces cyclooxygenase 2 (COX2) and synthesis of
prostaglandins, activating the peroxisome proliferator receptor
γ (PPARγ; NR1C3) (Figure 25.1) [155]. How bile acid signaling
is involved in crosstalk with SAA/FPRL1 and synthesis of
PPARγ agonists during cholestatic conditions is still unknown
(Figure 25.1) [150].
SPHINGOSINE 1 PHOSPHATE
RECEPTOR 2, CHOLINERGIC RECEPTOR
MUSCARINIC 3, AND EPIDERMAL
GROWTH FACTOR RECEPTOR
Several bile acid‐activated GPCRs share a common signaling
pathway which can result in activation of EGFR signaling and
are therefore summarized in this section. Bile acids such as
DCA can activate the sphingosine 1 phosphate receptor 2
(S1PR2), the muscarinic receptor 3 (CHMR3), or the epidermal
growth factor receptor (EGFR) in the cell membrane
(Figure 25.1). Moreover, S1PR2 and CHMR3 activate the phos-
pholipase Cβ (PLCβ) which in turn activates membrane metal-
loproteinase (MMP). releasing EGF from the membrane‐bound
precursor pro‐EGF, allowing EGF to bind to EGFR (Figure 25.1)
Sphingosine 1 phosphate receptor 2
Sphingosine can be phosphorylated by the sphingosine kinase 1
and 2 (SPHK1 and SPHK2) to generate bioactive S1P. S1P in
turn activates specific G protein‐coupled receptors known as
sphingosine 1 phosphate receptors (S1PR) 1 to 5. While S1PR1
is ubiquitously expressed, S1PR2 is mainly found in liver, kid-
ney, heart, brain, lung, and vascular smooth muscle cells [156].
S1PRs activate ERK, AKT, JNK, and SPHK [156]. S1PR2 acti-
vation increases pro-inflammatory cytokine (IL‐1β, TNF‐α)
secretion and S1pr2‐deficient mice have reduced TNF‐α and
IL‐1β levels [157]. After a meal, bile acid activation of hepatic
S1pr2 leads to nuclear S1P accumulation – via SphK2 activity –
inhibiting histone deacetylase‐regulating enzymes and NRs
 25:  Bile Acids as Signaling Molecules 307
involved in nutrient sensing and metabolism. In line with this
finding, S1pr2‐knockout mice develop a fatty liver under high‐
fat diet [158]. S1pr2 signaling also stimulates phospholipase C
β (PLCβ), which in turn activates MMP, releasing EGF from the
membrane, activating its receptor EGFR (Figure  25.1) [156,
159]. HDL carries the pro-inflammatory sphingolipid S1P
[160]. Interestingly, bile acid signaling (via FXR) reduces
apoA‐I production and thereby circulating HDL carrying S1P,
which could contribute to anti‐inflammatory bile acid effects by
inhibiting the S1PR2 pathway. Several S1PR modulators are
currently being developed to treat autoimmune disorders, such
as ozanimod in ulcerative colitis; although these do not target
S1PR2 [161], a bile acid‐based scaffold might be used to
­antagonize S1PR2.
Cholinergic receptor muscarinic 3
Bile acids can activate the cholinergic receptor muscarinic 3
(CHRM3), a G‐coupled receptor that is naturally activated by
acetylcholine [162]. Similar to S1PR2, CHMR3 and CHMR1
activate PLC [163] and thereby EGFR (Figure 25.1 and see ear-
lier). TCA was found to activate the CHMR2 in rat cardiomyo-
cytes, possibly contributing to bile acid‐induced arrhythmia
[164]. Moreover, CHMR3 signaling influences bile formation
and cholestasis: Mice lacking Chmr3 have a modestly reduced
bile flow and biliary bicarbonate secretion and absence of Chm3
enhances the susceptibility to cholestatic injury in DDC‐fed
mice. Moreover, treatment of Mdr2−/− with a CHMR3 agonist
decreased liver injury, indicating that CHMR3 signaling may
represent a therapeutic target in cholangiopathies [165].
Epidermal growth factor
EGF binds and activates the EGFR, stimulating cell growth and
differentiation; as such, EGFR signaling is also involved in hep-
atocarcinogenesis [166]. Hydrophobic bile acids such as DCA
stimulate hepatocyte apoptosis by inducing phosphorylation of
the death receptor CD95 via EGFR [167]. DCA can directly
activate EGFR signaling, probably involving plasma membrane
alterations leading to EGFR activation [168] (Figure 25.1), an
effect which is counteracted by UDCA [169]. Bile acid‐induced
EGFR activation in quiescent rat hepatic stellate cells (HSCs)
can trigger both proliferation and apoptosis. EGF stimulation of
quiescent primary rat HSCs results in increased migration with-
out proliferation [170], whereas stimulation of EGFR by hydro-
phobic bile acids leads to an NADPH oxidase‐driven ROS
generation followed by a Yes‐mediated EGFR activation and a
shift of the proliferative signal to an apoptotic signal when JNK
is costimulated [171]. These observations may be important in
understanding bile acid effects in the development of biliary
fibrosis.
INTEGRINS
Tauro‐UDCA (TUDCA) also signals via α5β1‐integrins, chang-
ing its β1 subunit conformation to activate the focal adhesion
kinases, subsequently increasing choleresis by inserting NTCP
and BSEP into the plasma membrane of hepatocytes [172].
TUDCA also raises cAMP levels and thus activates the protein
kinase A (PKA) protecting hepatocytes against apoptosis by
internalizing CD95, thus preventing its interaction and activa-
tion with pro‐apoptotic bile acids such as glyco‐CDCA [173].
KINASES: PHOSPHATIDYLINOSITOL
3‐KINASE AND PROTEIN KINASE C
Bile acids such as TCA activate phosphatidylinositol 3‐kinase
(PI3K), which promotes the insertion of transporters such as rat
Mrp2 and Bsep into the canalicular membrane, thereby increas-
ing biliary bile acid excretion [174, 175]. Since TCA activates
FXR as transcriptional regulator of MRP2 (see earlier), it would
be of interest to explore potential crosstalks between FXR and
PI3K in regulation of MRP2 and Bsep.
TUDCA stimulates calcium‐dependent exocytosis into bile
partially by activating PKCα [176] and enhances biliary excre-
tion by inserting Mrp2 into the canalicular membrane [177],
which has also been mechanistically linked to PKCα and PKA
activation [178].
CYTOTOXIC EFFECTS OF BILE ACIDS
Bile acid cytotoxicity depends not only on their concentrations
but also on their hydrophobicity [7]. Bile acids can induce cell
injury by stimulating apoptosis via the TNF‐related apoptosis‐
inducing ligand (TRAIL). GCDCA stimulates TRAIL
downstream signaling such as caspases 3, 8, and 10, and phos-
phorylates the cellular FLICE inhibitory protein (cFLIP) via
PKC, which then is bound to the Fas‐associated death domain
protein (FADD) and initiates cell apoptosis [179].
Secondary bile acid, such as DCA, activate the transmem-
brane receptor FASR, which results in mitochondrial damage
and ROS generation. [180, 181]. DCA activation of PLA2 leads
to production of arachidonic acid (AA), a pro-inflammatory
fatty acid generating ROS, while DCA stimulation of NADPH
oxidase also results in ROS [180, 181]. Activation of membrane
by DCA leads to inositol 1,4,5‐trisphosphate (IP3) and diacylg-
lycerol (DAG) production, which can promote endoplasmic
reticulum (ER) stress, release of the BCL2‐antagonist/killer 1
(BAK) and mitochondrial damage (Figure  25.4) [180, 181].
Bile acid accumulation in cholestasis also stimulates ERK
­activity, activating the early growth factor (EGR1) leading to
production of pro-inflammatory cytokines, adhesion mole-
cule expression, and amino acid synthesis all promoting liver
injury (Figure 25.4) [182].
Finally, bile acid can also induce necrosis. Human hepato-
cytes exposed to G‐CDCA at concentrations found in serum of
patients with cholestasis developed necrosis without increase in
apoptotic parameters such as cleaved cytokeratin 18. Patients
with cholestasis displayed elevations in serum full‐length
cytokeratin 18, high mobility group box 1 protein (HMGB1),
and acetylated HMGB1, all hallmarks of necrotic inflammation
[183]. Therefore, in contrast to mouse hepatocytes, human
308 THE LIVER:  REFERENCES

BA

BA

NADPHox PLA2 FASR PLC

ARA CASP8 IP3 ERK

Mitochondrial
ROS BAK ER stress
Cytosol damage

NFκB
Nucleus
EGR1

Inflammatory Inflammatory
genes genes
DNA DNA
damage repair systems

Figure 25.4  Bile acids (BA) induce ER stress, mitochondrial damage, and inflammation. Bile acid (e.g. DCA)‐mediated stimulation of NADPH
oxidase or PLA2 leads to reactive oxygen species (ROS) generation. BA activation of the receptor FASR releases caspase 8 (CASP8), leading to
mitochondrial damage with ROS production. ROS promotes DNA damage and activates NFκB, inhibiting DNA repair and promoting inflamma-
tion. BA stimulation of PLC releases inositol 1,4,5‐trisphosphate (IP3), driving endoplasmic reticulum (ER) stress, release of the BCL2‐antagonist/
killer 1 (BAK), and mitochondrial damage. BA also activate ERK, which stimulates the early growth factor (EGR1) transcriptional activity, which
induces pro-inflammatory cytokine secretion, adhesion molecules, and arachidonic acid synthesis.

hepatocytes are more resistant to human bile acid species and disease. Therefore, understanding bile acid signaling will most
are likely to preferentially die of necrosis. likely be highly relevant for understanding the pathobiology and
management of diseases of the liver and beyond.

CONCLUSIONS
The new insights into the signaling functions of bile acids via a
broad range of membrane and nuclear receptors has profoundly
changed our understanding of liver physiology and the patho-
physiology of liver diseases. Bile acid signaling is not only criti-
cal for maintaining bile acid homeostasis by providing feedback
and feedforward loops regulating bile acid transport, synthesis,
and metabolism, but as “enterohepatic hormones” bile acids
also coordinate several key aspects of lipid and glucose metabo-
lism, intestinal microbiota, and even some key aspects of
inflammation and immunity. Such broad implications of bile
acid signaling also have important potential implications for
treatment of cholestatic and metabolic liver diseases. However,
we have to keep in mind that many of the data rely either on in
vitro or mouse studies and in vivo data in humans are still scarce.
Therefore, the next challenge will be to elucidate and integrate
the contributions of individual bile acid signaling pathways in
human physiology and pathophysiology. Bile acid‐activated G
protein‐coupled receptors and NRs offer ample opportunities to
modulate hepatic metabolism, inflammation, and liver fibrosis
and several ligands have already shown promising effects in
preclinical or clinical studies. Moreover, changes in bile acid
composition and signaling are emerging as important biomark-
ers at the interface of host and microbial metabolism in liver
REFERENCES
1. Claudel, T., Zollner, G., Wagner, M., and Trauner, M. Role of nuclear recep-
tors for bile acid metabolism, bile secretion, cholestasis, and gallstone dis-
ease. Biochim Biophys Acta, 2011;1812(8):867–78.
2. Hofmann, A.F. and Hagey, L.R. Key discoveries in bile acid chemistry and
biology and their clinical applications: history of the last eight decades. J
Lipid Res, 2014;55(8):1553–95.
3. Takahashi, S., Fukami, T., Masuo, Y. et al. Cyp2c70 is responsible for the
species difference in bile acid metabolism between mice and humans. J Lipid
Res, 2016;57(12):2130–7.
4. Jia, W., Xie, G., and Jia, W. Bile acid‐microbiota crosstalk in gastrointestinal
inflammation and carcinogenesis. Nat Rev Gastroenterol Hepatol,
2018;15(2):111–28.
5. Wahlstrom, A., Sayin, S.I., Marschall, H.U., and Backhed, F. Intestinal
crosstalk between bile acids and microbiota and its impact on host metabo-
lism. Cell Metab, 2016;24(1):41–50.
6. Ikegami, T. and Honda, A. Reciprocal interactions between bile acids and gut
microbiota in human liver diseases. Hepatol Res, 2018;48(1):15–27.
7. Jansen, P.L., Ghallab, A., Vartak, N. et al. The ascending pathophysiology of
cholestatic liver disease. Hepatology, 2017;65(2):722–38.
8. Forman, B.M., Goode, E., Chen, J. et al. Identification of a nuclear recep-
tor that is activated by farnesol metabolites. Cell, 1995;81(5): 687–93.
9. Seol, W., Choi, H.S., and Moore, D.D. Isolation of proteins that interact spe-
cifically with the retinoid X receptor: two novel orphan receptors. Mol
Endocrinol, 1995;9(1):72–85.
10. Makishima, M., Okamoto, A.Y., Repa, J.J. et al. Identification of a nuclear
receptor for bile acids. Science, 1999;284(5418):1362–5.
 25:  Bile Acids as Signaling Molecules 309
11. Parks, D.J., Blanchard, S.G., Bledsoe, R.K. et al. Bile acids: natural ligands
for an orphan nuclear receptor. Science, 1999;284(5418):1365–8.
12. Wang, H., Chen, J., Hollister, K., Sowers, L.C., and Forman, B.M.
Endogenous bile acids are ligands for the nuclear receptor FXR/BAR. Mol
Cell, 1999;3(5):543–53.
13. Cui, J., Heard, T.S., Yu, J. et al. The amino acid residues asparagine 354 and
isoleucine 372 of human farnesoid X receptor confer the receptor with high
sensitivity to chenodeoxycholate. J Biol Chem, 2002;277(29):25963–9.
14. Downes, M., Verdecia, M.A., Roecker, A.J. et al. A chemical, genetic, and
structural analysis of the nuclear bile acid receptor FXR. Mol Cell,
2003;11(4):1079–92.
15. Zhao, A., Yu, J., Lew, J.L. et al. Polyunsaturated fatty acids are FXR ligands
and differentially regulate expression of FXR targets. DNA Cell Biol,
2004;23(8):519–26.
16. Cieslak, A., Trottier, J., Verreault, M. et al. N‐3 Polyunsaturated fatty acids
stimulate bile acid detoxification in human cell models. Can J Gastroenterol
Hepatol, 2018;2018:6031074.
17. Pellicciari, R., Fiorucci, S., Camaioni, E. et al. 6Alpha‐ethyl‐chenodeoxy-
cholic acid (6‐ECDCA), a potent and selective FXR agonist endowed with
anticholestatic activity. J Med Chem, 2002;45(17):3569–72.
18. Maloney, P.R., Parks, D.J., Haffner, C.D. et al. Identification of a chemical tool
for the orphan nuclear receptor FXR. J Med Chem, 2000;43(16): 2971–4.
19. Baghdasaryan, A., Claudel, T., Gumhold, J. et al. Dual farnesoid X receptor/
TGR5 agonist INT‐767 reduces liver injury in the Mdr2‐/‐ (Abcb4‐/‐) mouse
cholangiopathy model by promoting biliary HCO output. Hepatology,
2011;54(4):1303–12.
20. Liu, Y., Binz, J., Numerick, M.J. et al. Hepatoprotection by the farnesoid X
receptor agonist GW4064 in rat models of intra‐ and extrahepatic cholesta-
sis. J Clin Invest, 2003;112(11):1678–87.
21. Nevens, F., Andreone, P., Mazzella, G. et al. A placebo‐controlled trial of obet-
icholic acid in primary biliary cholangitis. N Engl J Med, 2016;375(7):631–43.
22. Mudaliar, S., Henry, R.R., Sanyal, A.J. et al. Efficacy and safety of the farnesoid
X receptor agonist obeticholic acid in patients with type 2 diabetes and nonalco-
holic fatty liver disease. Gastroenterology, 2013;145(3):574–82 e571.
23. Xu, Y. Recent progress on bile acid receptor modulators for treatment of
metabolic diseases. J Med Chem, 2016;59(14):6553–79.
24. Zavacki, A.M., Lehmann, J.M., Seol, W. et al. Activation of the orphan recep-
tor RIP14 by retinoids. Proc Natl Acad Sci U S A, 1997;94(15):7909–14.
25. Claudel, T., Sturm, E., Duez, H. et al. Bile acid‐activated nuclear receptor
FXR suppresses apolipoprotein A‐I transcription via a negative FXR
response element. J Clin Invest, 2002;109(7):961–71.
26. Barbier, O., Torra, I.P., Sirvent, A. et al. FXR induces the UGT2B4 enzyme
in hepatocytes: a potential mechanism of negative feedback control of FXR
activity. Gastroenterology, 2003;124(7):1926–40.
27. Chong, H.K., Infante, A.M., Seo, Y.K. et al. Genome‐wide interrogation of
hepatic FXR reveals an asymmetric IR‐1 motif and synergy with LRH‐1.
Nucleic Acids Res, 2010;38(18):6007–17.
28. Sinal, C.J., Tohkin, M., Miyata, M. et al. Targeted disruption of the nuclear
receptor FXR/BAR impairs bile acid and lipid homeostasis. Cell,
2000;102(6):731–44.
29. Kim, I., Ahn, S.H., Inagaki, T. et  al. Differential regulation of bile acid
homeostasis by the farnesoid X receptor in liver and intestine. J Lipid Res,
2007;48(12):2664–72.
30. Zhang, Y., Kast‐Woelbern, H.R., and Edwards, P.A. Natural structural vari-
ants of the nuclear receptor farnesoid X receptor affect transcriptional activa-
tion. J Biol Chem, 2003;278(1):104–10.
31. Gomez‐Ospina, N., Potter, C.J., Xiao, R. et al. Mutations in the nuclear bile
acid receptor FXR cause progressive familial intrahepatic cholestasis. Nat
Commun, 2016;7:10713.
32. Van Mil, S.W., Milona, A., Dixon, P.H. et al. Functional variants of the cen-
tral bile acid sensor FXR identified in intrahepatic cholestasis of pregnancy.
Gastroenterology, 2007;133(2):507–16.
33. Heni, M., Wagner, R., Ketterer, C. et al. Genetic variation in NR1H4 encod-
ing the bile acid receptor FXR determines fasting glucose and free fatty acid
levels in humans. J Clin Endocrinol Metab, 2013;98(7):E1224–9.
34. Otte, K., Kranz, H., Kober, I. et al. Identification of farnesoid X receptor beta
as a novel mammalian nuclear receptor sensing lanosterol. Mol Cell Biol,
2003;23(3):864–72.
35. Staudinger, J., Liu, Y., Madan, A., Habeebu, S., and Klaassen, C.D.
Coordinate regulation of xenobiotic and bile acid homeostasis by pregnane
X receptor. Drug Metab Dispos, 2001;29(11):1467–72.
36. Xie, W., Radominska‐Pandya, A., Shi, Y. et al. An essential role for nuclear
receptors SXR/PXR in detoxification of cholestatic bile acids. Proc Natl
Acad Sci U S A, 2001;98(6):3375–80.
37. Zhang, J., Huang, W., Qatanani, M., Evans, R.M., and Moore, D.D. The con-
stitutive androstane receptor and pregnane X receptor function coordinately
to prevent bile acid‐induced hepatotoxicity. J Biol Chem, 2004;279(47):
49517–22.
38. Schuetz, E.G., Strom, S., Yasuda, K. et al. Disrupted bile acid homeostasis
reveals an unexpected interaction among nuclear hormone receptors, trans-
porters, and cytochrome P450. J Biol Chem, 2001;276(42):39411–18.
39. Uppal, H., Toma, D., Saini, S.P. et  al. Combined loss of orphan receptors
PXR and CAR heightens sensitivity to toxic bile acids in mice. Hepatology,
2005;41(1):168–76.
40. Kliewer, S.A., Moore, J.T., Wade, L. et al. An orphan nuclear receptor acti-
vated by pregnanes defines a novel steroid signaling pathway. Cell,
1998;92(1):73–82.
41. Lehmann, J.M., McKee, D.D., Watson, M.A. et al. The human orphan nuclear
receptor PXR is activated by compounds that regulate CYP3A4 gene expres-
sion and cause drug interactions. J Clin Invest, 1998;102(5):1016–23.
42. Bertilsson, G., Heidrich, J., Svensson, K. et  al. Identification of a human
nuclear receptor defines a new signaling pathway for CYP3A induction.
Proc Natl Acad Sci U S, A, 1998;95(21):12208–13.
43. Blumberg, B., Sabbagh, W., Jr., Juguilon, H. et al. SXR, a novel steroid and
xenobiotic‐sensing nuclear receptor. Genes Dev, 1998;12(20):3195–205.
44. Watkins, R.E., Wisely, G.B., Moore, L.B. et al. The human nuclear xenobi-
otic receptor PXR: structural determinants of directed promiscuity. Science,
2001;292(5525):2329–33.
45. Baes, M., Gulick, T., Choi, H.S. et al. A new orphan member of the nuclear
hormone receptor superfamily that interacts with a subset of retinoic acid
response elements. Mol Cell Biol, 1994;14(3):1544–52.
46. Goodwin, B., Redinbo, M.R., and Kliewer, S.A. Regulation of cyp3a gene
transcription by the pregnane x receptor. Annu Rev Pharmacol Toxicol,
2002;42:1–23.
47. Willson, T.M. and Kliewer, S.A. PXR, CAR and drug metabolism. Nat Rev
Drug Discov, 2002;1(4):259–66.
48. Beilke, L.D., Aleksunes, L., Holland, R. et al. Car‐mediated changes in bile
acid composition contributes to hepatoprotection from Lca‐induced liver
injury in mice. Drug Metab Dispos, 2009;37(5):1035–45.
49. Jones, G., Strugnell, S.A., and DeLuca, H.F. Current understanding of the
molecular actions of vitamin, D. Physiol Rev, 1998;78(4):1193–231.
50. Makishima, M., Lu, T.T., Xie, W. et al. Vitamin D receptor as an intestinal
bile acid sensor. Science, 2002;296(5571):1313–16.
51. Nishida, S., Ozeki, J., and Makishima, M. Modulation of bile acid metabo-
lism by 1{alpha}‐hydroxyvitamin D3 administration in mice. Drug Metab
Dispos, 2009;37(10):2037–44.
52. Hollenberg, S.M., Weinberger, C., Ong, E.S. et  al. Primary structure and
expression of a functional human glucocorticoid receptor cDNA. Nature,
1985;318(6047):635–41.
53. Oakley, R.H., Jewell, C.M., Yudt, M.R., Bofetiado, D.M., and Cidlowski,
J.A. The dominant negative activity of the human glucocorticoid receptor
beta isoform. Specificity and mechanisms of action. J Biol Chem,
1999;274(39):27857–66.
54. Tanaka, H. and Makino, I. Ursodeoxycholic acid‐dependent activation of the
glucocorticoid receptor. Biochem Biophys Res Commun, 1992;
188(2):942–8.
55. Weitzel, C., Stark, D., Kullmann, F. et  al. Ursodeoxycholic acid induced
activation of the glucocorticoid receptor in primary rat hepatocytes. Eur J
Gastroenterol Hepatol, 2005;17(2):169–77.
56. Calmus, Y., Gane, P., Rouger, P., and Poupon, R. Hepatic expression of class
I and class II major histocompatibility complex molecules in primary biliary
cirrhosis: effect of ursodeoxycholic acid. Hepatology, 1990;11(1):12–15.
57. Trauner, M., Halilbasic, E., Kazemi‐Shirazi, L. et al. Therapeutic role of bile
acids and nuclear receptor agonists in fibrosing cholangiopathies. Dig Dis,
2014;32(5):631–6.
58. Takigawa, T., Miyazaki, H., Kinoshita, M. et  al. Glucocorticoid receptor‐
dependent immunomodulatory effect of ursodeoxycholic acid on liver lym-
phocytes in mice. Am J Physiol Gastrointest Liver Physiol,
2013;305(6):G427–38.
59. Sharma, R., Prichard, D., Majer, F. et al. Ursodeoxycholic acid amides as
novel glucocorticoid receptor modulators. J Med Chem,
2011;54(1):122–30.
310 THE LIVER:  REFERENCES
60. Inagaki, T., Choi, M., Moschetta, A. et al. Fibroblast growth factor 15 func-
tions as an enterohepatic signal to regulate bile acid homeostasis. Cell
Metab, 2005;2(4):217–25.
61. Holt, J.A., Luo, G., Billin, A.N. et al. Definition of a novel growth factor‐
dependent signal cascade for the suppression of bile acid biosynthesis.
Genes Dev, 2003;17(13):1581–91.
62. Babaknejad, N., Nayeri, H., Hemmati, R., Bahrami, S., and Esmaillzadeh, A.
An overview of FGF19 and FGF21: the therapeutic role in the treatment of
the metabolic disorders and obesity. Horm Metab Res, 2018;50(6):441–52.
63. Kir, S., Beddow, S.A., Samuel, V.T. et al. FGF19 as a postprandial, insulin‐
independent activator of hepatic protein and glycogen synthesis. Science,
2011;331(6024):1621–4.
64. Chiang, J.Y. Regulation of bile acid synthesis: pathways, nuclear receptors,
and mechanisms. J Hepatol, 2004;40(3):539–51.
65. Goodwin, B., Jones, S.A., Price, R.R. et  al. A regulatory cascade of the
nuclear receptors FXR, SHP‐1, and LRH‐1 represses bile acid biosynthesis.
Mol Cell, 2000;6(3):517–26.
66. Lu, T.T., Makishima, M., Repa, J.J. et al. Molecular basis for feedback regu-
lation of bile acid synthesis by nuclear receptors. Mol Cell,
2000;6(3):507–15.
67. Hayhurst, G.P., Lee, Y.H., Lambert, G., Ward, J.M., and Gonzalez, F.J.
Hepatocyte nuclear factor 4alpha (nuclear receptor 2A1) is essential for
maintenance of hepatic gene expression and lipid homeostasis. Mol Cell
Biol, 2001;21(4):1393–403.
68. Kir, S., Zhang, Y., Gerard, R.D., Kliewer, S.A., and Mangelsdorf, D.J.
Nuclear receptors HNF4alpha and LRH‐1 cooperate in regulating Cyp7a1 in
vivo. J Biol Chem, 2012;287(49):41334–41.
69. Kerr, T.A., Saeki, S., Schneider, M. et  al. Loss of nuclear receptor SHP
impairs but does not eliminate negative feedback regulation of bile acid syn-
thesis. Dev Cell, 2002;2(6):713–20.
70. Wang, L., Lee, Y.K., Bundman, D. et al. Redundant pathways for negative
feedback regulation of bile acid production. Dev Cell, 2002;2(6):721–31.
71. Mataki, C., Magnier, B.C., Houten, S.M. et al. Compromised intestinal lipid
absorption in mice with a liver‐specific deficiency of liver receptor homolog
1. Mol Cell Biol, 2007;27(23):8330–9.
72. Lee, Y.K., Schmidt, D.R., Cummins, C.L. et al. Liver receptor homolog‐1
regulates bile acid homeostasis but is not essential for feedback regulation of
bile acid synthesis. Mol Endocrinol, 2008;22(6):1345–56.
73. Jung, D. and Kullak‐Ublick, G.A. Hepatocyte nuclear factor 1 alpha: a key
mediator of the effect of bile acids on gene expression. Hepatology,
2003;37(3):622–31.
74. De Fabiani, E., Mitro, N., Gilardi, F. et al. Coordinated control of cholesterol
catabolism to bile acids and of gluconeogenesis via a novel mechanism of
transcription regulation linked to the fasted‐to‐fed cycle. J Biol Chem,
2003;278(40):39124–32.
75. Li, T. and Chiang, J.Y. Mechanism of rifampicin and pregnane X receptor
inhibition of human cholesterol 7 alpha‐hydroxylase gene transcription. Am
J Physiol Gastrointest Liver Physiol, 2005;288(1):G74–84.
76. Pineda Torra, I., Claudel, T., Duval, C. et al. Bile acids induce the expression
of the human peroxisome proliferator‐ activated receptor alpha gene via acti-
vation of the farnesoid X receptor. Mol Endocrinol, 2003;17(2):259–72.
77. Cuperus, F.J., Claudel, T., Gautherot, J., Halilbasic, E., and Trauner, M. The
role of canalicular ABC transporters in cholestasis. Drug Metab Dispos,
2014;42(4):546–60.
78. Li‐Hawkins, J., Lund, E.G., Turley, S.D., and Russell, D.W. Disruption of
the oxysterol 7alpha‐hydroxylase gene in mice. J Biol Chem,
2000;275(22):16536–42.
79. Sayin, S.I., Wahlstrom, A., Felin, J. et al. Gut microbiota regulates bile acid
metabolism by reducing the levels of tauro‐beta‐muricholic acid, a naturally
occurring FXR antagonist. Cell Metab, 2013;17(2):225–35.
80. Hu, X., Bonde, Y., Eggertsen, G., and Rudling, M. Muricholic bile acids are
potent regulators of bile acid synthesis via a positive feedback mechanism. J
Intern Med, 2014;275(1):27–38.
81. Halilbasic, E., Claudel, T., and Trauner, M. Bile acid transporters and regulatory
nuclear receptors in the liver and beyond. J Hepatol, 2013;58(1):155–68.
82. Barbier, O., Duran‐Sandoval, D., Pineda‐Torra, I. et al. Peroxisome prolif-
erator‐activated receptor alpha induces hepatic expression of the human bile
acid glucuronidating UDP‐glucuronosyltransferase 2B4 enzyme. J Biol
Chem, 2003;278(35):32852–60.
83. Trauner, M. and Boyer, J.L. Bile salt transporters: molecular characteriza-
tion, function, and regulation. Physiol Rev, 2003;83(2):633–71.
84. Jung, D., Hagenbuch, B., Fried, M., Meier, P.J., and Kullak‐Ublick, G.A.
Role of liver‐enriched transcription factors and nuclear receptors in regulat-
ing the human, mouse, and rat NTCP gene. Am J Physiol Gastrointest Liver
Physiol, 2004;286(5):G752–61.
85. Zollner, G., Fickert, P., Silbert, D. et  al. Induction of short heterodimer
partner 1 precedes downregulation of Ntcp in bile duct‐ligated mice. Am J
Physiol Gastrointest Liver Physiol, 2002;282(1):G184–91.
86. Geier, A., Zollner, G., Dietrich, C.G. et al. Cytokine‐independent repres-
sion of rodent Ntcp in obstructive cholestasis. Hepatology, 2005;
41(3):470–7.
87. Lee, Y.K., Dell, H., Dowhan, D.H., Hadzopoulou‐Cladaras, M., and Moore,
D.D. The orphan nuclear receptor SHP inhibits hepatocyte nuclear factor 4
and retinoid X receptor transactivation: two mechanisms for repression.
Mol Cell Biol, 2000;20(1):187–95.
88. Eloranta, J.J., Jung, D., and Kullak‐Ublick, G.A. The human Na+‐taurocho-
late cotransporting polypeptide gene is activated by glucocorticoid receptor
and peroxisome proliferator‐activated receptor‐gamma coactivator‐1alpha,
and suppressed by bile acids via a small heterodimer partner‐dependent
mechanism. Mol Endocrinol, 2006;20(1):65–79.
89. Zollner, G., Marschall, H.U., Wagner, M., and Trauner, M. Role of nuclear
receptors in the adaptive response to bile acids and cholestasis: pathoge-
netic and therapeutic considerations. Mol Pharm, 2006;3(3):231–51.
90. Ananthanarayanan, M., Balasubramanian, N., Makishima, M.,
Mangelsdorf, D.J., and Suchy, F.J. Human bile salt export pump promoter
is transactivated by the farnesoid X receptor/bile acid receptor. J Biol
Chem, 2001;276(31):28857–65.
91. Chow, E.C., Durk, M.R., Cummins, C.L., and Pang, K.S. 1α,25‐
Dihydroxyvitamin D3 up‐regulates P‐glycoprotein via the vitamin D recep-
tor and not farnesoid X receptor in both fxr(‐/‐) and fxr(+/+) mice and
increased renal and brain efflux of digoxin in mice in vivo. J Pharmacol
Exp Ther, 2011;337(3):846–59.
92. Keppler, D. and Konig, J. Hepatic secretion of conjugated drugs and endog-
enous substances. Semin Liver Dis, 2000;20(3):265–72.
93. Kast, H.R., Goodwin, B., Tarr, P.T. et al. Regulation of multidrug resist-
ance‐associated protein 2 (MRP2;ABCC2) by the nuclear receptors PXR,
FXR, and CAR. J Biol Chem, 2001;12:12.
94. Wagner, M., Fickert, P., Zollner, G. et al. Role of farnesoid X receptor in
determining hepatic ABC transporter expression and liver injury in bile
duct‐ligated mice. Gastroenterology, 2003;125(3):825–38.
95. Zollner, G., Fickert, P., Fuchsbichler, A. et  al. Role of nuclear bile acid
receptor, FXR, in adaptive ABC transporter regulation by cholic and urso-
deoxycholic acid in mouse liver, kidney and intestine. J Hepatol,
2003;39(4):480–8.
96. Teng, S., Jekerle, V., and Piquette‐Miller, M. Induction of ABCC3 (MRP3)
by pregnane X receptor activators. Drug Metab Dispos, 2003;
31(11):1296–9.
97. McCarthy, T.C., Li, X., and Sinal, C.J. Vitamin D receptor‐dependent regu-
lation of colon multidrug resistance‐associated protein 3 gene expression
by bile acids. J Biol Chem, 2005;280(24):23232–42.
98. Claudel, T., Staels, B., and Kuipers, F. The farnesoid X receptor: a molecu-
lar link between bile acid and lipid and glucose metabolism. Arterioscler
Thromb Vasc Biol, 2005;25(10):2020–30.
99. Gautier, T., de Haan, W., Grober, J. et al. Farnesoid X receptor activation
increases cholesteryl ester transfer protein expression in humans and trans-
genic mice. J Lipid Res, 2013;54(8):2195–205.
100. Claudel, T., Inoue, Y., Barbier, O. et al. Farnesoid X receptor agonists sup-
press hepatic apolipoprotein CIII expression. Gastroenterology,
2003;125(2):544–55.
101. Watanabe, M., Houten, S.M., Wang, L. et al. Bile acids lower triglyceride
levels via a pathway involving FXR, SHP, and SREBP‐1c. J Clin Invest,
2004;113(10):1408–18.
102. Hirokane, H., Nakahara, M., Tachibana, S., Shimizu, M., and Sato, R. Bile
acid reduces the secretion of very low density lipoprotein by repressing
microsomal triglyceride transfer protein gene expression mediated by
hepatocyte nuclear factor‐4. J Biol Chem, 2004;279(44):45685–92.
103. Sirvent, A., Claudel, T., Martin, G. et al. The farnesoid X receptor induces
very low density lipoprotein receptor gene expression. FEBS Lett,
2004;566(1–3):173–7.
104. Tregouet, D.A., Konig, I.R., Erdmann, J. et  al. Genome‐wide haplotype
association study identifies the SLC22A3‐LPAL2‐LPA gene cluster as a
risk locus for coronary artery disease. Nat Genet, 2009;41(3):283–5.
 25:  Bile Acids as Signaling Molecules 311
105. Chennamsetty, I., Claudel, T., Kostner, K.M. et  al. Farnesoid X receptor
represses hepatic human APOA gene expression. J Clin Invest,
2011;121(9):3724–34.
106. Chennamsetty, I., Claudel, T., Kostner, K.M., Trauner, M., and Kostner,
G.M. FGF19 signaling cascade suppresses APOA gene expression.
Arterioscler Thromb Vasc Biol, 2012;32(5):1220–7.
107. Papazyan, R., Liu, X., Liu, J. et al. FXR activation by obeticholic acid or
nonsteroidal agonists induces a human‐like lipoprotein cholesterol change
in mice with humanized chimeric liver. J Lipid Res, 2018;59(6):982–93.
108. Ma, K., Saha, P.K., Chan, L., and Moore, D.D. Farnesoid X receptor is essen-
tial for normal glucose homeostasis. J Clin Invest, 2006;116(4):1102–9.
109. Potthoff, M.J., Boney‐Montoya, J., Choi, M. et  al. FGF15/19 regulates
hepatic glucose metabolism by inhibiting the CREB‐PGC‐1alpha pathway.
Cell Metab, 2011;13(6):729–38.
110. Zhang, Y., Lee, F.Y., Barrera, G. et al. Activation of the nuclear receptor
FXR improves hyperglycemia and hyperlipidemia in diabetic mice. Proc
Natl Acad Sci U S A, 2006;103(4):1006–11.
111. Seyer, P., Vallois, D., Poitry‐Yamate, C. et al. Hepatic glucose sensing is
required to preserve beta cell glucose competence. J Clin Invest,
2013;123(4):1662–76.
112. Bishop‐Bailey, D., Walsh, D.T., and Warner, T.D. Expression and activation
of the farnesoid X receptor in the vasculature. Proc Natl Acad Sci U S A,
2004;101(10):3668–73.
113. Li, Y.T., Swales, K.E., Thomas, G.J., Warner, T.D., and Bishop‐Bailey, D.
Farnesoid x receptor ligands inhibit vascular smooth muscle cell inflamma-
tion and migration. Arterioscler Thromb Vasc Biol, 2007;27(12):2606–11.
114. Moraes, L.A., Unsworth, A.J., Vaiyapuri, S. et al. Farnesoid X receptor and
its ligands inhibit the function of platelets. Arterioscler Thromb Vasc Biol,
2016;36(12):2324–33.
115. Hughes, S.E. Localisation and differential expression of the fibroblast
growth factor receptor (FGFR) multigene family in normal and atheroscle-
rotic human arteries. Cardiovasc Res, 1996;32(3):557–69.
116. Cao, F., Wang, S., Cao, X. et  al. Fibroblast growth factor 21 attenuates
calcification of vascular smooth muscle cells in vitro. J Pharm Pharmacol,
2017;69(12):1802–16.
117. Zhang, R., Ran, H., Peng, L. et al. Farnesoid X receptor regulates vasoreac-
tivity via angiotensin II type 2 receptor and the kallikrein‐kinin system in
vascular endothelial cells. Clin Exp Pharmacol Physiol, 2016;43(3):327–34.
118. Schwabl, P., Hambruch, E., Seeland, B.A. et al. The FXR agonist PX20606
ameliorates portal hypertension by targeting vascular remodelling and sinu-
soidal dysfunction. J Hepatol, 2017;66(4):724–33.
119. Bozic, M., Alvarez, A., de Pablo, C. et al. Impaired vitamin D signaling in
endothelial cell leads to an enhanced leukocyte‐endothelium interplay: impli-
cations for atherosclerosis development. PLoS One, 2015;10(8):e0136863.
120. Molinari, C., Uberti, F., Grossini, E. et al. 1alpha,25‐dihydroxycholecalcif-
erol induces nitric oxide production in cultured endothelial cells. Cell
Physiol Biochem, 2011;27(6):661–8.
121. Kim, M.S., Shigenaga, J., Moser, A., Feingold, K., and Grunfeld, C.
Repression of farnesoid X receptor during the acute phase response. J Biol
Chem, 2003;278(11):8988–95.
122. Beigneux, A.P., Moser, A.H., Shigenaga, J.K., Grunfeld, C., and Feingold,
K.R. Reduction in cytochrome P‐450 enzyme expression is associated with
repression of CAR (constitutive androstane receptor) and PXR (pregnane X
receptor) in mouse liver during the acute phase response. Biochem Biophys
Res Commun, 2002;293(1):145–9.
123. Pramanik, R., Asplin, J.R., Lindeman, C. et al. Lipopolysaccharide nega-
tively modulates vitamin D action by down‐regulating expression of vita-
min D‐induced VDR in human monocytic THP‐1 cells. Cell Immunol,
2004;232(1–2):137–43.
124. Wang, Y.D., Chen, W.D., Wang, M. et al. Farnesoid X receptor antagonizes
nuclear factor kappaB in hepatic inflammatory response. Hepatology,
2008;48(5):1632–43.
125. Wang, Y.D., Chen, W.D., Li, C. et  al. Farnesoid X receptor antagonizes
JNK signaling pathway in liver carcinogenesis by activating SOD3. Mol
Endocrinol, 2015;29(2):322–31.
126. Balasubramaniyan, N., Ananthanarayanan, M., and Suchy, F.J. Nuclear factor‐
kappaB regulates the expression of multiple genes encoding liver transport
proteins. Am J Physiol Gastrointest Liver Physiol, 2016;310(8):G618–28.
127. Vavassori, P., Mencarelli, A., Renga, B., Distrutti, E., and Fiorucci, S. The
bile acid receptor FXR is a modulator of intestinal innate immunity. J
Immunol, 2009;183(10):6251–61.
128. Inagaki, T., Moschetta, A., Lee, Y.K. et  al. Regulation of antibacterial
defense in the small intestine by the nuclear bile acid receptor. Proc Natl
Acad Sci U S A, 2006;103(10):3920–5.
129. D’Aldebert, E., Biyeyeme Bi Mve, M.J., Mergey, M. et al. Bile salts control
the antimicrobial peptide cathelicidin through nuclear receptors in the
human biliary epithelium. Gastroenterology, 2009;136(4):1435–43.
130. Zhang, L., Xie, C., Nichols, R.G. et  al. Farnesoid X receptor signaling
shapes the gut microbiota and controls hepatic lipid metabolism. mSystems,
2016;1(5).
131. Kawamata, Y., Fujii, R., Hosoya, M. et  al. A G protein‐coupled receptor
responsive to bile acids. J Biol Chem, 2003;278(11):9435–40.
132. Maruyama, T., Miyamoto, Y., Nakamura, T. et al. Identification of mem-
brane‐type receptor for bile acids (M‐BAR). Biochem Biophys Res
Commun, 2002;298(5):714–19.
133. Maruyama, T., Tanaka, K., Suzuki, J. et al. Targeted disruption of G pro-
tein‐coupled bile acid receptor 1 (Gpbar1/M‐Bar) in mice. J Endocrinol,
2006;191(1):197–205.
134. Watanabe, M., Houten, S.M., Mataki, C. et  al. Bile acids induce energy
expenditure by promoting intracellular thyroid hormone activation. Nature,
2006;439(7075):484–9.
135. Broeders, E.P., Nascimento, E.B., Havekes, B. et al. The bile acid chenode-
oxycholic acid increases human brown adipose tissue activity. Cell Metab,
2015;22(3):418–26.
136. Thomas, C., Gioiello, A., Noriega, L. et al. TGR5‐mediated bile acid sens-
ing controls glucose homeostasis. Cell Metab, 2009;10(3):167–77.
137. Potthoff, M.J., Potts, A., H.e., T. et al. Colesevelam suppresses hepatic gly-
cogenolysis by TGR5‐mediated induction of GLP‐1 action in DIO mice.
Am J Physiol Gastrointest Liver Physiol, 2013;304(4):G371–80.
138. Wang, Y.D., Chen, W.D., Yu, D., Forman, B.M., and Huang, W. The G‐pro-
tein coupled bile acid receptor Gpbar1 (TGR5) negatively regulates hepatic
inflammatory response through antagonizing nuclear factor kappaB.
Hepatology, 2011;54(4):1421–32.
139. Yoneno, K., Hisamatsu, T., Shimamura, K. et al. TGR5 signalling inhibits
the production of pro‐inflammatory cytokines by in vitro differentiated
inflammatory and intestinal macrophages in Crohn’s disease. Immunology,
2013;139(1):19–29.
140. Lewis, N.D., Patnaude, L.A., Pelletier, J. et al. A GPBAR1 (TGR5) small
molecule agonist shows specific inhibitory effects on myeloid cell activa-
tion in vitro and reduces experimental autoimmune encephalitis (EAE) in
vivo. PLoS One, 2014;9(6):e100883.
141. Pols, T.W., Nomura, M., Harach, T. et al. TGR5 activation inhibits athero-
sclerosis by reducing macrophage inflammation and lipid loading. Cell
Metab, 2011;14(6):747–57.
142. Kida, T., Tsubosaka, Y., Hori, M., Ozaki, H., and Murata, T. Bile acid
receptor TGR5 agonism induces NO production and reduces monocyte
adhesion in vascular endothelial cells. Arterioscler Thromb Vasc Biol,
2013;33(7):1663–9.
143. McMahan, R.H., Wang, X.X., Cheng, L.L. et al. Bile acid receptor activa-
tion modulates hepatic monocyte activity and improves nonalcoholic fatty
liver disease. J Biol Chem, 2013;288(17):11761–70.
144. Keitel, V., Cupisti, K., Ullmer, C. et  al. The membrane‐bound bile acid
receptor TGR5 is localized in the epithelium of human gallbladders.
Hepatology, 2009;50(3):861–70.
145. Keitel, V., Ullmer, C., and Haussinger, D. The membrane‐bound bile acid
receptor TGR5 (Gpbar‐1) is localized in the primary cilium of cholangio-
cytes. Biol Chem, 2010;391(7):785–9.
146. Alemi, F., Kwon, E., Poole, D.P. et al. The TGR5 receptor mediates bile
acid‐induced itch and analgesia. J Clin Invest, 2013;123(4):1513–30.
147. Cao, W., Tian, W., Hong, J. et al. Expression of bile acid receptor TGR5 in
gastric adenocarcinoma. Am J Physiol Gastrointest Liver Physiol,
2013;304(4):G322–7.
148. Desai, M.S., Shabier, Z., Taylor, M. et  al. Hypertrophic cardiomyopathy
and dysregulation of cardiac energetics in a mouse model of biliary fibrosis.
Hepatology, 2010;51(6):2097–107.
149. Keitel, V., Gorg, B., Bidmon, H.J. et al. The bile acid receptor TGR5 (Gpbar‐1)
acts as a neurosteroid receptor in brain. Glia, 2010;58(15):1794–805.
150. Le, Y., Oppenheim, J.J., and Wang, J.M. Pleiotropic roles of formyl peptide
receptors. Cytokine Growth Factor Rev, 2001;12(1):91–105.
151. Chen, X., Yang, D., Shen, W. et al. Characterization of chenodeoxycholic
acid as an endogenous antagonist of the G‐coupled formyl peptide recep-
tors. Inflamm Res, 2000;49(12):744–55.
312 THE LIVER:  REFERENCES
152. Ferrari, C., Macchiarulo, A., Costantino, G., and Pellicciari, R.
Pharmacophore model for bile acids recognition by the FPR receptor. J
Comput Aided Mol Des, 2006;20(5):295–303.
153. Calmus, Y., Guechot, J., Podevin, P. et al. Differential effects of chenode-
oxycholic and ursodeoxycholic acids on interleukin 1, interleukin 6 and
tumor necrosis factor‐alpha production by monocytes. Hepatology,
1992;16(3):719–23.
154. Calmus, Y., Weill, B., Ozier, Y. et al. Immunosuppressive properties of che-
nodeoxycholic and ursodeoxycholic acids in the mouse. Gastroenterology,
1992;103(2):617–21.
155. Li, H., Ooi, S.Q., and Heng, C.K. The role of NF‐kappaB in SAA‐induced
peroxisome proliferator‐activated receptor gamma activation. Atherosclerosis,
2013;227(1):72–8.
156. Kwong, E., Li, Y., Hylemon, P.B., and Zhou, H. Bile acids and sphingo-
sine‐1‐phosphate receptor 2 in hepatic lipid metabolism. Acta Pharm Sin B,
2015;5(2):151–7.
157. Skoura, A., Michaud, J., Im, D.S. et  al. Sphingosine‐1‐phosphate recep-
tor‐2 function in myeloid cells regulates vascular inflammation and athero-
sclerosis. Arterioscler Thromb Vasc Biol, 2011;31(1):81–5.
158. Nagahashi, M., Takabe, K., Liu, R. et  al. Conjugated bile acid‐activated
S1P receptor 2 is a key regulator of sphingosine kinase 2 and hepatic gene
expression. Hepatology, 2015;61(4):1216–26.
159. Dent, P., Fang, Y., Gupta, S. et al. Conjugated bile acids promote ERK1/2
and AKT activation via a pertussis toxin‐sensitive mechanism in murine
and human hepatocytes. Hepatology, 2005;42(6):1291–9.
160. Sachinidis, A., Kettenhofen, R., Seewald, S. et al. Evidence that lipopro-
teins are carriers of bioactive factors. Arterioscler Thromb Vasc Biol,
1999;19(10):2412–21.
161. Dyckman, A.J. Modulators of sphingosine‐1‐phosphate pathway biology:
recent advances of sphingosine‐1‐phosphate receptor 1 (S1P1) agonists and
future perspectives. J Med Chem, 2017;60(13):5267–89.
162. Raufman, J.P., Chen, Y., Cheng, K. et al. Selective interaction of bile acids
with muscarinic receptors: a case of molecular mimicry. Eur J Pharmacol,
2002;457(2–3):77–84.
163. Raufman, J.P., Cheng, K., and Zimniak, P. Activation of muscarinic recep-
tor signaling by bile acids: physiological and medical implications. Dig Dis
Sci, 2003;48(8):1431–44.
164. Sheikh Abdul Kadir, S.H., Miragoli, M., Abu‐Hayyeh, S. et al. Bile acid‐
induced arrhythmia is mediated by muscarinic M2 receptors in neonatal rat
cardiomyocytes. PLoS One, 2010;5(3):e9689.
165. Durchschein, F., Krones, E., Pollheimer, M.J. et al. Genetic loss of the mus-
carinic M3 receptor markedly alters bile formation and cholestatic liver
injury in mice. Hepatol Res, 2018;48(3):E68–E77.
166. Lanaya, H., Natarajan, A., Komposch, K. et al. EGFR has a tumour‐pro-
moting role in liver macrophages during hepatocellular carcinoma forma-
tion. Nat Cell Biol, 2014;16(10):972–7.
167. Reinehr, R., Graf, D., and Haussinger, D. Bile salt‐induced hepatocyte
apoptosis involves epidermal growth factor receptor‐dependent CD95
tyrosine phosphorylation. Gastroenterology, 2003;125(3):839–53.
168. Qiao, L., Studer, E., Leach, K. et  al. Deoxycholic acid (DCA) causes
ligand‐independent activation of epidermal growth factor receptor (EGFR)
and FAS receptor in primary hepatocytes: inhibition of EGFR/mitogen‐
activated protein kinase‐signaling module enhances DCA‐induced apopto-
sis. Mol Biol Cell, 2001;12(9):2629–45.
169. Im, E. and Martinez, J.D. Ursodeoxycholic acid (UDCA) can inhibit deox-
ycholic acid (DCA)‐induced apoptosis via modulation of EGFR/Raf‐1/
ERK signaling in human colon cancer cells. J Nutr, 2004;134(2):483–6.
170. Yang, C., Zeisberg, M., Mosterman, B. et al. Liver fibrosis: insights into
migration of hepatic stellate cells in response to extracellular matrix and
growth factors. Gastroenterology, 2003;124(1):147–59.
171. Sommerfeld, A., Reinehr, R., and Haussinger, D. Bile acid‐induced epidermal
growth factor receptor activation in quiescent rat hepatic stellate cells can trig-
ger both proliferation and apoptosis. J Biol Chem, 2009;284(33):22173–83.
172. Gohlke, H., Schmitz, B., Sommerfeld, A., Reinehr, R., and Haussinger, D.
alpha5 beta1‐integrins are sensors for tauroursodeoxycholic acid in hepato-
cytes. Hepatology, 2013;57(3):1117–29.
173. Sommerfeld, A., Reinehr, R., and Haussinger, D. Tauroursodeoxycholate
protects rat hepatocytes from bile acid‐induced apoptosis via beta1‐integ-
rin‐ and protein kinase a‐dependent mechanisms. Cell Physiol Biochem,
2015;36(3):866–83.
174. Misra, S., Ujhazy, P., Gatmaitan, Z., Varticovski, L., and Arias, I.M. The
role of phosphoinositide 3‐kinase in taurocholate‐induced trafficking of
ATP‐dependent canalicular transporters in rat liver. J Biol Chem,
1998;273(41):26638–44.
175. Misra, S., Varticovski, L., and Arias, I.M. Mechanisms by which cAMP
increases bile acid secretion in rat liver and canalicular membrane vesicles.
Am J Physiol Gastrointest Liver Physiol, 2003;285(2):G316–24.
176. Beuers, U., Throckmorton, D.C., Anderson, M.S. et al. Tauroursodeoxycholic
acid activates protein kinase C in isolated rat hepatocytes. Gastroenterology,
1996;110(5):1553–63.
177. Beuers, U., Bilzer, M., Chittattu, A. et  al. Tauroursodeoxycholic acid
inserts the apical conjugate export pump, Mrp2, into canalicular mem-
branes and stimulates organic anion secretion by protein kinase C‐
dependent mechanisms in cholestatic rat liver. Hepatology,
2001;33(5):1206–16.
178. Wimmer, R., Hohenester, S., Pusl, T. et  al. Tauroursodeoxycholic acid
exerts anticholestatic effects by a cooperative cPKC alpha‐/PKA‐dependent
mechanism in rat liver. Gut, 2008;57(10):1448–54.
179. Higuchi, H., Yoon, J.H., Grambihler, A. et al. Bile acids stimulate cFLIP
phosphorylation enhancing TRAIL‐mediated apoptosis. J Biol Chem,
2003;278(1):454–61.
180. Barrasa, J.I., Olmo, N., Lizarbe, M.A., and Turnay, J. Bile acids in the colon,
from healthy to cytotoxic molecules. Toxicol In Vitro, 2013;27(2):964–77.
181. Bernstein, H., Bernstein, C., Payne, C.M., Dvorakova, K., and Garewal, H.
Bile acids as carcinogens in human gastrointestinal cancers. Mutat Res,
2005;589(1):47–65.
182. Allen, K., Jaeschke, H., and Copple, B.L. Bile acids induce inflammatory
genes in hepatocytes: a novel mechanism of inflammation during obstruc-
tive cholestasis. Am J Pathol, 2011;178(1):175–86.
183. Woolbright, B.L., Dorko, K., Antoine, D.J. et al. Bile acid‐induced necrosis
in primary human hepatocytes and in patients with obstructive cholestasis.
Toxicol Appl Pharmacol, 2015;283(3):168–77.
 Hepatic Adenosine
26 Triphosphate‐Binding
Cassette Transport Proteins
and Their Role in Physiology
Peter L.M. Jansen
Department of Hepatology and Gastroenterology, Academic Medical Center, Amsterdam, The Netherlands
and LiSyM Research Network, University of Freiburg, Germany
HISTORY
Bile flow in the liver is driven by active transport of organic ions
across the canalicular membrane of hepatocytes, followed by
secretion of water in the bile ductules [1]. Studies in rodents
identified bile salts as the main drivers of bile secretion [2]. Bile
flow generated by the active secretion of bile salts, is called
“bile salt dependent” bile flow [3]. A “bile salt independent”
fraction of bile flow has been proposed based on linear regres-
sion analysis indicating that when flow and bile salt secretion
are extrapolated to zero bile salt secretion, there still is consider-
able bile flow. This suggests a flow of bile [3] resulting from
secretion of glutathione and other non‐bile salt organic anions
across the canalicular membrane as well as bicarbonate and
water secretion in the small intrahepatic bile ductules [4]. The
magnitude of the bile salt independent fraction of bile flow is
species dependent and in humans is about one third of the
total bile flow [5]. However, the concept of “bile salt independent
bile flow” has been challenged on the basis that extrapolation
of bile flow and bile acid secretion may not be linear and fails to
account for the effect of aquaporins [6]. Intrahepatic water
secretion was first thought to occur via leaky tight junctions but
discovery of aquaporins revealed that water secretion into bile
ductules is by aquaporins driven by osmotic pressure differ-
ences [7].
Studies in which increasing amounts of bile salts were intra-
venously administered to intact animals or into isolated per-
fused livers, showed that bile salt secretion and bile flow
increased to a maximum called “biliary secretory maximum” or
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
apparent Tm (transport maximum) which suggests that bile salt
secretion is rate‐limiting and may involve specific transport pro-
teins [8]. Mathematical modeling revealed that these transport
processes behave like enzyme‐mediated activities and follow
Michaelis–Menten kinetics with a specific Km and Vmax (Km is
the substrate affinity at half maximal transport rate, Vmax the
maximal transport rate of a substrate).
Until 1989, the prevailing view was that bile acid secretion is
driven by the electrochemical gradient (−35mV) generated by
the Na+K+‐ATPase residing in the basolateral plasma domain of
the hepatocyte. A change in paradigm occurred following the
demonstration that ATP hydrolysis directly drives transport of
organic anions out of canalicular membrane vesicles [9]. Studies
with canalicular membrane vesicles revealed that bile acid and
organic anion transport is directly dependent on ATP hydrolysis
[10, 11]. In vivo studies in wild type and mutant sheep, monkeys
and rats confirmed the presence of distinct ATP‐dependent
transport processes for bile acids and non‐bile acid organic ani-
ons [12–14].
Proof for the functional activity of more than one canalicular
transport protein came from animals with natural or experimen-
tally induced mutations. For instance, sheep, rats, and monkeys
with genetically impaired canalicular secretion of bilirubin con-
jugates showed preserved bile salt and organic cation secretion
[11, 12, 14, 15]. Definitive proof had to wait for the cloning of
the affected transporters: ABCC2/MRP2 for non‐bile salt
organic anions, ABCB11/BSEP (sister of Pgp) for bile salts,
ABCB1 (Pgp, MDR1) [16–18], for organic cations and some
organic anions. Additional evidence for multiple transport
314 THE LIVER:  ABC TRANSPORTERS
proteins was obtained by the targeted deletion of genes encoding
separate proteins for transport of bilirubin, bile salts, and
phosphatidylcholine (PC) in mice [19–21].
PHYSIOLOGY OF BILE FLOW
The idea that organic ion transport drives bile flow needs refine-
ment in view of current insights in the 3D reconstruction of so‐
called liver “lobules”. Liver lobules consist of a number of liver
acini. A liver acinus is the smallest functional unit of the liver.
Acini are composed of hepatocytes, sinusoids, and bile canali-
culi. Blood flows from the periportal to the pericentral area of
the liver acinus and bile from the pericentral to the periportal
area. Whether bile really flows in the bile canaliculus remains
hypothetical until proved by techniques such as dynamic intra-
vital microscopy that enable observations at the subcellular
level in live tissues.
The hepatic artery and portal vein deliver blood to the liver
sinusoids. The sinusoidal perfusion of blood from the portal to
the central vein passes about 15 hepatocytes per acinus. These
hepatocytes are separated from the bloodstream by a layer of
fenestrated endothelium. This makes the liver sinusoids differ-
ent from capillaries in the general circulation. A single cell
RNAseq study revealed that hepatocytes along the axis of the
hepatic acini each have their own gene expression profile and
their own metabolic program. From these studies a picture
emerges with endogenous synthesis of chenodeoxycholic acid
(CDCA) mainly occurring in centrizonal hepatocytes, while the
12α‐hydroxylase CYP8B1, for synthesis of cholic acid (CA), is
mainly located in the midzonal cells and the bile salt conjugat-
ing enzyme coenzyme A:amino acid N‐acyltransferase (BAAT)
mainly in the periportal cells [22]. The functional consequence
of this is that unconjugated bile salts have to re‐enter the peri-
portal hepatocytes before being secreted as conjugates.
Currently this is merely a hypothesis based on the distribution
of enzymes. Final proof has to wait for detailed molecular meta-
bolic studies in which metabolites are identified in situ at the
subcellular level.
Bile canaliculi are the space between two hepatocytes where
they form an interconnected mesh within hepatocyte plates.
Each canaliculus is surrounded by the plasma membrane
domains of adjacent hepatocytes. Canaliculi can be visualized
by immune fluorescent microscopy using antibodies directed at
canalicular antigens (dipeptidyl peptidase IV, MDR1/ABCB1,
BSEP/ABCB11, MRP2/ABCC2) or by dynamic intravital
microscopy using fluorescent cholephilic tracers (e.g. bile salt
tracers such as cholyl‐lysylfluorescein [CLF] and organic anion
tracers such as 6‐carboxyfluorescein diacetate [6‐CFDA]) [23–
28]. An ideal fluorescent marker for BSEP/ABCB11‐mediated
transport is not available. The specificity of BSEP does allow
transport of bile acid analogues with a big fluorescein moiety.
However, these labeled molecules are also transported by
MRP2, the non‐bile salt organic anion transporting protein.
Raman imaging enables the study of transport and flow of natu-
ral bile salts without the need for fluorescent labeling.
Visualization of bile flow in the canaliculus is challenging.
Intravital microscopic studies suggest that the direction of bile
flow in the canaliculi is from the pericentral to the periportal
zone, from where bile is delivered to bile ductuli and ducts [23].
Other studies suggest that canalicular bile does not really flow
but represents a basin in which solutes move by diffusion. In
this model bile only flows in bile ductules and ducts (Vartak and
Hengstler, 2019, personal communication). Computational
modeling suggests that water is added to bile mainly in the peri-
portal area or in the bile ductules, where the biliary pressure is
low. Low biliary pressure allows the passive flow of water to
bile via water channels supported by the osmotic gradient cre-
ated by the high concentration of organic solutes in canalicular
bile [23].
To fully understand this model, the enterohepatic cycle has to
be taken into consideration. Bile salts are reabsorbed in the
intestine and delivered to the liver via the portal blood. Studies
using tracer dose radiolabeled taurocholate indicated that uptake
of bile salts mainly occurs in the periportal area of the hepatic
acinus. At higher concentrations, pericentral hepatocytes also
participate in bile acid uptake [29]. In a normal physiological
state, bile salts are delivered to the liver in the micromolar con-
centration range. At this concentration bile salt uptake is
expected to engage most hepatocytes in the hepatic acinus.
Transport via ABC transporters is dependent on sufficient sup-
ply of ATP and this is better guaranteed in the oxygen‐rich peri-
portal zone than in the oxygen‐poor pericentral zone of the
hepatic acinus. Therefore, the main ABC transporter activity is
in the periportal oxygen‐rich region of the hepatic acinus. If the
canalicular flow model is accepted it should be realized that the
pericentral‐to‐periportal pressure differences cannot entirely
explain the velocity of canalicular flow, particularly not in the
midzonal area. Experimental evidence indicates that peristaltic
contractions probably support canalicular bile flow [23] (see
Figure 26.1).
ABC TRANSPORTERS
Discovery, nomenclature, and structure
Victor Ling’s laboratory in 1976 identified a protein conferring
multidrug resistance by supporting the efflux of daunorubicin and
a number of other agents in Chinese ovary cells. This 170 kDa
protein was called P‐glycoprotein (Pgp) [18, 30, 31]. P‐glycopro-
tein belongs to the superfamily of ATP‐binding cassette (ABC)
transport proteins. The name ABC transporter indicates an evolu-
tionary conserved ATP‐binding domain and a transport function
[32, 33]. P‐glycoprotein is now known as ABCB1 or MDR1.
ABC transport proteins are expressed in prokaryotes as well
as eukaryotes which points to their evolutionary conserved
function as protectors of the interior milieu. For an overview see
Table 26.1.
The human ABC transporter superfamily comprises 48 genes
and proteins, classified in 7 subfamilies [34]. Members of the
ABC transport superfamily are expressed in liver, pancreas, kid-
ney, intestine, testis, and brain. These are tissues at the boundary of
the interior milieu and the external environment. ABC transporters
are ATP‐dependent pumps that support the active transport of a
wide variety of organic and inorganic anions, metals, peptides,
 26:  Hepatic Adenosine Triphosphate‐Binding Cassette Transport Proteins and Their Role in Physiology 315

Chol
CYP7A1 FXR
Bile salts Organic
H2O H2O H2O
CA anions Chol PC CDCA

aquaporins BSEP
MRP2 G5/G8 MDR3 BSEP

Bile ducts Bile

Periportal Pericentral
zone zone
Bile salts Organic Chol PC CDCA
H2O H2O H2O
CA anions

NTCP OATP

EHC: bile salts, C4 Organic anions

Periportal Midzonal Pericentral

Relative transporter 0.73 1.0 2.07
density
Intraluminal ~100 ~2500
pressure (Pa)
Bile velocity 1.86 0.39 0.11
(μm/sec)

Figure 26.1  Contribution of ABC transporters and water channels to bile secretion. A single bile canaliculus extends from the pericentral to
periportal area. According to the computational modeling study of Meyer et al. the relative density of transport proteins in the pericentral area is
about 2.8 times higher than in the periportal area. The calculated intraluminal pressure is about 25 times higher in the pericentral than in the peri-
portal area and the bile flow velocity is 17 times higher in the periportal than in the pericentral area [23]. Abbreviations: OATP, organic anion
transporting polypeptides OATP1B1 and OATP 1B3; NTCP, sodium taurocholate cotransporting polypeptide; EHC, enterohepatic cycle; Chol,
cholesterol; PC, phosphatidylcholine; CDCA, chenodeoxycholic acid, CA, cholic acid.

amino acids, and sugars. In liver, these ATP‐dependent pumps
transport solutes from liver to bile. In the intestine ABC transport-
ers mainly pump across the mucosa cell membrane toward the
intestinal lumen. By this orientation they form an active barrier
against the intestinal uptake of solutes, toxins, certain food ingre-
dients, and certain anti‐cancer medications (e.g. topotecan). Also,
in the blood–brain barrier ABC transport proteins pump agents out
of the brain thus presenting a barrier for uptake of drugs and toxic
agents, the so‐called blood–brain barrier. Most ABC transport pro-
teins support the transport of solutes across membranes but there
are exceptions. For instance, CFTR (cystic fibrosis transmem-
brane conductance regulator) structurally belongs to the ABC
transporter family but functionally acts as a chloride channel.
CFTR mutations are the cause of cystic fibrosis.
Structure
The molecular structure of the MDR (multidrug resistance) pro-
teins (ABCB1, ABCB4, ABCB11) comprises two transmem-
brane domains (TMD1 and TMD2) each with six membrane
spanning helices linked to an intracellular nucleotide‐binding
domain (NBD1 and NBD2) for ATP binding and hydrolysis
(Walker A and B and ABC signature sequence). Together TMD1
and TMD2 form a channel with sites for ligand binding and trans-
port. Upon ATP hydrolysis and ligand binding the channel
undergoes conformational changes wherein the intracellular
ligand‐accepting mode shifts to an extracellular efflux mode [35].
The molecular structure of the nine‐membered C family of
190 kDa MRP glycoproteins is similar to that of the B family
but includes an additional N‐terminal TMD (TMD0) with five
membrane‐spanning helices connected via a linker region to
TMD1. ABC C family members differ in the structure of the
TMD0 domain.
Proteins of the G family of “white” proteins only have one
TMD of six membrane spanning helices. These proteins act
as homo‐ (ABCG2) or hetero‐ (ABCG5/G8) dimers and as
functional transporters they resemble the B family proteins.
The mouse homologues of ABCB1 (MDR1) are Abcb1a and
Abcb2 (Mdr1a and Mdr1b); for ABCB4 (MDR3) this is
Abcb4 (Mdr2).
 Table 26.1  Human hepatic ABC transporters
Gene Protein Chromosome Typical substrates NHR Regulation Clinical phenotype Refs
ABCB1 MDR1 7q21.12 Drugs, chemotherapeutic agents PXR Neurotoxicity (mice and dogs) [77, 78]
ABCB4 MDR3 7q21.12 PC FXR, PPARα PFIC3, LPAC, ICP [21, 57]
ABCB11 BSEP 2q31.1 Bile salts FXR PFIC2, BRIC‐2, ICP [20, 25, 65, 165]
ABCC1 MRP1 16p13.11 LTC4, cyclic nucleotides, drugs PXR Increased chemosensitivity (mice) [109]
ABCC2 MRP2, cMOAT 10q24.2 Bilirubin glucuronides, GSH, glutathione conjugates, LTC4, anti‐HIV drugs FXR, PXR, CAR Dubin–Johnson syndrome [16, 90]
ABCC3 MRP3 7q21.33 Bilirubin glucuronides, bile salts, estradiol‐17 β glucuronide PXR, CAR Increased chemosensitivity (mice) [109, 169]
ABCC4 MRP4 13q32.1 Bile salts, bile salt sulfo‐conjugates, cyclic nucleotides (cAMP, cGMP), PXR, CAR Increased chemosensitivity (mice) [109, 111]
DHEA 3‐sulfate
ABCC6 MRP6 16p13.11 ATP HNF4α Pseudoxanthoma elasticum [116, 117]
ABCC7 CFTR 7q31.2 Chloride — Cystic fibrosis [118]
ABCG2 BCRP 4q22.1 Anticancer drugs, protoporphyrin IX — Diet‐induced phototoxicity (mice) [119]
ABCG5/ — 2p21 Cholesterol — Sitosterolemia, gallstones [41]
G8

c26.indd 316 03-11-2019 19:50:05

 26:  Hepatic Adenosine Triphosphate‐Binding Cassette Transport Proteins and Their Role in Physiology 317
Nomenclature
In this chapter we will follow the international literature in using
capitals for human ABC proteins and capital italics for the ABC
protein encoding genes. For ABC proteins and genes in rodents
we will use lower case.
From multidrug resistance to liver function
B‐, C‐, and G‐family ABC proteins have originally been discov-
ered as mediators of multidrug resistance of cancer cells [18,
36]. This is based on findings in the late eighties, indicating that
the expression of these proteins in cancer cells increases upon
exposure to chemotherapeutic agents, thus conferring drug
resistance. These findings were originally met with great enthu-
siasm as this was thought to provide an instrument to deal with
multidrug resistance in cancer. Meanwhile, however, it became
apparent that multidrug resistance in cancer is more complex
and does not only depend on the overexpression of a single pro-
tein [37]. Nevertheless, for liver research this burgeoning field
has been of considerable importance for the identification of
transport proteins supporting hepatic secretory function.
MDR1, BSEP, MDR3, MRP2, and ABCG5/G8 are located
in  the apical or canalicular membrane of the hepatocyte.
Cell  biological studies have elucidated the complex pathway
accounting for post-translational regulation and cellular locali-
zation (see Chapter 4). ABCC3 (MRP3) and ABCC4 (MRP4)
are located in the basolateral membrane of the hepatocyte.
These proteins are expressed during cholestasis and function to
protect the hepatocyte against high concentrations of toxic
metabolites by supporting active transport out of the cell into the
blood. Physiologically this is “reverse” transport. Agents leav-
ing the liver via this route are secreted by the kidney.
ABC‐MEDIATED TRANSPORT
IN THE LIVER
The apical or canalicular hepatocyte membrane is home to
ABCB1 (P‐glycoprotein, multidrug resistance protein 1
[MDR1]), ABCB4 (MDR3), ABCB11 (BSEP, bile salt export
protein), ABCG2 (BCRP, breast cancer resistance protein),
ABC G5/G8 (the cholesterol transport protein), and ABCC2
(MRP2, multidrug resistance‐associated protein 2, formerly
called cMOAT, canalicular multispecific organic anion transporter).
For normal liver function, MDR3, BSEP, MRP2, and ABCG5/
G8 are most relevant as they support the transport of the major
ingredients of bile: phosphatidylcholine (PC) (MDR3), bile
salts (BSEP), conjugated bilirubin, glutathione and glutathione‐
conjugates (MRP2), and cholesterol (ABCG5/G8). BSEP sup-
ports the transport of monovalent bile salts and, by action of the
solubilizing effect of bile salts at the canalicular membrane,
indirectly facilitates the secretion of cholesterol and PC [38].
MDR3 mediates the translocation of PC from the inner to the
outer leaflet of the canalicular membrane [39]. ABCG5/G8 sup-
ports the transport of cholesterol and MRP2 the transport of ani-
ons with at least two negative charges including bilirubin
conjugates [40–42]. The function of BSEP, MDR3, and ABCG5/
G8 is interdependent. Perturbation of any of these transporters
can give rise to unstable bile, prone to gallstone formation.
Although ABCG5/G8 is considered as a “gallstone gene”
(Lith9), malfunction of MDR3 is equally important for gall-
stone formation as in “low phospholipid‐associated cholelithia-
sis (LPAC)” [43–45]. LPAC is characterized by intrahepatic
microlithiasis, liver fibrosis, and gallstones.
The function of ABC transporters in the canalicular mem-
brane depends on the cholesterol content of microdomains or
“lipid rafts” in which these transporters are located. The stabil-
ity of these rafts is perturbed when phospholipid transport is
impaired as in MDR3 and ATP8B1 deficiency diseases.
Cholesterol lipid raft microdomains are essential for the proper
function of BSEP, MRP2, and ABCG2 [46–49].
ATP8B1
Although ATP8B1 (FIC1) does not belong to the ABC transport
protein family, it is discussed here since it is important for
understanding the hepatic secretory function. ATP8B1 is a P‐
type ATPase, encoded by the FIC1 gene. ATP8B1 supports the
reverse translocation of phosphatidylserine and phosphatidyle-
thanolamine from the outer to the inner leaflet of the canalicular
membrane domain [50]. FIC1 gene mutations have dramatic
effects on BSEP and MRP2 function (see Chapter 29). Children
with ATP8B1 deficiency have severe cholestatic liver disease
called progressive familial intrahepatic cholestasis (sporadic
PFIC type 1 in non‐Amish populations, endemic Byler’s disease
in the Amish population). Characteristically these patients have
a low serum gamma‐glutamyltransferase activity. Adults, hete-
rozygous for the FIC1 mutation, have a less severe disease
called benign recurrent intrahepatic cholestasis (BRIC type 1)
[51, 52].
ATP8B1−/− mice show an increased bile salt‐induced extrac-
tion of cholesterol into bile with, as a result, a reduced choles-
terol/PC ratio in the canalicular membrane and impaired BSEP
and MRP2 function [46]. Children with PFIC1 temporarily ben-
efit from partial external biliary drainage but most eventually
need to be transplanted. However, liver transplantation does not
completely resolve the clinical manifestations of this disease as
patients suffer from steatosis of the liver graft, diarrhea, pancre-
atic disease, and continue to have poor growth [53]. The effect
of external biliary drainage suggests that secondary bile salts
(deoxy‐ and lithocholic acid), play a role in the functional
defect. External biliary drainage causes enrichment of the bile
salt pool with primary bile salts and this may stabilize the cana-
licular membrane and improve the function of transport
proteins.
ABCB4 (MDR3, Mdr2)
The ABCB4/MDR3 gene encodes a 170 kDa MDR3 protein in
humans and Mdr2 in mice. The function of this apical hepato-
cyte protein became clear when Mdr2−/− mice became available
[21]. These Mdr2−/− mice have normal bile salt secretion but
lack PC in bile. This has major consequences for liver and bile
ducts. In normal bile, cytotoxic bile salts are packed in phospho-
lipid and cholesterol‐containing micelles or vesicles. This pro-
tects hepatocytes and cholangiocytes from the cytotoxic effect
318 THE LIVER:  ABCB1 (MDR1)
of bile salts. Without phospholipids bile is toxic and this may
underlie periductal inflammation and fibrosis as in Mdr2−/− mice
and children with MDR3 deficiency (PFIC type 3). The histopa-
thology in Mdr2−/− mice bears some resemblance to the
pathology in primary sclerosing cholangitis (PSC). PSC is
characterized by concentric periductular fibrosis, biliary cirrho-
sis, and an increased risk for cholangiocellular cancer. However,
apart from single nucleotide polymorphisms functional ABCB4/
MDR3 gene defects have not been detected in PSC patients
[54, 55].
Mutations of the ABCB4/MDR 3 gene underlie PFIC type 3
in children, low‐phospholipid‐associated cholelithiasis (LPAC)
and selected cases of intrahepatic cholestasis of pregnancy
(ICP) in adults [43, 56–60]. PFIC type 3 is a pediatric disease
characterized by cirrhosis, bile duct abnormalities, severe pruri-
tus, and elevated serum gamma‐glutamyltransferase. PFIC type
3 in humans is associated with a more severe histopathology
than that seen in Mdr2−/− mice. This is probably due to human
bile being more cytotoxic than mouse bile with more hydro-
philic bile salts (e.g. α‐ and β‐muricholic acid). Ursodeoxycholic
acid administration provides temporary relief in a subset of patients
but most patients eventually need a liver transplantation.
The MDR3 protein plays an important role in bile physiol-
ogy. It acts together with BSEP and ABCG5/G8 in producing
PC‐, bile salt‐, and cholesterol‐containing vesicles at the outer
leaflet of the canalicular membrane. Mdr2−/− mice produce bile
with bile salts in normal concentration but without PC and
reduced cholesterol. This shows that a normal MDR3 function
not only is required for PC secretion but cholesterol secretion
indirectly also depends on PC secretion [38]. Bile salts and PC
are needed for cholesterol solubilization. Hence a decreased PC
concentration in bile represents a risk for gallstone and biliary
sludge formation.
Mdr2 expression in mice is increased by treatment with
PPARα agonists [61]. Thus PPARα agonists are expected to
increase PC secretion in bile and therefore this may represent a
“bile detoxifying” therapy that may be of therapeutic value.
Norursodeoxycholic acid may also be a useful drug as it
increases bicarbonate secretion thus protecting bile duct epithe-
lium against bile salt toxicity [62].
Bile salt export pump, ABCB11 (BSEP)
A protein originally called “sister of P‐glycoprotein (SPGP)”
was identified as the principal transporter for glycine‐ and tau-
rine‐conjugated bile salts. SPGP or BSEP is located in the cana-
licular membrane domain of hepatocytes [17]. Sister of
P‐glycoprotein is generically called bile salt export pump
(BSEP) and formally ABCB11. Children with BSEP deficiency,
caused by ABCB11 mutations, have a severe cholestatic liver
disease called PFIC type 2. In mice the targeted deletion of the
Abcb11 gene causes cholestasis albeit not as severe as in humans
[63]. This is due to the extensive tetrahydroxylation of bile acids
and their subsequent renal and biliary excretion via upregulated
Mdr1a and Mrp2. Tetrahydroxylation of bile salts is a pathway
available in mice but not in humans [64]. Definite proof for
BSEP being a bile acid transporter came from studies by Gerloff
et al. who showed bile salt transport in Xenopus laevis oocytes
and Sf9 cells upon expression of the mouse Abcb11 gene [17].
A mutation affecting BSEP underlies PFIC type 2 in chil-
dren and benign recurrent intrahepatic cholestasis (BRIC 2) in
adults heterozygous for ABCB11/BSEP gene mutations [25,
65, 66] (see Chapter  29). In PFIC2 the accumulation of bile
salts causes inflammation and fibrosis of the liver and is asso-
ciated with an increased risk for hepatocellular carcinoma
[67]. Typically children with PFIC2 develop cholestasis with a
low serum gamma‐glutamyltransferase activity. Partial exter-
nal biliary diversion provides temporary relief particularly in
patients with some residual BSEP expression. Most patients
eventually need a liver transplantation. After liver transplanta-
tion patients usually do well and show a catch‐up growth [53].
Recurring PFIC2 has been described after liver transplantation
due to development of BSEP‐directed antibodies [68, 69].
Patients with residual BSEP activity may respond to ursode-
oxycholic acid therapy. A future option may be treatment with
chaperone molecules that help reparation of misfolded BSEP
protein and insertion into the apical membrane [70]. This
would help a subgroup of patients.
BSEP supports the transport of taurine‐ and glycine‐con-
jugated bile salts, unconjugated bile salts, and ursodeoxy-
cholic acid. Nor‐ursodeoxycholic acid and obeticholic acid
may also be BSEP substrates but this has not been tested yet.
BSEP has a preference for transporting bile salts but shows
low activity with pravastatin as substrate [71]. Cyclosporin
A, troglitazone, bosentan, glibenclamide, and rifampicin are
BSEP inhibitors and may cause drug‐induced liver disease
[72–74].
ABCB1 (MDR1)
MDR1/ABCB1 confers multidrug resistance in cancer. In the
liver ABCB1 is located in the hepatocyte canalicular mem-
brane domain where it supports the biliary secretion of
organic cations, hormones, and drugs [75, 76]. ABCB1 is
also expressed in epithelial cells of the bile duct, small intes-
tine, kidney, blood–brain barrier, and placenta. ABCB1 and
ABCB11 show structural similarity but considerable func-
tional differences. While ABCB1 mediates the transport
of a great variety of substrates (analgesics, anti‐arrhythmic
agents, antibiotics, anti‐cancer agents, antihistamines, anti‐
epileptic drugs, immunosuppressive agents, antilipidemic
agents), ABCB11 supports the transport of bile salts with a
preference for conjugated bile salts [35]. The role of ABCB1
in normal liver physiology is unclear. Natural mutations of
the Abcb1 gene in dogs or targeted deletion of Abcb1a in
mice cause sensitivity of the brain to drugs (e.g. ivermectin),
in otherwise healthy dogs and mice, suggesting a role of this
transporter in the blood–brain barrier [77, 78].
MDR1 may serve as a backup transport system. For instance
in Abcb11/Bsep−/− mice, Mdr1a expression is increased and sup-
ports transport of tetrahydroxylated bile salts [20]. Because of
this detoxifying pathway, Abcb11/Bsep−/− mice do not have a
real cholestatic phenotype. The varied phenotype in patients
with progressive familial intrahepatic cholestasis (PFIC type 2)
may be related to the expression level of MDR1 [79]. In hepato-
cellular carcinoma (HCC) MDR1 expression is variable and
 26:  Hepatic Adenosine Triphosphate‐Binding Cassette Transport Proteins and Their Role in Physiology 319
inversely correlated with disease free survival and response to
chemotherapy [80].
ABCC2 (MRP2)
The multidrug resistance‐associated proteins (MRPs) are 190
kDa glycoproteins expressed in various organs including liver,
MRP2 is encoded by the ABCC2 gene. Compared to proteins
of the ABCB family (the MDRs) the ABCC family proteins
(the MRPs) have an extra transmembrane domain comprised
of five helices connected by a linker region to the rest of the
protein. The function of this extra domain is unclear since
without this domain MRP1 is fully active and retains its mem-
brane localization [81].
MRP2 is the apical transporter of amphiphilic “multivalent”
organic anions, mainly agents with more than one negative
charge. Thus conjugated (mono‐ and diglucuronidated) biliru-
bin, oxidized and reduced glutathione (GSSG and GSH),
numerous drug conjugates, tetra‐hydroxylated, sulphated and
divalent glucuronidated bile salts, leukotriene C4, conjugated
hormones, endothelin I, and vasopressin are MRP2/mrp2 sub-
strates [15, 82–86]. Positively charged agents (e.g. metals) are
transported by MRP2 complexed to glutathione [87].
In bile, bile salts and glutathione are the solutes with the
highest concentration (millimolar range). Thus, glutathione acts
as an osmolyte contributing to the “bile‐salt independent bile
flow”. In TR− rats, with a genetic deletion of MRP2, bile flow is
reduced to ~50% [13, 15].
The MRP2 protein is expressed in the liver, intestine, renal
proximal tubule, gallbladder, and placenta [88, 89]. The absence
of Mrp2 caused by natural mutations in TR− rats, monkeys, and
sheep is associated with conjugated hyperbilirubinemia and a
phenotype similar to the human Dubin–Johnson syndrome
(DJS). Like humans with DJS, TR− rats have a brown‐black
lysosomal pigment in the liver upon feeding a tyrosine‐, tryp-
tophane‐, and phenylalanine‐enriched diet [13, 16, 90, 91]. The
targeted deletion of Abcc2 in mice causes mild conjugated
hyperbilirubinemia with defective biliary secretion of glu-
tathione and organic anions [92]. MRP2 is expressed in most
HCCs [80, 93, 94].
Jaundice
Jaundice is the clinical hallmark of cholestatic liver disease. On
a molecular level jaundice may be caused by MRP2 dysfunc-
tion, be it genetic or acquired. Jaundice of sepsis is a typical
example of acquired jaundice by reduced MRP2 expression.
Inflammatory cytokines suppress MRP2 expression to a level
that impairs the hepatobiliary transport with conjugated hyper-
bilirubinemia as a result [95, 96]. MRP2 downregulation may
be the underlying mechanisms of jaundice in a variety of liver
diseases [97, 98].
ABCC1, 3, 4 (MRP 1, 3, 4)
MRP1 is encoded by the ABCC1 gene. MRP1 is present in the
basolateral membrane of most epithelial cells where it supports
the transmembrane transport of leukotriene C4 (LTC4), reduced
and oxidized glutathione, and numerous drugs, glutathione‐ and
glucuronide‐conjugates. MRP1 expression in normal liver is
marginal but is increased in severe hepatic failure, in hepatic
progenitor cells, and in reactive hepatic ductules [99, 100].
Deletion of MRP1 has not been linked to a particular liver dis-
ease. MRP1 plays a role in the natural course of cancer and is
responsible for multidrug resistance in some patients. Among
the MRPs, MRP1/ABCC1 has the highest homology with
CFTR/ABCC7 and may act as a modifier gene in cystic fibrosis
[101]. While MRP2 is exclusively present in the canalicular
domain of hepatocytes and the apical domain of intestinal
mucosa and tubular epithelial cells, MRP1 is located in the
basolateral membrane. This difference in localization is due to a
PDZ motif in the C‐terminus of the MRP2 protein that is lack-
ing in MRP1[81, 102]. This motif interacts with other PDZ‐con-
taining proteins, directing MRP2 to the apical membrane.
MRP1 expression in HCC is associated with a more aggressive
phenotype [103].
MRP3 is an inducible ATP‐dependent transporter of organic
anions in the basolateral membrane of hepatocytes, bile duct
and intestine epithelial cells, adrenals, pancreas, kidney, and
prostate [100, 104–106]. MRP3 is upregulated when canalicular
MRP2 is not or poorly expressed as in Dubin–Johnson syn-
drome or when transport across the canalicular membrane is
impaired as in cholestasis [104, 105, 107]. The substrate spec-
trum of MRP3 is similar but not identical of that of MRP2 [108].
MRP3 transports bilirubin mono‐ and diglucuronide,
estradiol‐17β‐glucuronide, glutathione‐conjugates, bile salts
and their conjugates. In contrast to MRP2, MRP3 does mediate
the transport of reduced or oxidized glutathione and has a low
affinity for LTC4 [109]. MRP3 and/or MRP4 are highly
expressed in hepatic progenitor cells [99, 100].
Like MRP3, MRP4 is an inducible ATP‐dependent trans-
porter in the basolateral membrane of hepatocytes. MRP4 is
expressed in liver and proximal tubules of the kidney. MRP4
supports transport of conjugated, sulfated, and unconjugated
bile salts, cyclic nucleotides (cAMP), LTC4, dehydroepiandros-
terone (DHEA)‐3‐sulphate, estradiol‐17β‐glucuronide, and
cyclic nucleotides (cAMP) [110, 111]. MRP4 protein expres-
sion in rodent and human liver is induced by cholestasis [112–
114]. Basolateral MRP3 and MRP4 take over the function of
MRP2 and BSEP when the function of these canalicular trans-
porters is impaired as in cholestasis. They support “reverse”
transport from hepatocyte to blood leading to elevated serum
levels of conjugated bilirubin and bile salts.
ABCC6, MRP6
There is strong genetic linkage between the more than 300
­identified mutations of the ABCC6 gene and pseudoxanthoma
elasticum, a rare disease with abnormalities in skin, retina, car-
diovascular system, and gastrointestinal system [115]. ABCC6/
MRP6 is a transmembrane transporter of the ABC C family
expressed in liver and kidney. Pseudoxanthoma elasticum is
characterized by increased mineralization of connective tissue
and arteries. The relation between disease manifestations and
the gene defect has long been elusive. In particular the relation
between the gene product in the liver, and increased mineraliza-
tion in almost all tissues, was puzzling. Recent literature indi-
cates that ABCC6/MRP6 supports transport of ATP [116, 117].
320 THE LIVER:  REGULATION OF ABC TRANSPORTER EXPRESSION
ATP is degraded to inorganic pyrophosphate PPi, a potent inhib-
itor of mineralization. Reduced PPi in the circulation leads to
the precipitation of calcium‐phosphate in arteries, skin, and
retina. Oxidative stress in NASH may reduce HNF4α‐mediated
expression of MRP6 and thereby reduce ATP release and stimu-
late mineralization of the arteries leading to coronary athero-
sclerosis, a prominent cause of death in NASH [115].
ABCC7
ABCC7 deserves special mention. ABCC7 is the cystic fibrosis
transmembrane conductance regulator (CFTR) that, when
mutated, underlies cystic fibrosis [118]. This protein is excep-
tional as it is not an ATP‐dependent pump but a channel that
allows the cellular efflux of chloride and bicarbonate. CFTR is
expressed in lungs, pancreas, bile ducts, intestine, and vas defer-
ens. The molecular structure shows two transmembrane domains
each with six membrane‐spanning helices, two nucleotide bind-
ing domains, and a regulatory “R” domain activated by cAMP.
An inside negative electrochemical gradient drives chloride
efflux out of the cell. CFTR constitutes a gated channel that can
open and close upon ATP binding.
ABCG2
ABCG2 mediates transport of a number of chemotherapeutic
agents (hence the name breast cancer resistance protein 1),
estradiol, progesterone, and chlorophyll metabolites. ABCG2−/−
mice on an alfalfa‐containing diet accumulate phototoxic chlo-
rophyll metabolites [119].
Bcrp1−/− mice show increased topotecan serum levels after
oral administration of this anticancer drug showing that in nor-
mal mice Bcrp1 in the intestine acts as a barrier for orally
administered drugs (transporting drugs out of the organism
toward the intestine). In the liver, ABCG2, MRP3, and MRP4
are expressed in progenitor cells and may serve a cytoprotective
function [99, 100, 120].
ABCG5/G8
ABCG5/G8 expression in liver and intestine is important for
cholesterol transport. In the liver ABCG5/G8 supports the
biliary secretion of cholesterol, and in the small intestine
the protein mediates the efflux of cholesterol and plant ­sterols
thereby inhibiting their absorption. ABCG5 and G8 are
half  transporters, both proteins are needed for transport
­function. Patients with mutations in the ABCG8 gene develop
­hypercholesterolemia, premature coronary atherosclerosis, and
phytosterolemia but in meta‐analysis the association with
hypercholesterolemia was found not to be strong [41, 121].
Targeted deletion of the Abcg5 and Abcg8 genes in mice causes
impaired cholesterol secretion in bile and increased sensitivity
toward changes in the cholesterol content of their diet. On the
other hand, transgenic expression of ABCG5 and G8 causes
high cholesterol secretion in bile and lowered absorption of
­cholesterol in the intestine [122].
The function of the blood–brain barrier largely depends on
the presence of ABC transporters. ABC transporters actively
pump out neurotoxic compounds (e.g. ivermectin) thus
preventing accumulation in the brain. This indicates that muta-
tions of ABC transporters can affect the clearance of endoge-
nous and exogenous agents and, depending on the substrate,
allows the intestinal absorption or penetration of the blood–
brain barrier by agents that in normal wild‐type animals would
not be absorbed or cross the blood–brain barrier.
ABCG5/G8 are also known as gallstone susceptibility
genes (Lith9 gene in mice). In human populations the ABCG5‐
R50C and ABCG8‐D19H gene variants are associated with
gallstones [123]. The ABCG8‐D19H variant, when tested in
a  3H‐cholesterol export assay, showed increased transport
activity. The association between the D19H variant and an
increased gallstone risk (odds ratio 2.2; P = 1.4 × 10−14) has
been confirmed in German, Chilean, Chinese, and Indian
­populations [124].
REGULATION OF ABC TRANSPORTER
EXPRESSION
Transcriptional regulation
ABC transporter expression at the apical membrane can change
rapidly or more gradually. Transcriptional and post‐transcrip-
tional mechanisms underlie these regulations. A number of
elegant studies has shown that ABC transporter expression is
regulated by hepatocyte nuclear factor 4α (HNF4α), nuclear
hormone receptors (NHRs such as FXR [farnesoid X‐receptor]
and PXR [pregnane X‐receptor], PPARα and CAR [constitu-
tive androstane X‐receptor] [125–127 ]) (see Figure 26.2). The
NHRs act as intracellular sensors for bile salts, endogenous
metabolites, and xenobiotics. They reside in the cytosol and
when bound to their respective ligands they bind to the DNA of
target genes. In general they function to maintain intracellular
homeostasis by increasing outward transport and reducing
uptake. Bile salts provide a clear example. When intracellular
bile salt concentrations increase due to cholestasis, the expres-
sion of the FXR‐dependent bile salt transporter OSTαβ as well
as MRP3 and MRP4 at the basolateral domain increases [104,
107, 112, 114, 128]. MRP3 supports the cellular efflux of con-
jugated bilirubin from hepatocyte to blood thereby contributing
to elevated levels of conjugated bilirubin in serum and clinical
jaundice [129]. MRP4 supports the cellular efflux of conju-
gated and unconjugated bile salts and bile salt sulfate conju-
gates. MRP4−/− mice are more vulnerable and sensitive to
cholestasis [130].
FXR acts as a bile acid sensor that is activated upon bile
salt binding [131, 132]. This causes upregulation of SHP
(short heterodimer partner), a protein that blocks the LRH‐1
and HNF4α‐mediated regulation of CYP7A1, an enzyme
responsible for bile salt synthesis [133]. In addition, FXR
activation causes upregulation of BSEP, MRP3 and 4, and
downregulation of NTCP [134]. Also the expression of OSTαβ,
a non‐ATP‐dependent basolateral transporter, is increased in
cholestasis in humans [135].
The importance of FXR for biliary transport becomes appar-
ent when studying the phenotype of patients with genetic FXR
defects. Homozygous mutations of the NR1H4 gene, encoding
 26:  Hepatic Adenosine Triphosphate‐Binding Cassette Transport Proteins and Their Role in Physiology 321

Function
Chemotherapeutics Bilirubin glucuronides
Cyclosporin A Bile salts GSH/GSSG
(Bile salts) PC UDCA Organic anions Cholesterol Chemotherapeutics

MDR1 MDR3 BSEP MRP2 G5/G8 BCRP

Regulation

AhR
PXR MDR1 PPARα MDR3 FXR BSEP PXR MRP2 LXR G5/G8 BCRP

CAR ER

Fenofibrate CDCA Rifampicin Oxysterols Estradiol

Rifampicin
Benzofibrate Obeticholic acid Dexamethasone T0901317 Progesteron
Dexamethasone
PX-104 GW3965
miR 331-5 p miR 33 miR 328
Phenobarbital
miR 451 TCDD miR 181a
miR 298 miR 379 Benzopyrene miR 487a
miR 27a miR 133a miR 212
miR 122 miR 297 miR 200c
miR 137 miR 520h
miR 200c miR 519c
miR 1253
mR 145

Figure 26.2  Function and regulation of canalicular ABC transporters. Upper panel: Function of canalicular ABC transporters depicted with lead-
ing substrates. Lower panel: Regulation of ABC transporters by nuclear hormone receptors and inhibitory activity by miRNA [35, 144, 170–172].

FXR, causes severe neonatal cholestasis with undetectable
BSEP and preserved MDR3 expression. The disease in these
children is characterized by low serum gamma‐glutamyltrans-
ferase activity, severe jaundice, vitamin K dependent coagulop-
athy, and elevated α‐fetoprotein [136]. This shows that FXR not
only regulates BSEP expression at increased bile salt levels but
also plays a role in the household regulation of BSEP.
Regulatory adaptations in patients with progressive famil-
ial intrahepatic cholestasis (PFIC2) provide examples of
cytoprotection when export of toxic solutes is perturbed. In
PFIC2 BSEP function is reduced or absent, therefore normal
bile salt secretion across the canalicular membrane is not
possible and the intracellular bile salt concentration is ele-
vated. Bile salts bind FXR and this activates the transcrip-
tion of the ABCC4 gene. MRP4 and OSTαβ at the basolateral
membrane are upregulated where they support the efflux of
bile salts from liver to blood thereby lowering the intracel-
lular bile salt level.
FXR is regulated mainly by conjugated primary bile salts
while PXR is regulated by non‐bile salt ligands (e.g. rifampicin,
corticosteroids) and hydrophobic secondary bile salts. CAR
binds bilirubin and phenobarbital; VDR hydrophobic bile salts
and 1,25‐dihydroxyvitamin D3 [125, 137–140]. This knowl-
edge translates into clinical medicine. Phenobarbital lowers
serum bilirubin levels by increasing the activity of the conju-
gating enzyme UGT1A1 and the transporter MRP2; rifampicin
is used to enhance MRP2‐mediated transport in “persistent
hepatocellular secretory failure”, a clinical condition wherein
intrahepatic cholestasis persists despite the removal of bile
duct obstruction or the removal of cholestasis‐inducing medi-
cation [141].
MicroRNA
MicroRNAs are small non‐coding RNA molecules that
originate as small transcripts of hundreds of nucleotides.
These single‐stranded RNAs interfere with gene transcrip-
tion. MicroRNAs are studied as therapeutic targets to coun-
teract ABC‐transporter‐mediated multidrug resistance in
cancer chemotherapy [93]. However, single microRNAs
target multiple genes. Applications of microRNA (e.g. gene
silencing) may therefore be complicated by bystander
effects [142].
Haenisch et  al. reported that microRNA 379 inhibits
rifampicin‐induced MRP2 expression by post‐transcriptional
mechanisms [143]. In addition the expression of ABCB1,
BCRP/ABCG2, and MRP1/ABCC1 is affected by microRNAs
[144]. Although this is early work, eventually microRNAs may
be therapeutic targets that can be used in the treatment of chole-
static liver disease. For instance, studies in miR 155−/− mice
show that miR 155 protects against acetaminophen‐induced
liver injury in mice [145]. The use of microRNAs as biomarkers
in drug‐induced cholestasis and intrahepatic cholestasis of
pregnancy is debated [146–148].
322 THE LIVER:  CONCLUSION
Post‐transcriptional regulation
Transcriptional regulatory changes typically take minutes or
longer, post‐transcriptional regulations occur faster. In hepato-
cytes the BSEP protein is located in a subapical vesicular com-
partment from where it shuttles to and from the apical membrane
[149]. This process is stimulated by ursodeoxycholic acid, hor-
mones, oxidative stress, and cell swelling [150–153]. Proteins
that tether ABC transporters to the cytoskeleton play a role in
providing an anchor for these proteins that determine their loca-
tion in the canalicular membrane [154]. Studies in rat liver and
Madin–Darby canine kidney cells by Ortiz et al. indicate that
HAX‐1 and the actin‐protein cortactin are involved in the retrac-
tion of BSEP, MDR1, and MDR2 from the plasma membrane to
the cell interior, presumably the subapical compartment. HAX‐1
binds to the linker domain of BSEP [155]. HAX‐1 and cortactin
are known to participate in clathrin‐mediated endocytosis sug-
gesting that this mechanism plays a role in the retraction of
BSEP and the MDR proteins from the apical membrane.
During bile‐duct ligation‐ or estradiol 17β glucuronide (E2
17βG) ‐induced cholestasis in rats, the ABC transporters BSEP
and MRP2 have been shown to be redirected from their apical
orientation towards intrahepatic compartments by clathrin‐
mediated endocytosis [156, 157]. Underlying mechanisms for
BSEP and MRP2 relocalization in cholestasis differ. For its
localization in the apical domain Mrp2 needs to be associated
with a protein call radixin. Radixin belongs to a group called
ERM proteins (ezrin, radixin, moesin). These proteins cross‐
link integral membrane proteins to actin filaments in the
cytoskeleton. Radixin−/− mice have conjugated hyperbilirubine-
mia because of a disturbed orientation of Mrp2 [158].
Interestingly the localization of BSEP and MDR1 is not dis-
turbed in these radixin−/− mice. Studies in rats and humans, have
shown that in cholestasis MRP2 and radixin do not colocalize
anymore. As a result of this disconnection MRP2 loses its local-
ization in the canalicular membrane [159, 160]. In cholestasis,
MRP2 binds to ezrin and this results in the redirection of MRP2
to intracellular vesicles where the MRP2 protein is ubiquit-
inated and subjected to proteosomal breakdown. The phospho-
rylation of ezrin at a critical domain (T567) by protein kinase C
α, δ, and ε is key to this mechanism [161].
The MDR1, MDR3, and BSEP proteins differ from MRP2 in
that MRP2 contains a PDZ domain at the C‐terminus that deter-
mines its localization in the apical membrane by interacting
with other PDZ domain containing proteins. The MDR/ABCB
proteins do not have this PDZ domain. MDR1 and BSEP recy-
cle between similar subapical and apical compartments and in
case of BSEP deficiency MDR1 is often also mislocalized con-
tributing to the severity of the phenotype.
ABC TRANSPORTERS AND DRUG‐
INDUCED LIVER DISEASE
The list of drugs causing cholestasis and cholestatic hepatitis is
long. Frequently implicated cholestasis‐inducing drugs are
amoxicillin–clavulanic acid, flucloxacillin, rifampicin, oral
contraceptives, and diclofenac. Oral contraceptives and
anabolic steroid cause “bland cholestasis” and clavulanic acid,
fluoroquinolones, ketoconazole, itraconazole, diclofenac, and
tricyclic antidepressants are known to cause cholestatic hepati-
tis. MRP2 inhibitors giving rise to “bland jaundice” (conjugated
hyperbilirubinemia with or without mild AST/ALT elevation)
include the anticancer drug sorafenib and the antivirals saquina-
vir, ritonavir, tenofovir, emtricitabine, and lamivudine. It is not
always clear if drug‐induced cholestasis is due to interactions at
the level of canalicular ABC transporters, drug‐metabolizing
enzymes, or due to immune‐mediated effects. This limits the
predictive power of in vitro drug toxicity models and makes
drug‐induced cholestasis unpredictable. This poses a health haz-
ard and a considerable (financial) risk for drug development.
Interaction of drugs with ABC transporters in vivo and in ani-
mal experiments shows a complicated pattern. For instance,
studies in BSEP‐expressing Sf9 cells showed that ATP‐depend-
ent bile salt transport is inhibited by cyclosporin A, rifamycin,
rifampicin, and glibenclamide. This is called cis‐inhibition
(drugs acting on the cellular inside). E2 17βG only inhibits
BSEP‐mediated bile salt transport when MRP2 is also expressed.
E2 17βG first has to be transported by MRP2 before it inhibits
BSEP. This is called transinhibition (inhibition by interacting
with the transport protein on the canalicular side) [162].
Cyclosporin A, troglitazone, and bosentan are BSEP inhibitors
and cause cholestatic injury in humans [163]. The liver injury
caused by these drugs most probably results from the cytotoxic-
ity of accumulated bile salts. Comparing the severe clinical phe-
notype of patients with PFIC2 and PFIC3 and the mild phenotype
of patients with Dubin–Johnson syndrome allows the prediction
that drugs interfering with BSEP or MDR3 are more likely to
induce liver injury than drugs interfering with MRP2 function.
Since systematic studies are lacking it is unclear how often
drug‐induced cholestasis is due to genetic variants of ABCB11
and ABCB4. Both genes have been confirmed to be associated
with drug‐induced cholestasis and with intrahepatic cholestasis
of pregnancy [164–167]. Vanishing bile duct syndrome is the
most severe and potentially irreversible form of drug‐induced
liver injury. Implicated drugs are amoxycillin–clavulanic acid,
flucloxacillin, diclofenac, and ibuprofen [168]. How (pro-
longed) inhibition of ABC‐protein‐mediated transport leads to
disappearance of bile ducts is unclear.
CONCLUSION
ATP‐dependent ABC transporters support the secretory func-
tion of the liver. Genetic or acquired disturbances of ABC trans-
porter activity cause accumulation of conjugated bilirubin, bile
salts, and other solutes, both in blood, liver, and other organs.
Since the liver has transporters for uptake of bile salts, organic
anions, and cations hepatocytes suffer most from hepatotoxic
drugs, either directly or indirectly via increased cytotoxic bile
salts. Bile salts at concentrations of 200 μm and above are toxic
as they cause apoptosis and necrosis. Mitochondrial toxicity
occurs at lower concentrations and result in loss of hepatocel-
lular polarity (see Chapters 4 and 8).
Nuclear hormone receptors such as FXR, PXR, CAR, and
PPARα and δ act as transcriptional regulators of ABC
 26:  Hepatic Adenosine Triphosphate‐Binding Cassette Transport Proteins and Their Role in Physiology 323
transporters and bile salt synthesis. High affinity ligands for
these receptors are candidate drugs for the treatment of choles-
tatic liver disease (see Chapters 30 and 31). Ursodeoxycholic
acid and norursodeoxycholic acid act at the post‐transcriptional
level by facilitating the insertion of transport proteins into the
canalicular membrane. In particular ursodeoxycholic acid has
proved its value in the treatment of cholestatic liver disease.
ACKNOWLEDGEMENT
I thank Drs Jan Hengstler and Nachiket Vartak (Department of
Toxicology, Leibniz Research Centre for Working Environment
and Human Factors, Dortmund, Germany) for critically reading
the manuscript.
REFERENCES
1. Sperber, I. Secretion of organic anions in the formation of urine and bile.
Pharmacol Rev, 1959;11:109–34.
2. Javitt, N.B. and Emerman, S. Effect of sodium taurolithocholate on bile flow
and bile acid exeretion. J Clin Invest, 1968;47:1002–14.
3. Boyer, J.L. and Klatskin, G. Canalicular bile flow and bile secretory pres-
sure. Evidence for a non‐bile salt dependent fraction in the isolated perfused
rat liver. Gastroenterology, 1970;59:853–9.
4. Russell, T.R., Searle, G.L., and Jones, R.S. The choleretic mechanisms of
sodium taurocholate, secretin, and glucagon. Surgery, 1975;77:498–504.
5. Boyer, J.L. and Bloomer, J.R. Canalicular bile secretion in man. Studies uti-
lizing the biliary clearance of (14C)mannitol. J Clin Invest,
1974;54:773–81.
6. Javitt, N.B. History of hepatic bile formation: old problems, new approaches.
Adv Physiol Educ, 2014;38:279–85.
7. Marinelli, R.A. and LaRusso N.F. Aquaporin water channels in liver: their
significance in bile formation. Hepatology, 1997;26:1081–4.
8. O’Maille E.R. and Hofmann, A.F. Relatively high biliary secretory maxi-
mum for non‐micelle‐forming bile acid: possible significance for mecha-
nism of secretion. Q J Exp Physiol, 1986;71:475–82.
9. Kamimoto Y., Gatmaitan Z., Hsu J., and Arias, I.M. The function of Gp170,
the multidrug resistance gene product, in rat liver canalicular membrane
vesicles. J Biol Chem, 1989;264:11693–8.
10. Nishida T., Gatmaitan Z., Che M., and Arias, I.M. Rat liver canalicular mem-
brane vesicles contain an ATP‐dependent bile acid transport system. Proc
Natl Acad Sci U S A, 1991;88:6590–4.
11. Kitamura T., Jansen P., Hardenbrook C., Kamimoto Y., Gatmaitan Z., and
Arias, I.M. Defective ATP‐dependent bile canalicular transport of organic
anions in mutant (TR‐) rats with conjugated hyperbilirubinemia. Proc Natl
Acad Sci U S A, 1990;87:3557–61.
12. Schulman, F.Y. Montali, R.J., Bush, M. et al. Dubin‐Johnson‐like Syndrome
in Golden Lion Tamarins (Leontopithecus rosalia rosalia). Vet Pathol,
1991;30:8.
13. Jansen, P.L., Peters, W.H., and Lamers, W.H. Hereditary chronic conjugated
hyperbilirubinemia in mutant rats caused by defective hepatic anion trans-
port. Hepatology, 1985;5:573–9.
14. Alpert S., Mosher M., Shanske, A., and Arias, I.M. Multiplicity of hepatic
excretory mechanisms for organic anions. J Gen Physiol, 1969;53:238–7.
15. Jansen, P.L., Groothuis, G.M., Peters, W.H., and Meijer, D.F. Selective hepato-
biliary transport defect for organic anions and neutral steroids in mutant rats
with hereditary‐conjugated hyperbilirubinemia. Hepatology, 1987;7:71–6.
16. Paulusma, C.C., Bosma, P.J., Zaman, G.J. et al. Congenital jaundice in rats
with a mutation in a multidrug resistance‐associated protein gene. Science,
1996;271:1126–8.
17. Gerloff, T., Stieger, B., Hagenbuch, B. et  al. The sister of P‐glycoprotein
represents the canalicular bile salt export pump of mammalian liver. J Biol
Chem, 1998;273:10046–50.
18. Juliano, R.L. and Ling, V. A surface glycoprotein modulating drug permea-
bility in Chinese hamster ovary cell mutants. Biochim Biophys Acta,
1976;455:152–62.
19. Vlaming, M.L., Mohrmann, K., Wagenaar, E. et al. Carcinogen and antican-
cer drug transport by Mrp2 in vivo: studies using Mrp2 (Abcc2) knockout
mice. J Pharmacol Exp Ther, 2006;318:319–27.
20. Lam P., Wang R., and Ling, V. Bile acid transport in sister of P‐glycoprotein
(ABCB11) knockout mice. Biochemistry, 2005;44:12598–605.
21. Smit, J.J., Schinkel, A.H., Oude Elferink, R.P. et al. Homozygous disruption
of the murine mdr2 P‐glycoprotein gene leads to a complete absence of
phospholipid from bile and to liver disease. Cell, 1993;75:451–62.
22. Halpern, K.B., Shenhav, R., Matcovitch‐Natan, O. et al. Single‐cell spatial
reconstruction reveals global division of labour in the mammalian liver.
Nature, 2017;542:352–6.
23. Meyer, K., Ostrenko, O., Bourantas, G. et al. A predictive 3D multi‐scale model
of biliary fluid dynamics in the liver lobule. Cell Syst, 2017;4:277–90.
24. Fu, D., Wakabayashi, Y., Ido, Y., Lippincott‐Schwartz, J., and Arias, I.M.
Regulation of bile canalicular network formation and maintenance by AMP‐
activated protein kinase and LKB1. J Cell Sci, 2010;123:3294–302.
25. Jansen, P.L., Strautnieks, S.S., Jacquemin, E. et al. Hepatocanalicular bile
salt export pump deficiency in patients with progressive familial intrahepatic
cholestasis. Gastroenterology, 1999;117:1370–9.
26. Kruglov, E.A., Gautam, S., Guerra, M.T., and Nathanson, M.H. Type 2 ino-
sitol 1,4,5‐trisphosphate receptor modulates bile salt export pump activity in
rat hepatocytes. Hepatology, 2011;54:1790–9.
27. Mills, C.O., Milkiewicz, P., Muller, M. et al. Different pathways of canalicu-
lar secretion of sulfated and non‐sulfated fluorescent bile acids: a study in
isolated hepatocyte couplets and TR‐ rats. J Hepatol, 1999;31:678–84.
28. Porat‐Shliom, N., Tietgens, A.J., Van Itallie, C.M. et  al. Liver kinase B1
regulates hepatocellular tight junction distribution and function in vivo.
Hepatology, 2016;64:1317–29.
29. Groothuis, G.M., Hardonk, M.J., Keulemans, K.P., Nieuwenhuis, P., and
Meijer, D.K. Autoradiographic and kinetic demonstration of acinar heteroge-
neity of taurocholate transport. Am J Physiol, 1982;243:G455–62.
30. Ling, V., Kartner, N., Sudo, T., Siminovitch, L., and Riordan, J.R. Multidrug‐
resistance phenotype in Chinese hamster ovary cells. Cancer Treat Rep,
1983;67:869–74.
31. Kartner, N., Shales, M., Riordan, J.R., and Ling, V. Daunorubicin‐resistant
Chinese hamster ovary cells expressing multidrug resistance and a cell‐sur-
face P‐glycoprotein. Cancer Res, 1983;43:4413–9.
32. Higgins, C.F. ABC transporters: from microorganisms to man. Annu Rev
Cell Biol, 1992;8:67–113.
33. Childs, S. and Ling, V. The MDR superfamily of genes and its biological
implications. Important Adv Oncol, 1994:21–36.
34. Vasiliou, V., Vasiliou, K., and Nebert, D.W. Human ATP‐binding cassette
(ABC) transporter family. Hum Genomics, 2009;3:281–90.
35. Chen, Z., Shi, T., Zhang, L. et al. Mammalian drug efflux transporters of the
ATP binding cassette (ABC) family in multidrug resistance: a review of the
past decade. Cancer Lett, 2016;370:153–64.
36. Ueda, K., Cornwell, M.M., Gottesman, M.M., Pastan, I., Roninson, I.B., Ling,
V., and Riordan, J.R. The mdr1 gene, responsible for multidrug‐resistance,
codes for P‐glycoprotein. Biochem Biophys Res Commun, 1986;141:956–62.
37. Gottesman, M.M. and Pastan, I.H. The role of multidrug resistance efflux
pumps in cancer: revisiting a JNCI publication exploring expression of the
MDR1 (P‐glycoprotein) gene. J Natl Cancer Inst, 2015;107.
38. Crawford, A.R., Smith, A.J., Hatch, V.C., Oude Elferink, R.P., Borst, P., and
Crawford, J.M. Hepatic secretion of phospholipid vesicles in the mouse criti-
cally depends on mdr2 or MDR3 P‐glycoprotein expression. Visualization
by electron microscopy. J Clin Invest, 1997;100:2562–7.
39. Oude Elferink, R.P., Ottenhoff, R., van Wijland, M., Smit, J.J., Schinkel,
A.H., and Groen, A.K. Regulation of biliary lipid secretion by mdr2 P‐glyco-
protein in the mouse. J Clin Invest, 1995;95:31–8.
40. Lee, S.D., Thornton, S.J., Sachs‐Barrable, K., Kim, J.H., and Wasan, K.M.
Evaluation of the contribution of the ATP binding cassette transporter, P‐gly-
coprotein, to in vivo cholesterol homeostasis. Mol Pharm, 2013;10:3203–12.
41. Berge, K.E., Tian, H., Graf, G.A. et al. Accumulation of dietary cholesterol
in sitosterolemia caused by mutations in adjacent ABC transporters. Science,
2000;290:1771–5.
42. Elferink, R.P., Ottenhoff, R., Liefting, W., de Haan, J., and Jansen, P.L.
Hepatobiliary transport of glutathione and glutathione conjugate in rats with
hereditary hyperbilirubinemia. J Clin Invest, 1989;84:476–83.
324 THE LIVER:  REFERENCES
43. Rosmorduc, O., Hermelin, B., and Poupon, R. MDR3 gene defect in adults
with symptomatic intrahepatic and gallbladder cholesterol cholelithiasis.
Gastroenterology, 2001;120:1459–67.
44. Rebholz, C., Krawczyk, M., and Lammert, F. Genetics of gallstone disease.
J Clin Invest,Eur J Clin Invest, 2018;48:e12935.
45. Hirschfield, G.M., Chapman, R.W., Karlsen, T.H., Lammert, F., Lazaridis,
K.N., and Mason, A.L. The genetics of complex cholestatic disorders.
Gastroenterology, 2013;144:1357–74.
46. Paulusma, C.C., de Waart, D.R., Kunne, C., Mok, K.S., and Elferink, R.P.
Activity of the bile salt export pump (ABCB11) is critically dependent on
canalicular membrane cholesterol content. J Biol Chem, 2009;
284:9947–54.
47. Guyot, C., Hofstetter, L., and Stieger, B. Differential effects of membrane
cholesterol content on the transport activity of multidrug resistance‐associ-
ated protein 2 (ABCC2) and of the bile salt export pump (ABCB11). Mol
Pharmacol, 2014;85:909–20.
48. Szilagyi, J.T., Vetrano, A.M., Laskin, J.D., and Aleksunes, L.M. Localization
of the placental BCRP/ABCG2 transporter to lipid rafts: role for cholesterol
in mediating efflux activity. Placenta, 2017;55:29–36.
49. Telbisz, A., Hegedus, C., Varadi, A., Sarkadi, B., and Ozvegy‐Laczka, C.
Regulation of the function of the human ABCG2 multidrug transporter by
cholesterol and bile acids: effects of mutations in potential substrate and ster-
oid binding sites. Drug Metab Dispos, 2014;42:575–85.
50. Paulusma, C.C., Groen, A., Kunne, C. et  al. Atp8b1 deficiency in mice
reduces resistance of the canalicular membrane to hydrophobic bile salts and
impairs bile salt transport. Hepatology, 2006;44:195–204.
51. Bull, L.N., Carlton, V.E., Stricker, N.L. et  al. Genetic and morphological
findings in progressive familial intrahepatic cholestasis (Byler disease
[PFIC‐1] and Byler syndrome): evidence for heterogeneity. Hepatology,
1997;26:155–64.
52. Sinke, R.J., Carlton, V.E., Juijn, J.A. et  al. Benign recurrent intrahepatic
cholestasis (BRIC): evidence of genetic heterogeneity and delimitation of
the BRIC locus to a 7‐cM interval between D18S69 and D18S64. Hum
Genet, 1997;100:382–7.
53. Bull, L.N., Pawlikowska, L., Strautnieks, S. et al. Outcomes of surgical man-
agement of familial intrahepatic cholestasis 1 and bile salt export protein
deficiencies. Hepatol Commun, 2018;2:515–28.
54. Pauli‐Magnus, C., Kerb, R., Fattinger, K. et al. BSEP and MDR3 haplotype
structure in healthy Caucasians, primary biliary cirrhosis and primary scle-
rosing cholangitis. Hepatology, 2004;39:779–91.
55. Ji, S.G., Juran, B.D., Mucha, S. et al. Genome‐wide association study of pri-
mary sclerosing cholangitis identifies new risk loci and quantifies the genetic
relationship with inflammatory bowel disease. Nat Genet, 2017;49:269–73.
56. de Vree, J.M., Jacquemin, E., Sturm, E. et al. Mutations in the MDR3 gene
cause progressive familial intrahepatic cholestasis. Proc Natl Acad Sci U S A,
1998;95:282–7.
57. Smith, A.J., de Vree, J.M., Ottenhoff, R., Oude Elferink, R.P., Schinkel,
A.H., and Borst, P. Hepatocyte‐specific expression of the human MDR3 P‐
glycoprotein gene restores the biliary phosphatidylcholine excretion absent
in Mdr2 (‐/‐) mice. Hepatology, 1998;28:530–6.
58. Deleuze, J.F., Jacquemin, E., Dubuisson, C. et al. Defect of multidrug‐resist-
ance 3 gene expression in a subtype of progressive familial intrahepatic chol-
estasis. Hepatology, 1996;23:904–8.
59. Jacquemin, E., Cresteil, D., Manouvrier, S., Boute, O., and Hadchouel, M.
Heterozygous non‐sense mutation of the MDR3 gene in familial intrahepatic
cholestasis of pregnancy. Lancet, 1999;353:210–1.
60. Dixon, P.H., Weerasekera, N., Linton, K.J. et al. Heterozygous MDR3 mis-
sense mutation associated with intrahepatic cholestasis of pregnancy: evi-
dence for a defect in protein trafficking. Hum Mol Genet, 2000;9:1209–17.
61. Kok, T., Wolters, H., Bloks, V.W. et al. Induction of hepatic ABC transporter
expression is part of the PPARalpha‐mediated fasting response in the mouse.
Gastroenterology, 2003;124:160–71.
62. Fickert, P., Wagner, M., Marschall, H.U. et al. 24‐norUrsodeoxycholic acid
is superior to ursodeoxycholic acid in the treatment of sclerosing cholangitis
in Mdr2 (Abcb4) knockout mice. Gastroenterology, 2006;130:465–81.
63. Wang, R., Salem, M., Yousef, I.M. et al. Targeted inactivation of sister of P‐
glycoprotein gene (spgp) in mice results in nonprogressive but persistent
intrahepatic cholestasis. Proc Natl Acad Sci U S A, 2001;98:2011–6.
64. Hrycay, E., Forrest, D., Liu, L. et  al. Hepatic bile acid metabolism and
expression of cytochrome P450 and related enzymes are altered in Bsep (‐/‐)
mice. Mol Cell Biochem, 2014;389:119–32.
65. Strautnieks, S.S., Bull, L.N., Knisely, A.S. et al. A gene encoding a liver‐spe-
cific ABC transporter is mutated in progressive familial intrahepatic choles-
tasis. Nat Genet, 1998;20:233–8.
66. van Mil, S.W., van der Woerd, W.L., van der Brugge, G. et al. Benign recur-
rent intrahepatic cholestasis type 2 is caused by mutations in ABCB11.
Gastroenterology, 2004;127:379–84.
67. Knisely, A.S., Strautnieks, S.S., Meier, Y. et al. Hepatocellular carcinoma in
ten children five years of age with bile salt export deficience. Hepatology,
2006;44:478–86.
68. Siebold, L., Dick, A.A., Thompson, R. et al. Recurrent low gamma‐glutamyl
transpeptidase cholestasis following liver transplantation for bile salt export
pump (BSEP) disease (posttransplant recurrent BSEP disease). Liver
Transpl, 2010;16:856–63.
69. Kubitz, R., Droge, C., Kluge, S. et  al. High affinity anti‐BSEP antibodies
after liver transplantation for PFIC‐2 – successful treatment with immunoad-
sorption and B‐cell depletion. Pediatr Transplant, 2016;20:987–93.
70. Hayashi, H. and Sugiyama, Y. 4‐phenylbutyrate enhances the cell surface
expression and the transport capacity of wild‐type and mutated bile salt
export pumps. Hepatology, 2007;45:1506–16.
71. Hirano, M., Maeda, K., Hayashi, H., Kusuhara, H., and Sugiyama, Y. Bile
salt export pump (BSEP/ABCB11) can transport a nonbile acid substrate,
pravastatin. J Pharmacol Exp Ther, 2005;314:876–82.
72. Fattinger, K., Funk, C., Pantze, M. et  al. The endothelin antagonist bosentan
inhibits the canalicular bile salt export pump: a potential mechanism for hepatic
adverse reactions. Pharmacol Ther Clin Pharmacol Ther, 2001;69:223–31.
73. Funk, C., Ponelle, C., Scheuermann, G., and Pantze, M. Cholestatic potential
of troglitazone as a possible factor contributing to troglitazone‐induced
hepatotoxicity: in vivo and in vitro interaction at the canalicular bile salt
export pump (Bsep) in the rat. Mol Pharmacol, 2001;59:627–35.
74. Byrne, J.A., Strautnieks, S.S., Mieli‐Vergani, G., Higgins, C.F., Linton, K.J.,
and Thompson, R.J. The human bile salt export pump: characterization of
substrate specificity and identification of inhibitors. Gastroenterology,
2002;123:1649–58.
75. Nicolaou, M., Andress, E.J., Zolnerciks, J.K., Dixon, P.H., Williamson, C.,
and Linton, K.J. Canalicular ABC transporters and liver disease. J Pathol,
2012;226:300–15.
76. Tanigawara, Y., Okamura, N., Hirai, M. et al. Transport of digoxin by human
P‐glycoprotein expressed in a porcine kidney epithelial cell line (LLC‐PK1).
J Pharmacol Exp Ther, 1992;263:840–5.
77. Roulet, A., Puel, O., Gesta, S. et al. MDR1‐deficient genotype in Collie dogs
hypersensitive to the P‐glycoprotein substrate ivermectin. Eur J Pharmacol,
2003;460:85–91.
78. Schinkel, A.H., Smit, J.J., van Tellingen, O. et  al. Disruption of the
mouse mdr1a P‐glycoprotein gene leads to a deficiency in the blood‐
brain barrier and to increased sensitivity to drugs. Cell, 1994;
77:491–502.
79. Ellis, J.L., Bove, K.E., Schuetz, E.G. et al. Zebrafish abcb11b mutant reveals
strategies to restore bile excretion impaired by bile salt export pump defi-
ciency. Hepatology, 2018;67:1531–45.
80. Wicek, K. and Stieger, B. ATP‐binding cassette transporters in liver.
Biofactors, 2014;40:188–98.
81. Bakos, E., Evers, R., Szakacs, G. et al. Functional multidrug resistance pro-
tein (MRP1) lacking the N‐terminal transmembrane domain. J Biol Chem,
1998;273:32167–75.
82. Megaraj, V., Iida, T., Jungsuwadee, P., Hofmann, A.F., and Vore, M.
Hepatobiliary disposition of 3alpha,6alpha,7alpha,12alpha‐tetrahydroxy‐
cholanoyl taurine: a substrate for multiple canalicular transporters. Drug
Metab Dispos, 2010;38:1723–30.
83. Huber, M., Guhlmann, A., Jansen, P.L., and Keppler, D. Hereditary defect of
hepatobiliary cysteinyl leukotriene elimination in mutant rats with defective
hepatic anion excretion. Hepatology, 1987;7:224–8.
84. Cui, Y., Konig, J., Buchholz, J.K., Spring, H., Leier, I., and Keppler, D. Drug
resistance and ATP‐dependent conjugate transport mediated by the apical
multidrug resistance protein, MRP2, permanently expressed in human and
canine cells. Mol Pharmacol, 1999;55:929–37.
85. Kamisako, T., Leier, I., Cui, Y. et  al. Transport of monoglucuronosyl and
bisglucuronosyl bilirubin by recombinant human and rat multidrug resist-
ance protein 2. Hepatology, 1999;30:485–90.
86. Madon, J., Hagenbuch, B., Landmann, L., Meier, P.J., and Stieger, B.
Transport function and hepatocellular localization of mrp6 in rat liver. Mol
Pharmacol, 2000;57:634–41.
 26:  Hepatic Adenosine Triphosphate‐Binding Cassette Transport Proteins and Their Role in Physiology 325
87. Dijkstra, M., Havinga, R., Vonk, R.J., and Kuipers, F. Bile secretion of cad-
mium, silver, zinc and copper in the rat. Involvement of various transport
systems. Life Sci, 1996;59:1237–46.
88. Sandusky, G.E., Mintze, K.S., Pratt, S.E., and Dantzig, A.H. Expression of
multidrug resistance‐associated protein 2 (MRP2) in normal human tissues
and carcinomas using tissue microarrays. Histopathology, 2002;41:65–74.
89. Rost, D., Konig, J., Weiss, G., Klar, E., Stremmel, W., and Keppler, D.
Expression and localization of the multidrug resistance proteins MRP2 and
MRP3 in human gallbladder epithelia. Gastroenterology, 2001;121:1203–8.
90. Kartenbeck, J., Leuschner, U., Mayer, R., and Keppler, D. Absence of the
canalicular isoform of the MRP gene‐encoded conjugate export pump from
the hepatocytes in Dubin–Johnson syndrome. Hepatology, 1996;23:1061–6.
91. Kitamura, T., Alroy, J., Gatmaitan, Z. et al. Defective biliary excretion of
epinephrine metabolites in mutant (TR‐) rats: relation to the pathogenesis
of black liver in the Dubin–Johnson syndrome and Corriedale sheep with
an analogous excretory defect. Hepatology, 1992;15:1154–9.
92. Lagas, J.S., Vlaming, M.L., van Tellingen, O. et al. Multidrug resistance
protein 2 is an important determinant of paclitaxel pharmacokinetics. Clin
Cancer Res Cancer Res, 2006;12:6125–32.
93. Borel, F., Han, R., Visser, A. et al. Adenosine triphosphate‐binding cassette
transporter genes up‐regulation in untreated hepatocellular carcinoma is
mediated by cellular microRNAs. Hepatology, 2012;55:821–32.
94. Nies, A.T., Konig, J., Pfannschmidt, M., Klar, E., Hofmann, W.J., and
Keppler, D. Expression of the multidrug resistance proteins MRP2 and
MRP3 in human hepatocellular carcinoma. Int J Cancer, 2001;94:492–9.
95. Recknagel, P., Gonnert, F.A., Westermann, M. et al. Liver dysfunction and
phosphatidylinositol‐3‐kinase signalling in early sepsis: experimental stud-
ies in rodent models of peritonitis. PLoS Med, 2012;9:e1001338.
96. Trauner, M., Arrese, M., Soroka, C.J. et al. The rat canalicular conjugate
export pump (Mrp2) is down‐regulated in intrahepatic and obstructive chol-
estasis. Gastroenterology, 1997;113:255–64.
97. Hinoshita, E., Taguchi, K., Inokuchi, A. et al. Decreased expression of an
ATP‐binding cassette transporter, MRP2, in human livers with hepatitis C
virus infection. J Hepatol, 2001;35:765–73.
98. Zinchuk, V., Zinchuk, O., Akimaru, K., Moriya, F., and Okada, T. Ethanol con-
sumption alters expression and colocalization of bile salt export pump and mul-
tidrug resistance protein 2 in the rat. Histochem Cell Biol, 2007;127:503–12.
99. Ros, J.E., Libbrecht, L., Geuken, M., Jansen, P.L., and Roskams, T.A.
High expression of MDR1, MRP1, and MRP3 in the hepatic progenitor
cell compartment and hepatocytes in severe human liver disease. J
Pathol, 2003;200:553–60.
100. Ros, J.E., Roskams, T.A., Geuken, M. et  al. ATP binding cassette trans-
porter gene expression in rat liver progenitor cells. Gut, 2003;52:1060–7.
101. Silverton, L., Dean, M., and Moitra, K. Variation and evolution of the ABC
transporter genes ABCB1, ABCC1, ABCG2, ABCG5 and ABCG8: impli-
cation for pharmacogenetics and disease. Drug Metabol Drug Interact,
2011;26:169–79.
102. Harris, M.J., Kuwano, M., Webb, M., and Board, P.G. Identification of the
apical membrane‐targeting signal of the multidrug resistance‐associated
protein 2 (MRP2/MOAT). J Biol Chem, 2001;276:20876–81.
103. Vander Borght, S., Komuta, M., Libbrecht, L. et al. Expression of multid-
rug resistance‐associated protein 1 in hepatocellular carcinoma is associ-
ated with a more aggressive tumour phenotype and may reflect a progenitor
cell origin. Liver Int, 2008;28:1370–80.
104. Soroka, C.J., Lee, J.M., Azzaroli, F., and Boyer, J.L. Cellular localization
and up‐regulation of multidrug resistance‐associated protein 3 in hepato-
cytes and cholangiocytes during obstructive cholestasis in rat liver.
Hepatology, 2001;33:783–91.
105. Konig, J., Rost, D., Cui, Y., and Keppler, D. Characterization of the human
multidrug resistance protein isoform MRP3 localized to the basolateral
hepatocyte membrane. Hepatology, 1999;29:1156–63.
106. Scheffer, G.L., Kool, M., de Haas, M. et al. Tissue distribution and induc-
tion of human multidrug resistant protein 3. Lab Invest, 2002;82:193–201.
107. Donner, M.G. and Keppler, D. Up‐regulation of basolateral multidrug resist-
ance protein 3 (Mrp3) in cholestatic rat liver. Hepatology, 2001;34:351–9.
108. Hirohashi, T., Suzuki, H., and Sugiyama, Y. Characterization of the trans-
port properties of cloned rat multidrug resistance‐associated protein 3
(MRP3). J Biol Chem, 1999;274:15181–5.
109. Conseil, G., Deeley, R.G., and Cole, S.P. Polymorphisms of MRP1
(ABCC1) and related ATP‐dependent drug transporters. Pharmacogenet
Genomics, 2005;15:523–33.
110. Keppler, D. Cholestasis and the role of basolateral efflux pumps. Z
Gastroenterol, 2011;49:1553–7.
111. Zelcer, N., Reid, G., Wielinga, P. et al. Steroid and bile acid conjugates are
substrates of human multidrug‐resistance protein (MRP) 4 (ATP‐binding
cassette C4). Biochem, J., 2003;371:361–7.
112. Assem, M., Schuetz, E.G., Leggas, M. et al. Interactions between hepatic
Mrp4 and Sult2a as revealed by the constitutive androstane receptor and
Mrp4 knockout mice. J Biol Chem, 2004;279:22250–7.
113. Rius, M., Hummel‐Eisenbeiss, J., Hofmann, A.F., and Keppler, D. Substrate
specificity of human ABCC4 (MRP4)‐mediated cotransport of bile acids
and reduced glutathione. Am J Physiol Gastrointest Liver Physiol,
2006;290:G640–9.
114. Denk, G.U., Soroka, C.J., Takeyama, Y., Chen, W.S., Schuetz, J.D., and
Boyer, J.L. Multidrug resistance‐associated protein 4 is up‐regulated in
liver but down‐regulated in kidney in obstructive cholestasis in the rat.
J Hepatol, 2004;40:585–91.
115. Favre, G., Laurain, A., Aranyi, T. et  al. The ABCC6 transporter: a new
player in biomineralization. Int J Mol Sci, 2017;18. e1941.
116. Jansen, R.S., Duijst, S., Mahakena, S. et al. ABCC6‐mediated ATP secre-
tion by the liver is the main source of the mineralization inhibitor inorganic
pyrophosphate in the systemic circulation‐brief report. Arterioscler Thromb
Vasc Biol, 2014;34:1985–9.
117. Jansen, R.S., Kucukosmanoglu, A., de Haas, M. et  al. ABCC6 prevents
ectopic mineralization seen in pseudoxanthoma elasticum by inducing cel-
lular nucleotide release. Proc Natl Acad Sci U S A, 2013;110:20206–11.
118. Riordan, J.R., Rommens, J.M., Kerem, B. et al. Identification of the cystic
fibrosis gene: cloning and characterization of complementary DNA.
Science, 1989;245:1066–73.
119. Jonker, J.W., Buitelaar, M., Wagenaar, E. et al. The breast cancer resistance
protein protects against a major chlorophyll‐derived dietary phototoxin and
protoporphyria. Proc Natl Acad Sci U S A, 2002;99:15649–54.
120. Vander Borght, S., Libbrecht, L., Katoonizadeh, A. et  al. Breast cancer
resistance protein (BCRP/ABCG2) is expressed by progenitor cells/reac-
tive ductules and hepatocytes and its expression pattern is influenced by
disease etiology and species type: possible functional consequences. J
Histochem Cytochem, 2006;54:1051–9.
121. Jakulj, L., Vissers, M.N., Tanck, M.W. et al. ABCG5/G8 polymorphisms
and markers of cholesterol metabolism: systematic review and meta‐analy-
sis. J Lipid Res, 2010;51:3016–23.
122. Yu, L., Li‐Hawkins, J., Hammer, R.E. et al. Overexpression of ABCG5 and
ABCG8 promotes biliary cholesterol secretion and reduces fractional
absorption of dietary cholesterol. J Clin Invest, 2002;110:671–80.
123. von Kampen, O., Buch, S., Nothnagel, M. et  al. Genetic and functional
identification of the likely causative variant for cholesterol gallstone dis-
ease at the ABCG5/8 lithogenic locus. Hepatology, 2013;57:2407–17.
124. Buch, S., Schafmayer, C., Volzke, H. et al. A genome‐wide association scan
identifies the hepatic cholesterol transporter ABCG8 as a susceptibility fac-
tor for human gallstone disease. Nat Genet, 2007;39:995–9.
125. Makishima, M., Okamoto, A.Y., Repa, J.J. et al. Identification of a nuclear
receptor for bile acids. Science, 1999;284:1362–5.
126. Kliewer, S.A., Moore, J.T., Wade, L. et al. An orphan nuclear receptor acti-
vated by pregnanes defines a novel steroid signaling pathway. Cell,
1998;92:73–82.
127. Baes, M., Gulick, T., Choi, H.S., Martinoli, M.G., Simha, D., and Moore,
D.D. A new orphan member of the nuclear hormone receptor superfamily
that interacts with a subset of retinoic acid response elements. Mol Cell
Biol, 1994;14:1544–52.
128. Boyer, J.L., Trauner, M., Mennone, A. et al. Upregulation of a basolateral
FXR‐dependent bile acid efflux transporter OSTalpha‐OSTbeta in choles-
tasis in humans and rodents. Am J Physiol Gastrointest Liver Physiol,
2006;290:G1124–30.
129. Keppler, D. The roles of MRP2, MRP3, OATP1B1, and OATP1B3 in con-
jugated hyperbilirubinemia. Drug Metab Dispos, 2014;42:561–5.
130. Mennone, A., Soroka, C.J., Cai, S.Y. et al. Mrp4‐/‐ mice have an impaired
cytoprotective response in obstructive cholestasis. Hepatology,
2006;43:1013–21.
131. Parks, D.J., Blanchard, S.G., Bledsoe, R.K. et al. Bile acids: natural ligands
for an orphan nuclear receptor. Science, 1999;284:1365–8.
132. Wang, H., Chen, J., Hollister, K., Sowers, L.C., and Forman, B.M.
Endogenous bile acids are ligands for the nuclear receptor FXR/BA.R. Mol
Cell, 1999;3:543–53.
326 THE LIVER:  REFERENCES
133. Kir, S., Zhang, Y., Gerard, R.D., Kliewer, S.A., and Mangelsdorf, D.J.
Nuclear receptors HNF4alpha and LRH‐1 cooperate in regulating Cyp7a1
in vivo. J Biol Chem, 2012;287:41334–41.
134. Cai, S.Y. and Boyer, J.L. FXR: a target for cholestatic syndromes? Expert
Opin Ther Targets, 2006;10:409–21.
135. Schaap, F.G., van der Gaag, N.A., Gouma, D.J., and Jansen, P.L. High
expression of the bile salt‐homeostatic hormone fibroblast growth factor 19
in the liver of patients with extrahepatic cholestasis. Hepatology,
2009;49:1228–35.
136. Gomez‐Ospina, N., Potter, C.J., Xiao, R. et al. Mutations in the nuclear bile
acid receptor FXR cause progressive familial intrahepatic cholestasis. Nat
Commun, 2016;7:10713.
137. Staudinger, J.L., Goodwin, B., Jones, S.A. et al. The nuclear receptor PXR
is a lithocholic acid sensor that protects against liver toxicity. Proc Natl
Acad Sci U S A, 2001;98:3369–74.
138. Xie, W., Radominska‐Pandya, A., Shi, Y. et al. An essential role for nuclear
receptors SXR/PXR in detoxification of cholestatic bile acids. Proc Natl
Acad Sci U S A, 2001;98:3375–80.
139. Huang, W., Zhang, J., Chua, S.S. et al. Induction of bilirubin clearance by
the constitutive androstane receptor (CAR). Proc Natl Acad Sci U S A,
2003;100:4156–61.
140. Makishima, M., Lu, T.T., Xie, W. et al. Vitamin D receptor as an intestinal
bile acid sensor. Science, 2002;296:1313–6.
141. van Dijk, R., Kremer, A.E., Smit, W. et al. Characterization and treatment
of persistent hepatocellular secretory failure. Liver Int, 2015;35:1478–88.
142. Janssen, H.L., Reesink, H.W., Lawitz, E.J. et al. Treatment of HCV infec-
tion by targeting microRNA. N Engl J Med, 2013;368:1685–94.
143. Haenisch, S., Laechelt, S., Bruckmueller, H. et al. Down‐regulation of ATP‐
binding cassette C2 protein expression in HepG2 cells after rifampicin treat-
ment is mediated by microRNA‐379. Mol Pharmacol 2011; 80:314–20.
144. Haenisch, S., Werk, A.N., and Cascorbi, I. MicroRNAs and their relevance
to ABC transporters. Br J Clin Pharmacol, 2014;77:587–96.
145. Yuan, K., Zhang, X., Lv, L., Zhang, J., Liang, W., and Wang, P. Fine‐tuning
the expression of microRNA‐155 controls acetaminophen‐induced liver
inflammation. Int Immunopharmacol, 2016;40:339–46.
146. Church, R.J., Kullak‐Ublick, G.A., Aubrecht, J. et al. Candidate biomark-
ers for the diagnosis and prognosis of drug‐induced liver injury: an interna-
tional collaborative effort. Hepatology, 2019;69:760–73.
147. Howell, L.S., Ireland, L., Park, B.K., and Goldring, C.E. MiR‐122 and
other microRNAs as potential circulating biomarkers of drug‐induced liver
injury. Expert Rev Mol Diagn, 2018;18:47–54.
148. Rao, Z.Z., Zhang, X.W., Ding, Y.L., and Yang, M.Y. miR‐148a‐mediated
estrogen‐induced cholestasis in intrahepatic cholestasis of pregnancy: role
of PXR/MRP3. PLoS One, 2017;12:e0178702.
149. Kipp, H., Pichetshote, N., and Arias, I.M. Transporters on demand: intrahe-
patic pools of canalicular ATP binding cassette transporters in rat liver. J
Biol Chem, 2001;276:7218–24.
150. Beuers, U., Boyer, J.L., and Paumgartner, G. Ursodeoxycholic acid in chol-
estasis: potential mechanisms of action and therapeutic applications.
Hepatology, 1998;28:1449–53.
151. Crocenzi, F.A., Mottino, A.D., Cao, J. et al. Estradiol‐17beta‐D‐glucuron-
ide induces endocytic internalization of Bsep in rats. Am J Physiol
Gastrointest Liver Physiol, 2003;285:G449–59.
152. Perez, L.M., Milkiewicz, P., Elias, E., Coleman, R., Sanchez Pozzi, E.J.,
and Roma, M.G. Oxidative stress induces internalization of the bile salt
export pump, Bsep, and bile salt secretory failure in isolated rat hepatocyte
couplets: a role for protein kinase C and prevention by protein kinase A.
Toxicol Sci, 2006;91:150–8.
153. Haussinger, D., Kurz, A.K., Wettstein, M., Graf, D., Vom Dahl, S., and
Schliess, F. Involvement of integrins and Src in tauroursodeoxycholate‐
induced and swelling‐induced choleresis. Gastroenterology, 2003;
124:1476–87.
154. Elferink, R.P. and Paulusma, C.C. MRP2 in cholestasis: putting down the
anchor. J Hepatol, 2015;63:1309–10.
155. Ortiz, D.F., Moseley, J., Calderon, G., Swift, A.L., Li, S., and Arias, I.M.
Identification of HAX‐1 as a protein that binds bile salt export protein and
regulates its abundance in the apical membrane of Madin‐Darby canine
kidney cells. J Biol Chem, 2004;279:32761–70.
156. Paulusma, C.C., Kothe, M.J., Bakker, C.T. et al. Zonal down‐regulation and
redistribution of the multidrug resistance protein 2 during bile duct ligation
in rat liver. Hepatology, 2000;31:684–93.
157. Miszczuk, G.S., Barosso, I.R., Larocca, M.C. et al. Mechanisms of canali-
cular transporter endocytosis in the cholestatic rat liver. Biochim Biophys
Acta, 2018;1864:1072–85.
158. Kikuchi, S., Hata, M., Fukumoto, K. et al. Radixin deficiency causes con-
jugated hyperbilirubinemia with loss of Mrp2 from bile canalicular mem-
branes. Nat Genet, 2002;31:320–5.
159. Kojima, H., Sakurai, S., Uemura, M. et al. Disturbed colocalization of mul-
tidrug resistance protein 2 and radixin in human cholestatic liver diseases.
J Gastroenterol Hepatol, 2008;23:e120–8.
160. Kojima, H., Sakurai, S., Yoshiji, H., Uemura, M., Yoshikawa, M., and
Fukui, H. The role of radixin in altered localization of canalicular conjugate
export pump Mrp2 in cholestatic rat liver. Hepatol Res, 2008;38:202–10.
161. Chai, J., Cai, S.Y., Liu, X. et  al. Canalicular membrane MRP2/ABCC2
internalization is determined by Ezrin Thr567 phosphorylation in human
obstructive cholestasis. J Hepatol, 2015;63:1440–8.
162. Stieger, B., Fattinger, K., Madon, J., Kullak‐Ublick, G.A., and Meier, P.J.
Drug‐ and estrogen‐induced cholestasis through inhibition of the hepato-
cellular bile salt export pump (Bsep) of rat liver. Gastroenterology,
2000;118:422–30.
163. Ryan, J., Morgan, R.E., Chen, Y., Volak, L.P., Dunn, R.T., 2nd, and Dunn,
K.W. Intravital multiphoton microscopy with fluorescent bile salts in rats as
an in vivo biomarker for hepatobiliary transport inhibition. Drug Metab
Dispos, 2018;46:704–18.
164. Pauli‐Magnus, C., Meier, P.J., and Stieger, B. Genetic determinants of drug‐
induced cholestasis and intrahepatic cholestasis of pregnancy. Semin Liver
Dis, 2010;30:147–59.
165. Dixon, P.H., van Mil, S.W., Chambers, J. et  al. Contribution of variant
alleles of ABCB11 to susceptibility to intrahepatic cholestasis of preg-
nancy. Gut, 2009;58:537–44.
166. Pauli‐Magnus, C., Lang, T., Meier, Y. et al. Sequence analysis of bile salt
export pump (ABCB11) and multidrug resistance p‐glycoprotein 3
(ABCB4, MDR3) in patients with intrahepatic cholestasis of pregnancy.
Pharmacogenet, 2004;14:91–102.
167. Lang, C., Meier, Y., Stieger, B. et al. Mutations and polymorphisms in the
bile salt export pump and the multidrug resistance protein 3 associated with
drug‐induced liver injury. Pharmacogenet Genomics, 2007;17:47–60.
168. Sundaram, V. and Bjornsson, E.S. Drug‐induced cholestasis. Hepatol
Commun, 2017;1:726–35.
169. Zelcer, N., Saeki, T., Bot, I., Kuil, A., and Borst, P. Transport of bile acids
in multidrug‐resistance‐protein 3‐overexpressing cells co‐transfected with
the ileal Na+‐dependent bile‐acid transporter. Biochem J, 2003;
369:23–30.
170. Ghonem, N.S., Ananthanarayanan, M., Soroka, C.J., and Boyer, J.L.
Peroxisome proliferator‐activated receptor alpha activates human multid-
rug resistance transporter 3/ATP‐binding cassette protein subfamily B4
transcription and increases rat biliary phosphatidylcholine secretion.
Hepatology, 2014;59:1030–42.
171. Allen, R.M., Marquart, T.J., Albert, C.J. et al. miR‐33 controls the expres-
sion of biliary transporters, and mediates statin‐ and diet‐induced hepato-
toxicity. EMBO Mol Med, 2012;4:882–95.
172. To, K.K., Robey, R., Zhan, Z., Bangiolo, L., and Bates, S.E. Upregulation
of ABCG2 by romidepsin via the aryl hydrocarbon receptor pathway. Mol
Cancer Res, 2011;9:516–27.
 Basolateral Plasma
27 Membrane Organic Anion
Transporters
M. Sawkat Anwer1 and Allan W. Wolkoff2
1
Tufts Clinical and Translational Science Institute, Tufts University Cummings School of Veterinary
Medicine, Department of Biomedical Sciences, North, Grafton, MA, USA
2
Division of Hepatology, Marion Bessin Liver Research Center, Albert Einstein College of Medicine and
Montefiore Medical Center, Bronx, NY, USA

INTRODUCTION
The liver plays a central role in vectorial transport of solutes
from the portal vein to the canaliculus and this involves trans­
port of solutes via an array of transporters located at the baso­
lateral and the canalicular membrane of hepatocytes. Solutes
transported by hepatocytes include organic anions and cations,
inorganic ions, and neutral compounds. Many organic solutes
(such as bilirubin, bile acids) are albumin bound enabling
them to circulate in blood despite limited aqueous solubility.
The liver is ideally designed for extraction of protein‐bound
molecules using specific transporters. Apart from removing
solutes from the circulation for subsequent metabolic altera­
tions and reflux back to the circulation, vectorial transport of
solutes across the bile canaliculus provides the osmotic driv­
ing force for bile formation. Uptake of organic anions from
the  circulation across the basolateral plasma membrane and
subsequent biliary excretion, contributes significantly to
­ [8, 9]. A recent study based on clinical findings in two infants
bile ­formation. Organic anions are primarily transported
via  sodium‐dependent and sodium‐independent transporters
and include sodium‐taurocholate cotransporting polypeptide
(NTCP) and several members of the organic anion transport­
ing protein family (OATPs). This chapter will discuss func­
tions and regulations of these transporters.
SODIUM‐TAUROCHOLATE
COTRANSPORTING POLYPEPTIDE (NTCP)
Sodium‐dependent bile salt transport in rat hepatocytes was first
reported in 1978 [1] and NTCP was cloned from rat liver using
a functional expression cloning approach [2, 3]. It is the found­
ing member of the seven member SLC10A gene family [4, 5]
and is named Slc10a1 [6]. NTCP is a typical secondary active
transporter and strictly sodium‐dependent requiring the binding
of two sodium ions together with a bile salt molecule for the
transport step. The transport activity is electrogenic for monoan­
ionic bile salts, while it is electroneutral for the dianionic bile
salt, cholylglycylamidofluorescein [7]. Thus, the driving force
of the uptake can be the energy available from the chemical and/
or the electrochemical gradient of sodium.
Uptake of bile salts into hepatocytes occurs predominantly by
NTCP (SLC10A1) and to a lesser extent by OATPs (SLCOs)
with NTCP deficiency (based on Sanger sequencing of the
SLC10A1 gene) suggests the primary role of NTCP in hepatic
bile acid clearance [10]. While human NTCP (hNTCP) and rat
NTCP (rNTCP) are capable of transporting conjugated bile salts
as well as unconjugated bile acids [9], OATPs show a preference
for unconjugated bile acids [11–13].

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
328 THE LIVER:  SODIUM‐TAUROCHOLATE COTRANSPORTING POLYPEPTIDE (NTCP)
Transport functions of NTCP
The list of substrates transported by NTCP has been extensively
reviewed [4, 5, 9]. Solutes other than bile salts including steroid
hormones, thyroid hormones, drugs, and drugs conjugated
with  bile acids can be transported by hNTCP/rNTCP
(SLC10A1/Slc10a1) [5, 14]. In addition to endogenous sub­
strates, drugs including statins and an antifungal agent
micafungin [15], are transported by NTCP [5]. Interestingly,
rosuvastatin is transported by hNTCP, but not by rNTCP [16].
Thus, despite a large substrate overlap between rat, mouse, and
human NTCP [9] transport data should be extrapolated with
caution across different species. Clinical implications of trans­
port functions of NTCP is discussed below.
NTCP can serve as a transporter to deliver drugs into hepat­
ocytes. Hepatocellular carcinomas expressing NTCP mediate
high‐affinity Na+‐dependent uptake of chlorambucil‐taurocho­
late [17]. Thus, cytotoxic drugs conjugated with TC could pro­
vide a potential therapeutic strategy to deliver these drugs to
hepatocellular carcinomas. The role of transport systems in
liver function tests has been known for a long time [18] and is
increasingly appreciated today [19, 20]. Serum bile salt con­
centration is a good indicator of NTCP function and serves as
an indicator of liver function. Several exogenous compounds
used to assess dynamic liver functions in humans are taken up
by hNTCP among other transporters [21]. Thus, the functional
status of NTCP can be an important determinant of liver func­
tion testing and should be taken into consideration when
patients are treated with drugs known to interfere with NTCP
transport activity.
Drugs may inhibit solute transport by NTCP in two major
ways: (i) drugs may be transported by the NTCP resulting in
competitive inhibition and (ii) drugs may inhibit solute uptake
by NTCP without being themselves substrates for transport.
Indeed, there are drugs that inhibit hNTCP/rNTCP‐mediated
bile salt uptake, but are not transported by them [5, 9, 22]. Drugs
that inhibit solute transport by transporters with or without
being transported can have significant effects on the disposition
of endogenous substrates and drugs. These effects are important
considerations in drug development, as they may result in
adverse effects or provide therapeutic strategies. Thus, inhibi­
tion of NTCP can impair disposition of bile salts resulting in
higher systemic concentration of bile salts and consequent
adverse effects [23, 24].
Drug–drug interaction (DDI) is an important determinant of
drug safety and efficacy. It is becoming evident that transporters
often work together with drug metabolizing enzymes in drug
disposition. As a result, a role for transporters in drug dis­
position, therapeutic efficacy, and adverse drug reactions has
increasingly been recognized [25]. Some transporters are con­
sidered to be clinically important in drug absorption and dis­
position and therefore could mediate DDIs [26]. Although a
number of SLC transporters are considered to have clinical
importance in the absorption and disposition of drugs [26],
NTCP is not included. This is mainly because the clinical
importance of these transporters is not well established. Since
drugs can be transported by NTCP and can also inhibit solute
transport by NTCP, a potential role of NTCP in DDI is very
likely and has been reported [15, 27].
A novel function for NTCP as a receptor and transporter for
hepatitis B virus (HBV) and hepatitis D virus (HDV) has been
described [28] and human NTCP plays an important role in viral
entry and consequent infection [29]. Both HBV and HDV con­
tain three envelope proteins known as small (S), middle (M),
and large (L). These proteins share the same C‐terminal domain
corresponding to the S protein but differ at the N‐terminal
domains by the presence of the pre‐S2 domain in M protein and
pre‐S1 and pre‐S2 domains in L protein. The entry of both HBV
and HDV is determined by the pre‐S1 domain of L protein after
interacting with cellular receptor(s) on hepatocytes [30–32].
Myrcludex B, a drug based on myristoylated pre‐S1 domain,
has entered clinical trials [33] and an initial report shows prom­
ise [34]. However, viral entry is only one step in efficient viral
infection and replication in human hepatocytes and other fac­
tors, such as host components, are likely to be involved [35–38].
This novel function of NTCP raises interesting questions
about the mechanism of viral transport. One possibility is that
these viruses are transported by NTCP the same way as it trans­
ports bile salts. In that case, it would be important to know
whether the transport of HBV and HDV into hepatocytes is Na+
dependent and is inhibited by known substrates/inhibitors of
NTCP. A likely possibility is that NTCP is internalized by endo­
cytosis following binding to the virus allowing the virus to enter
the cell [35, 36]. Internalization of HBV/HDV may be initiated
by protein–protein interactions between the envelope protein of
the virus and NTCP leading to endocytosis. In that case, what
would be the signal for internalization? Calcium dependent con­
ventional protein kinase Cs (PKCs) and most likely PKCα
induces rNTCP retrieval [39, 40]. Whether PKCα or other sign­
aling molecules may be involved remains to be investigated.
Since NTCP traffics with the epidermal growth factor (EGF)
receptor via endocytic vesicles [41], HBV and HDV may use
this pathway to enter hepatocytes. Irrespective of the transport
mechanism involved, the finding that NTCP is involved in HBV
and HDV infection is expected to inspire further studies leading
to a better understanding of the role of NTCP in HBV/HDV
infection and development of agents to inhibit viral entry and
infection.
Transcriptional regulation
NTCP undergoes transcriptional as well as post‐transcriptional
regulation. Transcriptional regulation involving various nuclear
receptors (NRs) has been reviewed by others [4, 42]. Briefly,
NRs regulate NTCP expression by responding to changes in the
intracellular bile acid concentration (BAi). Increased BAi in
cholestasis activates the farnesoid X receptor (FXR) response
element in the small heterodimer partner (SHP) promoter lead­
ing to the expression of SHP, which in turn inhibits retinoid X
receptor (RXR)/retinoic acid receptor (RAR) activation of
NTCP (Figure 27.1). In addition, bile acid‐induced suppression
of hepatocyte nuclear factor (HNF1α), via activation of HNF4,
is involved in cholestasis‐induced downregulation of NTCP,
which is transcriptionally activated by HNF1α. Bile acids can
also suppress HNF1α via SHP. Downregulation of NTCP by
bile salts serves to decrease BAi and thereby limit or minimize bile
acid‐induced cellular toxicity in cholestasis. Proinflammatory
cytokines suppress RXR and may be involved in the loss of
 27:  Basolateral Plasma Membrane Organic Anion Transporters 329

NTCP BA

LRH-1 RXR FXR

↑SHP

RAR RXR
↓NTCP

HNF1α

HNF4α BA

Figure 27.1  Transcriptional regulation of NTCP by nuclear receptors. Bile acids (BA) stimulate expression of SHP by activating FXR response
elements in the SHP promoters. SHP, in turn, interferes with RXR : RAR transactivation of NTCP promoter, resulting in decreased expression of
NTCP. HNF1α transcriptionally activates NTCP. BA decreases NTCP expression also by suppressing HNF1α via its inhibitory effect on HNF4α.
SHP also inhibits HNF4α‐mediated transactivation of the HNF1α promoter and autoregulates its own promoter by repressing LRH‐1 (liver receptor
homologue).

RXR, RARα, and SHP in bile duct ligated rats [43]. The down­
regulation of NTCP in pregnancy and by endotoxin may also be
due to suppression of HNF1α and RXR : RAR [44, 45]. NTCP
expression is stimulated by the glucocorticoid receptor (GR),
the vitamin D receptor (VDR), the peroxisome proliferator‐acti­
vated receptor‐a (PPARa), signal transducer and activator of
transcription 5, as well as various transcription factors and hor­
mones [42, 46].
Post‐translational regulation
Plasma membrane transporters, after translation, need to be
translocated to the plasma membrane to carry out their intended
functions. The level of a transporter in the plasma membrane
may be increased or decreased based on the need to regulate
solute transport activity. This is a highly regulated post‐transla­
tional event requiring involvement of various cellular signaling
pathways. The post‐translational event also involves protein
quality control to select and target dysfunctional proteins for
degradation by the ubiquitin–proteasome system [47]. NTCP is
degraded by the ubiquitin–proteasome system at the level of
ER‐associated degradation. It is speculated that dysfunction of
this pathway may result in intracellular NTCP deposits in chole­
static livers [48].
Bile salt uptake by NTCP undergoes dynamic regulation.
Cyclic AMP (cAMP) stimulates TC uptake [49–52] and increases
the plasma membrane content of NTCP in rat hepatocytes and in
hepatic cells stably transfected with hNTCP [52–54]. In contrast,
cholestatic bile salts (taurochenodeoxycholate [TCDC] and
­taurolithocholate [TLC]) inhibit TC uptake [40, 55] and decrease
plasma membrane rNTCP [40, 56]. The increase in uptake by
cAMP, increased by glucagon, may allow hepatocytes to
­efficiently absorb bile salts returning from the intestine under
physiological conditions, while the decrease in uptake may
­protect hepatocytes from further increases in BAi observed in
cholestasis [57]. Functionally important polymorphisms in
NTCP may result in decreased plasma membrane expression.
Multiple single nucleotide polymorphisms in NTCP are present
in populations of European, African, Chinese, and Hispanic
Americans, and the altered transport activity of Ile223 to Thr
variant seen only in African Americans was due at least in part
to decreased plasma membrane expression [58]. The resulting
phenotype in people with these mutations, however, is not
known.
Studies to date suggest that the post‐translational regulation
of NTCP involves plasma membrane translocation/retrieval.
Several signaling pathways, including cAMP, intracellular Ca2+,
nitric oxide, phosphoinositide‐3‐kinase (PI3K), PKCs, and pro­
tein phosphatases are implicated in this process. Some of these
pathways have previously been reviewed [21, 59–61] and are
briefly described below.
The PI3K/Akt pathway plays an important role in cell ­survival
and translocation of hepatocellular transporters to the plasma
membrane, indicating a beneficial role of PI3K in hepatic cells
[60, 62–64]. PI3Ks are a family of lipid kinases (classes I, II,
and III) that phosphorylate the inositol ring of phosphatidylin­
ositides (PIs) at 3 position known as D3 phosphorylation [65].
The resulting phosphorylated PIs (PIPs), acting in concert
with phosphoinositide‐dependent kinases (PDKs), are involved
in the activation of downstream kinases, such as PKCζ/λ, Akt/
PKB, and p70S6K [66]. The PI3K/Akt pathway is involved in
cell swelling and cAMP‐induced increases in TC uptake and
330 THE LIVER:  SODIUM‐TAUROCHOLATE COTRANSPORTING POLYPEPTIDE (NTCP)

↑Ca2+

↑PI3K
↑Ca2+/CAM

↑PKCζ ↑PKCδ
↑PP2B

NTCP NTCP NTCP pNTCP

Rab4-GTP Rab4-GDP
Kinesin1

Mt

Figure 27.2  Signaling pathways regulating NTCP translocation to the basolateral plasma membrane. Insertion into the plasma membrane is medi­
ated via two major pathways; increase in intracellular Ca2+ and activation of PI3K. Increases in intracellular Ca2+ leads to the activation of Ca2+‐
dependent protein phosphatase 2B (PP2B) via Ca2+‐calmodulin kinase (Ca2+‐CAM). PP2B dephosphorylates NTCP (pNTCP to NTCP) allowing
translocation to the plasma membrane by vesicular trafficking. Activation of PI3K leads to activations of PKCδ and PKCζ, which in turn induce
translocation of NTCP from the intracellular compartment to the plasma membrane by stimulating vesicular trafficking. NTCP containing vesicles
colocalize with Rab4, which cycles between GTP bound (Rab4‐GTP) active form and GDP bound (Rab4‐GDP) inactive form. Activation of PKCδ
leads to the conversion of inactive Rab4‐GDP to active Rab4‐GTP, which moves NTCP containing vesicles toward the plus end of microtubules
(Mt) by interacting with kinesin‐1. This movement is further facilitated by PKCζ. On the other hand, activation of PKCα by a cholestatic bile acid,
taurochenodeoxycholate (TCDC), stimulates retrieval of NTCP from the plasma membrane by an unknown mechanism except that inhibition of
PI3K enhances the effect of TCDC (not shown).

rNTCP translocation to the plasma membrane [67, 68] and cAMP
inhibits TLC‐induced inhibition of TC uptake by activating the
PI3K pathway [56].
NTCP translocation to the plasma membrane is regulated by
phosphorylation and dephosphorylation, although kinase(s)
involved in phosphorylation of NTCP has yet to be identified.
NTCP is a serine/threonine phosphorylated protein [69] and
cAMP‐induced increases in TC uptake and plasma membrane
rNTCP is associated with dephosphorylation [70] at Ser226 [71].
It is proposed (Figure 27.2) that cAMP activates protein phos­
phatase 2B (PP2B) by increasing intracellular Ca2+ [50, 72] and
PP2B in turn dephosphorylates rNTCP facilitating its transloca­
tion to the plasma membrane (Figure 27.2). Interestingly, inhi­
bition of PP2B reverses TCDC‐induced retrieval of rNTCP [40].
Whether TCDC‐induced retrieval of rNTCP involves PP2B‐
mediated NTCP dephosphorylation remains to be established.
Protein kinase C isoforms differentially regulate NTCP trans­
location to and retrieval from the plasma membrane (Figure 27.2).
Thus, PKCδ [51, 73] and PKCζ [74] mediate cAMP‐induced
rNTCP translocation, while conventional PKCs (most likely
PKCα) mediate rNTCP retrieval [39, 40]. The activation of PKCδ
and PKCζ by cAMP, but not TCDC‐induced activation of PKCα
[40], is PI3K‐dependent [51, 74, 75]. On the other hand, TLC
induces retrieval of rNTCP [56], but inhibits PKCα [76]. Thus,
the retrieval of NTCP initiated by cholestatic bile acids may
involve mediators in addition to PKCα.
Limited studies suggest a role for the small GTPase Rab4
and microtubules in NTCP translocation by PKCs. Rab4 cycles
between GTP bound (Rab4‐GTP) active form and GDP bound
(Rab4‐GDP) inactive form [77]. PKCδ facilitates hNTCP
translocation by activating Rab4 [51] and PKCζ increases
motility of rNTCP‐containing vesicles along the microtubules
[78]. Other studies showed that Rab4 facilitates cAMP‐induced
NTCP translocation [52], NTCP‐containing vesicles colocalize
with Rab4 [78], cAMP‐induced rNTCP translocation is
dependent on PI3K [79], actin cytoskeleton and microtubules
[53, 79], and activation of PKCδ and PKCζ is PI3K dependent
[80]. Taken together, these results may suggest that activa­
tion of PI3K/PKCδ by cAMP leads to activation of Rab4 in
NTCP‐containing Rab4 vesicles and this is followed by
PI3K/PKCζ‐mediated increased motility of NTCP/Rab4‐
containing vesicles along the microtubules to the plasma
membrane (Figure 27.2).
S‐nitrosylation also affects plasma membrane localization of
NTCP. Treatment with thiol‐binding agents inhibits TC uptake
in isolated rat hepatocytes [81, 82] and by hNTCP [83] indicat­
ing a role for cysteine residues in transport function. Nitric
oxide (NO) inhibits TC uptake in rat hepatocytes [84], HuH7
cells stably transfected with hNTCP [85], and human hepato­
cytes [86]. One of the ways that NO exerts its cellular effects is
through a post‐translational modification termed S‐nitrosyla­
tion, which involves binding of NO to reactive thiol groups on
cysteine residues [87, 88]. Indeed, inhibition of TC uptake by
NO involves S‐nitrosylation of NTCP‐Cys96 and this is associ­
ated with a decrease in plasma membrane hNTCP [85, 89].
Such a mechanism may be involved in LPS‐induced cholestasis,
which is associated with a burst of NO production by iNOS in
liver cells [90]. NO also produces its cellular effects by nitration
of tyrosine residues [91] and increases tyrosine nitration of
NTCP [86]. Two tyrosine residues on the cytoplasmic tail of rat
NTCP appear to be involved in targeting to the basolateral mem­
brane [92]. Thus, tyrosine nitration of NTCP by NO may also
 27:  Basolateral Plasma Membrane Organic Anion Transporters 331
result in decreased translocation to the basolateral membrane
resulting in decreased TC uptake.
There are species differences in the regulation of plasma
membrane localization of hNTCP and rNTCP. TLC inhibits TC
uptake by hNTCP and rNTCP. However, TLC decreases plasma
membrane rNTCP but not hNTCP [56]. Like Bosentan [93],
TLC inhibits TC uptake by rNTCP and hNTCP non‐competi­
tively [55] and competitively [56], respectively. NO inhibits TC
uptake by hNTCP/rNTCP. However, NO decreases plasma
membrane level of hNTCP [85] and does not affect the plasma
membrane level of rNTCP [89]. In contrast, cAMP‐induced
increases in TC uptake are associated with translocation of
hNTCP as well as rNTCP. These differences should be taken into
account when extrapolating mechanisms regulating rNTCP to
hNTCP for the purpose of human drug development and also for
understanding the mechanistic relevance in human diseases.
ORGANIC ANION TRANSPORT PROTEINS
(OATPs)
Similar to hepatic uptake of bile acids, uptake of organic ani­
ons such as bilirubin and bromosulfophthalein (BSP) is rapid
and has carrier‐mediated kinetics [94]. Studies performed in
overnight cultured rat hepatocytes revealed saturable uptake
of BSP that was inhibited by bilirubin [95, 96]. Ligand was
taken up by cells, while albumin remained extracellular [96].
Of interest was the finding that isosmotic substitution of NaCl
in medium by sucrose resulted in an over 80% reduction of
BSP uptake [96]. This was not due to a requirement for extra­
cellular Na+, as substitution of Na+ by K+, Li+, or choline had
no effect. However, replacement of Cl− by HCO3‐ or gluconate
reduced BSP uptake by approximately 40% [96]. Similar
results were found for bilirubin uptake by cultured rat hepato­
cytes as well as by isolated perfused rat liver. The basis for
this chloride‐dependency is not clear. Studies performed with
36
Cl revealed that BSP uptake requires the presence of exter­
nal Cl− and is not stimulated by unidirectional Cl− gradients,
suggesting that BSP transport is not coupled to Cl− transport
[95]. These studies supported the existence of a hepatocyte
surface membrane organic anion transporter. Later studies
used the transport assay developed in cultured rat hepatocytes
as the basis for an expression cloning strategy in Xenopus lae-
vis oocytes [97]. Injection of oocytes with total rat liver poly
(A)+ RNA resulted in the functional expression of chloride‐
dependent BSP extraction from albumin. Using a subselection
cloning protocol, a single cDNA was isolated [98]. The
encoded protein was named organic anion transporting poly­
peptide 1 (oatp1), now known as oatp1a1. This protein medi­
ates uptake of BSP and also mediates transport of various bile
salts such as taurocholate in a Na+‐independent fashion. Since
the initial description of oatp1a1, over 20 additional members
of the oatp family have been described [99]. Amino acid
sequences of these proteins have a high degree of homology.
In hepatocytes oatps are distributed on the basolateral
­(sinusoidal) plasma membrane. They are of similar size and
have similar predicted membrane topologies and biochemical
characteristics.
Transport functions of OATPs
The oatps have broad and overlapping substrate specificities
[11, 100]. Macrolide antibiotics, antihistamines, and statins are
among the many drugs that are used clinically as well as various
xenobiotics that can be transported by members of the oatp fam­
ily. As oatps have ligand specificities that may overlap between
family members as well as other transporters (e.g. NTCP), stud­
ies to determine transport activity of a single member of the oatp
family performed in vitro may be difficult to extrapolate to what
actually occurs in vivo in a living organism. Studies performed
in specific oatp knockout mouse models have provided some
insight into their physiologic function, although the resulting
phenotypes are often subtle. For example, knockout of oatp1b2
resulted in reduced clearance of rifampicin, and lovastatin, but
not of pravastatin, simvastatin, rifamycin SV, or cerivastatin
[101]. Interestingly, reduced hepatic clearance of phalloidin and
microcystin resulted in reduced toxicity of these compounds in
oatp1b2 knockout mice, confirming that they are ligands for this
transporter [102]. Of note is the finding that there was a modest
increase in conjugated bilirubin in the circulation of these mice
and reduced clearance of dibromosulfophthalein (DBSP) [102,
103]. Reduced liver to plasma ratio of pravastatin was also noted
following continuous infusion to reach steady state, again show­
ing that changes in drug clearance can be subtle, requiring spe­
cial studies to demonstrate them [104]. Studies in mice in which
Oatp1a1 or Oatp1a4 were knocked out showed reduced uptake
of estradiol‐17beta‐D‐glucuronide, estrone‐3‐sulfate, and tauro­
cholic acid in primary hepatocytes prepared from both mouse
strains [104].
Mice, in which all members of the Oatp1a and Oatp1b were
simultaneously deleted without organ specificity had markedly
delayed plasma clearance of methotrexate and fexofenadine
[13]. Additional studies showed that doxorubicin clearance was
also reduced by 40% in these mice [105]. Interestingly, total
bile acid levels were increased in these mice and this was due
to increased plasma levels of unconjugated but not conjugated
bile acids [13]. However, another study in which oatp1a1 or
oatp1a4 were knocked out individually in mice revealed
reduced uptake of the conjugated bile acid taurocholic acid
[104]. Still other studies looking at transport mediated by rat
oatp1a1 showed that it was an avid transporter of several con­
jugated bile acids, including tauromuricholic and taurocholic
acids [106], the major bile acids found in mice [107]. The rea­
son for these seeming inconsistencies with respect to oatp‐
mediated bile acid transport remains unknown. The Oatp1a/
Oatp1b double knockout mice also had elevated levels of total
bilirubin that was due to increased conjugated but not unconju­
gated bilirubin [13]. Total bilirubin levels in single knockout
mice for Oatp1a1 or Oatp1a4 were normal [104], although as
noted above, single knockout of Oatp1b2 resulted in a modest
elevation of conjugated bilirubin in female mice [102]. Elevated
conjugated bilirubin levels in the double knockout mice was
phenotypically similar to the conjugated hyperbilirubinemia
that has been described in patients with Rotor syndrome, a rare
disorder characterized by elevated plasma conjugated bilirubin
with otherwise good health and normal routine liver function
tests [108, 109]. These individuals have now been found to have
simultaneous null mutations in the genes encoding OATP1B1
332 THE LIVER:  ORGANIC ANION TRANSPORT PROTEINS (OATPs)

and OATP1B3 [110]. It has been hypothesized that bilirubin
glucuronides are effluxed from hepatocytes back into the circu­
lation where they subsequently undergo reuptake mediated by
OATP1B1 or OATP1B3 [110]. In the absence of either one of
these transporters, the other is able to mediate conjugated
­bilirubin uptake. In the absence of both of these transporters,
levels of conjugated bilirubin in the circulation rise as is seen in
Rotor syndrome patients and the corresponding knockout mouse
model. Of note is the fact that the transporter(s) mediating
uptake of unconjugated bilirubin has not as yet been identified
[111, 112].
Regulation of non‐bile acid organic anion
transport
There are a number of physiologic states in which organic anion
transport may be perturbed. Many of these studies have been
performed in animal models. These include liver regeneration in
rats resulting from two‐thirds partial hepatectomy in which
influx of bilirubin and BSP, when determined independently of
reduced liver size, fall by approximately 50% within six hours
and normalize by four days [113, 114], correlating with changes
in oatp1a1 mRNA and protein levels [115]. Similar modulation
of transport and transporters is seen during development. For
example, BSP uptake is 70% lower in hepatocytes prepared
from three‐week‐old rats as compared to adults [116], consist­
ent with finding of reduced oatp1a1 protein and mRNA expres­
sion in liver for the first month after birth [116, 117]. In other
studies performed in rat hepatocytes, uptake of BSP was found
to be reduced in the presence of extracellular ATP [118].
Specifically, within minutes of exposure of cells to ATP, BSP
uptake is reduced by 80% [118, 119]. Pharmacokinetic analysis
revealed reduced Vmax and unchanged Km and this effect was
found to be mediated by a purinergic receptor, resulting in phos­
phorylation of oatp1a1 [119, 120]. Subsequent studies in trans­
fected cells and primary rat hepatocytes suggest that
phosphorylation of oatp1a1 results in its internalization and
consequent loss of transport activity [121].
Pharmacogenomic studies
A number of pharmacogenomic studies have identified single
nucleotide polymorphisms (SNPs) of OATPs that correlate with
episodes of drug toxicity or altered liver uptake [122–124].
Included among these are studies that implicate OATPs in clear­
ance of statins from the circulation. The importance of OATP
polymorphisms on statin pharmacokinetics was epitomized in a
study in which 85 patients out of 12 000 patients on simvastatin
developed myopathy [125]. A genome‐wide association study
of these individuals revealed a strong association with a poly­
morphism in the gene encoding OATP1B1. Previous studies
showed that this valine to alanine polymorphism (Val174Ala)
encodes a protein that traffics poorly to the plasma membrane
and accumulates within cells [126]. It is a reasonable presump­
tion that in these patients, reduced hepatocyte plasma membrane
levels of OATP1B1 resulted in reduced uptake of simvastatin
leading to elevated serum levels and a consequent increased
uptake by muscle cells causing the toxic effects. These data
point out the potential importance of transporter subcellular
trafficking on its transport activity. Importantly, other factors
that lead to reduced transporter on the hepatocyte surface could
potentially result in toxicity even in the absence of a mutation in
the transporter itself.
Regulation of oatp subcellular distribution
Seven of the eleven oatps that have been described in rat,
mouse, or human liver have PDZ consensus binding sites, as
defined by the sequences of their four C‐terminal amino acids
(Table 27.1) [127]. These binding sites can mediate interaction
of these oatps with one or more of the over 150 known PDZ
domain‐containing proteins [128]. Interestingly, each of these
PDZ domain‐containing proteins can have multiple binding
domains and can form functionally important complexes with
their protein ligands [128]. The presence of a potential PDZ
consensus binding site on a protein, such as a member of the
oatp family, does not mean that this protein necessarily binds
to a PDZ domain‐containing protein, nor does it help in iden­
tifying the potential binding protein. In particular the C‐termi­
nal four amino acids of rat oatp1a1, KTKL, form a potential
PDZ consensus binding sequence  [121, 127, 129]. Using a
synthetic peptide corresponding to the C‐terminal 16 amino
acids of oatp1a1 for affinity isolation, interacting proteins
from rat liver cytosol were purified. Protein mass fingerprint­
ing identified PDZK1 as the major interacting protein [127].
Using an otherwise identical synthetic peptide, but lacking the
C‐terminal KTKL that comprises the PDZ consensus binding
motif, no interacting proteins were resolved [127]. That
PDZK1 and oatp1a1 are bound to each other in vivo was shown

Table 27.1  Oatp family members found in rat, mouse, and human liver
NCBI accession C‐terminal Potential PDZ Plasma membrane
Species Original name New name number sequence consensus localization
Rat oatp1 oatp1a1 NM_017111 KTKL Yes Yes
Rat oatp2 oatp1a4 NM_131906 VTED No Yes
Rat oatp4 oatp1b2 NM_031650 ETPL Yes Yes
Rat oatp9 Oatp2b1 NM_080786 LQEL No ND
Mouse Oatp1 Oatp1a1 NM_013797 KTKL Yes Yes
Mouse Oatp2 Oatp1a4 NM_030687 KTKL Yes ND
Mouse Oatp4 Oatp1b2 NM_020495 ETPL Yes Yes
Human OATP‐A OATP1A2 NM_021094 KTKL Yes Yes
Human OATP‐C OATP1B1 NM_006446 ETHC No Yes
Human OATP8 OATP1B3 NM_019844 AAAN No Yes
Human OATP‐B OATP2B1 NM_007256 DSRV Yes Yes
 27:  Basolateral Plasma Membrane Organic Anion Transporters 333

Figure 27.3  Working model for trafficking of endocytic vesicle‐associated oatp1a1 on microtubules. The C‐terminal 4 amino acids of oatp1a1
(KTKL) represent a PDZ binding consensus sequence that binds to PDZK1. PDZK1 has 4 independent PDZ binding sites and oatp1a1 binds to sites
1 and 3 as indicated, leaving sites 2 and 4 available to bind other proteins [127]. Interestingly, PDZK1 itself has a PDZ consensus binding motif at
its C‐terminus that can bind internally to PDZ domain 1 as well as binding other scaffold proteins such as EBP50 [132]. In the presence of PDZK1
endocytic vesicles containing oatp1a1 can interact with microtubules and move toward their plus end (cell surface) via activity of the motor mole­
cule kinesin‐1. Endocytic vesicles in the absence of PDZK1 or those that contain constructs of oatp1a1 lacking the terminal 4 amino acids and can
no longer bind PDZK1 associate with the molecular motor dynein that moves towards the minus end of microtubules (cell interior). Endocytic vesi­
cles containing oatp1a1 can also be associated with other cargo and we hypothesize that this can include oatps that lack a PDZ consensus binding
site and co‐traffic to and from the plasma membrane.

by their coimmunoprecipitation from rat liver lysates. Mouse
oatp1a1 has an 82% amino acid identity to rat oatp1a1 includ­
ing the same KTKL amino acid sequence at its C‐terminus.
Immunofluorescence studies performed in wild‐type and
PDZK1 knockout mice revealed that plasma membrane distri­
bution of oatp1a1 was markedly reduced in the knockout mice
[127]. Total expression of oatp1a1 in PDZK1 knockout mice
was not reduced, as determined by Western blot, but it was
located predominantly in intracellular vesicles on immunoflu­
orescence analysis [127]. Corresponding to reduced plasma
membrane expression of oatp1a1 in the PDZK1 knockout
mouse, the rate of plasma disappearance of 35S‐sulfobromoph­
thalein (BSP), a ligand transported by oatp1a1, was reduced
by approximately 25% [127].
To validate that interaction of oatp1a1 with PDZK1 is
required for its optimal trafficking to the plasma membrane,
studies were performed in HEK 293T cells transiently trans­
fected with GFP‐oatp1a1 with or without co‐transfection with
PDZK1 [121]. These studies showed that oatp1a1 was present
on the plasma membrane only in the presence of PDZK1 expres­
sion. Transfection with an oatp1a1 construct that did not code
for the terminal four amino acids revealed that it was no longer
found on the cell surface even in the presence of PDZK1 [121].
Trafficking to the cell surface was also reduced with a phospho­
mimetic oatp1a1 construct, in which glutamic acid replaced the
serines at positions 634 and 635, irrespective of the presence of
PDZK1 [121]. This was related to increased internalization of
phosphomimetic oatp1a1 from the cell surface [121]. As
endocytic vesicles have been shown to traffic through cells on
­microtubules [41, 130], studies were performed to examine the
hypothesis that subcellular trafficking of oatp1a1 requires
movement of oatp1a1‐containing vesicles on microtubules
(Figure  27.3) [129]. Previous studies indicated that microtu­
bules in hepatocytes have their plus, rapidly growing ends near
the cell surface, and minus ends within the cell at a microtubule
organizing center [131].
Immunofluorescence studies performed in microchambers
showed the existence of an oatp1a1‐associated population of
endocytic vesicles prepared from wild‐type mouse liver [129].
These vesicles were also largely associated with PDZK1, bound
to polymerized microtubules in vitro, and moved approximately
equally toward the plus and minus microtubule ends upon addi­
tion of 50 μM ATP. A population of vesicles prepared from liv­
ers of PDZK1 knockout mice was also associated with oatp1a1.
These vesicles bound to microtubules in vitro but upon ATP
addition, their movement was highly biased toward the minus
end, consistent with movement of vesicles to the cell interior.
Further analysis showed that oatp‐associated vesicles were also
associated with microtubule‐based motors. The identity of
motors that were associated with these vesicles differed between
vesicles prepared from wild‐type vs PDZK1 knockout mice.
Kinesin‐1, a microtubule‐based plus end motor was highly asso­
ciated with wild‐type vesicles, with little associated with
­vesicles prepared from PDZK1 knockout mice. Dynein, a
microtubule‐based minus‐end motor, was largely associated
with PDZK1 knockout vesicles [129]. These studies suggest
that PDZK1 on oatp1a1‐containing vesicles results in recruit­
ment of microtubule‐based motors that are not associated with
these vesicles in its absence (Figure  27.3). Whether, in the
absence of PDZK1, oatp1a1 itself or other vesicle associated
proteins can recruit motors such as dynein is the subject of
ongoing studies.
334 THE LIVER:  REFERENCES

REFERENCES 25. Kramer, W. Transporters, Trojan horses and therapeutics: suitability of bile
acid and peptide transporters for drug delivery. Biol Chem, 2011;392:77–94.
26. Giacomini, K.M., Huang, S.M., Tweedie, D.J. et al. Membrane transporters
1. Anwer, M.S. and Hegner, D. Effect of Na+ on bile acid uptake by isolated rat
in drug development. Nat Rev Drug Discov, 2010;9:215–36.
hepatocytes. Hoppe‐Zeyler’s Z Physiol Chem, 1978;359:181–92.
27. Hebert, M.F., Townsend, R.W., Austin, S. et al. Concomitant cyclosporine
2. Hagenbuch, B., Lubbert, H., Stieger, B., and Meier, P.J. Expression of the
and micafungin pharmacokinetics in healthy volunteers. J Clin Pharmacol,
hepatocyte Na+/bile acid cotransporter in Xenopus laevis oocytes. J Biol
2005;45:954–60.
Chem, 1990;265:5357–60.
28. Yan, H., Zhong, G., Xu, G. et al. Sodium taurocholate cotransporting poly­
3. Hagenbuch, B., Stieger, B., Foguet, M. et al. Functional expression cloning
peptide is a functional receptor for human hepatitis B and D virus. Elife,
and characterization of the hepatocyte Na+/bile acid cotransport system. Proc
2012;1:1–28.
Natl Acad Sci U S A, 1991;88:10629–33.
29. Yan, H., Peng, B., He, W. et al. Molecular determinants of hepatitis B and d
4. Claro da, S.T., Polli, J.E., and Swaan, P.W. The solute carrier family 10
virus entry restriction in mouse sodium taurocholate cotransporting polypep­
(SLC10): beyond bile acid transport. Mol Aspects Med, 2013;34:252–69.
tide. J Virol, 2013;87:7977–91.
5. Doring, B., Lutteke, T., Geyer, J., and Petzinger, E. The SLC10 carrier fam­
30. Glebe, D. and Urban, S. Viral and cellular determinants involved in hepadna­
ily: transport functions and molecular structure. Curr Top Membr,
viral entry. World J Gastroenterol, 2007;13:22–38.
2012;70:105–68.
31. Gripon, P., Cannie, I., and Urban, S. Efficient inhibition of hepatitis B virus
6. Hagenbuch, B. and Dawson, P. The sodium bile salt cotransport family
infection by acylated peptides derived from the large viral surface protein.
SLC10. Pflugers Arch, 2004;447:566–70.
J Virol, 2005;79:1613–22.
7. Weinman, S.A. Electrogenicity of Na(+)‐coupled bile acid transporters. Yale
32. Meier, A., Mehrle, S., Weiss, T.S. et al. Myristoylated PreS1‐domain of the
J Biol Med, 1997;70:331–40.
hepatitis B virus L‐protein mediates specific binding to differentiated hepat­
8. Meier, P.J. and Stieger, B. Bile salt transporters. Annu Rev Physiol,
ocytes. Hepatology, 2013;58:31–42.
2002;64:635–61.
33. Petersen, J., Dandri, M., Mier, W. et al. Prevention of hepatitis B virus infec­
9. Stieger, B. The role of the sodium‐taurocholate cotransporting polypeptide
tion in vivo by entry inhibitors derived from the large envelope protein. Nat
(NTCP) and of the bile salt export pump (BSEP) in physiology and patho­
Biotechnol, 2008;26:335–41.
physiology of bile formation. Handb Exp Pharmacol, 2011;201:205–59.
34. Bogomolov, P., Alexandrov, A., Voronkova, N. et al. Treatment of chronic
10. Qiu, J.W., Deng, M., Cheng, Y. et  al. Sodium taurocholate cotransporting
hepatitis D with the entry inhibitor myrcludex B: first results of a phase Ib/
polypeptide (NTCP) deficiency: identification of a novel SLC10A1 mutation
IIa study. J Hepatol, 2016;65:490–8.
in two unrelated infants presenting with neonatal indirect hyperbilirubinemia
35. Xiao, F., McKeating, J.A., and Baumert, T.F. A bile acid transporter as a
and remarkable hypercholanemia. Oncotarget, 2017;8:106598–607.
candidate receptor for hepatitis B and D virus entry. J Hepatol,
11. Hagenbuch, B. and Stieger, B. The SLCO (former SLC21) superfamily of
2013;58:1246–8.
transporters. Mol Aspects Med, 2013;34:396–412.
36. Warner, N. and Locarnini, S. The new front‐line in hepatitis B/D research:
12. Csanaky, I.L., Lu, H., Zhang, Y. et al. Organic anion‐transporting polypep­
Identification and blocking of a functional receptor. Hepatology, 2013;58:9–12.
tide 1b2 (Oatp1b2) is important for the hepatic uptake of unconjugated bile
37. Schieck, A., Schulze, A., Gahler, C. et al. Hepatitis B virus hepatotropism is
acids: studies in Oatp1b2‐null mice. Hepatology, 2011;53:272–81.
mediated by specific receptor recognition in the liver and not restricted to
13. van de Steeg, E., Wagenaar, E., van der Kruijssen, C.M. et al. Organic anion
susceptible hosts. Hepatology, 2013;58:43–53.
transporting polypeptide 1a/1b‐knockout mice provide insights into hepatic
38. Lempp, F.A., Wiedtke, E., Qu, B. et al. Sodium taurocholate cotransporting
handling of bilirubin, bile acids, and drugs. J Clin Invest, 2010;120:2942–52.
polypeptide is the limiting host factor of hepatitis B virus infection in
14. Faber, K.N., Muller, M., and Jansen, P.L. Drug transport proteins in the liver.
macaque and pig hepatocytes. Hepatology, 2017;66:703–16.
Adv Drug Deliv Rev, 2003;55:107–24.
39. Stross, C., Helmer, A., Weissenberger, K. et  al. Protein kinase C induces
15. Yanni, S.B., Augustijns, P.F., Benjamin, D.K., Jr. et al. In vitro investigation
endocytosis of the sodium taurocholate cotransporting polypeptide. Am J
of the hepatobiliary disposition mechanisms of the antifungal agent
Physiol Gastrointest Liver Physiol, 2010;299:G320–8.
micafungin in humans and rats. Drug Metab Dispos, 2010;38:1848–56.
40. Muhlfeld, S., Domanova, O., Berlage, T. et al. Short‐term feedback regula­
16. Ho, R.H., Tirona, R.G., Leake, B.F. et al. Drug and bile acid transporters in
tion of bile salt uptake by bile salts in rodent liver. Hepatology,
rosuvastatin hepatic uptake: function, expression, and pharmacogenetics.
2012;56:2387–97.
Gastroenterology, 2006;130:1793–806.
41. Wang, X., Wang, P., Wang, W. et al. The Na(+)‐taurocholate cotransporting
17. Kullak‐Ublick, G.A., Glasa, J., Boker, C. et al. Chlorambucil‐taurocholate is
polypeptide traffics with the epidermal growth factor receptor. Traffic,
transported by bile acid carriers expressed in human hepatocellular carcino­
2016;17:230–44.
mas. Gastroenterology, 1997;113:1295–305.
42. Claudel, T., Zollner, G., Wagner, M., and Trauner, M. Role of nuclear recep­
18. Abel, J.J. and Rowntree, L.G. On the pharmacological action of some phtha­
tors for bile acid metabolism, bile secretion, cholestasis, and gallstone dis­
leins and their derivatives, with especial reference to their behavior as purga­
ease. Biochim Biophys Acta, 2011;1812:867–78.
tives. J Pharmacol Exp Ther, 1909;1:231–64.
43. Trauner, M. and Boyer, J.L. Bile salt transporters: molecular characteriza­
19. Hoekstra, L.T., de Graaf, W., Nibourg, G.A. et  al. Physiological and bio­
tion, function, and regulation. Physiol Rev, 2003;83:633–71.
chemical basis of clinical liver function tests: a review. Ann Surg,
44. Arrese, M., Trauner, M., Ananthanarayanan, M. et  al. Down‐regulation of
2013;257:27–36.
the Na+/taurocholate cotransporting polypeptide during pregnancy in the rat.
20. Stieger, B., Heger, M., de Graaf, W. et  al. The emerging role of transport
J Hepatol, 2003;38:148–55.
systems in liver function tests. Eur J Pharmacol, 2012;675:1–5.
45. Trauner, M., Arrese, M., Lee, H. et al. Endotoxin downregulates rat hepatic
21. Anwer, M.S. and Stieger, B. Sodium‐dependent bile salt transporters of the
ntcp gene expression via decreased activity of critical transcription factors.
SLC10A transporter family: more than solute transporters. Pflugers Arch,
J Clin Invest, 1998;101:2092–100.
2013;466:77–89.
46. Dawson, P.A., Lan, T., and Rao, A. Bile acid transporters. J Lipid Res,
22. Kouzuki, H., Suzuki, H., Stieger, B. et al. Characterization of the transport
2009;50:2340–57.
properties of organic anion transporting polypeptide 1 (oatp1) and Na(+)/
47. Claessen, J.H., Kundrat, L., and Ploegh, H.L. Protein quality control in the
taurocholate cotransporting polypeptide (Ntcp): comparative studies on the
ER: balancing the ubiquitin checkbook. Trends Cell Biol, 2012;22:22–32.
inhibitory effect of their possible substrates in hepatocytes and cDNA‐trans­
48. Kuhlkamp, T., Keitel, V., Helmer, A. et al. Degradation of the sodium tauro­
fected COS‐7 cells. J Pharmacol Exp Ther, 2000;292:505–11.
cholate cotransporting polypeptide (NTCP) by the ubiquitin–proteasome
23. Mita, S., Suzuki, H., Akita, H. et al. Inhibition of bile acid transport across
system. Biol Chem, 2005;386:1065–74.
Na+/taurocholate cotransporting polypeptide (SLC10A1) and bile salt export
49. Botham, K.M. and Suckling, K.E. The effect of dibutyryl cyclic AMP on the
pump (ABCB 11)‐coexpressing LLC‐PK1 cells by cholestasis‐inducing drugs.
uptake of taurocholic acid by isolated rat liver cells. Biochim Biophys Acta,
Drug Metab Dispos, 2006;34:1575–81.
1986;883:26–32.
24. McRae, M.P., Lowe, C.M., Tian, X. et  al. Ritonavir, saquinavir, and efa­
50. Grüne, S., Engelking, L.R., and Anwer, M.S. Role of intracellular calcium
virenz, but not nevirapine, inhibit bile acid transport in human and rat hepat­
and protein kinases in the activation of hepatic Na+/taurocholate cotransport
ocytes. J Pharmacol Exp Ther, 2006;318:1068–75.
by cyclic AMP. J Biol Chem, 1993;268:17734–41.
 27:  Basolateral Plasma Membrane Organic Anion Transporters 335
51. Park, S.W., Schonhoff, C.M., Webster, C.R., and Anwer, M.S. Protein kinase
Cdelta differentially regulates cAMP‐dependent translocation of NTCP and
MRP2 to the plasma membrane. Am J Physiol Gastrointest Liver Physiol,
2012;303:G657–65.
52. Schonhoff, C.M., Thankey, K., Webster, C.R. et  al. Rab4 facilitates cyclic
adenosine monophosphate‐stimulated bile acid uptake and Na(+)‐taurocholate
cotransporting polypeptide translocation. Hepatology, 2008;48:1665–70.
53. Dranoff, J.A., McClure, M., Burgstahler, A.D. et al. Short‐term regulation of
bile acid uptake by microfilament‐dependent translocation of ntcp to the
plasma membrane. Hepatology, 1999;30:223–9.
54. Mukhopadhayay, S., Ananthanarayanan, M., Stieger, B. et  al. cAMP
increases liver Na+‐taurocholate cotransport by translocating transporter to
plasma membranes. Am J Physiol, 1997;273, G842–G848.
55. Schwenk, M., Schwarz, L.R., and Greim, H. Taurolithocholate inhibits tau­
rocholate uptake by isolated hepatocytes at low concentrations. Naunyn
Schmiedebergs Arch Pharmacol, 1977;298:175–9.
56. Schonhoff, C.M., Yamazaki, A., Hohenester, S. et al. PKC(epsilon)‐depend­
ent and ‐independent effects of taurolithocholate on PI3K/PKB pathway and
taurocholate uptake in HuH‐NTCP cell line. Am J Physiol Gastrointest Liver
Physiol, 2009;297:G1259–67.
57. Prekeris, R., Klumperman, J., and Scheller, R.H. A Rab11/Rip11 protein
complex regulates apical membrane trafficking via recycling endosomes.
Mol Cell, 2000;6:1437–48.
58. Ho, R.H., Leake, B.F., Roberts, R.L. et al. Ethnicity‐dependent polymorphism in
Na+‐taurocholate cotransporting polypeptide (SLC10A1) reveals a domain criti­
cal for bile acid substrate recognition. J Biol Chem, 2004;279:7213–22.
59. Kosters, A. and Karpen, S.J. Bile acid transporters in health and disease.
Xenobiotica, 2008;38:1043–71.
60. Anwer, M.S. Cellular Regulation of hepatic bile acid transport in health and
cholestasis. Hepatology, 2004;39:581–9.
61. Bouscarel, B., Kroll, S.D., and Fromm, H. Signal transduction and hepato­
cellular bile acid transport: cross talk between bile acids and second mes­
sengers. Gastroenterology, 1999;117:433–52.
62. Rust, C., Bauchmuller, K., Fickert, P. et al. Phosphatidylinositol 3‐kinase‐
dependent signaling modulates taurochenodeoxycholic acid‐induced liver
injury and cholestasis in perfused rat livers. Am J Physiol Gastrointest Liver
Physiol, 2005;289:G88–94.
63. Webster, C.R., Usechak, P., and Anwer, M.S. cAMP inhibits bile acid‐
induced apoptosis by blocking caspase activation and cytochrome c release.
Am J Physiol Gastrointest Liver Physiol, 2002;283:G727–38.
64. Gates, A., Hohenester, S., Anwer, M.S., and Webster, C.R. cAMP‐GEF cyto­
protection by Src tyrosine kinase activation of phosphoinositide‐3‐kinase
p110 {beta}/{alpha} in rat hepatocytes. Am J Physiol Gastrointest Liver
Physiol, 2009;296:G764–74.
65. Cantley, L.C. he phosphoinositide 3‐kinase pathway. Science, 2002; 296:
1655–7.
66. Toker, A. Protein kinases as mediators of phosphoinositide 3‐kinase signal­
ing. Mol Pharmacol, 2000;57:652–8.
67. Webster, C.R., Srinivasulu, U., Ananthanarayanan, M. et al. rotein kinase B/
Akt mediates cAMP‐ and cell swelling‐stimulated Na+/taurocholate cotrans­
port and Ntcp translocation. J Biol Chem, 2002;277:28578–83.
68. Webster, C.R.L., Blanch, C.J., Philips, J., and Anwer, M.S. ell swelling‐
induced translocation of rat liver Na+/taurocholate cotransport polypeptide is
mediated via the phosphoinositide 3‐kinase signaling pathway. J Biol Chem,
2000;275:29754–60.
69. Mukhopadhayay, S., Ananthanarayanan, M., Stieger, B. et al. Sodium tauro­
cholate cotransporting polypeptide is a serine, threonine phosphoprotein and
is dephosphorylated by cyclic AMP. Hepatology, 1998;28:1629–36.
70. Mukhopadhayay, S., Webster, C.R.L., and Anwer, M.S. Role of protein
phosphatase in cyclic AMP‐mediated stimulation of hepatic Na+/taurocho­
late cotransport. J Biol Chem, 1998;273:30039–45.
71. Anwer, M.S., Gillin, H., Mukhopadhyay, S. et  al. Dephosphorylation of
Ser‐226 facilitates plasma membrane retention of Ntcp. J Biol Chem,
2005;280:33687–92.
72. Webster, C.R.L., Blanch, C., and Anwer, M.S. Role of PP2B in cAMP‐
induced dephosphorylation and translocation of NTCP. Am J Physiol
Gastrointest Liver Physiol, 2002;283:G44–50.
73. Schonhoff, C.M., Gillin, H., Webster, C.R., and Anwer, M.S. Protein kinase
Cdelta mediates cyclic adenosine monophosphate‐stimulated translocation
of sodium taurocholate cotransporting polypeptide and multidrug resistant
associated protein 2 in rat hepatocytes. Hepatology, 2008;47:1309–16.
74. McConkey, M., Gillin, H., Webster, C.R., and Anwer, M.S. Cross‐talk
between protein kinases Czeta and B in cyclic AMP‐mediated sodium tauro­
cholate co‐transporting polypeptide translocation in hepatocytes. J Biol
Chem, 2004;279:20882–8.
75. Jiang, Y., Chen, C., Li, Z. et al. Characterization of the structure and function
of a new mitogen‐activated protein kinase (p38beta). J Biol Chem,
1996;271:17920–6.
76. Beuers, U., Bilzer, M., Chittattu, A. et al. Tauroursodeoxycholic acid inserts
the apical conjugate export pump, Mrp2, into canalicular membranes and
stimulates organic anion secretion by protein kinase C‐dependent mecha­
nisms in cholestatic rat liver. Hepatology, 2001;33:1206–16.
77. Agola, J.O., Jim, P.A., Ward, H.H. et al. Rab GTPases as regulators of endo­
cytosis, targets of disease and therapeutic opportunities. Clin Genet,
2011;80:305–18.
78. Sarkar, S., Bananis, E., Nath, S. et al. PKCzeta is required for microtubule‐based
motility of vesicles containing the ntcp transporter. Traffic, 2006;7:1078–91.
79. Webster, C.R.L. and Anwer, M.S. Role of the PI3K/PKB signaling pathway
in cAMP‐mediated translocation of rat liver Ntcp. Am J Physiol,
1999;277:G1165–72.
80. Newton, A.C. Regulation of the ABC kinases by phosphorylation: protein
kinase C as a paradigm. Biochem J, 2003;370:361–71.
81. Blumrich, M. and Petzinger, E. Membrane transport of conjugated and
unconjugated bile acids into hepatocytes is susceptible to SH‐blocking rea­
gents. Biochim Biophys Acta, 1990;1029:1–12.
82. Blumrich, M. and Petzinger, E. Two distinct types of SH‐groups are neces­
sary for bumetanide and bile acid uptake into isolated rat hepatocytes.
Biochim Biophys Acta, 1993;1149:278–84.
83. Hallen, S., Fryklund, J., and Sachs, G. Inhibition of the human sodium/bile
acid cotransporters by side‐specific methanethiosulfonate sulfhydryl rea­
gents: substrate‐controlled accessibility of site of inactivation. Biochemistry,
2000;39:6743–50.
84. Song, I.S., Lee, I.K., Chung, S.J. et al. Effect of nitric oxide on the sinusoidal
uptake of organic cations and anions by isolated hepatocytes. Arch Pharm
Res, 2002;25:984–8.
85. Schonhoff, C.M., Ramasamy, U., and Anwer, M.S. Nitric oxide‐mediated
inhibition of taurocholate uptake involves S‐nitrosylation of NTCP. Am J
Physiol Gastrointest Liver Physiol, 2011;300:G364–70.
86. Gonzalez, R., Cruz, A., Ferrin, G. et al. Nitric oxide mimics transcriptional
and post‐translational regulation during alpha‐tocopherol cytoprotection
against glycochenodeoxycholate‐induced cell death in hepatocytes. J Hepatol,
2011;55:133–44.
87. Hess, D.T. and Stamler, J.S. Regulation by S‐nitrosylation of protein post­
translational modification. J Biol Chem, 2012;287(7):4411–8.
88. Schonhoff, C.M., Gaston, B., and Mannick, J.B. Nitrosylation of cytochrome
c during apoptosis. J Biol Chem, 2003;278:18265–70.
89. Ramasamy, U., Anwer, M.S., and Schonhoff, C.M. Cysteine 96 of Ntcp is
responsible for NO‐mediated inhibition of taurocholate uptake. Am J Physiol
Gastrointest Liver Physiol, 2013;305:G513–19.
90. Geier, A., Fickert, P., and Trauner, M. Mechanisms of disease: mechanisms
and clinical implications of cholestasis in sepsis. Nat Clin Pract Gastroenterol
Hepatol, 2006;3:574–85.
91. Ischiropoulos, H. Protein tyrosine nitration—an update. Arch Biochem
Biophys, 2009;484:117–21.
92. Sun, A.Q., Arrese, M.A., Zeng, L. et al. The rat liver Na(+)/bile acid cotrans­
porter. Importance of the cytoplasmic tail to function and plasma membrane
targeting. J Biol Chem, 2001;276:6825–33.
93. Leslie, E.M., Watkins, P.B., Kim, R.B., and Brouwer, K.L. Differential inhi­
bition of rat and human Na+‐dependent taurocholate cotransporting poly­
peptide (NTCP/SLC10A1)by bosentan: a mechanism for species differences
in hepatotoxicity. J Pharmacol Exp Ther, 2007;321:1170–8.
94. Scharschmidt, B.F., Waggoner, J.G., and Berk, P.D. Hepatic organic anion
uptake in the rat. J Clin Invest, 1975;56:1280–92.
95. Min, A.D., Johansen, K.L., Campbell, C.G., and Wolkoff, A.W. Role of chlo­
ride and intracellular pH on the activity of the rat hepatocyte organic anion
transporter. J Clin Invest, 1991;87:1496–502.
96. Wolkoff, A.W., Samuelson, A.C., Johansen, K.L. et al. Influence of Cl‐ on
organic anion transport in short‐term cultured rat hepatocytes and isolated
perfused rat liver. J Clin Invest, 1987;79:1259–68.
97. Jacquemin, E., Hagenbuch, B., Stieger, B. et al. Expression of the hepatocel­
lular chloride‐dependent sulfobromophthalein uptake system in Xenopus
laevis oocytes. J Clin Invest, 1991;88:2146–9.
336 THE LIVER:  REFERENCES
98. Jacquemin, E., Hagenbuch, B., Stieger, B. et al. Expression cloning of a rat
liver Na(+)‐independent organic anion transporter. Proc Natl Acad Sci U S
A, 1994;91:133–7.
99. Hagenbuch, B. and Gui, C. Xenobiotic transporters of the human organic
anion transporting polypeptides (OATP) family. Xenobiotica,
2008;38:778–801.
100. Zhao, W., Zitzow, J.D., Weaver, Y. et al. Organic anion transporting poly­
peptides contribute to the disposition of perfluoroalkyl acids in humans and
rats. Toxicol Sci, 2017;156:84–95.
101. Chen, C., Stock, J.L., Liu, X. et al. Utility of a novel Oatp1b2 knockout
mouse model for evaluating the role of Oatp1b2 in the hepatic uptake of
model compounds. Drug Metab Dispos, 2008;36:1840–5.
102. Lu, H., Choudhuri, S., Ogura, K. et  al. Characterization of organic
anion transporting polypeptide 1b2‐null mice: essential role in hepatic
uptake/toxicity of phalloidin and microcystin‐LR. Toxicol Sci, 2008;
103:35–45.
103. Zaher, H., Meyer zu Schwabedissen, H.E., Tirona, R.G. et al. Targeted dis­
ruption of murine organic anion‐transporting polypeptide 1b2 (Oatp1b2/
Slco1b2) significantly alters disposition of prototypical drug substrates
pravastatin and rifampin. Mol Pharmacol, 2008;74:320–9.
104. Gong, L., Aranibar, N., Han, Y.H. et al. Characterization of organic anion‐
transporting polypeptide (Oatp) 1a1 and 1a4 null mice reveals altered trans­
port function and urinary metabolomic profiles. Toxicol Sci, 2011;
122:587–97.
105. Lee, H.H., Leake, B.F., Kim, R.B., and Ho, R.H. Contribution of organic
anion‐transporting polypeptides 1A/1B to doxorubicin uptake and clear­
ance. Mol Pharmacol, 2017;91:14–24.
106. Hata, S., Wang, P., Eftychiou, N. et al. Substrate specificities of rat oatp1
and ntcp: implications for hepatic organic anion uptake. Am J Physiol
Gastrointest Liver Physiol, 2003;285:G829–39.
107. Wang, D.Q., Tazuma, S., Cohen, D.E., and Carey, M.C. Feeding natural
hydrophilic bile acids inhibits intestinal cholesterol absorption: studies in
the gallstone‐susceptible mouse. Am J Physiol Gastrointest Liver Physiol,
2003;285:G494–502.
108. Fu, D., Wakabayashi, Y., Lippincott‐Schwartz, J., and Arias, I.M. Bile acid
stimulates hepatocyte polarization through a cAMP‐Epac‐MEK‐LKB1‐
AMPK pathway. Proc Natl Acad Sci U S A, 2011;108:1403–8.
109. Wolkoff, A. The hyperbilirubinemias, in Harrison’s Principles of Internal
Medicine, (eds. H.L. Jameson et al.). McGraw Hill, New York, 2018.
110. van de Steeg, E., Stranecky, V., Hartmannova, H. et al. Complete OATP1B1
and OATP1B3 deficiency causes human Rotor syndrome by interrupting con­
jugated bilirubin reuptake into the liver. J Clin Invest, 2012;122:519–28.
111. Wang, P., Kim, R.B., Chowdhury, J.R., and Wolkoff, A.W. The human
organic anion transport protein SLC21A6 is not sufficient for bilirubin
transport. J Biol Chem, 2003;278:20695–9.
112. Cvorovic, J. and Passamonti, S. Membrane transporters for bilirubin and its
conjugates: a systematic review. Front Pharmacol, 2017;8:887.
113. Oleaga, A., Gonzalez, J., and Esteller, A. Effects of two‐thirds hepatectomy
on sulfobromophthalein handling by the rat liver. Comp Biochem Physiol A
Comp Physiol, 1987;87:13–9.
114. Gartner, U., Stockert, R.J., Morell, A.G., and Wolkoff, A.W. Modulation of
the transport of bilirubin and asialoorosomucoid during liver regeneration.
Hepatology, 1981;1:99–106.
115. Gerloff, T., Geier, A., Stieger, B. et al. Differential expression of basolateral
and canalicular organic anion transporters during regeneration of rat liver.
Gastroenterology, 1999;117:1408–15.
116. Angeletti, R.H., Bergwerk, A.J., Novikoff, P.M., and Wolkoff, A.W.
Dichotomous development of the organic anion transport protein in liver
and choroid plexus. Am J Physiol, 1998;275:C882–7.
117. Dubuisson, C., Cresteil, D., Desrochers, M. et al. Ontogenic expression of
the Na(+)‐independent organic anion transporting polypeptide (oatp) in rat
liver and kidney. J Hepatol, 1996;25:932–40.
118. Campbell, C.G., Spray, D.C., and Wolkoff, A.W. Extracellular ATP4‐ mod­
ulates organic anion transport by rat hepatocytes. J Biol Chem,
1993;268:15399–404.
119. Glavy, J.S., Wu, S.M., Wang, P.J. et al. Down‐regulation by extracellular
ATP of rat hepatocyte organic anion transport is mediated by serine phos­
phorylation of Oatp1. J Biol Chem, 2000;275:1479–84.
120. Xiao, Y., Nieves, E., Angeletti, R.H. et al. Rat organic anion transporting
protein 1A1 (Oatp1a1): purification and phosphopeptide assignment.
Biochemistry, 2006;45:3357–69.
121. Choi, J.H., Murray, J.W., and Wolkoff, A.W. PDZK1 binding and serine
phosphorylation regulate subcellular trafficking of organic anion transport
protein 1a1. Am J Physiol Gastrointest Liver Physiol, 2011;300:G384–93.
122. Ho, R.H., Choi, L., Lee, W. et al. Effect of drug transporter genotypes on
pravastatin disposition in European‐ and African‐American participants.
Pharmacogenet Genomics, 2007;17:647–56.
123. Miura, M., Satoh, S., Inoue, K. et al. Influence of SLCO1B1, 1B3, 2B1 and
ABCC2 genetic polymorphisms on mycophenolic acid pharmacokinetics
in Japanese renal transplant recipients. Eur J Clin Pharmacol,
2007;63:1161–9.
124. Gong, I.Y. and Kim, R.B. Impact of genetic variation in OATP transporters to
drug disposition and response. Drug Metab Pharmacokinet, 2013;28:4–18.
125. Link, E., Parish, S., Armitage, J. et al. SLCO1B1 variants and statin‐induced
myopathy—a genomewide study. N Engl J Med, 2008;359:789–99.
126. Tirona, R.G., Leake, B.F., Merino, G., and Kim, R.B. Polymorphisms in
OATP‐C: identification of multiple allelic variants associated with altered
transport activity among European‐ and African‐Americans. J Biol Chem,
2001;276:35669–75.
127. Wang, P., Wang, J.J., Xiao, Y. et al. Interaction with PDZK1 is required for
expression of organic anion transporting protein 1A1 on the hepatocyte sur­
face. J Biol Chem, 2005;280:30143–9.
128. Ye, F. and Zhang, M. Structures and target recognition modes of PDZ domains:
recurring themes and emerging pictures. Biochem J, 2013;455:1–14.
129. Wang, W.J., Murray, J.W., and Wolkoff, A.W. Oatp1a1 requires PDZK1 to
traffic to the plasma membrane by selective recruitment of microtubule‐
based motor proteins. Drug Metab Dispos, 2014;42:62–9.
130. Mukhopadhyay, A., Quiroz, J.A., and Wolkoff, A.W. Rab1a regulates sort­
ing of early endocytic vesicles. Am J Physiol Gastrointest Liver Physiol,
2014;306:G412–24.
131. Novikoff, P.M., Cammer, M., Tao, L. et al. Three‐dimensional organization
of rat hepatocyte cytoskeleton: relation to the asialoglycoprotein endocyto­
sis pathway. J Cell Sci, 1996;109(1):21–32.
132. LaLonde, D.P. and Bretscher, A. The scaffold protein PDZK1 undergoes a
head‐to‐tail intramolecular association that negatively regulates its interac­
tion with EBP50. Biochemistry, 2009;48:2261–71.
 Hepatic Nuclear Receptors
28 Raymond E. Soccio
University of Pennsylvania, Perelman School of Medicine, Department of Medicine, Division of
Endocrinology, Diabetes, and Metabolism, Institute for Diabetes, Obesity, and Metabolism, Philadelphia
Crescenz VA Medical Center, Philadelphia, PA, USA
INTRODUCTION
Nuclear receptors (NRs) are important regulators of develop-
ment and homeostasis, and many family members have vital
roles in the liver. NRs are transcription factors whose function is
reflected in their bipartite structure: the DNA‐binding domain to
directly bind regulatory DNA and the ligand‐binding domain to
sense hormones, metabolites, or drugs and modulate expression
of nearby genes [1–3]. Thus, NRs directly sense the environ-
ment to regulate expression of target genes. Their ligands are
small molecules falling into three classes: hormones which cir-
culate to affect distant organs, molecules generated locally by
normal metabolism, or exogenous compounds like drugs.
Sensing and responding to hormones, metabolites, and drugs is
an essential function of the liver to maintain homeostasis – the
constant internal environment  –  in the face of changes to the
external environment. One critical aspect of the external
environment to maintain life is nutrient availability, so many
liver NRs directly or indirectly regulate energy metabolism in
the fed or fasted state [2]. NRs are also notable for their pharma-
cology, with “druggable” ligand binding domains such that 13%
of all FDA‐approved drugs target NRs [1].
The 48 human NRs have been classified into subfamilies by
their sequence homology. This system is reflected by the names
(i.e. NR1A1) from the NR Nomenclature Committee [4], but most
NRs’ names and official gene symbols still relate to their original
discoveries. The field of NR study began in the 1980s, with the
molecular cloning of several hormone receptors revealing a com-
mon structure [1, 3]. For classic endocrine receptors, names accu-
rately describe the hormones that serves as ligands: thyroid
hormone, steroid hormones, and vitamin D. Other nuclear recep-
tors were first identified as “orphans” lacking known ligands,
though ligands have since been identified for many [2]. Due to
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
historical artifacts, many NR names are uninformative (like liver X
receptor) or even misleading. For example, farnesoid X receptor
(FXR) binds with high affinity to bile acids not farnesol deriva-
tives, and despite attempts to rename it bile acid receptor (BAR)
[5] the name FXR has persisted. This chapter will generally refer
to NR names in common use, and hepatic NRs will be broadly
divided into five classes based on their functions and ligands:
endocrine, xenobiotic, metabolic, circadian, and other (Table 28.1).
The sixth class of NRs have undetectable expression in liver and
are thus not covered here: GCNF is restricted to germ cells, TLX
to brain, and PNR to retina. Likewise, SF‐1 and DAX1 are
restricted to steroidogenic tissues, though their close relatives
LRH1 and SHP have key functions in liver.
The majority of the 48 human NRs are expressed in liver
(Figure 28.1). Human mRNA expression across multiple tissues
has been reported by several consortia, including the Human
Protein Atlas [6] and Genotype‐Tissue Expression (GTEx) [7].
Some NRs, such as the glucocorticoid receptor (GR), are
broadly expressed across all tissues, such that hepatic expres-
sion mediates the ligand action in liver. Other NRs have expres-
sion limited to liver and other tissues where the ligand acts, such
as FXR being mainly restricted to liver and other gastrointesti-
nal tissue related to bile acids their enterohepatic circulation.
One NR, the xenobiotic receptor CAR, is expressed primarily in
the liver with low or undetectable expression elsewhere, such
that its main effects are hepatic. In many NR groups, one family
member predominates in liver. For instance, the β isoform of
thyroid hormone receptor (THR) is the major liver form, and
only one of the three estrogen related receptors (ERRα) has
appreciable human liver expression. Finally, given that hepato-
cytes constitute >90% of liver mass, this chapter mainly
addresses NR functions in hepatocytes, though some important
roles in non‐parenchymal liver cells are also discussed.
 Class Group NRNC Name Gene HPA (TPM) GTEx (RPKM)
Thyroid Hormone NR1A2 THRβ Thyroid hormone receptor β THRB 11.6 2.7
Receptor NR1A1 THRα Thyroid hormone receptor α THRA 2.5 1.8
NR3C1 GR Glucocorticoid receptor NR3C1 20.6 4.9
3-ketosteroid NR3C2 MR Mineralocorticoid receptor NR3C2 4.4 1.4
Endocrine receptors NR3C3 PR Progesterone receptor PGR 0.1 0.0
NR3C4 AR Androgen receptor AR 21.6 8.3
NR3A1 ERα Estrogen receptor α ESR1 8.8 1.9
Estrogen Receptor
NR3A2 ERβ Estrogen receptor β ESR2 0.0 0.1
NR1I1 VDR Vitamin D receptor VDR 0.5 0.2
Vitamin D receptor-
NR1I2 PXR Pregnane X receptor NR1I2 20.0 23.3
Xenobiotic like
NR1I3 CAR Constitutive androstane receptor NR1I3 82.5 52.3
NR1H3 LXRα Liver X receptor α NR1H3 26.7 19.2
Liver X Receptor-like NR1H2 LXRβ Liver X receptor β NR1H2 11.0 14.3
NR1H4 FXR Farnesoid X receptor NR1H4 66.6 22.3
Peroxisome proliferator-activated
NR1C1 PPARα receptor α PPARA 14.7 11.7
Peroxisome
proliferator Peroxisome proliferator-activated
Activated receptor NR1C2 PPARβ/δ receptor δ PPARD 2.9 8.4
Peroxisome proliferator-activated
NR1C3 PPARγ receptor γ PPARG 8.6 2.0
NR1B1 RARα Retinoic acid receptor α RARA 11.8 9.8
Retinoic acid
Metabolic receptor NR1B2 RARβ Retinoic acid receptor β RARB 3.3 0.8
NR1B3 RARγ Retinoic acid receptor γ RARG 0.9 0.8
NR2B1 RXRα Retinoid X receptor α RXRA 7.2 22.7
Retinoid X receptor NR2B2 RXRβ Retinoid X receptor β RXRB 1.8 14.2
NR2B3 RXRγ Retinoid X receptor γ RXRG 0.0 0.1
NR5A2 LRH1 Liver receptor homolog 1 NR5A2 19.5 4.8
LRH1/SF1
NR5A1 SF1 Steroidogenic factor 1 NR5A1 0.0 0.0
NR0B2 SHP Small heterodimer partner NR0B2 48.8 46.9
Dosage-sensitive sex reversal,
SHP/DAX
adrenal hypoplasia critical region,
NR0B1 DAX1 on chromosome X, gene 1 NR0B1 0.0 0.0
NR1D1 REVERBα Rev-ErbAα NR1D1 6.7 14.1
Rev-erb
NR1D2 REVERBβ Rev-ErbAβ NR1D2 15.4 4.9
NF1F1 RORα RAR-related orphan receptor α RORA 14.5 4.6
RAR-related orphan
NF1F2 RORβ RAR-related orphan receptor β RORB 0.0 0.0
Circadian receptor
NF1F3 RORγ RAR-related orphan receptor γ RORC 43.0 18.4
NR3B1 ERRα Estrogen-related receptor α ESRRA 11.3 24.3
Estrogen Receptor
NR3B2 ERRβ Estrogen-related receptor β ESRRB 0.0 0.0
Related
NR3B3 ERRγ Estrogen-related receptor γ ESRRG 0.3 0.0
Hepatocyte nuclear NR2A1 HNF4α Hepatocyte nuclear factor 4 α HNF4A 51.0 39.5
factor 4 NR2A2 HNF4γ Hepatocyte nuclear factor 4 γ HNF4G 6.5 2.1
NR2C1 TR2 Testicular receptor 2 NR2C1 2.9 4.7
NR2C/TR
NR2C2 TR4 Testicular receptor 4 NR2C2 3.2 1.0
Chicken ovalbumin upstream
NR2F1 COUP-TF1 promoter-transcription factor I NR2F1 5.7 4.2
Other
NR2F/COUP Chicken ovalbumin upstream
NR2F2 COUP-TF2 promoter-transcription factor II NR2F2 16.6 6.4
NR2F6 EAR2 V-erbA-related NR2F6 12.2 33.9
NR4A1 NGF1B Nerve growth factor IB NR4A1 20.7 16.9
NR4A NR4A2 NURR1 Nuclear receptor related 1 NR4A2 5.3 6.2
NR4A3 NOR1 Neuron-derived orphan receptor 1 NR4A3 5.3 0.9
GCNF NR6A1 GCNF Germ cell nuclear factor NR6A1 0.5 0.9
Homologue of the Drosophila
Non-liver NR2E1 TLX tailless gene NR2E1 0.0 0.0
TLX/PNR
Photoreceptor cell-specific
NR2E3 PNR nuclear receptor NR2E3 0.0 0.0

Figure 28.1  48 Members of the human nuclear receptor superfamily. In this chapter, the 48 human NRs are divided into these six classes based
on their functions in liver. Based on sequence homology, the Nuclear Receptor Nomenclature Committee (NRNC) has divided NRs into 19 groups
and assigned standardized names. However, NRs are commonly called other names, and these abbreviated and full names are shown. The official
human gene symbols reflect either common or NRNC names. Gene expression profiling across human tissues has been reported by several consor-
tia including the Human Protein Atlas (HPA) and Genotype‐Tissue Expression (GTEx) [6]. The normalized expression of each human NR in liver
is thus quantified by RNA‐sequencing as either transcripts per million (TPM) or reads per kilobase per million mapped reads (RPKM). The color
scaling in these columns reflect hepatic expression levels (green: high, yellow: middle, red: low). NRs are listed in the order they are presented in
this chapter, with the 21 highlighted in yellow discussed most extensively due to their high expression and/or well‐studied roles in liver.
 28:  Hepatic Nuclear Receptors 339

Type 1 Type 2 Type 3 Type 4

Homodimers with
ligand-dependent
nuclear localization
RXR heterodimers with Exclusively homodimers
Monomers
ligand-dependent in nucleus with
(or Homodimers)
transcriptional activation uncertain ligand effects

LBD GR GR THR RXR HNF4 HNF4 ROR

DBD
DNA

Inverted Direct Direct Half Site

Repeat Repeat Repeat

Endocrine/Steroid: Endocrine: THR, VDR Other (NR2 family): Circadian:

GR, AR, ER Metabolic: LXR, FXR, HNF4, COUP-TF, TR Reverb, ROR, ERR
PPAR, RAR (RXR homodimers)
Xenobiotic:
CAR, PXR
Figure 28.2  Four mechanistic types of nuclear receptors (NRs). NRs have been classified into these four mechanistic types based on their dimeri-
zation, response elements, and ligand dependence. Notably, the four types correspond roughly to their functional classifications in liver. Type 1
receptors are the endocrine receptors for steroid hormones. Type 2 receptors include both the endocrine receptors for thyroid hormone and vitamin
D, as well as the four metabolic receptors with prominent roles in liver (LXRs, FXR, PPARs, and RARs). The endocrine RXR heterodimers are
“non‐permissive” such that RXR ligands are generally silent, while the metabolic ones are “permissive” such that they can also be activated by RXR
agonists. The xenobiotic receptors have features of both type 1 and type 2, the circadian receptors are generally type 4, and the other orphan recep-
tors of the NR2 family are type 3. LBD, ligand binding domain; DBD, DNA binding domain. Figure adapted from [3].

NRs have also been classified into mechanistic types based
on their DNA response elements, dimerization, and subcellu-
lar localization in the absence of ligand (Table 28.1, Figure
28.2). Each NR monomer binds to a core 6 nucleotide consen-
sus sequence (AGGTCA or related sequences), thus NR
dimers bind to response elements consisting of two such “half
sites”. For different dimers, the two half sites have variable
orientation (direct or inverted) and spacing between them (1–5
nucleotides), giving rise to names like direct repeat 1 (DR‐1)
and inverted repeat 2 (IR‐2). Steroid hormone receptors like
GR are type 1 NRs. In the absence of ligand, type 1 NRs reside
predominantly in the cytosol associated with heat shock pro-
teins (HSPs). Ligand binding triggers dissociation of HSPs,
homodimerization, and nuclear translocation to bind inverted
repeat DNA response elements and activate target genes. Other
endocrine NRs and the metabolic NRs are type 2, constitu-
tively nuclear and binding to various direct repeats as heterodi-
mers with retinoid X receptors (RXRs). In the absence of
ligand, these RXR heterodimers associate with corepressor
proteins and actively decrease expression of target genes – giv-
ing lower expression than even the basal level in the absence
of receptor. Ligand binding triggers a conformation change,
with release of corepressors and binding of coactivators. This
simplified “coregulator switch model” of type 2 NRs like THR
is well‐validated in vitro, though in vivo relevance remains to
be proven [2]. Type 3 NRs, mainly the NR2 family members
(HNF4s, COUP‐TFs, and TRs), are homodimers with consti-
tutive nuclear localization and direct repeat binding. Finally,
the circadian NRs (Rev‐erbs, RORs, ERRs) are type 4, which
can bind DNA as monomers to half sites and/or as dimers to
direct repeats.
ENDOCRINE NUCLEAR HORMONE
RECEPTORS IN LIVER
Thyroid hormone receptors (THRs)
Thyroid hormones exert effects throughout the body, including
prominent actions on energy metabolism and the cardiovascular
system. The liver plays key roles in thyroid hormone physiol-
ogy, both as a target of local effects and a mediator of systemic
effects [8]. Two THR genes encode the THRα and THRβ iso-
forms. While most tissues express both, THRβ is the major
receptor in liver (as well as brain), while THRα predominates in
the cardiovascular system and bone.
In the liver, thyroid hormone regulates many target genes
with key roles in lipid metabolism [9]. Some transcriptional
effects of thyroid hormone are direct via THRs, while oth-
ers may be indirect as THR can regulate other key transcrip-
tion factors in lipid metabolism including sterol regulatory
element binding proteins (SREBP1c regulating fatty acid
synthesis and SREBP2 cholesterol synthesis), carbohy-
drate‐responsive element binding proteins (ChREBPs), and
metabolic NRs (like LXRs, PPARs, FXRs). Some THR tar-
gets, such as the highly‐regulated THRSP (thyroid hormone
responsive protein, also called Spot14), are involved in de
novo lipogenesis, the synthesis of fatty acids from glucose.
However, thyroid hormones also activate fatty acid oxida-
tion, and this catabolic effect predominates in hyperthy-
roidism to reduce liver fat. Thyroid hormones similarly
affect both cholesterol synthesis and clearance, with the lat-
ter apparently predominating, including effects on the LDL
receptor activity and bile acid synthesis via cholesterol
 Table 28.1  Hepatic nuclear receptor (NR) ligands, dimerization, and response elements. For the five classes of hepatic nuclear receptor defined here, three key features are shown for each NR or
group of NR isoforms: ligands, dimerization status, and response elements. These features define the mechanistic type of NR (types 1–4, see Figure 28.2). In general, there are well‐defined ligands
for endocrine, xenobiotic, and metabolic NRs, with regulation by hormones, various drugs, and lipid metabolites, respectively. Most circadian and other NRs are still considered orphans (brackets
in table), either without a widely‐accepted ligand, or having a described ligand but without clear gene regulatory effects. The various NRs can function as monomers, homodimers, and/or heterodi-
mers with retinoid X receptors (RXRs). NRs bind to DNA motifs called response elements, consisting of a 6‐nucleotide half site, or two such sites in the same (direct repeat, DR) or opposite
(inverted repeat, IR) directions spaced by 1–5 nucleotides. Some NRs have well‐studied and specific direct target genes in liver, and examples are shown. However, many liver NRs directly or
indirectly regulate many often‐overlapping sets of genes, such that no specific example is shown. See text for details
Class Receptor(s) Ligand(s) Dimerization Response element(s) Type Example canonical liver target gene
Endocrine THRs T3 (active thyroid hormone) RXR heterodimer DR‐4 2 THRSP (thyroid hormone responsive protein)
GR Glucocorticoids Homodimer IR‐3 1 G6PC (Glucose 6‐phosphatase)
AR Androgens Homodimer IR‐3 1
ERs Estrogens Homodimer IR‐3 1
VDR Calcitriol (active vitamin D) RXR heterodimer DR‐3 2 CYP24A1 (vitamin D 24‐hydroxylase)
Xenobiotic CAR Many, including phenobarbital RXR heterodimer DR‐3 or DR‐4 1,2 CYP2B1 (phase 1 drug metabolism)
PXR Many, including rifampin RXR heterodimer DR‐3 or DR‐4 1,2 CYP3A4 (phase 1 drug metabolism)
Metabolic LXRs Oxysterols RXR heterodimer DR‐4 2 ABCG5/ABCG8 (cholesterol efflux)
FXR Bile acids RXR heterodimer IR‐1, DR‐5 2 ABCB11 (Bile salt export pump)
PPARs Fatty acid derivatives RXR heterodimer DR‐1 2 FGF21 (liver‐derived hormone)
RARs All‐trans retinoic acid (ATRA) RXR heterodimer DR‐5, DR‐2, others 2 CRABP1/2 (cellular retinoic acid binding proteins)
RXRs 9‐cis retinoic acid Heterodimer, homodimer DR‐1 (homodimer) 2,3
LRH1 (Phospholipids) Monomer, SHP heterodimer Half site 4 CYP7A1 (bile acid synthesis)
SHP (Orphan) LRH1, others N/A N/A
Circadian Reverbs (Heme) Monomers or homodimers Half site or DR‐2 4 BMAL1 (core clock gene)
RORs (Orphan) Monomers Half site 4 BMAL1 (core clock gene)
ERRs (Orphan) Monomers or homodimers Half site 4
Other NR2A/HNF4s (Linoleic acid) Homodimers DR‐1 3
NR2C/TRs (Orphan) Homodimers Various DRs 3
NF2F/COUP‐TFs (Orphan) Homodimers Various DRs 3
NR4As (Orphan) RXR heterodimers, monomers DR‐5 or half site 2,4

c28.indd 340 03-11-2019 19:50:15

 28:  Hepatic Nuclear Receptors 341
7α‐hydroxylase (CYP7A1). Hypothyroidism increases
serum cholesterol (particularly in low density lipoproteins)
and triglycerides, and also causes liver steatosis. Conversely,
thyroid hormone is an effective treatment for dyslipidemia
but limited by adverse cardiac effects (tachycardia) and
decreased bone mass. Selective activation of liver THRs,
either by TRβ specific agonists or selective hepatic delivery,
is thus a promising therapeutic approach to dyslipidemia
and non‐alcoholic fatty liver disease (NAFLD) and non‐
alcoholic steatohepatitis (NASH).
In addition to the liver being a target of thyroid hormone, it
alters systemic thyroid hormone levels by two main mecha-
nisms: deiodination and synthesis of binding proteins. The
thyroid gland, in response to the hypothalamic–pituitary axis,
releases the hormones T4 (thyroxine, with four iodines) and T3
(3,5,3′‐triiodothyronine, with one less iodine in the 5′ posi-
tion of its outer ring). THRs bind predominantly to T3, such
that T4 is essentially a prohormone acted upon by deiodinase
enzymes in tissues to generate T3. Importantly, the liver
expresses type 1 deiodinase enzyme (DIO1), which can deio-
dinate the outer ring to generate either active T3 or the inactive
reverse T3 (rT3), as well as the inner ring to generate other
inactive derivatives [10]. While DIO1 contributes to circulat-
ing levels of active T3, its main function in liver appears to be
thyroid hormone inactivation. In mouse models, DIO1 is a
target gene activated by thyroid hormone in liver, such that
liver TRβ‐deficiency decreases DIO1, while thyrotoxicosis
increases hepatic DIO1. Therefore, the liver mediates nega-
tive feedback whereby excess thyroid hormone, via TRβ,
stimulates its own inactivation by DIO1.
The liver is also the source of all three circulating thyroid
hormone binding proteins: thyroxine‐binding globulin (TBG),
transthyretin (TTR, also called prealbumin), and albumin [11].
While free thyroid hormone is the active fraction taken up by
cells, more than 99% of circulating thyroid hormone is protein‐
bound, and the two forms are in equilibrium. Steroid NR
ligands are well‐known to alter TBG, the major binding pro-
tein, and thus affect the total versus free fractions of serum T4:
androgens and glucocorticoids decrease TBG, while estrogens
increase TBG (though indirectly by affecting glycosylation).
Often these effects on binding proteins only affect total thyroid
hormone levels but not the active free levels, such that patients
remain clinically euthyroid, though selective effects on thyroid
hormone activity are plausible. Liver diseases like cirrhosis or
acute hepatitis can thus affect thyroid hormone binding pro-
teins or deiodination [8].
Glucocorticoid receptor (GR)
GR is present in nearly every cell type, and glucocorticoids exert
pleiotropic effects on the immune and cardiovascular systems, as
well as metabolism. Pharmacologic glucocorticoids are primar-
ily used as anti‐inflammatory and immunosuppressive agents,
but metabolic side‐effects are common. The name glucocorticoid
stems from their effects on carbohydrate metabolism to maintain
blood glucose in the fasted state, most prominently via stimula-
tion of gluconeogenesis in hepatocytes [12]. Thus, direct GR
target genes in liver include phosphoenolpyruvate kinase
(PCK1), encoding the rate‐limiting gluconeogenic enzyme, and
glucose‐6‐phosphatase (G6PC), which is required to liberate
free glucose. Glucocorticoids also stimulate hepatic synthesis
of glycogen, fatty acids, and cholesterol, particularly in the
fed state in the presence of insulin. Indeed, hepatic steatosis is
observed with glucocorticoid treatment, and this effect may
be due to direct cell‐autonomous effects on hepatocyte GR as
well as indirect effects upon adipose tissue, where glucocorti-
coids activate lipolysis releasing fatty acids for uptake in liver
[13]. Overall, glucocorticoid excess (Cushing’s syndrome,
either iatrogenic or endogenous due to overproduction), is
associated with metabolic disease including obesity, insulin
resistance, and hyperglycemia, some of which relate to GR
effects in liver.
In addition to gene activation, widespread gene repression is
also noted in response to glucocorticoids, particularly of inflam-
matory genes. As opposed to direct DNA binding by GR to acti-
vate genes, this gene repression is thought to result from
“transrepression,” whereby GR is brought to DNA indirectly by
tethering to inhibit activation by prominent inflammatory tran-
scription factors like AP‐1 and NF‐κB. Therefore, an elusive
goal in pharmacology has been developing glucocorticoid ana-
logs that stimulate transrepression of inflammatory genes with-
out the adverse metabolic effects of activating GR in liver,
muscle, and fat [14].
Cortisol is the main glucocorticoid in humans, produced by
the zona fasciculata of the adrenal cortex in response to a hypo-
thalamic–pituitary axis. 11β‐hydroxysteroid dehydrogenase
(HSD) enzymes interconvert cortisol and inactive cortisone,
with 11β‐HSD2 primarily inactivating cortisol (notably in renal
cells to protect the mineralocorticoid receptor from cortisol) and
11β‐HSD1 primarily reactivating cortisone. The liver is the
main site of 11β‐HSD1 expression to activate cortisone to corti-
sol, and this enzyme has been implicated strongly in metabolic
diseases like NAFLD [13]. Liver also expresses enzymes that
inactivate cortisol including 5α‐ and 5β‐reductases and 3α‐HSD,
and thyroid hormone increases hepatic cortisol clearance by
activating these pathways. Cortisol also circulates predomi-
nantly bound to a liver‐derived carrier protein, cortisol binding
globulin (CBG). Therefore, similar to the situation for thyroid
hormone, the liver is a direct target for glucocorticoid action on
glucose and lipid metabolism via GR, and it also plays a central
role in determining systemic glucocorticoid effects via cortisol
metabolism and CBG production.
Mineralocorticoid receptor (MR)
Expression of MR is detectable in liver, though it is uncertain
whether this is in hepatocytes, versus in endothelial cells or
macrophages where MR expression has been studied [15].
MR is closely related to GR, and in the absence of 11β‐HSD2
to inactivate cortisol, MR can be similarly activated by corti-
sol or the main mineralocorticoid aldosterone. Aldosterone is
produced by the zona glomerulosa of the adrenal cortex,
under control of the renin–angiotensin system, and it binds
MR in kidney to activate genes like the epithelial sodium
channel (ENaC) to regulate sodium and potassium homeostasis
and blood pressure. MR also has roles in the cardiovascular
system and gut, but little is known about any function for
hepatic MR.
342 THE LIVER:  XENOBIOTIC SENSING RECEPTORS: CAR AND PXR
Estrogen and androgen sex steroid receptors
(ER and AR)
Sex hormone receptors are widely expressed and not limited to
reproductive tissues. Like GR and MR, they are type 1 NRs with
ligand‐induced nuclear localization. AR and ERα are expressed in
liver in both sexes, while ERβ and PR have little to no hepatic
expression. Sex differences are observed in many liver diseases,
with male preponderance in steatosis, hepatitis, cirrhosis, and
hepatocellular carcinoma [16]. Of course, other factors besides sex
steroids drive many differences, including behaviors related to
viral infection and alcohol use. Sexual dimorphism has also been
found in the liver gene expression, with sex‐specific expression of
more than 1000 genes including some encoding drug‐metaboliz-
ing enzymes. Some of these differences had been attributed to
direct cell‐autonomous effects of ERα and AR in liver, but more
recent studies suggest brain control of this dimorphism via sex‐
specific temporal patterns of growth hormone release [17].
Estrogen is the primary female sex hormone, synthesized in
the ovary by the combined action of theca and granulosa cells,
the latter expressing aromatase enzyme. In males, much lower
amounts of estrogen are generated by the testes and peripheral
aromatase‐expressing tissues including fat. Clinical observa-
tions show that modifying estrogen levels, either increasing (i.e.
hormone replacement in post‐menopausal women) or decreas-
ing (suppression therapy in breast cancer), affects serum lipid
profiles, generally with estrogen lowering LDL cholesterol and
raising HDL cholesterol and triglycerides [18]. Furthermore,
mouse models with estrogen deficiency show hepatic steatosis
that is rescued by estrogen treatment, yet such studies are unable
to distinguish liver autonomous versus whole body effects of
estrogen. More recent liver‐specific knockout mice have shown
a role of hepatic ERα in glucose and lipid metabolism, with
hyperglycemia and liver steatosis [19]. Liver ERα has also been
linked to cholesterol and lipoprotein metabolism, via a func-
tional interaction with LXRα (see Liver X and farnesoid X
receptors (LXR and FXR). Indeed, estrogen can ameliorate the
fatty liver induced by LXR agonists, in a manner requiring
hepatic ERα [20]. Furthermore, ERs can even bind oxysterols
that are classically LXR agonists [21].
Androgens are synthesized in the testes and the zona reticula-
ris of the adrenal cortex, with the latter being the main source of
low levels in females. Androgen excess in females is part of the
polycystic ovarian syndrome (PCOS), with metabolic manifesta-
tions including obesity, insulin resistance, and NAFLD, though it
is unlikely that hepatic sex steroid receptors play a significant
role in PCOS pathophysiology. The effect of androgens on liver
steatosis is complex, with inconsistent and opposite effects in
various studies [16]. Liver‐specific AR knockout causes hepatic
steatosis in male but not female mice [22], though the mecha-
nism is unclear. Similar to ER, AR has been proposed to regulate
cholesterol metabolism by cross‐talk with LXR [23].
Like for thyroid and glucocorticoid hormones, the liver is also
a site for metabolism and clearance of sex steroids, and the syn-
thesis of sex hormone binding globulin (SHBG). Hepatic SHBG
expression is increased by estrogens and decreased by andro-
gens. Notably, in contrast to TBG and CBG, SHBG levels do
appear to alter levels of free sex hormones in blood, and SHBG
is an emerging biomarker in metabolic disease [24]. Overall, the
role of sex hormones and their receptors in liver energy
­homeostasis is complex and an area of active investigation.
Vitamin D receptor (VDR)
Unlike the steroid receptors, VDR is an RXR heterodimer type
2 NR like THR. VDR is activated by the hormone calcitriol, or
1,25‐dihydroxyvitamin D. The precursor vitamin D3, or chole-
calciferol, is obtained from the diet or synthesized in sun‐
exposed skin. 25‐hydroxylation of vitamin D3 occurs in liver,
via the P450 enzyme encoded by the CYP2R1 gene. However,
CYP2R1 is not regulated by VDR or other inputs, so constitutive
activity in liver results in inactive 25‐hydroxyvitamin D as the
primary circulating form reflecting overall vitamin D levels.
The final activation to calcitriol occurs via 1α‐hydroxylation in
the kidney, a step under tight regulation by parathyroid hormone
and phosphate levels. Conversely, calcitriol is inactivated by 24‐
hydroxylation, and the CYP24A1 gene encoding this 24‐hydroxy-
lase is highly regulated by VDR. Thus, calcitriol exerts negative
feedback on itself by VDR activation of CYP24A1.
VDR is expressed in most tissues, though overall hepatic
expression is extremely low compared to its levels in parathy-
roid, kidney, and GI tract [6]. Indeed, hepatocytes have little
to  no detectable VDR, but levels are much higher in non‐
parenchymal cells of liver particularly hepatic stellate cells
(HSCs), where VDR regulates CYP24A1 and other genes [25].
Importantly, HSCs mediate liver fibrosis in response to injury
and pathology, and an antifibrotic role of VDR in HSCs is
emerging [26]. VDR activation inhibits the main profibrotic
pathway in HSCs mediated by TGFβ, and VDR knockout mice
show spontaneous liver fibrosis. Thus, there is potential for
VDR‐targeted therapies in fibrotic liver diseases, especially if
liver‐specific ligands are developed to avoid hypercalcemia due
to excess VDR activity in bone, gut, and kidney.
XENOBIOTIC SENSING RECEPTORS:
CAR AND PXR
CAR and PXR are VDR‐related NRs that sense xenobiotics,
foreign substances like drugs and toxins that are not naturally
produced by the organism but require clearance. The names of
both NRs reflect historical identification of ligands, though
numerous chemical compounds have now been shown to serve
with variable affinities as agonists, antagonists, partial agonists,
and inverse agonists [27]. The structures of PXR and CAR
reveal large ligand binding pockets that can accommodate such
large and diverse ligands, with many being dual activators of
both CAR and PXR. In research, the model agonist ligands for
CAR and PXR are the anticonvulsant phenobarbital and the
antibiotic rifampicin, respectively. Both NRs activate genes
involved in various stages of the detoxification response: phase
1 modification (cytochrome P450 enzymes), phase 2 conjuga-
tion (glutathione S‐transferase [GST], UDP‐glucuronosyl-
transferases [UGTs], and sulfotransferases [SULTs]), and
phase 3 transport/efflux (organic anion‐transporting poly-
peptides [OATPs] and multidrug resistance proteins (MDRs]).
Classic targets for CAR are CYP2B genes, and for PXR CYP3A4,
 28:  Hepatic Nuclear Receptors 343
which encodes the most abundant human liver P450 responsible
for metabolizing over a third of clinically used drugs [28].
CAR is expressed almost exclusively in liver, while PXR is
also highly expressed in the gastrointestinal tract [6]. The litera-
ture on CAR mechanism is complex, suggesting features of
both type 1 and type 2 NRs. Like type 1 NRs, CAR can be
retained in the cytosol by HSPs and the cytoplasmic CAR reten-
tion protein (CCRP) [29], and it is reported not only to homodi-
merize but also to form an RXR heterodimer like the closely
related type 2 NR VDR. However, unlike other RXR heterodi-
mers that repress transcription in the absence of ligand, CAR/
RXR is reported to have constitutive ligand‐independent activa-
tion [30]. PXR is also an RXR heterodimer with apparent regu-
lation by nuclear localization [31].
In addition to being master regulators of the xenobiotic
response and drug metabolism, it has become clear that CAR
and PXR also have roles binding endobiotics and functioning in
normal physiology [32]. Reported endogenous ligands include
metabolites of bile acids, steroids, bilirubin, and lipids. CAR
may indirectly affect metabolic rate, as phase 2 SULT and UGT
enzymes are involved in the inactivation of thyroid hormones.
By this mechanism, CAR antagonism might be desirable, yet
other studies suggest instead that CAR activation improves met-
abolic health. Furthermore, there are conflicting results about
whether CAR activates or represses key pathways like gluco-
neogenesis and de novo lipogenesis. Both CAR and PXR have
been proposed to reduce the activation of glucose and fat syn-
thesis by other transcription factors like CREB, via an indirect
mechanism akin to transrepression. Overall, modulation of
CAR and PXR (in clinical studies with ligands, in knockout
mice, in cell models, and so on) has been consistently linked to
disturbances in energy homeostasis, but results have been mixed
and conflicted such that the precise metabolic role of xenobiotic
receptors remains quite uncertain, particularly in humans.
METABOLIC NUCLEAR RECEPTORS
Liver X and farnesoid X receptors (LXR
and FXR): sensing sterols and bile acids
The important effects of LXRs and FXR on hepatic lipid and
bile acid metabolism are described in detail in other chapters
and will be summarized here. LXRs and FXR are RXR heter-
odimer type 2 NRs, and both pathways are thought to be active
in the fed state in liver [2].
There are two isoforms of LXR, with LXRβ considered ubiq-
uitous and LXRα highest in liver, fat, gut, and immune cells, yet
the mRNAs for both are present in most tissues including liver
[6]. LXR agonists are oxysterols, oxygenated derivatives of
cholesterol including several specifically generated by cellular
hydroxylase enzymes (i.e. 24S‐, 25‐, and 27‐hydroxycholes-
terol), as well as other potent ligands of uncertain physiological
significance (i.e. 24S,25‐epoxycholesterol 22R‐ and 20S‐
hydroxylcholesterol). LXRs coordinately activate a program of
gene expression that removes cholesterol from peripheral cells
(such as macrophages in atherosclerotic lesions) via ABC1A1‐
mediated efflux to nascent HDL (APOA1 secreted by liver),
with “reverse cholesterol transport” to liver, where cholesterol is
taken up then secreted directly into bile or metabolized to bile
acids. Pharmacologic LXR agonists thus have anti‐atheroscle-
rotic effects in multiple studies, and LXRs have also been pro-
posed to have insulin‐sensitizing and anti‐inflammatory roles
[33]. LXRs also have a prominent role in stimulating hepatic de
novo lipogenesis, such that liver steatosis and hypertriglyceri-
demia have limited the therapeutic application of LXR agonists
[34]. This effect is mediated indirectly as LXR/RXR is
potent activator of SREBP‐1c, the master regulator of de novo
lipogenesis [35].
FXR expression is limited to liver, gut, kidney, and steroido-
genic tissues [6]. Only FXRα exists in humans, as FXRβ is a
pseudogene in primates (and rodents thus have 49 NR genes
compared to 48 in humans). Various bile acids serve as FXR
ligands, with the strongest natural agonist being the hydropho-
bic chenodeoxycholic acid (CDCA). Less potent agonists
include deoxycholic acid and lithocholic acid, while hydrophilic
bile acids like ursodeoxycholic acid and murocholic acids do
not activate FXR. Other sterol derivatives are also reported as
weak FXR agonists of uncertain relevance. Canonical FXR tar-
get genes in liver include the atypical NR SHP (discussed below
in the section on LRH1 and SHP) and the canalicular bile salt
export pump (ABCB11). In the distal small intestine, FXR tar-
gets include the ileal bile acid binding protein (I‐BABP) and,
importantly, the hormone FGF19 (FGF15 in rodents), which
provides negative feedback to hepatic bile acid synthesis. FXR
functions beyond bile acid metabolism have also been described
[36]. In both intestine and liver, anti‐inflammatory effects of
FXR have been reported, making FXR a candidate target in
inflammatory bowel disease and hepatitis. In the gut, FXR and
bile acids have complex crosstalk with the intestinal microbi-
ome, and in liver FXR has effects on fatty acid and glucose
metabolism. In contrast to the activation of SREBP1c and lipo-
genesis by LXR, FXR has the opposite via SHP‐mediated
repression. Indeed, FXR has been proposed to integrate liver
energy balance in the fed state, with PPARα as its counterpart in
the fasting state [37]. Bile acids and their derivatives are cur-
rently strong candidates for therapeutics, via activation of FXR
and/or the cell surface G protein coupled bile acid receptor
(GPBAR1, or TGR5). In particular, the FXR agonist obeticholic
acid is approved for use in primary biliary cholangitis, and it is
in phase 3 trials for NASH [38].
Peroxisome proliferator activated receptors
(PPARs): sensing lipids
The three PPAR nuclear receptors sense cellular lipids and regu-
late key transcriptional programs in lipid metabolism. The name
PPAR relates to the effects of PPARα agonists on rodent liver,
with induction of peroxisome proliferation, hepatomegaly, and
hepatocellular carcinoma. Notably, none of these adverse effects
occur with PPARα agonism in human liver, nor with ligands for
the other two PPARs. All three PPARs are type 2 NRs that func-
tion as RXR heterodimers to bind DR‐1 response elements. The
PPARs are expressed across multiple tissues, though PPARα is
highest in liver and PPARγ in adipose tissue [6]. The precise
endogenous ligands for PPARs remain uncertain, but there is
344 THE LIVER:  METABOLIC NUCLEAR RECEPTORS
general agreement that they are various fatty acids and their
metabolites, such as arachidonic acids and other polyunsatu-
rated fatty acids, or eicosanoids like 15‐hydroxyicosatetraenoic
acids (15‐HETE) and prostaglandins [39].
As the predominant PPAR in liver, PPARα functions as a
master regulator in hepatic lipid metabolism [40]. In the fasting
state, liver takes up free fatty acids released by lipolysis in adi-
pose tissue, and liver PPARα activates genes required for hepatic
fatty acid oxidation (FAO) and ketogenesis. FAO is also indi-
rectly required for gluconeogenesis in fasting, as the source of
cellular ATP to fuel glucose production and of mitochondrial
acetyl‐CoA, which activates pyruvate carboxylase (the first step
in gluconeogenesis). Therefore, mice deficient in PPARα appear
normal but respond poorly to fasting, with elevated serum free
fatty acids and severe hepatic steatosis, but defective ketogene-
sis and gluconeogenesis. PPAR target genes are also involved in
lipoprotein metabolism, such that PPARα fibrate drugs are used
to lower serum triglycerides. Another key transcriptional target
of PPARα is FGF21, a liver‐derived hormone that increases in
the fasted state and exerts effects throughout the body [41].
PPARγ has two protein isoforms generated by alternative
transcriptional initiation and splicing of the same gene: widely‐
expressed PPARγ1 and adipose‐selective PPARγ2, having an
additional 30 amino acids at the N‐terminus. Adipocytes express
both PPARγ1 and PPARγ2, and PPARγ is a master regulator of
fat cell differentiation. PPARγ agonist thiazolidinediones
(TZDs) are used clinically in type 2 diabetes to lower blood glu-
cose, and they function primarily in adipose tissue as insulin‐
sensitizers [39]. PPARγ1 is expressed in macrophages, thus
widespread low levels of PPARγ in some tissues may reflect
resident macrophages. While mostly PPARγ1 is present in nor-
mal liver, multiple rodent models with liver steatosis show
selective induction of the PPARγ2 isoform, consistent with a
model whereby hepatic PPARγ is lipogenic in liver like it is in
adipocytes [42]. However, while liver autonomous effects of
PPARγ may increase triglyceride storage, PPARγ agonist treat-
ment of intact animals generally reduces liver steatosis. This
may reflect the predominant effect of TZDs increasing lipid
storage in adipose tissue, thus resulting in “lipid steal” from
liver and other tissues [43]. Notably, multiple human trials have
shown that the PPARγ agonist pioglitazone improves liver his-
tology in NAFLD, and current guidelines support consideration
of pioglitazone in patients with biopsy‐proven NASH with or
without diabetes [44].
PPARδ (also called PPARβ) is widely‐expressed but best
studied in highly oxidative tissues like skeletal and cardiac mus-
cle. Studies of hepatic PPARδ are more limited and often con-
flicting, but they certainly suggest a role in regulating glucose
and lipid metabolism, overlapping with PPARα and PPARγ
[42]. However, it is clear from the phenotype of PPARα‐deficient
mice in fasting that neither PPARδ or PPARγ can rescue the key
role of PPARα in hepatic lipid metabolism.
There is great interest in drugs targeting PPARs for their met-
abolic and anti‐inflammatory effects. The clinical use of fibrate
PPARα and TZD PPARγ agonists is described above, and
PPARδ agonists have been proposed as exercise mimetics due to
their effects on muscle metabolism [45]. There are dual PPAR
agonists and pan‐PPAR agonists, as well as efforts to develop
selective and tissue‐specific drugs [46]. The dual PPARα/γ
agonist saroglitazar is used in India to treat diabetic dyslipi-
demia [47], while there is great interest in the dual PPARα/δ
agonist elafibrinor in late stage development for NASH [48].
The discussion above focused on PPARs in hepatocytes, but
they may also function in other liver cell types like HSCs related
to liver fibrosis [49].
Retinoic acid receptors (RARs): sensing
retinoic acid
RARs are activated by metabolites of vitamin A, the most potent
being all‐trans retinoic acid (ATRA). Vitamin A is a fat‐soluble
vitamin with diverse roles in development and physiology, and
~90% of total body stores are retinyl esters in lipid droplets of
HSCs. Metabolism of retinoids is complex and covered in detail
in other chapters. In liver, many classic target genes of RARs
function in retinoid metabolism, including binding proteins and
modifying enzymes.
RARα is the most abundant RAR isoform in whole liver [6],
though the RARβ and RARγ are present and could be enriched
in HSCs. Connections have been noted between vitamin A and
bile acid metabolism, with complex crosstalk among RAR,
FXR, RXR and their ligands, and bile acids required for absorp-
tion of fat soluble vitamins [50]. Liver diseases with cholestasis
are associated with vitamin A deficiency and disrupted retinoid
metabolism, and ATRA combined with bile acid therapy has
shown some promise in treating cholestasis in animal models
and a human pilot study [51].
During liver injury and fibrosis, HSCs transdifferentiate into
myofibroblasts with features of fibroblasts and smooth muscle
cells, which lose vitamin A storage capacity such that vitamin A
metabolism is disturbed in NAFLD [52]. Furthermore, the top
genetic variant associated with human liver steatosis is
PNPLA3‐I148M, and while the exact function of PNPLA3 is
unknown it has retinyl ester hydrolase activity (in addition to
triglyceride lipase activity) [53]. Therefore, vitamin A metabo-
lites also have potential roles in metabolic liver diseases, and
RAR agonists have shown favorable effects in mouse models of
NAFLD [54].
Retinoid X receptors (RXRs) and nuclear
receptor heterodimerization
Of the three RXR isoforms, RXRα expression predominates in
liver, with lower levels of RXRβ and very low or undetectable
RXRγ [6]. The best‐studied RXR ligand is 9‐cis retinoic acid,
though endogenous levels of this and other potential RXR
ligands in tissues (including liver) are a point of contention. All
three RXR isoforms can form homodimers (or heterodimers
with each other) to bind DR‐1 motifs, like the other type 3 NRs
in the NR2 family. However, the physiological roles of these
RXR homodimers, as well as the redundancy versus specific
function for each isoform, remain largely unknown.
RXRs are best known as the obligate heterodimeric partners
with the endocrine, metabolic, and xenobiotic NRs described
above. The potential benefits of heterodimerization are subject
to much teleological speculation, and RXR heterodimerization
has been described as the “big bang” in the evolution of the NR
 28:  Hepatic Nuclear Receptors 345
superfamily, as well as in the study of NRs [2]. Most RXR het-
erodimers are highly dependent on the ligand for the RXR part-
ner, while RXR ligands have variable effects. RXR heterodimers
are classified as “permissive” if RXR ligands generally activate,
or “non‐permissive” if RXR ligands appear silent [2]. Notably,
the non‐permissive group includes endocrine hormone recep-
tors like TR and VDR which respond mainly to their cognate
hormones, while the permissive group includes lipid‐sensing
NRs like LXRs, FXR, and PPARs. For these permissive heter-
odimers, synergistic activation has been observed with the com-
bination of both ligands compared to either alone, and
pharmacological pan‐RXR agonists called rexinoids have been
developed to activate this entire group [55]. In general, ligands
of a single NR can have unintended and pleiotropic effects due
to diverse and tissue‐selective gene activation, and this problem
could be compounded by activating all RXR heterodimers.
Liver receptor homolog 1 (LRH1) and small
heterodimer partner (SHP)
LRH1 is widely expressed with highest levels in liver, pancreas,
and intestine [6]. It was originally an orphan, but crystal struc-
tures revealed a large ligand binding pocket occupied by a phos-
pholipid. Mutagenesis and other studies support a model of
ligand‐independent transcriptional activation that can be
enhanced by phospholipid binding [56]. It is debated whether
phospholipid ligands simply serve a constitutive structural role,
or actively regulate LRH1 in normal physiology. Consistent
with the latter model of LRH1 as a phospholipid sensor, LRH1
was recently shown to have a key role in hepatic lipid storage
and phospholipid composition [57].
A key aspect of LRH1 function in liver is its binding to
SHP, an orphan NR with a similar tissue expression pattern.
SHP belongs to the distinctive NR0B group of NRs lacking
DNA binding domains, and is thus brought to DNA indirectly
by heterodimerization with other NRs (or even binding to
non‐NR transcription factors), where it has a repressive effect
on transcription [58]. SHP expression is highly regulated, and
activated by many NRs (particularly FXR) to mediate poten-
tial negative feedback loops. LRH1 is one of the best‐studied
binding partners for SHP, such that the two function together
to form a non‐traditional heterodimer with LRH1 providing
DNA binding and SHP providing repressive activity. A simi-
lar situation occurs in steroidogenic tissues (adrenal cortex,
testis, ovaries) with the LRH1‐related SF1 repressed by the
SHP‐related DAX1.
LRH1 regulates expression of bile acid biosynthetic
enzymes CYP7A1 and CYP8B1. It was originally proposed
that LRH1/SHP repression of CYP7A1, with SHP induced by
bile acids via FXR, was responsible for the negative feedback
of bile acids on their own synthesis [59]. Subsequent studies
showed that neither LRH1 or SHP was required for this nega-
tive feedback, indicating other possibly redundant mecha-
nisms, but loss of either NR in liver markedly alters bile acid
homeostasis [60, 61]. In addition to bile acid homeostasis,
hepatic LRH1 and SHP have also been proposed to regulate
many other genes involved in cholesterol, lipid, and glucose
metabolism (similar to LXRs, PPARs, and other NRs
described above) [56, 58].
CIRCADIAN NUCLEAR RECEPTORS
IN LIVER
Rev‐erbs: heme‐binding circadian nuclear
receptors
The two Rev‐erb NRs, of which Rev‐erbα is better studied, were
originally orphan receptors but have since been shown to bind
heme and regulate circadian rhythms [62]. Twenty‐four‐hour
diurnal rhythms are fundamental in biology, synchronizing
behavior and physiology with the day–night cycle. The elegant
molecular underpinnings of this clock have been worked out in
great detail [63]. The CLOCK/BMAL1 transcription factor is
the key activator of circadian genes, also activating the period
and cryptochrome (PER and CRY) factors which accumulate
and exert negative feedback on CLOCK/BMAL1. PER and
CRY (the “negative arm”) then decrease allowing re‐accumula-
tion of CLOCK/BMAL1 (the “positive arm”) in a 24‐hour
cycle, and this is the core circadian clock. Rev‐erbs are also acti-
vated by CLOCK/BMAL and thus have drastic diurnal rhythms
in their mRNA and protein levels, and this is not just an output
of the clock but a key input to the feedback system [64]. The
canonical target gene of Rev‐erbs is BMAL1 itself, and this tran-
scriptional repression serves as a second negative feedback
(“stabilizing loop”) on CLOCK/BMAL1 to maintain its 24‐hour
rhythm (Figure 28.3a). The central clock in the hypothalamus is
entrained by light, and mice lacking whole body Rev‐erbα and
Rev‐erbβ have profound disruptions in circadian behavior [65].
There are also autonomous peripheral clocks operating in most
tissues and cells, and Rev‐erbs are well‐studied as elements of
the hepatic clock. Rev‐erbs regulate many key genes in liver
metabolism, and liver‐specific loss of both Rev‐erbs results in
marked hepatic steatosis [66].
Uniquely among NRs, Rev‐erbs lack the C‐terminal α‐helix
12, which is required for coactivator binding, and thus serve as
transcriptional repressors by recruiting the NCoR/HDAC3 core-
pressor complex. As opposed to type 2 NRs in which ligand
stimulates corepressor release, heme binding to Rev‐erbα stabi-
lizes the interaction with NCoR to increase its repressive activ-
ity. It remains uncertain whether heme physiologically
modulates Rev‐erbs to control heme metabolism, or whether
heme binding is constitutive and structural (similar to the situa-
tion with phospholipids and LRH1 described in the section on
LRH1 and SHP). Rev‐erbs are type 4 NRs that bind as homodi-
mers or monomers to DNA, though only dimeric Rev‐erb can
recruit NCoR to actively repress genes. However, monomeric
Rev‐erb can repress genes by competing for the same half site
DNA elements with RAR‐related orphan receptors (RORs),
which are activators as described below. Repression of circadian
genes like BMAL1 in all tissues appears to involve direct Rev‐
erb binding to DNA and displacement of ROR (Figure 28.3b),
while repression of metabolic genes in liver instead often
involves indirect binding via tethering to other factors [67].
Circadian regulation of metabolism may serve to synchronize
expression of key gene products with anticipated daily periods
of rest and sleep versus activity and eating. This is a subject of
great interest, as circadian disruptions such as shift work are
associated with metabolic diseases, and alterations in circadian
gene expression occur in steatotic livers [68]. In liver and other
346 THE LIVER:  OTHER HEPATIC NUCLEAR RECEPTORS
(a)
Positive Arm
ROR
CORE
PER CLOCK BMAL1
CRY CLOCK
Stabilizing
Negative Arm
Loop
Figure 28.3  Nuclear receptors in the circadian clock. (a) There is a well‐described transcriptional–translational loop that maintains 24‐hour diur-
nal rhythms. The core clock has a positive arm mediated by BMAL1/CLOCK and a negative arm mediated by PER/CRY. This is stabilized via regu-
lation of BMAL1 by two opposing nuclear receptors groups, the Rev‐erbs and RORs. Rev‐erbs are activated by BMAL1, but then act as transcription
repressors of BMAL1 in a negative feedback loop. (b) At the BMAL1 promoter response element, the Rev‐erb repressors compete for binding with
the ROR activators. Figure adapted from [69].
tissues, Rev‐erbs provide a key link between circadian and met-
abolic pathways, such that therapeutics targeting Rev‐erb are
candidates to affect both processes.
RAR‐related orphan receptors (RORs)
RORs are closely related to Rev‐erbs, also regulating circadian
and metabolic genes in liver. RORs bind to DNA as monomers
only, occupying half sites rather than repeats, and are thought to
function solely as transcriptional activators even in the absence
of ligand. In fact, ligand binding may even decrease the basal
activation activity (inverse agonism) [69]. The identities of nat-
ural ROR ligands are controversial, with reports of binding to
various metabolites including oxysterols (like LXRs), vitamin
D metabolites (like VDR), and retinoids (like RARs). RORβ is
restricted to the central nervous system, while both RORα and
RORγ are more widely expressed including high expression in
liver. Prominent roles of these RORs have also been found in
immune cells related to autoimmune disease, particularly the
RORγt splice variant in the TH17 class of CD4 T cells [69].
In liver, both RORα and RORγ have diurnal rhythms in their
expression levels. As part of the stabilizing circadian loop
described above, the transcriptional activator RORs compete
with repressor Rev‐erbs for binding upstream of BMAL1, thus
having the opposing effects on the expression of this core clock
gene (Figure 28.3b) [70]. Similar competition occurs at other
genes, such that RORs also regulate metabolic genes in liver
[71], and drugs targeting RORs are likewise being developed
but remain far from clinical application.
Estrogen related receptors (ERRs)
and mitochondria
In mouse liver, marked circadian rhythms in expression have
been shown for all three ERRs [64]. In contrast to Rev‐erbs and
RORs, this appears not to be an input to the core clock, but
rather an output such that ERR targets also have diurnal rhythms.
ERRs are named based on homology to estrogen receptors, but
they do not bind estrogen or other steroids and remain orphan
receptors. ERRs are type 4 NRs that can bind DNA as mono-
mers or homodimers, and may even form heterodimers with two
different ERR isoforms. ERRα was the first described orphan
(b)
Rev-erb ROR
BMAL1
Rev-erb
Response
Element
NR, and it is abundant in human liver, while expression of ERRβ
and ERRγ is low or undetectable [6]. However, ERRγ is more
abundant in mouse liver, where it has been implicated in meta-
bolic pathways like gluconeogenesis [72].
ERRs have emerged as key regulators of fatty acid oxida-
tion and mitochondrial respiration, and many of these func-
tions are due to recruitment of PCG1 coactivators. Indeed,
PCG1α is well‐known to stimulate oxidative phosphoryla-
tion and mitochondrial biogenesis, and this activity requires
ERRα [73]. Therefore, ERR activation is a potential avenue
to increase energy expenditure. However, this is best studied
in highly oxidative tissues like skeletal and cardiac muscle,
yet in adipocytes ERRα can also bind the corepressor
RIP‐140 to stimulate lipid storage [74]. ERRα‐deficient
mice showed decreased body weight and resistance to high
fat diet‐induced obesity, but these mice lack the NR in all
tissues [75]. The specific role of ERRs in liver is uncertain,
and while ERR activation in other tissues may be desirable,
it appears from mouse models that hepatic ERR inhibition
improves glucose metabolism [76]. ERRs clearly play
important roles in mitochondrial energy metabolism, but
their roles as circadian factors in liver are only beginning to
be deciphered.
OTHER HEPATIC NUCLEAR RECEPTORS
Hepatocyte nuclear factor 4 (HNF4)
HNF4s are NR2As expressed mainly in liver, GI tract, kidney,
and pancreatic beta cells. Two HNF4 isoforms are encoded by
different genes, with HNF4γ predominating in small intestine
such that liver studies focus on HNF4α. The name HNF reflects
the discovery of a diverse and unrelated group of transcription
factors as liver proteins with DNA binding activity, but HNF4s
are the only NRs among these (HNF1s are POU‐homeodomain
proteins, HNF3s are winged helix forkhead box proteins, and
HNF6s are ONECUT proteins). HNF4s bind DNA as homodi-
mers to DR‐1 response elements [77]. HNF4s are considered
orphan nuclear receptors, though reversible binding of linoleic
acid has been reported without apparent effects on transcrip-
tional activity [78].
 28:  Hepatic Nuclear Receptors 347
In adult liver, HNF4α is thought to maintain hepatocyte dif-
ferentiated gene expression patterns. Mice completely lacking
HNF4α show embryonic lethality, likely due to the requirement
for HNF4α in liver development [79]. Liver‐specific loss of
HNF4α in adult livers causes hepatic steatosis, with increase
serum bile acids but reduced cholesterol and triglycerides, all
attributable to reduced hepatic expression of key genes in lipid
and bile acid metabolism [80]. Consistent with this is a small
study in people with heterozygous mutations in HNF4α, which
causes maturity onset diabetes of the young type 1 (MODY1)
due to effects in beta cells. While MODY1 does not have a clear
liver phenotype, subjects with a MODY1 mutation had reduced
serum levels of several liver‐derived apolipoproteins, consistent
with HNF4 maintaining expression of hepatocyte genes [81].
Even though HNF4 is mainly thought to maintain ligand‐inde-
pendent constitutive expression of these genes, it could cooper-
ate with other cofactors or NRs to mediate regulated expression
[82]. One example is the recent demonstration that HNF4α and
another transcription factor (PROX1) co‐recruit the HDAC3
corepressor complex to lipid metabolic genes [83], and another
is the drug‐metabolizing CYP3A4 gene which requires HNF4α
for its activation by xenobiotic NRs [84]. Other examples of
HNF4 crosstalk have been described for genes involved in bile
acid, sterol, fatty acid, and glucose metabolism, raising the pos-
sibility that HNF4 may yet emerge as a drug target [85].
The testicular receptor (TR) NR2C subfamily
Related to HNF4s are two NR2C family members also known
as testicular receptors: TR2 is NR2C1 and TR4 (also called
TAK1) is NR2C2. Despite the TR name, both are widely
expressed including in the liver [6]. Mice deficient in NR2C1
have normal testis and spermatogenesis, mice deficient in
NR2C2 have reduced sperm and other systemic problems, and
mice lacking both factors die in early embryonic development
[86]. Both are considered orphan receptors, though NR2C2 has
been reported to bind retinoids or fatty acids [87], consistent
with homology to RXRs and HNF4s. The two NR2Cs may
homodimerize or heterodimerize with one another and repress
transcription. Genome‐wide binding profiles of NR2C2 in four
human cell lines, including HepG2 liver carcinoma cells, sug-
gested roles in fundamental cellular processes like RNA metab-
olism and protein translation, rather than specific metabolic
pathways [88]. However, in mouse liver, NR2C2 has been
implicated in regulation of lipid metabolism particularly
stearoyl‐CoA desaturase 1 (SCD1) [89]. NR2C2 has also been
shown to be cAMP‐induced during fasting with a potential role
in hepatic gluconeogenesis [90]. Finally, NR2C2‐deficient mice
are protected against high fat diet‐induced metabolic disease
[91]. Similar to HNF4s, crosstalk between NR2Cs and other
NRs has also been proposed, particularly those which also bind
DR‐1 response elements like PPARs [92].
NR2F chicken ovalbumin upstream promoter‐
transcription factors (COUP‐TFs)
Also in the NR2 family related to HNF4s, TRs, and RXRs are the
three NR2Fs or COUP‐TF family (COUP‐TFI, COUP‐TFII, and
EAR2), all of which are widely expressed including in liver [6].
The COUP name is historical as these proteins were purified
based on binding to a chicken regulatory DNA. NR2F family
members form homodimers, but heterodimers are also reported
with each other and with RXR [93]. These NRs bind to direct
repeats of various spacing, and they function primarily as tran-
scriptional repressors. They are considered orphan receptors,
though their ligand binding domains are similar to RARs and
RXRs, and retinoic acid is a reported ligand COUP‐TFII [94].
COUP‐TFII is the best‐studied NR2F and strongly implicated in
many developmental processes. In contrast, COUP‐TFI and EAR2
are primarily studied in the nervous and immune systems respec-
tively, and little to nothing is reported about their roles in liver.
While COUP‐TFII deletion in adult mice does not give a discern-
able phenotype, roles in energy metabolism are nonetheless emerg-
ing [95]. For instance, hepatic COUP‐TFII was recently shown to
be up‐regulated by fasting and play a role in gluconeogenesis and
fatty acid oxidation [96]. Consistent with this, COUP‐TFII has
complex physical and functional interactions with HNF4α, PPARα,
and particularly GR to regulate metabolic genes [95].
The NR4A subfamily
The final group of hepatic NRs is the three closely related
NR4As, also called nerve growth factor IB (NGFIB, NUR77 or
NR4A1), nuclear receptor related 1 (NURR1, or NR4A2), and
neuron‐derived orphan receptor 1 (NOR1, or NR4A3). While
two of these historical names relate to neurons, the three NR4As
are widely expressed including in liver [6]. All have been
reported to bind DNA as monomers, and NR4A1 and NR4A2 as
RXR heterodimers [97]. No ligands are known or expected to be
found, as they are thought to be ligand‐independent NRs based
on the NR4A2 crystal structure showing a constitutively active
conformation with hydrophobic side‐chains filling the ligand‐
binding pocket. Rather than ligands, NR4A family members are
regulated by their expression levels. They are immediate early
response genes rapidly induced by environmental changes, and
implicated in diverse cellular processes [98].
In liver, all three NR4As are induced by cAMP, which is
downstream of glucagon and catecholamine signaling in fasting
[99]. These hormones trigger hepatic gluconeogenesis, and
NR4A members were strongly implicated in this response.
Overexpression of NR4A1 in mouse liver could mimic fasting
and activate gluconeogenic genes, while a dominant negative
NR4A1 (that antagonizes all three NR4As) could inhibit gluco-
neogenic genes and lower glucose in a diabetic model. In addi-
tion to this role in gluconeogenesis, NR4A1 has also been
implicated in hepatic lipid metabolism, so NR4A1 and it family
members thus could emerge as therapeutic targets [100].
OTHER ISSUES IN HEPATIC NUCLEAR
RECEPTOR BIOLOGY
While there are 48 NR genes, the NR proteome is even more
diverse. Most NR genes generate multiple mRNAs due to alter-
native transcriptional initiation and splicing (as described above
for PPARγ), and the proteins are subject to various reported
post‐translational modifications (phosphorylation, acetylation,
348 THE LIVER:  REFERENCES
SUMOylation, etc.) [101]. The relevance of these in liver is gen-
erally poorly understood and beyond the scope here.
A recent and profound advance in the study of NRs is the
advent of next generation sequencing approaches to identify their
DNA binding sites genome‐wide (cistromes), such as chromatin
immunoprecipitation followed by deep sequencing (ChIP‐seq).
Whereas previous efforts had identified only dozens of NR bind-
ing sites, mainly at response elements in promoters of well‐stud-
ied targets, ChIP‐seq typically identifies tens of thousands of
binding sites throughout the genome and near most expressed
genes. As described above, both direct binding and indirect teth-
ering is observed. Other unbiased methods like gene expression
profiling (transcriptomes, for instance in response to NR ligands,
knockdown, or overexpression) have also drastically increased
the number of potential targets genes, but not to the same extent.
Comparisons of cistromes and transcriptomes make it clear that
not all binding events have apparent functional consequences on
gene expression. Furthermore, regulated transcripts represent a
mixture of direct and indirect targets, regulated in both directions
(such that genes can be indirectly downregulated by ligands for
activating NR, and upregulated by a repressive NR). Advances in
genomic and computational methods have thus opened new win-
dows into the complexity of NR function [2].
Most regulatory elements occupied by NRs do not simply con-
sist of that single NR response element motif, but also binding
motifs for other hepatic transcription factors including other NRs.
The composite nature of these cis‐acting sites theoretically allows
distinct multiprotein complexes at different elements to fine‐tune
regulation based on crosstalk from multiple inputs. This complex
transcriptional regulation by crosstalk at composite elements is
only beginning to be understood. Given the abundance of impor-
tant hepatic NRs and their frequent co‐occupancy of regulatory
elements at target genes, the liver represents an excellent physio-
logical system to investigate these important issues.
This chapter focused on NRs themselves, but there are also
hundreds of NR coregulators, which do not bind DNA directly
but are recruited by transcription factors. They serve to modulate
the rate of transcription of nearby genes by various mechanisms,
such as altering chromatin structure (SWI/SNF), histone modifi-
cations (acetyl transferase [HAT] and deacylatase [HDAC] enzy-
matic activity), recruiting the basal transcription machinery with
RNA polymerase 2, and facilitating transcriptional elongation
versus pausing. A well‐studied co‐repressor of many NR target
genes is the NCoR/HDAC3 complex [102], while prominent co‐
activators include the histone acetyltransferase p300/CBP, the
steroid receptor coactivators (SRCs) family, and PGC1s [103].
Like NRs, manipulations of their co‐regulators can likewise affect
liver metabolic homeostasis, but targeting of co‐regulators thera-
peutically is even more complex. It as an area of active investiga-
tion to elucidate mechanisms by which NRs and co‐regulators
function at the genome to regulate transcription.
NUCLEAR RECEPTORS IN LIVER DISEASE
NRs have many potential roles in liver disease, as noted above.
Targeting NRs like VDR and RARs in HSCs may have antifibrotic
roles in cirrhosis, while targeting bile acid‐related NRs like
FXR, LRH1, and SHP could treat cholestatic diseases. Many
NRs exert anti‐inflammatory effects that could be relevant in
hepatitis. However, most hepatic NRs have clear roles in energy
metabolism, particularly lipid homeostasis. It is notable that,
while there are no approved therapeutics for NAFLD, one rec-
ommended for off‐label use in NASH is the PPARγ agonist
pioglitazone [44], and two agents most advanced in phase 3
­trials specifically for NASH are other metabolic NR ligands:
obeticholic acid targeting FXR [38], and elafibrinor targeting
PPARα/δ [48].
Human genetics has also implicated various NRs in disease
[104]. Rare PPARγ mutations cause familial partial lipodystro-
phy with severe liver steatosis, while FXR mutations cause
­progressive familial intrahepatic cholestasis. However, coding
mutations in NRs causing loss‐of‐function are expected to be
rare given the requirement of many NRs for normal develop-
ment and homeostasis. More common may be non‐coding
genetic variants in NR binding sites that affect regulation of spe-
cific target genes, rather than global NR activity. Such noncod-
ing regulatory variants would be predicted to have subtle effects
on complex disease risk rather than causing Mendelian syn-
dromes. Indeed, in human genome‐wide association studies
(GWAS), over 95% of causal genetic variants that are associated
with phenotype fall in noncoding regions, and these are thought
to function by affecting regulatory DNA. Human single nucleo-
tide polymorphisms (SNPs) that alter response elements for
NRs, as described for HNF4 in liver [85] and PPARγ in adipose
tissue [105], could determine differential target gene regulation
among individuals. Such variants would be relevant to precision
medicine approaches for individualized disease prediction and
treatments.
CONCLUSION
Liver NRs play key roles in homeostasis, mediating many
essential hepatic functions. NRs regulate the metabolism of glu-
cose, fatty acids, cholesterol, and bile acids, as well as the
detoxification and clearance of xenobiotics. By sensing hor-
mones, metabolites, and drugs, many NRs allow the liver to
respond appropriately to environmental changes. Other circa-
dian NRs allow the liver to anticipate such changes based on
diurnal rhythms. Existing and novel drugs targeting NRs will
play an increasing role in the treatment of liver diseases.
REFERENCES
Note: Important primary references were regretfully omitted due to constraints of
space. Many of the excellent review articles below include relevant citations of
primary literature.
1. Lazar, M.A. Maturing of the nuclear receptor family. J Clin Invest, 2017;
127(4):1123–5.
2. Evans, R.M., Mangelsdorf, D.J. Nuclear receptors, RXR, and the big bang.
Cell, 2014:157(1):255–66.
3. Mangelsdorf, D.J., Thummel, C., Beato, M. et al. The nuclear receptor super-
family: the second decade. Cell, 1995;83(6):835–9.
4. Nuclear Receptors Nomenclature Committee. A unified nomenclature system
for the nuclear receptor superfamily. Cell, 1999;97(2):161–3.
 28:  Hepatic Nuclear Receptors 349
5. Wang, H., Chen, J., Hollister, K., Sowers, L.C., and Forman, B.M.
Endogenous bile acids are ligands for the nuclear receptor FXR/BAR. Mol
Cell, 1999;3(5):543–53.
6. The Human Protein Atlas. https://www.proteinatlas.org/ (Accessed Feb 26,
2019).
7. Melé, M., Ferreira, P.G., Reverter, F. et al. Human genomics. The human tran-
scriptome across tissues and individuals. Science, 2015;348(6235):660–5.
8. Malik, R. and Hodgson, H. The relationship between the thyroid gland and
the liver. QJM Int J Med, 2002;95(9):559–69.
9. Sinha, R.A., Singh, B.K., and Yen, P.M. Direct effects of thyroid hormones
on hepatic lipid metabolism. Nat Rev Endocrinol, 2018;14(5):259–69.
10. Maia, A.L., Goemann, I.M., Meyer, E.L.S., and Wajner, S.M. Deiodinases:
the balance of thyroid hormone: type 1 iodothyronine deiodinase in human
physiology and disease. J Endocrinol, 2011;209(3):283–97.
11. Schussler, G.C. The thyroxine‐binding proteins. Thyroid, 2000;10(2):141–9.
12. Kuo, T., McQueen A., Chen T‐C., and Wang J‐C. Regulation of glucose home-
ostasis by glucocorticoids, in: Glucocorticoid Signaling: From Molecules to
Mice to Man, (eds. J‐C. Wang and C. Harris), Springer New York, New York,
NY, 2015, pp. 99–126. doi.org/10.1007/978‐1‐4939‐2895‐8_5.
13. Woods, C.P., Hazlehurst, J.M., and Tomlinson, J.W. Glucocorticoids and non‐
alcoholic fatty liver disease. J Steroid Biochem Mol Biol, 2015;154:94–103.
14. Newton, R. and Holden, N.S. Separating transrepression and transactivation:
a distressing divorce for the glucocorticoid receptor? Mol Pharmacol,
2007;72(4):799–809.
15. Cole, T.J. and Young, M.J. 30 years of the mineralocorticoid receptor: min-
eralocorticoid receptor null mice: informing cell‐type‐specific roles. J
Endocrinol, 2017;234(1):T83–92.
16. Ma, W‐L., Lai, H‐C., Yeh, S., Cai, X., and Chang, C. Androgen receptor
roles in hepatocellular carcinoma, fatty liver, cirrhosis and hepatitis. Endocr
Relat Cancer, 2014;21(3):R165–182.
17. Brie, B., Ramirez, M.C., De Winne C. et al. Brain control of sexually dimor-
phic liver function and disease: the endocrine connection. Cell Mol
Neurobiol, 2019;39(2):169–80.
18. Barros, R.P.A. and Gustafsson, J.‐Å. Estrogen receptors and the metabolic
network. Cell Metab, 2011;14(3):289–99.
19. Qiu, S., Vazquez, J.T., Boulger, E. et al. Hepatic estrogen receptor α is criti-
cal for regulation of gluconeogenesis and lipid metabolism in males. Sci Rep,
2017;7(1):1661.
20. Han, S., Komatsu, Y., Murayama, A. et al. Estrogen receptor ligands amelio-
rate fatty liver through a nonclassical estrogen receptor/liver X receptor path-
way in mice. Hepatology, 2014;59(5):1791–802.
21. Umetani, M. and Shaul, P.W. 27‐Hydroxycholesterol: the first identified
endogenous SERM. Trends Endocrinol Metab TEM, 2011;22(4):130–5.
22. Lin, H‐Y., Yu, I‐C., Wang, R‐S. et al. Increased hepatic steatosis and insulin
resistance in mice lacking hepatic androgen receptor. Hepatology,
2008;47(6):1924–35.
23. Krycer, J.R. and Brown, A.J. Cross‐talk between the androgen receptor and
the liver x receptor implications for cholesterol homeostasis. J Biol
Chem,2011;286(23):20637–47.
24. Thaler, M.A., Seifert‐Klauss, V. and Luppa, P.B. The biomarker sex hor-
mone‐binding globulin ‐ from established applications to emerging trends in
clinical medicine. Best Pract Res Clin Endocrinol Metab, 2015;29(5):749–60.
25. Gascon‐Barré, M., Demers, C., Mirshahi, A., Néron S., Zalzal, S., and
Nanci, A. The normal liver harbors the vitamin D nuclear receptor in non-
parenchymal and biliary epithelial cells. Hepatology, 2003;37(5):1034–42.
26. Ding, N., Liddle, C., Evans, R.M., and Downes, M. Hepatic actions of
Vitamin D receptor ligands: an unexpected solution to chronic liver disease?
Expert Rev Clin Pharmacol, 2013;6(6):597–9.
27. Yan, J. and Xie, W. A brief history of the discovery of PXR and CAR as
xenobiotic receptors. Acta Pharm Sin B, 2016;6(5):450–2.
28. Zanger, U.M., Turpeinen, M., Klein, K., Schwab, M. Functional pharmaco-
genetics/genomics of human cytochromes P450 involved in drug biotrans-
formation. Anal Bioanal Chem, 2008;392(6):1093–108.
29. Timsit, Y.E. and Negishi, M. Coordinated regulation of nuclear receptor
CAR by CCRP/DNAJC7, HSP70 and the ubiquitin–proteasome system.
PloS One, 2014;9(5):e96092.
30. Suino, K., Peng, L., Reynolds, R. et  al. The nuclear xenobiotic receptor
CAR: structural determinants of constitutive activation and heterodimeriza-
tion. Mol Cell, 2004;16(6):893–905.
31. Squires, E.J., Sueyoshi, T., and Negishi, M. Cytoplasmic localization of
pregnane X receptor and ligand‐dependent nuclear translocation in mouse
liver. J Biol Chem, 2004;279(47):49307–14.
32. Mackowiak, B., Hodge, J., Stern, S., and Wang, H. The roles of xenobiotic
receptors: beyond chemical disposition. Drug Metab Dispos Biol Fate Chem,
2018;46(9):1361–71.
33. Ma, Z., Deng, C., Hu, W. et al. Liver X receptors and their agonists: targeting
for cholesterol homeostasis and cardiovascular diseases. Curr Issues Mol
Biol, 2017;22:41–64.
34. Fessler, M.B. The challenges and promise of targeting the liver X receptors
for treatment of inflammatory disease. Pharmacol Ther, 2018;181:1–12.
35. Schultz, J.R., Tu, H., Luk, A. et al. Role of LXRs in control of lipogenesis.
Genes Dev, 2000;14(22):2831–8.
36. Lee, F.Y., Lee, H., Hubbert, M.L., Edwards, P.A., and Zhang, Y. FXR, a mul-
tipurpose nuclear receptor. Trends Biochem Sci, 2006;31(10):572–80.
37. Preidis, G.A., Kim, K.H., and Moore, D.D. Nutrient‐sensing nuclear recep-
tors PPARα and FXR control liver energy balance. J Clin Invest,
2017;127(4):1193–201.
38. Wang, H., He, Q., Wang, G., Xu, X., and Hao, H. FXR modulators for entero-
hepatic and metabolic diseases. Expert Opin Ther Pat, 2018;28(11):765–82.
39. Soccio, R.E., Chen, E.R., and Lazar, M.A. Thiazolidinediones and the promise
of insulin sensitization in type 2 diabetes. Cell Metab, 2014;20(4):573–91.
40. Kersten, S. Integrated physiology and systems biology of PPARα. Mol
Metab, 2014;3(4):354–71.
41. Vernia, S., Cavanagh‐Kyros, J., Garcia‐Haro, L. et al. The PPARα‐FGF21
hormone axis contributes to metabolic regulation by the hepatic JNK signal-
ing pathway. Cell Metab, 2014;20(3):512–25.
42. Chen, J., Montagner, A., Tan, N.S., and Wahli, W. Insights into the role of
PPARβ/δ in NAFLD. Int J Mol Sci, 2018;19(7).
43. Ye, J.‐M., Dzamko, N., Cleasby, M.E. et al. Direct demonstration of lipid
sequestration as a mechanism by which rosiglitazone prevents fatty‐acid‐
induced insulin resistance in the rat: comparison with metformin.
Diabetologia, 2004;47(7):1306–13.
44. Chalasani, N., Younossi, Z., Lavine, J.E. et al. The diagnosis and manage-
ment of non‐alcoholic fatty liver disease: Practice guidance from the
American Association for the Study of Liver Diseases. Hepatol Baltim Md,
2018;67(1):328–57.
45. Fan, W., Waizenegger, W., Lin, C.S. et al. PPARδ Promotes running endur-
ance by preserving glucose. Cell Metab, 2017;25(5):1186–93.
46. Heald, M. and Cawthorne, M.A. Dual acting and pan‐PPAR activators as
potential anti‐diabetic therapies. Handb Exp Pharmacol, 2011;
(203):35–51.
47. Joshi, S.R. Saroglitazar for the treatment of dyslipidemia in diabetic patients.
Expert Opin Pharmacother, 2015;16(4):597–606.
48. Ratziu, V., Harrison, S.A., Francque, S. et al. Elafibranor, an agonist of the
peroxisome proliferator‐activated receptor‐α and ‐δ, induces resolution of
nonalcoholic steatohepatitis without fibrosis worsening. Gastroenterology,
2016;150(5):1147–1159.e5.
49. Zardi, E.M., Navarini, L., Sambataro, G. et al. Hepatic PPARs: their role in liver
physiology, fibrosis and treatment. Curr Med Chem,. 2013;20(27):3370–96.
50. Saeed, A., Hoekstra, M., Hoeke, M.O., Heegsma, J., and Faber, K.N. The
interrelationship between bile acid and vitamin A homeostasis. Biochim
Biophys Acta Mol Cell Biol Lipids, 2017;1862(5):496–512.
51. Assis, D.N., Abdelghany, O., Cai, S.‐Y. et al. Combination therapy of all‐
trans retinoic acid with ursodeoxycholic acid in patients with primary scle-
rosing cholangitis: a human pilot study. J Clin Gastroenterol, 2017;
51(2):e11–6.
52. Saeed, A., Dullaart, R.P.F., Schreuder, T.C.M.A., Blokzijl, H., and Faber,
K.N. Disturbed vitamin A metabolism in non‐alcoholic fatty liver disease
(NAFLD). Nutrients, 2017;10(1).
53. Trépo, E., Romeo, S., Zucman‐Rossi, J., and Nahon, P. PNPLA3 gene in
liver diseases. J Hepatol, 2016;65(2):399–412.
54. Trasino, S.E., Tang X‐H., Jessurun, J., and Gudas, L.J. A retinoic acid recep-
tor β2 agonist reduces hepatic stellate cell activation in non‐alcoholic fatty
liver disease. J Mol Med Berl Ger, 2016;94(10):1143–51.
55. Wagner, C.E., Jurutka, P.W., Marshall, P.A., and Heck, M.C. Retinoid X
receptor selective agonists and their synthetic methods. Curr Top Med Chem,
2017;17(6):742–67.
56. Stein, S. and Schoonjans, K. Molecular basis for the regulation of the nuclear
receptor LRH‐1. Curr Opin Cell Biol, 2015;33:26–34.
57. Miranda, D.A., Krause, W.C., Cazenave‐Gassiot, A. et al. LRH‐1 regulates
hepatic lipid homeostasis and maintains arachidonoyl phospholipid pools
critical for phospholipid diversity. JCI Insight, 2018;3(5).
58. Zhang, Y., Hagedorn, C.H., and Wang, L. Role of nuclear receptor SHP in
metabolism and cancer. Biochim Biophys Acta, 2011;1812(8):893–908.
350 THE LIVER:  REFERENCES
59. Goodwin, B., Jones, S.A., Price, R.R. et  al. A regulatory cascade of the
nuclear receptors FXR, SHP‐1, and LRH‐1 represses bile acid biosynthesis.
Mol Cell, 2000;6(3):517–26.
60. Lee, Y‐K., Schmidt, D.R., Cummins, C.L. et al. Liver receptor homolog‐1
regulates bile acid homeostasis but is not essential for feedback regulation of
bile acid synthesis. Mol Endocrinol, 2008;22(6):1345–56.
61. Kerr, T.A., Saeki, S., Schneider, M. et  al. Loss of nuclear receptor SHP
impairs but does not eliminate negative feedback regulation of bile acid syn-
thesis. Dev Cell, 2002;2(6):713–20.
62. Yin, L., Wu, N., and Lazar, M.A. Nuclear receptor Rev‐erbalpha: a heme
receptor that coordinates circadian rhythm and metabolism. Nucl Recept
Signal, 2010;8:e001.
63. Huang, R.‐C. The discoveries of molecular mechanisms for the circadian
rhythm: the 2017 Nobel Prize in physiology or medicine. Biomed J,
2018;41(1):5–8.
64. Yang, X., Downes, M., Yu, R.T. et al. Nuclear receptor expression links the
circadian clock to metabolism. Cell, 2006;126(4):801–10.
65. Cho, H., Zhao, X., Hatori, M. et al. Regulation of circadian behaviour and
metabolism by REV‐ERB‐α and REV‐ERB‐β. Nature,
2012;485(7396):123–7.
66. Bugge, A., Feng, D., Everett, L.J. et al. Rev‐erbα and Rev‐erbβ coordinately
protect the circadian clock and normal metabolic function. Genes Dev,
2012;26(7):657–67.
67. Fang, B. and Lazar, M.A. Dissecting the Rev‐erbα cistrome and the mecha-
nisms controlling circadian transcription in liver. Cold Spring Harb Symp
Quant Biol, 2015;80:233–8.
68. Gerhart‐Hines, Z. and Lazar, M.A. Rev‐erbα and the circadian transcriptional
regulation of metabolism. Diabetes Obes Metab, 2015;17 Suppl 1:12–6.
69. Solt, L.A. and Burris, T.P. Action of RORs and their ligands in (patho)physi-
ology. Trends Endocrinol Metab, 2012;23(12):619–27.
70. Sato, T.K., Panda, S., Miraglia, L.J. et  al. A functional genomics strategy
reveals Rora as a component of the mammalian circadian clock. Neuron,
2004;43(4):527–37.
71. Zhang, Y., Papazyan, R., Damle, M. et al. The hepatic circadian clock fine‐
tunes the lipogenic response to feeding through RORα/γ. Genes Dev,
2017;31(12):1202–11.
72. Kim, D‐K., Ryu, D., Koh, M. et al. Orphan nuclear receptor estrogen‐related
receptor γ (ERRγ) is key regulator of hepatic gluconeogenesis. J Biol Chem,
2012;287(26):21628–39.
73. Schreiber, S.N., Emter, R., Hock, M.B. et al. The estrogen‐related receptor
alpha (ERRalpha) functions in PPARgamma coactivator 1alpha
(PGC‐1alpha)‐induced mitochondrial biogenesis. Proc Natl Acad Sci U S A,
2004;101(17):6472–7.
74. Hummasti, S. and Tontonoz, P. Adopting new orphans into the family of
metabolic regulators. Mol Endocrinol Baltim Md, 2008;22(8):1743–53.
75. Luo,, J., Sladek, R., Carrier, J., Bader J‐A., Richard, D., and Giguère, V.
Reduced fat mass in mice lacking orphan nuclear receptor estrogen‐related
receptor alpha. Mol Cell Biol, 2003;23(22):7947–56.
76. Audet‐walsh, É. and Giguére, V. The multiple universes of estrogen‐related
receptor α and γ in metabolic control and related diseases. Acta Pharmacol
Sin, 2015;36(1):51–61.
77. Fang, B., Mane‐Padros, D., Bolotin, E., Jiang, T., and Sladek, F.M.
Identification of a binding motif specific to HNF4 by comparative analysis
of multiple nuclear receptors. Nucleic Acids Res, 2012;40(12):5343–56.
78. Yuan, X., Ta, T.C., Lin, M. et  al. Identification of an endogenous ligand
bound to a native orphan nuclear receptor. PloS One, 2009;4(5):e5609.
79. Li, J., Ning, G., and Duncan, S.A. Mammalian hepatocyte differentiation
requires the transcription factor HNF‐4α. Genes Dev, 2000;14(4):464–74.
80. Hayhurst, G.P., Lee, Y.H., Lambert, G., Ward, J.M., and Gonzalez, F.J.
Hepatocyte nuclear factor 4alpha (nuclear receptor 2A1) is essential for
maintenance of hepatic gene expression and lipid homeostasis. Mol Cell
Biol, 2001;21(4):1393–403.
81. Shih, D.Q., Dansky, H.M., Fleisher, M., Assmann, G., Fajans, S.S., and
Stoffel, M. Genotype/phenotype relationships in HNF‐4alpha/MODY1:
haploinsufficiency is associated with reduced apolipoprotein (AII), apolipo-
protein (CIII), lipoprotein(a), and triglyceride levels. Diabetes, 2000;
49(5):832–7.
82. Gonzalez, F.J. Regulation of hepatocyte nuclear factor 4 alpha‐mediated
transcription. Drug Metab Pharmacokinet, 2008;23(1):2–7.
83. Armour, S.M., Remsberg, J.R., Damle, M. et al. An HDAC3‐PROX1 core-
pressor module acts on HNF4α to control hepatic triglycerides. Nat
Commun, 2017;8(1):549.
84. Tirona, R.G., Lee, W., Leake, B.F. et  al. The orphan nuclear receptor
HNF4alpha determines PXR‐ and CAR‐mediated xenobiotic induction of
CYP3A4. Nat Med, 2003;9(2):220–4.
85. Hwang‐Verslues, W.W. and Sladek, F.M. HNF4α‐‐role in drug metabolism
and potential drug target? Curr Opin Pharmacol, 2010;10(6):698–705.
86. Shyr, C.‐R., Kang, H.‐Y., Tsai, M.‐Y. et  al. Roles of testicular orphan
nuclear receptors 2 and 4 in early embryonic development and embryonic
stem cells. Endocrinology, 2009;150(5):2454–62.
87. Zhou, X.E., Suino‐Powell, K.M., Xu, Y. et al. The orphan nuclear receptor
TR4 is a vitamin A‐activated nuclear receptor. J Biol Chem,
2011;286(4):2877–85.
88. O’Geen H., Lin, Y.‐H., Xu, X. et al. Genome‐wide binding of the orphan
nuclear receptor TR4 suggests its general role in fundamental biological
processes. BMC Genomics, 2010;11:689.
89. Kim, E., Liu, N.‐C., Yu, I.‐C. et  al. Metformin inhibits nuclear receptor
TR4‐mediated hepatic stearoyl‐CoA desaturase 1 gene expression with
altered insulin sensitivity. Diabetes, 2011;60(5):1493–503.
90. Liu, N.‐C., Lin, W.‐J., Yu, I.‐C. et  al. Activation of TR4 orphan nuclear
receptor gene promoter by cAMP/PKA and C/EBP signaling. Endocrine,
2009;36(2):211–7.
91. Kang, H.S., Okamoto, K., Kim Y‐S. et al. Nuclear orphan receptor TAK1/
TR4‐deficient mice are protected against obesity‐linked inflammation,
hepatic steatosis, and insulin resistance. Diabetes, 2011;60(1):177–88.
92. Lin, S.‐J., Yang, D.‐R., Yang, G. et al. TR2 and TR4 orphan nuclear recep-
tors: an overview. Curr Top Dev Biol, 2017;125:357–73.
93. Avram, D., Ishmael, J.E., Nevrivy, D.J. et  al. Heterodimeric interactions
between chicken ovalbumin upstream promoter‐transcription factor family
members ARP1 and ear2. J Biol Chem, 1999;274(20):14331–6.
94. Kruse, S.W., Suino‐Powell, K., Zhou, X.E. et al. Identification of COUP‐
TFII orphan nuclear receptor as a retinoic acid‐activated receptor. PLoS
Biol, 2008;6(9):e227.
95. Ashraf, U.M., Sanchez, E.R. and Kumarasamy, S. COUP‐TFII revisited: Its
role in metabolic gene regulation. Steroids, 2019;141:63–9.
96. Planchais, J., Boutant, M., Fauveau, V. et al. The role of chicken ovalbumin
upstream promoter transcription factor II in the regulation of hepatic fatty
acid oxidation and gluconeogenesis in newborn mice. Am J Physiol
Endocrinol Metab. 2015;308(10):E868–78.
97. Kurakula, K., Koenis, D.S., van Tiel, C.M., and de Vries, C.J.M. NR4A
nuclear receptors are orphans but not lonesome. Biochim Biophys Acta,
2014;1843(11):2543–55.
98. Maxwell, M.A. and Muscat GE.O. The NR4A subgroup: immediate early
response genes with pleiotropic physiological roles. Nucl Recept Signal,
2006;4:e002.
99. Pei, L., Waki, H., Vaitheesvaran, B., Wilpitz, D.C., Kurland, I.J., and
Tontonoz, P. NR4A orphan nuclear receptors are transcriptional regulators
of hepatic glucose metabolism. Nat Med, 2006;12(9):1048–55.
100. Zhang, L., Wang, Q., Liu, W., Liu, F., Ji, A., and Li, Y. The orphan nuclear
receptor 4A1: a potential new therapeutic target for metabolic diseases. J
Diabetes Res, 2018;2018:9363461.
101. Berrabah, W., Aumercier, P., Lefebvre, P., and Staels, B. Control of nuclear
receptor activities in metabolism by post‐translational modifications. FEBS
Lett, 2011;585(11):1640–50.
102. Emmett, M.J. and Lazar, M.A. Integrative regulation of physiology by his-
tone deacetylase 3. Nat Rev Mol Cell Biol, 2019;20(2):102–15.
103. Dasgupta, S., Lonard, D.M., and O’Malley, B.W. Nuclear receptor coacti-
vators: master regulators of human health and disease. Annu Rev Med,
2014;65:279–92.
104. Achermann, J.C., Schwabe, J., Fairall, L., and Chatterjee, K. Genetic disor-
ders of nuclear receptors. J Clin Invest, 127(4):1181–92.
105. Soccio, R.E., Chen, E.R., Rajapurkar, S.R. et al. Genetic variation deter-
mines PPARγ function and anti‐diabetic drug response in vivo. Cell.
2015;162(1):33–44.
 Molecular Cholestasis
29 Paul Gissen1 and Richard J. Thompson2
1
UCL Great Ormond Street Institute of Child Health and Great Ormond Street Hospital for Children,
London, UK
2
Institute of Liver Studies, King’s College London, UK
The molecular bases of cholestasis have been unraveled through
the detailed investigation of individual patients and cohort
studies. The most obvious examples have resulted from
­
genomic and molecular biology investigation. Major scientific
advances that participated in this increase included elucidation
of the cellular mechanisms of synthesis and secretion of vari-
ous constituents of bile, facilitated by improved genomic meth-
ods that have led to identification of genes mutated in patients
with several familial forms of cholestasis. Discovery of the
genetic causes in patients with severe clinical manifestations
has gradually led to an appreciation of the scope of the phe-
nomena associated with abnormalities in individual genes.
More recently, the aim of clinical research has been to elucidate
how genes mutated in severe forms of childhood cholestasis
contribute to multifactorial, later onset, liver diseases. In this
chapter we discuss the molecular and clinical features of the
heritable forms of cholestasis.
INHERITED FORMS OF INTRAHEPATIC
CHOLESTASIS
Cholic acid (CA), chenodeoxycholic acid (CDCA), the primary
bile acids (BAs), are synthesized in the liver from cholesterol by
a number of steps occurring in several organelles, including
­peroxisomes. Failure of primary BA synthesis leads to accumu-
lations of toxic intermediate metabolites, with hepatocellular
damage and cholestasis [1]. For this reason, cholestasis can be a
feature of single enzyme defects and peroxisomal biogenesis
disorders. Some defects of cholesterol biogenesis, such as 3β‐
hydroxysteroid Δ7‐reductase deficiency, also may cause choles-
tasis due to production of abnormal BA [2, 3].
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
Disorders of primary BA synthesis and
conjugation presenting with infantile
cholestasis
Deficiencies of enzymes involved in primary BA synthesis
cause some instances of neonatal cholestasis in which serum
gamma‐glutamyl transpeptidase (GGT) values do not rise,
despite conjugated hyperbilirubinemia (see Table  29.1). As at
least 16 enzymes are involved in conversion of cholesterol into
primary BA, disorders of primary BA synthesis are not uncom-
mon [4]. Bile acid synthesis disorders can be diagnosed by iden-
tification of changes in BA composition within serum and urine
(lack of primary BA, high concentrations of intermediary
metabolites). Such disorders are important to recognize because
many can be easily treated by supplementing the diet with pri-
mary BA, which can be lifesaving. The fed BA, through the
action of the nuclear receptor FXR, downregulate BA synthesis,
and fewer toxic intermediate species are generated [5].
Disorders associated with abnormalities
of hepatocyte function
Transmembrane transporter proteins are important for normal bile
secretion, therefore transporter defects play a central role in the
pathogenesis of cholestasis. Many of the proteins crucial to bile‐
component transport belong to the family of ATP‐binding cassette
(ABC) transporters. Some of these defects manifest in early life
and others may become apparent only in certain situations such as
pregnancy or after ingestion of specific pharmaceutical com-
pounds. Defects in different transports are considered below and
summarized in Table 29.2.
Bile constituents (BA, bilirubin, cholesterol, phospholipids,
and other products of metabolism) are secreted into biliary
352 THE LIVER:  INHERITED FORMS OF INTRAHEPATIC CHOLESTASIS

Table 29.1  Disorders of primary BA biosynthesis

Enzyme defect
CYP27A1 deficiency
(cerebrotendinous xanthomatosis)
3β‐Hydroxy‐C27‐steroid
oxidoreductase deficiency
Δ4‐3‐oxosteroid‐5β‐reductase
deficiency
Oxysterol 7α‐hydroxylase
deficiency
2‐methylacyl‐CoA racemase
deficiency
Bile acid‐CoA ligase deficiency
Bile acid‐CoA: Amino acid N‐
acyltransferase deficiency (familial
hypercholanemia)
Response to treatment with
Clinical presentation primary BA Reference
Neonatal cholestasis in some patients; juvenile Good [6, 7]
cataracts, tendinous xanthomata, and progressive
neuropsychiatric signs including cerebellar
ataxia, spastic paraparesis, and dementia
Cholestasis in infancy, hepatomegaly, variable Good [8–12]
course. Commonest BA synthesis disorder
Early neonatal cholestasis, rapid progression of Good [13]
liver disease
Infantile cholestasis progressing to cirrhosis and Poor, requiring liver [14, 15]
liver failure (single patient identified). Same gene transplantation (single known
(SPG5A) mutated in hereditary spastic paraplegia patient, CDCA given)
Infantile mild cholestasis; adult sensorimotor Good, may require restriction [16–18]
neuropathy of phytanic and pristanic acids
Neonatal cholestasis, growth failure, fat‐soluble May require conjugated BA [1, 19]
vitamin deficiency
Growth failure and coagulopathy of fat‐soluble May require conjugated BA [20]
vitamin deficiency, without jaundice

NA, not applicable; CA, cholic acid; CDCA, chenodeoxycholic acid; GGT, gamma glutamyl transpeptidase.

c­analiculi in an energy‐dependent manner. Transmembrane
transporter proteins mediate the secretory function of hepato-
cytes and biliary epithelial cells. Genes encoding some of these
transporters were found to be mutated in progressive familial
intrahepatic cholestasis (PFIC), a clinical syndrome manifest by
nonremitting conjugated hyperbilirubinemia, often with low or
normal serum GGT values and liver scarring; typically requir-
ing liver transplantation before adulthood. Many patients with
PFIC also have reduced concentrations of primary BA in bile
[21, 22]. Mutations in two genes, ATP8B1 and ABCB11, were
initially found to cause ­respectively FIC1 and bile salt export
pump (BSEP) deficiencies. A spectrum of disease severity is
recognized in association with mutation in ATP8B1 and
ABCB11. Whilst some patients have a classical PFIC pheno-
type, others have cholestasis that completely resolves between
bouts, a phenotype sometimes termed “benign” recurrent intra-
hepatic cholestasis [23, 24]. However, patients with apparently
benign disease may progress to PFIC after some years [25, 26].
In truth, therefore, both FIC1 and BSEP deficiencies are spectra,
and late or relapsing onset does not always predict a benign
clinical course. Heterozygotes for ATP8B1 and ABCB11 muta-
tions are predisposed to intrahepatic cholestasis of pregnancy
(ICP) [27, 28]. Intrahepatic cholestasis of pregnancy, a third‐tri-
mester disorder, is characterized by pruritus and elevated serum
concentrations of BA [29] and causes fetal disease, with fetal
distress, premature birth, and stillbirth [29]. GGT values are
normal in 70% of patients with ICP, implying that at least two
different processes can lead to ICP [30].
FIC1 deficiency
Previously, clinically severe FIC1 deficiency was often termed
Byler disease; Byler is the family name of the extended Amish
kindred first described with this condition [31]. ATP8B1 encodes
FIC1, a member of the type 4 subfamily of P‐type ATPases [32],
and hence is not an ABC transporter. FIC1 is present in the
­apical membrane of many epithelial cells, including hepatocytes
and enterocytes. Its role in cholestasis is only now becoming
clear [33–37]. FIC1 appears to translocate aminophospholipids
such as phosphatidylserine (PS) from the outer to the inner
­leaflet of the plasma membrane bilayer [36].
In mice homozygous for an Atp8b1 missense substitution
mutation (mimicking clinically severe FIC1 deficiency), PS,
cholesterol, and ectoenzymes were easily lost from canalicular
membranes during increased BA flux [38]. Membrane instabil-
ity in FIC1 deficiency may impair bile salt transport and
decrease hepatocyte resistance to bile salts. FIC1 also may, via
FXR, upregulate transcription of ABCB11 [37].
Patients with severe FIC1 deficiency present in the first year
of life with pruritus and cholestatic jaundice [31, 35]. Extrahepatic
manifestations include diarrhea, episodes of pancreatitis, fat‐
soluble vitamin deficiencies secondary to malabsorption, sensori-
neural deafness, delay in growth and puberty, and elevated sweat
chloride concentration [39–43]. Although cholestasis resolves
after liver transplantation, marked hepatic steatosis is common in
the graft. Severe diarrhea and growth restriction persist [35]. The
etiology of these extrahepatic symptoms is unclear; patients have
been shown to have normal pancreatic secretion [44].
Liver biopsy at presentation in patients with FIC1 deficiency
shows predominantly centrilobular cholestasis that is canalicu-
lar rather than hepatocellular. Swelling and multinucleation of
hepatocytes is not common, neither is portal‐tract cholestasis.
With progression of disease, fibrosis and portal–portal bridging
develops. Unlike canalicular bile of patients with BSEP defi-
ciency, that of FIC1 patients is coarsely granular on transmis-
sion electron microscopy [22]. This may reflect shedding of
canalicular wall into lumen, as may deficiency of ectoenzyme
expression along canaliculi [38].
In one large study ATP8B1 mutations were detected in 30%
of PFIC (39/130) and 40% (20/50) of patients with relapsing
cholestasis. Genotype–phenotype correlation included a greater
likelihood of missense substitutions in milder disease than in
PFIC, with nonsense mutations only in heterozygous form in
the former phenotype; an exception, a homozygous frameshift
mutation, left most of ATP8B1 intact, with loss of only the ter-
minal 11 amino acid residues [45].
 29:  Molecular Cholestasis 353

BSEP deficiency anion transporter). It exports anionic glutathione, glucuronate,

or sulphated conjugates (including bilirubin) from hepatocytes
Bile salt export pump, encoded by ABCB11, is the protein
into canaliculi [61, 62]. MRP2 is expressed on apical mem-
responsible for the ATP‐dependent transport of salts of primary
branes of epithelia other than hepatocytes (proximal renal
BA across the canalicular membrane [46]. BSEP is a member of
tubules, gallbladder, small intestine, bronchi, and placenta) [63,
the ABC family of proteins and also a member of the P glyco-
64]. MRP2 is now regarded as a crucial efflux pump, eliminat-
protein/multidrug resistance (MDR/ABCB) subfamily of trans-
ing conjugates of various toxins and carcinogens from hepato-
porters [47]. As with FIC1 disease, phenotypes of patients with
cytes into bile, from kidney proximal tubular epithelium into
mutations in ABCB11 represent a continuum between mild and
urine, and from intestinal enterocytes into chyme and stool [65].
severe disease [24, 26, 48–50].
Mutations in ABCC2 result in Dubin–Johnson syndrome, a con-
Liver biopsies from patients with severe BSEP deficiency dis-
dition characterized by episodes of cholestatic jaundice without
play a neonatal hepatitis picture at presentation [22]. Canalicular
other clinical biochemistry indication of hepatobiliary injury
bile on ultrastructural study may appear dense, amorphous, or
[66]. Microscopy of liver shows intrahepatocyte deposits of
finely filamentous, but is not loosely and coarsely granular as in
dark brown pigment but no other abnormalities. Patients may
FIC1 disease [22, 51]. Clinical features include intermittent or
present at any age; choleretics such as ursodeoxycholic acid
persistent cholestatic jaundice, pruritus, and growth failure, but
(UDCA) and phenobarbitone are recommended only in particu-
extrahepatic features such as diarrhea and pancreatitis typically
larly severe neonatal cases  [67]. The relatively mild, or even
are absent [35]. Genotype–phenotype correlation in a large
completely absent, clinical phenotype despite absence of MRP2
cohort of patients with severe BSEP deficiency found less severe
may be explained by upregulation of other transporters such as
disease when missense mutations, even in heterozygous form,
MRP3; this compensates for loss of MRP2 as an anion efflux
were present, and suggested that truncating mutations conferred
pump [68]. Abcc2‐deficient rats and mice may be useful in stud-
increased risk of hepatobiliary malignancy [50].
ying pharmacokinetics of various drugs and drug elimination
Concordant with this were the findings [52, 53] that the range
mechanisms [69]. A partial phenocopy of Dubin–Johnson syn-
of severity of clinical phenotype in patients with ABCB11 muta-
drome exists in mice lacking radixin, a cytoskeletal protein
tions can be explained by intracellular localization and differ-
found in canalicular membranes, without which poorly anchored
ences in stability of mutated BSEP and in retention of
Abcc2 may not function optimally [70].
taurocholate transport function. Mutations associated with mild
Rotor syndrome, phenotypically very similar to Dubin–
BSEP deficiency were trafficked normally to the apical domain
Johnson syndrome, also follows an autosomal recessive mode of
of the plasma membrane, where they retained some transport
inheritance. No intrahepatocyte pigment deposits are found in
activity, whilst most mutant BSEP forms associated with severe
Rotor syndrome patients, however, and Dubin–Johnson and
disease were not trafficked to the apical membrane domain and
Rotor syndromes can also be distinguished by studies of bromo-
lacked most transport activity [52].
sulfophthalein and coproporphyrin excretion [71]. Allelism with
Like ATP8B1, carriers of ABCB11 mutations are prone to devel-
Dubin–Johnson syndrome has been excluded in Rotor syndrome
oping ICP [54, 55]. Cholestasis associated with the administration
[72]; instead it has been suggested that a combined impairment
of hormonal agents such as oral contraceptives also is more frequent
of hepatocellular uptake transporters is lacking [73].
in carriers of ABCB11 mutations; inhibition of BSEP by estrogen
and progestogen metabolites may be contributory [56, 57]. ABCB11
mutations, unlike ATP8B1 mutations, predispose to gallstones. This MDR3 deficiency
complication is particularly common with milder disease; an inci-
This disorder results from mutations in ABCB4, which encodes
dence as high as 7/11 was found in benign recurrent intrahepatic
the MDR3, or multidrug resistance protein 3, a P‐glycoprotein
cholestasis‐2 (BRIC‐2) [24]. Although ABCB11 lies in the murine
that flops phospholipids from the internal to external leaflet of
Lith1 major susceptibility locus for cholelithiasis, a study of 810
the canalicular membrane [74, 75]. In mice Abcb4 encodes the
German patients and 718 controls showed no association between
orthologous protein mdr2, and homozygous mutants are a good
ABCB11 mutation and cholelithiasis [58].
model for the study of pathogenesis and testing of therapies in
In mice, knockout of Abcb11 resulted in only mild cholesta-
human MDR3 deficiency [76, 77].
sis unless CA was fed [59]. Higher excretion of BA than
MDR3 deficiency results in impaired translocation of phos-
expected was traced to complementation of Abcb11 by Abcb1,
phatidylcholine (PC) into bile and is associated with a wide
which may transport salts of BA in mice [60].
spectrum of phenotypes ranging from neonatal cholestasis to
Hepatocellular carcinoma seems to be a particularly common
biliary cirrhosis in adults [38, 78–80]. PC in bile is an essential
feature of BSEP deficiency. Complete absence of the protein is
component of the mixed micelles into which salts of BA are
associated with the highest incidence. The same patients are
emulsified, reducing BA detergent activity. Deficiency of
also at the greatest risk of inhibitory anti‐BSEP antibodies after
MDR3 permits BA unchaperoned by PC to damage hepatocytes
liver transplantation. This phenomenon has not been described
and cholangiocytes [75, 78, 81]. MDR3 disease is usually char-
in similar conditions, and can be problematic.
acterized by high serum GGT values and, on biopsy, ductular
reaction and fibrosis, a picture similar to that seen in Abcb4−/−
Dubin–Johnson syndrome
mice [74]. Liver disease progresses to fibrosis and cirrhosis,
The protein encoded by ABCC2, multidrug‐associated protein sometimes with ductopenia, and may require liver transplanta-
2 (MRP2), is a member of the ABC transporter superfamily and tion. Response to UDCA may be seen [79, 80, 82], particularly
also known as the cMOAT (canalicular multispecific organic in milder disease. Obligate carriers of ABCB4 mutations
354 THE LIVER:  INHERITED FORMS OF INTRAHEPATIC CHOLESTASIS
(parents of infants with ABCB4 mutations ablating MDR3 func-
tion and liver disease presenting as neonatal hepatitis [NH]) are
prone to ICP and gallstones [81]. ABCB4 mutations similarly
have been found in a subgroup of patients with ICP and
high  GGT values, and some common ABCB4 polymorphisms
­conduce to development of ICP [83–85]. Certain ABCB4 poly-
morphisms also may mark adverse prognosis in primary biliary
cholangitis [86].
Low phospholipid‐associated cholelithiasis (LPAC) is a form
of gallstone disease occurring in younger patients which is asso-
ciated with ABCB4 mutations, recurs after cholecystectomy,
and sometimes to respond well to UDCA [87]. Homozygous
missense mutations or heterozygous nonsense or frameshift
ABCB4 mutations may be found in LPAC patients [88]. UDCA
may act by upregulating residual expression of MDR3 at the
canalicular membrane, exerting a protective effect against
hydrophobic endogenous BA, and increasing the pool of protec-
tive hydrophilic BA [87]. Hepatolithiasis with intractable chol-
angitis may, however, be associated with MDR3 disease in some
patients [89, 90].
As in BSEP disease, MDR3 disease is associated with drug‐
induced cholestasis (DIC) following administration of amoxi-
cillin, clavulanic acid, and risperidone [91]. Not enough data are
available to explain the pathophysiology of DIC associated with
the mutations and polymorphisms so far identified in these
genes. Although inhibition of ABCB4 expression has been sug-
gested, no evidence of this effect has yet been found [92].
TJP and BAAT deficiencies
Two forms of familial hypercholanemia (FHCA) were
described in Amish families. These involved mutations in
TJP2, encoding tight junction protein 2 (TJP2), and BAAT,
encoding bile acid‐CoA:amino acid N‐acyltransferase (BAAT)
[20]. Affected individuals usually have markedly increased
serum primary BA concentrations, severe pruritus, fat malab-
sorption leading to coagulopathy and rickets, and failure to
thrive [93]. While abnormal BA species are not present in
serum, in one form of FHCA, that owing to lesions in BAAT,
BA are inadequately conjugated [20, 93]. Patients intermit-
tently have episodes of hepatitis, and liver biopsy may find a
nonspecific pattern of portal and lobular inflammation with
spotty hepatocyte necrosis and mild ductular proliferation [93].
Serum BA concentrations fall and pruritus improves with
UDCA treatment [93]. The clinical course is unpredictable and
symptoms may resolve spontaneously [20].
The molecular pathology of familial hypercholanemia is
complex. A homozygous TJP2 missense mutation predicted to
yield the amino acid residue substitution p.V48A was identi-
fied in 11 out of 17 subjects [20]. TJP2 is a tight junction‐­
associated protein and biophysical studies suggested that the
altered protein conformation caused by the p.V48A mutation
may lead to increased paracellular permeability, with reflux
of  BA into plasma. Morphological abnormalities of tight
junctions were identified in patients’ liver tissues [20].
­ tion. Although other patients have been described previously
Homozygosity for this mutation was also found in three unaf-
fected siblings, suggesting lack of penetrance for this variant.
Hence, the investigators searched for further genes conferring
susceptibility to development of cholestasis and identified five
patients with a homozygous mutation in BAAT predicted to
yield the amino acid residue substitution p.M76V missense
substitution in BAAT, which conjugates BA with glycine
and  taurine [20]. The p.M67V mutation was associated in
dose‐dependent fashion with decreased BA conjugation.
Furthermore, 6 of 16 clinically‐affected individuals homozy-
gous for a mutation in either TJP2 or BAAT also were
­heterozygous for a mutation in the second gene [20], thus
­suggesting oligogenic inheritance in FHCA.
Whilst FHCA is associated with seemingly mild missense
variation in TJP2, more damaging variants have been associated
with a quite different phenotype [94]. In patients with a PFIC
phenotype, similar to FIC1 deficiency, protein truncating on
both alleles of TJP2 have been identified. These patients have
early onset progressive disease requiring liver transplantation.
In several cases hepatocellular carcinomata have been described
[95]. As TJP2 is widely expressed in other epithelia it should
not be surprising that extrahepatic manifestations have become
evident. It is probably more surprising that they are not more
severe, especially as the homozygous knockout mouse is embry-
onic lethal.
MYO5B associated cholestasis
MYO5B is an atypical myosin involved in intracellular vesicu-
lar transport. As such it is clearly important in apical membrane
assembly [96–99]. The first patients described with mutations in
the MYO5B gene (encoding MYO5B) were those with microvil-
lous atrophy (MVA) also known as microvillous inclusion dis-
ease (MVID) [100–103]. This is a severe form of intestinal
failure, requiring parenteral nutrition (PN). Whilst intestinal
failure associated with liver disease is common in children
requiring PN from infancy, it was clear that cholestasis was a
particular feature of these children [103]. Reduced expression
of BSEP in the apical membrane of hepatocytes in patients with
MYO5B associated MVA has been demonstrated. More recently
patients have been described with little or no intestinal disease,
but instead isolated cholestasis, with mutation in MYO5B. No
genotype–phenotype correlations are evident, and it is not yet
clear why intestinal failure can predominate in some patients
and not others.
Neonatal sclerosing cholangitis
Sclerosing cholangitis, of neonatal onset (NSC), was first
described many years ago. There was a strong suggestion that
many cases had a genetic etiology. A recent study found that a
quarter of such patients harbored mutations in the DCDC2 gene
[104, 105]. This gene encodes a member of the doublecortin
family of proteins, and has previously been implicated in dys-
lexia and shown to be mutated in nephronophthisis [106, 107].
The protein appears to be important in the flagellary transport
body, itself a component of cilial assembly. DCDC2‐associated
disease is therefore a ciliopathy. The majority of patients have
rapidly progressive liver disease, requiring early transplanta-
with a renal phenotype, many NSC have normal renal function
and imaging. Several such patients have undergone uneventful
liver transplantation with use of calcineurin inhibitors, without
obvious complications.
 29:  Molecular Cholestasis 355

Table 29.2  Disorders associated with abnormalities of hepatocyte function

Diseases and genes Typical presentations and clinical features
FIC1 deficiency 1. Infantile cholestasis, pruritus, normal GGT; bland, canalicular cholestasis on liver biopsy at
Mutations in ATP8B1 presentation, advancing to cirrhosis. Bile coarse and loosely granular on transmission electron
microscopy. Diarrhea, episodes of pancreatitis, malabsorption, sensorineural deafness, delay in growth
and puberty, short stature, elevated sweat chloride concentrations. Severe mutations on both alleles.
2. Periodic bouts of cholestasis, intense pruritus. Classically cholestasis resolves between the relapses,
but may progress. Milder mutations, usually on both alleles.
3. Cholestasis of pregnancy. Resolution of symptoms after delivery of the baby. Usually associated with
monoallelic mutation.
BSEP deficiency 1. Infantile cholestasis, pruritus, normal GGT. Giant‐cell hepatitis on liver biopsy at presentation,
Mutations in ABCB11 advancing to cirrhosis. Growth failure. Increased risk of hepatocellular carcinoma and
cholangiocarcinoma. Severe mutations on both alleles.
2. Periodic bouts of cholestasis, intense pruritus. Classically cholestasis resolves between the relapses,
but may progress. Milder mutations, usually on both alleles.
3. Cholestasis of pregnancy. Usually symptoms resolve after delivery of the baby. Usually associated
with monoallelic mutation.
4. Cholestasis induced by the administration of hormonal contraceptives, and other drugs. Usually
associated with monoallelic mutation.
Dubin–Johnson syndrome Recurrent episodes of cholestatic jaundice without liver disease. Mutations on both alleles
Mutations in ABCC2
MDR3 deficiency 1. Early onset cholestatic jaundice with high GGT. Severe mutations on both alleles.
Mutations in ABCB4 2. Cholangiopathy, similar to “small duct primary sclerosing cholangitis”. Rapidity of progression
dependent on degree of functional loss. May be associated with severe mutation on one allele, or
milder mutations on both.
3. Low phospholipid‐associated cholelithiasis, with hepatolithiasis a well as cholecystolithiasis. Usually
monoalleleic mutation.
4. Cholestasis of pregnancy with raised GGT. Usually associated with monoallelic mutations, but can be
first presentation of more severe disease.
5. Cholestasis induced by the administration of hormonal contraceptives and other drugs. Usually
associated with monoallelic mutation.
TJP2 deficiency 1. Early onset, normal GGT cholestasis. Bland, canalicular cholestasis on liver biopsy at presentation,
Mutations in TJP2 advancing rapidly to cirrhosis. Severe mutations on both alleles.
2. Episodic cholestasis, often associated with drug administration. Milder mutations on both alleles.
3. Hypercholanaemia. No liver disease. Usually with specific missense change on both alleles.
MYO5B deficiency 1. Microvillous inclusion disease, causing intestinal failure, often with particularly severe cholestasis.
Mutations in MYO5B Mutations of both alleles.
2. Isolated cholestasis, with variable age of onset and duration. Mutations of both alleles.
Neonatal sclerosing cholangitis Early onset cholangiopathy with rapid progression in most cases. Renal involvement in some cases. Severe
Mutations in DCDC2 in 25% of cases mutations on both alleles.

Syndromic forms of inherited cholestasis
Multisystem disorders with a prominent feature of cholestasis
are delineated below and summarized in Table 29.3.
Alagille syndrome
Alagille syndrome (ALGS) is an autosomal dominant disorder
which affects development of multiple organ systems, including
cardiovascular, musculoskeletal, gastrointestinal, central nerv-
ous, and renal [108]. A clinical diagnosis of ALGS is made when
intrahepatic bile duct paucity is associated with three
­distinguishing clinical features such as cardiac disease, abnormal
vertebral bodies, typical ocular abnormalities, and characteristic
facial features [109]. Hepatocellular carcinoma is a recognized
complication, occurring in children as young as four years old
[110, 111]. The incidence of ALGS is estimated between 1 in
30 000 to 1 in 70 000 [112]. Mutations in JAG1 are found in more
than 90% of patients with ALGS [113, 114]. JAG1 encodes a
cell‐surface protein, JAG1, a ligand for NOTCH receptors. The
notch signaling pathway, including both JAGGED1 and
NOTCH2, is important in morphogenesis and is conserved in
evolution [115–119]. Clinical ­features in ALGS vary dramati-
cally and intrafamilial differences among penetrance of various
clinical features are substantial. In fact, detailed study of the
clinical manifestations that characterize ALGS in JAG1 muta-
tion‐positive relatives of classical ALGS patients revealed that
liver and cardiac disease exhibited 31% and 41% penetrance,
respectively, among mutation carriers [112]. This suggested that
other factors, such as genetic modifier loci, in addition to JAG1
mutations determine the development of clinical features in this
condition. Complete knockout of Jag1 was embryonically lethal
in mice and haploinsufficiency for murine Jag1 did not cause a
complete ALGS‐like phenotype. A knockout of Notch2 pro-
duced an ALGS mouse model; heterozygous mutations in
NOTCH2 were identified in several JAG1 mutation‐­negative
patients with ALGS [120–122]. Further studies in various Notch2
and Jag1 mouse knockouts have extended ­knowledge of the cru-
cial role of this pathway, in particular with respect to gene dos-
age, in biliary morphogenesis [123].
Arthrogryposis, renal dysfunction and cholestasis syndrome
Arthrogryposis, renal dysfunction, and cholestasis (ARC)
syndrome is an autosomal recessive multisystem disorder
which commonly presents with NH. Patients usually have
associated neurogenic arthrogryposis and renal tubular acido-
sis; however, as these features may be mild, they are easily
missed early in evaluation. ARC syndrome is caused by
356 THE LIVER:  INHERITED FORMS OF INTRAHEPATIC CHOLESTASIS
mutations in either VPS33B, which encodes VPS33B, or
VIPAS39 encoding VIPAR [124, 125]. VPS33B and VIPAR
form a stable protein complex with a function in intracellular
protein trafficking and biogenesis of various lysosome‐related
organelles such as keratinocyte lamellar bodies or platelet
alpha granules [126–130].
VPS33B and VIPAR are homologous to VPS33A and
VPS16 proteins that form parts of the HOPS and CORVET
complexes involved in membrane fusion in endocytic pathway
and biogenesis of early and late endosomes and autophago-
somes. Thus, it is hypothesized that VPS33B and VIPAR com-
plexes may be involved in a similar process to HOPS but in
different intracellular trafficking pathways. For example, it
was found that integrin beta subunit directly interacted with
VPS33B and Vps33b knockout mice had abnormalities in
αIIbβ3‐mediated fibrinogen endocytosis.
ARC patients’ liver biopsy features include bile duct paucity,
lipofuscin deposition, and giant‐cell transformation of hepato-
cytes [131]. A striking feature of ARC syndrome is abnormal
localization of specific apical membrane proteins (e.g. alanyl
aminopeptidase/CD13, dipeptidyl peptidase/CD26, and carci-
noembryonic antigen / CD66) along biliary canaliculi and in
renal tubular epithelium [124, 132]. Liver disease in ARC is
associated with normal gGT, and gGT in hepatocytes also is
abnormally sited in ARC [124]. Impaired delivery of apical
membrane proteins to their destination could explain renal tubu-
lar leak of various molecules including glucose, amino acids,
and bicarbonate and might underlie the failure of serum gGT
values to rise despite cholestasis. Vps33b knockout mice dis-
played abnormal localization of multiple ABC transporters
including BSEP and ABCG8 but not MDR2 suggesting involve-
ment of VPS33B‐VIPAR complex in the Rab11a‐dependent
protein trafficking route in the hepatocytes [98, 103, 133]. Wang
and colleagues showed that VPS33B is downregulated in the
samples from human and mouse hepatocellular carcinoma and
Vps33b liver‐specific knockout mice developed hepatocellular
carcinoma following progressive hepatitis and fibrosis [134].
In addition to mild dysmorphism, hypotonia (more than 90%
of ARC syndrome patients), and developmental delay, abnor-
malities (hypoplasia or absence) of the corpus callosum occur in
~20% [135]. It is now recognized that sensorineural deafness
can be identified in the majority of ARC patients. Other symp-
toms include hypothyroidism and congenital cardiac defects.
Ichthyosis is seen in most patients and recurrent febrile illnesses
occur in many, although the cause of the fevers is often not
found. Interestingly, Akbar and colleagues found that VPS33B
is necessary for clearance of endosomes containing internalized
pattern‐recognition receptors following microbial recognition
[136]. VPS33B deficiency led to enhanced signaling and expres-
sion of inflammatory mediators which may explain the febrile
episodes in ARC.
Absence of platelet alpha granules can be detected on elec-
tron microscopy; platelets appear hypogranular and large in
approximately 25% of patients with ARC [127, 128, 137]. An
increased risk of severe bleeding after liver and kidney biopsies
is probably due to abnormal platelet aggregation. Severe failure
to thrive is a prominent feature of ARC syndrome. Death occurs
before the first birthday in most patients, usually from sepsis,
severe dehydration, and acidosis [135].
CLAUDIN‐1 deficiency (NISCH syndrome)
Neonatal ichthyosis and sclerosing cholangitis syndrome
(NISCH) is a rare monogenic form of sclerosing cholangitis
associated with recessive mutations in CLDN1 encoding the
tight junction protein claudin‐1 [138, 139]. Sixteen patients and
three different mutations have been described so far in families
from Morocco and Switzerland [140]. There is significant clini-
cal heterogeneity in patients with NISCH. Whilst some may have
early normal gGT, typically children develop neonatal cholesta-
sis with high gGT and hepatomegaly [141]. Liver biopsy find-
ings include extensive fibrosis without fatty infiltration or
ductular proliferation and a mixture of extracellular and intracel-
lular cholestasis. Transparietal cholangiography may show scle-
rosing cholangitis and portal hypertension after the age of five.
Interestingly, in some patients resolution of liver disease may
occur with symptomatic treatment [140]. The cutaneous features
of the syndrome include dry and hard skin with large scales par-
ticularly prominent on the abdomen and limbs [141]. In addition,
patients have substantial alopecia. Other reported clinical mani-
festations included intracytoplasmic vacuoles in peripheral blood
eosinophils and enamel dysplasia. Interestingly, all cutaneous
phenomena subsided after liver transplantation.
Primary tight junction abnormalities also have been identi-
fied in a form of familial hypercholanemia (see TJP and BAAT
deficiencies), with secondary abnormalities in animal models of
cholestatic disease [142–146]. Tight junctions are essential to
maintain the separation between tissue layers and electrochemi-
cal gradients across epithelial cell monolayers. Thus, intact tight
junction structure is important to prevent bile leaking from bil-
iary canaliculi into blood [147]. Claudin‐1 is expressed at high
levels in mouse liver and kidney; however, a Claudin‐1 deficient
mouse was unsuitable for the investigation of cholestasis as skin
abnormalities led to death on the first postnatal day with severe
dehydration [148].
GRACILE syndrome
Several genetically distinct disorders affecting the mitochon-
drial respiratory chain can cause liver disease with or without
associated renal tubular abnormalities [149]. GRACILE syn-
drome (growth retardation, aminoaciduria, cholestasis, iron
overload, lactic acidosis, and early death) is a recessively inher-
ited lethal disease to date found only in Finnish populations and
known to be caused only by homozygous inheritance of a mis-
sense 232A>G (S78G) mutation in BSC1L [150, 151]. BCS1L
encodes a mitochondrial inner‐membrane protein necessary for
the assembly of mitochondrial respiratory chain complex III.
Typically, patients with GRACILE have fulminant lactic acido-
sis soon after birth, with nonspecific aminoaciduria and choles-
tasis. Prominent siderosis is noted in the hepatocytes and
reticuloendothelial system but typical histopathologic features
of mitochondrial dysfunction such as microvesicular steatosis
are not found. GRACILE syndrome patients have no neurologi-
cal abnormalities or dysmorphic features and normal respiratory
chain function, including normal complex III activity.
Other mutations described in BSC1L cause a variety of phe-
notypes including neurological disease, generalized mitochon-
driopathies and, most recently, Bjørnstad syndrome, a form of
sensorineural hearing loss associated with pili torti [152–156].
 29:  Molecular Cholestasis 357
Severity of effects of BSC1L mutation appears correlated with
increased generation of reactive oxygen species [154]. A recent
review showed that whilst the collection of symptoms in
GRACILE is specific to the 232A>G (S78G) mutation, other
mutations in BCS1L can cause liver disease. The early diagno-
sis is important as data shows that a ketogenic diet may improve
hepatic symptoms [157].
SLC25A13 disease (citrullinemia type II, neonatal
intrahepatic cholestasis caused by citrin
deficiency, NICCD)
SLC25A13 (CITRIN) deficiency was initially described in
adult‐onset citrullinemia type II (CTLN2), an autosomal
recessive disorder with variable age of presentation character-
ized by intermittent encephalopathy with hyperammonemia
[158–160]. CTLN2 is due to decreased activity of hepatic
argininosuccinate synthase, which leads to hyperammonemia,
coma, and may cause death due to acute metabolic decompen-
sation within a few years of onset [160]. More recently muta-
tions in SLC25A13 were discovered in patients with an
inherited form of NH, now known as neonatal intrahepatic
cholestasis caused by CITRIN deficiency (NICCD). Although
initially described predominantly in Japanese patients it was
then recognized in patients from neighboring Korea, Taiwan,
and China. We now appreciate that this disorder occurs in all
ethnic groups and is important to diagnose as specialized
management may not only cause clinical improvement but
could prevent the onset of the adult form of the disease [161,
162]. A summary of the clinical characteristics of NICCD in
75 patients reported intrahepatic cholestasis with raised gGT,
acholic stools, and poor weight gain in most. A NH syndrome
was associated with hypocoagulability, hypoglycemia, distur-
bances of liver synthetic function, galactosuria, and, rarely,
hyperammonemia. A typical but transient feature in NICCD
was found to be abnormal plasma amino acid concentrations
[161]. Most patients had raised citrulline, methionine, argi-
nine, and threonine to serine ratio values. Raised tyrosine and
lysine values are also seen in some patients [163]. Timely
diagnosis of NICCD is important in view of specific dietary
treatment (increased protein intake, initial treatment with a
galactose‐free diet, and avoidance of high carbohydrate
intake). The standard dietary management of NH could be
partially toxic in this condition, precipitating deterioration of
liver disease and even requiring transplantation [161]. NICCD
is typically not severe although some cases required liver
transplantation [163]. The proportion of patients who pro-
gress from NICCD to CTLN2 is unknown. An additional
­benefit of accurate diagnosis may be in the introduction of
cancer surveillance and early intervention as hepatocellular
carcinoma and cholangiocarcinoma are known to complicate
CITRIN deficiency [164].
Citrin is a liver‐specific mitochondrial aspartate‐glutamate
carrier. Investigation of the Slc25a13 knockout (Ctrn−/−) mouse
model revealed markedly decreased activities in aspartate
transport and in the malate‐aspartate mitochondrial shuttle.
Deficits in ureagenesis from ammonia and in gluconeogenesis
from lactate also could be demonstrated [165]. Ctrn−/− mice
failed to show CTLN2‐like symptoms; this suggests that citrin
deficiency alone may not be sufficient to produce a CTLN2‐
like phenotype in mice. These observations are in parallel with
the variable age of onset and incomplete penetrance of CTLN2,
where involvement of additional environmental or genetic
­factors is suspected.
CIRH1A deficiency (North American Indian
childhood cirrhosis, NAIC)
NAIC is a form of neonatal cholestasis that progresses to
cirrhosis. It has been recognized only in the Ojibway‐Cree
­
­population from northwestern Quebec [166, 167]. Children may
initially present with transient neonatal jaundice which pro-
gresses to cirrhosis. Management of the liver disease requires
liver transplantation in childhood or early adulthood. Patients
have elevated gGT levels and the characteristic histological
appearance of this condition, with portal‐tract fibrosis, may
resemble that of extrahepatic biliary atresia. A missense muta-
tion in CIRH1A, which encodes CIRHIN, was identified in
these children [167]. It was found that CIRHIN is a ribonucleo-
protein that is required for maturation of the 18S rRNA of the
ribosome. The NAIC mutation is likely to affect the function of
CIRHIN in ribosome biogenesis and pre‐ribosome assembly by
disrupting interaction with NOL11 [168, 169].
Lymphoedema‐cholestasis syndrome
Lymphoedema‐cholestasis syndrome (LCS, also known as
Aagenaes syndrome) was originally described in Norwegian
patients [170], and refers to an autosomal recessive disorder
characterized by severe neonatal cholestasis with an elevated
­
serum gGT. Although in most patients investigated liver disease
progressed to cirrhosis, some requiring liver transplantation in
childhood, more than 50% of patients survive into adulthood [171,
172]. In addition, patients develop severe lymphoedema, which
may be manifest at birth or appear later in childhood [173].
Norwegian families with LCS share genetic linkage to LCS1 locus
on chromosome 15 [173], however, the causal gene mutations
have not yet been identified. Mutations in CCBE1 (collagen and
calcium‐binding EGF domain‐1) were also reported in pediatric
and adult patients with LCS from different ethnic backgrounds
[174, 175]. CCBE1 is known to be important for embryonic lym-
phangiogenesis and mutations in CCBE1 were identified in cases
of lymphatic dysplasia and hydrops fetalis [176, 177].
Other inherited conditions associated
with cholestasis
A number of other conditions not mentioned here but thor-
oughly reviewed elsewhere may present with neonatal and
infantile cholestasis. The list of the disorders includes alpha‐1
antitrypsin deficiency, Niemann–Pick disease type C, galacto-
semia, hereditary fructose deficiency, fatty acid oxidation
defects, cystic fibrosis, and indeed many other infective or
immune‐mediated conditions in which hepatocellular injury
leads nonspecifically to impairment of pathways involved in
handling of the constituents of bile [7, 178]. Each of the condi-
tions mentioned above has specific distinguishing features and
baseline metabolic investigations should help in diagnosis.
358 THE LIVER:  INFANTILE CHOLESTATIC SYNDROMES WITHOUT OBVIOUS GENETIC CAUSE
Table 29.3  Syndromic forms of cholestasis
Clinical
syndromes Hepatic manifestations
Alagille Neonatal giant‐cell hepatitis with high GGT in
syndrome some patients. Chronic cholestasis with bile duct
loss in most patients with fully penetrant
phenotype.
Hepatocellular carcinoma may occur as early as
four years of age
NISCH Neonatal cholestasis with high gGT and
hepatomegaly. Liver biopsy shows extensive fibrosis
and cholangiography shows sclerosing cholangitis.
Portal hypertension detected after age five years
NICCD NH with high GGT. Liver biopsy features of
Citrullinemia hepatocyte microvesicular steatosis and cholestasis.
type II Patients may develop hepatocellular carcinoma
(adult onset)
GRACILE Hepatocellular cholestasis and giant‐cell changes,
syndrome progressing to intralobular fibrosis; prominent
siderosis. Bile duct paucity may be seen
ARC syndrome Neonatal cholestasis with low GGT, abnormal
bile excretion on hepatobiliary scan with
trimethylbromoimino‐diacetic acid (TBIDA)
scan, lipofuscin granule accumulation in
hepatocytes and interlobular bile duct hypoplasia.
Patients typically have very severe failure to
thrive
Lymphoedema‐ Transient neonatal cholestasis with high GGT.
cholestasis Liver disease progresses to cirrhosis in most.
syndrome Transient episodes of cholestasis in adulthood
NAIC Transient neonatal cholestasis progressing to
cirrhosis. High GGT. Liver transplantation is
required in childhood/early adulthood.
Histopathological features of severe
cholangiopathy
INFANTILE CHOLESTATIC SYNDROMES
WITHOUT OBVIOUS GENETIC CAUSE
Cholestatic jaundice is the commonest presentation of liver
­disease in infancy, affecting approximately 1 in every 2500
infants. Cholestasis has multiple etiologies with only subtle
clinical differences among various diseases. Obstruction to bile
flow – whether mechanical, as in extrahepatic biliary atresia, or
functional, as in PFIC  –  increases intrahepatocyte bile acid
­concentration and causes damage which, if not corrected, leads
to cell death. Jaundice, acholic stools, hepatomegaly, and the
consequences of fat malabsorption are among the manifesta-
tions, joint or several, of cholestasis [179, 180].
The most populated categories of cholestatic jaundice in the
first months of life, historically accounting for up to 70% of
cases [179, 180], are the syndromes of extrahepatic biliary atresia
(EHBA) and NH. In most patients with either syndrome the cause
was then unknown, although some cases were already linked to
genetic, infectious, or other environmental factors [181, 182].
EHBA is the most frequent form of neonatal cholestasis, occur-
ring in approximately 1 in 10 000 births, and is more common in
Extrahepatic manifestations Genetic defect
Cardiovascular: typically peripheral pulmonary stenosis, JAG1 or NOTCH2
various other congenital cardiac malformations. Monoallelic mutation,
Renal: possibly more common in NICCD. Various forms mostly with complete
of dysplasia, cystic kidneys, renal tubular insufficiency loss of function
(hematuria, proteinuria), renal tubular acidosis.
Ocular: posterior embryotoxon.
Facial: pointed chin, straight nose, deep set eyes, broad
forehead.
Skeletal: butterfly vertebrae
Dry and scaly skin. Ichthyosis most prominent on CLDN1
abdomen and limbs. Hypotrichosis and alopecia. Mutation of both alleles
Intracytoplasmic vacuoles in peripheral blood eosinophils.
Significant improvement after liver transplantation
Intermittent hyperammonemia, confusion, coma, SLC25A13 (CITRIN)
cerebral edema occurs in the adult form of CITRIN Mutation of both alleles
deficiency
Growth retardation, lactic acidosis, aminoaciduria, and BCSL1
iron deposition in hepatocytes and reticuloendothelial Mutation of both alleles
system. Pancreatic fibrosis. Death in infancy
Neuromuscular: neurogenic arthrogryposis. The severity VPS33B or VIPAS39
may vary from unilateral talipes to multiple joint Mutation of both alleles
involvement. Dysplasia of corpus callosum in at least 20%.
Renal: renal fanconi syndrome and nephrogenic diabetes
insipidus. Both can vary in severity.
Hematology: deficiency of alpha granules resulting in
hypogranular appearance of platelets. Abnormal
platelet aggregation.
Skin: ichthyosis, cutis laxa.
Dysmorphism: low set ears, hirsutism, proximal
insertion of thumbs
Peripheral lymphoedema LCS1 locus on
chromosome 15q26.1
CCBE1 mutations, on
both alleles, in patients
without linkage to LCS1
Not known CIRH1A
Mutation of both alleles
girls. Concentrations of GGT activity in serum rise in EHBA.
EHBA is characterized by complete fibrotic obliteration of the
lumen of all or part of the extrahepatic biliary tree within the first
three postnatal months [183]. Luminal obliteration and fibrosis of
intrahepatic bile ducts also may be seen [184]. Approximately
20% of patients with biliary atresia also have at least one other
major congenital anomaly. This finding suggests that genetically
determined defects underlie some of these cases and that the same
genes are involved in regulation of development of both the biliary
tract and other organs [185, 186]. Polysplenia syndrome (polys-
plenia, midline liver, interrupted inferior vena cava, situs inversus,
preduodenal portal vein, and malrotation of the intestine) in
particular is present in 10% of all children with biliary atresia
[187, 188]. Abnormal situs coupled with bile‐duct disease sug-
gests that genes involved in shaping the laterality of thoracic
and abdominal organs are involved in bile duct development.
Numerous authors have suggested virus‐induced mecha-
nisms in the etiopathogenesis of EHBA. The rationale for this
is reviewed elsewhere [189, 190].
Interestingly, EHBA has been reported as a rare association
of primary ciliary dyskinesia [191]. Indeed, the inv mouse, with
 29:  Molecular Cholestasis 359
a defect in the murine orthologue of INVS, has EHBA associ-
ated with situs inversus [192, 193]. INVS mutations were iden-
tified as one of the causes of nephronophthisis, an autosomal
recessive syndrome of cystic kidney disease often associated
with congenital hepatic fibrosis and cholestasis [194, 195].
Many other inherited renal cystic disorders are associated with
hepatic fibrosis, which suggests analogous pathways for the
development of tubular structures in the liver and kidneys [196,
197]. Primary cilia act not only to move fluid: apart from
mucociliary clearance, cilia also take part in pattern formation
during embryonic development, in left–right axis orientation
and in retinal photoreception [198, 197]. A particularly impor-
tant role in development of bile ducts and renal tubules is
reserved for primary cilia; most of the known genes inactivated
in hepatorenal cystic diseases are associated with primary cilia
function [199].
Defects in other molecular processes, such as tight junction
protein complex formation and intercellular signaling during
bile‐duct morphogenesis, may also be among contributors to
EHBA. Deficiency of the proteins involved in these processes
causes specific forms of neonatal sclerosing cholangitis (absence
of claudin‐1 in NISCH syndrome, [138]) and paucity of inter-
lobular bile ducts (haploinsufficiency for JAG1/NOTCH2 in
Alagille syndrome [113, 122]. Assessment of synthesis of dys-
functional forms of these proteins in patients with EHBA
remains to be undertaken. A description of villin deficiency
associated with EHBA‐like disease [200] has not been followed
by a report of a genetic correlate.
The proportion of patients with a final diagnosis of NH is
getting smaller, as a result of the identification of the diseases
described above. NH is evidently a syndrome, or clinicopatho-
logic diagnosis of last resort. NH is diagnosed if cholestasis
cannot be assigned to infection, genetic/metabolic defect, or
­ oxisomal disorders. J Lipid Res, 2002;43(3):438–44.
mechanical impediment to bile flow and if microscopy of liver
finds widespread giant‐cell transformation of hepatocytes. Giant‐
cell change of hepatocytes is common, however, to many choles-
tatic disorders in neonates, including EHBA. Most instances of
NH resolve without sequelae; they may represent constitutive
dysfunction of bile‐handling mechanisms unmasked by perinatal
stress, a hypothesis that awaits experimental assessment.
Previously, 20% of cases of idiopathic NH that are progressive,
appeared familial, and had a poor prognosis [184]. This group in
particular has been eroded by the discoveries described above,
combined with the widespread use of genetic testing [201–203].
REFERENCES
1. Heubi, J.E., Setchell, K.D., and Bove, K.E. Inborn errors of bile acid metabo-
lism. Semin Liver Dis, 2007;27(3):282–94.
2. Rossi, M., Vajro, P., Iorio, R. et al. Characterization of liver involvement in
defects of cholesterol biosynthesis: long‐term follow‐up and review. Am
J Med Genet A, 2005;132A(2):144–51.
3. Steinberg, S.J., Dodt, G., Raymond, G.V., Braverman, N.E., Moser, A.B., and
Moser H.W. Peroxisome biogenesis disorders. Biochim Biophys Acta,
2006;1763(12):1733–48.
4. Russell, D.W. The enzymes, regulation, and genetics of bile acid synthesis.
Annu Rev Biochem, 2003;72:137–74.
5. Hofmann, A.F. and Hagey, L.R. Bile acids: chemistry, pathochemistry, biology,
pathobiology, and therapeutics. Cell Mol Life Sci, 2008;65(16):2461–83.
  6. Sedel, F., Tourbah, A., Fontaine, B. et al. Leukoencephalopathies associated
with inborn errors of metabolism in adults. J Inherit Metab Dis, 2008;31(3):
295–307.
  7. Clayton, P.T., Verrips, A., Sistermans, E., Mann, A., Mieli‐Vergani, G., and
Wevers, R. Mutations in the sterol 27‐hydroxylase gene (CYP27A) cause
hepatitis of infancy as well as cerebrotendinous xanthomatosis. J Inherit
Metab Dis, 2002;25(6):501–13.
  8. Clayton, P.T., Leonard, J.V., Lawson, A.M. et al. Familial giant cell hepatitis
associated with synthesis of 3 beta, 7 alpha‐dihydroxy‐ and 3 beta,7 alpha,
12 alpha‐trihydroxy‐5‐cholenoic acids. J Clin Invest, 1987;79(4):1031–8.
 9. Ichimiya, H., Egestad, B., Nazer, H., Baginski, E.S., Clayton, P.T., and
Sjövall, J. Bile acids and bile alcohols in a child with hepatic 3 beta‐hydroxy‐
delta 5‐C27‐steroid dehydrogenase deficiency: effects of chenodeoxycholic
acid treatment. J Lipid Res, 1991;32(5):829–41.
10. Buchmann, M.S., Kvittingen, E.A., Nazer, H. et al. Lack of 3 beta‐hydroxy‐
delta 5‐C27‐steroid dehydrogenase/isomerase in fibroblasts from a child
with urinary excretion of 3 beta‐hydroxy‐delta 5‐bile acids. A new inborn
error of metabolism. J Clin Invest, 1990;86(6):2034–7.
11. Jacquemin, E., Setchell, K.D., O’Connell N.C. et al. A new cause of progres-
sive intrahepatic cholestasis: 3 beta‐hydroxy‐C27‐steroid dehydrogenase/
isomerase deficiency. J Pediatr, 1994;125(3):379–84.
12. Cheng, J.B., Jacquemin, E., Gerhardt, M. et al. Molecular genetics of 3beta‐
hydroxy‐delta5‐C27‐steroid oxidoreductase deficiency in 16 patients with
loss of bile acid synthesis and liver disease. J Clin Endocrinol Metab,
2003;88(4):1833–41.
13. Setchell, K.D., Suchy, F.J., Welsh, M.B., Zimmer‐Nechemias, L., Heubi, J.,
and Balistreri W.F. Delta 4–3‐oxosteroid 5 beta‐reductase deficiency
described in identical twins with neonatal hepatitis. A new inborn error in
bile acid synthesis. J Clin Invest, 1988;82(6):2148–57.
14. Setchell, K.D., Schwarz, M., O’Connell N.C. et al. Identification of a new
inborn error in bile acid synthesis: mutation of the oxysterol 7alpha‐hydrox-
ylase gene causes severe neonatal liver disease. J Clin Invest, 1998;102(9):
1690–703.
15. Tsaousidou, M.K., Ouahchi, K., Warner, T.T. et  al. Sequence alterations
within CYP7B1 implicate defective cholesterol homeostasis in motor‐neu-
ron degeneration. Am J Hum Genet, 2008;82(2):510–5.
17. Ferdinandusse, S., Rusch, H., van Lint, A.E., Dacremont, G., Wanders, R.J.,
and Vreken, P. Stereochemistry of the peroxisomal branched‐chain fatty acid
alpha‐ and beta‐oxidation systems in patients suffering from different per-
18. Van Veldhoven, P.P., Meyhi, E., Squires, R.H. et al. Fibroblast studies docu-
menting a case of peroxisomal 2‐methylacyl‐CoA racemase deficiency: pos-
sible link between racemase deficiency and malabsorption and vitamin K
deficiency. J Clin Invest Eur J Clin Invest, 2001;31(8):714–22.
19. Setchell, K.D., Heubi, J.E., O’Connell N.C., Hofmann, A.F., and Levine J.A.
Absence of bile acid conjugation: a new inborn error of bile acid metabolism.
Hepatol, 1996;24(4):371A.
20. Carlton, V.E., Harris, B.Z., Puffenberger, E.G., Batta, A.K., Knisely, A.S.,
Robinson, D.L. et al. Complex inheritance of familial hypercholanemia with
associated mutations in TJP2 and BAAT. Nat Genet, 2003;34(1):91–6.
21. Tazawa, Y., Yamada, M., Nakagawa, M., Konno, T., and Tada, K. Bile acid
profiles in siblings with progressive intrahepatic cholestasis: absence of bil-
iary chenodeoxycholate. J Pediatr Gastroenterol Nutr, 1985;4(1):32–7.
22. Bull, L.N., Carlton, V.E., Stricker, N.L. et  al. Genetic and morphological
findings in progressive familial intrahepatic cholestasis (Byler disease
[PFIC‐1] and Byler syndrome): evidence for heterogeneity. Hepatol, 1997;
26(1):155–64.
23. Houwen, R.H., Baharloo, S., Blankenship, K. et al. Genome screening by
searching for shared segments: mapping a gene for benign recurrent intrahe-
patic cholestasis. Nat Genet, 1994;8(4):380–6.
24. van Mil, S.W., van der Woerd, W.L., van der Brugge, G. et al. Benign recur-
rent intrahepatic cholestasis type 2 is caused by mutations in ABCB11.
Gastroenterol, 2004;127(2):379–84.
25. van Ooteghem, N.A., Klomp, L.W., van Berge‐Henegouwen, G.P., and
Houwen R.H. Benign recurrent intrahepatic cholestasis progressing to pro-
gressive familial intrahepatic cholestasis: low GGT cholestasis is a clinical
continuum. J Hepatol, 2002;36(3):439–43.
26. Takahashi, A., Hasegawa, M., Sumazaki, R. et al. Gradual improvement of
liver function after administration of ursodeoxycholic acid in an infant with
a novel ABCB11 gene mutation with phenotypic continuum between BRIC2
and PFIC2. Eur J Gastroenterol Hepatol, 2007;19(11):942–6.
360 THE LIVER:  REFERENCES
27. Meier, Y., Zodan, T., Lang, C. et al. Increased susceptibility for intrahepatic
cholestasis of pregnancy and contraceptive‐induced cholestasis in carriers of
the 1331T>C polymorphism in the bile salt export pump. World J Gastroenterol,
2008;14(1):38–45.
28. Müllenbach, R., Bennett, A., Tetlow, N. et al. ATP8B1 mutations in British
cases with intrahepatic cholestasis of pregnancy. 1. Gut, 2005;54(6):829–34.
29. Lammert, F., Marschall, H.U., Glantz, A., and Matern, S. Intrahepatic chol-
estasis of pregnancy: molecular pathogenesis, diagnosis and management. J
Hepatol, 2000;33(6):1012–21.
30. Milkiewicz, P., Gallagher, R., Chambers, J., Eggington, E., Weaver, J., and Elias,
E. Obstetric cholestasis with elevated gamma glutamyl transpeptidase: incidence,
presentation and treatment. J Gastroenterol Hepatol, 2003;18(11):1283–6.
31. Clayton, R.J., Iber, F.L., Ruebner, B.H., and McKusick V.A. Byler disease.
Fatal familial intrahepatic cholestasis in an Amish kindred. Am J Dis Child,
1969;117(1):112–24.
32. Bull, L.N., van Eijk, M.J., Pawlikowska, L. et al. A gene encoding a P‐type
ATPase mutated in two forms of hereditary cholestasis. Nat Genet,
1998;18(3):219–24.
33. Eppens, E.F., van Mil, S.W., de Vree, J.M. et al. FIC1, the protein affected in
two forms of hereditary cholestasis, is localized in the cholangiocyte and the
canalicular membrane of the hepatocyte. J Hepatol, 2001;35(4):436–43.
34. Ujhazy, P., Ortiz, D., Misra, S. et al. Familial intrahepatic cholestasis 1: stud-
ies of localization and function. Hepatol, 2001;34(4 Pt 1):768–75.
35. van Mil, S.W., Houwen, R.H., and Klomp L.W. Genetics of familial intrahe-
patic cholestasis syndromes. J Med Genet, 2005;42(6):449–63.
36. Paulusma, C.C., Folmer, D.E., Ho‐Mok, K.S. et  al. ATP8B1 requires an
accessory protein for endoplasmic reticulum exit and plasma membrane lipid
flippase activity. Hepatol, 2008;47(1):268–78.
37. Frankenberg, T., Miloh, T., Chen, F.Y. et al. The membrane protein ATPase
class I type 8B member 1 signals through protein kinase C zeta to activate the
farnesoid X receptor. Hepatol, 2008;44(1):195–204.
39. Demeilliers, C., Jacquemin, E., Barbu, V. et al. Altered hepatobiliary gene
expressions in PFIC1: ATP8B1 gene defect is associated with CFTR down-
regulation. Hepatol, 2006;43(5):1125–34.
40. Knisely, A.S., Agostini, R.M., Zitelli, B.J., Kocoshis, S.A., and Boyle J.T.
Byler’s syndrome. Arch Dis Child, 1997;77(3):276–7.
41. Oshima, T., Ikeda, K., and Takasaka, T. Sensorineural hearing loss associated
with Byler disease. Tohoku J Exp Med, 1999;187(1):83–8.
42. Winklhofer‐Roob, B.M., Shmerling, D.H., Solèr, R., and Briner, J. Progressive
idiopathic cholestasis presenting with profuse watery diarrhoea and recurrent
infections (Byler’s disease). Acta Paediatr, 1992;81(8):637–40.
43. Knisely A.S. Progressive familial intrahepatic cholestasis: a personal per-
spective. Pediatr Dev Pathol, 2000;3(2),113–25.
44. Walkowiak, J., Jankowska, I., Pawlowska, J., Bull, L., Herzig, K.H., and
Socha, J. Normal pancreatic secretion in children with progressive familial
intrahepatic cholestasis type 1. Scand J Gastroenterol, 2006;41(12):1480–3.
45. Klomp, L.W., Vargas, J.C., van Mil, S.W. et al. Characterization of mutations in
ATP8B1 associated with hereditary cholestasis. Hepatol, 2004;40(1):27–38.
46. Strautnieks, S.S., Bull, L.N., Knisely et al. A gene encoding a liver‐specific
ABC transporter is mutated in progressive familial intrahepatic cholestasis.
Nat Genet, 1998;20(3):233–8.
47. Stieger, B., Meier, Y., and Meier, P.J. The bile salt export pump. Pflugers
Arch, 2007;453(5):611–20.
48. Noe, J., Kullak‐Ublick, G.A., Jochum, W. et  al. Impaired expression and
function of the bile salt export pump due to three novel ABCB11 mutations
in intrahepatic cholestasis. J Hepatol, 2005;43(3):536–43.
49. Lam, C.W., Cheung, K.M., Tsui, M.S., Yan, M.S., Lee, C.Y., and Tong, S.F.
A patient with novel ABCB11 gene mutations with phenotypic transition
between BRIC2 and PFIC2. J Hepatol, 2006;44(1):240–2.
50. Strautnieks, S.S., Byrne, J.A., Pawlikowska, L. et  al. Severe bile salt export
pump deficiency: 82 different ABCB11 mutations in 109 families. Gastroenterol,
2008;134(4):1203–14.
51. Chen, H.L., Chang, P.S., Hsu, H.C. et  al. FIC1 and BSEP defects in
Taiwanese patients with chronic intrahepatic cholestasis with low gamma‐
glutamyltranspeptidase levels. J Pediatr, 2002;140(1):119–24.
52. Kagawa, T., Watanabe, N., Mochizuki, K. et  al. Phenotypic differences in
PFIC2 and BRIC2 correlate with protein stability of mutant Bsep and
impaired taurocholate secretion in MDCK II cells. Am J Physiol Gastrointest
Liver Physiol, 2008;294(1):G58–67.
53. Lam, P., Pearson, C.L., Soroka, C.J., Xu, S., Mennone, A., and Boyer, J.L.
Levels of plasma membrane expression in progressive and benign mutations
of the bile salt export pump (Bsep/Abcb11) correlate with severity of chole-
static diseases. Am J Physiol Cell Physiol, 2007;293(5):C1709–16.
54. Dixon, P.H., Wadsworth, C.A., Chambers, J. et al. A comprehensive analysis
of common genetic variation around six candidate loci for intrahepatic chol-
estasis of pregnancy. Am J Gastroenterol, 2014;109(1):76–84.
55. Dixon, P.H., Sambrotta, M., Chambers, J. et al. An expanded role for hete-
rozygous mutations of ABCB4, ABCB11, ATP8B1, ABCC2 and TJP2 in
intrahepatic cholestasis of pregnancy. Sci Rep, 2017;7(1):11823.
56. Vallejo, M., Briz, O., Serrano, M.A., Monte, M.J., and Marin J.J. Potential
role of trans‐inhibition of the bile salt export pump by progesterone metabo-
lites in the etiopathogenesis of intrahepatic cholestasis of pregnancy. J
Hepatol, 2006;44(6):1150–7.
57. Iwanaga, T., Nakakariya, M., Yabuuchi, H., Maeda, T., and Tamai, I.
Involvement of bile salt export pump in flutamide‐induced cholestatic hepa-
titis. Biol Pharm Bull, 2007;30(4):739–44.
58. Schafmayer, C., Tepel, J., Franke, A. et al. Investigation of the Lith1 candi-
date genes ABCB11 and LXRA in human gallstone disease. Hepatol,
2006;44(3):650–7.
59. Wang, R., Lam, P., Liu, L. et al. Severe cholestasis induced by cholic acid
feeding in knockout mice of sister of P‐glycoprotein. Hepatol, 2003;
38(6):1489–99.
60. Lam, P., Wang, R., and Ling, V. Bile acid transport in sister of P‐glycoprotein
(ABCB11) knockout mice. Biochemistry, 2005;44(37):12598–605.
61. Paulusma, C.C., Bosma, P.J., Zaman, G.J. et al. Congenital jaundice in rats
with a mutation in a multidrug resistance‐associated protein gene. Science,
1996;271(5252):1126–8.
62. Ito, K., Suzuki, H., Hirohashi, T., Kume, K., Shimizu, T., and Sugiyama, Y.
Molecular cloning of canalicular multispecific organic anion transporter
defective in EHBR. Am J Physiol, 1997;272(1 Pt 1):G16–22.
63. Paulusma, C.C., Kool, M., Bosma, P.J. et al. A mutation in the human cana-
licular multispecific organic anion transporter gene causes the Dubin–
Johnson syndrome. Hepatol, 1997;25(6):1539–42.
64. Ito, K., Suzuki, H., Hirohashi, T., Kume, K., Shimizu, T., and Sugiyama, Y.
Functional analysis of a canalicular multispecific organic anion transporter
cloned from rat liver. J Biol Chem, 1998;273(3):1684–8.
65. Nies, A.T. and Keppler, D. The apical conjugate efflux pump ABCC2
(MRP2). Pflugers Arch, 2007;453(5):643–59
66. Dubin, I.N. and Johnson F.B. Chronic idiopathic jaundice with unidentified
pigment in liver cells; a new clinicopathologic entity with a report of 12
cases. Medicine (Baltimore), 1954;33(3):155–97.
67. Lee, J.H., Chen, H.L., Chen, H.L., Ni, Y.H., Hsu, H.Y., and Chang M.H.
Neonatal Dubin–Johnson syndrome: long‐term follow‐up and MRP2 muta-
tions study. Pediatr Res, 2006;59(4 Pt 1):584–9.
68. Donner, M.G. and Keppler, D. Up‐regulation of basolateral multidrug resist-
ance protein 3 (Mrp3) in cholestatic rat liver. Hepatol, 2001;34(2):351–9.
69. Chu, X.Y., Strauss, J.R., Mariano, M.A. et al. Characterization of mice lack-
ing the multidrug resistance protein MRP2 (ABCC2). J Pharmakos Exp
Ther, 2006;317(2):579–89.
70. Kikuchi, S., Hata, M., Fukumoto, K. et al. Radixin deficiency causes conju-
gated hyperbilirubinemia with loss of Mrp2 from bile canalicular mem-
branes. Nat Genet, 2002;31(3):320–5.
71. Wolpert, E., Pascasio, F.M., Wolkoff, A.W., and Arias I.M. Abnormal sulfo-
bromophthalein metabolism in Rotor’s syndrome and obligate heterozy-
gotes. N Engl J Med, 1977;296(19):1099–101.
72. Hrebícek, M., Jirásek, T., Hartmannová, H. et al. Rotor‐type hyperbilirubi-
naemia has no defect in the canalicular bilirubin export pump. Liver Int,
2007;27(4):485–91.
73. van de Steeg, E., Stránecký, V., Hartmannová, H. et al. Complete OATP1B1
and OATP1B3 deficiency causes human Rotor syndrome by interrupting con-
jugated bilirubin reuptake into the liver. J Clin Invest, 2012;122(2):519–28.
74. Deleuze, J.F., Jacquemin, E., Dubuisson, C. et al. Defect of multidrug‐resist-
ance 3 gene expression in a subtype of progressive familial intrahepatic chol-
estasis. Hepatol, 1996;23(4):904–8.
75. Oude Elferink, R.P. and Paulusma, C.C. Function and pathophysiological impor-
tance of ABCB4 (MDR3 P‐glycoprotein). Pflugers Arch, 2007;453(5):601–10.
76. Lamireau, T., Bouchard, G., Yousef, I.M. et  al. Dietary lecithin protects
against cholestatic liver disease in cholic acid‐fed Abcb4‐deficient mice.
Pediatr Res, 2007;61(2):185–90.
77. Baghdasaryan, A., Fickert, P., Fuchsbichler, A. et al. Role of hepatic phos-
pholipids in development of liver injury in Mdr2 (Abcb4) knockout mice.
Liver Int, 2008;28(7):948–58.
 29:  Molecular Cholestasis 361
  78. Smit, J.J., Schinkel, A.H., Oude Elferink, R.P. et al. Homozygous disrup-
tion of the murine mdr2 P‐glycoprotein gene leads to a complete absence of
phospholipid from bile and to liver disease. Cell, 1993;75(3):451–62.
  79. Ziol, M., Barbu, V., Rosmorduc, O. et al. ABCB4 heterozygous gene muta-
tions associated with fibrosing cholestatic liver disease in adults.
Gastroenterol, 2008;135(1):131–41.
  80. Gotthardt, D., Runz, H., Keitel, V. et al. A mutation in the canalicular phos-
pholipid transporter gene, ABCB4, is associated with cholestasis, ductope-
nia, and cirrhosis in adults. Hepatol, 2008;48(4):1157–66.
  81. Jacquemin, E. Role of multidrug resistance 3 deficiency in pediatric and
adult liver disease: one gene for three diseases. Semin Liver Dis,
2001;21(4):551–62
  82. de Vree, J.M., Jacquemin, E., Sturm, E. et al. Mutations in the MDR3 gene
cause progressive familial intrahepatic cholestasis. Proc Natl Acad Sci
USA, 1998;95(1):282–7.
  83. Dixon, P.H., Weerasekera, N., Linton, K.J. et al. Heterozygous MDR3 mis-
sense mutation associated with intrahepatic cholestasis of pregnancy: evi-
dence for a defect in protein trafficking. Hum Mol Genet,
2000;9(8):1209–17.
  84. Pauli‐Magnus, C., Lang, T., Meier, Y. et al. Sequence analysis of bile salt
export pump (ABCB11) and multidrug resistance p‐glycoprotein 3
(ABCB4, MDR3) in patients with intrahepatic cholestasis of pregnancy.
Pharmacogenet, 2004;14(2):91–102.
  85. Wasmuth, H.E., Glantz, A., Keppeler, H. et al. Intrahepatic cholestasis of
pregnancy: the severe form is associated with common variants of the hepa-
tobiliary phospholipid transporter ABCB4 gene. Gut, 2007;56(2):265–70.
  87. Rosmorduc, O. and Poupon, R. Low phospholipid associated cholelithiasis:
association with mutation in the MDR3/ABCB4 gene. Orphanet J Rare
Dis, 2007;2:29.
 88. Rosmorduc, O., Hermelin, B., Boelle, P.Y., Parc, R., Taboury, J., and
Poupon, R. ABCB4 gene mutation‐associated cholelithiasis in adults.
Gastroenterol, 2003;125(2):452–9.
  89. Kano, M., Shoda, J., Sumazaki, R., Oda, K., Nimura, Y., and Tanaka, N.
Mutations identified in the human multidrug resistance P‐glycoprotein 3
(ABCB4) gene in patients with primary hepatolithiasis. Hepatol Res,
2004;29(3):160–6.
  90. Shoda, J., Oda, K., Suzuki, H. et  al. Etiologic significance of defects in
cholesterol, phospholipid, and bile acid metabolism in the liver of patients
with intrahepatic calculi. Hepatol, 2001;33(5):1194–205.
  91. Lang, C., Meier, Y., Stieger, B. et al. Mutations and polymorphisms in the
bile salt export pump and the multidrug resistance protein 3 associated with
drug‐induced liver injury. Pharmacogenet Genomics, 2007;17(1):47–60.
  92. Bramow, S., Ott, P., Thomsen Nielsen, F., Bangert, K., Tygstrup, N., and
Dalhoff, K. Cholestasis and regulation of genes related to drug metabolism
and biliary transport in rat liver following treatment with cyclosporine A
and sirolimus (rapamycin). Pharmakon Toxicol, 2001;89(3):133–9.
  93. Morton, D.H., Salen, G., Batta, A.K. et al. Abnormal hepatic sinusoidal bile
acid transport in an Amish kindred is not linked to FIC1 and is improved by
ursodiol. Gastroenterol, 2000;119(1):188–95.
  94. Sambrotta, M., Strautnieks, S., Papouli, E. et al. Mutations in TJP2 cause
progressive cholestatic liver disease. Nat Genet, 2014;46(4):326–8
  95. Zhou, S., Hertel, P.M., Finegold, M.J. et al. Hepatocellular carcinoma associ-
ated with tight‐junction protein 2 deficiency. Hepatol, 2015;62(6):1914–6.
  96. Knowles, B.C., Roland, J.T., Krishnan, M. et  al. Myosin Vb uncoupling
from RAB8A and RAB11A elicits microvillus inclusion disease. J Clin
Invest, 2014;124(7):2947–62.
  97. Lapierre, L.A., Kumar, R., Hales, C.M. et al. Myosin vb is associated with
plasma membrane recycling systems. Mol Biol Cell, 2001;12(6):1843–57.
  98. Wakabayashi, Y., Dutt, P., Lippincott‐Schwartz, J. et al. Rab11a and myosin
Vb are required for bile canalicular formation in WIF‐B9 cells. Proc Natl
Acad Sci USA, 2005;102(42):15087–92.
  99. Wakabayashi, Y., Lippincott‐Schwartz, J., and Arias I.M. Intracellular traf-
ficking of bile salt export pump (ABCB11) in polarized hepatic cells: con-
stitutive cycling between the canalicular membrane and rab11‐positive
endosomes. Mol Biol Cell, 2004;15(7):3485–96.
100. Muller, T., Hess, M.W., Schiefermeier, N. et al. MYO5B mutations cause
microvillus inclusion disease and disrupt epithelial cell polarity. Nat Genet,
2008;40(10):1163–5.
101. Ruemmele, F.M., Muller, T., Schiefermeier, N. et al. Loss‐of‐function of
MYO5B is the main cause of microvillus inclusion disease: 15 novel muta-
tions and a CaCo‐2 RNAi cell model. Hum Mutat, 2010;31(5):544–51.
102. Thoeni, C.E., Vogel, G.F., Tancevski, I. et al. Microvillus inclusion disease:
loss of Myosin vb disrupts intracellular traffic and cell polarity. Traffic,
2014;15(1):22–42.
103. Girard, M., Lacaille, F., Verkarre, V. et  al. MYO5B and bile salt export
pump contribute to cholestatic liver disorder in microvillous inclusion dis-
ease. Hepatol, 2014;60(1):301–10.
104. Grammatikopoulos, T., Sambrotta, M., Strautnieks, S. et al. Mutations in
DCDC2 (doublecortin domain containing protein 2) in neonatal sclerosing
cholangitis. J Hepatol, 2016;65(6):1179–1187.
105. Girard, M., Bizet, A.A., Lachaux, A. et al. DCDC2 mutations cause neona-
tal sclerosing cholangitis. Hum Mutat, 2016;37(10):1025–9.
106. Carrion‐Castillo, A., Franke, B., and Fisher S.E. Molecular genetics of dys-
lexia: an overview. Dyslexia, 2013;19(4):214–40.
107. Schueler, M., Braun, D.A., Chandrasekar, G. et  al. DCDC2 mutations
cause a renal‐hepatic ciliopathy by disrupting Wnt signaling. Am J Hum
Genet, 2015;96(1):81–92.
107. Shueler, M., Braun, D.A., Chandrasekar, G. et al. DCDC2 mutations cause
a renal‐hepatic ciliopathy by disrupting Wnt signaling. Am J Hum Genet,
2015;96(1):82–92.
108. Alagille, D., Odievre, M., Gautier, M., and Demerges J.P. Hepatic ductular
hypoplasia associated with characteristic facies, vertebral malformations,
retarded physical, mental and sexual development, and cardiac murmur. J
Pediatr 1975;86:63–71.
109. Emerick, K.M., Rand, E.B., Goldmuntz, E., Krantz, I.D., Spinner, N.B.,
and Piccoli D.A. Features of Alagille syndrome in 92 patients: frequency
and relation to prognosis. Hepatol, 1999;29(3):822–9.
110. Chiaretti, A., Zampino, G., Botto, L., and Polidori, G. Alagille syndrome
and hepatocarcinoma: a case report. Acta Paediatr, 1992;81(11):937.
111. Békássy, A.N., Garwicz, S., Wiebe, T., Homestand, I., and Jensen O.A.
Hepatocellular carcinoma associated with arteriohepatic dysplasia in a 4‐
year‐old girl. Med Pediatr Oncol, 1992;20(1):78–83.
112. Kamath, B.M., Bason, L., Piccoli, D.A., Krantz, I.D., and Spinner N.B.
Consequences of JAG1 mutations. J Med Genet, 2003;40(12):891–5.
113. Oda, T., Elkahloun, A.G., Pike, B.L. et al. Mutations in the human Jagged1
gene are responsible for Alagille syndrome. Nat Genet,
1997;16(3):235–42.
114. Warthen, D.M., Moore, E.C., Kamath, B.M. et al. Jagged1 (JAG1) muta-
tions in Alagille syndrome: increasing the mutation detection rate. Hum
Mutat, 2006;27(5):436–43.
115. Lanford, P.J., Lan, Y., Jiang, R. et al. Notch signalling pathway mediates hair
cell development in mammalian cochlea. Nat Genet, 1999;21(3):289–92.
116. Davies, J.A. and Fisher C.E. Genes and proteins in renal development. Exp
Nephrol, 2002;10(2):102–13.
117. Hu, Q.D., Cui, X.Y., Ng, Y.K., and Xiao, Z.C. Axoglial interaction via the
notch receptor in oligodendrocyte differentiation. Ann Acad Med Singapore,
2004;33(5):581–8.
118. Bray, S.J. Notch signalling: a simple pathway becomes complex. Nat Rev
Mol Cell Biol, 2006;7(9):678–89.
119. Ehebauer, M., Hayward, P., and Arias A.M. Notch, a universal arbiter of
cell fate decisions. Science, 2006;314(5804):1414–5.
120. McCright B., Gao, X., Shen, L. et al. Defects in development of the kidney,
heart and eye vasculature in mice homozygous for a hypermorphic Notch2
mutation. Development, 2001;128(4):491–502.
121. McCright B., Lozier, J., and Gridley, T. A mouse model of Alagille syn-
drome: Notch2 as a genetic modifier of Jag1 haploinsufficiency.
Development, 2002;129(4):1075–82.
122. McDaniell R., Warthen, D.M., Sanchez‐Lara, P.A. et al. NOTCH2 muta-
tions cause Alagille syndrome, a heterogeneous disorder of the notch sign-
aling pathway. Am J Hum Genet, 2006;79(1):169–73.
123. Lozier, J., McCright B., and Gridley, T. Notch signaling regulates bile duct
morphogenesis in mice. PLoS ONE, 2008;3(3):e1851.
124. Gissen, P., Johnson, C.A., Morgan, N.V. et al. Mutations in VPS33B., encoding
a regulator of SNARE‐dependent membrane fusion, cause arthrogryposis‐renal
dysfunction‐cholestasis (ARC) syndrome. Nat Genet, 2004;36(4):400–4.
125. Cullinane, A.R., Straatman‐Iwanowska, A., Zaucker, A. et al. Mutations in
VIPAR cause an arthrogryposis, renal dysfunction and cholestasis syn-
drome phenotype with defects in epithelial polarization. Nat Genet,
2010;42(4):303–12.
126. Banushi, B., Forneris, F., Straatman‐Iwanowska, A. et  al. Regulation of
post‐Golgi LH3 trafficking is essential for collagen homeostasis. Nat
Commun, 2016;7:12111.
362 THE LIVER:  REFERENCES
127. Lo, B., Li, L., Gissen, P. et al. Requirement of VPS33B., a member of the
Sec1/Munc18 protein family, in megakaryocyte and platelet alpha‐granule
biogenesis. Blood, 2005;106(13):4159–66.
128. Bem, D., Smith, H., Banushi, B. et al. VPS33B regulates protein sorting
into and maturation of α‐granule progenitor organelles in mouse megakar-
yocytes. Blood, 2015;126(2):133–43.
129. Hershkovitz D., Mandel, H., Ishida‐Yamamoto, A. et al. Defective lamellar
granule secretion in arthrogryposis, renal dysfunction, and cholestasis syn-
drome caused by a mutation in VPS33B. Arch Dermatol, 2008;144(3):
334–40.
130. Rogerson, C. and Gissen, P. VPS33B and VIPAR are essential for epider-
mal lamellar body biogenesis and function. Biochim Biophys Acta Mol
Basis Dis, 2018;1864(5):1609–21.
131. Horslen, S.P., Quarrell, O.W., and Tanner M.S. Liver histology in the
arthrogryposis multiplex congenita, renal dysfunction, and cholestasis
(ARC) syndrome: report of three new cases and review. J Med Genet,
1994;31(1):62–4.
132. Bull, L.N., Mahmoodi, V., Baker, A.J. et  al. VPS33B mutation with ich-
thyosis, cholestasis, and renal dysfunction but without arthrogryposis:
incomplete ARC syndrome phenotype. J Pediatr, 2006;148(2):269–71.
133. Hanley, J., Dhar, D.K., Mazzacuva, F. et al. Vps33b is crucial for structural
and functional hepatocyte polarity. J Hepatol, 2017;66(5):1001–11.
134. Wang C., Cheng, Y., Zhang, X. et  al. Vacuolar protein sorting 33B is a
tumor suppressor in hepatocarcinogenesis. Hepatol, 2018;68(6):2239–53.
135. Gissen, P., Tee, L., Johnson, C.A. et al. Clinical and molecular genetic fea-
tures of ARC syndrome. Hum Genet, 2006;120(3):396–409.
136. Akbar, M.A., Mandraju, R., Tracy, C., Hu, W., Pasare, C., and Krämer, H.
ARC syndrome‐linked Vps33B protein is required for inflammatory endo-
somal maturation and signal termination. Immunity, 2016;45(2):267–79.
137. Deal, J.E., Barratt, T.M., and Dillon, M.J. Fanconi syndrome, ichthyosis,
dysmorphism, jaundice and diarrhoea‐‐a new syndrome. Pediatr Nephrol,
1990;4(4):308–13.
138. Hadj‐Rabia, S., Baala, L., Vabres, P. et  al. Claudin‐1 gene mutations in
neonatal sclerosing cholangitis associated with ichthyosis: a tight junction
disease. Gastroenterol, 2004;127(5):1386–90.
139. Feldmeyer, L., Huber, M., Fellmann, F., Beckmann, J.S., Frenk, E., and
Hohl, D. Confirmation of the origin of NISCH syndrome. Hum Mutat,
2006;27(5):408–10.
140. Szepetowski, S., Lacoste, C., Mallet, S., Roquelaure, B., Badens, C., and
Fabre, A. NISCH syndrome, a rare cause of neonatal cholestasis: a case
report. Arch Pediatr, 2017;24(12):1228–34.
141. Baala, L., Hadj‐Rabia, S., Hamel‐Teillac, D. et al. Homozygosity mapping
of a locus for a novel syndromic ichthyosis to chromosome 3q27‐q28.
J Invest Dermatol, 2002;119(1):70–6.
142. Mazzon, E., Puzzolo, D., Caputi, A.P., and Cuzzocrea, S. Role of IL‐10 in
hepatocyte tight junction alteration in mouse model of experimental colitis.
Mol Med, 2002;8(7):353–66.
143. Maly, I.P. and Landmann, L. Bile duct ligation in the rat causes upregula-
tion of ZO‐2 and decreased colocalization of claudins with ZO‐1 and occlu-
din. Histochem Cell Biol, 2008;129(3):289–99.
144. Fickert, P., Fuchsbichler, A., Marschall, H.U. et al. Lithocholic acid feeding
induces segmental bile duct obstruction and destructive cholangitis in mice.
Am J Pathol, 2006;168(2):410–22.
145. Anderson, J.M. Leaky junctions and cholestasis: a tight correlation.
Gastroenterol, 1996;110(5):1662–5.
146. Mottino, A.D., Hoffman, T., Crocenzi, F.A., Sánchez Pozzi, E.J., Roma,
M.G., and Vore, M. Disruption of function and localization of tight junc-
tional structures and Mrp2 in sustained estradiol‐17beta‐D‐glucuronide‐
induced cholestasis. Am J Physiol Gastrointest Liver Physiol, 2007;293(1):
G391–402.
147. Kojima, T., Yamamoto, T., Murata, M., Chiba, H., Kokai, Y., and Sawada,
N. Regulation of the blood‐biliary barrier: interaction between gap and
tight junctions in hepatocytes. Med Electron Microsc, 2003;36(3):157–64.
148. Furuse, M., Hata, M., Furuse, K. et  al. Claudin‐based tight junctions
are crucial for the mammalian epidermal barrier: a lesson from claudin‐1‐
deficient mice. J Cell Biol, 2002;156(6):1099–111.
149. Smeitink, J., van den Heuvel, L., and DiMauro S. The genetics and pathol-
ogy of oxidative phosphorylation. Nat Rev Genet, 2001;2(5):342–52.
150. Visapää, I., Fellman, V., Vesa, J. et al. GRACILE syndrome, a lethal meta-
bolic disorder with iron overload, is caused by a point mutation in BCS1L.
Am J Hum Genet, 2002;71(4):863–76.
151. Fellman, V., Lemmelä, S., Sajantila, A., Pihko, H., and Järvelä, I. Screening
of BCS1L mutations in severe neonatal disorders suspicious for mitochon-
drial cause. J Hum Genet, 2008;53(6):554–8.
152. Baker, R.A., Priestley, J.R.C., Wilstermann, A.M., Reese, K.J., and Mark
P.R. Clinical spectrum of BCS1L Monopathies and their underlying struc-
tural relationships. Am J Med Genet A, 2018.
153. Fernandez‐Vizarra, E., Bugiani, M., Goffrini, P. et al. Impaired complex III
assembly associated with BCS1L gene mutations in isolated mitochondrial
encephalopathy. Hum Mol Genet, 2007;16(10):1241–52.
154. Hinson, J.T., Fantin, V.R., Schönberger, J. et al. Missense mutations in the
BCS1L gene as a cause of the Björnstad syndrome. N Engl J Med, 2007;
356(8):809–19.
155. De Meirleir, L., Seneca, S., Damis, E. et al. Clinical and diagnostic charac-
teristics of complex III deficiency due to mutations in the BCS1L gene. Am
J Med Genet A, 2003;121A(2):126–31.
156. Selvaag, E. Pili torti and sensorineural hearing loss. A follow‐up of
Bjørnstad’s original patients and a review of the literature. Eur J Dermatol,
2000;10(2):91–7.
157. Purhonen J,, Rajendran, J., Mörgelin, M. et al. Ketogenic diet attenuates
hepatopathy in mouse model of respiratory chain complex III deficiency
caused by a Bcs1l mutation. Sci Rep,. 2017;7(1):957.
158. Sase, M., Kobayashi, K., Imamura, Y. et al. Level of translatable messenger
RNA coding for argininosuccinate synthetase in the liver of the patients
with quantitative‐type citrullinemia. Hum Genet, 1985;69(2):130–4.
159. Saheki, T., Nakano, K., Kobayashi, K. et al. Analysis of the enzyme abnor-
mality in eight cases of neonatal and infantile citrullinaemia in Japan.
J Inherit Metab Dis, 1985;8(3):155–6.
160. Kobayashi, K., Sinasac, D.S., Iijima, M. et al. The gene mutated in adult‐
onset type II citrullinaemia encodes a putative mitochondrial carrier pro-
tein. Nat Genet, 1999;22(2):159–63.
161. Ohura, T., Kobayashi, K., Tazawa, Y. et al. Clinical pictures of 75 patients
with neonatal intrahepatic cholestasis caused by citrin deficiency (NICCD).
J Inherit Metab Dis, 2007;30(2):139–44.
162. Tabata, A., Sheng, J.S., Ushikai, M. et al. Identification of 13 novel muta-
tions including a retrotransposal insertion in SLC25A13 gene and fre-
quency of 30 mutations found in patients with citrin deficiency. J Hum
Genet, 2008;53(6):534–45.
163. Kobayashi, K., Iijima, M., Ushikai, M., Ikeda, S., and Saheki, T. Citrin
deficiency. J Jpn Pediatr Soc, 2006;110:1047–59.
164. Soeda, J., Yazaki, M., Nakata, T. et al. Primary liver carcinoma exhibiting
dual hepatocellular‐biliary epithelial differentiations associated with citrin
deficiency: a case report. J Clin Gastroenterol, 2008;42(7):855–60.
165. Sinasac, D.S., Moriyama, M., Jalil, M.A. et  al. Slc25a13‐knockout mice
harbor metabolic deficits but fail to display hallmarks of adult‐onset type II
citrullinemia. Mol Cell Biol, 2004;24(2):527–36.
166. Weber, A.M., Tuchweber, B., Yousef, I. et al. Severe familial cholestasis in
North American Indian children: a clinical model of microfilament dys-
function? Gastroenterol, 1981;81(4):653–62.
167. Chagnon, P., Michaud, J., Mitchell, G. et al. A missense mutation (R565W)
in cirhin (FLJ14728) in North American Indian childhood cirrhosis. Am
J Hum Genet, 2002;71(6):1443–9.
168. Freed, E.F. and Baserga, S.J. The C‐terminus of Utp4, mutated in childhood
cirrhosis, is essential for ribosome biogenesis. Nucleic Acids Res,
2010;38(14):4798–806.
169. Freed, E.F., Prieto, J.L., McCann, K.L., McStay, B., and Baserga, S.J.
NOL11, implicated in the pathogenesis of North American Indian childhood
cirrhosis, is required for pre‐rRNA transcription and processing. PLoS
Genet, 2012;8(8):e1002892.
170. Aagenaes, O., Sigstad, H., and Bjorn‐Hansen, R. Lymphoedema in
hereditary recurrent cholestasis from birth. Arch Dis Child,1970;
­
45(243):690–5.
171. Drivdal, M., Trydal, T., Hagve, T.A., Bergstad, I., and Aagenaes, O.
Prognosis, with evaluation of general biochemistry, of liver disease in lym-
phoedema cholestasis syndrome 1 (LCS1/Aagenaes syndrome). Scand
J Gastroenterol, 2006;41(4):465–71.
172. Iversen, K., Drivdal, L.M., Billaud Feragen, K.J., and Geirdal, A.Ø. Quality
of life in adults with lymphedema cholestasis syndrome 1. Health Qual Life
Outcomes, 2018;16(1):146.
173. Bull, L.N., Roche, E., Song, E.J. et al. Mapping of the locus for cholestasis‐
lymphedema syndrome (Aagenaes syndrome) to a 6.6‐cM interval on
­chromosome 15q. Am J Hum Genet, 2000;67(4):994–9.
 29:  Molecular Cholestasis 363
174. Viveiros, A., Reiterer, M., Schaefer, B. et al. CCBE1 mutation causing scle-
rosing cholangitis: Expanding the spectrum of lymphedema‐cholestasis
syndrome. Hepatol, 2017;66(1):286–288.
175. Shah, S., Conlin, L.K., Gomez, L. et al. CCBE1 mutation in two siblings,
one manifesting lymphedema‐cholestasis syndrome, and the other, fetal
hydrops. PLoS One,. 2013;8(9):e75770.
176. Alders, M., Hogan, B.M., Gjini, E. et al. Mutations in CCBE1 cause gener-
alized lymph vessel dysplasia in humans. Nat Genet, 2009;41(12):1272–4.
177. Hogan, B.M., Bos, F.L., Bussmann, J. et al. Ccbe1 is required for embryonic
lymphangiogenesis and venous sprouting. Nat Genet, 2009;41(4):396–8.
178. Clayton, P.T. Diagnosis of inherited disorders of liver metabolism. J Inherit
Metab Dis, 2003;26(2–3):135–46.
179. Balistreri, W.F. Neonatal cholestasis. J Pediatr, 1985;106(2):171–84.
180. Dick, M.C. and Mowat, A.P. Hepatitis syndrome in infancy – an epidemio-
logical survey with 10 year follow up. Arch Dis Child,1985;60(6):512–6.
181. Ozkan, T.B., Mistik, R., Dikici, B., and Nazlioglu H.O. Antiviral therapy in
neonatal cholestatic cytomegalovirus hepatitis. BMC Gastroenterol, 2007;7:9.
182. Sokol, R.J., Shepherd, R.W., Superina, R., Bezerra, J.A., Robuck, P., and
Hoofnagle J.H. Screening and outcomes in biliary atresia: summary of a
National Institutes of Health workshop. Hepatol, 2007;46(2):566–81.
183. Balistreri, W.F., Grand, R., Hoofnagle, J.H., Suchy, F.J., Ryckman, F.C.,
Perlmutter, D.H., et al. Biliary atresia: current concepts and research direc-
tions. Summary of a symposium. Hepatol, 1996;23(6):1682–92.
184. Sokol, R.J., Mack, C., Narkewicz, M.R., and Karrer F.M. Pathogenesis and
outcome of biliary atresia: current concepts. J Pediatr Gastroenterol Nutr,
2003;37(1):4–21.
185. Sokol, R.J. and Mack, C. Etiopathogenesis of biliary atresia. Semin Liver
Dis, 2001;21(4):517–24.
186. Ohi, R. Surgery for biliary atresia. Liver, 2001;21(3):175–82.
187. Davenport, M., Savage, M., Mowat, A.P., and Howard E.R. Biliary atresia
splenic malformation syndrome: an etiologic and prognostic subgroup.
Surgery, 1993;113(6):662–8.
188. Karrer, F.M. and Lilly J.R. Correction of biliary atresia and jejunal atresia
in an infant. J Pediatr Surg, 1993;168(4):469–76.
190. Roach, J.P. and Bruny J.L. Advances in the understanding and treatment of
biliary atresia. Curr Opin Pediatr, 2008;20(3):315–9.
191. Gershoni‐Baruch, R., Gottfried, E., Pery, M., Sahin, A., and Etzioni, A.
Immotile cilia syndrome including polysplenia, situs inversus, and extrahe-
patic biliary atresia. Am J Med Genet, 1989;33(3):390–3.
192. Mazziotti, M.V., Willis, L.K., Heuckeroth, R.O. et al. Anomalous development
of the hepatobiliary system in the Inv mouse. Hepatol, 1999;30(2):372–8.
193. Shimadera, S., Iwai, N., Deguchi, E., Kimura, O., Fumino, S., and
Yokoyama, T. The inv mouse as an experimental model of biliary atresia. J
Pediatr Surg, 2007;42(9):1555–60.
194. Otto, E.A., Schermer, B., Obara, T., O’Toole J.F., Hiller, K.S., Mueller,
A.M. et al. Mutations in INVS encoding inversin cause nephronophthisis
type 2, linking renal cystic disease to the function of primary cilia and left‐
right axis determination. Nat Genet, 2003;34(4):413–20.
195. Olbrich, H., Fliegauf, M., Hoefele, J. et  al. Mutations in a novel gene,
NPHP3, cause adolescent nephronophthisis, tapeto‐retinal degeneration
and hepatic fibrosis. Nat Genet, 2003;34(4):455–9.
196. Johnson, C.A., Gissen, P., and Sergi, C. Molecular pathology and genetics
of congenital hepatorenal fibrocystic syndromes. J Med Genet, 2003;40(5):
311–9.
197. Adams, M., Smith, U.M., Logan, C.V., and Johnson C.A. Recent advances
in the molecular pathology, cell biology and genetics of ciliopathies. J Med
Genet, 2008;45(5):257–67.
198. Ibañez‐Tallon, I., Heintz, N., and Omran, H. To beat or not to beat: roles of
cilia in development and disease. Hum Mol Genet, 2003;12(1):R27–35.
199. Lina, F. and Satlinb, L.M. Polycystic kidney disease: the cilium as a com-
mon pathway in cystogenesis. Curr Opin Pediatr, 2004;16(2):171–6.
200. Phillips, M.J., Azuma, T., Meredith, S.L. et al. Abnormalities in villin gene
expression and canalicular microvillus structure in progressive cholestatic
liver disease of childhood. Lancet, 2003;362(9390):1112–9.
201. Nicastro, E. and D’Antiga, L. Next generation sequencing in pediatric
hepatology and liver transplantation. Liver Transpl, 2018;24(2):282–93.
202. Chen, H.L., Li, H.Y., Wu, J.F. et al. Panel‐based next‐generation sequenc-
ing for the diagnosis of cholestatic genetic liver diseases: clinical utility and
challenges. J Pediatr, 2018;171:171–7.
203. Togawa, T., Sugiura, T., Ito, K. et al. Molecular genetic dissection and neona-
tal/infantile intrahepatic cholestasis using targeted next‐generation sequenc-
ing. J Pediatr, 2016;171:171–7.
 Pathophysiologic Basis for
30 Alternative Therapies for
Cholestasis
Claudia D. Fuchs, Emina Halilbasic, and Michael Trauner
Hans Popper Laboratory of Molecular Hepatology, Division of Gastroenterology and Hepatology,
Department of Internal Medicine III, Medical University of Vienna, Austria
INTRODUCTION
Cholestasis is characterized by impaired bile formation with
insufficient amounts of bile reaching the duodenum resulting in
intrahepatic and systemic accumulation of bile acids (BAs) and
other potentially toxic cholephiles [1]. Important causes include
disturbances of hepatocellular and/or cholangiocellular bile
secretion and destruction/obstruction of smaller and/or larger
bile ducts by mechanical processes (e.g. stones, tumors) or
immune‐mediated fibrosing cholangiopathies such as primary
biliary cholangitis (PBC) and primary sclerosing cholangitis
(PSC). Ideally, causal therapy of cholestasis resolves the
­underlying cholestatic insult and promotes recovery which is
currently possible only in some cases (e.g. removal of stone,
surgical removal or stenting of tumor/stricture, delivery in intra-
hepatic cholestasis of pregnancy, discontinuation of causative
drug in drug‐induced liver injury [DILI]). Even when the under-
lying cause has been resolved, recovery can be slow (e.g. persis-
tent hepatocellular secretory failure after DILI) requiring
supportive therapy. Since the etiologies of many cholestatic dis-
orders such as PBC and PSC are not fully understood and causal
therapies are still lacking, stimulation of bile secretion and
adaptive mechanisms may attenuate liver injury irrespective of
the cause of cholestasis. Of note, targeting the immune mecha-
nisms underlying or at least contributing to the pathogenesis of
PBC and PSC has so far been rather disappointing.
The first‐line therapy for many cholestatic liver diseases is
ursodeoxycholic acid (UDCA), a hydrophilic BA that reduces
the toxicity of the endogenous BA pool and has several
­additional beneficial therapeutic mechanisms [2]. As a potent
intracellular signaling molecule UDCA stimulates hepatobiliary
secretion (mainly by promoting vesicular targeting of
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
transporters in a post‐transcriptional manner) thereby counter-
acting cholestasis [2]. In addition, UDCA stimulates biliary
bicarbonate (HCO3−) secretion supporting the “biliary HCO3−
umbrella” [3], an important protective mechanism for hepato-
cytes and cholangiocytes against highly toxic BA monomers
in  bile by maintaining alkaline pH along the apical surface.
In  addition, UDCA may also promote biliary phospholipid
secretion thereby facilitating the formation of mixed micelles
and limiting the amounts of free, non‐micellar bound BAs. As
chaperone UDCA also reduces endoplasmatic reticulum stress,
exerts anti‐apoptotic effects, and its activation of the glucocorti-
coid receptor (GR/NR3C1) may be responsible for some direct
anti‐inflammatory effects, it further contributes to the cytopro-
tective action of UDCA in cholestatic liver disease [2].
As such, UDCA has found application in the treatment of a
broad range of cholestatic diseases but it has been approved
only as first‐line treatment for PBC, where UDCA improves
survival [4]. In other cholestatic disorders, such as PSC, UDCA
often improves serum liver tests, without proven survival bene-
fit. Increasing its dose (28–30 mg kg−1/day) was even harmful in
PSC [2] although the mechanisms are still incompletely under-
stood. However, amelioration in liver enzymes upon UDCA
treatment (in moderate dose) and their exacerbation after UDCA
withdrawal may point toward potential beneficial effects. The
field has made impressive progress since introduction of UDCA
[2]. The identification of specific BA transport systems and
dedicated BA receptors [5] now allows a more specific targeting
of impaired BA homeostasis (including adaptive/alternative
transport, metabolic pathways, and regulatory networks) irre-
spective of the etiology of cholestasis. This chapter will focus
on the pathophysiologic basis for alternative therapies for chol-
estasis beyond UDCA, the first available drug for cholestasis.
 30:  Pathophysiologic Basis for Alternative Therapies for Cholestasis 365
NUCLEAR RECEPTORS
Nuclear receptors (NRs) are the largest family of transcription
factors comprising 48 members in humans and 49 members in
mice, which share a common structural organization and are
activated upon ligand binding [6]. They coordinate several key
hepatic functions including the regulation of BA synthesis and
hepatobiliary excretory function, glucose and lipid metabolism,
inflammatory processes, fibrosis as well as liver regeneration,
and tumorigenesis [7]. After ligand (e.g. BA) binding, NRs
change their conformation, facilitating the recruitment of
­cofactors and dissociation of corepressors, implementing DNA
binding and stimulating transcription of genes typically involved
in metabolism and/or transport of the ligand (e.g. BAs) thus
constituting the molecular basis for feedback regulation of
metabolism [6]. Due to their central role in hepatic (patho)phys-
iology, NRs have become attractive targets for anticholestatic
drugs serving as ligands for these receptors. The most relevant
NRs involved in BA signaling and hepatobiliary homeostasis
include the farnesoid X receptor (FXR/NR1H4), pregnane
X  receptor (PXR/NR1I2), constitutive androstane receptor
(CAR/NR1I3), and vitamin D receptor (VDR/NR1I1).
Furthermore, the glucocorticoid receptor (GR/NR3C1) and the
fatty acid activated peroxisome proliferator‐activated receptors
(PPARα, PPARγ, and PPARδ) as regulators of inflammation,
fibrosis, and energy metabolism may also impact on biliary
homeostasis and thereby counteract cholestatic liver injury.
Notably the current standard therapy of cholestasis, UDCA,
does not activate FXR (but may even counteract its activation by
reducing the relative concentration of BAs with more potent
FXR ligand function [8]), and has low affinity to GR and VDR
mediating its anti‐inflammatory effects and may indirectly acti-
vate PXR after being metabolized into lithocholic acid (LCA) [2].
Nuclear bile acid receptor (FXR)
FXR serves as the main NR for BAs [8–10] regulating their
­synthesis and hepatic uptake as well as elimination from the
liver. It forms heterodimers with the retinoid X receptor (RXR,
NR2B), binding to the inverted repeat (IR)‐1 within the pro-
moter sequence of target genes. FXR is predominantly present
in liver, gastrointestinal tract, kidney, and adrenal glands acting
as a central regulator of BA homeostasis, lipid, glucose [11],
and amino acid metabolism [12] as well as inflammation [7] and
liver regeneration [13].
FXR has a key role in controlling BA synthesis by regulating
the expression of rate limiting enzymes of both classical
(CYP7A1) and alternative pathways (CYP8B1 and CYP27A1).
This is mediated by induced expression of two key regulators,
an orphan nuclear receptor, small heterodimer partner (SHP,
NR0B2) [11] and an intestinal hormone, fibroblast growth fac-
tor 15/19 (human FGF19, mouse Fgf15) [14, 15]. BA‐activated
FXR induces the hepatic expression of SHP, an atypical nuclear
receptor lacking a DNA binding domain that functions as a
potent transcriptional repressor, which in turn, interacts with
transcription factors such as hepatocyte nuclear factor (HNF)‐4α/
NR2A1 and liver receptor homolog (LRH)‐1/NR5A2 to inhibit
their function and suppress CYP7A1 expression [16]. The ileal
hormone FGF15/19 is secreted in the portal circulation upon
BA activation of FXR in enterocytes and reaches the liver,
where it binds to fibroblast growth factor receptor 4 (FGFR4)/β‐
Klotho complex and activates c‐Jun N‐terminal kinase‐­
dependent pathway which ultimately leads to CYP7A1
repression [15] (Figure  30.1). Of note, SHP is required for
FGF15/19 to efficiently repress BA synthesis and mice lacking
SHP are refractory to the inhibitory effects of either FXR
­agonists or FGF15/19 on Cyp7a1 expression [7]. FXR further
fine‐tunes BA levels through induction of other target genes,
including transcription factor V‐Maf avian musculoaponeu-
rotic  fibrosarcoma oncogene homolog G (MAFG), a global
transcriptional repressor of BA synthesis, lysine‐specific his-
tone demethylase (LSD1), a key repressive epigenetic compo-
nent in a SHP complex, and β‐Klotho, the obligate co‐receptor
for FGF15/19 [17].
Apart from regulation of BA synthesis, the intrahepatic
BA  concentration is controlled by FXR‐mediated inhibition
of  the basolateral BA uptake transporter sodium/taurocholate
co‐transporting polypeptide (NTCP/SLC10A1) and upregula-
tion of the canalicular BA transporter bile salt export pump
(BSEP/ABCB11) in hepatocytes (Figure 30.1), promoting BA
elimination via induction of bile flow. Beside this induction of
BA dependent bile flow, FXR also increases BA‐independent
bile flow promoting biliary HCO3− transport by facilitating
complex formation of carboanhydrase 14 and anion exchanger 2
(AE2) [18]. FXR also induces alternative basolateral BA trans-
port in hepatocytes via organic solute transporter alpha and beta
(OSTα/β; SLC51A/SLC51B) and detoxification via enzymes
such as Cyp3a11/CYP3A4, Sult2a1/SULT2a1 (sulfotransferase
family 2A member 1), and Ugt2b4/UGT2b4 [7] further prevent-
ing the accumulation of BAs within target cells (Figure 30.1).
In the liver, FXR is also present in cells other than hepato-
cytes. In cholangiocytes FXR regulates ductular bile formation
by alkalization and fluidization via induction of vasoactive
intestinal polypeptide receptor 1 (VPAC‐1) [7]. Moreover, FXR
activation in cholangiocytes induces expression of the antimi-
crobial peptide cathelicidin, indicating an important role of
BAs/FXR in maintaining sterility of biliary tree and protection
against bile duct inflammation [19].
Activation of FXR has antifibrotic effects [18], the underly-
ing mechanisms and FXR expression in hepatic stellate cells
(HSC) are discussed controversially. FXR has been reported to
have a direct antifibrotic effect via activation of SHP in HSC
[7], but other studies, however, have revealed only very low or
even no FXR and SHP expression in human HSCs and murine
periductal myofibroblasts [20], arguing for a more indirect anti-
fibrotic effect through control of cholestasis and inflammation.
Present also in liver sinusoidal endothelial cells [21], FXR
ameliorates portal hypertension not by targeting fibrosis, but by
directly interfering with endothelial function. FXR activation
reduces intrahepatic vascular resistance by upregulation of
eNOS, cystathionase, and dimethylaminohydrolase 1, promoting
the generation of vasodilators such as nitric oxide and hydrogen
sulfide, as well as a reduction of intrahepatic vasoconstrictors
endothelin‐1 and p moesin [18].
Mice lacking FXR display increased hepatic inflammation
at  baseline and are more susceptible to lipopolysaccharide
(LPS)‐induced inflammation [18]. FXR may have direct anti‐
inflammatory effects by interference with nuclear factor
366 THE LIVER:  NUCLEAR RECEPTORS

MRP3

MRP4

OSTβ
OSTα
Hepatic stellate cell
FXR PXR PPARα
PXR PXR CAR
GR FXR
Inflammation
e.g. NFκB
Alternative BA export
FXR secretion of hydroxylated &
FGF βKlotho FGF BSEP BAs sulfated BAs
15/19
FGFR4 15/19 PPARγ PXR VDR

FXR
GR Fibrosis
MDR2/3 PL
PPARγ: MCP1, migration, proliferation
JNK PPARα HCO3 - AE2 PXR: TGF1β, transdifferentiation, proliferation
BAs NTCP
VDR: SMAD pathway

FXR FXR PXR CAR PPAR

FXR conjugated α/δ Immune cells

MRP2
Bilirubin BA detoxification
SHP PPAR PXR CYPs, SULTs, UGTs
BA synthesis α/δ
CYP7A1 Macrophage CD4+ T cell
Hepatocyte
PPARα PPARγ PPARδ VDR
PPARγ
FXR VDR GR
Reactive cholangiocyte Inflammation
Phenotype
Cholangioprotective PPARα: phosphorylation of p38 in CD4+ T cells
VCAM1 VPAC1, AE2, Cathelecithin PPARγ: cytocine secretion from macrophages
PPARδ: M2 activation of macropages
VDR: production of Tregs
Cholangiocyte

Figure 30.1  Nuclear receptors as therapeutic targets of alternative therapies in cholestasis. Nuclear receptor activation in hepatocytes maintains
the balance between BA synthesis, detoxification, uptake, and excretion by regulation of expression of hepatobiliary transporters and metabolic
pathways. FXR (in a SHP dependent manner) represses hepatic BA uptake (NTCP) and BA synthesis (CYP7A1). Moreover intestine‐derived
FGF15/19 to the FGFR4/bKlotho dimer counteracts CYP7A1 expression. Of note, human hepatocytes (in addition to enterocytes) also express
FGF19. FXR promotes BA secretion via induction of canalicular transporters (BSEP, MRP2, and MDR2/3) and induces BA elimination via alterna-
tive basolateral BA transporter OSTα/β. CAR and PXR regulate the adaption to increased intracellular BA concentrations via upregulation of MRP3
and 4 (alternative BA export) as well as induction of detoxification/hydroxylation enzymes. PPARα stimulates phospholipid secretion (via MDR2/3)
and is also involved in regulation of detoxification pathways. Activation of AE2 expression by GR stimulates biliary bicarbonate secretion, thus
reducing bile toxicity. Apart from regulating BA homeostasis, NRs have additional anti‐inflammatory and antifibrotic effects (right‐hand panel).
Their activation may result in induction of defensive mechanisms in bile duct epithelial cells. Activation of PPARγ in cholangiocytes reduces vas-
cular cell adhesion molecule (VCAM‐1) expression, thereby counteracting reactive cholangiocyte phenotype. Activation of FXR, VDR, and GR in
cholangiocytes exert cholangioprotective effects via upregulation of VPAC1, AE2, and Cathelicidin, respectively. Antifibrotic effects of NRs are
related to their activation in hepatic stellate cells. PPARγ, PXR, and VDR reduce expression of profibrogenic genes such as MCP1, TGFβ1, and
SMAD, respectively. Furthermore, they reduce migration, proliferation, as well as transdifferentiation of HSC into myofibroblasts. Anti‐inflamma-
tory effects of NRs are related to their activation in immune cells such as macrophages and CD4+ T cells. Green arrows indicate stimulatory effects
and red lines indicate suppressive effects. AE, anion exchanger; BAs, bile acids; Bili‐glu, bilirubin glucuronide; BSEP, bile salt export pump; CAR,
constitutive androstane receptor; CYP7A1, cholesterol‐7α‐hydroxylase; CYPs, cytochrome P450 enzymes; FGF, fibroblast growth factor; FXR,
farnesoid X receptor; GR, glucocorticoid receptor; HSC, hepatic stellate cell; MDR3, multidrug resistance protein 3; MRP2, multidrug resistance‐
associated protein 2; MRP3, multidrug resistance‐associated protein 3; MRP4, multidrug resistance‐associated protein 4; NTCP, sodium taurocho-
late cotransporting polypeptide; OSTα/β, organic solute transporter α and β; PC, phosphatidylcholine; PXR, pregnane X receptor; PPARα,
peroxisome proliferator‐activated receptor α; PPARγ, peroxisome proliferator‐activated receptor γ; SHP, small heterodimer partner; SULTs, sulfa-
tation enzymes; UGTs, glucuronidation enzymes; VDR, vitamin D receptor.

kappa‐b (NFκB) [18]. Recently, FXR has also been implicated
in activation of hepatic natural killer T cells and hepatic accu-
mulation of myeloid‐derived suppressor cells, counteracting
immune‐mediated liver injury in rodents [18]. Anti‐­
inflammatory effects of FXR were not only observed in liver,
but also in intestine since pharmacological FXR activation
improved intestinal inflammation and permeability in colitis
models [7]. Reciprocal interactions between BAs and immune
cells may exist, since the BA receptors FXR and TGR5/GBAR‐1
(Takeda G protein‐coupled receptor/G protein‐coupled bile acid
receptor 1) are expressed in macrophages [22]. BAs have been
identified on one hand as damage‐associated molecular patterns
(DAMPs), triggering inflammatory response by binding to
inflammasomes [22] and on the other hand specific BA species
(LCA) have been shown to inhibit inflammasome activation [23].
Beside its direct anti‐inflammatory properties and modulation
of immune cell function, FXR represents a central molecular
switch controlling intestinal microbiota by protecting the gut
from bacterial overgrowth and disruption of the epithelial bar-
rier [24]. Conversely, FXR dysfunction drives bacterial translo-
cation, facilitated by increased intestinal permeability and
intestinal inflammation [18]. Pharmacological activation of
FXR improved gene expression of tight junction proteins such
as claudin‐1 and occludin [25] as well as expression of induci-
ble isoform nitric oxide synthase (iNOS) and angiopoietin
1 (Ang1), genes involved in antibacterial defense by producing
 30:  Pathophysiologic Basis for Alternative Therapies for Cholestasis 367
antimicrobial peptides [26]. Activation (or intestinal overex-
pression [24]) of FXR decreased bacterial overgrowth and
improved liver injury in the bile duct ligation (BDL) [24] as well
as in the α‐naphthyl‐isothiocyanate (ANIT) model [27]. In line,
obstructive jaundice with virtually complete absence of BAs
from the intestinal lumen promotes bacterial translocation in
patients [28].
Loss of FXR results in severe disruption of BA homeostasis
as reflected by increased BA synthesis as well as increased BA
uptake and reduced hepatobiliary BA secretion in FXR knock-
out (KO) mice [29]. Although metabolic adaption to cholestatic
conditions does not entirely rely only on FXR and – at least in
mice – can be partially compensated by CAR/PXR‐dependent
overexpression of detoxifying enzymes and alternative efflux
systems [18, 30], a loss‐of‐function mutation in the FXR gene
results in severe neonatal progressive familial intrahepatic chol-
estasis 5 (PFIC5) [31]. While impaired function of FXR leads to
profound dysregulation of BA homeostasis, pharmacological
activation of FXR has shown multiple beneficial effects in pre-
clinical models of cholestasis. Collectively, FXR: (i) suppresses
hepatic BA synthesis, reduces intestinal and hepatic BA uptake,
while stimulating canalicular and basolateral BA efflux, thus
resulting in a net reduction of the toxic hepatic BA overload; (ii)
alters bile composition by increasing phospholipid and HCO3−
secretion finally resulting in less toxic bile; (iii) mediates direct
anti‐inflammatory effects in hepatocytes and non‐parenchymal
liver cells, together with immunomodulatory effects in the adap-
tive immune system; (iv) impacts on the gut–liver axis by induc-
tion of FGF15/19, a suppressor of BA synthesis; (v) reduces
bacterial overgrowth and intestinal permeability; (vi) amelio-
rates the complications of advanced liver disease such as portal
hypertension and possibly carcinogenesis. These features make
FXR ligands attractive compounds for alternative therapies in
cholestasis and immune‐mediated disorders such as PBC and
PSC including associated/underlying inflammatory bowel
­disease (IBD).
Various steroidal (BA derived) and non‐steroidal (non‐BA
derived) FXR activators have been developed and have shown a
range of beneficial effects in animal models of cholestasis.
More specifically, the non‐steroidal FXR agonist GW4064 and
the steroidal FXR agonist obeticholic acid (OCA/INT747  –  a
6‐ethyl derivative of the primary BA chenodeoxycholic acid
[18]) showed positive effects in chemically induced cholestatic
liver injury (α‐naphthyl isothiocyanate (ANIT) and estradiol
induced cholestasis), in bile duct ligated animals as well as in
the Mdr2 KO mouse model of sclerosing cholangitis [18].
Based on these encouraging preclinical data, FXR ligands
have already entered clinical development. Add‐on treatment
with OCA in PBC patients, who biochemically did not respond
sufficiently to UDCA or OCA monotherapy, improved liver
enzymes and markers of inflammation [7, 32, 33]. These studies
led to conditional approval of OCA as a second‐line therapy for
the treatment of PBC patients with incomplete biochemical
response or intolerance to standard therapy with UDCA, which
is already reflected by current guidelines [4]. Long‐term clinical
effects of OCA treatment in PBC patients are currently under
investigation. Since OCA functions as systemic FXR ligand
undergoing enterohepatic circulation the risk of overdosing in
cirrhotic PBC patients needs to be taken into account. The role
of OCA in other cholestatic liver disorders such as PSC is
undergoing clinical evaluation [34].
Non‐steroidal FXR agonists (such as LJN452/Tropifexor and
GS‐9674) have different pharmacokinetic properties favoring
intestinal (instead of hepatic) FXR activation, and also do not
undergo enterohepatic circulation. Differential targeting of FXR
in hepatocytes, cholangiocytes, and/or enterocytes could criti-
cally determine the pharmacological effects of steroidal versus
non‐steroidal FXR ligands. Whether non‐steroidal FXR ligands
lack possible BA‐structure‐related side‐effects, such as pruritus,
needs to be determined [35]. An open question remains whether
activation of intestinal FXR resulting mainly in inhibition of
BA synthesis via FGF15/19 signaling is sufficient to counteract
cholestatic liver disease or whether more systemic anti‐­
inflammatory effects are necessary.
Another important G‐protein coupled BA receptor, TGR5, is
expressed in cholangiocytes, gallbladder epithelium, as well as
endothelial cells and Kupffer cells (but not in hepatocytes),
where it regulates biliary HCO3− secretion and has anti‐inflam-
matory effects [18] (also see Chapter  24 for more details). In
contrast to FXR [18], TGR5 stimulates production of intestinal
glucagon like peptide 1 (GLP‐1) [7]. GLP‐1 signaling may pro-
tect cholangiocytes from apoptosis and thus modulate their
adaptive response to cholestasis in mice [18]. However, TGR5
expression is reduced in Mdr2 KO mouse model of sclerosing
cholangitis [36] and a pure TGR5 agonist did not improve liver
disease in this model [18]. Additional concerns include a
­potential role of TGR5 in development of pruritus [18], poten-
tial pro‐proliferative and anti‐apoptotic effects in isolated
human cholangiocytes and cholangiocellular carcinoma cells
[18] as well as TGR5 overexpression in human cholangiocarci-
noma [37] (for more details see Chapter 24).
Fibroblast growth factor 19 (FGF19)
Activation of FXR stimulates production of endogenous
FGF15/19 in the ileum, which is transported to the liver where
it binds to FGFR4 and its co‐receptor β‐Klotho in hepatocytes,
contributing to suppression of hepatic BA synthesis via c‐Jun
N‐terminal kinase (JNK) and SHP, in addition to other effects
on lipid and glucose metabolism [15]. Interestingly, FGF19 may
also exert anti‐inflammatory effects by counteracting NFκB
signaling [38]. Selective overexpression of intestinal FXR
improved liver injury in the Mdr2 KO mouse model of scleros-
ing cholangitis via induction of intestinal FGF15 expression and
subsequent reduction of BA pool size. In line, FGF19 applica-
tion to bile duct ligated mice reduced total BA pool size [39],
improved jaundice and serum levels of liver transaminases as
well as necrosis. However, stimulation of endogenous FGF15/19
could also have side‐effects such as pro‐proliferative or pro‐car-
cinogenic effects. Endogenous FGF19 and its receptor FGFR4
have been implicated in development of hepatocellular carci-
noma (HCC) and inhibitors of FGFR4‐related signaling are
even developed as anticancer drugs [18]. As such, FGF19 is
amplified in human HCC and in mice ectopic overexpression of
FGF19 accelerates tumor development [18]. Importantly,
FGF19 mimetics such as M70/NGM282 have amino acid modi-
fications at the N‐terminus allowing the dissection of metabolic
from pro‐proliferative actions thus maintaining their BA
368 THE LIVER:  NUCLEAR RECEPTORS
metabolic regulatory properties while lacking their pro‐­
proliferative effects. In line, adeno‐associated virus mediated
expression of M70 in Mdr2 KO mouse model of sclerosing
cholangitis decreased hepatic inflammation and biliary fibrosis
[40]. Notably, prolonged M70 treatment did not accelerate HCC
formation in the Mdr2 KO mouse and even had tumor preven-
tive effects, while FGF19 application promoted hepatic carcino-
genesis [40]. In PBC patients with insufficient response to
UDCA NGM282 improved cholestatic liver enzymes [40].
Diarrhea, headache, and nausea were observed as side‐effects,
suggesting that FGF19 may also impact on intestinal motility.
Of note, NGM282 has been shown to improve markers of
hepatic fibrosis without reducing cholestatic liver enzymes such
as alkaline phosphatase in patients with PSC [40] suggesting
that FGF19 could also mediate potential direct anti‐inflammatory
and antifibrotic actions. To understand the full (patho)physio-
logical and therapeutic implications of FGF19 it has to be kept
in mind that  –  in contrast to rodents  –  human FGF19 is also
produced by hepatocytes, bile ducts, and gallbladder epithelium
[41]. Interestingly, serum FGF19 correlates positively with
severity of the cholestasis [18]. The extent to which patients
with elevated endogenous levels of FGF19 may benefit
from  exogenous FGF19 mimetic application remains to be
­elucidated. For example, the FGF19 mimetic could also com-
pete with endogenous FGF19 at the FGFR4 binding site [18]
potentially counteracting development of HCC, while maintain-
ing the beneficial metabolic effects on BA homeostasis.
Peroxisome proliferator‐activated receptors
(PPARs)
PPARα (NR1C1), PPARγ (NR1C3), and PPARδ (NR1C2), are
structural homologues which are activated by endogenous fatty
acids and their derivatives [7]. PPARδ is ubiquitously expressed,
while PPARα is predominantly found in organs responsible for
fatty acid catabolism such as liver, heart, kidney, brown adipose
tissue, small and large intestine. PPARγ is highly expressed in
adipose tissue and immune cells [42]. The main function of
PPARs is the regulation of lipid and energy homeostasis.
Beyond the realm of metabolism, PPARs also exert direct anti‐
inflammatory effects.
Apart from being a key regulator of hepatic lipid metabolism,
PPARα is also involved in BA metabolism and BAs can indi-
rectly stimulate human PPARα via FXR [43]. PPAR activation
via bezafibrate represses BA synthesis (through reduction of
HNF4α binding to the CYP7A1 promoter) [7] as well as hepa-
tocellular BA uptake (reduced Ntcp expression by direct mecha-
nisms) [44]. Bezafibrates (PPARα ligand with additional
broader pan PPAR activity) also induce expression of enzymes
involved in BA detoxification (Sult2a1, Ugt2b4, and Ugt1a3)
[7]. In vitro, the PPARα agonist ciprofibrate also induced the
BA uptake transporter ASBT (apical sodium dependent bile salt
transporter) in human cholangiocytes and Caco2 cells [2].
PPARα stimulates expression of multidrug resistant 3 (MDR3)
[44] thereby potentially contributing to elevated biliary
­phospholipid (PL) secretion and counteracting BA toxicity by
promoting formation of mixed micelles.
In addition, PPARα exerts anti‐inflammatory effects by
­transrepressing NFκB signaling via induction of IκBα (which
sequesters NFκB) [45]. Fenofibrate (a more selective PPARα
ligand) downregulates the IL6 receptor subunits gp80 and gp130
in the liver [46], thereby reducing phosphorylation of STAT3
and c‐Jun [47].
In addition to PPARα, PPARγ and PPARδ activation reduces
production of inflammatory mediators and cytokines in various
immune cells including monocytes/macrophages, dendritic
cells, lymphocytes, as well as cholangiocytes [7].
Activation of all PPAR isoforms also has antifibrotic effects.
Although HSC do not show pronounced PPARα expression, its
activation led to reduced HSC activation and repression of
fibrotic markers [48]. PPARγ is expressed only in quiescent
HSC and has antifibrotic effects by suppressing their activation,
proliferation, and migration [49]. Accordingly, loss of PPARγ is
a typical feature of HSC activation [49].Treatment with thiazo-
lidinedione (a PPARγ agonist) improved bile duct proliferation
and hepatic fibrosis in bile duct ligated rats [7]. The plant extract
curcumin, the yellow pigment of the spice turmeric, also acti-
vated PPARγ in cholangiocytes, decreased infiltration of inflam-
matory cells, and improved portal inflammation and fibrosis in
Mdr2 KO mouse model [7]. In contrast to PPARα and PPARγ,
PPARδ activation had pro‐ as well as antifibrotic effects.
Application of KD3010, but not GW501516, (two different
PPARδ agonists) was hepatoprotective and antifibrotic in mice
subjected to BDL and CCl4. In vitro that KD3010, in contrast to
GW501516, increased the expression of several CYP enzymes
resulting in reduced reactive oxygen species production, possi-
bly counteracting hepatic fibrosis [50].
Patients with obstructive jaundice undergoing percutaneous
transhepatic biliary drainage that received bezafibrate treatment
showed increased biliary PL secretion. However, the same study
showed that MDR3 expression is already increased in PBC
patients at baseline and was not further upregulated by bezafi-
brate treatment [51], indicating that bezafibrate may induce
beneficial clinical effects via PL‐independent mechanisms.
Fibrates such as bezafibrate and fenofibrate have shown ben-
eficial effects in several smaller, mostly uncontrolled studies in
patients with PBC (not responding to UDCA) and  –  to lesser
extent – in PSC [52]. Recently, a larger randomized controlled
trial demonstrated pronounced improvement in liver enzymes,
liver stiffness, and pruritus by bezafibrate in PBC patients with
insufficient response to UDCA [53] and fibrates may be consid-
ered a second‐line treatment in PBC [54]. Future long‐term
studies will have to explore the safety and efficacy of fibrates in
PBC and other cholestatic disorders.
A selective and potent PPARδ agonist MBX‐8025, Seladelpar,
reduced alkaline phosphatase in a phase II proof‐of‐concept
study in PBC [55] and further clinical trials are ongoing.
Another promising PPARα/δ dual agonist, elafibranor, with
beneficial effects in non‐alcoholic steatohepatitis (NASH) and
diabetes [56], is currently also being tested in PBC.
Glucocorticoid receptor (GR)
GR/NR3C1 is ubiquitously expressed. Apart from its well‐
known anti‐inflammatory and immunosuppressive function, GR
regulates several metabolic pathways including not only carbo-
hydrate and protein, but also BA homeostasis [2]. Glucocorticoids
(via GR) induce BA uptake systems such as NTCP, ASBT, and
 30:  Pathophysiologic Basis for Alternative Therapies for Cholestasis 369
OSTα/β [2]. Ligand‐activated GR also upregulates AE2 expres-
sion, which should result in increasing cholangiocellular bicar-
bonate secretion and restoring the bicarbonate umbrella [57].
This observation is of particular interest since AE2 expression is
reduced in liver and inflammatory cells of PBC patients [2],
possibly responsible for making cholangiocytes more suscepti-
ble to an autoimmune first hit. Besides glucocorticoids, UDCA
also activates GR. The combination of UDCA and another GR
ligand dexamethasone (but not UDCA or dexamethasone alone)
increased expression and function of AE2 via interaction with
HNF1 and GR on the AE2 alternate promoter [57]. UDCA‐­
activated GR is also anti‐inflammatory through suppressing
NFκB‐dependent transcription by preventing GR – p65 interac-
tion [2] representing another example for the dichotomous
effects of NRs on both metabolism and inflammation.
Glucocorticoids signal not only via GR, but also other NRs
such as CAR, and glucocorticoids increase expression of PXR and
RXRα [43]. Budesonide also activates PXR in mice and humans,
while dexamethasone directly activates PXR in mice but not in
humans [58]. Clinically prednisolone and budesonide were tested
in combination with UDCA in PBC patients. While the combina-
tion of prednisolone with UDCA was not superior to UDCA mon-
otherapy [2], budesonide UDCA combination significantly
improved serum liver tests compared to UDCA alone [2].
Xenobiotic sensors PXR and CAR
Pregnane X receptor (PXR/NR1I2) and constitutive androstane
receptor (CAR/NR1I3) are key regulators of enzymes involved
in BA hydroxylation/detoxification [2] and alternative BA
export. Both receptors are low affinity, broad specificity xenosen-
sors, activated by a range of compounds which are structurally
unrelated such as antibiotics including rifampicin, clotrimazole,
the antidepressant St. John’s wort, and synthetic steroids such as
5b‐pregnane‐3,20‐dione, pregnenolone 16a‐carbonitrile (PCN)
and dexamethasone [2, 59]. In addition to xenobiotics, also
potentially toxic endogenous cholephiles such as secondary BAs
(e.g. LCA) and bilirubin activate these receptors [60].
PXR and CAR mitigate potentially harmful effects of toxic
BAs by coordinated activation of detoxification pathways such
as hydroxylation (Cyp3a11/CYP3A4, and CYP2B10), sulfation
(SULT2a1/Sult2a1), conjugation (BACS/Bacs, BAT/Bat), and
subsequent elimination via basolateral export pumps multidrug
resistance‐associated protein 3 (MRP3) and MRP4 (Figure 30.1)
[43]. In addition, they regulate bilirubin metabolism in the liver
inducing its glucuronidation (UGT1A1/Ugt1a1) and canalicular
excretion (MRP2) (Figure 30.1) [43]. PXR was shown to inhibit
the interaction between PGC1α and HNF4α, thereby repressing
CYP7A1 gene transcription and BA synthesis [7]. Interestingly,
PXR was identified as FXR target gene [61], pointing toward a
potential NR crosstalk in the protection from BA toxicity.
Besides beneficial effects on BA detoxification and elimina-
tion, PXR may also exert direct anti‐inflammatory and antifi-
brotic effects. In mice, activation of PXR inhibits inflammation
and fibrosis, while PXR KO mice are more susceptible to
inflammatory stimuli [7]. In human HSCs, activation of PXR
inhibits expression of the major profibrogenic cytokine TGF‐1β,
preventing their transdifferentiation to fibrogenic myofibro-
blasts and slows down proliferation [62].
CAR ligands such as phenobarbital and PXR ligands such as
rifampicin have been used to treat pruritus for decades [43]. These
antipruritogenic effects could be explained by their impact on BA
detoxification and alternative BA elimination outlined above. Of
note, rifampicin is recommended as second‐line treatment for
pruritus by the current guidelines [4]. Moreover, PXR and CAR
activation may not only alleviate pruritus, but also have antichole-
static effects. As such, rifampicin (PXR ligand) improved liver
function in PBC patients [63] and p­ henobarbital (CAR ligand)
reduced serum BA concentration [43]. Interestingly, traditional
Chinese medicines which were used to treat neonatal jaundice by
enhancing bilirubin clearance have been identified as CAR
ligands (e.g. Yin Chin/Artemisia capillaris) [64].
Stimulation of an adaptive response to cholestasis may
­particularly be useful when the underlying cause cannot be
improved. Moreover, persistent secretory failure after DILI and
prolonged mechanical cholestasis may be overcome by PXR
ligands such as rifampicin [2]. Stimulation of detoxification
(e.g. via CAR and PXR agonists) may complement other thera-
pies (e.g. FXR agonists) targeting mainly hepatobiliary flow
and excretion of BAs and other cholephiles. Therefore, more
specific and less toxic CAR/PXR agonists could be a valuable
addition to the therapeutic armamentarium in cholestasis.
Vitamin D and A receptors
Vitamin D receptor (VDR/NR1I1) is not only involved in the
regulation of calcium homeostasis, but also in cell proliferation
and differentiation. Expression of VDR is high in intestine and
low in liver where it is mainly found in components of the
immune system including Kupffer cells, monocytes, natural
killer cells, T and B lymphocytes, and dendritic cells, endothe-
lial cells, cholangiocytes, and HSCs [43]. Vitamin D represses
Th1 and Th17 response and increases differentiation of
­regulatory T cells and NK cells, implying a critical role in the
pathogenesis of autoimmune diseases [65]. In line, VDR poly-
morphisms are associated with several immune‐mediated liver
diseases such as PBC and autoimmune hepatitis [7]. VDR,
­acting as a potential BA sensor along the gut–liver axis, also
functions as a receptor for the secondary BA LCA, which is
rather hydrophobic/cytotoxic. Activation of VDR by LCA
­stimulates Cyp3a11 (and its human homologue of CYP3A4)
expression, an enzyme that detoxifies LCA in liver and intestine
[7]. Additionally, SULT2A1/Sult2a1 (an enzyme catalyzing
­sulfation of BAs) as well as basolateral export pump MRP3 was
upregulated by vitamin D application in enterocytes [7]. The
role of the VDR pathway in BA detoxification and elimination
may have also emerged to shield first‐pass tissues from the
­cytotoxic effects of BA overload.
Notably, VDR activation in cholangiocytes also increased
expression of the antimicrobial peptide cathelicidin, which
could contribute to the anti‐inflammatory properties of VDR in
immune‐mediated bile duct injury [7]. Cathelicidin is able to
neutralize the deleterious effect of LPS which accumulates in
the biliary tree in fibrosing cholangiopathies [19]. Furthermore,
activation of VDR ameliorated fibrosis, possibly by epigenetic
modifications of the SMAD pathway in HSCs [2]. However,
administration of vitamin D to Mdr2 KO mice did not improve
hepatic fibrosis despite ameliorating serum liver enzymes [2].
370 THE LIVER:  TRANSPORTER MODULATION
Low vitamin D levels have been linked to progression of fibro-
sis in various liver diseases including cholestatic liver injury [7].
Low serum vitamin D level is also an indicator of advanced dis-
ease and a predictor of UDCA non‐response in PBC patients
[66]. Whether (high dose) vitamin D supplementation may also
improve hepatic immune‐tolerance, thereby reducing biliary
injury and subsequent fibrosis in PBC remains to be evaluated.
Vitamin A is the umbrella term for diverse retinoids and their
precursor forms. Retinoids are important regulators of cell pro-
liferation and differentiation that bind to two families of NRs,
retinoic acid receptors (RARα, β, and γ/NR1B1, B2, and B3)
and retinoid X receptors (RXRα, β, and γ/NR2B1, B2, and B3)
[65]. A biologic active form of vitamin A  –  all‐trans retinoic
acid (ATRA) and its stereoisomer, 9‐cis‐retinoic acid – activate
RARs, whereas RXRs are mainly activated by 9‐cis‐retinoic
acid [67, 68]. RARs and RXRs are known to be key regulators
in activation of HSC, transdifferentiating them from vitamin
A‐storing cells to myofibroblasts, which are proliferative, con-
tractile, inflammatory, and chemotactic and characterized by
enhanced matrix production [69]. In line with loss of vitamin A,
activation of HSC is accompanied by downregulation of RAR
and RXR expression [70]. On the other hand, activation of RAR
and RXR may attenuate HSC activation and proliferation [71],
and has been shown to suppress TGFβ and IL6 mRNA expres-
sion, thereby reducing fibrosis [72]. ATRA also interferes with
the innate immune system and inhibits proinflammatory
cytokine expression of macrophages [73].
ATRA also repressed CYP7a1 promoter activity [74], which
was complemented by UDCA [75], indicating potential syner-
gistic effects. In line, combination of ATRA with UDCA attenu-
ated biliary fibrosis, bile duct proliferation, as well as hepatic
inflammation in BDL rats and Mdr2 KO mice [72, 76]. However,
in a clinical pilot study in PSC combination of ATRA and
UDCA did not achieve the primary endpoint (reduction of alka-
line phosphatase) despite reduction of the bile acid intermediate
C4 and inflammation [77].
TRANSPORTER MODULATION
Since the enterohepatic circulation of BAs depends on active BA
transport systems in liver and intestine, inhibition or stimulation
of their transport function not only alters the BA load to the liver
but also BA signaling. Key transporter targets include the apical
sodium dependent bile salt transporter (ASBT/SLC10A2) for
intestinal reabsorption, the sodium taurocholate cotransporting
polypeptide (NTCP/SLC10A1) for hepatocellular reuptake of
conjugated BAs and organic anion transporter (OATPs/SLCO/
SCL21) mainly for unconjugated BAs, and perhaps under certain
circumstances also bile salt export pump (BSEP/ABCB11) as
rate limiting canalicular BA transport system.
ASBT inhibitors and BA sequestrants
Interruption of enterohepatic circulation of BAs  –  either
­pharmacologically with ASBT inhibitors or with BA seques-
trants – may be advantageous not only for treating pruritus but
also the underlying cholestatic liver disease since depleting BAs
from the enterohepatic circulation reduces hepatic and biliary
BA concentrations [18, 78]. Surgical interruption of the entero-
hepatic BA circulation has been a longstanding therapeutic strat-
egy in PFIC [79]. In addition, increasing fecal BA concentrations
and exposure of gut epithelial cells to BAs can stimulate poten-
tial beneficial colonic BA signaling such as TGR5‐­regulated
GLP‐1 secretion from enteroendocrine L‐cells [80]. GLP‐1 has
cholangioprotective effects protecting these cells from apoptosis
and attenuating the reactive cholangiocyte phenotype [18].
Application of highly potent and selective ASBT inhibitors
(Lopixipat and A4250) to Mdr2 KO mice increased fecal BA
loss which cannot be compensated by increased endogenous BA
synthesis. Thus, despite increased Cyp7a1 and repressed Fgf15
expression by ASBT inhibition, hepatic as well as biliary BA
concentration and bile flow were reduced, leading to a beneficial
BA/PL ratio, resulting in an improvement of liver and bile duct
injury in this model of biliary toxicity [18, 78]. Furthermore,
ASBT inhibitor treatment also reduced recruitment of Kupffer
cells and neutrophils to the liver and promoted expansion of anti‐
inflammatory monocytes [18]. Similarly, application of the BA
sequestrant colesevelam completely reversed liver and bile duct
injury [78], further indicating that interruption of enterohepatic
circulation of BAs may have therapeutic potential in attenuating
cholestatic liver disease beyond their role in treatment of pruri-
tus. Although both therapeutic options result in reduced hepatic
BA uptake and thereby lowering hepatobiliary BA toxicity, their
mechanism of action may differ profoundly since enhanced con-
version of primary into secondary BAs (seen under colesevelam,
but not under ASBT inhibitor treatment), may change intestinal
BA signaling. GLP‐1 expression in enteroendocrine L‐cells was
increased in colesevelam treated Mdr2 KO mice, pointing toward
elevated TGR5 activity [78]. Enteroendocrine L‐cell FXR activ-
ity and TGR5 mediated GLP‐1 secretion appear to be inversely
related [18]. Resin‐bound BAs may activate colonic TGR5 but
not intracellular FXR [80], while non‐bound BAs signal intracel-
lularly via FXR [18]. This may have critical implications on
GLP‐1 secretion (stimulated by TGR5, but suppressed by FXR).
In addition to the differences in GLP‐1 signaling between ASBT
inhibitors and BA sequestrants hepatic and biliary BA composi-
tion also varies. Interestingly, colesevelam (but not ASBT inhibi-
tor) resulted in a “hydrophilization” (thereby detoxification) of
the biliary BA composition [18, 78]. Potential mechanistic dif-
ferences between these two treatment strategies may deserve
­further clinical evaluation.
Several ASBT inhibitors have recently been investigated in
early phase clinical trials for treatment of pruritus in PBC and a
range of pediatric cholestatic syndromes. While pharmacologi-
cal target engagement such as reduced serum BA levels, reduced
FGF‐19 and increased levels of C4 (as marker of BA synthesis)
was mostly observed, the clinical effects on pruritus were rather
variable (often not superior to placebo) and confounded by GI
side‐effects such as diarrhea [81, 82].
NTCP blockers
Reduction of NTCP function may also reduce intrahepatic
BA  levels, subsequently changing hepatobiliary BA excretion
and, thereby, also hepatobiliary toxicity. BAs can induce
­inflammation‐based liver injury by stimulating the expression
and secretion of inflammatory cytokines in cultured mouse
 30:  Pathophysiologic Basis for Alternative Therapies for Cholestasis 371
hepatocytes [83], thereby initiating hepatic neutrophil recruit-
ment which contributes to progression of cholestatic liver dis-
ease [84]. Pharmacological inhibition of NTCP may reduce
intracellular accumulation of BAs, resulting in amelioration of
cholestasis and subsequent hepatocellular driven inflammatory
response. Genetic absence of NTCP in mice and humans results
in increased levels of unconjugated BAs in blood, but is well
tolerated and no pruritus, fat malabsorption, or liver dysfunction
were observed [18]. Myrcludex B, is a NTCP blocker, a small
peptide inhibitor designed to prevent hepatitis B virus uptake
which, like BAs, need NTCP to enter hepatocytes [18]. In DDC
(3,5‐diethoxycarbonyl‐1,4‐dihydrocollidine) fed mice as model
for sclerosing cholangitis NTCP inhibition improved serum bio-
chemistry as well as hepatic inflammation and fibrosis. This
protective effect may be based on increased biliary PL/BA ratio,
making the bile less toxic. Consistent with this hypothesis, no
beneficial effects of NTCP inhibition were seen in the Mdr2 KO
mouse model lacking biliary PL secretion to begin with [85].
Based on these preclinical findings, combination therapy with
Myrcludex B (to reduce hepatocellular BA uptake) and PPARα
agonists (upregulation of Mdr2/MDR3) might be of particular
interest to counteract biliary toxicity (at least in conditions with
residual MDR2/3 functionality). Besides Myrcludex B, several
drugs have been identified to prevent hepatitis B virus (HBV)
infection via inhibition of NTCP, however, only some such as
rosiglitazone, zafirlukast, and sulfasalazine inhibited simultane-
ously NTCP‐mediated BA uptake, making them attractive
­therapeutic approaches for cholestasis [86, 87].
Potential indirect benefits of BSEP blockage
in mice
Although in humans BSEP deficiency results in the very severe
liver disease such as progressive familial intrahepatic cholesta-
sis (PFIC) type 2, mice lacking BSEP are surprisingly protected
from development of severe cholestatic liver injury [18].
Moreover, these mice are also resistant to acquired cholestasis
induced either by bile duct ligation or DDC feeding [18].
Already at baseline, BSEP KO mice exhibit increased ­expression
of enzymes Cyp2b10 and Cyp3a11, involved in BA hydroxyla-
tion, resulting in a very hydrophilic, less toxic BA pool compo-
sition, mainly consisting of tetrahydroxylated BAs [18]. This
metabolic preconditioning with a  –  potential non‐toxic  –
­hydrophilic BA pool, may protect these animals from develop-
ment of cholestatic liver injury. Thus, tetrahydroxylated BAs or
activation of enzymes involved in formation of these BAs may
have therapeutic potential to treat cholestatic liver disease. This
hypothesis is in line with the fact that liver injury in bile duct‐
ligated mice can be attenuated with agonists for CAR and PXR
(activating enzymes involved in BA hydroxylation) [43].
Under rare circumstances, direct transient blockage of BSEP
export function could even directly protect bile ducts from
BA‐induced injury. This might be relevant in the context of liver
transplantation, where BSEP expression recovers faster than
MDR3, resulting in less favorable biliary BA/PL ratio espe-
cially when MDR3 variants are present [18]. Altered bile
­composition after liver transplantation has been associated with
the development of non‐anastomotic biliary strictures [88].
In  such exceptional conditions transient inhibition of BSEP
together with activation of MDR3 and bicarbonate secretion
could theoretically help to protect bile ducts from harmful BA
concentration during the immediate post‐transplant period [18],
although this highly controversial hypothesis still needs to be
tested. Interestingly, drugs such as cyclosporine A used for
immune suppression inhibit BSEP function [89].
Chaperones
Impaired bile secretory function in cholestasis does not only
result from changes in gene expression/transcription but may
also be due to post‐transcriptional changes with disturbed cell
polarity and impaired targeting and function of transporters to
the canalicular (and – to a lesser degree – the basolateral) mem-
brane [5]. Therefore, stimulation of transporter targeting and/or
function may also have important therapeutic implications.
Chaperones are mainly used as therapies for hereditary trans-
port defects, but could also be of interest in counteracting
acquired cholestasis. They may directly interact with the variant
of the ABC transporter proteins such as ABCB11/BSEP or
ABCC2/MRP2, thereby decreasing its degradation and increas-
ing its folding rate (correctors), or impact indirectly on protein
synthesis via interaction with endogenous ER protein chaper-
ones (potentiators) [90]. Interestingly, tauroUDCA has been
shown to enhance the secretory capacity of cholestatic hepatocytes
by stimulation of exocytosis and insertion of canalicular trans-
port proteins into the apical membrane via PKC dependent
mechanisms (Figure 30.2) [91]. In MDCKII cells, 4‐phenylbutyrate
(4‐PBA) was shown to increase cell surface expression and
transport capacity not only from BSEP variants but also from
the wild‐type form [91]. Treatment of rats with 4‐PBA enhanced
expression of wild‐type ABCB11 at the canalicular membrane
[93]. In patients with ABCB11 mutations (PFIC‐2) 4‐PBA treat-
ment increased canalicular expression of ABCB11 [94]. Further
promising observations are made with ivacaftor, originally
known as a cystic fibrosis transmembrane conductance regula-
tor (CFTR) potentiator [95] which has been shown to increase
phosphatidyl concentration in patients carrying certain muta-
tions in the MDR3 gene [95]. Chaperones targeting ABC
transporters such as ABCB11/BSEP and ABCB4/MDR2/3
­
could therefore also be of particular interest to improve acquired
cholestasis.
Tight junction sealers
Hepatocyte tight junctions are crucial for maintaining the struc-
ture and function of bile canaliculi and osmotic gradients gener-
ated by active transport ultimately driving bile flow, while bile
duct epithelium tight junctions are required to maintain epithe-
lial cell polarity and prevent leakage of potentially toxic bile
into the periportal space [96]. Loss of tight junction integrity
may result in regurgitation of bile and bile duct injury [96].
Apart from several experimental models, disruption of tight
junctions have also been observed in PBC, PSC, and biliary
atresia [97]. Notably, restoration of abnormal ZO1 staining was
observed in PBC patients treated by UDCA [97]. Sulfasalazine,
an anti‐inflammatory agent which is extensively used in long‐
term therapy for chronic IBD increases expression of (intesti-
nal) tight junctions in vivo [98] and protects against BA‐induced
372 THE LIVER:  TARGETING GUT–LIVER AXIS FOR CHOLESTATIC LIVER DISEASE

TUDCA
NTCP
Hepatocyte
TUDCA
& cAMP εPKC

Wild type

Subapical endosomal
recycling compartement
Golgi

Potentiator
Corrector
WT TLC Mutant
Transcription Translation
Folding

mRNA mRNA canaliculus

Misfolding
Mutant
Nucleus
Corrector Folded
mutant
ER

Figure 30.2  Therapeutic targeting of transporter trafficking and function at the post‐transcriptional level. After post‐translational modification in
the Golgi apparatus, wild‐type (WT, blue boxes) and mutated forms (green boxes) of transporter are shuttled to a subapical recycling compartment.
TauroUDCA, entering the hepatocytes via NTCP, and cAMP enhances the secretory capacity of cholestatic hepatocytes by stimulation of
­exocytosis and insertion of transport proteins into the apical membrane (green arrow) via εPKC‐dependent mechanisms. Conversely, transporter
internalization (back to the subapical recycling compartment) is accelerated by taurolithocholic acid (TLC, red arrows). Correctors can partially
rescue misfolding by improving folding at the ER and delaying the turnover to the plasma membrane with a presently poorly understood m ­ echanism.
Although rescued mutants retain in part their function, they are conformationally unstable and might be eliminated by ubiquitination. Potentiators
(e.g. Ivacaftor) may correct this phenotype.

apoptosis [99]. Therefore, this drug might be an attractive thera-
peutic for cholestatic liver diseases, especially for PSC patients
also suffering from IBD and this is also currently being studied
in ongoing clinical trials.
norUDCA – A CHOLEHEPATIC DRUG
24‐norursodeoxycholic acid (norUDCA) is a side‐chain short-
ened derivative of UDCA lacking a methyl group and confer-
ring resistance to taurine and glycine conjugation, facilitating its
penetration into cholangiocytes [100]. This results in cholehe-
patic shunting between hepatocytes and cholangiocytes (bypass-
ing the classical enterohepatic circulation), and stimulates
biliary bicarbonate secretion, resulting in alkalization of bile
(Figure 30.3). NorUDCA is actively secreted in its anionic form
(norUDCA−) into the canaliculus. NorUDC is converted to the
protonated form (norUDCA) by accepting a hydrogen cation
(generated by ionization of carbonic acid into hydrogen and
bicarbonate) (Figure  30.3). Protonated norUDCA is then
absorbed passively into cholangiocytes (where it gets deproto-
nated again) subsequently entering periductular capillary plexus
and returns back to the sinusoid to be resecreted into the canali-
culus in its anionic form (norUDCA−). Acceptance of the proton
of carbonic acid by norUDCA− results in increased alkalization
of the bile by generation of bicarbonate [101]. Alkalization of
bile by bicarbonate enrichment may be a key protective mecha-
nism (“bicarbonate umbrella”) against BA toxicity to cholangi-
ocytes [3], preventing uncontrolled membrane permeation of
protonated (in particular glycine) conjugated BAs [2]. Taurine
conjugated and bi‐norUDCA, (lacking two methyl groups) do
not undergo cholehepatic shunt and exert less (Tauro-norUDCA)
or no (bi‐norUDCA) effect on biliary bicarbonate secretion
compared to unconjugated norUDCA, emphasizing the unique
therapeutic properties of norUDCA in the Mdr2 KO mouse
model of sclerosing cholangitis [2]. Anti‐inflammatory effects
of norUDCA were seen on isolated cholangiocytes and mac-
rophages, where it directly interferes with NFκB signaling
[102]. In several mouse models (Mdr2 KO, BDL, NEMO KO,
Schistosomiasis) norUDCA treatment reduced hepatic infiltra-
tion of inflammatory cells [102], indicating that norUDCA may
even have direct anti‐inflammatory effects. Moreover, norUDCA
also exerts antifibrotic and antiproliferative effects [103],
including downregulation of cyclins and inhibition of mTOR
signaling with induction of autophagy [18]. Notably a recent
phase 2 study demonstrated improvement of cholestatic liver
enzymes independent from the previous exposure and response
to UDCA [104], encouraging an ongoing phase 3 study.
TARGETING GUT–LIVER AXIS FOR
CHOLESTATIC LIVER DISEASE
In healthy conditions an enormous amount of gut‐derived mol-
ecules reach the liver via the portal blood without triggering any
inflammatory response. Under pathophysiological conditions,
when the intestinal barrier function is defective or in case of
imbalanced gut bacterial homeostasis, an altered composition of
 30:  Pathophysiologic Basis for Alternative Therapies for Cholestasis 373

BA detoxification Basolateral export

norUDC– norUDC– CYPs MRP3

BAs
SULTs, UGTs
MRP4

Blood
Hepatocyte
norUDC–

H2O + CO2
Kidney
H2CO3
Anti-proliferative norUDC–
Anti-inflammatory H+ + HCO3–
Anti-fibrotic
norUDCA +
HCO3–
norUDC–

AE2 HCO3–
Cholehepatic shunting
Bile Cholangiocyte
HCO3–

Figure 30.3  Therapeutic mechanisms of norUDCA. The anionic form of norUDCA (norUDC−) is secreted into the canaliculus by unknown
mechanisms. norUDC− accepts a hydrogen molecule deriving from ionization of carbonic acid (H2CO3). The acceptance of the proton of H2CO3 by
norUDC− leaves behind a HCO3− in the ductular lumen. NorUDCA is passively absorbed into cholangiocytes and transported to the periductular
capillary plexus, returning to the sinusoids and to being resecreted into the canaliculus, thereby completing a cholehepatic shunt, allowing “ductal
targeting” of anti‐inflammatory, antifibrotic, and antiproliferative effects to injured bile ducts. Through the process of cholehepatic shunting
norUDCA also increases biliary bicarbonate (HCO3−) concentration, thereby stabilizing the bicarbonate umbrella and facilitates BA detoxification
and elimination via basolateral efflux pumps facilitating their subsequent renal excretion. BAs, bile acid.

gut‐derived products or even aberrant gut‐primed lymphocytes
can reach the liver, where it may elicit or exacerbate hepatic
inflammation. PBC and PSC are cholestatic autoimmune liver
diseases in which the end results are caused by immune‐­
mediated bile duct injury. Both conditions are frequently associ-
ated with gut inflammation, PSC with IBD and PBC with
coeliac disease [105], and dysbiosis [106]. This clinical obser-
vation has stimulated several intriguing pathogenic concepts in
which gut commensals, pathogens, and intestinal antigens have
been implicated in causing bile duct injury, in particular in PSC.
One of the first theories connecting intestinal inflammation with
liver injury is passive leakage of proinflammatory microbial
components to the portal circulation and the possibility of an
antigenic trigger of microbial origin. A study in rats with small
bowel bacterial overgrowth described significant hepatic inflam-
mation leading to fibrosis due to toxin translocation [107]. On
the other side, obstructive cholestasis with absence of intralumi-
nal BAs leads to bacterial overgrowth promoting bacterial trans-
location [61]. This observation could be explained by the
well‐known direct antibacterial effects of BAs. Furthermore, it
has been recently shown that BAs exert also indirect effects of
microbiota via activation of intestinal FXR mediating the
expression of anti‐inflammatory genes such as Ang1, iNOS, and
IL18 [24]. The fact that BAs are able to shape microbiota com-
munities by growth inhibition of certain bile sensitive bacteria
[108] as well as by anti‐inflammatory effects via FXR activation
open opportunities for BA‐based therapies in liver diseases
associated with intestinal inflammation.
In addition to intestinal inflammation‐related liver injury
changes in microbiota (dysbiosis) have been described in a wide
range of liver diseases. Several studies using high‐throughput
sequencing technology reported reduced diversity and signifi-
cant shifts in the overall composition of the gut microbiota in
PSC patients and appear to be distinct from IBD in the absence
of PSC [61]. The role of dysbiosis in pathogenesis of cholangio-
pathies has also been emphasized by recent reports of a changed
microbiome in PBC [106]. Several animal models demonstrated
that enteric dysbiosis and/or administration of bacterial antigens
can result in hepatobiliary inflammation with features resem-
bling PSC [61]. Interestingly, a recent GWAS uncovered several
new PSC risk loci related to immune regulation including fuco-
syltransferase‐2, known to influence microbiota and affecting
susceptibility to microbial infection [61]. The importance of the
presence of gut microbiota was reflected in aggravation of
cholestatic liver phenotype and cholangiocyte senescence in
germ free Mdr2 KO mice (as an established model of PSC) [61].
However, the key question remains whether the alterations in
gut microbiota cause liver and bile duct disease or only appear
to be secondary to cholestasis or when the disease has already
progressed. Since reduced diversity dysbiosis is observed in
multiple inflammatory and metabolic diseases it is tempting to
assume that these aspects of observation are secondary to dis-
ease and disease‐related inflammation. Future evidence for the
potential cause–effect relationship between the microbiome and
liver disease may come from interventional studies targeting gut
microbiota.
Several absorbable and non‐absorbable antibiotics modulat-
ing microbiota have been tested in PSC [109] and revealed bio-
chemical improvement, with vancomycin being one of the most
promising agents [110]. In addition probiotics have been
374 THE LIVER:  CONCLUSIONS AND OUTLOOK

explored but so far not recalled convincing clinical results [111].rationale (indicated by association of PBC and genetic variants
Finally, there has been an increasing interest in fecal microbiota in IL12A and IL12RB2 loci [117]) and good efficacy in other
transplantation which is currently being tested as a treatment diseases such as psoriasis in Crohn’s disease [118].
option for PSC [112]. In small studies including PBC patients with an incomplete
Given the key role of BA in gut microbiome homeostasis an response to UDCA, selective B cell depletion with rituximab
effect of UDCA on the microbiota profile could be suspected. In had limited biochemical efficacy although a significant decrease
an interventional trial in PBC [106] UDCA treatment partially of autoantibody production was observed [119]. In addition,
relieved microbiome dysbiosis after six months [106]. since fatigue in PBC has been thought to be driven by muscle
Interestingly, it was also found that patients with PBC and inad- bioenergetics, abnormality related to AMA was reduced by
equate response to UDCA had increased abundance of disease‐ rituximab treatment; rituximab was also tested for fatigue in
associated genera compared with those with adequate response. PBC, but no effect was observed [120]. Moreover, other
In line it has also been shown that UDCA attenuated disease and biologicals such as abatacept (chimeric CTLA4 protein),
­
normalized microbiota changes in mouse models of colitis [113]. FFP104 (anti‐CD40 antibody), NI0801 (anti‐Cxcl10 antibody),
Homing of gut primed T cells has been implicated in the E6011 (anti‐Cxcl1 antibody), and Etrasimod (S1P receptor
pathogenesis of PSC [61]. Effector T cells activated during colitismodulator) are tested in PBC.
express receptors such as α4β7 integrin for MADCAM and Cenicriviroc (CVC) is a novel antagonist for both C‐C
CCR9 for CCL25, facilitating homing to the gut and from the gut chemokine receptors 2 and 5 (CCR2 and 5) that are expressed in
to the liver [61]. Vedolizumab, a gut‐specific α4β7 integrin‐neu- many inflammatory cells including neutrophils, T cells, and
tralizing monoclonal antibody and CCX282‐B, a small molecule monocytes [121]. Peribiliary‐recruited inflammatory cells may
inhibiting CCR9, have been developed for treatment of IBD. contribute to the pathogenesis of cholangiopathies. CVC treat-
Activation of VAP‐1 stimulates hepatic expression of ment improved cholestatic liver injury in the Mdr2 KO mouse
MADCAM‐1 and CCL25, subsequently initiating the recruit- model of sclerosing cholangitis by reducing macrophage
ment of mucosal T cells to the liver [61]. Recently, serum levels recruitment to the liver [122]. Combining treatments that reduce
of VAP‐1 have been shown to correlate with severity of disease in BA pool size with the one that blocks inflammation may have
PSC patients [61]. Therefore, the drugs that may regulate migra- superior effects in the treatment of cholestasis. Combining CVC
tion of activated gut mucosal lymphocytes to the liver appear to and ATRA in cholestatic models (BDL rats and Mdr2 KO
be an attractive option for treating PSC. Several small studies that
mouse) potentiated the effect of monotherapies, indicating a
tested anti‐TNFα agents on reducing biliary inflammation have multitargeted therapy as an important paradigm for treating
been disappointing [114]. A retrospective study in IBD patients cholestatic liver injury [123]. CVC is currently being tested in
with PSC, who have been treated with vedolizumab, did not have patients with PSC [124].
any beneficial effect on liver enzymes [114]. Given its role in Among the emerging antifibrotic inhibitors αVβ6 and lysyl
mediating α4β7/MADCAM‐1 interactions VAP‐1 inhibition may oxidase homolog 2 (LOXL2) are attractive targets. αVβ6 is
also represent a therapeutic target for counteracting cholestatic expressed during epithelial repair in cholestatic hepatocytes and
liver disease [61]. VAP‐1 antibody BTT1023 is currently being cholangiocytes and has a critical role in activation of TGFb1.
clinically tested in PSC patients [115]. Protective effects of αVβ6 inhibition (STX100) were observed
in the Mdr2 KO mouse model of sclerosing cholangitis as well
as in bile duct ligated animals. The humanized monoclonal
anti‐αVβ6 antibody STX100 is under clinical investigation for
TARGETING LIVER INFLAMMATION idiopathic pulmonary fibrosis and chronic allograft nephropathy
AND FIBROSIS [125]. LOXL2, promoting collagen and elastin cross‐linking
[126], as well as cholangiocellular tight junction permeability
Apart from their detergent and proapoptotic properties, BAs [127], has been shown to be critical for development of fibrosis.
cause liver injury in cholestasis by stimulating cytokine‐­ However, despite promising preclinical findings the clinical
mediated inflammatory response and fibrosis [83]. In this results in simtuzumab (LOXL2 inhibitor) treated PSC patients
­process increased hepatocellular BA levels initiated the event by were negative [128].
releasing proinflammatory cytokines (e.g. by activating Egr‐1)
subsequently attracting inflammatory cells which then lead to
tissue injury. Conclusively, reduction of BA load to the liver
and/or inflammatory response by different treatment strategies CONCLUSIONS AND OUTLOOK
would diminish cholestatic liver injury.
Since PSC and PBC are considered to be immune‐mediated Our progress in understanding the molecular pathophysiology
bile duct diseases, several classic immunosuppressive and newer of bile formation and cholestasis has led to the development
immunomodulatory approaches have been tested over time with of  novel therapeutic options as alternative and second‐line
rather disappointing results in regard to efficacy and/or safety ­therapies to UDCA. While UDCA acts mainly at a post‐­
profile [116]. Newer approaches have targeted IL12 [117, 118] transcriptional level, some of these new alternative therapies
and IL23 [118] which have both been implicated in the develop- target key regulatory transcription factors such as FXR, GR, and
ment of PBC [117]. However, ustekinumab, a monoclonal PPARs. Other candidates include key regulators of adaptive BA
­antibody targeting the shared subunit IL12p40, showed rather metabolic and transport pathways such as the xenobiotic sensors
disappointing effects in PBC [118], despite a good scientific CAR and PXR, the FXR downstream target FGF19, inhibitors
 30:  Pathophysiologic Basis for Alternative Therapies for Cholestasis 375
of BA uptake systems within the enterohepatic circulation (e.g.
ASBT), or novel BA derivatives (e.g. norUDCA) undergoing
cholehepatic shunting even bypassing the enterohepatic circula-
tion. Direct immunological approaches with immunosuppres-
sive or biological agents have so far been rather disappointing in
treating cholestatic disorders such as PBC and PSC, but novel
approaches target the gut–liver axis, proinflammatory mono-
cytes, and integrin‐signaling in fibrosis. Moreover, several of
the novel alternative therapeutic approaches targeting BA trans-
port and metabolism exert not only anticholestatic but also pro-
found anti‐inflammatory as well as immune‐modulatory actions
which could be key in treating immune‐mediated cholangiopa-
thies such as PBC/PSC and inflammatory/immune responses
secondary to cholestasis irrespective of their underlying etiol-
ogy. Finally, the gut–liver axis offers additional opportunities
for therapeutic modulation of the microbiome (e.g. in PSC and
IBD). A major challenge will be to clinically test and apply
these expanding therapeutic opportunities and their combina-
tions in an individualized personalized medicine approach.
ACKNOWLEDGMENTS
This work was supported by grants F3517‐B20 and I 2755 from
the Austrian Science Foundation (to MT).
REFERENCES
1. Erlinger, S. What is cholestasis in 1985? J Hepatol, 1985;1:687–93.
2. Beuers, U., Trauner, M., Jansen, P., and Poupon, R. New paradigms in the
treatment of hepatic cholestasis: from UDCA to FXR PXR and beyond.
J Hepatol, 2015;62:S25–37.
3. Beuers, U., Hohenester, S., de Buy Wenniger, L.J., Kremer, A.E., Jansen,
P.L., and Elferink, R.P. The biliary HCO3(‐) umbrella: a unifying hypothesis
on pathogenetic and therapeutic aspects of fibrosing cholangiopathies.
Hepatol, 2010;52:1489–96.
4. European Association for the Study of the Liver. EASL clinical practice
guidelines: the diagnosis and management of patients with primary biliary
cholangitis. J Hepatol, 2017;67:145–72.
5. Trauner, M. and Boyer, J.L. Bile salt transporters: molecular characteriza-
tion, function, and regulation. Physiol Rev, 2003;83:633–71.
6. Evans, R.M. and Mangelsdorf, D.J. Nuclear receptors, RXR and the big
bang. Cell, 2014;157:255–66.
7. Halilbasic, E., Baghdasaryan, A., and Trauner, M. Nuclear receptors as drug
targets in cholestatic liver diseases. Clin Liver Dis 2013;17:161–189.
8. Parks, D.J., Blanchard, S.G., Bledsoe, R.K. et al. Bile acids: natural ligands
for an orphan nuclear receptor. Science, 1999;284:1365–8.
9. Makishima, M., Okamoto, A.Y., Repa, J.J. et al. Identification of a nuclear
receptor for bile acids. Science, 1999;284:1362–5.
10. Wang, H., Chen, J., Hollister, K., Sowers, L.C., and Forman, B.M.
Endogenous bile acids are ligands for the nuclear receptor FXR/BAR. Mol
Cell, 1999;3:543–53.
11. Thomas, C., Pellicciari, R., Pruzanski, M., Auwerx, J., and Schoonjans, K.
Targeting bile‐acid signalling for metabolic diseases. Nat Rev Drug Discov,
2008;7:678–93.
12. Massafra, V., Milona, A., Vos, H.R. et  al. Farnesoid X receptor activation
promotes hepatic amino acid catabolism and ammonium clearance in mice.
Gastroenterol, 2017;152:1462–76.
13. Huang, W., Ma, K., Zhang, J. et al. Nuclear receptor‐dependent bile acid sign-
aling is required for normal liver regeneration. Science, 2006;312:233–6.
14. Holt, J.A., Luo, G., Billin, A.N. et al. Definition of a novel growth factor‐
dependent signal cascade for the suppression of bile acid biosynthesis.
Genes Dev, 2003;17:1581–91.
15. Inagaki, T., Choi, M., Moschetta, A. et  al. Fibroblast growth factor 15
functions as an enterohepatic signal to regulate bile acid homeostasis.
­
Cell Metab, 2005;2:217–25.
16. Lu, T.T., Makishima, M., Repa, J.J. et al. Molecular basis for feedback regu-
lation of bile acid synthesis by nuclear receptors. Mol Cell, 2000;6:507–15.
17. Byun, S., Kim, D.H., Ryerson, D. et al. Postprandial FGF19‐induced phos-
phorylation by Src is critical for FXR function in bile acid homeostasis.
Nat Commun, 2018;9:2590.
18. Trauner, M., Fuchs, C.D., Halilbasic, E., and Paumgartner, G. New therapeu-
tic concepts in bile acid transport and signaling for management of cholesta-
sis. Hepatol, 2017;65:1393–404.
19. D’Aldebert, E., Biyeyeme Bi Mve, M.J., Mergey, M. et al. Bile salts control
the antimicrobial peptide cathelicidin through nuclear receptors in the human
biliary epithelium. Gastroenterol, 2009;136:1435–43.
20. Fickert, P., Fuchsbichler, A., Moustafa, T. et al. Farnesoid X receptor criti-
cally determines the fibrotic response in mice but is expressed to a low extent
in human hepatic stellate cells and periductal myofibroblasts. Am J Pathol,
2009;175:2392–405.
21. Schwabl, P., Hambruch, E., Seeland, B.A. et al. The FXR agonist PX20606
ameliorates portal hypertension by targeting vascular remodelling and
­sinusoidal dysfunction. J Hepatol, 2017;66:724–33.
22. Hao, H., Cao, L., Jiang, C. et  al. Farnesoid X receptor regulation of the
NLRP3 inflammasome underlies cholestasis‐associated sepsis. Cell Metab,
2017;25:856–67.
23. Guo, C., Xie, S., Chi, Z. et al. bile acids control inflammation and metabolic
disorder through inhibition of NLRP3 inflammasome. Immunity,
2016;45:802–16.
24. Inagaki, T., Moschetta, A., Lee, Y.K. et  al. Regulation of antibacterial
defense in the small intestine by the nuclear bile acid receptor. Proc Natl
Acad Sci USA, 2006;103:3920–5.
25. Verbeke, L., Farre, R., Verbinnen, B. et al. The FXR agonist obeticholic acid
prevents gut barrier dysfunction and bacterial translocation in cholestatic
rats. Am J Pathol, 2015;185:409–19.
26. Ding, L., Yang, L., Wang, Z., and Huang, W. Bile acid nuclear receptor FXR
and digestive system diseases. Acta Pharm Sin B, 2015;5:135–44.
27. Liu, Y., Binz, J., Numerick, M.J. et al. Hepatoprotection by the farnesoid X
receptor agonist GW4064 in rat models of intra‐ and extrahepatic cholesta-
sis. J Clin Invest, 2003;112:1678–87.
28. Kuzu, M.A., Kale, I.T., Col, C., Tekeli, A., Tanik, A., and Koksoy, C.
Obstructive jaundice promotes bacterial translocation in humans.
Hepatogastroenterology, 1999;46:2159–64.
29. Sinal, C.J., Tohkin, M., Miyata, M., Ward, J.M., Lambert, G., and Gonzalez,
F.J., Targeted disruption of the nuclear receptor FXR/BAR impairs bile acid
and lipid homeostasis. Cell, 2000;102:731–44.
30. Wagner, M., Fickert, P., Zollner, G. et  al. Role of farnesoid X receptor in
determining hepatic ABC transporter expression and liver injury in bile duct‐
ligated mice. Gastroenterol, 2003;125:825–38.
31. Gomez‐Ospina, N., Potter, C.J., Xiao, R. et al. Mutations in the nuclear bile
acid receptor FXR cause progressive familial intrahepatic cholestasis. Nat
Commun, 2016;7:10713.
32. Nevens, F., Andreone, P., Mazzella, G. et  al. A placebo‐controlled trial
of  obeticholic acid in primary biliary cholangitis. N Engl J Med,
2016;375:631–43.
33. Hirschfield, G.M., Mason, A., Luketic, V. et al. Efficacy of obeticholic acid
in patients with primary biliary cholangitis and inadequate response to
­ursodeoxycholic acid. Gastroenterol, 2015;148:751–61.
34. Kris, V., Kowdley, C.L.B., Levy, C. et al. The AESOP trial: a randomized,
double‐blind, placebo‐controlled, phase 2 study of obeticholic acid in
patients with primary sclerosing cholangitis. Hepatol, 2017;66:1254A.
35. Papazyan, R., Liu, X., Liu, J. et al. FXR activation by obeticholic acid or
nonsteroidal agonists induces a human‐like lipoprotein cholesterol change in
mice with humanized chimeric liver. J Lipid Res, 2018;59:982–93.
36. Reich, M., Spomer, L., Hoehne, J. et al. Expression of the bile acid receptor
TGR5 in livers of PSC patients and Mdr2−/− (Abcb4−/−) mice. J Hepatol,
2017;66:S555.
37. Reich, M., Klindt, C., Deutschmann, K., Spomer, L., Haussinger, D., and
Keitel, V. Role of the G protein‐coupled bile acid receptor TGR5 in liver
damage. Dig Dis, 2017;35:235–40.
38. Drafahl, K.A., McAndrew, C.W., Meyer, A.N., Haas, M., and Donoghue,
D.J. The receptor tyrosine kinase FGFR4 negatively regulates NF‐kappaB
signaling. PLoS One, 2010;5:e14412.
376 THE LIVER:  REFERENCES
39. Modica, S., Petruzzelli, M., Bellafante, E. et  al. Selective activation of
nuclear bile acid receptor FXR in the intestine protects mice against choles-
tasis. Gastroenterol, 2012;142:355–65.
40. Hirschfield, G.M., Chazouilleres, O., Drenth, J.P. et al. Effect of NGM282, a
FGF19 analogue, in primary sclerosing cholangitis: a multicentre, rand-
omized, double‐blind, placebo‐controlled phase 2 trial. J Hepatol,
2019;70(3):483–93.
41. Zweers, S.J., Booij, K.A., Komuta, M. et  al. The human gallbladder
secretes fibroblast growth factor 19 into bile: towards defining the role of
fibroblast growth factor 19 in the enterobiliary tract. Hepatology,
2012;55:575–83.
42. Varga, T., Czimmerer, Z., and Nagy, L. PPARs are a unique set of fatty acid
regulated transcription factors controlling both lipid metabolism and inflam-
mation. Biochim Biophys Acta, 2011;1812:1007–22.
43. Zollner, G. and Trauner, M. Nuclear receptors as therapeutic targets in chole-
static liver diseases. Br J Pharmacol, 2009;156:7–27.
44. Honda, A., Ikegami, T., Nakamuta, M. et  al. Anticholestatic effects of
­bezafibrate in patients with primary biliary cholangitis treated with ursode-
oxycholic acid. Hepatology, 2013;57:1931–41.
45. Delerive, P., Gervois, P., Fruchart, J.C., and Staels, B. Induction of
IkappaBalpha expression as a mechanism contributing to the anti‐inflamma-
tory activities of peroxisome proliferator‐activated receptor‐alpha activators.
J Biol Chem, 2000;275:36703–7.
46. Gervois, P., Kleemann, R., Pilon, A. et  al. Global suppression of IL‐6‐
induced acute phase response gene expression after chronic in vivo treatment
with the peroxisome proliferator‐activated receptor‐alpha activator fenofi-
brate. J Biol Chem, 2004;279:16154–60.
47. Ricote, M. and Glass, C.K. PPARs and molecular mechanisms of transre-
pression. Biochim Biophys Acta, 2007;1771:926–35.
48. Ip, E., Farrell, G.C., Robertson, G., Hall, P., Kirsch, R., and Leclercq, I.
Central role of PPARalpha‐dependent hepatic lipid turnover in dietary stea-
tohepatitis in mice. Hepatology, 2003;38:123–32.
49. Marra, F., Efsen, E., Romanelli, R.G. et al. Ligands of peroxisome prolifera-
tor‐activated receptor gamma modulate profibrogenic and proinflammatory
actions in hepatic stellate cells. Gastroenterol, 2000;119:466–78.
50. Iwaisako, K., Haimerl, M., Paik, Y.H. et al. Protection from liver fibrosis by
a peroxisome proliferator‐activated receptor delta agonist. Proc Natl Acad
Sci USA, 2012;109:E1369–76.
51. Nakamuta, M., Fujino, T., Yada, R. et al. Therapeutic effect of bezafibrate
against biliary damage: a study of phospholipid secretion via the PPARalpha‐
MDR3 pathway. Int J Clin Pharmacol Ther, 2010;48:22–8.
52. Ghonem, N.S., Assis, D.N., and Boyer, J.L. Fibrates and cholestasis.
Hepatology, 2015;62:635–43.
53. Corpechot, C., Chazouilleres, O., Rousseau, A. et al. A placebo‐controlled
trial of bezafibrate in primary biliary cholangitis. N Engl J Med,
2018;378:2171–81.
54. Lindor, K.D., Bowlus, C.L., Boyer, J., Levy, C., and Mayo, M. Primary bil-
iary cholangitis: 2018 practice guidance from the American Association for
the Study of Liver Diseases. Hepatology, 2019;69(1):394–419.
55. Jones, D., Boudes, P.F., Swain, M.G. et al. Seladelpar (MBX‐8025), a selec-
tive PPAR‐delta agonist, in patients with primary biliary cholangitis with an
inadequate response to ursodeoxycholic acid: a double‐blind, randomised,
placebo‐controlled, phase 2, proof‐of‐concept study. Lancet Gastroenterol
Hepatol, 2017;2:716–26.
56. Sookoian, S. and Pirola, C.J. Elafibranor for the treatment of NAFLD: one
pill, two molecular targets and multiple effects in a complex phenotype. Ann
Hepatol, 2016;15:604–9.
57. Arenas, F., Hervias, I., Uriz, M., Joplin, R., Prieto, J., and Medina, J.F.
Combination of ursodeoxycholic acid and glucocorticoids upregulates
the  AE2 alternate promoter in human liver cells. J Clin Invest,
2008;118:695–709.
58. Zimmermann, C., van Waterschoot, R.A., Harmsen, S., Maier, A., Gutmann,
H., and Schinkel, A.H. PXR‐mediated induction of human CYP3A4
and  mouse Cyp3a11 by the glucocorticoid budesonide. Eur J Pharm Sci,
2009;36:565–71.
59. Moore, L.B., Parks, D.J., Jones, S.A. et al. Orphan nuclear receptors consti-
tutive androstane receptor and pregnane X receptor share xenobiotic and
steroid ligands. J Biol Chem, 2000;275:15122–7.
60. Huang, W., Zhang, J., Chua, S.S., Qatanani, M., Han, Y., Granata, R., and
Moore, D.D. Induction of bilirubin clearance by the constitutive androstane
receptor (CAR). Proc Natl Acad Sci USA, 2003;100:4156–61.
61. Hov, J.R. and Karlsen, T.H. The microbiome in primary sclerosing cholangitis:
current evidence and potential concepts. Semin Liver Dis, 2017;37:314–31.
62. Haughton, E.L., Tucker, S.J., Marek, C.J. et al. Pregnane X receptor activa-
tors inhibit human hepatic stellate cell transdifferentiation in vitro.
Gastroenterology, 2006;131:194–209.
63. Bachs, L., Pares, A., Elena, M., Piera, C., and Rodes, J. Comparison of
rifampicin with phenobarbitone for treatment of pruritus in biliary cirrhosis.
Lancet, 1989;1:574–6.
64. Huang, W., Zhang, J., and Moore, D.D. A traditional herbal medicine
enhances bilirubin clearance by activating the nuclear receptor CAR. J Clin
Invest, 2004;113:137–43.
65. Yang, C.Y., Leung, P.S., Adamopoulos, I.E., and Gershwin, M.E. The impli-
cation of vitamin D and autoimmunity: a comprehensive review. Clin Rev
Allergy Immunol, 2013;45:217–26.
66. Guo, G.Y., Shi, Y.Q., Wang, L. et  al. Serum vitamin D level is associated
with disease severity and response to ursodeoxycholic acid in primary biliary
cholangitis. Aliment Pharmacol Ther, 2015;42:221–30.
67. Levin, A.A., Sturzenbecker, L.J., Kazmer, S. et al. 9‐cis retinoic acid stereoi-
somer binds and activates the nuclear receptor RXR alpha. Nature,
1992;355:359–61.
68. Idres, N., Marill, J., Flexor, M.A., and Chabot, G.G. Activation of retinoic
acid receptor‐dependent transcription by all‐trans‐retinoic acid metabolites
and isomers. J Biol Chem, 2002;277:31491–8.
69. Tsuchida, T. and Friedman, S.L. Mechanisms of hepatic stellate cell activa-
tion. Nat Rev Gastroenterol Hepatol, 2017;14:397–411.
70. Ohata, M., Lin, M., Satre, M., and Tsukamoto, H. Diminished retinoic acid
signaling in hepatic stellate cells in cholestatic liver fibrosis. Am J Physiol,
1997;272:G589–96.
71. Sharvit, E., Abramovitch, S., Reif, S., and Bruck, R. Amplified inhibition of
stellate cell activation pathways by PPAR‐gamma, RAR and RXR agonists.
PLoS One, 2013;8:e76541.
72. He, H., Mennone, A., Boyer, J.L., and Cai, S.Y. Combination of retinoic acid
and ursodeoxycholic acid attenuates liver injury in bile duct‐ligated rats and
human hepatic cells. Hepatol, 2011;53:548–57.
73. Wang, X., Allen, C., and Ballow, M. Retinoic acid enhances the production
of IL‐10 while reducing the synthesis of IL‐12 and TNF‐alpha from LPS‐
stimulated monocytes/macrophages. J Clin Immunol, 2007;27:193–200.
74. Crestani, M., Sadeghpour, A., Stroup, D., Galli, G., and Chiang, J.Y. The
opposing effects of retinoic acid and phorbol esters converge to a common
response element in the promoter of the rat cholesterol 7 alpha‐hydroxylase
gene (CYP7A). Biochem Biophys Res Commun, 1996;225:585–92.
75. Le Vee, M., Jouan, E., Stieger, B., and Fardel, O. Differential regulation of
drug transporter expression by all‐trans retinoic acid in hepatoma HepaRG
cells and human hepatocytes. Eur J Pharm Sci, 2013;48:767–74.
76. Cai, S.Y., Mennone, A., Soroka, C.J., and Boyer, J.L. All‐trans‐retinoic acid
improves cholestasis in alpha‐naphthylisothiocyanate‐treated rats and
Mdr2−/− mice. J Pharmacol Exp Ther, 2014;349:94–8.
77. Assis, D.N., Abdelghany, O., Cai, S.Y. et al. Combination therapy of all‐trans
retinoic acid with ursodeoxycholic acid in patients with primary sclerosing
cholangitis: a human pilot study. J Clin Gastroenterol, 2017;51:e11–16.
78. Fuchs, C.D., Paumgartner, G., Mlitz, V. et al. Colesevelam attenuates chole-
static liver and bile duct injury in Mdr2(−/−) mice by modulating composition,
signalling and excretion of faecal bile acids. Gut, 2018;67:1683–91.
79. Whitington, P.F., Freese, D.K., Alonso, E.M., Schwarzenberg, S.J., and
Sharp, H.L. Clinical and biochemical findings in progressive familial intra-
hepatic cholestasis. J Pediatr Gastroenterol Nutr, 1994;18:134–41.
80. Harach, T., Pols, T.W., Nomura, M., Maida, A., Watanabe, M., Auwerx, J.,
and Schoonjans, K. TGR5 potentiates GLP‐1 secretion in response to ani-
onic exchange resins. Sci Rep, 2012;2:430.
81. Al‐Dury, S. and Marschall, H.U. Ileal bile acid transporter inhibition for the
treatment of chronic constipation, cholestatic pruritus, and NASH. Front
Pharmacol, 2018;9:931.
82. Shneider, B.L., Spino, C., Kamath, B.M. et  al. Placebo‐controlled rand-
omized trial of an intestinal bile salt transport inhibitor for pruritus in
Alagille syndrome. Hepatol Commun, 2018;2:1184–98.
83. Cai, S.Y., Ouyang, X., Chen, Y. et  al. Bile acids initiate cholestatic liver
injury by triggering a hepatocyte‐specific inflammatory response. JCI
Insight, 2017;2:e90780.
84. Saito, J.M. and Maher, J.J. Bile duct ligation in rats induces biliary expres-
sion of cytokine‐induced neutrophil chemoattractant. Gastroenterology,
2000;118:1157–68.
 30:  Pathophysiologic Basis for Alternative Therapies for Cholestasis 377
85. Slijepcevic, D., Roscam Abbing, R.L.P. et al. Na(+) ‐taurocholate cotrans-
porting polypeptide inhibition has hepatoprotective effects in cholestasis in
mice. Hepatol, 2018. doi: 10.1002/hep.29888.
86. Donkers, J.M., Zehnder, B., van Westen, G.J.P. et al. Reduced hepatitis B
and D viral entry using clinically applied drugs as novel inhibitors of the
bile acid transporter NTCP. Sci Rep, 2017;7:15307.
87. Kaneko, M., Futamura, Y., Tsukuda, S. et al. Chemical array system, a plat-
form to identify novel hepatitis B virus entry inhibitors targeting sodium
taurocholate cotransporting polypeptide. Sci Rep, 2018;8:2769.
88. Buis, C.I., Geuken, E., Visser, D.S. et  al. Altered bile composition after
liver transplantation is associated with the development of nonanastomotic
biliary strictures. J Hepatol, 2009;50:69–79.
89. Byrne, J.A., Strautnieks, S.S., Mieli‐Vergani, G., Higgins, C.F., Linton,
K.J., and Thompson, R.J. The human bile salt export pump: characteriza-
tion of substrate specificity and identification of inhibitors.
Gastroenterology, 2002;123:1649–58.
90. Martin, G.M., Chen, P.C., Devaraneni, P., and Shyng, S.L., Pharmacological
rescue of trafficking‐impaired ATP‐sensitive potassium channels. Front
Physiol, 2013;4:386.
91. Beuers, U., Bilzer, M., Chittattu, A. et al. Tauroursodeoxycholic acid inserts
the apical conjugate export pump, Mrp2, into canalicular membranes and
stimulates organic anion secretion by protein kinase C‐dependent mecha-
nisms in cholestatic rat liver. Hepatology, 2001;33:1206–116.
92. Hayashi, H. and Sugiyama, Y. 4‐phenylbutyrate enhances the cell surface
expression and the transport capacity of wild‐type and mutated bile salt
export pumps. Hepatology, 2007;45:1506–16.
93. Hayashi, H. and Sugiyama, Y. Short‐chain ubiquitination is associated with
the degradation rate of a cell‐surface‐resident bile salt export pump (BSEP/
ABCB11). Mol Pharmacol, 2009;75:143–50.
94. Gonzales, E., Grosse, B., Schuller, B. et al. Targeted pharmacotherapy in
progressive familial intrahepatic cholestasis type 2: Evidence for improve-
ment of cholestasis with 4‐phenylbutyrate. Hepatology, 2015;62:558–66.
95. Vauthier, V., Housset, C., and Falguieres, T. Targeted pharmacotherapies
for defective ABC transporters. Biochem Pharmacol, 2017;136:1–11.
96. Rao, R.K. and Samak, G. Bile duct epithelial tight junctions and barrier
function. Tissue Barriers, 2013;1:e25718.
97. Sakisaka, S., Kawaguchi, T., Taniguchi, E. et al. Alterations in tight junc-
tions differ between primary biliary cholangitis and primary sclerosing
cholangitis. Hepatology, 2001;33:1460–8.
98. Xu, B., Li, Y.L., Xu, M. et  al. Geniposide ameliorates TNBS‐induced
experimental colitis in rats via reducing inflammatory cytokine release and
restoring impaired intestinal barrier function. Acta Pharmacol Sin,
2017;38:688–98.
99. Rust, C., Bauchmuller, K., Bernt, C. et al. Sulfasalazine reduces bile acid
induced apoptosis in human hepatoma cells and perfused rat livers. Gut,
2006;55:719–27.
100. Hofmann, A.F. and Hagey, L.R., Key discoveries in bile acid chemistry and
biology and their clinical applications: history of the last eight decades. J
Lipid Res, 2014;55:1553–95.
101. Yoon, Y.B., Hagey, L.R., Hofmann, A.F., Gurantz, D., Michelotti, E.L., and
Steinbach, J.H. Effect of side‐chain shortening on the physiologic proper-
ties of bile acids: hepatic transport and effect on biliary secretion of 23‐nor‐
ursodeoxycholate in rodents. Gastroenterology, 1986;90:837–52.
102. Halilbasic, E., Steinacher, D., and Trauner, M. Nor‐ursodeoxycholic acid as
a novel therapeutic approach for cholestatic and metabolic liver diseases.
Dig Dis, 2017;35:288–92.
103. Fickert, P., Wagner, M., Marschall, H.U. et al. 24‐norUrsodeoxycholic acid
is superior to ursodeoxycholic acid in the treatment of sclerosing cholangi-
tis in Mdr2 (Abcb4) knockout mice. Gastroenterology, 2006;130:465–81.
104. Fickert, P., Hirschfield, G.M., Denk, G. et  al. norUrsodeoxycholic acid
improves cholestasis in primary sclerosing cholangitis. J Hepatol,
2017;67:549–58.
105. Trivedi, P.J. and Adams, D.H. Mucosal immunity in liver autoimmunity: a
comprehensive review. J Autoimmun, 2013;46:97–111.
106. Tang, R., Wei, Y., Li Y. et  al. Gut microbial profile is altered in primary
biliary cholangitis and partially restored after UDCA therapy. Gut,
2018;67:534–41.
107. Lichtman, S.N., Sartor, R.B., Keku, J., and Schwab, J.H. Hepatic inflam-
mation in rats with experimental small intestinal bacterial overgrowth.
Gastroenterology, 1990;98:414–23.
108. Watanabe, M., Fukiya, S., and Yokota, A. Comprehensive evaluation of the
bactericidal activities of free bile acids in the large intestine of humans and
rodents. J Lipid Res, 2017;58:1143–52.
109. Elfaki, D.A. and Lindor, K.D. Antibiotics for the treatment of primary scle-
rosing cholangitis. Am J Ther, 2011;18:261–5.
110. Tabibian, J.H., Weeding, E., Jorgensen, R.A. et  al. Randomised clinical
trial: vancomycin or metronidazole in patients with primary sclerosing
cholangitis – a pilot study. Aliment Pharmacol Ther, 2013;37:604–12.
111. Shen, J., Zuo, Z.X., and Mao, A.P. Effect of probiotics on inducing remis-
sion and maintaining therapy in ulcerative colitis, Crohn’s disease, and
pouchitis: meta‐analysis of randomized controlled trials. Inflamm Bowel
Dis, 2014;20:21–35.
112. Ali, A.H., Carey, E.J., and Lindor, K.D. The microbiome and primary scle-
rosing cholangitis. Semin Liver Dis, 2016;36:340–8.
113. Van den Bossche, L., Hindryckx, P., Devisscher, L. et al. Ursodeoxycholic
acid and its taurine‐ or glycine‐conjugated species reduce colitogenic dys-
biosis and equally suppress experimental colitis in mice. Appl Environ
Microbiol, 2017;83.
114. Tse, C.S., Loftus, E.V., Jr., Raffals, L.E., Gossard, A.A., and Lightner, A.L.
Effects of vedolizumab, adalimumab and infliximab on biliary inflamma-
tion in individuals with primary sclerosing cholangitis and inflammatory
bowel disease. Aliment Pharmacol Ther, 2018;48:190–5.
115. Arndtz, K., Corrigan, M., Rowe, A. et al. Investigating the safety and activity
of the use of BTT1023 (Timolumab), in the treatment of patients with pri-
mary sclerosing cholangitis (BUTEO): a single‐arm, two‐stage, open‐label,
multi‐centre, phase II clinical trial protocol. BMJ Open, 2017;7:e015081.
116. Mousa, H.S., Carbone, M., Malinverno, F., Ronca, V., Gershwin, M.E., and
Invernizzi, P. Novel therapeutics for primary biliary cholangitis: toward a
disease‐stage‐based approach. Autoimmun Rev, 2016;15:870–6.
117. Hirschfield, G.M., Liu, X., Xu, C. et al. Primary biliary cirrhosis associated
with HLA, IL12A, and IL12RB2 variants. N Engl J Med,
2009;360:2544–55.
118. Mousa, H.S., Lleo, A., Invernizzi, P., Bowlus, C.L., and Gershwin, M.E.
Advances in pharmacotherapy for primary biliary cholangitis. Expert Opin
Pharmacother, 2015;16:633–43.
119. Myers, R.P., Swain, M.G., Lee, S.S., Shaheen, A.A., and Burak, K.W. B‐
cell depletion with rituximab in patients with primary biliary cholangitis
refractory to ursodeoxycholic acid. Am J Gastroenterol, 2013;108:933–41.
120. Khanna, A., Jopson, L., Howel, D. et al. Rituximab is ineffective for treat-
ment of fatigue in primary biliary cholangitis: a phase 2 randomized con-
trolled trial. Hepatology, 2018.
121. Mack, M., Cihak, J., Simonis, C. et al. Expression and characterization of
the chemokine receptors CCR2 and CCR5 in mice. J Immunol,
2001;166:4697–4704.
122. Guicciardi, M.E., Trussoni, C.E., Krishnan, A. et al. Macrophages contrib-
ute to the pathogenesis of sclerosing cholangitis in mice. J Hepatol,
2018;69:676–86.
123. Yu, D., Cai, S.Y., Mennone, A., Vig, P., and Boyer, J.L. Cenicriviroc, a
cytokine receptor antagonist, potentiates all‐trans retinoic acid in reducing
liver injury in cholestatic rodents. Liver Int, 2018;38:1128–38.
124. Bennett, L.D., Fox, J.M., and Signoret, N. Mechanisms regulating
chemokine receptor activity. Immunology, 2011;134:246–56.
125. Dyson, J.K., Hirschfield, G.M., Adams, D.H. et al. Novel therapeutic tar-
gets in primary biliary cholangitis. Nat Rev Gastroenterol Hepatol,
2015;12:147–58.
126. Barry‐Hamilton, V., Spangler, R., Marshall, D. et al. Allosteric inhibition of
lysyl oxidase‐like‐2 impedes the development of a pathologic microenvi-
ronment. Nat Med, 2010;16:1009–17.
127. Pollheimer, M.J., Racedo, S., Mikels‐Vigdal, A. et al. Lysyl oxidase‐like
protein 2 (LOXL2) modulates barrier function in cholangiocytes in choles-
tasis. J Hepatol, 2018;69:368–77.
128. Muir, A.J., Levy, C., Janssen, H.L.A. et al. Simtuzumab for primary scle-
rosing cholangitis: phase 2 study results with insights on the natural history
of the disease. Hepatology, 2019;69(2):684–98.
 Adaptive Regulation
31 of Hepatocyte Transporters
in Cholestasis
James L. Boyer
Department of Internal Medicine and Liver Center, Yale University School of Medicine, New Haven, CT, USA
INTRODUCTION
Bile secretion is a unique and vital function of the liver. It deliv-
ers bile acids to the intestine for the solubilization and absorption
of dietary lipids and serves as an excretory pathway for endoge-
nous metabolic products including cholesterol, bilirubin, and
porphyrins as well as foreign compounds and xenobiotics. Many
liver disorders can impair this secretion, resulting in retention of
bile and the syndrome of cholestasis. In turn, cholestatic liver
injury results in adaptive responses in the expression of mem-
brane transport proteins and hepatic enzymes that synthesize and
excrete bile acids, bilirubin, and other solutes in an attempt to
reduce the hepatic accumulation of these products and thus
attenuate liver tissue injury. Ultimately these adaptive responses
are not sufficient and chronic cholestatic injury leads to the
development of biliary cirrhosis and progressive liver failure.
Developing therapeutic strategies that might enhance these
intrinsic adaptations is an unmet medical need. Here we review
current knowledge of the mechanism of these adaptive responses.
The major membrane transport systems that result in the forma-
tion of bile are illustrated in Figure 31.1. Their gene designations,
membrane locations, and primary functions are briefly summa-
rized in Table 31.1. Hepatic enzymes that are involved in the regu-
lation of bile salt synthesis and metabolism are summarized in
Table  31.2. A more detailed set of references may be found in
Chapter  23 of the earlier 5th edition. My apologies to authors
whose articles were not able to be included in this 6th edition.
Steps in bile formation
The overall process of hepatic bile secretion can be divided
functionally into four different phases: Phase 0: Hepatic uptake
mechanisms, Phase I: Hydroxylating enzyme reactions, Phase
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
II: Conjugating enzyme reactions, for example, sulfation, glucu-
ronidation, and amidation, and Phase III: Hepatic efflux mecha-
nisms. Each of these steps is regulated by a series of transporters
(Phase 0 and III) or enzyme reactions (Phase I and II).
This chapter focuses on the molecular adaptations in these
hepatocyte membrane transporters and enzymatic reactions that
serve to attenuate cholestatic liver injury and its systemic
effects. Cellular events that impair signal transduction path-
ways, cytoskeleton structures, tight junction and gap junctional
proteins, and the targeting of intracellular vesicles to the apical
canalicular domain that maintains the secretory polarity of the
hepatocyte all contribute to the cholestatic phenotype [1] but are
not considered here.
Much information has been obtained from animal models of
cholestasis and from discoveries of the genetic basis of several
hereditary and acquired human cholestatic disorders. Studies in
human hepatocyte cell lines and cholestatic liver disorders, as
well as mouse knockout models have contributed significantly
to advances in this field.
GENERAL OVERVIEW – ACQUIRED
DEFECTS IN BILE TRANSPORT
PROTEINS
The molecular characterization of hepatic transport systems led
to analyses of their response to cholestatic liver injury resulting
in the following paradigm: Determinants of bile formation
adapt to cholestatic liver injury in order to minimize hepatic
injury by: (i) diminishing the hepatic uptake of bile salts and
other solutes, (ii) reducing bile acid synthesis, (iii) augmenting
bile acid detoxification mechanisms, and (iv) upregulating
 31:  Adaptive Regulation of Hepatocyte Transporters in Cholestasis 379
Na+
BA–
BA–, OA–
OC+
OC+
OA–
Figure 31.1  Membrane transporters that determine the uptake and excretion of bile acids and other organic solutes in hepatocytes (see tables for
full terminology).
Table 31.1  Nomenclature, location, and function of the major hepatocyte plasma membrane bile acid and organic solute transporters involved
in bile secretion (Phase 0 and III)
Name
Sodium‐taurocholate
cotransporter polypeptide
Organic anion‐transporting
polypeptides
Organic solute transporter
alpha/beta
Organic cation
transporter‐1
Organic anion transporter
2 hepatocytes
Multidrug‐resistance
protein-1
(P‐glycoprotein)a
Multidrug‐resistance
protein-3
(phospholipid transporter)a
Bile salt export pumpa
Multidrug‐resistance‐
associated protein 2
(canalicular multispecific
organic anion transporter)a
Multidrug‐resistance‐
associated protein 3a
Multidrug‐resistance‐
associated protein 4a
Multidrug‐resistance‐
associated protein‐6a
Breast cancer resistance
proteina
Sterolin‐1 and 2a
Multidrug and toxin
extrusion protein 1
a
 These transporters are members of the ATP‐binding cassette superfamily.
BSEP
BA FIC1 MRP4
NTCP PS
BA-S–
MRP2
OATPs OA– MDR1
GSH MRP3
OC+
GSH Bil-G–
HCO3– PL
BCRP OA–S
MDR3
OCT1 H+,Na+
OSTα-β
Chol
OATs ABCG5/8
OC+
? MATE-1
Abbreviation (gene) Phase Location Function
NTCP (SLC10A1) 0 Basolateral membrane of Primary carrier for conjugated bile salt
hepatocytes uptake from portal blood
OATPs (SLCO1B1,1B3 0 and III Basolateral membrane of Broad substrate carriers for sodium‐
and 3A1) hepatocytes independent uptake of bile salts, organic
anions, and other amphipathic organic
solutes from portal blood. OATP3A1 is
induced in cholestasis
OSTα/β (SLC51A and III Basolateral membrane of Heteromeric solute carrier for facilitated
SLC51B) hepatocytes, cholangiocytes, transport of bile acids across basolateral
ileum, and proximal tubule membrane of ileum. Expression induced in
of kidney liver in cholestasis
OCT‐1 (SLC22A1) 0 Basolateral membrane of Facilitates sodium‐independent hepatic
hepatocytes uptake of small organic cations
OAT‐2 (SLC22A7) 0 Basolateral membrane of Facilitates sodium‐independent hepatic
uptake of drugs and prostaglandins
MDR1 (ABCB1) III Canalicular and ATP‐dependent excretion of various organic
cholangiocyte apical cations, xenobiotics, and cytotoxins into
membrane bile; barrier function in cholangiocytes
MDR3 (ABCB4) III Canalicular membrane ATP‐dependent translocation of
phosphatidylcholine from inner to outer
leaflet of membrane bilayer (a floppase)
BSEP (ABCB11) III Canalicular membrane ATP‐dependent bile salt transport into bile;
stimulates bile salt‐dependent bile flow
MRP2 (ABCC2) III Canalicular membrane Mediates ATP‐dependent multispecific
organic anion transport (e.g. bilirubin
diglucuronide) into bile; contributes to bile
salt‐independent bile flow by GSH transport
MRP3 (ABCC3) III Basolateral membrane of Expression induced in cholestasis.
hepatocytes and Transports bilirubin and bile acid
cholangiocytes glucuronide conjugates
MRP4 (ABCC4) III Basolateral membrane of Expression induced in
hepatocyte; apical membrane cholestasis – transports sulfated bile acid
of proximal tubule of kidney conjugates and cyclic nucleotides
MRP6 (ABCC6) III Basolateral membrane of ATP‐dependent transport of organic anions
hepatocyte and small peptides. Mutations of MRP6
gene result in pseudoxanthoma elasticum
BRCP (ABCG2) III Canalicular membrane, ATP‐dependent multispecific drug
proximal tubule of kidney transporter, particularly sulfate conjugates;
protoporphyrins are endogenous substrates.
Substrates overlap with MRP2
ABCG5/G8 III Canalicular membrane and Heteromeric ATP dependent transporter for
apical membrane of intestine cholesterol and plant sterols (stilbestrol)
MATE‐1 (SLC47A1) III Canalicular membrane and Organic cation/H+ exchanger extrudes
brush border of kidney cationic xenobiotics
380 THE LIVER: GENERAL OVERVIEW – ACQUIRED DEFECTS IN BILE TRANSPORT PROTEINS

Table 31.2  Hepatic enzymes involved in the regulation of bile salt synthesis and bile salt and bilirubin metabolism (Phase I and II) and their
adaptive responses to cholestasis
Name
Cytochrome P450 7A1
Cytochrome P450 8B1
Cytochrome P450 27A1
Cytochrome P450 3A4
Sulfotransferase
Uridine glucuronyl
transferase
Bile acid CoA synthetase
Bile acid CoA:amino
acid N‐acyltransferase
a
 The adaptive response in cholestasis is not known.
Abbreviation
(gene) Function Adaptive response in cholestasisa
(CYP7A1) Cholesterol 7α‐hydroxylase, the rate‐limiting step in Downregulation limits the synthesis
bile acid synthesis from cholesterol of bile acids
(CYP8B1) Sterol 12α‐hydroxylase, the pathway to cholic acid More favored pathway resulting in
synthesis. Controls ratio of cholic acid to increased % of bile acids as cholic
chenodeoxycholic acid acid
(CYP27A1) Sterol 27‐hydroxylase, the mitochondrial “acidic” Unchanged
alternative pathway of bile acid side‐chain oxidation.
Favors chenodeoxycholic acid formation
(CYP3A4) Mediates Phase I bile acid hydroxylation Unchanged or stimulated; may
facilitate ability of bile acids to
undergo Phase II conjugations
SULT2A1 Mediates Phase II conjugation with sulfate Unchanged or stimulated;
facilitates bile acid conjugation to
substrates for MRP2 and MRP4
(UGT1A1/ Mediates Phase II conjugation with glucuronic acid Unchanged or stimulated: mediates
B4/B7) bile acid and bilirubin conjugation
to substrates for MRP2 and MRP3
BACS Forms bile acid CoA‐thioesters, substrates for BAAT —a
BAAT/Bat Mediates taurine and glycine conjugation of bile —a; reduced in sepsis
acids (amidation)

mechanisms that facilitate the export of bile salts and other
toxic substances from hepatocytes [2]. Thus, transport proteins
on the basolateral sinusoidal membrane which normally func-
tion to remove bile salts and other cholephiles selectively from
portal blood are usually downregulated during cholestatic liver
injury by both transcriptional and post‐transcriptional mecha-
nisms [3, 4]. In contrast, some canalicular transport proteins,
particularly the multidrug resistance protein (MDR) homologs,
are either not severely impaired or may actually be upregulated;
these include MDR1/Mdr1, BSEP/Bsep (bile salt export pump),
and MDR3/Mdr2. These later findings suggest that the choles-
tatic hepatocyte attempts to maintain canalicular export func-
tion. Perhaps more importantly, transporters such as MRP3/
Mrp3, MRP4/Mrp4 (5–10) and OSTα‐OSTβ (the organic solute
transporter) [11] that are expressed at the basolateral sinusoidal
membrane at low levels in normal liver are substantially upregu-
lated in cholestatic hepatocytes, facilitating the efflux of hydro-
phobic bile salts and other products back to the circulation
where they may be cleared in part by the kidney [12]. Thus,
cholestasis results in a partial reversal of bile secretory polarity.
Less is known concerning molecular responses in cholangio-
cytes during cholestasis. However, bile duct proliferation is
characteristic of most cholestatic disorders. The apical sodium‐
dependent bile salt transporter (ASBT), originally identified
in  the ileum, is also expressed on the luminal membranes of
cholangiocytes and the proximal tubules of the kidney and func-
tions to remove bile salts from bile and glomerular filtrate,
respectively [13]. Because the cholestatic hepatocyte continues
to excrete bile salts, albeit at a reduced rate, ASBT may remove
bile salts from an obstructed biliary tree. MRP3, MRP4,
OSTα‐β, and OATP3A1 are also located on the blood side of the
cholangiocyte and are upregulated in the cholestatic liver [5, 11,
14]. MRP3/Mrp3 has a high affinity for glucuronidated conju-
gates and may be the means by which bilirubin glucuronide is
effluxed back into the blood in cholestatic liver. MRP4/Mrp4
may provide the same role for sulfated conjugates [15, 16].
Sulfated bile salt conjugates in particular are formed in the
cholestatic human liver albeit to a lesser extent in mouse liver.
OSTα‐OSTβ, which mirrors the tissue expression of ASBT, also
plays a major role in bile acid efflux from the cholestatic hepat-
ocyte in human liver [11]. Downregulation of Asbt in the renal
proximal tubule results in diminished absorption of bile acids
from the glomerular filtrate, thereby facilitating bile salts excre-
tion in the urine [17]. Mrp2 and Mrp4 (expressed on the apical
luminal membrane of the proximal tubule) may also facilitate
the tubular excretion of bile acid and other divalent conjugates.
Less is known about the response of intestinal bile acid transport
proteins in cholestasis, although ASBT is generally downregu-
lated on the luminal brush border of the ileum, reducing the
return of bile acids to the liver in the enterohepatic circulation
[18], and intestinal MRP2/Mrp2 is diminished in both humans
and rats with obstructive cholestasis [19]. Adaptations also
occur in hepatic enzymes that determine the synthesis and
metabolism of bile acids. In general, these changes result in a
smaller and less hydrophobic bile salt pool, although significant
species differences exist in composition and response. For
example, in cholestatic mice, bile acid pools are enriched in the
highly hydrophilic bile acid, muricholic acid. Enzymes that
regulate bile acid and bilirubin conjugation are also significantly
upregulated in cholestasis. The major enzymes and their func-
tion and adaptive responses are summarized in Table 31.1.
Thus, a complex pattern of adaptive responses in transporter
expression and metabolism occurs in hepatocytes, cholangio-
cytes, kidney, and intestine, that attempt to mitigate tissue dam-
age from the retention of bile salts and other toxic substrates.
Much of this adaptive regulatory response is mediated by
transcriptional events that are regulated by liver‐enriched tran-
scription factors (hepatocyte nuclear factors [HNFs]) and
nuclear receptors (NRs) [4, 2]. Members of the HNF family
tend to regulate constitutive expression of genes whereas NRs
induce adaptive responses in gene expression and are activated
by specific ligands. In cholestasis, these ligands are bile acids,
 31:  Adaptive Regulation of Hepatocyte Transporters in Cholestasis 381

bilirubin, oxysterols, and drugs or xenobiotics [21]. Chief
among these are bile acids which regulate gene expression pri-
marily via the farnesoid X receptor (FXR). When bile acids
accumulate in the cholestatic liver, the expression of genes that
are regulated by FXR are usually enhanced. Other NRs that play
an important role in regulating gene expression are the pregnane
X receptor (PXR), the constitutive androstane receptor (CAR),
the liver X receptor (LXR) [22], the vitamin D receptor (VDR),
and peroxisome proliferator‐activated receptor alpha (PPARα)
[23]. Knockout animals for FXR, PXR, CAR, and LXR are each
more susceptible to cholestatic injury than their respective wild‐
type [4, 22–24]. Other ligand stimulated NRs involved in the
adaptive response to cholestasis include the retinoic X receptor
alpha (RXRα) and the glucocorticoid receptor (GR). RXRα
plays a particularly important role in these responses since it is
the obligate heterodimeric partner for the class II low‐affinity
NRs that include FXR, PXR, LXR, CAR, and PPARα.
Cholestasis also can affect the expression of NRs. Studies of
inflammatory and fibrotic models of cholestasis indicate that
RXRα expression is reduced and may be translocated out of the
nucleus into the cytoplasm, where it may be degraded, thus
influencing the expression of genes dependent on its NR part-
ners [25]. Reductions in mRNA levels of FXR, and its target
gene SHP (short heterodimeric partner), have also been reported
in patients with cholestasis from PFIC1 and 2 (progressive
familial intrahepatic cholestasis type 1 and type 2) and biliary
atresia [26, 27]. Reductions in PXR and CAR have also been
reported in late stages of biliary atresia [27]. PPAR agonists
(fibrates) stimulate MDR3, inhibit bile acid synthesis, have anti‐
inflammatory effects [28], and improve liver function in patients
with primary biliary cholangitis (PBC) [29, 30].
Several other regulators of transcription for which no specific
ligands are known, but that also influence the expression of bile
transporters and enzymes, include the small heterodimer part-
ner‐1 (SHP‐1), liver receptor homologue‐1 (LRH‐1), and
HNF‐4α. HNF‐4α is a primary determinant of HNF‐1α expres-
sion, which determines the transcription of genes that encode for
cytochrome P4507A (CYP7A), NTCP, and OATPs (organic
anion transporting polypeptides) as examples  [31]. Table  31.3
lists these NRs and their activating ligands and key transcription

Table 31.3  Nuclear receptors, important ligands, transcription factors, and target genes regulating the adaptive response in cholestasis
(? = from rodents only)a
Nuclear receptors Ligands Major target genes Anticipated function in cholestasis
FXR (farnesoid X Bile acids, GW4064, BSEP, OATP1B3, MRP2, MDR3, Inhibition of bile acid synthesis and ileal
receptor) – NR1H4 6α‐ethyl‐CDCA, OSTα/β, CYP3A4, UGT2B4 and bile acid uptake via SHP and FGF‐15/19.
fexaramines B7, SULT2A1, BACS and BAAT, Upregulation of Phase I and II bile acid
SHP, I‐BABP, FGF‐15/19, PXR hydroxylation and conjugation, induction
of bile acid canalicular and basolateral
transporters
PXR (pregnane X Xenobiotics, LCA, CYP3A4, OATP1B1?, SULT2A1, Induction of Phase I and II bile acid and
receptor) – NRII ursodeoxycholic acid, UGT1A1 MRP2, MRP3, MDR1, bilirubin conjugation reactions and bile
rifampicin GST acid and bilirubin alternative export pumps
CAR (constitutive androstane Xenobiotics, bilirubin, CYP2B, CYP3A4, OATP1B1, Induction of Phase I and II bile acid and
receptor) – NR3 phenobarbital, Yin Chin, MRP2, MRP4, UGT1A1, bilirubin conjugation reactions and bile
TCPOBOP SULT2A1, GSTs acid and bilirubin alternative export pumps
LXRα (liver X Oxysterols (metabolites CYP7A1, CYP8B, UGT1A3 Inhibits bile acid synthesis while
receptor) – NR1H3 of cholesterol) ABC5/8?, SHP, OSTα/β?, LRH‐1 stimulating Phase II and III steps
RARα (retinoic acid All‐trans‐retinoic acid NTCP?, MRP2?, ASBT, MRP3 ? Induction of RXR partners via
receptor) – NR1B1 metabolism to 9‐cis‐retinoic acid; loss in
cholestasis upregulates MRP3
VDR (vitamin D Vitamin D, lithocholic CYP3A4, SULTs?, MRP3? Induction of Phase I and II bile acid
receptor) – NRIII acid hydroxylation
GR (glucocorticoid receptor) Corticosteroids NTCP, ASBT, MRP2, BSEP, AE2 ? Induction of AE2 (together with
ursodeoxycholate)
RXRα (retinoic X 9‐cis‐Retinoic acid Obligate heterodimeric partner for Reduction of RXR in cholestasis has
receptor) – NR2B1 all class II NRs variable effects depending on the gene
Others
SHP‐1 small heterodimer None; Upregulated by Interacts with LRH‐1 to suppress Suppresses bile acid synthesis, and hepatic
partner) – NR0B2 FXR CYP7A1, CYP8B1, CYP27A1. and ileal bile acid uptake
Also inhibits NTCP?, and ASBT?,
OATP1B1
LRH‐1(FTF) – liver receptor ? Blocked by SHP CYP7A1, CYP8B1. MRP3, Inhibition of bile acid synthesis (CYP8B1)
homologue‐1; fetal ASBT? and ileal uptake (ASBT?) via SHP;
transcription factor – NR5A2 upregulates MRP3
HNF‐4α (hepatocyte nuclear None HNF‐1, CYP7A1, CYP8B1, Inhibition by SHP via FXR of bile acid
factor‐4) – NR2A1 CYP27A1 OATP1B1, NTCP? synthesis and bile acid uptake
PPARα (peroxisome Fatty acids, eicosanoids, BACS and BAAT SULT2A1, Induces Phase II bile acid and bilirubin
proliferator‐activated fibrates, statins UGT2B4, MDR2/3 conjugation reactions. Inhibits bile acid
receptor‐alpha) – NR1C1 CYP7A synthesis
a
 Transporter and enzyme gene abbreviations include: ABCG5 and G8 (sterolin 1 and 2); ASBT (apical sodium‐dependent bile acid transporter, SLC10A2); BAAT (bile acid
CoA:amino acid N‐acyltransferase); BACS (bile acid CoA‐synthase); BSEP (bile salt export pump, ABCB11); CYP7A (cytochrome P450 7A); CYP8B (cytochrome P450 8B);
CYP27A1 (cytochrome P450 27A1); FGF‐15/19 (fibroblast growth factor 15 or 19); GSTs (glutathione S‐transferases); I‐BABP (ileal bile acid binding protein); MDR1 or 2/3
(multidrug resistance protein 1 or 2/3, ABCB1or 4); MRP2, 3, and 4 (multidrug resistance associated protein 2, 3, or 4, ABCC2,3, or 4); NTCP (sodium taurocholate
cotransporting polypeptide, SLC10A1); OATP‐1B1, 1B3 (organic anion transporting polypeptide C or 8, SLC01B1 or 1B3); OSTα‐OSTβ (organic solute transport protein alpha
and beta); SHP‐1 (small heterodimer partner 1); SULTs (sulfotransferases); UGTs (uridine 5′‐glucuronosyl‐transferases).
382 THE LIVER: GENERAL OVERVIEW – ACQUIRED DEFECTS IN BILE TRANSPORT PROTEINS
factors and illustrates their major target genes and also summa-
rizes the anticipated regulatory responses in cholestasis. Further
details of these interactions are provided below. However, because
of differences in gene expression between species, it is difficult to
generalize. For additional information on adaptive responses in
animal models of cholestasis and the role of NRs, the reader is
also referred to several recent comprehensive reviews [4, 32].
Regulation of transporter genes
in the cholestatic hepatocyte
Ion transporters: (Na+K+‐ATPase and Na+/H+ exchange)
Several major ion transport proteins are located on the basolat-
eral membrane and regulate important homeostatic housekeep-
ing functions in the hepatocyte, including maintenance of the
electrical potential, cell volume, and intracellular pH. These
transporters include Na+/K+‐ATPase and the Na+/H+ exchanger‐
isoform 1 (NHE‐1). Although many cholestatic agents have
been shown to inhibit the sodium pump in vitro, the molecular
expression of the sodium pump is either not significantly
affected or somewhat upregulated, as discerned from studies in
several different animal models of cholestasis [33–35]. This
adaptive response could result from an attempt to counteract
increased sodium entry and cell swelling that may result from
the detergent properties of retained bile salts.
Na+/H+ exchange is also upregulated at both transcriptional
and post‐transcriptional levels following common bile duct liga-
tion in the rat, resulting in an increase in intracellular pH. This
response may also contribute to sodium entry and cell swelling
in the cholestatic liver [36].
Transporters on the basolateral membrane
of the hepatocyte (phase 0)
Sodium taurocholate cotransporting polypeptide,
NTCP/Ntcp (SLC10A1/Slc10a1)
Sodium taurocholate cotransporting polypeptide (Ntcp) and its
human homolog NTCP is the major determinant of the selective
hepatic uptake of conjugated bile salts from the portal circula-
tion and is substantially downregulated in both PBC [27, 37],
biliary atresia [7], PFIC, and animal models of cholestasis [33,
35, 38]. NTCP mRNA is also reduced in patients with alcoholic
hepatitis [37] and other inflammatory conditions. suggesting
mediation by cytokines.
The transcriptional regulation of NTCP/Ntcp differs
between species [39]. In human tissue, the mechanism is com-
plex as bile acids induce SHP via FXR, which reduces HNF4α
binding to bile acid response elements (BAREs) in the NTCP
promoter which, in turn, inhibits its transactivating effects on
HNF1α [31]. HNF1α expression is highly dependent on acti-
vation by HNF4α, which is a main regulator of NTCP expres-
sion [31]. Bile acids also have SHP‐independent effects on
HNF4α binding. However, SHP has no direct effect on NTCP/
Ntcp ­promoter activity, and bile acid‐induced signaling path-
ways via c‐Jun N‐terminal kinase may be involved. Other
studies suggest that the human NTCP gene is also activated by
the glucocorticoid receptor and PPARγ coactivator‐1α, and
can be suppressed by bile acids via a small heterodimer part-
ner‐dependent ­mechanism [40].
In contrast, the rat Ntcp promoter contains binding regions
for the transactivators, HNF1, and retinoid receptors (RARα/
RXRα) [41, 42]. Bile acids downregulate rat Ntcp via FXR‐
dependent Shp expression and subsequent inhibition of retinoid
activation of RARα/RXRα [43]. Several additional upstream
regions are involved in cytokine responses. Cholestasis pro-
duced by administration of endotoxin (LPS) results in loss of
HNF1 and the RXRα : RARα heterodimer [42], which decreases
Ntcp expression. LPS also results in release of cytokines (tumor
necrosis factor alpha [TNF‐α] and IL‐1β) that may contribute to
the diminished Ntcp expression [35]. In mice, Ntcp repression
by common bile duct ligation (CBDL) and cholic acid or tauro-
cholate feeding is mediated by FXR and does not depend on
cytokines, whereas Ntcp repression by LPS is independent of
FXR. Although impairment of Ntcp in rodents and NTCP in
humans can explain the reduced Na+‐dependent uptake of con-
jugated bile salts, sodium‐independent mechanisms for hepatic
bile salt uptake persist due to continued expression of several
other sinusoidal membrane organic anion transporters such as
Oatp2/Oatp1a4 and Oatp4/Oatp1b2 in mice and possibly
OATP1B3 in humans [44, 45].
Organic anion transporting polypeptides,
OATPs/Oatps (SLCO/Slco)
The OATP/Oatps are members of a large super family of
transporters (solute carrier organic anion transporter family
[SLCO]) with broad substrate specificity that are capable of
translocating a wide range of organic anions, including
unconjugated and conjugated bile salts, bulky organic cations,
and even certain uncharged organic substrates [46]. These
proteins appear to function as anion exchangers, exchanging
the extracellular anion with either intracellular bicarbonate or
glutathione [41,42], although the exact driving forces are still
not known. OATPs in human liver consist of OATP1B1
(SLCO1B1, ­formally OATP‐C), OATP1B3 (SLCO1B3, for-
mally OATP8), OATP1A2 (SLCO1A2, formally OATP‐A),
OATP2B1 (SLCO2B1, formally OATP‐B), and OATP3A1.
Human OATP1B1 is the most widely studied and the major
sodium‐independent uptake mechanism for bile acids. Its
expression is reduced in alcoholic hepatitis [37] and PBC,
particularly in later stages of the disease [47]. Although tran-
scriptional mechanisms are not known for most OATPs in
cholestatic patients, in vitro studies indicate that  OATP1B1,
like NTCP, is regulated by both FXR/SHP‐dependent and ‐
independent mechanisms where HNF1α is also a primary
transcription factor [33]. In contrast to OATP1B1, studies in
primary sclerosing cholangitis suggested that OATP1A2,
(formally OATP‐A), mRNA expression is actually increased
[44]. However, OATP1A2 is expressed in cholangiocytes in
human liver and not hepatocytes [48]. Bile acids also stimu-
late expression of OATP1B3 via FXR, suggesting that
OATP1B3 might function in reverse and extrude organic ani-
ons from the cholestatic human hepatocyte. Recently,
OATP3A1 was found to be upregulated in the liver of patients
with obstructive cholestasis and in primary human hepato-
cytes via FGF19 activation of transcription factors SP1 and
 31:  Adaptive Regulation of Hepatocyte Transporters in Cholestasis 383
nuclear factor‐kappa B where it functions as a bile acid efflux
transporter [14].
Several cholestatic animal models have been used to evaluate
the expression of Oatp1al (Slco1a1, Oatp1), Oatp1a4 (Slc1a4,
Oatp2), Oatp1b2 (Slco1b2), and Oatp 4. CBDL, ethinylestradiol
(EE), and LPS all result in downregulation of Oatp1al, although
mRNA levels remain unchanged after EE treatment, suggesting
that the mechanism of EE cholestasis is mainly post‐transcrip-
tional (see Chapter 23 in the 5th edition for specific references).
Estrogen‐induced cholestasis results in downregulation of all
basolateral Oatps. Cytokines, particularly IL‐1β and TNF‐α,
play a major role in the regulation of OATP/Oatps and also other
bile acid transporters in cholestasis [49]. Basolateral and canali-
cular transporter systems are downregulated by both cytokines.
Decreased binding activities of nuclear receptor heterodimers
may also be explained in part by a reduction of the ubiquitous
heterodimerization partner, RXR‐α.
Organic cation transporter (OCT‐1, SLC22a1)
OCT‐1 (organic cation transporter) is the major hepatic uptake
transporter for small organic cations. OCT‐1 expression in
human cholestatic liver has not been assessed but animal models
of cholestasis (CBDL and LPS) indicate downregulation of
rOct‐1 mRNA and protein and impaired uptake of Oct‐1 sub-
strates [50]. The human OCT‐1 gene is transactivated by
HNF‐4α and, like other basolateral hepatic uptake transporters,
is suppressed by bile acids via SHP [51]. Thus, hOCT‐1 is also
likely to be downregulated in cholestasis affecting uptake of
substrates like metformin.
Organic anion transporter 2 and 3 (OAT‐2 and ‐3,
Slc22a6 and Slc22a8)
These two members of the OAT family are expressed in liver
and kidney, and transport a variety of organic anions including
bile acids. However, the effects of cholestasis on the expression
of these transporters in liver is not well studied.
Bile acid synthesis
Bile acids are formed from cholesterol in the liver by a series
of enzymatic pathways consisting of the classic (neutral) path-
way or the alternative (acidic) pathway [52]. The former is
mediated by CYP7A1 and is rate limiting in the overall pro-
cess. This pathway also involves CYP8B1, which leads to the
production of CA and which determines the relative hydropho-
bicity of bile by determining the ratio of chenodeoxycholic
acid (CDCA) to CA that normally is approximately equal.
CYP27A1 mediates the alternative pathway which leads
­primarily to the formation of CDCA. All bile acid species are
conjugated (amidated) with either glycine or taurine by bile
acid CoA‐synthase (BACS) and bile acid CoA:amino acid N‐
acyltransferase (BAAT). These bile acid conjugates are then
excreted into bile via the BSEP and can undergo both deami-
dation and 7α‐dehydroxylation by intestinal bacteria, produc-
ing deoxycholic acid and lithocholic acid. Unconjugated bile
acids are reamidated when they recycle back to the liver in the
enterohepatic circulation [52].
During cholestasis, elevated hepatic levels of bile acids
a­ ctivate FXR, which induces SHP and inhibits CYP7A1 gene
transcription and bile acid synthesis [53]. In the intestine reduc-
tions in bile salt concentrations in the intestinal lumen decrease
production of FGF15/19, an FXR regulated hormone in the
ileum, that interacts with the FGFR4 receptor in hepatocytes in
conjunction with β‐klotho [54]. Reductions in FGF15/19 lead to
an upregulation of CYP7A1/Cyp7a1. A decline in FGF15/19
also reduces gallbladder filling and contractions [55]. FGF19 is
upregulated in extrahepatic obstruction in human liver, inhibit-
ing bile acid synthesis [56]. In addition, proinflammatory
cytokines (TNFα and IL‐1β) also play a role via c‐Jun N‐termi-
nal kinase (JNK)/cJun post‐transcriptional signal transduction
pathways [57].
Thus, multiple transcriptional and post‐transcriptional mech-
anisms exist to reduce the formation of bile acids during choles-
tasis. However, the master switch appears to be at the level
of  the intestine, where bile acid homeostasis is exquisitely
­regulated via FGF15/19.
In rat models of CBDL and α‐naphthyl isothiocyanate
(ANIT) administration, Cyp8b1 expression, but not Cyp7a1,
was significantly inhibited [58] whereas 17α‐ethinylestradiol‐
induced cholestasis decreased Cyp7a1 but not the acidic path-
way via Cyp27a. In patients with primary biliary cholangitis,
CYP7A1 mRNA decreased to 10–20% of control levels with no
changes in CYP27A or CYP8B1 [9]. Thus, adaptive changes
that serve to reduce bile acid synthesis occur in both rodents and
patients with cholestasis.
Bile acid hydroxylation (phase I)
Studies in bile duct ligated rats indicate that most Cyp450‐­mediated
reactions are diminished. In vitro studies show that unconjugated
hydrophobic bile acids are more potent inhibitors than conjugated
bile acids [59]. However, during cholestasis hydroxylation reac-
tions mediated predominantly by CYP3A4 attempt to decrease
the hydrophobicity of the bile acid pool accounting for the signifi-
cant levels of (poly)hydroxylated bile acids in the urine of chole-
static animals and patients [60]. CYP3A4 is regulated by the NRs
PXR, FXR [60], VDR, and CAR [61]. Bile acids also induce
Cyp3a11 after CBDL in mice [60]. However, CYP3A4 mRNA is
only mildly altered in patients with PBC [9].
Bile acid conjugation (phase II)
Glucuronidation, sulfation, and amidation of bile acids (major
Phase II reactions) also diminish their toxicity during cholesta-
sis and facilitates their excretion back into blood via MRP3 and
MRP4 with subsequent elimination by the kidney. Human
­uridine glucuronyl transferase 1A3 (UGT1A3) forms acyl che-
nodeoxycholic acid 24‐glucuronide. Cytosolic sulfotransferase
2A (SULT2A) also plays a significant role in sulfation reac-
tions, particularly in females. This enzyme may be upregulated
by Car‐dependent mechanisms [61] in rodents. In contrast,
SULT2A is negatively regulated through CDCA‐mediated FXR
activation in mice and humans [62].
Amidation reactions for bile acids are determined by both
BACS and BAAT and both are regulated by FXR [63]. Few
384 THE LIVER: GENERAL OVERVIEW – ACQUIRED DEFECTS IN BILE TRANSPORT PROTEINS
studies have examined these enzymes during cholestatic liver
injury. Mutations in BAAT have been described in hypercholane-
mia in childhood [64].
Hepatic efflux mechanisms (phase III)
Basolateral membrane of the hepatocyte (the multidrug
resistance associated proteins MRPS/Mrps (ABCC)
and the organic solute transporter, OSTα‐OSTβ
Cholestatic liver injury results in reversal of the secretory polar-
ity of the hepatocyte. This phenomenon is associated with
the  adaptive upregulation of transporters on the basolateral
membrane that facilitates the extrusion of bile salts and other
organic solutes back into the hepatic sinusoids where they can
be eliminated in the urine.
Mrp1, ‐3, and ‐4 and Ostα‐β are normally expressed at low
levels in normal liver [65, 66] but are generally upregulated
during cholestasis [5, 6, 11]. Mrp1, ‐5, and ‐6 are of little
functional importance in the cholestatic liver while Mrp3 and
‐4 are capable of preferentially transporting glucuronidated
and sulfated bile acids, respectively [15], and their induction
during cholestasis may account for the appearance of these
bile acid conjugates in the urine [12]. Induction of Mrp3 after
BDL in mice is due to TNF‐α dependent upregulation of
Lrh‐1 with increased binding of Lrh‐1 to the Mrp3 promoter
[67]. Hepatic injury is more severe in bile duct ligated TNF‐
receptor knockout mice [67].
Studies in HepG2 cells indicate that human MRP3 expression
is repressed by RXR α : RARα which occupies Sp1 activator sites
in the MRP3 promoter. Because RXRα : RARα expression is
diminished by cholestatic liver injury, loss of RXRα : RARα may
lead to upregulation of MRP3/Mrp3 expression by derepressing
Sp1 activation in cholestatic disorders [68]. More recent studies
in Mrp4 and Mrp3 knockout mice suggest that Mrp4, rather than
Mrp3, may be more important in protecting the hepatocytes from
bile acid toxicity [8, 69]. MRP4 protein, but not mRNA, levels are
significantly increased in patients with stage III and IV primary
biliary cholangitis [9], PFIC1 [7] and late‐stage biliary atresia
[27], suggesting post‐transcriptional regulation [9]. Mrp3 in mice
is regulated by CAR, PXR, and VDR whereas Mrp4 is induced
primarily by CAR and PPARα [70].
Both CAR and aryl hydrocarbon receptor (AHR) response
elements are present in the MRP4 promoter so that bilirubin, a
CAR activator, as suggested by studies in Car knockout mice
[70], may stimulate MRP4 expression. There is no evidence that
other basolateral Mrps expressed in the liver, including Mrp5
and ‐6, play any protective role in the adaptive response in
cholestatic liver injury.
OSTα‐β/Ostα‐β is a facilitated heteromeric transporter of
bile acids and other organic solutes where the direction of
transport is dependent on the electrochemical gradient
between the cell and blood. It is weakly expressed in the
liver in rodents but is significantly upregulated by FXR reg-
ulated mechanisms at both mRNA and protein levels primar-
ily in patients with stage III and IV PBC, biliary atresia, and
in bile duct ligated mice and rats [11, 27]. Ostα‐β is also
regulated by Lxr in the mouse which shares a DR‐1 binding
site with Fxr in the mouse promoter [71]. Paradoxically
hepatic injury is reduced after bile duct obstruction in Ostα
knockout mice [72].
Canalicular membrane of the hepatocyte (MDR1/
Mdr1a,b; MDR3/Mdr2; MRP2/Mrp2; BSEP/Bsep;
BCRP/Bcrp; ABCG5/G8;Abcg5/8)
Despite downregulation of hepatic uptake mechanisms and
upregulation of basolateral efflux transporters that function to
retard the accumulation of bile acids and other toxic substrates,
hepatic levels of bile acids and other constituents of bile con-
tinue to accumulate in the cholestatic liver and are primary sub-
strates for the apical canalicular membrane efflux pumps. These
transporters are all members of the ABC superfamily and
the  inability of these rate‐limiting transporters to effectively
excrete these substrates into bile becomes the major determinant
of the cholestatic phenotype.
Canalicular organic solute transporters
Multidrug resistance protein, MDR1/Mdr1a,b
(ABCB1/Abcb1)
MDR1 encodes for the drug efflux pump also known as P‐gly-
coprotein 170 (Pgp‐170). MDR1 transports a variety of drugs,
xenobiotics and lipids although its importance in excretion of
endogenous substrates remains unclear. Mdr1 is capable of
transporting bile acids, albeit with a much lower affinity than
Bsep [73]. Disease causing MDR1 mutations have not been
found. However, polymorphisms have significant effects on
drug absorption, excretion, and toxicity [74]. Most forms of
cholestasis result in significant upregulation of Pgp‐170
Mdr1a/b mRNA in both animal models and humans with the
level of expression correlating with the severity of the cholesta-
sis [75]. MDR1 protein is increased in late‐stage PBC and in
biliary atresia [9, 27]. Molecular regulation of MDR1/ Mdr1 is
thought to be mediated by NF‐κB transcriptional mechanisms
[76]. Both CAR and PXR activators can also stimulate MDR1
expression [77].
MDR3/Mdr2 (ABCB4/Abcb4)
The importance of this phospholipid export pump in the patho-
genesis of cholestasis is dramatically demonstrated by muta-
tions in the MDR3/Mdr2 gene that result in PFIC3 in children
[78] and biliary cirrhosis in the mouse knockout model, Mdr2−/−
[79]. MDR3/ Mdr2/is a phospholipid floppase. In its absence
phosphatidylcholine cannot be excreted into bile, and bile salts
cannot form mixed micelles resulting in progressive injury to
the bile duct epithelium and, in the mouse, histological findings
reminiscent of primary sclerosing cholangitis [80]. Over time a
biliary‐type cirrhosis develops and, in some cases, hepatocellu-
lar carcinoma.
Children with PFIC3 also have defective phospholipid excre-
tion in bile, bile duct proliferation by histological examination,
and elevated γ‐glutamyl transferase, thus distinguishing them
from PFIC1 and 2, where bile duct proliferation is absent and
γ‐glutamyl transferase is normal [81].
Phospholipid excretion is partially deficient in heterozygotes
who normally do not have a cholestatic phenotype [78, 82].
 31:  Adaptive Regulation of Hepatocyte Transporters in Cholestasis 385
However, mothers of PFIC3 patients are obligate heterozygotes
(MDR3+/−) and are at risk for developing cholestasis during the
third trimester of pregnancy when high levels of estrogens are
present [83, 84]. Polymorphisms and mutations in MDR3 can
predispose patients to cholestatic liver injury when exposed to
other potential cholestatic agents, including drugs and environ-
mental toxins [85]. A missense mutation in ABCB4 has been
described in ductopenic cholestatic liver disease in adults.
MDR3‐deficient patients present a wide spectrum of disease
including intrahepatic cholestasis of pregnancy, low phospho-
lipid associated cholelithiasis (LPAC), cholestatic cirrhosis, and
death in childhood and adults [86].
MDR3/Mdr2 is upregulated in most forms of cholestasis,
including bile duct obstruction and ANIT [58] and in patients
with biliary atresia [27]. MDR3/Mdr2 is regulated in large part
by Fxr/FXR‐mediated mechanisms [87], as well as PPARs in
mice and in humans [88]. Trials with fibrates in patients with
PBC are based in part on this finding.
MRP2/Mrp2 (ABCC2/Abcc2)
MRP2 encodes for the conjugate drug export pump, also known
as multidrug resistance protein 2 or the canalicular multidrug
organic anion transporter (cMOAT) [65, 66]. This ABC trans-
porter is the major export pump for bilirubin diglucuronides,
glutathione conjugates, and divalent bile acids conjugated with
sulfates and glucuronides. It is a major determinant of bile acid
independent bile flow and drug conjugate excretion. Expression
varies considerably in human liver [89] and single nucleotide
polymorphisms (SNPs) affect drug clearance. Mutations in the
Mrp2 gene result in a stop codon and premature termination of
protein translocation in the mutant TR–/GY [90] and the Eisai
hyperbilirubinemic rat (EHBR) [91]. Mutations in the MRP2
gene in humans result in the Dubin–Johnson syndrome [92] a
genetically determined cause of conjugated hyperbilirubinemia.
Although not cholestatic by definition, since bile salt excretion
is normal, this mutation results in impaired excretion of a vari-
ety of amphipathic organic anions, including leukotrienes,
conjugated bilirubin, divalent bile acid conjugates, and
­ the  transporter remains at the canalicular membrane [96].
coproporphyrin isomer series 1, in addition to a variety of other
compounds including bromosulfophthalein (BSP), indocyanine
green, and oral cholecystographic agents [65]. Antibiotics, such
as ampicillin and ceftriaxone, and heavy metals are also excreted
by MRP2. Although it is not known if bile secretion is impaired
in the Dubin–Johnson syndrome, bile salt independent bile flow
is reduced in the rat model as a result of impaired glutathione
excretion [93]. Glutathione is a low‐affinity substrate for Mrp2 ,
the major pathway for excretion of glutathione, oxidized GSSG
and glutathione conjugates [94]. Polymorphisms in MRP2 can
result in diminished glutathione excretion and thus may contrib-
ute to other forms of toxic and cholestatic liver injury [89, 95].
Indeed, the mRNA and protein expression levels of Mrp2, are
markedly downregulated in animal models of cholestasis,
including CBDL, EE, and particularly following administration
of LPS [38]. The latter explains the increase in serum conju-
gated bilirubin characteristic of sepsis‐induced jaundice
[38,  96]. Post‐transcriptional events can result in retrieval of
Mrp2 from the canalicular membrane to a submembranous
localization early in cholestasis prior to changes in mRNA
expression. MRP2 protein is also markedly reduced in liver
biopsies from patients with inflammatory cholestasis [37].
The Mrp2 promoter is activated by RXRα : RARα, and
depressed by IL‐1β following bile duct ligation in rats [42].
Thus, Mrp2 expression in obstructive cholestasis is associated
with cytokine‐dependent alterations in the RXRα : RARα NRs.
Response elements for CAR, PXR, FXR, and AHR also are pre-
sent in the MRP2/Mrp2 promoter so the molecular regulation of
this transporter in cholestasis remains complex [4].
Since bilirubin and glutathione are antioxidants, their hepatic
retention in cholestasis may have cytoprotective effects.
The bile salt export pump (BSEP/Bsep, ABCB11/Abcb11)
This gene encodes for the “sister of P‐glycoprotein”. the canali-
cular membrane bile salt export pump and the major, if not the
sole determinant of bile salt dependent bile flow. Mutations in
BSEP result in PFIC2 [97], benign recurrent intrahepatic chol-
estasis (BRIC), and intrahepatic cholestasis of pregnancy in
adults. More than 100 different mutations have been described
in families with these disorders [97]. PFIC2 resembles the
Byler’s disease (PFIC1) phenotype, with the absence of bile
duct proliferation and normal γ‐glutamyl transferase levels in
serum. Bsep knockout mice also demonstrate impaired bile salt
transport into bile. Altogether the evidence is compelling that
BSEP/Bsep is the canalicular transporter that determines bile
salt dependent bile formation.
Experimental observations suggest that BSEP/Bsep is varia-
bly preserved during cholestatic liver injury. Bsep is only mod-
erately impaired in several rat models of cholestasis [96].
Although a variety of cholestatic agents, including LPS, EE,
cyclosporin A, and rifamycin, profoundly inhibit ATP depend-
ent bile salt transport in vitro in isolated rat liver canalicular
membrane vesicles [98], downregulation in vivo is less pro-
nounced. After 3 days of CBDL in the rat Bsep mRNA and
­protein expression are inhibited by ~30 and 50%, respectively,
but after 7–14 days recover to ~60 and 80% of control
­values, respectively. Immunofluorescence studies indicate that
Furthermore, bile salt excretion continues in the face of com-
plete bile duct obstruction, albeit at a reduced rate [96] and is
mediated by TNF‐α and Il‐1β [99]. LPS and EE administration
in vivo to rats also results in only partial inhibition of Bsep
expression [100]. Although reduced staining of BSEP protein is
seen in patients with inflammatory cholestasis [37], BSEP
expression is maintained in patients with primary biliary chol-
angitis [47] and is reduced in early but not late stages of biliary
atresia [7, 27], where it is maintained in its normal amount and
location [101]. BSEP/Bsep is strongly regulated by FXR/Fxr in
human, rat, and mouse [102] and Bsep induction by bile acids is
absent in the FXR knockout mouse [103]. Thus, differences in
levels of BSEP/Bsep expression in cholestasis may be related in
part to differences in the levels of expression of this nuclear
receptor. For example, while initially reduced, FXR and BSEP
levels return to normal in late‐stage cases of biliary atresia [27].
LRH‐1 also can regulate BSEP expression and thus may play a
supporting role to FXR in maintaining hepatic bile acid levels
and in coordinating the expression of both CYP7A1 and BSEP
to determine bile acid synthesis and excretion [104]. The nuclear
386 THE LIVER:  REFERENCES
factor erythroid 2‐related factor 2 (NRF2) also transactivates the
human BSEP promoter by binding to a Maf recognition element
(MARE). NRF2 is activated by oxidative stress which can result
from the accumulation of “toxic” bile acids [105].
Taurocholate transport is competitively cis‐inhibited by ­various
cholestatic drugs, including cyclosporin A, rifamycin, rifamycin
SV, and glibenclamide in Sf9 cells expressing rat Bsep [98].
Altogether these findings suggest that the cholestatic effects of
these compounds may be determined in part by the extent to
which the canalicular export pumps continue to function as export
pumps. Interestingly, the cholestatic metabolite estradiol‐17β‐
glucuronide inhibited ATP‐dependent taurocholate transport only
when Mrp2 was co‐expressed in Sf9 cells, suggesting that this
compound results in trans‐inhibition of Bsep‐mediated bile salt
transport only after excretion into the bile canaliculi by Mrp2
[98]. Thus, some drugs may produce cholestasis by inhibiting
BSEP only after excretion into bile. Polymorphisms in BSEP,
particularly V444A, also affect the level of expression of BSEP in
human liver [89] and may predispose to some forms of drug‐
induced cholestasis and cholestasis of pregnancy [85, 106].
Mutations of BSEP are associated with cholestatic liver diseases
of varying severity including PFIC2 and BRIC2. Genetic poly-
morphisms are also linked to intrahepatic cholestasis of preg-
nancy (ICP) and drug‐induced liver injury [106].
Other canalicular solute and lipid transporters/flippases
(BCRP, ABCG2) and ABCG5/8, MATE‐1, and FIC1)
The breast cancer resistance protein, BRCP (ABCG2), is also
expressed on the canalicular membrane. BRCP shares broad sub-
strate specificity with MRP2 and excretes sulfated drug conjugates
[107]. Bcrp does not appear to have a significant role in the adap-
tive response to cholestasis in the liver but may be more important
for solute export in the kidney and intestine. BRCP is downregu-
lated in the duodenum of patients with obstructive cholestasis but
not in bile duct ligated rats [108]. Polymorphisms in BRCP are
speculated to play a role in troglitazone sulfate excretion and could
possibly result in cholestasis induced by this metabolite.
Sterolin 1 and 2 (ABCG5/8) are two ABC transporters that
form a heterodimer at the canalicular membrane and account in
large part for the excretion of cholesterol and plant sterols [109].
Little is known about its role in cholestasis, although estrogen‐
induced cholestasis results in diminished mRNA expression [110].
Reduced expression of ABCG5/8 would be expected to contribute
to the hypercholesterolemia that develops during cholestasis.
The multidrug and toxic compound extrusion protein‐1
(MATE‐1) is also expressed in the liver on the apical canalicular
membrane. Members of the MATE family are organic cation
exporters that excrete metabolic or xenobiotic organic cations
from the body by way of a H+ or Na+ exchange mechanism
[111]. However, little is known about its specific role in the liver
and whether it is regulated during cholestasis.
The gene product of familial intrahepatic cholestasis‐1 (FIC1,
ATP8B1) is a P‐type ATPase which functions as a phosphatidyl-
serine flippase at the canalicular membrane [112]. Mutations in
FIC1 result in PFIC‐1, also called Byler’s disease [113]. PFIC‐1
is phenotypically similar to PFIC‐2. Whether FIC1’s function is
impaired in other forms of cholestasis and contributes further to
cell injury is not clear.
Canalicular ion transporters (AE2, also SLC4A2)
Anion Exchanger‐2 (AE2)
This transporter encodes for the canalicular Cl−/HCO3−
exchanger that regulates the excretion of bicarbonate, a partial
determinant of canalicular bile salt independent bile formation
[114]. Anion exchanger‐2 (AE2) is also expressed on the lumi-
nal membrane of the cholangiocyte and is a determinant of
bicarbonate excretion from this epithelium, providing a protec-
tive “biliary bicarbonate umbrella” [115]. Thus, impairment of
AE2 would be expected to reduce bile flow and thus might
­predispose to other cholestatic insults.
Initial studies showed that AE2 gene expression and protein
immunoreactivity at the canaliculus and in bile ducts were
reduced in patients with PBC but not in other cholestatic and
non‐cholestatic liver diseases [116]. Reduced expression of
AE2 mRNA has also been observed in salivary glands in patients
with PBC and sicca syndrome, suggesting that there may be a
generalized deficiency of this transporter in this disease.
Measurements of the Cl−/HCO3− anion exchanger (AE) in iso-
lated cholangiocytes showed that the cAMP‐stimulated AE
activity is diminished in PBC compared to both healthy and dis-
eased controls. MicroRNA‐506 (miR‐506) is upregulated in
cholangiocytes of PBC patients and AE2 may be a target of
miR‐506 [117]. Stimulation of the activity of AE2 may be one
of several mechanisms by which ursodeoxycholic acid (UDCA)
may improve outcomes in PBC patients.
CONCLUSION
This chapter has focused on the adaptive response that mem-
brane transporters in hepatocytes play in inherited and acquired
forms of cholestasis and the role of NRs in this process.
Information about the mechanisms of these transcriptional
events is rapidly expanding, revealing the complexity and the
interrelatedness of these responses in the form of extended net-
works. Adaptative mechanisms also occur in cholangiocytes,
kidney, and intestine that contribute to the ability to modulate
this disorder but are beyond the scope of this review. In the
future, progress will come from new therapeutic strategies, most
likely combinations of novel nuclear receptor ligands that stim-
ulate these protective pathways.
ACKNOWLEDGMENTS
Publications cited from this laboratory were supported by
USPHS DK 25636 and DK P30–34989.
REFERENCES
1. Trauner, M. and Boyer, J.L. Bile salt transporters: molecular characterization,
function and regulation. Physiol Rev, 2003;83:633–71.
2. Boyer, J.L. New perspectives for the treatment of cholestasis: lessons from
basic science applied clinically. J Hepatol, 2007;46:365–71.
 31:  Adaptive Regulation of Hepatocyte Transporters in Cholestasis 387
 3. Stahl, S., Davies, M.R., Cook, D.I. and Graham, M.J. Nuclear hormone
receptor‐dependent regulation of hepatic transporters and their role in the
adaptive response in cholestasis. Xenobiotica, 2008;38:725–77.
 4. Halibasic, E.C.T. and Trauner, M. Bile acid transporters and regulatory
nuclear receptors in the liver and beyond. J Hepatol, 2013;58:155–68.
  5. Soroka, C.J., Lee, J.M., Azzaroli, F. and Boyer, J.L. Cellular localization and
up‐regulation of multidrug resistance‐associated protein 3 in hepatocytes
and cholangiocytes during obstructive cholestasis in rat liver. Hepatology,
2001;33:783–91.
  6. Denk, G.U., Soroka, C.J., Takeyama, Y. et al. Multidrug resistance‐associ-
ated protein 4 is upregulated in liver but downregulated in kidney in obstruc-
tive cholestasis in the rat. J Hepatol, 2004;40:585–91.
  7. Keitel, V., Burdelski, M., Waqrskulat, U. et al. Expression and localization of
hepatobiliary transport proteins in progressive familial intrahepatic cholesta-
sis. Hepatology, 2005;41:1160–72.
  8. Mennone, A., Soroka, C.J., Cai, S.Y. et al. Mrp4‐/‐ mice have an impaired cyto-
protective response in obstructive cholestasis. Hepatology, 2006;43:1013–21.
  9. Zollner, G., Wagner, M., Fickert, P. et al. Expression of bile acid synthesis
and detoxification enzymes and the alternative bile acid efflux pump MRP4
in patients with primary biliary cirrhosis. Liver Int, 2007;27:920–9.
10. Takeyama, Y., Uehara, Y., Inomata, S. et al. Alternative transporter pathways
in patients with untreated early‐stage and late‐stage primary biliary cirrhosis.
Liver Int, 2009;29:406–14.
11. Boyer, J.L., Trauner, M., Mennone, A. et al. Upregulation of a basolateral
FXR‐dependent bile acid efflux transporter OSTalpha‐OSTbeta in cholesta-
sis in humans and rodents. Am J Physiol Gastrointest Liver Physiol,
2006;290:G1124–30.
12. Makino, I., Hashimoto, H., Shinozaki, K., Yoshino, K., and Nakagawa, S.
Sulfated and nonsulfated bile acids in urine, serum, and bile of patients with
hepatobiliary diseases. Gastroenterol, 1975;68:545–53.
13. Lee, J., Azzaroli, F., Wang, L. et al. Adaptive regulation of bile salt transport-
ers in kidney and liver in obstructive cholestasis in the rat. Gastroenterology,
2001;121:1473–84.
14. Pan, Q., Zhang, X., Zhang, L. et al. Solute carrier organic anion transporter
family member 3A1 is a bile acid efflux transporter in cholestasis.
Gastroenterology, 2018; 155(5):1578–92.
15. Kruh, G.D., Belinsky, M.G., Gallo, J.M., and Lee, K. Physiological and phar-
macological functions of Mrp2, Mrp3 and Mrp4 as determined from recent
studies on gene‐disrupted mice. Cancer Metastasis Rev, 2007;26:5–14.
16. Zamek‐Gliszczynski, M.J., Hoffmaster, K.A., Nezasa, K., Tallman, M.N.
and Brouwer, K.L. Integration of hepatic drug transporters and phase II
metabolizing enzymes: mechanisms of hepatic excretion of sulfate, glucuro-
nide, and glutathione metabolites. Eur J Pharm Sci, 2006;27:447–86.
17. Lee, J.M., Azzaroli, F., Gigliozzi, A., Mennone, A., and Boyer, J.L. Adaptive
regulation of the ileal sodium‐dependent bile salt transporter (ISBT), and the
multispecific organic anion transporter (MRP2), in kidney and cholangio-
cytes in chronic cholestasis  –  alternative pathways for bile salt excretion.
Hepatology, 1999;30:417A.
18. Hruz, P., Zimmermann, C., Gutmann, H. et  al. Adaptive regulation of the
ileal apical sodium dependent bile acid transporter (ASBT) in patients with
obstructive cholestasis. Gut, 2006;55:395–402.
19. Dietrich, C.G., Geier, A., Salein, N. et al. Consequences of bile duct obstruc-
tion on intestinal expression and function of multidrug resistance‐associated
protein 2. Gastroenterology, 2004;126:1044–53.
20. Chiang, J.Y. Bile acid metabolism and signaling. Compr Physiol, 2013;3:
1191–1212.
21. Boyer, J.L. Nuclear receptor ligands: rational and effective therapy for
chronic cholestatic liver disease? Gastroenterology, 2005;129:735–40.
22. Guo, G.L., Lambert, G., Negishi, M. et al. Complementary roles of farnesoid
X receptor, pregnane X receptor, and constitutive androstane receptor in pro-
tection against bile acid toxicity. J Biol Chem, 2003;278:45062–71.
23. Trottier, J., Milkiewicz, P., Kaeding, J., Verreault, M., and Barbier, O.
Coordinate regulation of hepatic bile acid oxidation and conjugation by
nuclear receptors. Mol Pharm, 2006;3:212–22.
24. Stedman, C.A., Liddle, C., Coulter, S.A. et al. Nuclear receptors constitutive
androstane receptor and pregnane X receptor ameliorate cholestatic liver
injury. Proc Natl Acad Sci USA, 2005;102:2063–8.
25. Zimmerman, T.L., Thevananther, S., Ghose, R., Burns, A.R., and Karpen,
S.J. Nuclear export of retinoid X receptor alpha in response to interleukin‐
1beta‐mediated cell signaling: roles for JNK and SER260. J Biol Chem,
2006;281:15434–40.
26. Demeilliers, C., Jacquemin, E., Barbu, V. et al. Altered hepatobiliary gene
expressions in PFIC1: ATP8B1 gene defect is associated with CFTR down-
regulation. Hepatology, 2006;43:1125–34.
27. Chen, H.L., Liu, Y.J., Chen, H.L. et al. Expression of hepatocyte transporters
and nuclear receptors in children with early and late‐stage biliary atresia.
Pediatr Res, 2008.
28. Ghonem, N.S., Assis, D.N., and Boyer, J.L. Fibrates and cholestasis.
Hepatology, 2015;62:635–43.
29. Levy, C. and Lindor, K.D. Editorial: itching to know: role of fibrates in PBC.
Am J Gastroenterol, 2018;113:56–7.
30. Levy, C., Peter, J.A., Nelson, D.R. et al. Pilot study: fenofibrate for patients
with primary biliary cirrhosis and an incomplete response to ursodeoxy-
cholic acid. Aliment Pharmacol Ther, 2011;33:235–42.
31. Jung, D. and Kullak‐Ublick, G.‐A. Hepatocyte nuclear factor 1a: a key
­mediator of the effect of bile acids on gene expression. Hepatology, 2003;37
:622–31.
32. Staudinger, J.L., Woody, S., Sun, M., and Cui, W. Nuclear‐receptor‐mediated
regulation of drug‐ and bile‐acid‐transporter proteins in gut and liver. Drug
Metab Rev,2013;45:48–59.
33. Gartung, C., Ananthanarayanan, M., Rahman, M.A. et al. Down‐regulation
of expression and function of the rat liver Na+/bile acid cotransporter in
extrahepatic cholestasis. Gastroenterology, 1996;110:199–209.
34. Landmann, L. Cholestasis‐induced alterations of the trans‐ and paracellular
pathways in rat hepatocytes. Histochemistry,1995;103:3–9.
35. Green, R.M., Beier, D., and Gollan, J.L. Regulation of hepatocyte bile salt
transporters by endotoxin and inflammatory cytokines in rodents.
Gastroenterology, 1996;111:193–8.
36. Elsing, C., Reichen, J., Marti, U., and Renner, E.L. Hepatocellular Na+/H+
exchange is activated at transcriptional and posttranscriptional levels in rat
biliary cirrhosis. Gastroenterology, 1994;107:468–78.
37. Zollner, G., Fickert, P., Zenz, R. et al. Hepatobiliary transporter expression
in percutaneous liver biopsies of patients with cholestatic liver diseases.
Hepatology, 2001;33:633–46.
38. Trauner, M., Arrese, M., Lee, H., Boyer, J.L., and Karpen, S. Endotoxin
downregulates rat hepatic Ntcp gene expression via decreased activity of
critical transcription factors. J Clin Invest, 1998;101:2092–100.
39. Jung, D., Hagenbuch, B., Fried, M., Meier, P.J., and Kullak‐Ublick, G.A.
Role of liver‐enriched transcription factors and nuclear receptors in regulat-
ing the human, mouse, and rat NTCP gene. Am J Physiol Gastrointest Liver
Physiol, 2004;286:G752–61.
40. Eloranta, J.J., Jung, D., and Kullak‐Ublick, G.A. The human Na+‐taurocho-
late cotransporting polypeptide gene is activated by glucocorticoid receptor
and peroxisome proliferator‐activated receptor‐gamma coactivator‐1alpha,
and suppressed by bile acids via a small heterodimer partner‐dependent
mechanism. Mol Endocrinol,2006;20:65–79.
41. Karpen, S.J., Sun, A., Kudish, B. et al. Multiple factors regulate the rat liver
basolateral sodium‐dependent bile acid cotransporter gene promoter. J Biol
Chem, 1996;271:15211–21.
42. Denson, L.A., Auld, K.L., Schiek, D.S., McClure, M.H., Mangelsdorf, D.J.,
and Karpen, S.J. Interleukin‐1β suppresses retinoid transactivation of two
hepatic transporter genes involved in bile formation. J Biol Chem, 2000;275:
8835–43.
43. Denson, L.A., Sturm, E., Echevarria, W. et al. The orphan nuclear receptor,
shp, mediates bile acid‐induced inhibition of the rat bile acid transporter,
Ntcp. Gastroenterology, 2001;121:140–7.
44. Kullak‐Ublick, G.‐A., Beuers, U., Fahney, C., Hagenbuch, B., Meier, P.J.,
and Paumgartner, G. Identification and functional characterization of the
promoter region of the human organic anion transporting polypeptide gene.
Hepatology, 1997;26:991–7.
45. Jung, D., Podvinec, M., Meyer, U.A. et al. Human organic anion transport-
ing polypeptide 8 promoter is transactivated by the farnesoid X receptor/bile
acid receptor. Gastroenterology, 2002;122:1954–66.
46. Hagenbuch, B. and Stieger, B. The SLCO (former SLC21) superfamily of
transporters. Mol Aspects Med, 2013;34:396–412.
47. Zollner, G., Fickert, P., Silbert, D. et al. Adaptive changes in hepatobiliary
transporter expression in primary biliary cirrhosis. J Hepatol, 2003;38:
717–27.
48. Lee, W., Glaeser, H., Smith, L.H. et  al. Polymorphisms in human organic
anion‐transporting polypeptide 1A2 (OATP1A2): implications for altered
drug disposition and central nervous system drug entry. J Biol Chem,
2005;280:9610–7.
388 THE LIVER:  REFERENCES
49. Geier, A., Dietrich, C.G., Voigt, S. et  al. Effects of proinflammatory
cytokines on rat organic anion transporters during toxic liver injury and chol-
estasis. Hepatology, 2003;38:345–54.
50. Denk, G.U., Soroka, C.J., Mennone, A., Koepsell, H., Beuers, U., and Boyer,
J.L. Down‐regulation of the organic cation transporter 1 of rat liver in
obstructive cholestasis. Hepatology, 2004;39:1382–9.
51. Saborowski, M., Kullak‐Ublick, G.A., and Eloranta, J.J. The human organic
cation transporter 1 gene is transactivated by hepatocyte nuclear factor‐
4{alpha}. J Pharmacol Exp Ther, 2006;317(2):778–85.
52. Hofmann, A.F. and Hagey, L.R. Bile acids: chemistry, pathochemistry, biol-
ogy, pathobiology, and therapeutics. Cell Mol Life Sci, 2008;65:2461–83.
53. Goodwin, B., Jones, S.A., Price, R.R. et  al. A regulatory cascade of the
nuclear receptors FXR, SHP‐1, and LRH‐1 represses bile acid biosynthesis.
Mol Cell, 2000;6:517–26.
54. Lin, B.C., Wang, M., Blackmore, C., and Desnoyers, L.R. Liver‐specific
activities of FGF19 require Klotho beta. J Biol Chem, 2007;282:27277–84.
55. Inagaki, T., Choi, M., Moschetta, A. et al. Fibroblast growth factor 15 func-
tions as an enterohepatic signal to regulate bile acid homeostasis. Cell
Metab, 2005;2:217–25.
56. Schaap, F.G., van der Gaag, N.A., Gouma, D.J., and Jansen, P.L. High expres-
sion of the bile salt‐homeostatic hormone fibroblast growth factor 19 in the
liver of patients with extrahepatic cholestasis. Hepatology, 2009;49:1228–35.
57. Li, T., Jahan, A., and Chiang, J.Y. Bile acids and cytokines inhibit the human
cholesterol 7 alpha‐hydroxylase gene via the JNK/c‐jun pathway in human
liver cells. Hepatology, 2006;43:1202–10.
58. Liu, Y., Binz, J., Numerick, M.J. et al. Hepatoprotection by the farnesoid X
receptor agonist GW4064 in rat models of intra‐ and extrahepatic cholesta-
sis. J Clin Invest, 2003;112:1678–87.
59. Chen, J. and Farrell, G.C. Bile acids produce a generalized reduction of the
catalytic activity of cytochromes P450 and other hepatic microsomal
enzymes in vitro: Relevance to drug metabolism in experimental cholestasis.
J Gastroenterol Hepatol, 1996;11:870–7.
60. Marschall, H.U., Wagner, M., Bodin, K. et al. Fxr(‐/‐) mice adapt to biliary
obstruction by enhanced phase I detoxification and renal elimination of bile
acids. J Lipid Res, 2006;47:582–92.
61. Saini, S.P., Sonoda, J., Xu, L. et al. A novel constitutive androstane recep-
tor‐mediated and CYP3A‐independent pathway of bile acid detoxification.
Mol Pharmacol, 2004;65:292–300.
62. Miyata, M., Matsuda, Y., Tsuchiya, H. et al. Chenodeoxycholic acid‐mediated
activation of the farnesoid X receptor negatively regulates hydroxysteroid
sulfotransferase. Drug Metab Pharmacokinet, 2006;21:315–23.
63. Pircher, P.C., Kitto, J.L., Petrowski, M.L. et al. Farnesoid X receptor regu-
lates bile acid‐amino acid conjugation. J Biol Chem, 2003;278:27703–11.
64. Carlton, V.E., Harris, B.Z., Puffenberger, E.G. et al. Complex inheritance of
familial hypercholanemia with associated mutations in TJP2 and BAAT. Nat
Genet, 2003;34:91–6.
65. Konig, J., Nies, A.T., Cui, Y., Leier, I., and Keppler, D. Conjugate export
pumps of the multidrug resistance protein (MRP) family: localization, sub-
strate specificity, and MRP2‐mediated drug resistance. Biochim Biophys
Acta, 1999;1461:377–94.
66. Borst, P., Evers, R., Kool, M., and Wijnholds, J. The multidrug resistance
protein family. Biochim Biophys Acta, 1999;1461:347–57.
67. Bohan, A., Chen, W.S., Denson, L.A., Held, M.A., and Boyer, J.L. Tumor
necrosis factor alpha‐dependent up‐regulation of Lrh‐1 and Mrp3(Abcc3)
reduces liver injury in obstructive cholestasis. J Biol Chem,
2003;278:36688–98.
68. Chen, W., Cai, S.Y., Xu, S., Denson, L.A., Soroka, C.J., and Boyer, J.L.
Nuclear receptors RXRalpha:RARalpha are repressors for human MRP3
expression. Am J Physiol Gastrointest Liver Physiol, 2007;292:G1221–27.
69. Belinsky, M.G., Dawson, P.A., Shchaveleva, I. et al. Analysis of the in vivo
functions of Mrp3. Mol Pharmacol, 2005;68:160–8.
70. Huang, W., Zhang, Z., Chua, S.S. et al. Induction of bilirubin clearance by
the constitutive androstane receptor (CAR). PNAS, 2003;100:4156–61.
71. Okuwaki, M., Takada, T., Iwayanagi, Y. et  al. LXR alpha transactivates
mouse organic solute transporter alpha and beta via IR‐1 elements shared
with FXR. Pharm Res, 2007;24:390–8.
72. Soroka, C.J., Mennone, A., Hagey, L.R., Ballatori, N., and Boyer, J.L. Mouse
organic solute transporter alpha deficiency enhances renal excretion of bile
acids and attenuates cholestasis. Hepatology, 2010;51:181–90.
73. Lam, P., Wang, R., and Ling, V. Bile acid transport in sister of P‐glycoprotein
(ABCB11) knockout mice. Biochemistry, 2005;44:12598–605.
74. Maeda, K. and Sugiyama, Y. Impact of genetic polymorphisms of transport-
ers on the pharmacokinetic, pharmacodynamic and toxicological properties
of anionic drugs. Drug Metab Pharmacokinet, 2008;23:223–35.
75. Schrenk, D., Gant, T.W., Preisegger K‐H., Silverman, J.A., Marino, P.A., and
Thorgeirsson, S.S. Induction of multidrug resistance gene expression during
cholestasis in rats and nonhuman primates. Hepatology, 1993;17:854–60.
76. Bentires‐Alj, M., Barbu, V., Fillet, M. et al. NF‐kappaB transcription factor
induces drug resistance through MDR1 expression in cancer cells. Oncogene,
2003;22:90–7.
77. Geick, A., Eichelbaum, M., and Burk, O. Nuclear receptor response elements
mediate induction of intestinal MDR1 by rifampin. J Biol Chem, 2001;276:
14581–7.
78. Deleuze, J.F., Jacquemin, E., Dubuisson, C. et al. Defect on multidrug‐resist-
ance 3 gene expression in a subtype of progressive familial intrahepatic chol-
estasis. Hepatology, 1996;23:904–8.
79. Smit, J.J., Schinkel, A.H., Oude Elferink, R.P. et al. Homozygous disruption
of the murine mdr2 P‐glycoprotein gene leads to a complete absence of
phospholipid from bile and to liver disease. Cell, 1993;75:451–62.
80. Fickert, P., Fuchsbichler, A., Wagner, M. et al. Regurgitation of bile acids
from leaky bile ducts causes sclerosing cholangitis in Mdr2 (Abcb4) knock-
out mice. Gastroenterology, 2004;127:261–74.
81. Bull, L.N., Carlton VE.H., Stricker, N.L. et al. Genetic and morphological find-
ings in progressive familial intrahepatic cholestasis (Byler disease [PFIC‐1] and
Byler syndrome): evidence for heterogeneity. Hepatology, 1997;26:155–64.
82. de Vree, J.M., Jacquemin, E., Sturm, E. et al. Mutations in the MDR3 gene
cause progressive familial intrahepatic cholestasis. Proc Natl Acad Sci USA,
1998;95:282–7.
83. Jacquemin, E., Cresteil, D., Manouvrier, S., Boute, O., and Hadchouel, M.
Heterozygous non‐sense mutation of the MDR3 gene in familial intrahepatic
cholestasis of pregnancy. Lancet, 1999;353:210–11.
84. Pauli‐Magnus, C., Lang, T., Meier, Y. et  al. Sequence analysis of bile salt
export pump (ABCB11) and multidrug resistance p‐glycoprotein 3 (ABCB4,
MDR3) in patients with intrahepatic cholestasis of pregnancy. Pharmacogenet,
2004;14:91–102.
85. Lang, C., Meier, Y., Stieger, B. et al. Mutations and polymorphisms in the
bile salt export pump and the multidrug resistance protein 3 associated with
drug‐induced liver injury. Pharmacogenet Genomics, 2007;17:47–60.
86. Gotthardt, D., Runz, H., Keitel, V. et al. A mutation in the canalicular phos-
pholipid transporter gene, ABCB4, is associated with cholestasis, ductope-
nia, and cirrhosis in adults. Hepatology, 2008;48(4):1157–66.
87. Huang, L., Zhao, A., Lew, J.L. et al. Farnesoid X receptor activates transcrip-
tion of the phospholipid pump MDR3. J Biol Chem, 2003;278:51085–90.
88. Ghonem, N.S., Ananthanarayanan, M., Soroka, C.J., and Boyer, J.L.
Peroxisome proliferator‐activated receptor alpha activates human multidrug
resistance transporter 3/ATP‐binding cassette protein subfamily B4 tran-
scription and increases rat biliary phosphatidylcholine secretion. Hepatology,
2014;59:1030–42.
89. Meier, Y., Pauli‐Magnus, C., Zanger, U.M. et al. Interindividual variability of
canalicular ATP‐binding‐cassette (ABC)‐transporter expression in human
liver. Hepatology, 2006;44:62–74.
90. Buechler, M., Koenig, J., Brom, M. et al. cDNA cloning of the hepatocyte
canalicular isoform of the multidrug resistance protein, cMrp, reveals a novel
conjugate export pump deficient in hyperbilirubinemic mutant rats. J Biol
Chem, 1996;271:15091–8.
91. Ito, K., Suzuki, H., Hirohashi, T., Kazuhiko, K., Shimizu, T., and Sugiyama,
Y. Molecular cloning of canalicular multispecific organic anion transporter
defective in EHBR. Am J Physiol, 1997;272:G16–22.
92. Kartenbeck, J., Leuschner, U., Mayer, R., and Keppler, D. Absence of the
canalicular isoform of the MRP gene‐encoded conjugate export pump from
the hepatocytes in Dubin–Johnson syndrome. Hepatology, 1996;23:1061–6.
93. Oude Elferink, R.P.J., Ottenhoff, R., Liefting, W., Haan, J.D., and Jansen,
P.L.M. Hepatobiliary transport of glutathione and glutathione conjugate in
rats with hereditary hyperbilirubinemia. J Clin Invest, 1989;84:478–83.
94. Paulusma, C., van Geer, M., Evers, R. et al. Canalicular multispecific organic
anion transporter/multidrug resistance protein 2 mediates low‐affinity trans-
port of reduced glutathione. Biochem J, 1999;338:393–401.
95. Choi, J.H., Ahn, B.M., Yi, J. et al. MRP2 haplotypes confer differential sus-
ceptibility to toxic liver injury. Pharmacogenet Genomics,2007;17:403–15.
96. Lee, J.M., Trauner, M., Soroka, C.J., Stieger, B., Meier, P.J., and Boyer, J.L.
Expression of the bile salt export pump is maintained after chronic cholesta-
sis in the rat. Gastroenterology, 2000;118:163–72.
 31:  Adaptive Regulation of Hepatocyte Transporters in Cholestasis 389
  97. Strautnieks, S.S., Byrne, J.A., Pawlikowska, L. et al. Severe bile salt export
pump deficiency: 82 different ABCB11 mutations in 109 families.
Gastroenterology, 2008;134:1203–14.
  98. Stieger, B., Fattinger, K., Madon, J., Kullak‐Ublick G‐A., and Meier, P.J. Drug‐
and estrogen‐induced cholestasis through inhibition of the hepatocellular bile
salt export pump (Bsep) of rat liver. Gastroenterology, 2000;118:422–30.
  99. Donner, M.G., Schumacher, S., Warskulat, U., Heinemann, J., and Haussinger,
D. Obstructive cholestasis induces TNF‐{alpha}‐ and IL‐1 ‐mediated peripor-
tal downregulation of Bsep and zonal regulation of Ntcp, Oatp1a4, and
Oatp1b2. Am J Physiol Gastrointest Liver Physiol, 2007;293:G1134–46.
100. Lee, J.M., Trauner, M., Soroka, C., Steiger, B., Meier, P.J., and Boyer, J.L.
The molecular expression of the bile salt excretory pump, sister of P‐glyco-
protein (SPGP), is selectively preserved in cholestatic liver injury.
Hepatology, 1998;28(4):429A.
101. Kogan, D., Ananthanarayanan, M., Emre, S., Suchy, F.J., and Shneider,
B.L. The bile salt excretory pump (BSEP/SPGP) is not down‐regulated in
human cholestasis associated with extrahepatic biliary atresia (EHBA).
Hepatology, 1999;30:468A.
102. Ananthanarayanan, M., Balasubramanian, N., Makishima, M., Mangelsdorf,
D.J., and Suchy, F.J. Human bile salt export pump promoter is transactivated by
the farnesoid X receptor/bile acid receptor. J Biol Chem, 2001;276:28857–65.
103. Sinal, C.J., Tohkin, M., Miyata, M., Ward, J.M., Lambert, G., and Gonzalez,
F.J. Targeted disruption of the nuclear receptor FXR/BAR impairs bile acid
and lipid homeostasis. Cell, 2000;102:731–44.
104. Song, X., Kaimal, R., Yan, B., and Deng, R. Liver receptor homolog 1 tran-
scriptionally regulates human bile salt export pump expression. J Lipid Res,
2008;49:973–84.
105. Weerachayaphorn, J., Cai, S.Y., Soroka, C.J., and Boyer, J.L. Nuclear factor
erythroid 2‐related factor 2 is a positive regulator of human bile salt export
pump expression. Hepatology, 2009;50:1588–96.
106. Kubitz, R., Droge, C., Stindt, J., Weissenberger, K., and Haussinger, D. The
bile salt export pump (BSEP) in health and disease. Clin Res Hepatol
Gastroenterol,2012;36:536–53.
107. Suzuki, M., Suzuki, H., Sugimoto, Y., and Sugiyama, Y. ABCG2 transports sul-
fated conjugates of steroids and xenobiotics. J Biol Chem, 2003;278:22644–9.
108. Mennone, A., Soroka, C.J., Harry, K.M., and Boyer, J.L. Role of breast
cancer resistance protein in the adaptive response to cholestasis. Drug
Metab Dispos, 2010;38:1673–8.
109. Graf, G.A., Yu, L., Li, W.P. et al. ABCG5 and ABCG8 are obligate heter-
odimers for protein trafficking and biliary cholesterol excretion. J Biol
Chem, 2003;278:48275–82.
110. Kamisako, T. and Ogawa, H. Alteration of the expression of adenosine
triphosphate‐binding cassette transporters associated with bile acid and
cholesterol transport in the rat liver and intestine during cholestasis. J
Gastroenterol Hepatol, 2005;20:1429–34.
111. Omote, H., Hiasa, M., Matsumoto, T., Otsuka, M., and Moriyama, Y. The
MATE proteins as fundamental transporters of metabolic and xenobiotic
organic cations. Trends Pharmacol Sci, 2006;27:587–93.
112. Paulusma, C.C., Groen, A., Kunne, C. et  al. Atp8b1 deficiency in mice
reduces resistance of the canalicular membrane to hydrophobic bile salts
and impairs bile salt transport. Hepatology, 2006;44:195–204.
113. Bull, L.N., van Eijk, M.J., Pawlikowska, L. et al. A gene encoding a P‐type
ATPase mutated in two forms of hereditary cholestasis. Nat Genet,
1998;18:219–24.
114. Boyer, J.L. Bile duct epithelium: frontiers in transport physiology. Am
J Physiol, 1996;270:G1–5.
115. Hohenester, S., Wenniger, L.M., Paulusma, C.C. et  al. A biliary HCO3‐
umbrella constitutes a protective mechanism against bile acid‐induced
injury in human cholangiocytes. Hepatology, 2012;55:173–83.
116. Medina, J.F., Martinez‐Anso, E., Vazquez, J.J., and Prieto, J. Decreased
anion exchanger 2 immunoreactivity in the liver of patients with primary
biliary cirrhosis. Hepatology, 1997;25:12–7.
117. Concepcion, A.R., Lopez, M., Ardura‐Fabregat, A., and Medina, J.F. Role of
AE2 for pHi regulation in biliary epithelial cells. Front Physiol, 2013;4:413.
SECTION D:
NON‐HEPATOCYTE CELLS
 Cholangiocyte Biology
32 and Pathobiology
Massimiliano Cadamuro1,2, Romina Fiorotto2,3, and Mario Strazzabosco2,3
1
Department of Molecular Medicine, University of Padua, Padova, Italy
2
International Center for Digestive Health (ICDH), University of Milan‐Bicocca, Monza, Italy
3
Liver Center and Section of Digestive Diseases, Department of Internal Medicine, Section of Digestive
Diseases, Yale University School of Medicine, New Haven, CT, USA

INTRODUCTION CoH connect the hepatocellular canalicular network, carrying

Cholangiocytes are the epithelial cells that line the intrahepatic
and extrahepatic biliary tree. The last two decades have wit-
nessed a significant expansion of our understanding of the func-
tion and dysfunction of this important epithelium. We have also
understood that cholangiocytes are important actors in liver
repair and progression of liver fibrosis and play a fundamental
role in liver immunobiology. Several chapters and reviews have
been recently written on a number of aspects of cholangiocyte
biology. In this review we elected to focus on those mechanisms
that are more pathophysiologically relevant to human diseases
and that may represent possible avenues for basic and transla-
tion research in the next few years.
FUNCTIONAL ANATOMY
OF THE BILIARY TREE
The biliary system is a complex network of tubules coated by
epithelial cells, or cholangiocytes, which starts from the canals
of Hering in the liver lobules (intrahepatic biliary tree), contin-
ues outside the liver (extrahepatic biliary tree), and terminates
into the ampulla of Vater. The best known function of the biliary
tree is the transport and modification of the primary bile secreted
by the hepatocytes, but it is now widely recognized that the bil-
iary tree is actually a functionally complex structure endowed
with many different biological functions reflected in a morpho‐
functional specialization of cholangiocytes [1, 2]. The first
structure belonging to the biliary tree is represented by the
canals of Hering (CoH), that is composed 50% by cholangio-
cytes and 50% by hepatocytes juxtaposed with each other [3, 4];
the primary bile, with the intraportal ramifications of the biliary
tree. CoH is also considered the site of the putative liver stem
cell niche. The intrahepatic biliary tree gradually merges from
cholangioles, interlobular, septal, areal, and segmental ducts
that end in the two principal hepatic ducts and then the common
hepatic duct. The latter receives the cystic duct coming from the
gallbladder and connects the liver and gallbladder to the intes-
tine. The extrahepatic biliary network is surrounded by a capil-
lary plexus and by peribiliary glands, a stem cell niche of hepatic
progenitor cells that differs from that of the CoH. Notably, the
intra‐ and extrahepatic biliary tree has a different embryonic ori-
gin [5, 6].
Cholangiocytes belonging to different compartments of the
biliary tree are characterized by a specialized morphology,
reflecting distinct physiological function. The smallest biliary
epithelial cells lining the CoH and the distal branches of the
biliary tree have a cuboidal shape with a round basal nucleus,
and are quickly reactive to liver and/or biliary damage. The
large cholangiocytes, lining the bile ducts of higher diameter,
are usually columnar and, thanks to their transport abilities,
mediate alkalinization, hydration, and modification of the bile
[1, 7]. Cholangiocytes in fact, have both secretory and absorp-
tive functions, being involved in the recirculation of bile acids,
glucose, and water. Notably, about 40% of human bile is gener-
ated by biliary epithelial cells, following modification of the
primary product of the hepatocytes. Data in animals suggest that
while secretory functions of large cholangiocytes are regulated
mostly by cyclic adenosine monophosphate (cAMP) signaling,
the most important second messenger in small cholangiocytes is
[Ca2+]i [7, 8]. These distinctions are, however, less actual now
that we have a better understanding of the microdomain restric-
tion of second messenger signaling.

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
394 THE LIVER:  DEVELOPMENT OF THE BILIARY TREE
DEVELOPMENT OF THE BILIARY TREE
The extra‐ and intrahepatic branches of the biliary tree originate
from different embryological areas, thus explaining their mor-
phological and biological properties. The extrahepatic biliary
tree originates from Hex−/Pdx1+/Sox17+ cells in the caudal part
of the ventral foregut [9, 10]. Studies conducted on Alagille syn-
drome demonstrate that perturbation of the Notch pathway, one
of the morphogenetic signaling pathways involved in the biliary
specification of the hepatoblasts, causes intrahepatic ductopenia
without affecting the extrahepatic biliary structures, and sug-
gests that different signals govern the formation of the intra‐ and
extrahepatic structures [11].
Conversely, intrahepatic bile ducts derive from Hex+/Pdx‐1−/
Sox17−/AFP+ hepatoblasts, bipotent cells present in the fetal
liver bud. Starting from the eighth gestational week (GW),
hepatoblasts in contact with mesenchymal cells located in the
nascent portal tract, form a continue single cells rim of keratin
K8+/18+/19+ cells called the ductal plate (DP). Between twelfth
and sixteenth GW, the DP duplicates in discrete areas, represses
the expression of hepatocellular markers and begins to express
K7 and K19, that is, markers of commitment toward a biliary
phenotype. The duplicating DP progressively forms a lumen,
while the residual biliary monolayers are gradually reabsorbed
by apoptosis or by retro‐differentiation to periportal hepato-
cytes, as recently shown by lineage tracking experiments [12].
Duplicating DPs are gradually incorporated into the portal mes-
enchyma where they are joined by a peribiliary capillary plexus,
which nourishes the nascent biliary tree, and by vascular struc-
tures that will become the portal arteries [13].
The concomitant elongation and maturation of biliary epithe-
lium and vascular structures requires the cooperation of differ-
ent cell types equipped with several specific ligands, receptors,
and other morphogens, among them angiogenic growth factors,
such as vascular endothelial growth factor (VEGF) A, platelet‐
derived growth factor (PDGF), and the angiopoietins, Ang‐1
and Ang‐2, and their specific receptors, VEGFR1, VEGFR2,
PDGFRβ, and Tie‐2 [13, 14]. In addition, a range of morphoge-
netic signaling and transcription factors, including Notch and
WNT/β‐catenin pathways, transforming growth factor‐β
(TGF‐β), and hepatocyte nuclear factors (HNFs) are activated
[9, 15]. Furthermore, recent studies outlined the importance of
miRNAs as regulatory mediators coordinating the interactions
among all those peptides. It is important for the experimental as
well as the clinical hepatologist to acquire a working knowledge
of these mechanisms, since an altered development of the bil-
iary tract is responsible for a number of liver malformative dis-
eases, among which the so‐called DP malformations (DPM) are
of particular interest [16–18].
A fundamental family of growth factors involved in biliary
fate determination is TGFβ, and its receptors; in particular,
TGFβ type II‐R (TβRII). All TGFβ ligands are highly expressed
around the nascent portal area, and induce the transition of
hepatoblasts into cholangiocytes, by binding to TβRII, which is
transiently expressed by the monostratified DP cells. This effect
is, at least in part, mediated by the upregulation of Jag1 and by
the modulation of the transcription factors Hes1 and Hey1, two
downstream effectors of Notch signaling. As demonstrated by
studies on different animal models (zebrafish and mice)
[19–21], periportal hepatoblasts expressing Notch2 are induced
by Jag1+ portal mesenchymal cells to acquire cholangiocyte‐like
morphology and phenotype (DP cells). The interaction between
Jag1 and its receptor Notch2 induces the cleavage and nuclear
import of the Notch intracellular domain (NICD) that, in coop-
eration with TGFβ, binds to the nuclear transcription factor
recombinant signal binding protein for immunoglobulin kappa J
(RBP‐Jk) and activates Hes1 that, in turn promotes the expres-
sion of the biliary epithelial cell‐specific transcription factors
Hnf1β, Sox4, and Sox9. Notably, mutations on Jag1 or Notch2,
cause Alagille syndrome, a multiorgan disease characterized by
vanishing of intrahepatic biliary structures, and by the accumu-
lation of cells retaining both hepatic and cholangiocyte phe-
notype (intermediate hepatobiliary cells) [11, 22]; jaundice,
itching, and hypercholesterolemia characterize the hepatic
phenotype.
Recent work by Diehl’s group has clarified the important role
of the Hedgehog (Hh) pathway in biliary morphogenesis, repair,
and cancer. This signaling is normally repressed by the interac-
tion between Patched (Ptc) and Smoothened (Smo) [23–25].
Once the Hh ligand (i.e. Sonic, SHh) binds its receptor Patched
(Ptc), the signal pathway is de‐repressed and Gli3A induces the
transcription of the downstream effectors Ptc, Glioblastoma
(Gli) 1 and Gli2. Recent data support an important role of Hh in
bile duct morphogenesis; in fact, in the fetal mouse liver, SHh
expression together with Gli1, increase at the earliest gesta-
tional days (E11.5) and rapidly decrease overtime. Its reported
association with the Meckel syndrome, a lethal DP malforma-
tion, highlights the morphogenetic properties of Hh signaling.
Hh also has modulatory effects on two other fundamental tran-
scription factors, Yes‐associated protein (YAP) and Sox9. In
mice, YAP is expressed at E15.5 by several cells accumulated in
the periportal areas around the nascent portal tract; at E18.5 the
clear majority of YAP+ cells co‐express Sox9, a marker of
­biliary differentiation, which is also downstream of Notch
signaling [26]. The role of YAP as a master gene involved in
differentiation of cholangiocyte precursors is also confirmed by
the essential role played by the Hippo pathway in regulating
liver size during liver regeneration after partial hepatectomy
(PHx) [27]. Furthermore, recent data [28] show that Sox9 and
Sox4, are involved in the maintenance of the apical–basal
­polarity of the epithelial sheet, in the correct development of the
primary cilium and, as well as in the normal formation, elongation
and branching of the biliary tree. This mechanism, which is
known as “planar cell polarity” (PCP), is primarily governed by
the non‐canonical Wnt/β‐catenin pathway, and is altered in
some kidney and liver ciliopathies (diseases deriving from a
faulty assembly of primary cilia or from the malfunction of
­proteins expressed in cilia) [29].
In summary, the available information highlight the intensive
cross‐talk mechanisms among the different morphogens and
likely a redundancy of the signals regulating biliary tree archi-
tecture, in health and in the reparative/regenerative response to
liver damage. More recently, miRNAs have been suggested to
play a role in biliary tree specification and development; Rogler
and colleagues showed that the treatment of mice fetuses at
E16.5 with miR‐23b, miR‐27b, and miR‐24 antagonists per-
turbed the correct development of the fetal liver and stimulated
the expression of keratin 19 in the liver parenchyma due to the
 32:  Cholangiocyte Biology and Pathobiology 395
upregulation of the TGFβ signaling [30]. Those miRNAs in fact
target different components of the TGFβ pathway, and in par-
ticular at the level of the small mother against decapentaplegic
(SMAD) [31].
PHYSIOLOGICAL FUNCTIONS
OF THE BILIARY TREE
Within the biliary tree there is a functional and morphological
specialization between small and large ducts; in particular, the
main secretory functions are sustained by cholangiocytes lining
interlobular, septal, and major ducts, while the ability to react to
liver damage and possibly function as bipotential progenitor
cells resides in the small ducts and canals of Hering, respec-
tively [32–36].
SECRETION AND BILE PRODUCTION
The best understood function of cholangiocytes is to modify the
volume, fluidity, and alkalinity of the primary bile secreted by
the hepatocytes. In addition, the biliary epithelium may reab-
sorb water, glucose, glutathione, bile acids, and electrolytes
[37]. Actually, biliary epithelial cells are able to modify the
composition of the primary bile secreted by the hepatocytes into
the bile canaliculi and in humans, up to 40% of the bile is pro-
duced by the biliary epithelium, depending on the digestive
phases. Primary bile flows through the CoH to the biliary net-
work where cholangiocytes lining the larger biliary structures
modify it. Changes in bile volume and composition are modu-
lated by a complex interplay among different pro‐secretory hor-
monal and paracrine stimuli, among which secretin and ATP are
important examples [38, 39], and anti‐secretory factors such as
endothelin‐1 (ET‐1) [40], gastrin, and somatostatin [41]. These
secretory and anti‐secretory stimuli mostly work by increasing
or decreasing the levels of cellular cAMP as will be described in
detail below.
The above described secretory mediators are integrated by a
number of membrane‐bound or soluble adenylyl cyclase (ACs),
such as AC4, 5, 6, 7, 8, 9, and SAC (soluble adenyl cyclase),
resulting in given levels of intracellular levels of cAMP, and
activation of PKA or epithelial Na+ channel (ENAC). Recent
works have highlighted the importance of microdomains in
cAMP/PKA activation. Some of these microdomains may be
associated with a specific ACs that binds PKA to selected pro-
teins. Activation of PKA stimulates the cystic fibrosis trans-
membrane conductance regulator (CFTR) channel [42] that
mediates the luminal secretion of Cl−. Cl− secreted into the
lumen creates the driving force for the activation of an apically
located Na+‐independent Cl−/HCO3− exchanger (AE2) that
extrudes HCO3− in exchange for Cl− from the apical surface of
the biliary epithelial cells. The concentration of anions into the
lumen of the bile ducts creates an osmotic gradient that attracts
water into the bile ducts, through aquaporin‐1 and ‐4. To pre-
serve cellular homeostasis, the secretion of bicarbonate is coun-
terbalanced by the intracellular import of anions through the
Na+‐dependent Cl−/HCO3− exchanger (NCHE), and by the Na+/
HCO3− cotransporter (NCB1) present at the basolateral side of
the cholangiocytes. The difference of potential between the
intra‐ and extracellular sides of the epithelium is maintained by
two Na+ transport systems: the Na+/K+/2Cl− cotransporter
(NKCC1) and the Na+/K+ adenosine triphosphate (ATP)‐ase [8].
These coordinated mechanisms are necessary to increase the
alkalinity and the volume of the bile and their dysfunction can
be responsible for pathologic conditions. In cystic fibrosis (CF),
for example, a genetic disease caused by a genetically transmit-
ted defect in the function of CFTR, biliary complications are
increasingly being recognized. In cystic fibrosis liver disease
(CFLD) defective CFTR function alters the coordinated activa-
tion of AE2 and the secretion of bicarbonate and fluid into the
bile. Alterations of the bile composition precipitates a cascade
of events such as the formation of bile plugs, accumulation of
toxicants (i.e. toxic bile acids, endotoxins) that damage the epi-
thelium and causes focal biliary inflammation progressing to
sclerosing cholangitis and eventually cirrhosis [43]. Recent
findings (revised in sections on Barrier Function of
Cholangiocytes and in Cholangiocytes and Immunity) have
highlighted a novel function of CFTR as a regulator of endo-
toxin tolerance in cholangiocytes, thanks to its interaction with
other proteins (i.e. Src tyrosine kinases), and this will change
our view of CFLD from a classic channelopathy to an inflam-
matory disease (Figure 32.1).
Bicarbonate secretion is not only necessary to drive bile
secretory processes but it is also an important defense mecha-
nism developed by cholangiocytes to protect themselves from
several toxicants present in the bile and by the action of apolar
protonated hydrophobic bile acids. The vast majority of human
bile salts is glycine‐conjugated and are usually partially proto-
nated, and apolar, becoming membrane‐permeable at pH 7.4
[44]. Thus, hydrophobic bile salts accumulate into the cell cyto-
plasm and may cause cell damage, including apoptosis at micro-
molar concentrations. Starting from these observations, it was
hypothesized that the secretion of HCO3− above the apical mem-
brane of cholangiocytes provides a sort of local bicarbonate
shield (or umbrella) able to protect the epithelium. In physio-
logic conditions, several mechanisms are at play to generate and
maintain this bicarbonate rich microenvironment [45]. In nor-
mal cholangiocytes, the primary bile acid chenodeoxycholic
acid (CDCA) interacts with its receptor TGR5, a membrane‐
bound bile acid receptor coupled to a stimulatory G‐protein,
localized both on the apical membrane and on the primary cil-
ium. Activation of TGR5 is thought to increase cAMP concen-
tration and activate the excretion of Cl−, together with ATP
through CFTR. ATP released into the bile binds specific surface
purinergic receptor P2Y2 that increases intracellular Ca2+ levels
and stimulates by an autocrine loop the secretion of Cl− from
calcium‐dependent Cl‐channels (i.e. TMEM16A). Cl− is trans-
ported back into the cytoplasm by the AE2 exchanger as
described in the section on Secretion and Bile Production.
Finally, the action of alkaline phosphatases, located in the apical
glycocalyx in close proximity to the P2Y receptors ensures the
correct amount of bicarbonate secretion by degradation of ATP
in ADP and AMP [46].
It is hypothesized that a “destabilization” of the biliary bicar-
bonate shield could generate changes in the luminal pH that
396 THE LIVER:  BILIARY PRIMARY CILIA AND “CILIOPATHIES”

Figure 32.1  Bile secretion and modification in cholangiocytes. Cholangiocytes, through a complex network of pumps, ion exchangers, and
channels are able to modify bile composition, altering its volume, pH, and hydration. Prosecretory hormones (i.e. secretin) bind their respective
receptors stimulating the intracellular increase of cAMP due to the activity of ACs, leading to the activation of CFTR that extrude Cl− anions in the
luminal space together with ATP. Cl− is, in turn, reabsorbed by AE2 extruding HCO3−, responsible for the alkalinization of the bile. cAMP levels are
further sustained by the activation of TGR5 on the apical side of cholangiocytes. The increased osmotic gradient at the apical side of the cholangio-
cytes stimulates H2O flow passively through AQPs (AQP1, AQP4), as well as through a paracellular pathway; ATP, also secreted into the bile by
CFTR, activates the P2Y purinergic receptor and Cl− secretion. Cellular ion homeostasis is maintained by the presence of several ion transporters
at the basal side of cholangiocytes, such as NCHE, NCB1, and Na+/H− pump. This mechanism leads to HCO3− secretion into the bile, which together
with the presence of a glycocalyx, constitutes a well‐known defensive mechanism that is also at the basis of the cholehepatic shunting of weak acids
and bile acids (see [127]), dubbed the “bicarbonate umbrella”. Abbreviations: cAMP, cyclic adenosine monophosphate; ACs, adenylyl cyclase;
CFTR, cystic fibrosis transmembrane conductance regulator; AE2, Na+‐independent Cl−/HCO3− exchanger; AQPs, aquaporins; ENAC, epithelial
Na+ channel; NCB1, Na+/HCO3− cotransporter; NCHE, Na+‐dependent Cl−/HCO3− exchanger; NHE‐1, sodium‐hydrogen antiporter 1.

alter biliary homeostasis and bile salt physicochemical status
and causes toxic effects to the biliary epithelium, thereby trig-
gering a chronic cholangitis and scarring and sclerosing cholan-
gitis akin to the pathophysiological sequence described in the
Mdr3‐knockout (KO) mouse [44, 45, 47].
In chronic cholangitis, AE2 expression seems to be reduced
due to the ex novo expression of miR‐506 that suppresses the
translation of AE2 by binding its 3′UTR region. The reduced
AE2 activity has a dual effect, that on one side destabilizes the
biliary umbrella and the buffering effect on bile salts, and sec-
ondarily leads to an intracellular accumulation of HCO3− that
activates the soluble AC (sAC) and sensitizes cholangiocytes to
apoptosis. Treatment with ursodeoxycholic acid (UDCA) or the
homolog norUDCA could in part activate the luminal secretion
of HCO3−, through a Ca2+/cPKCα/PKA mechanism and the
P2Y‐dependent signaling and thus reduce the necroinflamma-
tory damage [48].
BILIARY PRIMARY CILIA
AND “CILIOPATHIES”
The apical surface of cholangiocytes presents a non‐motile
primary cilium. Recent studies have highlighted its importance
in cholangiocyte biology and pathobiology. Primary, or sensory,
cilia are non‐motile structures, characterized by a 9 + 0 configu-
ration, in which, nine doublet microtubules, or axoneme, are
 32:  Cholangiocyte Biology and Pathobiology 397
anchored to the basal body [49]. Primary cilia express several
proteins, including polycystin 1 and 2 (PC1, and PC2), and
fibrocystin (FPC) that could participate in a number of intracel-
lular signaling cascades. Primary cilia can act as osmo‐ or
mechano‐receptors in biliary structures. Cilia sense the direc-
tion of the bile flow and, by bending, activate calcium channels
allowing the intracellular influx of Ca2+ ions. These structures
are likely involved in cell proliferation and senescence, activa-
tion of progenitor cell compartment, regeneration, and develop-
ment through the modulation of the Hh and canonical WNT
pathways. Furthermore, signaling mechanisms located in cilia
may supervise the correct cell planarity and polarity through
non‐canonical WNT signaling [50, 51]. During biliary develop-
ment, the presence and the correct conformation of the primary
cilia is necessary for the correct assembly of the biliary struc-
tures; in fact, its absence is observed in three rare but lethal syn-
dromes of infants, the Meckel [52], Joubert [53], and Jeune [54]
syndromes. These are multiorgan diseases and the liver is char-
acterized by cyst‐like, dysmorphic biliary structures surrounded
by a dense fibrotic stroma, resembling congenital hepatic fibro-
sis (CHF). Several liver diseases, known as ciliopathies, may be
related to primary cilia defects, including CHF, Caroli’s disease
(CD), autosomal dominant polycystic kidney disease (ADPKD).
ADPKD is a genetic liver and kidney disease due to mutation on
the gene encoding for PC1 (80%) and PC2 (20%) characterized
by progressive and massive enlargement of cyst covered by bil-
iary epithelia; patients with ADPKD may develop severe com-
plications such as mass effect, cyst hemorrhage, rupture, or
infection requiring urgent liver transplantation. CHF, CD, and
autosomal recessive polycystic kidney disease (ARPKD), are
rare inherited diseases of the renal tubular and biliary epithe-
lium, characterized by enlargement of the biliary tree with cyst‐
like features associated with portal fibrosis and inflammation.
These diseases, all due to mutations in the Pkhd1 gene, encod-
ing for FPC cause severe portal hypertension and may be com-
plicated by acute cholangitis, intrahepatic lithiasis, and also
cholangiocarcinoma. The influences of ciliary protein dysfunc-
tion in the pathogenesis of cystic liver diseases will be explained
in the following chapters. Interestingly, mice carrying mutations
for the Itf88 gene, also known as Tg737, encoding for the ciliary
protein polaris, display several malformations of the kidney and
biliary cystic dilatation, accompanied by pericystic fibrosis
resembling the ARPKD phenotype [55].
BARRIER FUNCTION
OF CHOLANGIOCYTES
A fundamental function of epithelial cells, including cholangio-
cytes, is to selectively control the diffusion of ions and mole-
cules through the epithelial barrier. Biliary epithelial cells,
thanks to their ability to secrete ions in a polarized fashion and
to their selective permeability to solute and water, actively
maintain liver homeostasis. Moreover, the biliary epithelium
functions as a barrier against the back‐diffusion of xenobiotics,
toxic metabolites, and bile salts from the bile to the interstitial
tissue. Primary or secondary changes in the correct formation
and polarization of the epithelial sheet have been reported to
play a role the pathogenesis of several liver diseases, including
primary biliary cholangitis (PBC), primary sclerosing cholangi-
tis (PSC), and CFLD [56–58].
The tight junctions (TJs) are among the main structures
responsible for epithelial barrier function and are of primary
importance for the correct conformation of the biliary struc-
tures; unfortunately, to date, very little is known about mecha-
nisms controlling their function at the level of canaliculi and the
bile ducts. In general, assembling and disassembling of TJ are
modulated by different signaling mechanisms, including protein
kinases, and G‐proteins; probiotics, glutamine, and growth fac-
tor protect the correct assembly of TJs, while disruption of TJ
are induced by pathogens, toxins, and inflammatory cytokines.
Experiments on polarized monolayers of cholangiocytes dem-
onstrated that treatment with TNFα, as well as with lipopolysac-
charides (LPS) and nitric oxide (NO) increase the paracellular
permeability.
In particular, in human and mouse in vitro cultures of cholan-
giocytes carrying mutations for CFTR treated with LPS
increases permeability to dextrans by inducing a derangement
of the normal F‐actin cytoskeletal shape, and the delocalization
of E‐cadherin, a cytoskeletal structural protein typical of differ-
entiated epithelial cells [58]. In normal cholangiocytes, CFTR is
localized on the apical membrane where it functions as a chan-
nel protein but also interacts in a multiprotein complex and
regulates the function of other proteins. Examples are proteins
that negatively regulate the activity of Src family tyrosine
kinases (SFK). In cholangiocytes that lack CFTR at the mem-
brane, Src is indeed more active and phosphorylates the LPS
receptor TLR4. An aberrant activation of TLR4 in response to
LPS and the increased downstream activation of NF‐κB with
production of inflammatory mediators are directly responsible
for the incorrect reshaping of the F‐actin cytoskeleton and con-
sequent increased permeability [58, 59]. Notably, the treatment
with PP2, an SFK inhibitor, prevents the improper assembly of
the cytoskeleton and rescues the paracellular permeability of
CF‐KO cells. In contrast to Src, activation of tyrosine kinases
(TKs), such as EGFR, have a protective effect on TJ conforma-
tion following hepatic noxae. The activation of the pathway
mediated by EGFR inhibits Src phosphorylation by increasing
concentration and activation of the phospholipase Cγ (PLCγ)/
PKC axis. Altered TJ permeability and back‐diffusion of bile
acids may play a pathogenic role in experimental conditions, as
well as in human biliary diseases.
CHOLANGIOCYTE REACTION
TO BILIARY DAMAGE
Biliary epithelial cells are usually quiescent, however, following
a liver insult, cholangiocytes activate and/or proliferate as a part
of the so‐called “hepatic reparative complex”. A typical element
of the hepatic repair response to liver damage is the “ductular
reaction” (DR), a stereotyped histopathological lesion of the bil-
iary epithelium, which plays a fundamental role in the progres-
sion of hepatic fibrosis. DR is characterized by a marked
proliferation of cholangiocytes with poor cytoplasm, arranged
in cell cords without a lumen or in richly anastomosed small
398 THE LIVER:  CHOLANGIOCYTE PROLIFERATION
diameter ducts (<10 μm) with almost unrecognizable lumens
[51]. An infiltrate composed by innate and adaptive immune
cells is invariably present. Reactive ductular cells (RDC) are
activated epithelial cells that secrete a vast array of factors,
including cytokines, chemokines, growth factors, and angio-
genic factors [60]. Different theories, not mutually exclusive,
have been proposed to explain the genesis of RDCs: it was
hypothesized that they may derive from hepatocytes undergoing
a process of ductular metaplasia, or from the activation of the
hepatic progenitor cells (HPC) compartment, and/or from pro-
liferation and de‐differentiation of preexisting cholangiocytes
[51]. HPC is a population of bipotent stem cells, whose niche is
localized at the level of the CoH, the boundary structure con-
necting the hepatocytic bile canaliculi with the terminal bile
ducts. Although this concept is currently being challenged,
HPCs are thought to be able to expand and differentiate into
hepatocytes, through the formation of cells with intermediate
phenotype (intermediate hepatobiliary cell, IHBC), or in chol-
angiocytes, or in RDC as a response to non‐resolving damage
[51]. RDCs’ phenotype is less differentiated than normal chol-
angiocytes and expresses molecular markers of immaturity such
as chromogranin A, neural cell adhesion molecule (NCAM),
and B‐cell protein leukemia lymphoma‐2 (Bcl‐2). Interestingly,
NCAM mediates the interaction between cells and matrix dur-
ing the development of different epithelial tissues, while Bcl‐2
inhibits apoptosis during development and the growth of new
epithelial structures  [61]. The increase in RDCs is associated
with a significant increase in inflammatory infiltrate and portal
fibrosis, as demonstrated in viral hepatic diseases [62], meta-
bolic disorders [63], or disorders of embryogenesis, such as
Alagille syndrome and biliary atresia [22, 63]. Deposition of
fibrosis is based on a paracrine cross‐talk mediated by the abil-
ity of RDCs to secrete profibrotic and proinflammatory growth
factors including interleukin‐6 (IL‐6), TGFβ2, endothelin‐1
(ET‐1), monocyte chemotactic protein‐1 (MCP‐1), platelet‐
derived growth factor‐B (PDGF‐B), TNF‐α, NO, vascular
endothelial growth factor (VEGF) [8, 60], and connective tissue
growth factors (CTGF). These are among many other factors
and cyto/chemokines that send signals able to recruit inflamma-
tory and immune cells, mesenchymal cells, and endothelial
cells. RDC are the masterminds of these dynamic interactions
and are therefore considered the “pacemakers” of portal fibrosis
[64]. In addition to inflammatory cells, of relevance is the cross‐
talk with cells of mesenchymal origin, in particular Kupffer
cells and portal fibroblasts, which have the role as the main
effectors of fibrosis, as stimulators of the deposition of extracel-
lular matrix (ECM). In addition, RDC also establish paracrine
communications with endothelial cells that provide the vascular
support necessary for the growth and arborization of the ductal
structures themselves [13]. Most of these factors are also tran-
siently expressed by the DP during fetal development, an obser-
vation in line with the hypothesis that the DR recapitulates,
morphogenetic signaling patterns observed during hepatic
ontogenesis [13] (Figure 32.2).
An important novel stem cell niche is represented by the
“peribiliary glands”. These glands are located in the area of the
common hepatic duct at the hepatic hilum, in the cystic duct,
and in the hepatic papilla [65]. These glands contain cells
expressing a number of markers and transcription factors
consistent with their stem cell nature, such as Pdx1, Sox9,
Sox17, and EpCAM, together with markers of mature epithelial
cells, such as keratin 7 [66, 67]. Notably, these putative stem
cells could be isolated through magnetic immunoselection and
when cultured by using specific matrix and growth factors, they
could commit to hepatocyte, cholangiocyte, and pancreatic
lineage.
CHOLANGIOCYTE PROLIFERATION
Proliferation of cholangiocytes is fundamental not only for the
maintenance of the normal homeostasis of the biliary tree, but
also in response to liver damage. Proliferation of biliary epithe-
lial cells (BEC) is mediated by the autocrine and/or paracrine
action of several growth factors and hormones released in the
local microenvironment by inflammatory and mesenchymal
cells or by epithelial cells themselves. One of the critical
cytokines that stimulate cholangiocyte proliferation is IL6,
together with other members of the IL6 family, such as
Oncostatin M [68, 69]. Other growth factors significantly
involved in BEC proliferation include hepatocyte growth factor
(HGF) [70], insulin‐like growth factor (IGF)‐1 [71], epidermal
growth factor (EGF) [72], and VEGF‐A [13, 14]. This response
is sustained also by different hormones such as secretin, a major
activator of intracellular cAMP, or estrogen, testosterone, prol-
actin, and follicle stimulating hormone (FSH). This mechanism
is finely counterbalanced by the action of antiproliferative pep-
tides such as gastrin, melatonin, somatostatin, serotonin, and
gamma aminobutyric acid (GABA) [73].
Genetic diseases of the biliary epithelium, such as ADPKD
may be considered as a model to study the role played by some
of these growth factors in stimulating cholangiocyte prolifera-
tion. VEGF, angiopoietins, and their related receptors are over-
expressed in the biliary epithelium of ADPKD. ADPKD is
caused by mutations in either PKD1 or PKD2 genes, encoding
for PC1 and PC2, and are characterized by the formation of
multiple cysts in the kidney and in 90% of the cases of bile duct‐
derived cysts [74]. The production of angiogenic factors by the
cystic epithelium is a recapitulation of biliary development and
a feature of de‐differentiation. In fact, during the DP stage, the
developing epithelium secretes VEGF that acts on the endothe-
lial cell precursors and promotes the peribiliary vascularization.
In the same way, in ADPKD the growing cystic epithelium
secretes VEGF to support the expansion of the surrounding
­vascular supply [14].
The relationship between angiogenic signaling and cholan-
giocyte proliferation in ADPKD was demonstrated by studies in
PC2 defective mice. PC2 is present on the surface of the primary
cilium and of different organelles of the cholangiocytes of
unclear function, but is involved also in calcium homeostasis in
the endoplasmic reticulum (ER), cilium, and cytoplasm. In
physiologic conditions, bile flow stimulates the activation of
Ca2+ signaling by a rapid release of Ca2+ from the ER mediated
by IP3R, followed by a sustained Ca2+ entry from the plasma
membrane. This mechanism is also known as store operated
Ca2+ entry (SOCE). When PC2 is defective, store activated cal-
cium signaling is altered and cytoplasmic Ca2+ levels are lower.
 32:  Cholangiocyte Biology and Pathobiology 399

Figure 32.2  Morphogens involved in epithelial mesenchymal interactions in ductular reaction. In response to liver damage, HPC and RDC com-
partments start to proliferate to restore the normal structure of the liver; the abnormal increase of these structures are responsible for the accumula-
tion of peribiliary fibrosis and of portal inflammatory infiltrates that in turn further stimulate this anomalous process. MF surrounding RDC
stimulates the development of an increased dysmorphic biliary bed by the activation of the Notch2/Jag1 signaling and by secreting Hh ligands that
dock to their receptor Ptc on the surface of the cholangiocytes. Moreover, TGFβ exerts its trophic properties on RDC by binding its specific receptor
TGFβR. Abbreviations: CoH, canal of Hering; Hep, hepatocyte; Hh, hedgehog; HPC, hepatic progenitor cells; BD, bile duct; RDC, reactive ductu-
lar cell; MF, myofibroblast; Ptc, Patched.

This activates adenyl cyclase 5 (AC5) a Ca2+ inhibitable AC,
which, through PKA, phosphorylates ERK1/2 that upregulates
HIF1α, a potent inducer of VEGF‐A production by BEC.
VEGF‐A, in turn, through an autocrine loop binds its receptor
VEGFR2 present on the surface of the PC2 defective cholangio-
cytes, which further sustains the proliferation of the BEC acti-
vating the Raf/MEK/ERK1/2 pathway. Furthermore, PKA
activation inhibits tuberin, a repressor of mammalian target of
rapamycin (mTOR), a downstream effector of the IGF‐1R that,
following activation by its ligand IGF‐1, stimulates cyclins to
promote BEC proliferation [75–77].
A novel transcription factor controlling cholangiocyte prolif-
eration (as well as liver size) is YAP, a transcription factor
belonging to the Hippo pathway. YAP is present in a phospho-
rylated and inactive state in the cytosol of cholangiocytes. Once
dephosphorylated by LATS1/2 in response to cytoskeletal dis-
tress or other stimuli, it translocates into the nucleus and stimu-
lates the transcription of several downstream effectors, such as
CTGF, leading to the activation of mitogen‐activated protein
kinase (MAPK) signaling that sustains cell proliferation [78].
Notably, perturbation of YAP‐mediated signaling leads to
­biliary paucity in mouse embryos at E18.5, whilst its hyperacti-
vation, using mutant YAP constitutively expressed into the
nucleus, stimulates an uncontrolled liver growth following
­partial hepatectomy in mice [26].
CHOLANGIOCYTES AND IMMUNITY
Cholangiocytes are a first line of defense in liver innate immu-
nity and can present the antigen to immune cells, can be a target
of immune‐mediated aggression, or be the initiators of an
inflammatory reaction that then progresses to adaptive immune
activation. Contribution of biliary epithelial cells to liver
immune responses was believed to be limited to the secretion of
400 THE LIVER:  CHOLANGIOCYTES AND IMMUNITY
immunoglobulin (Ig) A into the bile [79–80], but it is now clear
that the role of cholangiocytes in the immune response is far
more complex [81, 82]. In the liver, IgAs are synthesized by
plasma cells located along the biliary tree, bind to polymeric
immunoglobulin receptor (pIgR) present on the surface of the
basolateral membrane of BEC, and then are transported into
the cytoplasm of cholangiocytes where they are secreted into the
bile by vesicular transport [83].
The biliary epithelium stands as a first line of defense against
bacteria, fungi, and other pathogens by secreting antimicrobial
peptides, such as defensin and cathelicidin; normal cholangio-
cytes, in fact, express high concentration of β‐defensins hBD1
and hBD3. hBD1 is particularly effective in exerting an antimi-
crobial effect against Gram negative bacteria and Pseudomonas
aeruginosa, whereas hBD3 have a pronounced activity against
Helicobacter pylori and Staphylococcus aureus [84]. In con-
trast, hBD2 is induced by an inflammatory status or by exposure
to inflammatory peptides such as TNFα and IL1β, or to infec-
tious agents, like Cryptosporidium parvum, through a TLR2/
TLR4‐dependent NF‐κB activation [85].
A major role in epithelial innate immunity in cholangiocytes
is played by toll‐like receptors (TLRs) [85], and by nuclear
receptors (NR) [86]. TLRs can recognize pathogen‐associated
molecular patterns (PAMPs), namely bacterial structural ele-
ments such as LPS, DNA, RNA fragments, and flagellin but
also respond to endogenous components or damage‐associated
molecular patterns (DAMPs), such as hyaluronan and HMGB1
[87–89] that are released from damaged cells.
The TLRs family is composed by nine (1 to 9) transmem-
brane receptors with an intracellular domain, Toll‐IL1 receptor
(TIR) that can bind different adaptors, such as MyD88, Mal,
TRIF, and TRAM and thereby activate the NF‐κB signaling
which is then responsible for the secretion of a number of
inflammatory peptides [87–89]. Normal cholangiocytes consti-
tutively express TLR2, 3, 4, 5, and 9 [90]: TLR2 is a receptor for
the lipoteichoic acid (LTA), a component of the bacterial mem-
brane, TLR3 senses viral dsRNA, TLR4 is the main sensor for
LPS and DAMPS, and TLR9 is a receptor for CpG DNA, a hall-
mark of bacterial invasion.
TLR4‐mediated signaling is better known and studied in
cholangiocytes; once activated by LPS or other ligands, TLR4
activates two different pathways, one mediated by NF‐κB and
an alternative one via MAPK/AP‐1. The first one stimulates the
expression of a number of proinflammatory cyto‐ and
chemokines including TNFα, IL1, IL6, IL8, IL12, G‐CSF, and
LIX [89, 91]. The second pathway requires the nuclearization of
the AP‐1 complex [92].
In normal cholangiocytes, TLR4 signaling is repressed by
­protective mechanisms intended to maintain a level of “LPS toler-
ance” such as the expression of negative regulators (i.e. interleu-
kin‐1 receptor‐associated kinase M (IRAK‐M)) and PPARγ or by
post‐translational regulation of the receptor (i.e. Src mediated
tyrosine phosphorylation) (see below). Also, upregulation of
miR‐146 in response to LPS has been described as a mechanism
that partially controls the activation of TLR4‐induced NF‐κB‐
mediated cytokine secretion [93]. Since the biliary epithelium is
continuously in contact with bacterial products coming from the
intestine, changes in one or more regulatory checkpoints may
elicit an exaggerated inflammatory response in the liver.
The biliary epithelium is often a target for inflammation or
immune‐mediated injury. In inflammatory cholangiopathies,
as  discussed above, innate and adaptive immune responses
are sequentially or simultaneously involved. Innate immune
responses can be triggered by exogenous or endogenous insults
to cholangiocytes or nearby hepatocytes and if the inflammatory
reaction is protracted, it can be perpetuated through adaptive
immune mechanisms. A number of studies have shown that cell
dysfunction caused by genetic defects, causing cellular or tissue
malfunction, can generate a chronic inflammation of low mag-
nitude, defined as “parainflammation”, that is an adaptive
response to a persistent cell dysfunction shifting the homeo-
static set points aimed at restoring a normal cell/tissue homeo-
stasis [94]. As proposed by Medzihitov, parainflammation in
spite of its homeostatic nature may become maladaptive and
may stimulate a fibrotic response. In this scenario, epithelial
cells secrete a number of molecules and factors able to instruct
immune cells or to generate an inflammatory reaction (the
epi‐immunome).
An example is CFLD, where a decrease in the LPS tolerance
plays a major role in the development of the disease. Treatment
of mice with dextran sulfate sodium (DSS), a chemical that, by
inducing colitis, increases intestinal permeability and transloca-
tion of gut endotoxins to the liver, causes a peribiliary inflamma-
tion with infiltration of CD45+ cells, mainly neutrophils and
macrophages that damage the biliary epithelium and promotes
DR in CFTR‐KO mice but not in their wild‐type (WT) litter-
mates. Cholangiocytes isolated from CFTR‐KO mice and
exposed to LPS show an aberrant activation of TLR4/NF‐κB
signaling that leads to the increased secretion of proinflamma-
tory cytokines such as MCP‐1, G‐CSF, MIP‐2, KC, and lipopol-
ysaccharide‐induced CXC chemokine (LIX) [95]. As discussed,
in normal conditions CFTR acts as a docking protein that main-
tains proteins (i.e. Cbp, Csk) involved in the negative regulation
of Src tyrosine kinase at the membrane. CFTR mutations pre-
venting the expression of CFTR at the apical membrane cause
the disaggregation of the complex, and the activation of the
kinase that phosphorylates TLR4 and increases its response to
LPS. Notably, the treatment of CF‐KO mice exposed to DSS
with PP2, an inhibitor of Src family kinases, reduces the accu-
mulation of CD45+ cells and the extent of DR. Similar mecha-
nisms were further confirmed in human cholangiocytes derived
from induced pluripotent stem cells (iPSC) of a cystic fibrosis
patient with the common mutation ΔF508 CFTR [96]. Human
CF cholangiocytes show an increased activation of Src, aberrant
activation of NF‐κB, sustained secretion of Mcp‐1 and IL‐8, and
cytoskeletal defects. Treatment with PP2 improves these patho-
logic features and remarkably, increases the efficacy of VX‐770
and VX‐809, two small compounds used in clinic to correct the
ΔF508 defect, to restore the secretory function [96] (Figure 32.3).
A group of factors whose involvement in the immune
response of cholangiocytes has been recognized only recently,
are the nuclear receptors (NRs) that bind the retinoic X receptor
(RXR). This superfamily is composed of several members,
among which are the glucocorticoid receptor (GR), the retinoic
acid receptor (RAR), the vitamin D receptor (VDR), the liver X
receptors (LXRs), and the peroxisome proliferator‐activated
receptors (PPARs). NRs control several cell functions including
cell proliferation and apoptosis, cell metabolism, cell–cell
Figure 32.3  Control of innate immune‐mediated inflammation in cholangiocytes is altered in cystic fibrosis cholangiocytes. Healthy cholangiocytes (A), once stimulated by pathogenic noxa, such
as bacterial degradation products (PAMPs), activate a TLR4‐mediated response that transiently allows the secretion of profibrotic and proinflammatory mediators. This response is downmodulated by
several factors, among which, miR‐146, that inhibits TLR4, and by PPARγ, that through its effector IκBα, hinders the nuclear translocation of the p65 subunit of the NF‐κB complex. Furthermore,
polarized normal cholangiocytes express CFTR at the apical membrane and CFTR is able to assemble a macromolecular complex with EBP‐50, Cbp, and Csk, that are able to prevent the tyrosine
kinase Src form phosphorylating TLR4. In CFLD (B), PAMPs (i.e. LPS, or LTA) and DAMPs (i.e. HMGB1) coming from the gut or released by nearby cells, stimulate the secretion of profibrotic and
proinflammatory cyto‐ and chemokines through the TLR4/NF‐κB axis. The absence of CFTR sustains and amplifies this response, allowing the aberrant phosphorylation of Src that, by binding TLR4,
continuously stimulates the secretion of fibroinflammatory mediators and of hBD2. Similarly to TLR4, TLR2 induces the disaggregation of p65, its nuclear import, and its transcriptional activity.
Abbreviations: BEC, biliary epithelial cells; CF, cystic fibrosis; TLR, toll‐like receptor; DAMPs, damage‐associated molecular patterns; PAMPs, pathogen‐associated molecular patterns; LPS,
lipopolysaccharides; LTA, lipoteichoic acid; hBD2, human β‐defensin 2.
402 THE LIVER:  PORTAL AND PERIBILIARY FIBROSIS
interaction, detoxification from bile acids (VDR, FXR), and bile
secretion (GR, FXR) [86]. Interest in their role in inflammation
and innate immunity is raised, particularly for the PPARs (i.e.
PPARα, PPARβ/δ, and PPARγ). Among the different isoforms,
PPARγ has been shown to negatively regulate TLR4‐NF‐κB
signaling in CF cholangiocytes. Activation of PPARγ by piogl-
itazone and rosiglitazone inhibits LPS‐induced activation of
NF‐κB and cytokine secretion in Cftr‐KO cholangiocytes and
reduces biliary damage and inflammation in Cftr‐KO mice
treated with DSS. PPARγ can inhibit the transcriptional activity
of NF‐κB in two different manners, that is, by inhibiting p65
nuclear entry directly binding the p65/p50 complex, or
indirectly, by stimulating IkBα expression that retain p65 in the
cytoplasm [97].
It is worthy to note that autoimmune responses may be con-
sequences of alterations of innate immunity. In fact, continuous
stress in the absence of correct modulation of the TLR‐mediated
responses could be the trigger for a chronic inflammatory or
autoimmune response. Similarly, PSC that was typically consid-
ered as a disease with autoimmune traits may in reality be an
autoinflammatory disease on the basis of its increased activation
of NF‐κB transcription factor and its exaggerated activation of
the Th17 response to pathogens [98]. Another necroinflamma-
tory disease characterized by an increased Th17 response is bil-
iary atresia (BA), a severe pediatric disease characterized by
fibro‐obliterative vanishing of the bile ducts and inflammation
whose etiopathogenesis is still unclear. In BA, the serum levels
of IL17a and IL23, two typical Th17 response‐related ILs are
augmented, together with the expression of the IL17 transcrip-
tion factor ROR‐γt, IL‐17α, IL‐1β, IL‐6 [99]. Together with the
Th17 response, in the case of Ross River virus (RRV) infection,
BA patients could mount a Th2 adaptive response, theoretically
able to respond to the viral infection, but also to cause BA.
Recent data have shown that IL‐33 secreted by cholangiocytes
in responses to RRV infection stimulate innate lymphoid cells
type 2 (ILC2) to secrete a range of ILs and growth factors such
as IL‐13, osteopontin, and TGFβ. IL‐13 and IL‐33 behave in a
paracrine loop exerting trophic effects on cholangiocytes and
increasing the extension of ductular reaction in the disease
[100, 101], while local release of TGFβ stimulates proliferation
and secretion of ECM proteins by myofibroblasts (MF) leading
to the obliteration of bile ducts.
In addition, presence of DAMPs and PAMPs could promote
cellular senescence. Cell senescence is a mechanism of irrevers-
ible cell arrest in G1 stage induced by different stimuli, in par-
ticular, by oncogenic stressors. The main causes responsible for
the onset of senescence on cells include DNA damage, in par-
ticular, but not exclusively, to the telomers, the activation of
mitogenic signals induced by the activation of oncogenes, epi-
genetic modifications, and expression of tumor suppressor
genes. All these signals lead to different physiological responses
generally leading to tumor suppression, but, in other cases could
also promote cancer development, induce a fibrosing response,
and mediate age‐related degenerative diseases. Regardless of
the stimulus, cell senescence is mediated by two major signal
pathways, p16INK4a/pRB and the p53/p21CIP1/WAF1 [102]. In par-
ticular, genomic or epigenomic damages could induce a persis-
tent DNA damage response (DDR) that leads, on one side, to the
direct activation of the p53/p21pathway, and on the other side to
the activation of the P38MAPK/PKC signaling leading to the
expression of p16/pRB. Once senescent, cells not only stop their
proliferation, but assume a senescence‐associated secretory
phenotype (SASP) characterized by the secretion of a plethora
of peptides with profibrogenic, proinflammatory, and tumori-
genic properties, indicating that senescence could not only act
as a barrier to tumor growth, but also paracrinally stimulate
the activation of aberrant reparative/regenerative responses
[103, 104].
Cholangiocytes can act as antigen presenting cells (APC).
Cholangiocytes normally express only low levels of human leu-
cocyte antigen (HLA) class I, while HLA class II and the mem-
brane co‐receptors CD80 and CD86 are undetectable. However,
when cholangiocytes are challenged in vitro with the proinflam-
matory cytokines IL1, IFNγ, and TNFα or infected by cytomeg-
alovirus, HLA class I are upregulated and HLA class II but not
CD80 and CD86 can be expressed de novo, allowing the presen-
tation of the antigens to CD4+ and CD8+ T cells. Interestingly,
contrary to in vitro cell cultures, human specimens from PBC
and PSC show the neo‐expression of CD86 by RDC indicating
that cholangiocytes can orchestrate the necroinflammatory fea-
tures of these two diseases [105]. More recently, two papers
show that cholangiocytes from BA display the expression of
MHC class I and II, CD40, the co‐stimulatory molecules B7‐1
and B7‐2 [106], and MHC class I‐like molecule CD1d [107].
PORTAL AND PERIBILIARY FIBROSIS
Liver fibrosis is the result of an aberrant reparative response to
liver injury, and the mechanism supporting the clinical and his-
topathological progression of chronic liver diseases regardless
of their etiology. Fibrosis is a complex and highly integrated
pathophysiologic process, which is achieved through the inter-
action of multiple cell types and various molecular factors, lead-
ing to an exaggerated and qualitatively altered deposition of
ECM proteins [108].
Understanding the fine mechanisms that regulate portal fibro-
genesis, is a key step in the effort to identify new therapeutic
targets for biliary diseases (Figure  32.4). In the context of a
chronic hepatic insult, the persistent necroinflammatory dam-
age determines the progressive loss of parenchyma by apoptosis
or necrosis and acts as a stimulus for fibrosis. Necrotic and
apoptotic cells release cytokines and chemokines and DAMPs
that induce the recruitment of inflammatory and mesenchymal
cells, which in turn secrete a wide range of profibrotic and pro-
inflammatory factors, including TGFβ, CTGF, and TNFα [108,
109]. In addition, the necrotic cells release genetic material and
debris directly into the intercellular space, where they activate
stellate cells (HSC) and portal fibroblasts that proliferate and
secrete ECM proteins [110]. More information on epithelial–
mesenchymal cross‐talk in biliary disease can be found in recent
reviews [111, 112].
In hepatic fibrogenesis, the main actor is TGFβ1, which is
secreted mainly by fibroblasts and macrophages and functions
to activate HSC, responsible for the deposition of about 80% of
the matrix proteins [109]. Under physiological conditions the
ECM is composed mainly of collagens (the most important of
 32:  Cholangiocyte Biology and Pathobiology 403

Figure 32.4  Mechanism of peribiliary fibrosis deposition and ECM modification in cholangiopathies. Following cholangiocyte damage, biliary
epithelial cells secrete a wide range of mediators stimulating, in a paracrine fashion, the proliferation, transdifferentiation, and activation of cells of
mesenchymal origin (i.e. HSC and PF) responsible for the deposition of fibrosis. Furthermore, cholangiocytes secrete chemokines involved in
immune cell recruitment. Biliary epithelial cells are also able to modify the structure of the hepatic scaffold by secreting a wide range of metallo-
proteinases and proteolytic enzymes. Abbreviations: DAMPs, damage‐associated molecular patterns; PAMPs, pathogen‐associated molecular pat-
terns; HSC, hepatic stellate cells; PF, portal fibroblast; MF, myofibroblast; Φ, macrophage.

which are types I, III, IV, and VI) and non‐collagen proteins
such as laminin, fibronectin, thrombospondin, and tenascin, and
of various classes of proteoglycans consisting of a protein core
covalently linked to glycosaminoglycan (GAG) such as chon-
droitin‐, dermatan‐ and keratan‐sulfate, heparin and heparan‐
sulfate, and hyaluronic acid (HA) [113]. In response to the
damage, quiescent HSCs transdifferentiate into MF, character-
ized by the neo‐expression of α‐actin smooth muscle (α‐SMA),
by sensitivity to fibrogenic, chemoattractant and mitogenic
stimuli, and by the ability to actively secrete ECM proteins
[109, 114], in particular laminin and fibrillar collagen (type I
and III). This causes an alteration of the normal composition of
the matrix, which is important for the different tissue functions,
including cell–matrix interactions and the deposition of growth
factors [113]. Following intoxication with CCl4 or DDC, com-
position of liver ECM components changes dramatically, simi-
lar to what is observed in human diseases. In particular collagens
I, IV, and V, and fibronectin were increased while content of
elastin decreased [115]. Human ARPKD/CHF/CD (see below)
also exhibits a derangement of the normal composition of ECM
due to an increased and aberrant secretion of fibrosis produced
by activated portal myofibroblasts, macrophages (mainly M2),
and cholangiocytes too [116, 117], and the erosion of the basal
membrane, composed of collagen IV, laminin, and reticulin,
surrounding the biliary epithelium [118]. These changes could
be persistent as in late fibrosis or cirrhosis, or could be transient
as in early stages of alcoholic liver disease [119]. In addition to
the ECM proteins, MFs are able to produce a broad spectrum of
proteins involved in matrix degradation, such as metallopro-
teases (MMPs) (in particular MMP‐2 and ‐9) and their inhibi-
tors (TIMPs), which with a finely regulated mechanism, lead to
continuous remodeling of the matrix during chronic hepatic
injury. In the early stages of hepatic injury, the MFs do not
express TIMPs, thus allowing the degradation of the normal
matrix that will be replaced by the altered one [120].
Cholangiocytes are able to secrete MMPs, in particular
MMP2, 3, 13, and MT1‐MMP and other enzymes involved in the
modification of the hepatic 3D scaffold, such as uPA; notably, in
a rodent model of fibropolycystic liver diseases, the treatment of
8‐week‐old PCK rats with marimastat, a broad spectrum MMP
404 THE LIVER:  REFERENCES
inhibitor, was able to reduce both cyst area and col1a1 expres-
sion [120]. At a certain point, the production of MMPs is blocked,
so the remodeling ceases to the advantage of the excessive depo-
sition. In addition, MF further release growth factors that main-
tain fibrogenesis, and also intervene in the activation of
neoangiogenesis. However, HSCs are not the only mesenchymal
cells capable of giving rise to MF: alternative sources exist, such
as activated portal fibroblasts, circulating mesenchymal precur-
sors of medullary origin (5–7%) and fibrocytes (4–6%) [114].
Whether epithelial–mesenchymal transition (EMT) [121]
plays a role in liver fibrosis similar to other organs like the kid-
ney has been hotly debated. Most evidence indicates that during
liver repair, cholangiocytes lose part of their epithelial charac-
teristics to acquire some biological and immunophenotypic
properties of mesenchymal cells, such as motility and the ability
to depose collagen, but there is not a transition to a full mesen-
chymal phenotype [122] and cholangiocytes simply acquire a
few functional properties that favor the repair of the wound.
CTGF is another growth factor capable of modulating the
transduction of extracellular signals by interacting with mem-
brane glycoproteins (integrins), growth factors (TGFβ1), and
with different components of the extracellular matrix (fibronec-
tin). Once secreted, CTGF can stimulate the proliferation, adhe-
sion, and recruitment of different cell types [123], including
macrophages; CTGF can play a fundamental role in supporting
the inflammatory response driven by macrophages. PBC and PSC
are characterized by increased CTGF secretion by different cell
milieu, among which are cholangiocytes, in which expression
sustains the deposition of ECM proteins and fibrosis. Recently, a
paper by Pi and colleagues [124] demonstrates that in the liver,
CTGF, in association with integrin αvβ6, is expressed by HPCs
and RDCs after experimental biliary damage, where they regulate
the activation of HPC and deposition of fibrosis, by interacting
with fibronectin and TGF‐β1. Notably, in a mouse model of CCl4
intoxication, the treatment with siRNA against CTGF was able to
revert fibrosis and reduce collagen deposition [125].
Among genetic cholangiopathies of particular interest are
fibropolycystic diseases, as they are characterized by exuberant
biliary fibrosis, but in the absence of cholangiocyte necrosis/
apoptosis. This is a group of monogenic diseases with autoso-
mal recessive transmission and variable phenotype that includes
the autosomal recessive disease of polycystic kidney (ARPKD),
congenital hepatic fibrosis (CHF), and Caroli’s disease (CD),
all due to the mutation, in different loci, of the same gene,
which encodes for fibrocystin (FPC). Fibropolycystic liver dis-
eases are characterized by ductal dysgenesia resulting in biliary
microhamartomas and segmental dilations, with a phenotype
retaining a fetal configuration (DP malformations). Biliary
dysgenesis is typically associated with progressive accumula-
tion of portal fibrosis eventually leading to portal hypertension.
Interestingly, these diseases are characterized by “parainflam-
mation”, a smoldering and unresolved inflammatory process
indicating a maladaptive response to persistent tissue dysfunc-
tion [94]. Increased nuclear expression of β‐catenin is respon-
sible for the secretion of several chemokines, among which are
CXCL1, CXCL10, and CXCL12. Chemokines secreted in the
hepatic microenvironment are responsible for the recruitment
of inflammatory cells around the cysts, and in particular of M1
and M2 macrophages. Macrophages locally secrete TGFβ and
TNFα that stimulate the de novo expression of αVβ6 integrin
on the surface of the biliary structures. This integrin is able to
activate the latent form of TGFβ by the cleavage of the latency‐
associated peptide (LAP). Notably, the treatment of FPC‐KO
mice with clodronate, a bisphosphonate that inhibits the mono-
cyte–macrophage differentiation, improves the main patho-
logic readouts of the disease, namely, cyst enlargement, fibrosis
accumulation, and development of portal hypertension [116].
Interestingly, the secretion of CXCL10 was induced not only
by a β‐catenin‐dependent mechanism but also by the secretion
of IL1β, thanks to the NF‐κB dependent activation of the
NLRP3 inflammasome complex [126]. Moreover, the treat-
ment of Pkhd1del4/del4 mice with AMG‐487 an inhibitor of the
receptor for CXCL10 is able to reduce the extent of pericystic
macrophages together with cystic areas and peribiliary fibrosis
in a similar manner to clodronate, demonstrating that mac-
rophage accumulation is one of the major drivers of the disease
in CHF [126].
CONCLUSIONS
In this chapter, we have summarized some of the recent
advances in the understanding of cholangiocyte biology. We
elected to focus our discussion on those mechanisms that have
relevance for understanding the pathophysiology of the most
relevant genetic or acquired cholangiopathies. We believe that
the reader will have a working understanding of the function
and dysfunction of this important epithelium of the liver and
to focus their research and interest on finding novel therapeutic
targets that are urgently needed. We expect that future research
into the pathophysiology of these fascinating diseases will be
rewarding.
ACKNOWLEDGMENTS
Fondazione Cariplo, grant number 2014–1099, to M. Cadamuro;
NIH grant DK‐034989 (Silvio O. Conte Digestive Diseases
Research Core Centers) to M. Strazzabosco; NIH grants DK‐079005
and DK‐096096, and PSC Partners Seeking a Cure and Connecticut
Innovations (16‐RMA‐YALE‐26) to M. Strazzabosco.
REFERENCES
1. Strazzabosco, M., Spirlí, C., and Okolicsanyi L. Pathophysiology of the intra-
hepatic biliary epithelium. J Gastroenterol Hepatol, 2000;15:244–53.
2. Cheung, A.C., Lorenzo Pisarello, M.J., and LaRusso, N.F. Pathobiology of
biliary epithelia. Biochim Biophys Acta, 2018;1864:1220–1231.
3. Crawford, A.R., Lin, X.Z., and Crawford, J.M. The normal adult human liver
biopsy: a quantitative reference standard. Hepatology, 1998;28:323–31.
4. Roskams, T.A., Theise, N.D., Balabaud, C. et al. Nomenclature of the finer
branches of the biliary tree: canals, ductules, and ductular reactions in human
livers. Hepatology, 2004;39:1739–45.
5. Lemaigre, F.P. Molecular mechanisms of biliary development. Prog Mol Biol
Transl Sci, 2010;97:103–26.
6. Ober, E.A. and Lemaigre, F.P. Development of the liver: insights into organ
and tissue morphogenesis. J Hepatol, 2018;68:1049–62.
 32:  Cholangiocyte Biology and Pathobiology 405
7. Lazaridis, K.N., Strazzabosco, M., and Larusso N.F. The cholangiopathies:
disorders of biliary epithelia. Gastroenterol, 2004;127:1565–77.
8. Strazzabosco, M., Fabris, L., and Spirli, C. Pathophysiology of cholangiopa-
thies. J Clin Gastroenterol, 2005;39:S90–102.
9. Raynaud, P., Carpentier, R., Antoniou, A., and Lemaigre, F.P. Biliary differentia-
tion and bile duct morphogenesis in development and disease. Int J Biochem
Cell Biol, 2011;43:245–56.
10. Gordillo, M., Evans, T., and Gouon‐Evans V. Orchestrating liver develop-
ment. Development, 2015;142:2094–108.
11. Libbrecht, L., Spinner, N.B., Moore, E.C., Cassiman, D., Van Damme‐
Lombaerts, R., and Roskams, T. Peripheral bile duct paucity and cholestasis
in the liver of a patient with Alagille syndrome: further evidence supporting
a lack of postnatal bile duct branching and elongation. Am J Surg Pathol,
2005;29:820–6.
12. Carpentier, R., Suñer, R.E., van Hul, N. et al. Embryonic ductal plate cells
give rise to cholangiocytes, periportal hepatocytes, and adult liver progenitor
cells. Gastroenterol, 2011;141:1432–8.
13. Fabris, L., Cadamuro, M., Libbrecht, L. et al. Epithelial expression of angio-
genic growth factors modulate arterial vasculogenesis in human liver devel-
opment. Hepatology, 2008;47:719–28.
14. Fabris, L., Cadamuro, M., Fiorotto, R. et  al. Effects of angiogenic factor
overexpression by human and rodent cholangiocytes in polycystic liver diseases.
Hepatology, 2006;43:1001–12.
15. Lemaigre F.P. Mechanisms of liver development: concepts for understanding
liver disorders and design of novel therapies. Gastroenterol, 2009;137:62–79.
16. Nakanuma, Y., Harada, K., Sato, Y., and Ikeda, H. Recent progress in the
etiopathogenesis of pediatric biliary disease, particularly Caroli’s disease
with congenital hepatic fibrosis and biliary atresia. Histol Histopathol,
2010;25:223–35.
17. Desmet, V.J. Ductal plates in hepatic ductular reactions. Hypothesis and
implications. II. Ontogenic liver growth in childhood. Virchows Arch,
2011;458:261–70.
18. Desmet, V.J. Ductal plates in hepatic ductular reactions. Hypothesis
and implications. III. Implications for liver pathology. Virchows Arch,
2011;458:271–9.
19. Simons, M., Gloy, J., Ganner, A. et al. Inversin, the gene product mutated in
nephronophthisis type, I.I., functions as a molecular switch between Wnt
signaling pathways. Nat Genet, 2005;37:537–43.
20. Lozier, J., McCright, B., and Gridley, T. Notch signaling regulates bile duct
morphogenesis in mice. PLoS One, 2008;3:e1851.
21. Zong, Y., Panikkar, A., Xu, J. et al. Notch signaling controls liver development
by regulating biliary differentiation. Development, 2009;136:1727–39.
22. Fabris, L., Cadamuro, M., Guido, M. et al. Analysis of liver repair mecha-
nisms in Alagille syndrome and biliary atresia reveals a role for notch signal-
ing. Am J Pathol, 2007;171:641–53.
23. Omenetti, A. and Diehl A.M. The adventures of sonic hedgehog in develop-
ment and repair. II. Sonic hedgehog and liver development, inflammation,
and cancer. Am J Physiol Gastrointest Liver Physiol, 2008;294:G595–8.
24. Choi, S.S., Omenetti, A., Syn, W.K., and Diehl A.M. The role of Hedgehog
signaling in fibrogenic liver repair. Int J Biochem Cell Biol, 2011;43:238–44.
25. Machado, M.V. and Diehl A.M. Hedgehog signalling in liver pathophysiol-
ogy. J Hepatol, 2018;68:550–62.
26. Yi, J., Lu, L., Yanger, K. et al. Large tumor suppressor homologs 1 and 2 regu-
late mouse liver progenitor cell proliferation and maturation through antago-
nism of the coactivators YAP and TAZ. Hepatology, 2016;64:1757–72.
27. Swiderska‐Syn, M., Xie, G., Michelotti, G.A. et al. Hedgehog regulates yes‐
associated protein 1 in regenerating mouse liver. Hepatology, 2016;
64:232–44.
28. Poncy, A., Antoniou, A., Cordi, S., Pierreux, C.E., Jacquemin, P., and
Lemaigre F.P. Transcription factors SOX4 and SOX9 cooperatively control
development of bile ducts. Dev Biol, 2015;404:136–48.
29. Mansini, A.P., Peixoto, E., Thelen, K.M., Gaspari, C., Jin, S., and Gradilone,
S.A. The cholangiocyte primary cilium in health and disease. Biochim
Biophys Acta, 2018;1864:1245–53.
30. Rogler, C.E., Matarlo, J.S., Kosmyna, B., Fulop, D., and Rogler, L.E.
Knockdown of miR‐23, miR‐27, and miR‐24 alters fetal liver development
and blocks fibrosis in mice. Gene Expr, 2017;17:99–114.
31. Rogler, C.E., Levoci, L., Ader, T., Massimi, A., Tchaikovskaya, T., Norel, R.,
and Rogler L.E. MicroRNA‐23b cluster microRNAs regulate transforming
growth factor‐beta/bone morphogenetic protein signaling and liver stem cell
differentiation by targeting Smads. Hepatology, 2009;50:575–84.
32. Kanno, N., LeSage, G., Glaser, S., Alvaro, D., and Alpini, G. Functional
heterogeneity of the intrahepatic biliary epithelium. Hepatology, 2000;
31:555–61.
33. Marzioni, M., Glaser, S.S., Francis, H., Phinizy, J.L., LeSage, G., and Alpini,
G. Functional heterogeneity of cholangiocytes. Semin Liver Dis,
2002;22:227–40.
34. Glaser, S., Francis, H., Demorrow, S. et al. Heterogeneity of the intrahepatic
biliary epithelium. World J Gastroenterol, 2006;12:3523–36.
35. Strazzabosco, M. and Fabris, L. Functional anatomy of normal bile ducts.
Anat Rec, 2008;291:653–60.
36. Han, Y., Glaser, S., Meng, F. et al. Recent advances in the morphological and func-
tional heterogeneity of the biliary epithelium. Exp Biol Med, 2013;238:549–65.
37. Boyer J.L. Bile formation and secretion. Compr Physiol, 2013;3:1035–78.
38. Nathanson, M.H., Burgstahler, A.D., Mennone, A., and Boyer J.L.
Characterization of cytosolic Ca2+ signaling in rat bile duct epithelia. Am J
Physiol, 1996;271:G86–96.
39. Zsembery, A., Spirlì, C., Granato, A., et al. Purinergic regulation of acid/base trans-
port in human and rat biliary epithelial cell lines. Hepatology, 1998;28:914–20.
40. Caligiuri, A., Glaser, S., Rodgers, R.E. et al. Endothelin‐1 inhibits secretin‐
stimulated ductal secretion by interacting with ETA receptors on large chol-
angiocytes. Am J Physiol, 1998;275:G835–46.
41. Gong, A.Y., Tietz, P.S., Muff, M.A. et al. Somatostatin stimulates ductal bile
absorption and inhibits ductal bile secretion in mice via SSTR2 on cholan-
giocytes. Am J Physiol Cell Physiol, 2003;284:C1205–14.
42. Mennone, A., Biemesderfer, D., Negoianu, D. et al. Role of sodium/hydro-
gen exchanger isoform NHE3 in fluid secretion and absorption in mouse and
rat cholangiocytes. Am J Physiol Gastrointest Liver Physiol, 2001;280:
G247–54.
43. Elborn, J.S. Cystic fibrosis. Lancet, 2016;388:2519–31.
44. Beuers, U., Hohenester, S., de Buy Wenniger, L.J., Kremer, A.E., Jansen,
P.L., and Elferink, R.P. The biliary HCO(3)(‐) umbrella: a unifying hypothesis
on pathogenetic and therapeutic aspects of fibrosing cholangiopathies.
Hepatology, 2010;52:1489–96.
45. Hohenester, S., Maillette de Buy Wenniger, L., Jefferson, D.M., Oude Elferink,
R.P., and Beuers, U. Biliary bicarbonate secretion constitutes a protective
mechanism against bile acid‐induced injury in man. Dig Dis, 2011;29:62–5.
46. Beuers, U., Maroni, L., and Elferink, R.O. The biliary HCO(3)(‐) umbrella:
experimental evidence revisited. Curr Opin Gastroenterol, 2012;28:253–7.
47. Hohenester, S., Wenniger, L.M., Paulusma, C.C. et  al. A biliary HCO3‐
umbrella constitutes a protective mechanism against bile acid‐induced injury
in human cholangiocytes. Hepatology, 2012;55:173–83.
48. Beuers, U., Trauner, M., Jansen, P., and Poupon, R. New paradigms in the
treatment of hepatic cholestasis: from UDCA to FXR., PXR and beyond.
J Hepatol, 2015;62:S25–37.
49. Ainsworth, C. Cilia: tails of the unexpected. Nature, 2007;448:638–41.
50. Bisgrove, B.W. and Yost H.J. The roles of cilia in developmental disorders
and disease. Development, 2006;133:4131–43.
51. Strazzabosco, M. and Fabris, L. Development of the bile ducts: essentials for
the clinical hepatologist. J Hepatol, 2012;56:1159–70.
52. Clotman, F., Libbrecht, L., Killingsworth, M.C., Loo, C.C., Roskams, T., and
Lemaigre, F.P. Lack of cilia and differentiation defects in the liver of human
foetuses with the Meckel syndrome. Liver Int, 2008;28:377–84.
53. Kagan, K.O., Dufke, A., and Gembruch, U. Renal cystic disease and associated
ciliopathies. Curr Opin Obstet Gynecol, 2017;29:85–94.
54. Schmidts M. Clinical genetics and pathobiology of ciliary chondrodyspla-
sias. J Pediatr Genet, 2014;3,46–94.
55. Lehman, J.M., Michaud, E.J., Schoeb, T.R., Aydin‐Son, Y., Miller, M., and
Yoder, B.K. The Oak Ridge polycystic kidney mouse: modeling ciliopathies
of mice and men. Dev Dyn, 2008;237:1960–71.
56. Sambrotta, M., Strautnieks, S., Papouli, E. et al. Mutations in TJP2 cause
progressive cholestatic liver disease. Nat Genet, 2014;46:326–28.
57. Lee, S.H. Intestinal permeability regulation by tight junction: implication on
inflammatory bowel diseases. Intest Res, 2015;13:11–8.
58. Fiorotto, R., Villani, A., Kourtidis, A. et al. The cystic fibrosis transmembrane
conductance regulator controls biliary epithelial inflammation and permeabil-
ity by regulating Src tyrosine kinase activity. Hepatology, 2016;64:2118–34.
59. Fiorotto, R., Scirpo, R., Trauner, M. et  al. Loss of CFTR affects biliary
epithelium innate immunity and causes TLR4‐NF‐κB‐mediated inflamma-
tory response in mice. Gastroenterol, 2011;141:1498–508.
60. Fabris, L. and Strazzabosco, M. Epithelial‐mesenchymal interactions in
biliary diseases. Semin Liver Dis, 2011;31:11–32.
406 THE LIVER:  REFERENCES
61. Fabris, L., Strazzabosco, M., Crosby, H.A. et al. Characterization and isola-
tion of ductular cells coexpressing neural cell adhesion molecule and Bcl‐2
from primary cholangiopathies and ductal plate malformations. Am J Pathol,
2000;156:1599–612.
62. Clouston, A.D., Powell, E.E., Walsh, M.J., Richardson, M.M., Demetris,
A.J., and Jonsson J.R. Fibrosis correlates with a ductular reaction in hepatitis
C: roles of impaired replication, progenitor cells and steatosis. Hepatology,
2005;41:809–18.
63. Richardson, M.M., Jonsson, J.R., Powell, E.E. et al. Progressive fibrosis in
non‐alcoholic steatohepatitis: association with altered regeneration and a
ductular reaction. Gastroenterol, 2007;133:80–90.
64. Desmet, V.J. and Roskams, T. Cirrhosis reversal: a duel between dogma and
myth. J Hepatol, 2004;40:860–7.
65. de Jong, I.E.M., van Leeuwen, O.B., Lisman, T., Gouw, A.S.H., and Porte,
R.J. Repopulating the biliary tree from the peribiliary glands. Biochim
Biophys Acta, 2018;1864:1524–31.
66. Cardinale, V., Wang, Y., Carpino, G. et al. Multipotent stem/progenitor cells
in human biliary tree give rise to hepatocytes, cholangiocytes, and pancreatic
islets. Hepatology, 2011;54:2159–72.
67. Carpino, G., Cardinale, V., Renzi, A. et  al. Activation of biliary tree stem
cells within peribiliary glands in primary sclerosing cholangitis. J Hepatol,
2015;63:1220–8.
68. Meng, F., Yamagiwa, Y., Ueno, Y., and Patel, T. Over‐expression of interleu-
kin‐6 enhances cell survival and transformed cell growth in human malig-
nant cholangiocytes. J Hepatol, 2006;44:1055–65.
69. Znoyko, I., Sohara, N., Spicer, S.S., Trojanowska, M., and Reuben, A.
Expression of oncostatin M and its receptors in normal and cirrhotic human
liver. J Hepatol,2005;43:893–900.
70. Joplin, R., Hishida, T., Tsubouchi, H et  al. Human intrahepatic biliary
epithelial cells proliferate in vitro in response to human hepatocyte growth
factor. J Clin Invest, 1992;90:1284–9.
71. Alvaro, D., Metalli, V.D., Alpini, G., et al. The intrahepatic biliary epithe-
lium is a target of the growth hormone/insulin‐like growth factor 1 axis. J
Hepatol, 2005;43:875–83.
72. Glaser, S.S., Gaudio, E., Miller, T., Alvaro, D., and Alpini, G. Cholangiocyte
proliferation and liver fibrosis. Expert Rev Mol Med, 2009;11:e7.
73. Hall, C., Sato, K., Wu, N. et al. Regulators of cholangiocyte proliferation.
Gene Expr, 2017;17:155–71.
74. Strazzabosco, M. and Somlo, S. Polycystic liver diseases: congenital disorders
of cholangiocyte signaling. Gastroenterol, 2011;140:1855–9.
75. Gatto, M., Drudi‐Metalli, V., Torrice, A. et al. Insulin‐like growth factor‐1
isoforms in rat hepatocytes and cholangiocytes and their involvement in pro-
tection against cholestatic injury. Lab Invest, 2008;88:986–94.
76. Spirli, C., Okolicsanyi, S., Fiorotto, R. et  al. ERK1/2‐dependent vascular
endothelial growth factor signaling sustains cyst growth in polycystin‐2
defective mice. Gastroenterol, 2010;138:360–71.
77. Spirli, C., Okolicsanyi, S., Fiorotto, R. et al. Mammalian target of rapamycin
regulates vascular endothelial growth factor‐dependent liver cyst growth in
polycystin‐2‐defective mice. Hepatology, 2010;51:1778–88.
78. Piccolo, S., Dupont, S., and Cordenonsi, M. The biology of YAP/TAZ: hippo
signaling and beyond. Physiol Rev, 2014;94:1287–312.
79. Nagura, H., Smith, P.D., Nakane, P.K., and Brown, W.R. IGA in human bile
and liver. J Immunol, 1981;126:587–95.
80. Sugiura, H. and Nakanuma, Y. Secretory component and immunoglobulins
in the intrahepatic biliary tree and peribiliary gland in normal livers and
hepatolithiasis. Gastroenterol Jpn, 1989;24:308–14.
81. Chuang, Y.H., Lan, R.Y., and Gershwin, M.E. The immunopathology
of human biliary cell epithelium. Semin Immunopathol, 2009;31:
323–31.
82. Syal, G., Fausther, M., Dranoff J.A. Advances in cholangiocyte immunobiol-
ogy. Am J Physiol Gastrointest Liver Physiol, 2012;303:G1077–86.
83. Wu, C.T., Davis, P.A., Luketic, V.A., and Gershwin, M.E. A review of the
physiological and immunological functions of biliary epithelial cells: targets
for primary biliary cholangitis, primary sclerosing cholangitis and drug‐
induced ductopenias. Clin Dev Immunol, 2004;11:205–13.
84. Menendez, A. and Brett Finlay, B. Defensins in the immunology of bacterial
infections. Curr Opin Immunol, 2007;19:385–91.
85. Chen, X.M., O’Hara, S.P., Nelson, J.B. et al. Multiple TLRs are expressed in
human cholangiocytes and mediate host epithelial defense responses to
Cryptosporidium parvum via activation of NF‐kappaB. J Immunol,
2005;175:7447–56.
86. Firrincieli, D., Zuniga, S., Poupon, R., Housset, C., and Chignard, N. Role
of nuclear receptors in the biliary epithelium. Dig Dis, 2011;29:52–7.
87. Barton, G.M. and Medzhitov, R. Toll‐like receptors and their ligands. Curr
Top Microbiol Immunol, 2002; 270:81–92.
88. Iwasaki, A. and Medzhitov, R. Toll‐like receptor control of the adaptive
immune responses. Nat Immunol, 2004; 5:987–995.
89. Seki, E. and Brenner, D.A. Toll‐like receptors and adaptor molecules in
liver disease: update. Hepatol, 2008; 48:322–35.
90. Di Gioia, M. and Zanoni, I. Toll‐like receptor co‐receptors as master regula-
tors of the immune response. Mol Immunol, 2015;63:143–52.
91. Akira, S., Takeda, K., and Kaisho T. Toll‐like receptors: critical proteins
linking innate and acquired immunity. Nat Immunol, 2001;2:675–80.
92. Whitmarsh, A.J. and Davis, R.J. Transcription factor AP‐1 regulation by
mitogen‐activated protein kinase signal transduction pathways. J Mol Med
(Berl), 1996;74:589–607.
93. Soares, J.B., Pimentel‐Nunes, P., Roncon‐Albuquerque, R., and Leite‐
Moreira, A. The role of lipopolysaccharide/toll‐like receptor 4 signaling in
chronic liver diseases. Hepatol Int. 2010;4:659–72.
94. Medzhitov, R. Origin and physiological roles of inflammation. Nature,
2008;454, 428–35.
95. Scirpo, R., Fiorotto, R., Villani, A., Amenduni, M., Spirli, C., and
Strazzabosco, M. Stimulation of nuclear receptor peroxisome proliferator‐
activated receptor‐γ limits NF‐κB‐dependent inflammation in mouse cystic
fibrosis biliary epithelium. Hepatol, 2015;62:1551–62.
96. Fiorotto, R., Amenduni, M., Mariotti, V., Fabris, L., Spirli, C., and
Strazzabosco, M. Src kinase inhibition reduces inflammatory and cytoskel-
etal changes in ΔF508 human cholangiocytes and improves cystic fibrosis
transmembrane conductance regulator correctors efficacy. Hepatology,
2018;67:972–88.
97. Baeuerle, P.A. and Baltimore, D. NF‐kappa B: ten years after. Cell,
1996;87:13–20.
98. Katt, J., Schwinge, D., Schoknecht, T. et  al. Increased T helper type 17
response to pathogen stimulation in patients with primary sclerosing chol-
angitis. Hepatology, 2013;58:1084–93.
99. Yang, Y., Liu, Y.J., Tang, S.T. et al. Elevated Th17 cells accompanied by
decreased regulatory T cells and cytokine environment in infants with bil-
iary atresia. Pediatr Surg Int, 2013;29:1249–60.
100. Li, J., Razumilava, N., Gores, G.J. et al. Biliary repair and carcinogenesis
are mediated by IL‐33‐dependent cholangiocyte proliferation. J Clin Invest,
2014;124:3241–51.
101. Gieseck, R.L. 3rd, Ramalingam, T.R., Hart, K.M. et al. Interleukin‐13 acti-
vates distinct cellular pathways leading to ductular reaction, steatosis, and
fibrosis. Immunity, 2016;45:145–58.
102. d’Adda di Fagagna, F. Living on a break: cellular senescence as a DNA‐
damage response. Nat Rev Cancer, 2008;8:512–22.
103. Muñoz‐Espín, D. and Serrano, M. Cellular senescence: from physiology to
pathology. Nat Rev Mol Cell Biol, 2014;15:482–96.
104. Strazzabosco, M., Fiorotto, R., Cadamuro, M. et  al. Pathophysiologic
implications of innate immunity and autoinflammation in the biliary epithe-
lium. Biochim Biophys Acta, 2018;1864:1374–9.
105. Tsuneyama, K., Yasoshima, M., Harada, K., Hiramatsu, K., Gershwin,
M.E., and Nakanuma, Y. Increased CD1d expression on small bile duct
epithelium and epithelioid granuloma in livers in primary biliary cholangi-
tis. Hepatology, 1998;28:620–3.
106. Barnes, B.H., Tucker, R.M., Wehrmann, F., Mack, D.G., Ueno, Y., and
Mack C.L. Cholangiocytes as immune modulators in rotavirus‐induced
murine biliary atresia. Liver Int, 2009;29:1253–61.
107. Schrumpf, E., Tan, C., Karlsen, T.H. et al. The biliary epithelium presents anti-
gens to and activates natural killer T cells. Hepatology, 2015;62:1249–59.
108. Weiskirchen, R., Weiskirchen, S., and Tacke, F. Organ and tissue fibrosis:
molecular signals, cellular mechanisms and translational implications. Mol
Aspects Med, 2018;65:2–15.
109. Kisseleva, T. and Brenner, D.A. Mechanisms of fibrogenesis. Exp Biol Med
(Maywood), 2008;233:109–22.
110. Mehal, W. and Imaeda, A. Cell death and fibrogenesis. Semin Liver Dis,
2010;30:226–31.
111. Gressner, O.A. and Gao, C. Monitoring fibrogenic progression in the liver.
Clin Chim Acta, 2014;10;433:111–22.
112. Fabris, L., Spirli, C., Cadamuro, M., Fiorotto, R., and Strazzabosco, M.
Emerging concepts in biliary repair and fibrosis. Am J Physiol Gastrointest
Liver Physiol, 2017;313:G102–16.
 32:  Cholangiocyte Biology and Pathobiology 407
113. Batzios, S.P., Zafeiriou, D.I., and Papakonstantinou, E. Extracellular matrix
components: an intricate network of possible biomarkers for lysosomal
storage disorders? FEBS Lett, 2013;587:1258–67.
114. Mallat, A. and Lotersztajn, S. Cellular mechanisms of tissue fibrosis. 5. Novel
insights into liver fibrosis. Am J Physiol Cell Physiol, 2013;305:C789–99.
115. Klaas, M., Kangur, T., Viil. J. et al. The alterations in the extracellular matrix
composition guide the repair of damaged liver tissue. Sci Rep, 2016;6:27398.
116. Locatelli, L., Cadamuro, M., Spirlì, C. et al. Macrophage recruitment by
fibrocystin‐defective biliary epithelial cells promotes portal fibrosis in con-
genital hepatic fibrosis. Hepatology, 2016;63:965–82.
117. Sato, Y., Harada, K., Ozaki, S. et  al. Cholangiocytes with mesenchymal
features contribute to progressive hepatic fibrosis of the polycystic kidney
rat. Am J Pathol, 2007;171:1859–71.
118. Yasoshima, M., Sato, Y., Furubo, S. et  al. Matrix proteins of basement
membrane of intrahepatic bile ducts are degraded in congenital hepatic
fibrosis and Caroli’s disease. J Pathol, 2009;217:442–51.
119. Poole, L.G. and Arteel, G.E. Transitional remodeling of the hepatic extracellu-
lar matrix in alcohol‐induced liver injury. Biomed Res Int, 2016;2016:3162670.
120. Urribarri, A.D., Munoz‐Garrido, P., Perugorria, M.J. et  al. Inhibition of
metalloprotease hyperactivity in cystic cholangiocytes halts the develop-
ment of polycystic liver diseases. Gut, 2014;63:1658–67.
121. Zeisberg, M., Yang, C., Martino, M. et al. Fibroblasts derive from hepato-
cytes in liver fibrosis via epithelial to mesenchymal transition. J Biol Chem,
2007;282:23337–47.
122. Kisseleva, T. and Brenner, D.A. Is it the end of the line for the EMT?
Hepatology, 2011;53:1433–5.
123. Charrier, A., Chen, R., Kemper, S., and Brigstock, D.R. Regulation of pan-
creatic inflammation by connective tissue growth factor (CTGF/CCN2).
Immunology, 2014;141:564–76.
124. Pi, L., Robinson, P.M., Jorgensen, M. et al. Connective tissue growth factor
and integrin αvβ6: a new pair of regulators critical for ductular reaction and
biliary fibrosis in mice. Hepatology, 2015;61:678–91.
125. Li, G., Xie, Q., Shi, Y. et al. Inhibition of connective tissue growth factor by
siRNA prevents liver fibrosis in rats. J Gene Med, 2006;8:889–900.
126. Kaffe, E., Fiorotto, R., Pellegrino, F. et al. β‐Catenin and interleukin‐1β‐
dependent chemokine (C‐X‐C motif) ligand 10 production drives progres-
sion of disease in a mouse model of congenital hepatic fibrosis.
Hepatology, 2018;67:1903–19.
127. Strazzabosco, M., Sakisaka, S., Hayakawa, T., and Boyer, J.L. Effect of
UDCA on intracellular and biliary pH in isolated rat hepatocyte couplets
and perfused livers. Am J Physiol, 1991;260(1):G58–69.
 Polycystic Liver Diseases:
33 Genetics, Mechanisms,
and Therapies
Tatyana Masyuk, Anatoliy Masyuk, and Nicholas LaRusso
Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, MN, USA
INTRODUCTION
Polycystic liver disease (PLD), a genetic cholangiociliopathy,
is characterized by the presence of multiple liver cysts of dif­
ferent shape and size filled with cystic fluid [1]. PLD exists as
an extra‐renal manifestation of autosomal dominant polycys­
tic kidney disease (ADPKD) and autosomal recessive PKD
(ARPKD) or as isolated autosomal dominant PLD (ADPLD).
In PLD associated with ADPKD and ARPKD liver cysts
coexist with kidney cysts while in isolated ADPLD the majority
of the cysts are localized to the liver with very few or no cysts
in the kidney. Clinically, PLD is arbitrarily defined by the
presence of more than 20 liver cysts as measured by ultra­
sonography, computed tomography, or magnetic resonance
imaging [2]. Liver volume and number of cysts in patients
with isolated ADPLD is generally larger than in patients with
ADPKD‐associated PLD [2].
Familial studies indicate that around 20% of patients with
mutations in PLD‐related genes have mild or no disease [2]. In
the remaining 80% of patients, progressive enlargement of liver
volume results in PLD‐related symptoms caused by liver com­
pression of adjacent organs. Based on the volume of the liver,
PLD is classified as mild (height‐corrected liver volume less
than 1600 mL), moderate (height‐corrected liver volume
between 1600–3200 mL), and severe disease (height‐corrected
liver volume more than 3200 mL) [3]. Severity of disease
­correlates with symptom burden and declined quality of life.
Patients with mild PLD seldomly develop complications.
Patients with moderate and especially with severe disease
­experience pressure‐related symptoms such as pain, abdominal
distention and bloating, gastroesophageal reflux, nausea, and
dyspnea. Multiple complications might occur including cyst
infection, hemorrhage, and rupture [2]. Currently available
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
therapeutic interventions in PLD are aimed at reducing liver
volume and improving quality of life. These procedures include
aspiration, sclerotherapy, cyst resection, fenestration, and liver
transplantation [4]. However, while these procedures might
relieve the symptoms and improve quality of life, they are inva­
sive, only partially effective, have high recurrence risk, and do
not block disease progression.
The only currently available pharmacological therapy in PLD
is off‐label use of somatostatin analogs which reduce total liver
volume and improve quality of life. While the beneficial value
of somatostatin analogs was shown in multiple clinical trials,
their effects are modest, not all patients respond to treatment,
and the cost is very high [5, 6]. Clinical testing of UDCA did not
show reduction in the total liver volume in human subjects with
isolated ADPLD, however, liver cyst volume was decreased in
patients with ADPKD‐associated PLD [7]. Therefore, the search
for new therapies is necessary and ongoing.
The discovery of PLD‐related genes and the identification of
mechanisms underlying cyst progression have provided a better
understanding of the pathogenesis of this disease leading to new
perspectives for therapeutic interventions. The current knowl­
edge related to the genetics of PLD, the mechanisms involved,
and potential therapeutic targets identified are discussed in this
chapter.
GENETICS
As of today, there are nine PLD‐causative genes that could be
classified as: (i) exclusively mutated in ADPKD‐related PLD
(i.e. PKD1 and PKD2), (ii) exclusively mutated in isolated
ADPLD (PRKCSH, SEC63, ALG8, LRP5, and SEC61B), and
 33:  Polycystic Liver Diseases: Genetics, Mechanisms, and Therapies 409

Figure 33.1  PLD‐related genes and proteins. Hepatic cystogenesis in ADPKD‐associated PLD, ARPKD‐associated PLD, and in isolated ADPLD
is triggered by mutations in nine genes. PKD1 and PKD2 are causative for ADPKD‐associated PLD. Mutations in GANAB are associated with both
ADPKD‐associated PLD and isolated ADPLD. PKHD1 is implicated in development of both ADPKD and ADPLD. Five genes (i.e. PRKCSH, SEC63,
ALG8, LRP5, and SEC61B) are responsible for hepatic cystogenesis in ADPLD. Despite the genetic heterogeneity, in all forms of PLD multiple cysts
occupied liver parenchyma. Photograph of the resected liver from Pkhd1del2/del2 mice, light microscopic image of the liver section from the PCK rat and
scanning electron micrographs of the liver from Pkd2WS25/− mice demonstrate the presence of cysts in different animal models of PLD.

(iii) commonly mutated in PKD‐associated PLD and in isolated
ADPLD (i.e. GANAB and PKHD1) [2, 4, 8, 9]. Despite the
genetic heterogeneity and complexity, phenotypically all PLDs
are characterized by the presence of numerous liver cysts of
variable size and shape that occupy the majority of the liver
parenchyma (Figure 33.1).
ADPKD‐associated PLD
The most common PLD is associated with ADPKD. The prev­
alence of ADPKD is 1 : 400 and up to 90% of ADPKD patients
develop PLD [2, 9, 10]. The kidney and liver cysts in this dis­
order are the result of mutations in three genes  –  PKD1,
PKD2, and GANAB. PKD1 and PKD2 genes encode, respec­
tively, polycystin 1 (PC1) and polycystin 2 (PC2) proteins
(Figure  33.1). PC1 is expressed on the plasma membrane
and along primary cilia (Figure  33.2a). PC2 is present in
endoplasmic reticulum (ER), plasma membrane, and primary
cilia (Figure 33.2a). The exact functions of PC1 and PC2 are
not well understood. PC1 is known to interact with PC2
­regulating cell/cell–matrix interactions, cell proliferation, and
G‐­protein‐coupled signal‐transduction pathways [4, 8, 9]. GANAB
encodes the catalytic alpha subunit of glucosidase II (GΙΙα,
also known as PKD3) and is a one of the member of the gly­
cosyl hydrolase family of 31 proteins. GIIα resides in ER and
plays a role in protein folding and quality control (Figures 33.1
and 33.2a) [4, 8, 9].
ARPKD‐associated PLD
ARPKD (prevalence is 1 : 10 000–20 000) causes significant
renal and liver‐related morbidity and mortality in children.
ARPKD is characterized by enlarged kidneys with various
degrees of renal dysfunction, bile duct dilatation, and congenital
hepatic fibrosis that lead to portal hypertension [11]. The num­
ber of affected individuals surviving the neonatal period that
might require liver transplantation increases significantly with
age. The disease is caused by mutation in a single PKHD1 gene
which encodes the protein fibrocystin (FC, Figure 33.1). FC is
expressed at plasma membrane and primary cilia (Figure 33.2a).
While the functions of FC remained largely unclear, experimen­
tal data suggest that this protein may act as a membrane‐bound
receptor involved in regulation of the cell–cell adhesion and
proliferation [11, 12].
410 THE LIVER:  GENETICS

Figure 33.2  Interactions of genes and proteins in PLD. (a) Schematic diagram illustrates the site of expression of PLD‐related proteins. PKD‐
associated PLD is known as a cholangiociliopathy indicating the nature of disease origin (i.e. cholangiocytes) and organelles involved (i.e.
­primary cilia). Indeed, the protein products of the genes mutated in PKD‐associated PLD are all localized to primary cilia. Proteins encoded by
the genes involved in isolated ADPLD are localized to ER but play an important role in targeting of PC1, PC2, and FC to the plasma membrane
and cilia. (b) Mutations in ADPLD‐related genes reduce functional dosage of PC1 resulting in hepatic cystogenesis. Abbreviations: FC, ­fibrocystin;
PC1, polycystin 1; PC, polycystin 2; LRP5, low‐density lipoprotein receptor‐related protein 5; GIIα, glucosidase II subunit alpha; GIIβ, glucosi­
dase II subunit beta; ALG8, alpha‐1,3‐glucosyltransferase; GANAB, glucosidase II alpha subunit; SEC63, protein‐transporting protein SEC63;
SEC61, translocon subunit SEC61; PRKCSH, protein kinase C substrate 80K‐H; PKHD1, polycystic kidney hepatic disease 1. Reproduced from [9]
with permission.

Isolated ADPLD
Isolated ADPLD (prevalence 1 : 100 000) is characterized by
the  predominant presence of liver cysts [5, 13]. ADPLD is a
result of heterozygous loss‐of‐function mutations in at least five
genes  –  PRKCSH, SEC63, GANAB, ALG8, and SEC61B
(Figure  33.1). There is some evidence that mutations in
PRKCSH (responsible for ~15% of clinical cases) and SEC63
are associated with early onset of disease and more severe clini­
cal symptoms [2].
The protein products of PRKCSH (the regulatory beta‐subu­
nit of glucosidase II, GIIβ), SEC63 (SEC63), GANAB (GIIα),
ALG8 (α‐1,3‐glucosyltransferase), and SEC61B (beta‐subunit
of the SEC61 protein translocation apparatus) are integral ER
membrane residents (Figure 33.2a) that facilitate translocation
and transmembrane insertion of multiple glycoproteins
­including PC1 and PC2 [8, 14–18]. In addition, LRP5 has been
implicated in development of isolated ADPLD; however, no
clear loss‐of‐function variants in the LRP5 gene have been
found in ADPLD patients [19]. The protein product of LRP5
(i.e. LRP5) is located to the cell membrane where it interacts
with Frizzled in the canonical Wnt signaling pathway
(Figures 33.1 and 33.2a). Upon binding of Wnt to the Frizzled/
LRP5 complex, the canonical Wnt/beta‐catenin pathway is
activated subsequently enhancing cholangiocyte proliferation.
Finally, mutations in the ARPKD‐associated gene, PKHD1,
was shown as disease‐causative not only in ARPKD but in
ADPLD as well (Figure 33.1) [20].
Seven genes associated with isolated ADPLD account for
approximately 50% of genetically identifiable cases; therefore
the list of genes causative for this form of PLD is expected to
grow. In line with this, a recent study in mice demonstrated that
deletion of the mitochondrial uncoupling protein 2 gene (i.e.
Ucp2) led to a spontaneous development of hepatic cysts that
pathologically resembled human PLD [21].
Genetic interaction network in PLD
As mentioned, the genetic landscape of PLD includes nine
causative genes. Recent data clearly demonstrate a genetic
interaction network within PLD‐related genes [9, 22]. First,
mutations in GANAB cause both ADPKD and isolated ADPLD
while PKHD1 was found to be mutated in patients with
ARPKD and ADPLD (Figure 33.1) [8, 16, 23]. Second, stud­
ies in animal models demonstrate that loss‐of‐function in
PRKCSH, SEC63, GANAB, ALG8, and SEC61B genes com­
bined with the functional dosage of PC1 act as a predictor of
severity of hepatic and renal cystogenesis (Figure 33.2b) [8,
16, 22, 23]. Third, GANAB is crucial for PC1 and PC2 matu­
ration and localization to the cell surface and cilia [16].
Fourth, GΙΙβ and SEC63 are required for formation of PC1/
PC2 functional complex with PC1 being the rate‐limiting
component of this process [23]. Fifth, mutations in ALG8,
GANAB, and SEC61B reduce steady‐state levels of PC1;
among them, SEC61B causes the most notable quantitative
reduction of PC1 (Figure 33.2b) [8]. Sixth, mutations in the
PKHD1 gene likely contribute to hepatic cystogenesis through
aberrant interaction of PC1 with FC without affecting biogen­
esis of PC1 [8]. Finally, missense variants in LRP5 coexist
with PKHD1 variants suggesting that interaction of these two
genes contribute to hepatic cystogenesis in patients with
mutated PKHD1 [8].
 33:  Polycystic Liver Diseases: Genetics, Mechanisms, and Therapies 411
MECHANISMS OF HEPATIC
CYSTOGENESIS
Cholangiocytes lining liver cysts are enlarged with thickened
basement membrane [24–26]. While the majority of hepatic
cysts are lined by single layer of cholangiocytes, around 10% of
cysts in animal models of PLD (i.e. the PCK rat and
Pkd2WS25/−mice) and around 20% of cysts in ADPKD patients
are multilayered [25].
Ductal plate malformation
Ductal plate malformation (i.e. the embryological arrest of
ductal plate development) contributes to hepatic cystogenesis in
PLD [5, 27]. Incomplete development of ductal plates leads to
the presence of their remnants along the portal tract margins in
ARPKD‐associated PLD [28]. ADPKD‐associated PLD and
isolated ADPLD are also linked to ductal plate malformation;
however, the pathological mechanisms involved are unknown
[5, 27]. In addition, many regulators of ductal plate remodeling
(such as growth factors, transcriptional factors, and miRNAs)
are abnormally expressed in cystic cholangiocytes, therefore
further connecting ductal plate malformation and hepatic cys­
togenesis [5, 27].
Cholangiocyte primary cilia
PKD‐associated PLD is often referred as a cholangiociliopathy
[29, 30] accentuating the cell of origin and affected organelle in
hepatic cystogenesis. In control cholangiocytes primary cilia
(around 7 μm in length) extend from the apical membrane into
the bile duct lumen and function as chemo‐, osmo‐ and mecha­
nosensors detecting and transducing extracellular stimuli into the
cell interior to regulate cholangiocyte proliferation, planar cell
polarity, intracellular cyclic adenosine monophosphate (cAMP)
and calcium signaling and many other cellular functions [29, 30].
The protein products of genes that cause ADPKD‐ and
ARPKD‐linked PLD (i.e. PKD1, PKD2, and PKHD1) are all
expressed in primary cilia and play an important role in preser­
vation of ciliary structure and functional integrity (Figure 33.2a)
[31]. Indeed, mutations in PKD1, PKD2, and PKHD1 are asso­
ciated with various ciliary abnormalities including aberrantly
shaped, shortened, and malformed cholangiocyte cilia in vivo
and in vitro (Figure 33.3a,b) [12, 24, 25]. In addition, basal bodies
(i.e. the nucleation site for the growth of primary cilia) and cen­
trosomes are also defective in PLD as evident by their aberrant
positioning and elongated shape [26]. The presence of supernu­
merary centrosomes in cholangiocytes within hepatic cysts
results in the formation of extra cilia in up to 15% of the cells
[26]. Malfunctioned planar cell polarity (i.e. improper align­
ment of the mitotic spindle during cell division which leads to
longitudinal rather than horizontal cell growth) has also been
implicated [32]. On the other hand, the protein products of
genes which are causative for isolated ADPLD are not present in
cilia (Figure  33.2a); however, several studies clearly demon­
strated that they are crucial for biogenesis of proteins associated
with PKD‐related PLD (i.e. PC1, PC2, and FC) and their proper
translocation to the cholangiocyte primary cilia [8, 16, 23].
Morphological abnormalities in primary cilia, basal bodies,
and/or centrosomes coupled with disturbances in ciliary func­
tions accelerate cholangiocyte proliferation and cyst expan­
sion [5, 9, 27, 31]. Peculiarities in primary cilia and centrosomes
are also accompanied by dysregulated cell cycle machinery
and anomalous expression of the cell cycle‐related proteins
[26, 33, 34].
Intracellular cAMP and [Ca2+]i signaling
pathways
Elevated levels of cAMP in PLD cholangiocytes is considered
to be one of the major mechanisms that drives hepatic cystogen­
esis [5, 14, 31]. This second messenger was shown to regulate
many cellular functions including autophagy, the cell cycle,
fluid secretion, and cell proliferation via its downstream effec­
tors Rap guanine nucleotide exchange factor 3 also known as
exchange factor directly activated by cAMP (EPAC), protein
kinase A (PKA), and mitogen‐activated protein kinase/extracel­
lular signal‐regulated kinase (MAPK/ERK) [29, 35–38]. The
mechanisms behind cAMP elevations in PLD cholangiocytes
are still not well understood. However, recent data suggest that
the overexpression of cAMP‐linked bile acid‐responsive
G‐­protein coupled receptor, TGR5, combined with previously
published observations that the concentration of endogenous
TGR5 ligands is increased in cystic cholangiocytes, might con­
tribute to cAMP overproduction in PLD [38, 39].
PLD cholangiocytes are also characterized by the reduced
concentration of intracellular calcium ([Ca2+]i) believed to result
from a dysfunctional PC1/PC2 ciliary complex [32]. Mutational
defects in PC2 reduce entry of extracellular calcium that leads
to depletion of cytoplasmic and ER calcium stores altering
homeostasis of this signaling molecule [27, 32, 40]. Decreased
[Ca2+]i may contribute to cAMP elevations by activating c­ alcium
inhibitable adenylyl cyclase 5/6 (AC5/6) and by preventing
cAMP degradation by phosphodiesterase 1 (PDE1) and PDE3
[41]. Alternatively, due to existing crosstalk between these two
signaling pathways, cAMP can influence calcium signaling via
PKA activation [10, 27].
Cholangiocyte autophagy
Recently we discovered that autophagy, an evolutionary con­
served pathway involved in regulation of cellular homeostasis
via degradation and recycling of cytoplasmic constituents, is
abnormal in PLD [37]. The morphologic appearance of orga­
nelles involved in autophagy is strikingly different from those in
control cholangiocytes; for example, the number of autophago­
somes and their size was around 2–3 times greater in cultured
PCK rat and human ADPKD cholangiocytes (Figure 33.3c) as
well as in cholangiocytes lining liver cysts in animal models of
PLD and human patients with PLD compared to respective con­
trols [37]. Changes in autophagosome morphology was accom­
panied by similar changes (i.e. increased number and size) in
lysosomes and autolysosomes, the organelles critical for proper
functioning of autophagic machinery [37]. Global analysis of
the PLD cholangiocyte transcriptome followed by functional
pathway cluster analysis showed that the autophagy–lysosomal
Figure 33.3  PLD cholangiocytes are characterized by a number of morphological abnormalities. (a) In cholangiocytes lining liver cysts in an
animal model of PLD, the PCK rat, primary cilia are shorter and malformed as assessed by scanning electron microscopy (SEM), transmission
electron microscopy (TEM), and immunofluorescent confocal microscopy (IMF) compared to cholangiocytes lining bile ducts in wild‐type rat.
(b) (Consistent with in vivo data, shorter and defective cilia are also present in cultured cholangiocytes derived from the PCK rat as compared to
cholangiocytes derived from wild‐type rat (NRC). Reproduced from [101] with permission. (c) Transmission electron microscopic images demon­
strate that the number and size of autophagosomes (depicted by black arrows) are greater in cystic cholangiocytes derived from the PCK rat and
human patients with ADPKD compared to control cholangiocytes derived, respectively, from the wild‐type rat (NRC) and healthy humans (NHC).
 33:  Polycystic Liver Diseases: Genetics, Mechanisms, and Therapies 413
pathway is one of the most altered pathways in PLD. Cystic
cholangiocytes have aberrantly expressed autophagy‐related
proteins and enhanced autophagic flux. Furthermore, altered
autophagy in PLD cholangiocytes is accompanied by activation
of the cAMP–PKA–cAMP response element binding protein
(CREB) signaling pathway and enhanced cell proliferation [37].
It has been also reported that mutations in PRKCSH induces
autophagy although a different mechanism (i.e. mTOR‐­
mediated) was implicated [42]. In line with our observations,
the absence of endogenous hepatocystin (i.e. the protein product
of PRKCSH gene) or the presence of mutated hepatocystin in
an  in  vitro experimental system induces autophagy [42].
Together  these data emphasize the important contribution of
autophagy in hepatic cystogenesis and suggest autophagy as a
therapeutic target.
Cholangiocyte proliferation
Cholangiocyte proliferation is considered to be one of the
important mechanisms of hepatic cystogenesis and in many pre­
clinical studies was shown to be a marker of disease progression
or regression [24, 29, 33–35, 39, 43–51]. Under basal con­
ditions, cholangiocytes are dormant but become hyperprolifera­
tive during cyst expansion as indicated by their positive staining
for proliferating cell nuclear antigen (PCNA) [48–50, 52].
However in well‐established disease, proliferation of cystic
cholangiocytes diminishes [27, 53]. Different cytokines and
growth factors (such as interleukin 6 [IL‐6], epidermal growth
factor [EGF], vascular epidermal growth factor [VEGF], and
insulin‐like growth factor [IGF]) accelerate cholangiocyte pro­
liferation via ERK and phosphatidylinositol‐4,5‐bisphosphate
3‐kinase (PI3‐kinase)–protein kinase B (AKT)–mTOR path­
ways. These cytokines and growth factors are present in the
cystic fluid and/or secreted by cystic cholangiocytes [27, 32]. In
PLD linked to ARPKD, a proliferation‐independent mechanism
(i.e. excessive differentiation of biliary precursors during ductal
plate remodeling) appears to play a role in early stages of cyst
development followed by proliferation‐dependent cyst expan­
sion in the postnatal period [54].
Fluid secretion
Increased secretion of fluid into the lumen of hepatic cysts is
also an important contributing mechanism in PLD progression.
In cystic cholangiocytes of animal models and humans with
PLD, cystic fibrosis transmembrane conductance regulator
(CFTR; i.e. the chloride channel), anion exchange protein 2
(AE2; i.e. the chloride/bicarbonate exchanger), and aquaporin 1
(AQP1; i.e. the water channel) are all overexpressed and mislo­
calized. Activation of cAMP machinery in cultured cholangio­
cytes expedites insertion of these proteins into the apical
membrane and thus facilitates water influx [43]. In line with this
observation, an overexpression of AQP1 was recently reported
in PLD patients [55].
The concentration of bile acids is increased in the liver of
PCK rats and in cystic fluid of PLD patients [39]. Given the
importance of these molecular species in bile flow regulation
and the overexpression of bile acid‐responsive receptor TGR5
in cystic cholangiocytes, it has been suggested that bile acids
might contribute to enhanced fluid secretion in PLD. However,
no such experimental evidence exists so far.
miRNAs
As of today a single experimental study described the role of
miRNAs in PLD progression [34]. It was shown that miR‐15a is
one of the most downregulated miRNA in rat cystic cholangio­
cytes. This miRNA targets the cell cycle master phosphatase,
cell division cycle 25 (Cdc25A), the expression of which is
markedly increased in PLD. Experimental upregulation of miR‐
15a in cystic cholangiocytes decreases levels of Cdc25A, inhib­
its cell proliferation, and reduces cyst growth in vitro. In
contrast, downregulation of miR‐15a in control cholangiocytes
increases expression of Cdc25A, enhances cell proliferation,
and accelerates cyst growth in culture [34]. Therefore, these
data indicate that alterations in miR‐15a/Cdc25A axis contrib­
ute to the molecular pathogenesis of PLD.
We also found that besides miR‐15a, expression of additional
67 miRNAs was dysregulated in PLD cholangiocytes [56].
Importantly, the mRNA targets of these miRNAs include those
related to cycle machinery, fluid secretion, cAMP pathway, and
calcium intracellular signaling further indicating that miRNAs
are specific regulators of affected intracellular pathways in PLD.
Other mechanisms of hepatic cystogenesis
In addition to the above mentioned mechanisms implicated in PLD
pathogenesis, many more have been experimentally proven to pro­
mote hepatic cystogenesis (Figure  33.4a). These mechanisms
include, in particular: (i) enhanced angiogenesis and cyst wall neo­
vascularization, (ii) extracellular matrix remodeling, and (iii) global
changes in mRNA and protein expression [10, 32, 49, 57].
We recently performed mRNA profiling analysis of human
cholangiocytes derived from healthy individuals and from
patients with ADPKD and ADPLD using the NexGen sequenc­
ing (NGS) technology. The expression patterns of mRNAs dif­
fer strikingly in control and cystic cholangiocytes. Many
transcripts are up‐ or downregulated in PLD. Importantly,
ADPKD and ADPLD cholangiocytes share a number of com­
monly expressed genes (Figure  33.4b). Clustering of these
genes by functional pathway cluster analysis reveals previously
identified cellular pathways and exposes some novel pathways
as well (Figure 33.4c).
THERAPIES
Accumulating evidence suggest that mutations in PLD‐causa­
tive genes initiate the formation of hepatic cysts that continue to
grow because of disturbances in cellular functions and signaling
pathways. The efficacy of targeting of these processes for poten­
tial therapeutic applications has been tested in cultured cystic
cholangiocytes and in animal models of PLD [2, 4, 9, 10, 27].
Several of these preclinical studies have been translated into
clinical trials. Specifically, the role of the somatostatin analogs,
octreotide and pasireotide, and the effects of UDCA were exam­
ined in patients with ADPKD‐associated PLD and isolated
414 THE LIVER:  SINGLE DRUG THERAPIES

Figure 33.4  Cholangiocyte morphological defects and abnormal signaling pathways underlie hepatic pathology in PLD. (a) Cystic cholangio­
cytes are characterized by structurally deformed primary cilia, increased number of autophagic organelles and mitochondria, and numerous dys­
regulated cellular functions. (b) Using the NextGen sequencing approach, 2310 and 1914 dysregulated transcripts were detected, respectively, in
cholangiocytes derived from human patients with ADPKD‐associated PLD and isolated ADPLD. These genes were differentially expressed (i.e.
up‐ and downregulated) in cystic cholangiocytes compared to cholangiocytes derived from healthy individuals (fold changed great than 2.0, P
< 0.05). Using a Venn diagram analysis, 1078 transcripts were commonly present in cystic cholangiocytes. (c) The pathway cluster analysis of these
1078 genes identified pathways associated with PLD pathogenesis.

ADPLD [2, 4–7, 58]. In clinical trials, total liver volume is a
prognostic marker of disease and the primary endpoint for
assessing drug effects in patients. In preclinical trials, liver
cystic areas and the rate of cholangiocyte proliferation are usu­
ally used to assess disease progression/regression. The majority
of preclinical studies have focused on the evaluation of a single
drug; however, studies that examine combinational strategies
have also begun to emerge. Here we discuss current promising
monotherapies and combinational therapies in PLD treatment
which are summarized in Table 33.1.
SINGLE DRUG THERAPIES
Somatostatin analogs
Synthetic analogs of somatostatin currently represent the main
pharmacological treatment options for PLD patients and are
prescribed off‐label [5, 27, 58]. Out of five somatostatin recep­
tors (SSTR) known to be present in cholangiocytes octreotide
binds to SSTR2, SSTR3, and SSTR5. Another somatostatin
analog, pasireotide, is more potent than octreotide, has a broader
 33:  Polycystic Liver Diseases: Genetics, Mechanisms, and Therapies 415

Table 33.1  Therapeutic targets in polycystic liver disease (PLD)

Target Studies Drug Outcome Ref
SINGLE DRUG THERAPIES
cAMP Preclinical Octreotide, pasireotide Decreased cholangiocyte proliferation in vitro and hepatic [33, 47]
cystogenesis in PCK rats and Pkd2WS25/− mice
Clinical Octreotide, lanreotide Decreased total liver volume and increased quality of life [59, 62–69,
in patients with isolated ADPLD and ADPKD‐associated 71, 73, 74]
PLD
Pasireotide A clinical trial involving patients with isolated ADPLD
and ADPKD‐associated PLD is underway at the Mayo
Clinic (NCT01670110107)
Pasireotide No improvement in cyst reduction after aspiration [75]
sclerotherapy
Intracellular Preclinical UDCA Decreased cholangiocyte proliferation in vitro and hepatic [39]
calcium cystogenesis in PCK rats
Clinical UDCA No changes in total liver volume; decreased liver cystic [7]
volume in patients with ADPKD‐associated PLD (but not
in ADPLD patients)
Tendency to decrease the total liver volume in patients [80]
with PLD
Autophagy Preclinical HCQ Decreased cholangiocyte proliferation in vitro and hepatic [37]
cystogenesis in PCK rats
TGR5 Preclinical SBI‐115 Decreased cholangiocyte proliferation and cysts growth [38]
in vitro
AC5 Preclinical SQ22536 Decreased cholangiocyte proliferation in vitro and hepatic [83]
cystogenesis in PC2‐defective mice
PKA, MEK Preclinical PKA & MEK inhibitors Decreased cholangiocyte proliferation and cysts growth [29, 84]
in vitro
FSHR Preclinical Reduction by shRNA Decreased cholangiocyte proliferation in vitro [85, 100]
VR2 Preclinical Tolvaptan, OPC‐31260 Decreased cholangiocyte proliferation in vitro [86]
Single case Tolvaptan Decreased total liver volume in a single patient with [88]
report severe PLD
TRPV4 Preclinical GSK1016790A Decreased cholangiocyte proliferation in vitro and hepatic [98]
cystogenesis in PCK rats
BNP Preclinical BNP Decreased cholangiocyte proliferation in vitro and hepatic [89]
cystogenesis in PCK rats
HSP90 Preclinical STA‐2842 Decreased cholangiocyte proliferation in vitro and hepatic [90]
cystogenesis in Pkd1KO mice
CDC25A Preclinical Menadione, PM‐20 Decreased cholangiocyte proliferation in vitro and hepatic [48]
cystogenesis in PCK rats and Pkd2WS25/− mice
VEGF Preclinical SU5416 Decreased cholangiocyte proliferation in vitro and hepatic [50, 92]
cystogenesis in PC2‐deficient mice
PPARγ Preclinical Pioglitazone, telmisartan Decreased hepatic cystogenesis in PCK rats [93, 94]
HDAC6 Preclinical Tubastatin‐A, tubacin, Decreased cholangiocyte proliferation in vitro and hepatic [35, 44]
panobinostat, ACY‐1215, cystogenesis in PCK rats
ACY‐738, ACY‐241
MMP Preclinical Marimastat Decreased cholangiocyte proliferation in vitro and hepatic [95]
cystogenesis in PCK rats
mTOR Preclinical Rapamycin Decreased cholangiocyte proliferation in vitro and hepatic [49]
cystogenesis in Pkd2KO mice
Sirolimus No effects on hepatic cystogenesis in PCK rats [96]
Everolimus Decreased hepatic cystogenesis in PCK rats [97]
c‐Src Preclinical SKI‐606 Decreased hepatic cystogenesis in PCK rats [51]
COMBINATIONAL DRUG THERAPIES
cAMP Clinical Everolimus No synergistic effects on total liver volume [98]
+ +
mTOR octreotide
cAMP Preclinical Octreotide Octreotide abolished the paradoxical effects of sorafenib [84]
+ + on hepatic cystogenesis in PC2‐deficient mice
RAF sorafenib
cAMP Preclinical Octreotide No additive effects on hepatic cystogenesis in PCK rats [99]
+
pasireotide
cAMP Preclinical Pasireotide Additive effects on cholangiocyte proliferation and cyst [38]
+ growth in vitro
SBI‐115
cAMP Preclinical Pasireotide Additive effects on cholangiocyte proliferation in vitro and [37]
+ + hepatic cystogenesis in PCK rats
autophagy HCQ
cAMP Preclinical Pasireotide Additive effects on cholangiocyte proliferation in vitro and [35]
+ + hepatic cystogenesis in PCK rats
HDAC6 ACY‐1215
416 THE LIVER:  SINGLE DRUG THERAPIES
spectrum of binding (i.e. binds to SSTR1, SSTR2, SSTR3, and
SSTR5) and has around six times the half‐life [33, 58]. In
­clinical trials, three somatostatin analogues (namely, octreotide,
lanreotide, and pasireotide) have been used. A dose of 120 mg
of lanreotide is equivalent to 60 mg of octreotide [59].
Preclinical studies
In cultured rat and human cystic cholangiocytes, both octreotide
and pasireotide decreased cAMP levels and cell proliferation with
pasireotide being more effective [33, 47]. Consistent with this, a
more substantial reduction of hepatic cystogenesis was observed
in animal models of PLD (i.e. PCK rats and Pkd2WS25/− mice) after
treatment with pasireotide than with octreotide [33, 47]. The
reduced cyst growth was coupled with decreased cAMP levels
and rate of cholangiocyte proliferation. These studies provided a
strong justification for assessing the potential benefits of somato­
statin analogs in humans.
Clinical trials
A number of clinical trials have evaluated the efficacy of soma­
tostatin analogs in patients with isolated ADPLD and ADPKD‐
associated PLD. The overall summary of the findings is as
follows. First, octreotide and lanreotide modestly (by around
5%) decreased total liver volume when administrated for 6–12
months; this decrease remained stable for 12 months post‐ther­
apy. Second, with continuous octreotide usage, the reduction is
liver volumes is maintained; however, when treatment is
stopped, liver volume begins to increase toward baseline. Third,
patients with the most severe PLD appear to respond to treat­
ment more efficiently than those with modest disease. Fourth, in
some patients (around 15%) treatment has no effects on total
liver volume and quality of life. Fifth, somatostatin analogues
are well tolerated and improve quality of life. Unfortunately,
somatostatin analogs are very expensive, US$7000  –  US$11
000 per month for patients with health insurance who have to
obtain pre‐approval usually on a case‐by‐case basis [6, 59–74].
A clinical trial (NCT01670110107) to examine the efficacy
of a more potent somatostatin analog pasireotide in ADPKD‐
associated PLD and isolated ADPLD is now underway at the
Mayo Clinic College of Medicine, USA. In line with this, in
PLD patients who underwent aspiration sclerotherapy of a large,
single, symptomatic, hepatic cyst, pasireotide did not improve
the reduction in cyst size. Inadequate dose of pasireotide or
­limited timing of drug administration (administered two weeks
before and two weeks after procedure) might explain the nega­
tive result [75].
Ursodeoxycholic acid
Bile acids have been recognized as signaling molecules regulat­
ing cell proliferation, apoptosis, and liver regeneration [76].
Ursodeoxycholic acid (UDCA; 3α,7β‐dihydroxy‐5β‐cholanoic
acid), an endogenous hydrophilic bile acid, stimulates a bicar­
bonate‐rich choleresis protecting both hepatocytes and cholan­
giocytes against the cytotoxicity of hydrophobic bile acids by
increasing levels of intracellular calcium [77–79]. Taking into
consideration that cystic cholangiocytes are characterized by
decreased levels of intracellular calcium a potential role of
UDCA as therapeutic agent in PLD has been tested in preclini­
cal and clinical studies.
Preclinical studies
In PCK rats (that have increased cAMP, decreased intracellular
calcium, and elevated concentration of toxic bile acids), UDCA
inhibits hepatic cystogenesis, fibrosis, and cholangiocyte
hyperproliferation and improves their physical fitness. UDCA
also normalized bile acid concentrations in bile. These effects
overlap with restoration of the intracellular calcium in cystic
cholangiocytes via a PI3K–AKT–MEK/ERK‐dependent mech­
anism [39].
Clinical trials
Given the positive outcome of preclinical studies, a clinical trial
was performed to assess the effects of UDCA in patients with
ADPKD‐associated PLD and isolated ADPLD. UDCA was
well tolerated with no major side‐effects. While no changes
in  total liver volume were observed in response to UDCA
­treatment, liver cystic volume was decreased in patients with
ADPKD‐associated PLD (but not in ADPLD patients). The
levels of the γ‐glutamyl transpeptidase (GGT) and alkaline
­
phosphatase (ALP) were reduced in all patients [7]. In another
small trial, UDCA treatment also reduced the levels of biliary
enzymes (i.e. GGT and ALP) and had a tendency (although not
statistically significant) to decrease the total liver volume [80].
Thus, UDCA might be a pharmacological option for patients
with ADPKD‐associated PLD [5, 7, 57, 80].
Autophagy
Recent data strongly suggest that altered autophagy in PLD
cholangiocytes contributes to disease progression and might
serve as a potential therapeutic target [37, 42]. Indeed, treatment
of cultured rat and human cholangiocytes with inhibitors of
autophagy, bafilomycin A1 and hydroxychloroquine (HCQ),
decreased their proliferation and cyst growth in vitro. However,
in PLD cholangiocytes with genetically reduced expression of
one of the major autophagy‐related proteins, ATG7, these effects
of bafilomycin A1 and HCQ were abolished. Finally, in PCK
rats HCQ treatment attenuated hepatic cystogenesis by decreas­
ing cAMP levels and proliferation of cholangiocytes lining liver
cysts [37].
cAMP‐linked bile acid receptor TGR5
TGR5 is expressed in different cell types regulating multiple
cellular processes and recently has been considered as a poten­
tial therapeutic target in metabolic, inflammatory, and digestive
diseases [38]. Under basal conditions TGR5 is localized to dif­
ferent cholangiocyte compartments including primary cilia and
apical membrane [36, 81]. In PLD cholangiocytes, TGR5 is
overexpressed and the concentrations of endogenous TGR5
agonists (e.g. lithocholic, chenodeoxycholic, deoxycholic, and
cholic acids) are increased in livers and serum of PCK rats
[38, 39]. Activation of cultured rat and human cystic cholangio­
cytes with TGR5 natural or synthetic agonists further increased
cAMP levels, cell proliferation, and cyst growth. Furthermore,
 33:  Polycystic Liver Diseases: Genetics, Mechanisms, and Therapies 417
the TGR5 agonist, xenobiotic oleanolic acid, worsened hepatic
cystogenesis in PCK rats and increased proliferation of cystic
cholangiocytes in vivo. Genetic reduction of TGR5 in vitro or
complete elimination of this G protein‐coupled receptor (GPCR)
in vivo by generating TGR5−/−; Pkhd1del2/del2 double mutant mice
prevented the effects of TGR5 activation of cyst growth and
reduced cAMP production. Together, these data indicate that
TGR5 contributes to cAMP elevation and to hepatic cystogene­
sis in PLD representing a potential therapeutic target. Indeed, a
novel small molecule TGR5 antagonist, SBI‐115, decreases
cAMP levels, rate of proliferation, and cyst growth in cultured
cholangiocytes emphasizing the beneficial potential of TGR5
inhibition in PLD [38].
Adenylyl cyclase 5 (AC5)
Several studies implicated AC5/6 in the pathogenesis of PLD.
Cholangiocyte primary cilia possess a protein complex consist­
ing of AC5/6 and the A kinase anchor protein AKAP150.
Activation of this complex by luminal flow triggers changes in
intracellular calcium and cAMP levels [82]. In PC2‐defective
cells, inhibition of AC5 reduced cAMP production, expression
of cAMP downstream effector ERK, VEGF secretion, and
growth of biliary organoids. Finally, an AC5 inhibitor, SQ22536,
decreased liver cystic area and cell proliferation in PC2‐­
defective mice [83].
cAMP downstream effectors
cAMP machinery has been shown to pay an important role in
PLD pathogenesis. One of the PKA regulatory subunits, PKA‐
RIb, is overexpressed in cystic cholangiocytes and its activation
triggers cell proliferation and cyst growth in vitro [29]. These
PKA‐stimulated effects were blocked by the PKA‐ and MEK‐spe­
cific inhibitors decreasing ERK phosphorylation, rate of prolifera­
tion, and cyst growth in 3D cultures [29, 84]. Given the recent
recognition that PKA targeting might be beneficial in other prolif­
erative diseases such as cancer, it would be of value to test the
effects of PKA inhibitors on hepatic cystogenesis in vivo.
Follicle‐stimulating hormone receptor (FSHR)
FSHR is a cAMP‐associated receptor involved in follicular
growth. FSHR is present in cystic cholangiocytes derived from
livers of patients with ADPKD‐associated PLD and its stimula­
tion increases cell proliferation in a cAMP–ERK‐dependent
manner; these effects were abolished after genetic reduction of
FSHR. Nevertheless, therapeutic efficacy of FSHR has not yet
been tested in animal models of PLD [85].
Vasopressin
Preclinical studies
Recently, vasopressin receptor 2 (VR2; one of the three known
arginine vasopressin receptors – VR1A, VR1B, and VR2) was
found to be present in control cholangiocytes [86]. Moreover,
overexpressed VR2 was detected in cystic cholangiocytes of
ADPKD patients. Activation of this cAMP‐linked receptor with
its agonist, vasopressin, increased cell proliferation and cAMP
levels which were blocked by preincubation of cholangiocytes
with two VR2 antagonists, tolvaptan and OPC‐31260 [86]. As
of today, effects of VR2 targeting on hepatic cystogenesis have
not been tested preclinically.
Single clinical case
An inhibitor of VR2, tolvaptan, was shown to decrease kidney
volume in patients with PKD in clinical trials and has recently
been approved by the Food and Drug Administration (FDA) for
treatment of renal (but not hepatic) cystogenesis in ADPKD
patients [87]. A single clinical report also demonstrates that
tolvaptan decreased total liver volume after 17 months of treat­
ment in a patient with severe PLD [88]. As of today, no other
data related to efficacy of tolvaptan in PLD patients have been
reported.
TRPV4
The calcium entry channel transient receptor potential vanilloid 4
(TRPV4) is overexpressed at the mRNA and protein levels in
PLD cholangiocytes. Activation of TRPV4 in these cells increased
levels of intracellular calcium subsequently inhibiting cell prolif­
eration and cyst growth in vitro by activating AKT and inhibiting
the BRAF–ERK signaling pathway. Partial silencing of TRPV4
by small interfering RNA blocked effects of TRPV4 activation on
these processes. Finally, a TRPV4 activator, GSK1016790A,
decreased hepatic cystogenesis in PCK rats [44].
BNP B‐type natriuretic peptide (BNP)
BNP, a guanylyl cyclase A agonist, triggers cyclic guanosine
monophosphate (cGMP) production demonstrating antifibrotic,
antihypertensive, and vasopressin‐suppressive properties. BNP
decreased renal cystic epithelial cell proliferation, suppressed
fibrotic gene expression, and increased intracellular calcium in
vitro. Continuous BNP production in PCK rats (in response to
adeno associated viral serotype 9 [AAV9] vectors that carry
cytomegalovirus [CMV]‐driven codon‐optimized proBN­
PcDNA) inhibited hepatic cystogenesis. Thus, targeting cGMP
in PLD may provide a novel therapeutic strategy [89].
HSP90
Heat shock protein 90 (HSP90) is a chaperone protein that con­
trols proper folding of its clients preventing their degradation. In
conditional Pkd1 KO mice, inhibitor of HSP90 STA‐2842 (a
resorcinolic triazole molecule) prevented growth of hepatic
cysts and decreased cholangiocyte proliferation. In addition,
expression and activity of HSP90 client proteins, EGFR and
ERK1/2, was reduced [90].
CDC25A
In PLD cholangiocytes overexpression of the master cell cycle
regulator Cdc25A is associated with cell cycle dysregulation
and increased numbers of cells with multiple centrosomal
abnormalities [26, 34]. Genetic depletion of Cdc25A from
418 THE LIVER:  COMBINATIONAL DRUG THERAPIES
cultured cholangiocytes or complete elimination of this protein
by generating Cdc25A+/−;Pkhd1del2/del2 double mutant mice nor­
malized these processes [26, 48]. In addition, suppression of
Cdc25A by increasing levels of its regulator, miR‐15A, inhib­
ited cholangiocyte proliferation and cyst growth in vitro [34].
These observations provided the rationale to study the therapeu­
tic potential of Cdc25A targeting in PCK rats and Pkd2WS25/−
mice. Two Cdc25A inhibitors (i.e. menadione and PM‐20)
decreased cyst growth in these models of PLD by reducing
activity and expression of Cdc25A [48].
VEGF
Vascular endothelial growth factor (VEGF)/VEGF receptor 2
(VEGFR2) plays an important role in hepatic cystogenesis.
Decreased levels of intracellular calcium and increased
­production of cAMP in PLD cholangiocytes stimulate mTOR/
hypoxia‐inducible factor 1‐alpha (HIFα) axis via the PKA–
RAS–RAF–ERK pathway leading to overproduction of VEGF
[49, 50]. The VEGFR1 and VEGFR2 are also upregulated in
liver cysts and their activation enhanced cholangiocyte prolifera­
tion in a cAMP‐dependent fashion [91]. In contrast, an inhibitor
of VEGFR2, SU5416, reduced hepatic cystogenesis in PC2‐defi­
cient mice [50, 92].
PPARγ
Several studies suggest that pioglitazone (a full agonist of
PPARγ) and telmisartan (a partial PPARγ agonist) decreased
liver cyst growth and elevated blood pressure in PCK rats
which is accompanied by inhibited expression of the trans­
forming growth factor beta 1 (TGFβ) in cystic cholangiocytes
[93, 94].
Histone deacetylase 6 (HDAC6)
HDAC6 (in contrast to other members of the HDAC family that
epigenetically control gene expression) is predominantly local­
ized in the cytoplasm and is involved in regulation of multiple
cellular process including cell cycle progression, autophagy,
and ciliary disassembly. HDAC6 is upregulated in PLD cholan­
giocytes and its inhibition by tubastatin‐A, tubacin, and
ACY‐1215 decreased cell proliferation in vitro in a dose‐ and
time‐dependent manner. Several HDAC6 inhibitors (i.e. panobi­
nostat, ACY‐1215, ACY‐738, and ACY‐241) reduced hepatic
cystogenesis in PCK rats; however, ACY‐1215 exhibits stronger
suppressive effects [35, 44].
Metalloproteinases (MMP)
Increased levels of different MMPs have been observed in PLD
cholangiocytes. Enhanced secretion of IL‐6, IL‐8, and 17β‐
estradiol by cystic cholangiocytes promote in an autocrine/par­
acrine fashion hyperactivity of MMPs that, in turn, stimulate
the digestion and remodeling of extracellular matrix [47]. When
secretion of IL‐6 and/or IL‐8 was blocked, enhanced
­proliferation of cystic cholangiocytes was reduced. Importantly,
MMP inhibitor, marimastat, attenuated hepatic cystogenesis in
PCK rats [95].
mTOR
The mammalian target of rapamycin (mTOR) has been impli­
cated in the regulation of a variety of fundamental cellular pro­
cesses including cell growth, metabolism, protein synthesis and
autophagy. In PLD, a dysregulated mTOR signaling network
contributes to disease progression by affecting a spectrum of
intracellular pathways, including growth factor signaling [49].
Overexpression of mTOR in cystic cholangiocytes is accompa­
nied by upregulation of VEGF, IGF1, IGF1R, and PKA.
Rapamycin, an inhibitor of mTOR, lessened cyst growth in
Pkd2 KO mice by decreasing cholangiocyte proliferation. The
effects of rapamycin were coupled with accumulation of HIF1α
(the major transcription factor of VEGF) and inhibition of
VEGF secretion with involvement of the MEK–ERK pathway
[49]. On the other hand, the mTOR inhibitor, sirolimus, had no
effects on hepatic and renal cystogenesis in PCK rats [96].
Finally, an alternative mTOR inhibitor, everolimus, halted
hepatic cystogenesis and reduced hepatic fibrosis in PCK rats
via PI3K–AKT–mTOR signaling [97]. In this study, the conse­
quences of everolimus treatment on disease progression were
evaluated by magnetic resonance imaging approach using liver
volume as endpoint in contrast to previous studies in which
cystic areas were measured in thin liver sections. Therefore, the
role of mTOR inhibitors as a therapeutic entity in PLD remains
unclear due to limited number of experimental reports and con­
troversial outcomes.
Proto‐oncogene tyrosine‐protein kinase Src
(c‐Src)
c‐Src, which phosphorylates tyrosine residues of different pro­
teins, was shown to promote angiogenesis and proliferation in
different cell types. Cystic cholangiocytes of PCK rats have
increased c‐Src activity and a pharmacologic c‐Src inhibitor,
SKI‐606, ameliorated biliary duct abnormalities in this model of
PLD [51].
COMBINATIONAL DRUG THERAPIES
cAMP and mTOR
The effects of the somatostatin analog, lanreotide, and the
mTOR inhibitor, everolimus, on disease pathogenesis were
evaluated in patients with ADPKD‐associated PLD after kidney
transplants. While lanreotide decreased total liver volume as
previously reported [64], no synergistic reduction on disease
progression was achieved in response to drug combination [98].
cAMP and RAF
Given that in PC2‐deficient mice cAMP together with its down­
stream effector PKA activates RAS–RAF–MEK–ERK pathway
resulting in overgrowth of hepatic cysts, a therapeutic potential
of RAF inhibition by sorafenib was examined. Unexpectedly,
sorafenib further accelerated hepatic cystogenesis and ERK
phosphorylation by activating RAF1 in PC2‐deficient mice.
 33:  Polycystic Liver Diseases: Genetics, Mechanisms, and Therapies 419
Combinational administration of octreotide and sorafenib abol­
ished the paradoxical effects of the latter [84].
cAMP
Octreotide + pasireotide
The effects of coadministration of octreotide and pasireotide on
hepatic cystogenesis were examined in PCK rats. As previously
reported [33, 47], each somatostatin analog decreased hepatic
cystogenesis; however, no additive effects in reduction of cyst
growth were observed in response to drug combination [99].
Pasireotide + SBI‐115
We recently reported that elevated cAMP in PLD cholangio­
cytes might be a result of overexpression of the cAMP‐linked
GPCR, TGR5 [38]. We also showed that TGR5 antagonist,
SBI‐115, decreased cholangiocyte proliferation and growth of
cystic structures in vitro. Therefore, we tested how targeting of
cAMP machinery by the somatostatin analog, pasireotide, and
the TGR5 inhibitor, SBI‐115, affected these processes. More
substantial inhibition of cAMP levels, rate of proliferation, and
cyst growth in vitro was observed in response to concurrent drug
administration compared to each drug alone [38]. Genetic
depletion of TGR5 in cultured cholangiocytes abolished the
effects of SBI‐115. While this result is encouraging, further
studies in animal models are needed.
cAMP and autophagy
When combined, pasireotide and HCQ reduced cAMP levels
and decreased proliferation of cultured cystic cholangiocytes to
a greater extent compared to each drug alone. Consistent with
this, co‐treatment of PCK rats with pasireotide and HCQ showed
additive effects of the drug combination on reduction in liver
weight, hepatic cystic and fibrotic areas, and rate of cell prolif­
eration within the liver cysts [37].
cAMP and HDAC6
Among several HDAC6 inhibitors (i.e. panobinostat, ACY‐1515,
ACY‐738, and ACY‐241) tested in PCK rats, ACY‐1215 was
most effective. Therefore, the synergistic effects of ACY‐1215
and HDAC6 in combination on disease progression in vitro and
in vivo were tested. In cultured cystic cholangiocytes, a mixture
of ACY‐1215 and pasireotide more effectively decreased cell
proliferation and reduced cAMP levels compared to each drug
alone. In PCK rats, co‐administration of two drugs synergisti­
cally attenuated hepatic cystogenesis and increased length of
primary cilia showing additive effects of combinational therapy
on these processes [35].
CONCLUSION
Considerable progress in our understanding of the genetic
landscape in PLD and identification of dysregulated signal­
ing pathways and abnormal cellular functions of hepatic
cystogenesis facilitated the discovery of multiple potential
therapeutic targets for disease treatment. Many of these tar­
gets have been tested preclinically showing beneficial
effects on suppression of cell proliferation and cyst growth
in different experimental systems (i.e. cultured cystic chol­
angiocytes, in vitro models of cystogenesis, and animal
models of PLD). However, these targets still have to be eval­
uated in clinical trials. Mounting evidence suggests that
drug combinations might be more effective that monothera­
pies in PLD treatment. Clinically, somatostatin analogs and
UDCA have been evaluated in patients with ADPKD‐­
associated PLD and isolated ADPLD, yet only somatostatin
analogs showed beneficial effects and are used as pharmaco­
therapy in PLD.
REFERENCES
1. Masyuk, T., Masyuk, A., and LaRusso, N. Cholangiociliopathies: genetics,
molecular mechanisms and potential therapies. Curr Opin Gastroenterol,
2009;25(3):265–71.
2. van Aerts, R.M.M. et  al. Clinical management of polycystic liver disease.
J Hepatol, 2017.
3. Neijenhuis, M.K. et al. Impact of liver volume on polycystic liver disease‐
related symptoms and quality of life. United European Gastroenterol J,
2018;6(1):81–8.
4. van de Laarschot, L.F.M. and Drenth, J.P.H. Genetics and mechanisms of
hepatic cystogenesis. Biochim Biophys Acta, 2018;1864(4 Pt B):1491–7.
5. Santos‐Laso, A.,et al. New advances in polycystic liver diseases. Semin Liver
Dis, 2017;37(1):45–55.
6. Khan, S., Dennison, A., and Garcea, G. Medical therapy for polycystic liver
disease. Ann R Coll Surg Engl, 2016;98(1):18–23.
7. D’Agnolo, H.M. et  al. Ursodeoxycholic acid in advanced polycystic liver
disease: a phase 2 multicenter randomized controlled trial. J Hepatol,
2016;65(3):601–7.
8. Besse, W. et  al. Isolated polycystic liver disease genes define effectors of
polycystin‐1 function. J Clin Invest, 2017;127(9):3558.
9. Masyuk, T.V., Masyuk, A.I., and LaRusso, N.F. Polycystic liver disease: the
interplay of genes causative for hepatic and renal cystogenesis. Hepatology,
2018;67(6):2462–4.
10. Masyuk, T.V., Masyuk, A.I., and LaRusso, N.F. Therapeutic targets in poly­
cystic liver disease. Curr Drug Targets, 2017;18(8):950–957.
11. Bergmann, C. Genetics of autosomal recessive polycystic kidney disease and
its differential diagnoses. Front Pediatr, 2017;5:221.
12. Woollard, J.R. et  al. A mouse model of autosomal recessive polycystic
­kidney disease with biliary duct and proximal tubule dilatation. Kidney Int,
2007;72(3):328–36.
13. Wills, E.S., Roepman, R., and Drenth, J.P. Polycystic liver disease:
ductal  plate malformation and the primary cilium. Trends Mol Med,
2014;20(5):261–70.
14. Cnossen, W.R. and Drenth, J.P. Polycystic liver disease: an overview of
pathogenesis, clinical manifestations and management. Orphanet J Rare
Dis, 2014;9:69.
15. Davila, S. et al. Mutations in SEC63 cause autosomal dominant polycystic
liver disease. Nat Genet, 2004;36(6):575–7.
16. Porath, B. et  al. Mutations in GANAB, encoding the glucosidase II alpha
subunit, cause autosomal‐dominant polycystic kidney and liver disease.
Am J Hum Genet, 2016;98(6):1193–207.
17. Drenth, J.P. et  al. Germline mutations in PRKCSH are associated with
­autosomal dominant polycystic liver disease. Nat Genet, 2003;33(3):345–7.
18. Waanders, E. et  al. Cysts of PRKCSH mutated polycystic liver disease
patients lack hepatocystin but express Sec63p. Histochem Cell Biol,
2008;129(3):301–10.
19. Cnossen, W.R., et al. Whole‐exome sequencing reveals LRP5 mutations and
canonical Wnt signaling associated with hepatic cystogenesis. Proceedings
of the National Academy of Science,Science,,s of the United States of
America, 2014;111(14):5343–8.
420 THE LIVER:  REFERENCES
20. Besse, W. et al. A noncoding variant in GANAB explains isolated polycystic
liver disease (PCLD) in a large family. Hum Mutat, 2018;39(3):378–82.
21. Hirose, M. et  al. The mitochondrial uncoupling protein 2 gene is causal
for  the spontaneous polycystic liver diseases in mice. Mitochondrion,
2017;42:50–3.
22. Bergmann, C. and Weiskirchen, R. It’s not all in the cilium, but on the road
to it: genetic interaction network in polycystic kidney and liver diseases and
how trafficking and quality control matter. J Hepatol, 2012;56(5):1201–3.
23. Fedeles, S.V. et al. A genetic interaction network of five genes for human
polycystic kidney and liver diseases defines polycystin‐1 as the central deter­
minant of cyst formation. Nature Genetics, 2011;43(7):639–47.
24. Masyuk, T.V. et  al. Biliary dysgenesis in the PCK rat, an orthologous
model  of autosomal recessive polycystic kidney disease. Am J Pathol,
2004;165(5):1719–30.
25. Stroope, A. et  al. Hepato‐renal pathology in pkd2ws25/‐ mice, an animal
model of autosomal dominant polycystic kidney disease. Am J Pathol,
2010;176(3):1282–91.
26. Masyuk, T.V. et  al. Centrosomal abnormalities characterize human and
rodent cystic cholangiocytes and are associated with Cdc25A overexpres­
sion. Am J Pathol, 2014;184(1):110–21.
27. Perugorria, M.J. et al. Polycystic liver diseases: advanced insights into the
molecular mechanisms. Nat Rev Gastroenterol Hepatol, 2014;11(12):
750–61.
28. Raynaud, P. et  al. Biliary differentiation and bile duct morphogenesis in
development and disease. Int J Biochem Cell Biol, 2011;43(2):245–56.
29. Banales, J.M. et al. The cAMP effectors Epac and protein kinase a (PKA) are
involved in the hepatic cystogenesis of an animal model of autosomal reces­
sive polycystic kidney disease (ARPKD). Hepatology, 2009;49(1):160–74.
30. Masyuk, A.I., Masyuk, T.V., and LaRusso, N.F. Cholangiocyte primary cilia
in liver health and disease. Dev Dyn, 2008;237(8):2007–12.
31. Masyuk, T.V. et al. Defects in cholangiocyte fibrocystin expression and cili­
ary structure in the PCK rat. Gastroenterol, 2003;125(5):1303–10.
32. Strazzabosco, M. and Somlo, S. Polycystic liver diseases: congenital disor­
ders of cholangiocyte signaling. Gastroenterol, 2011;140(7):1855–9.
33. Masyuk, T.V. et al. Pasireotide is more effective than octreotide in reducing
hepatorenal cystogenesis in rodents with polycystic kidney and liver dis­
eases. Hepatology, 2013;58(1):409–21.
34. Lee, S.O. et al. MicroRNA15a modulates expression of the cell‐cycle regula­
tor Cdc25A and affects hepatic cystogenesis in a rat model of polycystic
kidney disease. J Clin Invest, 2008;118(11):3714–24.
35. Lorenzo Pisarello, M. et al. Combination of a histone deacetylase 6 inhibitor
and a somatostatin receptor agonist synergistically reduces hepatorenal cys­
togenesis in an animal model of polycystic liver disease. Am J Pathol,
2018;188(4):981–94.
36. Masyuk, A.I. et al. Ciliary subcellular localization of TGR5 determines the
cholangiocyte functional response to bile acid signaling. Am J Physiol
Gastrointest Liver Physiol, 2013;304(11):G1013–24.
37. Masyuk, A.I. et  al. Cholangiocyte autophagy contributes to hepatic cys­
togenesis in polycystic liver disease and represents a potential therapeutic
target. Hepatology, 2018;67(3):1088–108.
38. Masyuk, T.V. et al. TGR5 contributes to hepatic cystogenesis in rodents with
polycystic liver diseases through cyclic adenosine monophosphate/Galphas
signaling. Hepatology, 2017;66(4):1197–218.
39. Munoz‐Garrido, P. et al. Ursodeoxycholic acid inhibits hepatic cystogenesis
in experimental models of polycystic liver disease. J Hepatol, 2015;63(4):
952–61.
40. Masyuk, A.I., Gradilone, S.A., and LaRusso, N.F. Calcium signaling in cilia
and ciliary‐mediated intracellular calcium signaling: are they independent or
coordinated molecular events? Hepatology, 2014;60(5):1783–5.
41. Spirli, C. et al. Altered store operated calcium entry increases cyclic 3′,5′‐
adenosine monophosphate production and extracellular signal‐regulated
kinases 1 and 2 phosphorylation in polycystin‐2‐defective cholangiocytes.
Hepatology, 2012;55(3):856–68.
42. Yang, J. et  al. Deficiency of hepatocystin induces autophagy through an
mTOR‐dependent pathway. Autophagy, 2011;7(7):748–59.
43. Banales, J.M. et al. Hepatic cystogenesis is associated with abnormal expres­
sion and location of ion transporters and water channels in an animal
model  of autosomal recessive polycystic kidney disease. Am J Pathol,
2008;173(6):1637–46.
44. Gradilone, S.A. et al. HDAC6 is overexpressed in cystic cholangiocytes and
its inhibition reduces cystogenesis. Am J Pathol, 2014;184(3):600–8.
45. Gradilone, S.A. et al. Cholangiocyte cilia express TRPV4 and detect changes
in luminal tonicity inducing bicarbonate secretion. Proc Natl Acad Sci U S A,
2007;104(48):19138–43.
46. Gradilone, S.A. et  al. Activation of Trpv4 reduces the hyperproliferative
phenotype of cystic cholangiocytes from an animal model of ARPKD.
­
Gastroenterol, 2010;139(1):304–14.
47. Masyuk, T.V. et  al. Octreotide inhibits hepatic cystogenesis in a rodent
model of polycystic liver disease by reducing cholangiocyte adenosine 3’,5’‐
cyclic monophosphate. Gastroenterol, 2007;132(3):1104–16.
48. Masyuk, T.V. et al. Inhibition of Cdc25A suppresses hepato‐renal cystogen­
esis in rodent models of polycystic kidney and liver disease. Gastroenterol,
2012;142(3):622–33.
49. Spirli, C. et al. Mammalian target of rapamycin regulates vascular endothe­
lial growth factor‐dependent liver cyst growth in polycystin‐2‐defective
mice. Hepatology, 2010;51(5):1778–88.
50. Spirli, C. et al. ERK1/2‐dependent vascular endothelial growth factor signal­
ing sustains cyst growth in polycystin‐2 defective mice. Gastroenterol,
2010;138(1):360–71.
51. Sweeney, W.E., Jr. et al. Src inhibition ameliorates polycystic kidney d­ isease.
J Am Soc Nephrol, 2008;19(7):1331–41.
52. Alvaro, D. et al. Morphological and functional features of hepatic cyst epi­
thelium in autosomal dominant polycystic kidney disease. Am J Pathol,
2008;172(2):321–32.
53. Waanders, E. et  al. Disrupted cell adhesion but not proliferation mediates
cyst formation in polycystic liver disease. Mod Pathol, 2008;21(11):
1293–302.
54. Beaudry, J.B. et  al. Proliferation‐independent initiation of biliary cysts in
polycystic liver diseases. PLoS One, 2015;10(6):e0132295.
55. Li, D. et al. Overexpression of aquaporin 1 on cysts of patients with polycys­
tic liver disease. Rev Esp Enferm Dig, 2016;108(2):71–8.
56. Masyuk, T., Masyuk, A., and LaRusso, N., MicroRNAs in cholangiociliopa­
thies. Cell Cycle, 2009;8(9):1324–8.
57. Perugorria, M.J. et  al. Bile acids in polycystic liver diseases: triggers of
­disease progression and potential solution for treatment. Dig Dis, 2017;35(3):
275–81.
58. Larusso, N.F., Masyuk, T.V., and Hogan, M.C. Polycystic liver disease:
the  benefits of targeting cAMP. Clin Gastroenterol Hepatol, 2016;14(7):
1031–4.
59. Hogan, M.C. et al. Randomized clinical trial of long‐acting somatostatin for
autosomal dominant polycystic kidney and liver disease. J Am Soc Nephrol,
2010;21(6):1052–61.
60. Gevers, T.J. and Drenth, J.P., Somatostatin analogues for treatment of poly­
cystic liver disease. Curr Opin Gastroenterol, 2011;27(3):294–300.
61. Temmerman, F. et al. Safety and efficacy of different lanreotide doses in the
treatment of polycystic liver disease: pooled analysis of individual patient
data. Aliment Pharmacol Ther, 2013;38(4):397–406.
62. Pisani, A. et al. Long‐term effects of octreotide on liver volume in patients
with polycystic kidney and liver disease. Clin Gastroenterol Hepatol,
2016;14(7):1022–30.
63. van Keimpema, L., de Man, R.A. and Drenth, J.P., Somatostatin analogues
reduce liver volume in polycystic liver disease. Gut, 2008;57(9):1338–9.
64. van Keimpema, L. et al. Lanreotide reduces the volume of polycystic liver:
a  randomized, double‐blind, placebo‐controlled trial. Gastroenterol,
2009;137(5):1661–8.
65. Caroli, A. et  al. Reducing polycystic liver volume in ADPKD: effects of
somatostatin analogue octreotide. Clin J Am Soc Nephrol, 2010;5(5):
783–9.
66. Hogan, M.C. et al. Somatostatin analog therapy for severe polycystic liver
disease: results after 2 years. Nephrology, dialysis, transplantation: Nephrol
Dial Transplant, 2012;27(9):3532–9.
67. Neijenhuis, M.K. et al. Somatostatin analogues improve health‐related qual­
ity of life in polycystic liver disease: a pooled analysis of two randomised,
placebo‐controlled trials. Aliment Pharmacol Ther, 2015;42(5):591–8.
68. Gevers, T.J. et  al. Effect of lanreotide on polycystic liver and kidneys in
autosomal dominant polycystic kidney disease: an observational trial. Liver
Int, 2015;35(5):1607–14.
69. Hogan, M.C. et al. Efficacy of 4 years of octreotide long‐acting release ther­
apy in patients with severe polycystic liver disease. Mayo Clin Proc,
2015;90(8):1030–7.
70. Drenth, J.P. et al. Medical and surgical treatment options for polycystic liver
disease. Hepatology, 2010;52(6):2223–30.
 33:  Polycystic Liver Diseases: Genetics, Mechanisms, and Therapies 421
71. Chrispijn, M. et al. The long‐term outcome of patients with polycystic liver
disease treated with lanreotide. Aliment Pharmacol Ther, 2012;35(2):266–74.
72. Gevers, T.J. et al. Young women with polycystic liver disease respond best to
somatostatin analogues: a pooled analysis of individual patient data.
Gastroenterol, 2013;145(2):357–65.
73. Temmerman, F. et al. Safety and efficacy of different lanreotide doses in the
treatment of polycystic liver disease: pooled analysis of individual patient
data. Aliment Pharmacol Ther, 2013;38(4):397–406.
74. Temmerman, F. et  al. Lanreotide reduces liver volume, but might not
improve muscle wasting or weight loss, in patients with symptomatic poly­
cystic liver disease. Clin Gastroenterol Hepatol, 2015;13(13):2353–9.
75. Wijnands, T.F. et al. Efficacy and safety of aspiration sclerotherapy of simple
hepatic cysts: a systematic review. Am J Roentgenol, 2017;208(1):201–7.
76. Marin, J.J. et al. Bile acids in physiology, pathology and pharmacology. Curr
Drug Metab, 2015;17(1):4–29.
77. Poupon, R. Ursodeoxycholic acid and bile‐acid mimetics as therapeutic
agents for cholestatic liver diseases: an overview of their mechanisms of
action. Clin Res Hepatol Gastroenterol, 2012;36(1):S3–12.
78. Beuers, U. et al. New paradigms in the treatment of hepatic cholestasis: from
UDCA to FXR, PXR and beyond. J Hepatol, 2015;62(1):S25–37.
79. Glaser, S. et al. Adrenergic receptor agonists prevent bile duct injury induced
by adrenergic denervation by increased cAMP levels and activation of Akt.
Am J Physiol Gastrointest Liver Physiol, 2006;290(4):G813–26.
80. Iijima, T. et al. Ursodeoxycholic acid for treatment of enlarged polycystic
liver. Ther Apher Dial, 2016;20(1):73–8.
81. Keitel, V. and Haussinger, D. TGR5 in cholangiocytes. Curr Opin
Gastroenterol, 2013;29(3):299–304.
82. Masyuk, A.I. et al. Cholangiocyte cilia detect changes in luminal fluid flow
and transmit them into intracellular Ca2+ and cAMP signaling. Gastroenterol,
2006;131(3):911–20.
83. Spirli, C. et al. Adenylyl cyclase 5 links changes in calcium homeostasis to
cAMP‐dependent cyst growth in polycystic liver disease. J Hepatol,
2017;66(3):571–80.
84. Spirli, C. et al. Cyclic AMP/PKA‐dependent paradoxical activation of Raf/
MEK/ERK signaling in polycystin‐2 defective mice treated with sorafenib.
Hepatology, 2012;56(6):2363–74.
85. Onori, P. et al. Role of follicle‐stimulating hormone on biliary cyst growth in
autosomal dominant polycystic kidney disease. Liver Int, 2013;33(6):914–25.
86. Mancinelli, R. et al. Vasopressin regulates the growth of the biliary epithe­
lium in polycystic liver disease. Lab Invest, 2016;96(11):1147–55.
87. Edwards, M.E. et al. Long‐term administration of tolvaptan in autosomal dom­
inant polycystic kidney disease. Clin J Am Soc Nephrol, 2018;13(8):1153–61.
88. Mizuno, H. et al. Tolvaptan for the treatment of enlarged polycystic liver
disease. Case Rep Nephrol Dial, 2017;7(3):108–11.
89. Holditch, S.J. et al. B‐type natriuretic peptide overexpression ameliorates
hepatorenal fibrocystic disease in a rat model of polycystic kidney disease.
Kidney Int, 2017;92(3):657–68.
90. Smithline, Z.B. et al. Inhibiting heat shock protein 90 (HSP90) limits the
formation of liver cysts induced by conditional deletion of Pkd1 in mice.
PLoS One, 2014;9(12):e114403.
91. Fabris, L. et al. Effects of angiogenic factor overexpression by human and
rodent cholangiocytes in polycystic liver diseases. Hepatology,
2006;43(5):1001–12.
92. Amura, C.R. et  al. VEGF receptor inhibition blocks liver cyst growth in
pkd2(WS25/‐) mice. Am J Physiol Cell Physiol, 2007;293(1):C419–28.
93. Yoshihara, D. et al. Telmisartan ameliorates fibrocystic liver disease in an
orthologous rat model of human autosomal recessive polycystic kidney
­disease. PloS One, 2013;8(12):e81480.
94. Yoshihara, D. et al. PPAR‐gamma agonist ameliorates kidney and liver dis­
ease in an orthologous rat model of human autosomal recessive polycystic
kidney disease. Am J Physiol Renal Physiol, 2011;300(2):F465–74.
95. Urribarri, A.D. et al. Inhibition of metalloprotease hyperactivity in cystic
cholangiocytes halts the development of polycystic liver diseases. Gut,
2014;63(10):1658–67.
96. Renken, C. et  al. Inhibition of mTOR with sirolimus does not attenuate
progression of liver and kidney disease in PCK rats. Nephrol Dial
Transplan, 2011;26(1):92–100.
97. Temmerman, F. et  al. Everolimus halts hepatic cystogenesis in a rodent
model of polycystic‐liver‐disease. World J Gastroenterol, 2017;23(30):
5499–507.
98. Chrispijn, M. et al. Everolimus does not further reduce polycystic liver vol­
ume when added to long acting octreotide: results from a randomized con­
trolled trial. J Hepatol, 2013;59(1):153–9.
99. Kugita, M. et al. Beneficial effect of combined treatment with octreotide
and pasireotide in PCK rats, an orthologous model of human autosomal
recessive polycystic kidney disease. PLoS One, 2017;12(5):e0177934.
100. Mancinelli, R. et al. Follicle‐stimulating hormone increases cholangiocyte
proliferation by an autocrine mechanism via cAMP‐dependent phospho­
rylation of ERK1/2 and Elk‐1. Am J Physiol Gastrointest Liver Physiol,
2009;297(1):G11–26.
101. Muff, M., Masyuk, T., Stroope, A. et al. Development and characterization
of a cholangiocyte cell line from the PCK rat, an animal model of autoso­
mal recessive polycystic kidney disease. Laboratory Investigation 2006;
86:940–950. PMID:16783394.
 The Liver Sinusoidal
34 Endothelial Cell: Basic
Biology and Pathobiology
Karen K. Sørensen and Bård Smedsrød
Vascular Biology Research Group, Department of Medical Biology, UiT The Arctic University of Norway,
Tromsø, Norway
INTRODUCTION
Endothelial cells constitute a heterogenous cell population
with distinct structural and functional characteristics depend-
ing on tissue type. Different from other microvascular
endothelial cells, the liver sinusoidal endothelial cells
(LSECs) are highly perforated and lack an organized basal
lamina [1]. This signature morphological feature of the LSEC
allows free passage of macromolecules and colloids, such as
lipoproteins, across the sinusoidal lining. The functional char-
acteristics of LSECs include their active endocytosis recep-
tors endowing the cells with an extraordinary high endocytic
capability [2]. This makes these cells the body’s most effec-
tive scavenger cells, clearing blood‐borne macromolecules
and nanosized compounds, thus contributing importantly to
maintaining homeostasis. A special metabolic feature of
LSECs is their very high production of lactate signifying their
largely anaerobic metabolism, which is also reflected in their
low number of mitochondria. LSECs represent an important
player in immunity: they express the endocytic Fc‐gamma
receptor IIb2 (CD32b, FcγRIIb2), several pattern recognition
receptors (PRRs), for example, various toll‐like receptors
(TLRs), as well as the mannose receptor (MR) and several
scavenger receptors (SRs), of which stabilin‐1 and stabilin‐2
represent the most important LSEC SRs. LSECs further dis-
play features of adaptive immunity, contributing to hepatic
immune tolerance [3].
In liver pathobiology, LSECs play central roles in liver fibro-
sis [4, 5], cancer metastasis [6, 7], and liver toxicology [8]. In
addition to being involved in the etiology of pathologies of the
liver, it is noteworthy that the LSECs are also believed to have a
role in the development of complications at extrahepatic sites,
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
for example, atherosclerosis, systemic lupus erythematosus
(SLE), and glomerular fibrotic nephrosis.
This update starts by outlining central aspects of the basic
biology of the LSEC. The importance of LSECs in the etiology
of selected pathologies of the liver and extrahepatic sites will
then be described. A section on LSEC‐mediated liver toxicity is
also included.
STRUCTURE–FUNCTION
CONSIDERATIONS
The hepatocytes represent the major cell type in the liver (92.5%
of the total liver cell volume), with LSECs, Kupffer cells, and
stellate cells making up 3.3%, 2.5%, and 1.7% of the cell v­ olume
[9]. In spite of the volume‐wise predominance of the hepato-
cytes, the minor cell populations play pivotal functional roles
due to the anatomical and morphological organization of the
organ. The liver is highly vascularized, receiving about a quarter
of cardiac output. The blood that enters liver is a mixture of 25%
oxygen‐rich blood from the hepatic artery and 75% oxygen‐poor
blood that enters via the portal vein. The portal vein carries gut‐
derived nutrients and potentially or overtly harmful s­ ubstances
that will be surveilled and if needed eliminated by the liver,
ensuring that blood leaving the liver is compatible with normal
homeostasis. This efficient filtering capability of the liver can be
mainly ascribed to the following three structure/function features
of the liver sinusoid:
1. The cells facing the blood flowing through the sinusoids
include mainly the LSECs that make up the wall of the
 34:  The Liver Sinusoidal Endothelial Cell: Basic Biology and Pathobiology 423

(a)

SC
LSEC WBC

RBC
KC

Space of Disse

HC

(b)
KC
LSEC

Virus Bacteria
Oligonucleotides
Nanoparticles

Proteins Red blood cells

1nm 10 nm 100 nm 1μm 10 μm

Figure 34.1  (a) The liver sinusoid with its main cell types. The cartoon illustrates the localization of hepatocytes (HC), liver sinusoidal endothelial
cells (LSEC), Kupffer cells (KC), and stellate cells (SC). The Kupffer cells are normally located at the luminal aspect of the sinusoid, whereas the
stellate cells are located between the LSECs and hepatocytes, in the space of Disse. RBC, red blood cell; WBC, white blood cell. Blue and green
dots represent soluble macromolecules. The figure is reproduced from [2] with permission. (b) Share of scavenging workload between LSECs and
Kupffer cells. Together LSECs and Kupffer cells make up the largest population of scavenger cells in the body [10]. LSECs are geared to effective
clathrin‐mediated endocytosis and avidly endocytose a wide range of soluble macromolecules and nanoparticles. LSECs are essentially non‐phago-
cytic, and larger particles, such as bacteria and aged and damaged red blood cells are therefore cleared by the Kupffer cells, illustrating “the dual
cell principle of waste handling” [11]. The figure is a courtesy from Dr Kjetil Elvevold, D’Liver AS Tromsø, Norway.

numerous hepatic sinusoids (Figure 34.1a), and the Kupffer
cells, representing by far the largest population of mac-
rophages in our body. Both of these cell types have unsur-
passed ability to recognize and endocytose soluble and
particulate material that is incompatible with normal homeo-
stasis. Together the two cell types make up the largest scaven-
ger cell system in the body [10, 11] (Figure 34.1b).
2. The LSECs are perforated with pores 100–150 nm in
diameter, commonly referred to as fenestration, and occu-
pying as much as 20% of the LSEC [12]. The LSECs lack
a basal lamina and form the only fenestrated endothelium
that offers open passage of macromolecules and nanopar-
ticles between the blood and underlying parenchyma
(Figure 34.2).
3. Also contributing to the high blood filtering capability of the
liver is the large total surface area of the sinusoidal endothe-
lium. Data from stereological analysis of the rat liver [13]
can be used to calculate that the total LSEC surface area of
an average normal adult human liver is approx. 210 m2 (i.e.
80% of a tennis court).
THE LSEC SCAVENGER FUNCTION
AND ITS IMPLICATIONS
Historical considerations
It was only after detailed ultrastructural studies carried out by
Eddie Wisse in the early 1970s that LSECs were distinguished
as a cell type distinct from the Kupffer cells [1]. Development of
a method for mass isolation of LSECs, Kupffer cells, and hepat-
ocytes from a single rat liver, allowing establishment of primary
cultures of purified cells, enabled the discovery that the extra-
cellular matrix polysaccharide hyaluronan was avidly and spe-
cifically endocytosed and degraded mainly in LSECs, but not in
other types of liver cells [14]. This assigned a role for the LSEC
as a novel type of physiologically important scavenger cell [15].
Over the years that followed, the list of macromolecules found
to be cleared mainly by LSECs increased steadily, and it is now
generally accepted that LSECs are far more important than
Kupffer cells in the clearance of soluble macromolecules and
nanomaterial from the circulation [2] (Table 34.1). This notion
424 THE LIVER:  THE LSEC SCAVENGER FUNCTION AND ITS IMPLICATIONS

(a) (b)

HC

LSEC SD HC

HC

LSEC
1µm HC 1µm

(c) (d)

PV

50 µm 2 µm

Figure 34.2  Morphology of the liver sinusoidal endothelial cell. (a) Scanning electron micrograph of mouse liver showing the high porosity of
the liver sinusoidal endothelium. The fenestrae of the LSECs are typically organized in groups, termed sieve plates (encircled by a white dashed
line). HC, hepatocyte; SD, space of Disse. (b) Transmission electron micrograph of a rat liver sinusoid. The very thin cytoplasm of the LSEC con-
tains many fenestrations (arrow heads), and there is no organized basal membrane beneath the endothelium. The fenestrae provide open channels
allowing bidirectional traffic of fluids, solutes, and small particles between the blood and hepatocytes. (c) Immune fluorescence micrograph of
mouse liver, showing uptake of BK polyomavirus‐like particles (green fluorescence, arrow heads) in the sinusoidal endothelium 15 minutes after
intravenous injection, visualized by an anti‐BK polyomavirus antibody. PV, portal vein. The experiment was carried out as part of the study
described in [69]. (d) Scanning electron micrograph of rat LSEC. Freshly isolated LSECs were plated on collagen coated tissue culture plastic and
fixed one hour after seeding. The cells are highly fenestrated; the white dashed line encircles a sieve plate.

came as a surprise to many researchers in the field because it
deviated from the paradigm that hepatic blood clearance is a
function of the reticuloendothelial system (RES), a term that has
been erroneously used synonymously with the macrophage, or
the mononuclear phagocyte system. The name and notion of
RES, launched nearly 100 years ago by Ludwig Aschoff [16],
was based on the consistent observation that different types of
vital stains (i.e. colloidal, nanosized compounds) accumulated
in specific anatomical locations, and most notably along the
hepatic sinusoids. Although Aschoff himself did not link RES
specifically to macrophage or phagocytosis, scientists gradually
came to view the RES as a cellular system responsible for
phagocytic uptake of blood‐borne colloidal vital stains mainly
or exclusively in macrophages. The reason why this happened
must be understood in light of the fact that although the terms
macrophage and phagocytosis had been introduced by
Metchnikoff [17] several years prior to Aschoff’s publication,
the perception of these terms have been considerably developed
up through the years. Yet, even at the present time it is not
uncommon to find scientific papers and updated textbooks
where hepatic uptake is described merely as phagocytic uptake
in Kupffer cells. To resolve this misconception experiments
have been performed in exactly the same way as described by
Aschoff and his contemporaries, and with vital stains prepared
according to the original protocols as those used 100–140 years
ago. Thus revisiting these early studies about 75 years after
Aschoff launched his concept of RES revealed that the uptake of
the most frequently used vital stain, lithium‐carmine, was pre-
dominantly in LSECs [18].
The dual cell principle of waste clearance
Thanks to the studies reported in the above paragraph we know
today that the liver scavenger system consists of both the Kupffer
cells, that are geared mainly to phagocytic uptake of particles
greater than 200 nm, and LSECs that clear blood‐borne macromol-
ecules and nanoparticles less than 200 nm (Figure 34.1b). This con-
cept, referred to as the dual cell principle of waste clearance [11]
must be taken into consideration when aiming to control or predict
uptake of blood‐borne compounds in liver. Of note, the principle
applies to animal species of all vertebrate classes. However, the
scavenger endothelial cells are mainly in liver in land‐based
 34:  The Liver Sinusoidal Endothelial Cell: Basic Biology and Pathobiology 425

Table 34.1  Ligands cleared from the blood circulation mainly by the LSECs

Ligand Receptor Reference
Tissue turnover waste products Hyaluronan Stabilin‐2 [14, 25]
Chondroitin sulfate Stabilin‐2 [14, 139]
Nidogen SR [140]
Heparin Stabilin‐2 [139, 141]
Serglycin SR [142]
N‐terminal propeptides of procollagen (I, III) SR, stabilin‐2 [25, 143]
Collagen alpha chains (I, II, III, IV, V, XI) Mannose receptora [15, 39, 144]
C‐terminal propeptide of procollagen type I Mannose receptor [36]
Tissue plasminogen activator Mannose receptor [37]
Lysosomal enzymes Mannose receptor [22]
Salivary amylase Mannose receptor [145]
Modified proteins and Oxidized LDL SR, stabilin‐1, stabilin‐2 [31, 146]
lipoproteins AGE‐albumin SR, stabilin‐2 (stabilin‐1b) [132, 147]
Immune complexes Soluble IgG immune complexes Fc‐γRIIb2 [44]
Ligands of non‐mammalian CpG oligodeoxynucleotides SR [33]
origin Invertase Mannose receptor [148]
Mannan Mannose receptor [41]
Ovalbumin Mannose receptor [21]
Ricin Mannose receptor [68]
Horseradish peroxidase Mannose receptor [149]
Aminated β(1–3) glucan Unknown [150]
Non‐physiological model ligands Formaldehyde‐treated albumin SR, stabilin‐1, stabilin‐2 [31, 151]
Acetylated LDL SR (stabilin‐1/ stabilin‐2b) [152]
Agalacto‐orosomucoid Mannose receptor [149]
Ahexosamino‐orosomucoid Mannose receptor [149]
Lithium carmine Unknown [18]
The table is modified from [11] with permission. SR, scavenger receptor; AGE, advanced glycation end‐product; LDL, low density lipoprotein. Stabilin‐1 synonyms: SR‐H1,
FEEL‐1 (fasciclin endothelial growth factor (EGF)‐like, laminin‐type EGF‐like and link domain‐containing scavenger receptor 1) [153], CLEVER‐1 (common lymphatic endothelial
and vascular endothelial receptor‐1) [154]. Stabilin‐2 synonyms: SR‐H2, FEEL‐2 [155], HA/SR (hyaluronan/SR)[25], HARE (hyaluronan receptor for endocytosis) [106].
a
 The reported LSEC receptor for collagen α‐chains is identical to the mannose receptor on these cells [39]. Collagen α‐chains and mannose‐terminated ligands have affinity to
non‐overlapping binding sites on the receptor [38].
b
 Affinity of AGE‐albumin to stabilin‐1, and acetylated LDL to stabilin‐1 and stabilin‐2 have been studied in transfected cell lines [153, 155].

vertebrates (i.e. mammals, birds, reptiles, and amphibia), whereas
in phylogenetically older vertebrates the scavenger endothelial
cells are located in heart atrium and/or kidney (bony fishes) or gills
(cartilaginous fishes, lamprey, and hagfish) [10].
Additional evidence that LSECs are geared
to effective uptake of blood‐borne
macromolecules and nanosized material
LSECs are packed with organelles involved in endocytosis, and
intracellular transport and endo‐lysosomal processing of inter-
nalized cargo. In spite of their low proportion of the total liver
cell volume (3.3%), LSECs contain an impressive 45% of the
organ’s total mass of pinocytic vesicles and 17% of the lysoso-
mal volume [13]. Moreover, LSECs, compared to hepatocytes
and Kupffer cells, contain twice as many clathrin‐coated pits per
membrane unit [19]. The fact that LSECs contain high amounts
of proteins involved in membrane traffic (clathrin, alpha and
beta‐adaptins, Rab4, Rab5, Rab7, and rabaptin‐5) [19, 20]
strengthens the notion that the cells are indeed geared to clath-
rin‐mediated endocytosis. Another factor that is rarely dis-
cussed, but nevertheless contributes to the superefficient
endocytosis in LSECs is the much more rapid receptor recycling
time in LSEC (seconds) [21] compared to most other cell types
(minutes). These features, along with the fact that the specific
activity of several lysosomal enzymes is higher in LSECs than
in hepatocytes and Kupffer cells [22, 23] indicate that LSECs
play a major role in the hepatic blood clearance function.
Receptors important for LSEC scavenger
function
Several LSEC receptor species have been reported to bind
blood‐borne macromolecules. In this section only major endo-
cytically active receptors known to play a chief role in the
LSEC‐mediated blood clearance will be highlighted
(Table 34.1). Other LSEC receptors for which important scav-
enger function is not obvious will be briefly mentioned at the
end of the section.
Scavenger receptors (SRs)
Scavenger receptors (SRs) represent important LSEC clearance
receptors. The concept of SR has evolved considerably since the
receptor for acetylated low‐density lipoprotein (LDL) was dis-
covered in macrophages and denoted “scavenger receptor” [24].
The common denominator of all SRs is their affinity for
­polyanionic ligands. The LSEC is reported to express SRs class
A, B, E, and H [11]. Of these, SR‐H1 (stabilin‐1) and SR‐H2
(stabilin‐2) are instrumental for LSEC‐mediated scavenging
[11]. In the liver stabilin‐2 is expressed on LSECs only, and thus
is a signature receptor than can be used as a specific LSEC
marker [25–28]. Interestingly the Stab2 gene is highly con-
served, and antibodies to human stabilin‐2 stain the same protein
in fish (Atlantic cod) [11]. Moreover, in zebrafish where the
Stab2 gene had been knocked out, clearance of ligands for stabi-
lin‐2 in scavenger endothelial cells was completely lost [29].
Blood‐borne ligands cleared by LSEC stabilins include several
426 THE LIVER:  THE LSEC SCAVENGER FUNCTION AND ITS IMPLICATIONS
molecules that are products of physiological turnover of connec-
tive tissue [2] (Table  34.1). Physiologically modified proteins
such as advanced glycation end products (AGEs) and oxidized
LDL (oxLDL) are also cleared by LSEC stabilins [30, 31]. Of
note, mildly oxLDL (the major circulating form of oxLDLs) is
endocytosed in LSECs only, mainly via stabilin‐1, whereas non‐
physiological heavily oxidized oxLDL is endocytosed by both
LSECs and Kupffer cells [31], pointing to the importance of
LSECs in preventing cholesterol accumulation. SR class B, type
I (SR‐BI), responsible for mediating the selective uptake of high‐
density lipoprotein (HDL) cholesteryl esters in hepatocytes, has
been identified in LSECs as well, where the receptor may be
responsible for the efflux of cellular cholesterol to HDL, deliver-
ing cholesterol from LSECs to hepatocytes [32]. Nucleotides, for
example, oligonucleotides containing unmethylated CpG motifs
[33], anti‐sense‐oligonucleotides [34], and plasmid DNA [35]
represent a different group of molecules that are cleared by SR‐
mediated uptake in LSECs.
The mannose receptor
The mannose receptor (MR, CD206, Mrc1), a 175–180 kDa
type I integral membrane protein belonging to the C‐type lec-
tin family, has traditionally been associated with macrophages.
However, several other cell types express this receptor, not the
least the LSECs, which is the main carrier of this receptor in
liver [2]. The importance of the LSEC MR is reflected by the
consistent observation that intravenously administered solu-
ble/macromolecular MR ligands are rapidly cleared from the
circulation mainly by uptake in the LSECs [11]. The reason
for the remarkably efficient MR‐mediated uptake by LSECs is
its rapid receptor recycling time (only 10 seconds) [21]. The
MR is highly versatile in its function as a clearance receptor:
several C‐type lectin‐like domains displayed by the receptor
protein are responsible for its binding to mannose and
N‐acetylglucosamine (GlcNAc), enabling LSECs to scavenge
an array of ­different macromolecules carrying mannose or
GlcNAc in the glycosylation terminal position. The binding
specificity of the MR is essentially independent of the class of
macromolecule that carries the glycosylation side‐chains.
Important physiological ligands include C‐terminal propep-
tides of collagens [36] and tissue plasminogen activator [37].
Interestingly, MR‐mediated uptake of blood‐borne lysosomal
enzymes represents a physiological route to maintain the high
specific activity of these enzymes in LSECs [22]. The MR also
expresses a fibronectin‐type II domain, which mediates the
scavenging of blood‐borne collagen alpha‐chains [38], result-
ing from physiological turnover processes of bone and other
connective tissues (estimated to be in the order of grams per
day) [39]. Yet a third ligand binding site on the MR is the
­outermost cysteine‐rich N‐terminal domain, which recognizes
specific sulfated sugars [40]. Cytokines and inflammatory
stimuli affect the expression of LSEC MR: upregulation has
been reported after exposure to IL‐1 [6, 41, 42], IL‐10, or IL‐4
in combination with IL‐13 [43]. In a mouse colon carcinoma
liver metastasis model MR expression was observed to increase
along the sinusoids, which was suggested to contribute to the
hepatic prometastatic effects of IL‐1, cyclic oxygenase‐2, and
ICAM‐1 [6].
The Fc‐gamma receptor IIb2
The Fc‐gamma receptor IIb2 (FcγRIIb2, CD32b) is highly and
uniquely expressed in LSECs in liver, enabling efficient endocy-
tosis of soluble IgG‐antigen complexes [44]. The first hint that
not only Kupffer cells, but also LSECs bind IgG immune com-
plexes was reported in the early 1980s, when it was found that
isolated LSECs formed rosettes with IgG‐coated red blood cells
[45]. Another study showed that LSECs at high capacity could
bind, but not phagocytose the same test particles [46]. The find-
ing that soluble IgG immune complexes could compete with
IgG‐red blood cells for binding to the LSECs strengthened the
notion that the binding of IgG‐red cells was FcγR‐mediated
[47]. At the same time it was reported that small circulating IgG
immune complexes were taken up in LSECs, and it was noted
that the relative proportions removed by Kupffer cells and
LSECs depended on the size of the immune complex: the larger
the complex the more was directed to the Kupffer cells [48]. Of
note, the LSECs account for at least 70% of the uptake of small
IgG immune complexes in mouse liver sinusoids [49]. Thus
FcγR‐mediated uptake agrees with the dual cell principle of
waste clearance, in that small soluble immune complexes are
cleared by LSECs, whereas larger immune complexes are endo-
cytosed by Kupffer cells. More recent studies have revealed that
FcγRIIb2 is the main receptor for the efficient uptake of immune
complexes in rat LSECs, via clathrin‐mediated endocytosis
[44]. The authors reported that the receptor was present on
LSECs only, but not in Kupffer cells or other types of liver cells.
Consistent with these findings in rats it has later been shown
that FcγRIIb2 in LSECs is responsible for the clearance of solu-
ble immune complexes in mice as well, with 75% and 90% of
total body and total liver FcγRIIb2 expressed in LSECs [50].
Due to the consistent observation that FcγRIIb2 removes blood‐
borne soluble immune complexes, several studies have impli-
cated the LSECs as important in the etiology of autoimmune
diseases, for example, systemic lupus erythematosus (SLE).
This will be dealt with in the section on the Role of LSECs in
the etiology of autoimmune disorders.
Other endocytosis receptors in LSECs
Compared to the relatively rich literature on LSEC SRs (stabi-
lin‐1 and stabilin‐2), MR, and FcγRIIb2, little is known about
the functional significance of other LSEC receptors, including
LYVE‐1 (lymphatic vessel endothelial hyaluronan receptor‐1)
[51], and the C‐type lectins L‐SIGN (liver/lymph node‐specific
ICAM‐3 grabbing nonintegrin; syn. DC‐SIGNR) and LSECtin
(liver and lymph node sinusoidal endothelial cell C‐type lectin;
syn. DC‐SIGN) [52, 53]. L‐SIGN is involved in recognition and
uptake of human immune deficiency virus and hepatitis C virus
in LSEC [52, 54], whereas LSECtin regulates viral immune
responses by interacting with L‐SIGN [55], and has also been
reported to have a role in adherence of cancer metastasis in liver
[7]. LYVE‐1 is a hyaluronan binding protein expressed in lym-
phatic vascular endothelium and sinusoidal endothelial cells of
lymph nodes, liver, and spleen [28]. In adult liver, LYVE‐1 is
only expressed in the LSECs [51], and is distributed all along
the sinusoids [56], making it a good LSEC marker. LYVE ‐1
clears hyaluronan from lymph, but its function in liver is unclear.
Stabilin‐2 shows a high capacity for hyaluronan‐uptake in
 34:  The Liver Sinusoidal Endothelial Cell: Basic Biology and Pathobiology 427
LSECs, and the relative contribution of LYVE‐1 and stabilin‐2
to this clearance is unknown.
LSEC markers
The gold standard of LSEC markers is the characteristic open
fenestrae that, until recently, could only be detected in the elec-
tron microscope [5]. In spite of the small pore size of the
fenestrae, preventing their detection in conventional light and
fluorescence microscopes, recent development has made it
­possible to observe fenestrae in specially equipped light micro-
scopes, for example, structured illumination microscopy, direct
stochastic optical reconstruction microscopy, atomic force
microscopy, and stimulated emission depletion (STED) micros-
copy [57–59]. Structured illumination microscopy, atomic force
microscopy, and STED microscopy offer the important possibil-
ity of studying the dynamics of the fenestrae opening and clos-
ing [57–59]. In spite of these new developments, the most
reliable way of quantifying the number, size, and membrane
organization of fenestrae is still examination in the electron
microscope. Apart from using the fenestration as an identifier,
LSECs may be distinguished by a set of markers that can be
detected in the light/fluorescence microscope. These markers
include antigens that are specifically expressed on the LSEC
surface and can be marked by specific antibodies. In addition,
LSEC signature endocytosis receptors may be utilized function-
ally to determine specific uptake of macromolecular ligands
known to be taken up by these receptors. It is however, utterly
important to be aware of the challenges of interpreting the
results using these markers, of which some are quite specific for
LSECs, whereas others are not [60]. Moreover, there is a clear
need to standardize methods for identification and purity assess-
ment of isolated LSECs, as well for identification of LSECs in
the intact liver. In a recently published update on LSECs it was
advocated that cell identification should include validation of
the identity of isolated cells assessed by both ultrastructure and
purity based on evidence of high‐affinity endocytosis of specific
ligands [61].
Von Willebrand factor, a marker of vascular endothelial
cells, is not normally expressed in LSECs in young individuals
[15], but may be expressed in old livers [62]. CD31 is expressed
in all endothelia, including human [63], mouse [64], and rat
LSECs [27], but at variance from other endothelia the expres-
sion of CD31 on the LSEC surface is low or varying in young
individuals, but upregulated in liver fibrosis [65]. Hence other
LSEC markers should be chosen. Stabilin‐2, FcγRIIb2, and
LYVE‐1 are uniquely expressed by LSECs in adult liver, and
are to date the most reliable surface markers available for
immune‐identification of LSECs [2]. Other proteins that can be
used for LSEC identification are VEGFR3 (Flt‐4) [66] and
coagulation factor VIII, the latter being uniquely synthesized
by LSEC in the liver [67]. The MR is highly expressed on
LSECs but also shows a variable expression in Kupffer cells,
which may be species dependent. MR expression has been
reported to be absent in human and mouse Kupffer cells [22,
28] and expressed at a low level in rat Kupffer cells [68] com-
pared to LSECs.
Since the functional hallmark of LSECs is their super active
and specific endocytosis via their signature clearance receptors,
a useful way of identifying the cells is to determine their ability
to accumulate ligands known to be recognized by any of these
receptors. To this end FITC‐labeled formaldehyde‐modified
serum albumin (FITC‐FSA), a ligand for stabilin‐1 and stabi-
lin‐2 [31], has been frequently used [60]. In liver FITC‐FSA
accumulates predominantly in LSECs when administered
intravenously at low dose (< 2 μg g−1 body weight), and moni-
tored within 10–15 minutes [69]. LSECs with an intact ability
to internalize bound ligand and process it intracellularly, will
accumulate the probe and stain positively. Hence, uptake of
FSA or other ligands that are specifically endocytosed via any
of the LSEC signature receptors will yield information on both
LSEC identity and integrity. To assess the identity and purity of
LSECs in vitro it is important to use FSA at low concentration
(< 10 μg mL−1) and short incubation times (10–30 minutes).
Unspecific staining of non‐LSECs may result if FITC‐FSA is
applied at higher concentrations or longer incubation times.
Uptake of acetylated LDL (AcLDL) is sometimes used to
identify LSECs. However, this ligand is taken up by many
types of endothelial cells and is not a totally reliable LSEC‐
specific probe [5].
LSEC heterogeneity of marker molecules across
the liver lobule
LSEC features such as cell size, number, and size of fenestra-
tion, expression of intracellular and surface markers, expres-
sion of lectins and von Willebrand factor, endocytosis, and
susceptibility to acetaminophen toxicity vary across the
hepatic lobule [61]. Lobular zonation has not been reported for
the LSEC markers stabilin‐2 and MR in human or rodent mod-
els, but heterogeneic expression along sinusoids in the human
liver was reported for CD32/FcγRIIb, ICAM‐1, and LYVE‐1,
with low or absent expression of these markers in acinar zone
1 [70]. The authors mentioned the possibility that the low
staining intensity of CD32/FcγRIIb in zone 1 may be the result
of systemic endothelial cells radiating from the portal tract. It
is important to know whether positive selection using an
LSEC‐specific marker like CD32b will exclude LSECs from
acinar zone 1, and if there are species differences. In rat,
FcγRIIb2 expression was shown to be continuous along the
sinusoids and similar to stabilin‐2 [71], when using the
FcγRIIb2‐specific SE‐1 antibody [72, 73]. In accordance with
this finding, studies in mice on binding of IgG immune com-
plexes to localize sinusoidal FcγR showed continuous pres-
ence of the probe along the sinusoidal wall, but no binding to
the vascular endothelium of the central and portal areas. The
finding that far more probe bound to LSECs than to Kupffer
cells, combined with present knowledge that FcγRIIb2 is the
only FcγR present on LSEC, indicate continuous presence of
FcγRIIb2 along the sinusoid [74, 75], but that the density of
receptors may vary in acinar zone 1.
The common leukocyte antigen CD45 is often used as a
negative selection marker when isolating human and mouse
LSECs by fluorescent‐ or magnetic‐activated cell sorting.
However, rat LSEC is reported to be CD45 positive [76–78]
with high expression in zone 1, low in zone 2, and negative in
zone 3 [76]. Studies in rat have also shown recruitment of
428 THE LIVER:  THE LSEC SCAVENGER FUNCTION AND ITS IMPLICATIONS
CD133+ CD45+ CD31+ bone marrow‐derived progenitor cells
for LSECs after liver injury [79, 80]. Using CD45 as a negative
selection criterion when isolating LSECs may therefore lead to
a biased selection of cells. Clearly, the CD45 expression pat-
tern in LSECs needs to be further investigated in different
species.
LSEC metabolism
Less than 20% of ATP generated by LSECs and Kupffer cells
from exogenous substrates is derived from glucose metabo-
lism, whereas the main energy sources in these cells are oxida-
tion of glutamine and palmitate [81]. Studies on the metabolic
fate of endocytosed macromolecules following their endocyto-
sis by LSECs revealed that lactate and acetate are major degra-
dation products that are transported out of the cells. It has been
speculated that the very high release of lactate from the LSECs
serves to fuel the ATP production in hepatocytes [82]. On a per
cell basis hepatocytes oxidize 136 times and 45 times more lac-
tate than do LSECs and KCs, meaning that practically all oxi-
dation of lactate in the liver is by the hepatocytes. This is in
accordance with the report that 98.8% of the mitochondria
in liver is in the hepatocytes, with 0.5% in LSECs and 0.4% in
Kupffer cells [13].
The tension of O2, an important modulator of metabolic pro-
cesses, ranges from 60–70 mmHg periportally to 25–35 mmHg
pericentrally. Most studies with cultured LSECs have been per-
formed in incubators using a standard gas mixture of 5% CO2
and 95% air, which represents 150 mmHg O2 (as in the hepatic
artery). Hence, cultivation of LSECs in a conventional incubator
represents a hyperoxic environment for these cells. A study
­performed to determine the effect of lowering O2 tension on
­primary cultured LSECs gave the following results [83]: (i) cell
longevity in culture increased, (ii) endocytic activity was pre-
served for longer periods, (iii) proinflammatory IL‐6 was
reduced, (iv) anti‐inflammatory cytokine IL‐10 was increased,
(v) production of lactate doubled, (vi) generation of H2O2 was
drastically reduced. Based on these results we and some other
laboratories routinely cultivate LSECs at normoxic conditions
(5%, or 52 mmHg O2).
Role of LSECs in host defense and immune
regulatory functions
Innate immune functions
PRRs that recognize pathogen associated molecular patterns
(PAMPs) or own molecules that may cause damage, called
damage‐associated molecular patterns (DAMPs), are hall-
marks of cells involved in innate immune functions [84]. PRRs
in LSECs include the MR and stabilin‐1 and ‐2, which mediate
the clearance of an array of PAMPs and DAMPs from the
­circulation. Additional PRRs expressed by LSECs are TLR2,
3, 4, 6/2, 8, and 9 [33, 85, 86]. Of note, LSECs produce cytokines
TNF‐alpha, IL‐6, and IL‐1beta when challenged with TLR
agonists. The inflammasome molecules NLRP‐1, NLRP‐3,
and AIM2 are highly expressed in both LSECs and Kupffer
cells, with much lower, or absent expression in other types of
liver cells [87].
Adaptive immune functions
Instead of eliciting an immune response in the liver intraportally
administered antigens results in specific tolerance [88]. Without
a mechanism that elicits tolerogenic response in the liver, the
many waste molecules from the intestines and the general circu-
lation would readily cause devastating immune responses in the
organ. Several studies have reported on the role of the LSECs in
the generation of hepatic immune tolerance [3, 89–92]. One
hypothesis suggests that the LSECs express MHC II molecules
enabling them to present exogenous antigens resulting in toler-
ance [92, 93]. However, some studies have failed to detect the
presence of MHC II in LSECs [47, 90]. Another hypothesis
advocates LSEC in cross‐presentation, where LSECs present
exogenous antigens via MHC I to CD8+ T helper cells [91].
Role of LSECs as a sink for blood‐borne virus
A vast number of virus particles gain access to our circulation
[94]. From the gut alone it is estimated that every day 31 × 109
phage particles find their way across the epithelial cell layer
[95]. Yet few virus particles accumulate in the circulation.
Hence there must be a very effective clearance mechanism elim-
inating these virus particles from the blood. It has been hypoth-
esized that blood‐borne virus is cleared by non‐phagocytic
uptake in LSECs, that are geared to endocytosis of material of
less than approx. 200 nm. And indeed, two recent studies have
shown that intravenously administered adenovirus and BK and
JC polyomavirus‐like particles are efficiently taken up in the
mouse liver, with LSECs as the main cellular site of uptake [69,
96] (Figure  34.2c). If uptake of blood‐borne virus in LSECs
turns out to represent a general route of virus elimination, it will
assign an important novel role for these cells in the antivirus
defense apparatus.
Studies on how duck hepatitis B virus infects hepatocytes have
shown that virus is first endocytosed by the LSECs, before a few
virus particles are rescued from the endo‐lysosomal route and trans-
ferred to the hepatocytes [97]. Future studies on the early interaction
of blood‐borne viruses with LSECs should be carried out to gain
more knowledge on how some viruses escape metabolism by
LSECs and produce infection of liver and other organs.
Origin and renewal of LSECs
There appears to be two populations of LSEC progenitor cells:
bone marrow‐derived progenitors and resident or intrahepatic
LSEC progenitors [79, 80]. Normal LSEC turnover is main-
tained by the liver resident population of LSEC progenitor cells;
in addition the liver is able to recruit bone marrow‐derived cells
to help replenish the LSEC population when needed [98].
After  induction of liver injury or partial hepatectomy LSEC
­repopulation with CD133+ CD45+ CD31+ bone marrow‐derived
progenitor occurred rapidly [77, 79, 80]. Hepatic VEGF regu-
lates the recruitment of bone marrow cells to liver [80].
Hepatocyte growth factor (HGF) is essential for liver regenera-
tion. In normal liver, the stellate cells are the main producer of
HGF, with little expression in LSECs. However, bone marrow
LSECs repopulating the liver after liver injury are rich in HGF,
suggesting an essential role of bone marrow‐derived cells in
liver regeneration [79].
 34:  The Liver Sinusoidal Endothelial Cell: Basic Biology and Pathobiology 429
PATHOBIOLOGY
In this part of the chapter we have chosen to focus mainly on
selected pathological conditions known to involve signature
structure and functions of LSECs, namely their fenestration and
their scavenger activity.
LSEC function in inflammation
and liver fibrosis
A recent, excellent review has outlined in detail how LSECs
influence the immune microenvironment within the liver and
discusses their contribution to immune‐mediated liver diseases
and the complications of fibrosis and carcinogenesis [99]. This
topic will therefore be only briefly discussed here.
Fibrosis is a reversible scarring process resulting from chronic
liver injury and characterized by excess deposition of extracellular
matrix (ECM) [100]. Stellate cells are the primary source of ECM
both in normal and fibrotic liver, whereas LSECs and Kupffer cells
stimulate this process through TGFβ‐activation and signaling, and
production of proinflammatory molecules [101, 102]. Liver fibrosis
is associated with decreased LSEC fenestration, and appearance of
an organized basal lamina in the space of Disse, a process called
capillarization. The capillarization of LSECs is reported to precede
the onset of liver fibrosis and suggested to act as a gatekeeper event
for the progression of fibrosis [5]. Normally differentiated LSECs
prevent hepatic stellate cell activation and promote reversion to qui-
escence, whereas capillarized LSECs do not [103].
In normal liver, LSECs are tolerogenic [91]. Following fibrotic
hepatotoxic injury, LSECs develop a proinflammatory phenotype
able to induce an immunogenic T cell phenotype and enhance cyto-
toxic T‐cell responses [64]. During the progression of liver fibrosis
the profile of adhesion molecules expressed by LSECs changes [99,
104]. In liver inflammation the majority of leukocytes adhere in
the sinusoids. Different from the general mechanism for leukocyte
recruitment in postcapillary venules, adhesion to the sinusoidal
endothelium is selectin independent [105] and adherence and trans-
migration occur via integrin‐dependent and integrin‐independent
mechanisms, involving VCAM‐1, ICAM‐1, MADCAM, VAP‐1,
hyaluronan, chemokine ligand, and stabilin‐mediated binding
of LSECs to leukocytes [99, 104].
Few studies have been reported on how inflammation affects
the LSEC endocytic function in vivo, or vice versa. Uptake of
hyaluronan, a ligand for stabilin‐2 [25, 106] was impaired in
cirrhotic rat liver [107], whereas a study in mouse showed
increased capacity of fibrotic LSECs to capture ­antigen (man-
nosylated albumin) via the MR, compared to LSECs from nor-
mal liver [64]. In accordance with this, MR expression and
function in LSEC was upregulated by cytokines involved in the
progression of inflammation and fibrosis, that is, IL1, IL‐10, or
IL‐4 in combination with IL‐13 [6, 41–43]. From this it may be
hypothesized that the effect of chronic inflammation on LSEC
endocytosis varies depending on the receptor involved.
Altered LSEC fenestration – a proposed cause
of atherosclerosis
Due to the fenestrae the LSEC lining functions like a semiperme-
able membrane between the sinusoidal blood and the hepato-
cytes, filtering blood‐borne material of size less than the diameter
of the fenestral pores and preventing larger particles from passing
through the pores. This filtering is responsible for the production
of 50% of all lymph produced in the body [108]. LSEC porosity
is key to understanding the lipoprotein traffic between the blood
and the hepatocytes [12]. Spherical particles of triglyceride‐rich
spherical lipoproteins, generated by intestinal epithelium from
dietary lipids, are too big (100–1000 nm) to pass through the
fenestral pores [109], and thus stay in the circulation. The 30–80
nm chylomicron remnant particles resulting from size reduction
by lipoprotein lipase at the surface of extrahepatic endothelium,
pass through the LSEC fenestrae, and are endocytosed by the
hepatocytes [110]. This explains why reduced porosity, as in liver
cirrhosis, diabetes mellitus, and old age leads to prolonged post-
prandial lipoproteinemia and increased circulatory cholesterol
levels, with consequential increased risk for development of
atherosclerosis [110]. Due to the apparent correlation between
reduced LSEC porosity and probability of developing atheroscle-
rosis, effort has been invested into studying the dynamics and
molecular structure of the pores. Cholinergic agonists, vasoactive
intestinal peptide, glucagon, and increased intrasinusoidal blood
pressure dilate fenestrae, whereas alpha‐adrenergic agonists and
serotonin give constriction [12, 111]. Such studies are also
­anticipated to yield results that can help controlling the delivery
of drugs to hepatocytes.
Role of LSECs in liver toxicology
and drug distribution
Although hepatotoxicity is generally associated with the effect
of toxic compounds on the hepatocytes, LSECs are sometimes
the initial target of injury in a condition referred to as sinusoidal
obstruction syndrome (SOS, formerly hepatic veno‐occlusive
disease, VOD). SOS denotes a change of the sinusoid that may
lead to hepatocyte hypoxia, with liver dysfunction and disruption
of the portal circulation [112]. In the two major causes of SOS,
dietary ingestion of pyrrolizidine alkaloids and chemo‐irradiation‐
induced injury, the LSEC is primarily affected. The chemo-
therapeutic drugs most clearly associated with SOS at conventional
doses are gemtuzumab ozogamicin, a­ctinomycin D, dacar-
bazine, cytosine arabinoside, mithramycin, 6‐­thioguanine, and
urethane. Another common cause of SOS is the myeloablative
conditioning therapy used to prepare for hematopoietic stem
cell  transplantation for malignancy, particularly when cyclo-
phosphamide is included. A hallmark of SOS is that circulatory
disruption precedes hepatocyte failure. In vitro studies show
that drugs and toxins implicated in SOS can be detoxified by
glutathione and are selectively more toxic to LSECs than
to  hepatocytes. Studies with dacarbazine and monocrotaline
(a pyrrolizidine alkaloid) have shown that the selective toxicity
to LSECs is related to metabolic activation of the drug.
Cyclophosphamide, a commonly used drug in preparative
­regimens for myeloablative hematopoietic stem cell transplanta-
tion, has among the highest incidences of SOS. The drug must be
activated by P450 in hepatocytes to be toxic to LSECs.
The sequence of events from administration of monocrotaline
until fulminant SOS is similar to the hepatotoxic action of aceta-
minophen [98]. In both cases initial LSEC injury precedes
hepatocyte necrosis. The initial histological manifestation of
both is centrilobular hemorrhagic necrosis, but in SOS sinusoi-
dal fibrosis evolves in the late phase, whereas acetaminophen
430 THE LIVER:  ACKNOWLEDGMENT

toxicity does not have chronic histological sequelae. The fol- are cleared in the mouse mainly via the FcγRIIb2 of LSECs
lowing important biochemical features are shared between SOS [44], along with the observation that deletion of same receptor
and acetaminophen toxicity: preserving NO levels is protective, causes spontaneous autoimmunity and SLE‐like disease in mice
and inhibition of MMP‐2 and MMP‐9 reduces circulatory injury [120], point to a pivotal role of LSEC FcγRIIb2 in the disease
and subsequent parenchymal necrosis. Of note, SOS injury lasts mechanism of SLE. Moreover, the finding that scavenging of
much longer and may have higher fatality. blood‐borne DNAs is chiefly by SR‐mediated uptake in LSECs
The mechanism of hepatoxicity caused by high doses of aceta- [35], along with the fact that SLE is associated with generation
minophen has been studied in mice [8]. Soon after administration of anti‐DNA antibody, lends additional support to the hypothe-
of acetaminophen LSECs swell and begin to lose their scavenger sis that LSECs participate in the onset of SLE.
activity, as probed with a ligand for stabilin‐1/2. Already after two
hours gaps through the LSEC are formed by the destruction of LSEC function in old liver
fenestrae. These events take place before any evidence of injury to
Since early 2000, a series of studies have reported distinct age‐
parenchymal cells. The sinusoid then collapses, reducing the
related changes in the morphology and function of the hepatic
blood flow. Ethanol binged animals subsequently treated with
microvasculature of human, non‐human primates, and rodents
acetaminophen show the most severe LSEC damage.
[62]. Aging is associated with increased endothelial thickness and
The observation that serum hyaluronan levels increased in SOS
reduced LSEC porosity, and the gradual formation of an organ-
following bone marrow transplantation [113] or monocrotaline
ized basal lamina in the space of Disse [62, 121–128]. Most stud-
treatment [114] suggests either decreased clearance or increased
ies also report LSEC upregulation of von Willebrand factor [62],
production of hyaluronan. Similarly, severe hepatotoxicity due to
an antigen not normally expressed in LSECs of young livers [15].
acetaminophen ingestion correlated with increased serum hyalu-
A common finding is also the presence of highly fat‐filled stellate
ronan [115]. Moreover, SOS following bone marrow transplanta-
cells [62]. Importantly, in primary liver aging the stellate cells
tion showed significantly increased serum tissue plasminogen
remain quiescent, in contrast to liver fibrosis where these cells are
activator (tPA) [116]. Since circulating hyaluronan and tPA are
highly activated. The specific change in sinusoidal morphology
normally cleared by stabilin‐1/2 and MR mainly in LSECs, it is
due to high age has been termed age‐related pseudocapillarization
not far‐fetched to hypothesize that increased serum levels of these
[123, 124], to distinguish it from capillarization associated with
molecules mainly reflect decreased uptake due to LSEC injuries.
liver fibrosis. Postprandial hypertriglyceridemia linked to
The finding that clearance of intravenously administered marker
decreased LSEC porosity, a common condition in old age which
ligands for the LSEC stabilin‐1/2 was decreased in mice following
predisposes for the development of atherosclerosis, has been dealt
acetaminophen treatment strengthens this notion [8].
with in the section on Altered LSEC fenestration  –  a proposed
Common to many therapeutic IgG immunoglobulins is their ulti-
cause of atherosclerosis [110, 124].
mate clearance mediated by the LSEC FcγRIIb2. This pathway may
Aging also affects the LSEC scavenger function [128]. Most
result in the delivery of high concentrations of the immunoconjugate
studies show impaired endocytic capacity of LSECs at old age
and the cytotoxic agent directly to LSECs, which is why LSEC tox-
[127–130] whereas no changes were reported in [131].
icity may be observed with these compounds. Other members of the
Of note, several of the substances normally cleared by
new generation of drugs include to a large extent other types of large
LSECs, such as AGE‐modified proteins [132, 133] and oxidized
molecule compounds and nano formulations, of which many must
LDLs [31, 134] are proinflammatory, and thus have to be effec-
be administered intravascularly. Since LSECs are geared to active
tively eliminated. Reduced LSEC endocytosis as seen with
uptake of this group of compounds [2] it is not surprising that these
aging may therefore increase the risk of vascular pathologies
cells may be easily intoxicated by off‐target mechanisms, causing
within and outside the liver.
subsequent hepatotoxicity [117]. Hence it is utterly important that
The underlying mechanism of the observed changes in the sinu-
developers of large molecule drugs and nano formulations, as well
soids of old liver is not clear, but long‐term effects of gut‐derived
as regulatory authorities take into consideration the possibility that
toxins, several of them known to have negative effects on LSEC
hepatotoxicity elicited by these compounds may be initiated by inju-
fenestration, have been suggested as a driver of the age‐related
ries to LSEC due to their tendency to accumulate very high concen-
pseudocapillarization [62]. A recent study on the effect of dietary
trations of drugs via their specialized clearance receptors.
macronutrients on the liver microcirculation in ad lib fed mice
suggests that diet influences LSEC fenestration, and both diet and
Role of LSECs in the etiology of autoimmune the gut microbiome may have implications for liver function into
disorders old age [135]. Aging in liver may also be linked to the general
Reduced Fc receptor function, and the resulting formation of increase in age‐related inflammatory markers [136], which affect
soluble immune complexes may be important in the etiology of the vasculature of various organs [137], including the hepatic sinu-
autoimmune diseases such as systemic lupus erythematosus soids [127], whereas caloric restriction, which reduces oxidative
(SLE) and Sjögren’s syndrome [118]. The decreased sequestra- cell stress, prevents age‐related LSEC defenestration [138].
tion of immune complexes increases the probability of their
deposition in the kidney or other vulnerable parenchymal
organs, which may be detrimental to the host. Indeed, abnor- ACKNOWLEDGMENT
malities in the clearance of and inflammatory response to
immune complexes are characteristic features of SLE [119]. Jaione Simon‐Santamaria is highly acknowledged for help with
The finding that small soluble IgG‐antigen immune complexes the figures.
 34:  The Liver Sinusoidal Endothelial Cell: Basic Biology and Pathobiology 431

REFERENCES 24. Goldstein, J.L., Ho, Y.K., Basu, S.K., and Brown, M.S. Binding site on mac-
rophages that mediates uptake and degradation of acetylated low density
lipoprotein, producing massive cholesterol deposition. Proc Natl Acad Sci U
1. Wisse, E. An ultrastructural characterization of the endothelial cell in the rat
S A, 1979;76:333–7.
liver sinusoid under normal and various experimental conditions, as a contri-
25. McCourt, P.A., Smedsrød, B.H., Melkko, J., and Johansson, S.
bution to the distinction between endothelial and Kupffer cells. J Ultrastruct
Characterization of a hyaluronan receptor on rat sinusoidal liver endothelial
Res, 1972;38(5):528–62.
cells and its functional relationship to scavenger receptors. Hepatology,
2. Sørensen, K.K., Simon‐Santamaria, J., McCuskey R.S., and Smedsrød, B.
1999;30(5):1276–86.
Liver sinusoidal endothelial cells. Compr Physiol, 2015;5(4):1751–74.
26. Schledzewski, K., Geraud, C., Arnold, B. et al. Deficiency of liver sinusoidal
3. Knolle, P.A. and Wohlleber, D. Immunological functions of liver sinusoidal
scavenger receptors stabilin‐1 and ‐2 in mice causes glomerulofibrotic
endothelial cells. Cell Mol Immunol, 2016;13(3):347–53.
nephropathy via impaired hepatic clearance of noxious blood factors. J Clin
4. Ni Y., Li J.M., Liu, M.K. et  al. Pathological process of liver sinusoidal
Invest, 2011;121(2):703–14.
endothelial cells in liver diseases. World J Gastroenterol, 2017;23(43):
27. Geraud, C., Schledzewski, K., Demory, A. et  al. Liver sinusoidal
7666–77.
­endothelium: a microenvironment‐dependent differentiation program in rat
5. DeLeve, L.D. Liver sinusoidal endothelial cells in hepatic fibrosis.
including the novel junctional protein liver endothelial differentiation‐­
Hepatology, 2015;61(5):1740–6.
associated protein‐1. Hepatology, 2010;52(1):313–26.
6. Arteta, B., Lasuen, N., Lopategi, A., Sveinbjörnsson, B., Smedsrød, B., and
28. Martens, J.H., Kzhyshkowska, J., Falkowski‐Hansen, M. et al. Differential
Vidal‐Vanaclocha, F. Colon carcinoma cell interaction with liver sinusoidal
expression of a gene signature for scavenger/lectin receptors by endothelial
endothelium inhibits organ‐specific antitumor immunity through interleu-
cells and macrophages in human lymph node sinuses, the primary sites of
kin‐1‐induced mannose receptor in mice. Hepatology, 2010;51(6):2172–82.
regional metastasis. J Pathol, 2006;208(4):574–89.
7. Zuo, Y., Ren, S., Wang, M. et al. Novel roles of liver sinusoidal endothelial
29. Campbell, F., Bos, F.L., Sieber, S. et al. Directing nanoparticle biodistribu-
cell lectin in colon carcinoma cell adhesion, migration and in‐vivo metastasis
tion through evasion and exploitation of Stab2‐dependent nanoparticle
to the liver. Gut, 2013;62(8):1169–78.
uptake. ACS Nano, 2018;12(3):2138–50.
8. McCuskey, R.S. Sinusoidal endothelial cells as an early target for hepatic
30. Hansen, B., Longati, P., Elvevold, K. et al. Stabilin‐1 and stabilin‐2 are both
toxicants. Clin Hemorheol Microcirc, 2006;34(1–2):5–10.
directed into the early endocytic pathway in hepatic sinusoidal endothelium
9. Pertoft, H. and Smedsrød, B. Separation and characterization of liver cells, in
via interactions with clathrin/AP‐2, independent of ligand binding. Exp Cell
Cell Separation: Methods and Selected Applications. 4, (eds. T.G. Pretlow
Res, 2005;303(1):160–73.
and T.P. Pretlow), Academic Press, New York, 1987, pp. 1–24.
31. Li, R., Oteiza, A., Sørensen, K.K. et al. Role of liver sinusoidal endothelial
10. Seternes, T., Sørensen, K., and Smedsrød, B. Scavenger endothelial cells
cells and stabilins in elimination of oxidized low‐density lipoproteins. Am J
of vertebrates: a nonperipheral leukocyte system for high‐capacity elimi-
Physiol Gastrointest Liver Physiol, 2011;300(1):G71–81.
nation of waste macromolecules. Proc Natl Acad Sci U S A, 2002;99(11):
32. Malerod, L., Juvet, K., Gjoen, T., and Berg, T. The expression of scavenger
7594–7.
receptor class B type I (SR‐BI) and caveolin‐1 in parenchymal and non-
11. Sørensen, K.K., McCourt, P., Berg, T. et al. The scavenger endothelial cell: a
parenchymal liver cells. Cell Tissue Res, 2002;307(2):173–80.
new player in homeostasis and immunity. Am J Physiol Regul Integr Comp
33. Martin‐Armas, M., Simon‐Santamaria, J., Pettersen, I., Moens, U.,
Physiol, 2012;303(12):R1217–30.
Smedsrød, B., and Sveinbjørnsson, B. Toll‐like receptor 9 (TLR9) is present
12. Cogger, V.C. and Le Couteur, D.G. Fenestrations in the liver sinusoidal
in murine liver sinusoidal endothelial cells (LSECs) and mediates the effect
endothelial cell, in The Liver: Biology and Pathobiology, 5th edn, (ed. I.M.
of CpG‐oligonucleotides. J Hepatol, 2006;44(5):939–46.
Arias) Wiley‐Blackwell; 2009.
34. Bijsterbosch, M.K., Manoharan, M., Rump, E.T. et al. In vivo fate of phos-
13. Blouin, A., Bolender, R.P. and Weibel, E.R. Distribution of organelles and
phorothioate antisense oligodeoxynucleotides: predominant uptake by scav-
membranes between hepatocytes and nonhepatocytes in the rat liver paren-
enger receptors on endothelial liver cells. Nucleic Acids Res,
chyma. A stereological study. J Cell Biol, 1977;72(2):441–55.
1997;25(16):3290–6.
14. Smedsrød, B., Pertoft, H., Eriksson, S., Fraser, J.R., and Laurent, T.C.
35. Hisazumi, J., Kobayashi, N., Nishikawa, M., and Takakura, Y. Significant
Studies in vitro on the uptake and degradation of sodium hyaluronate in rat
role of liver sinusoidal endothelial cells in hepatic uptake and degradation of
liver endothelial cells. Biochem J, 1984;223(3):617–26.
naked plasmid DNA after intravenous injection. Pharm Res, 2004;21(7):
15. Smedsrød, B., Pertoft, H., Gustafson, S., and Laurent, T.C. Scavenger func-
1223–8.
tions of the liver endothelial cell. Biochem J, 1990;266(2):313–27.
36. Smedsrød, B., Melkko, J., Risteli, L., and Risteli, J. Circulating C‐terminal
16. Aschoff, L. Das reticulo‐endotheliale System. Ergebnisse die innere
propeptide of type I procollagen is cleared mainly via the mannose receptor
Medizin, 1924;26:1–118.
in liver endothelial cells. Biochem J, 1990;271(2):345–50.
17. Metchnikoff, E. Uber eine Sprosspilzkrankheit der Daphnien; Beitrag zur
37. Smedsrød, B., Einarsson, M., and Pertoft, H. Tissue plasminogen activator is
Lehre uber den Kampf der Phagocyten gegen Krankheitserreger. Virchows
endocytosed by mannose and galactose receptors of rat liver cells. Thromb
Archiv fur pathologische Anatomie und Physiologie, und fur klinische
Haemost, 1988;59(3):480–4.
Medicin, Berlin. 1884;96:177–95.
38. Napper, C.E., Drickamer, K., and Taylor, M.E. Collagen binding by the man-
18. Kawai, Y., Smedsrød, B., Elvevold, K., and Wake, K. Uptake of lithium car-
nose receptor mediated through the fibronectin type II domain. Biochem J,
mine by sinusoidal endothelial and Kupffer cells of the rat liver: new insights
2006;395(3):579–86.
into the classical vital staining and the reticulo‐endothelial system. Cell
39. Malovic, I., Sørensen, K.K., Elvevold, K.H. et al. The mannose receptor on
Tissue Res, 1998;292(2):395–410.
murine liver sinusoidal endothelial cells is the main denatured collagen
19. Kjeken, R., Mousavi, S.A., Brech, A., Gjøen, T., and Berg, T. Fluid phase
clearance receptor. Hepatology, 2007;45(6):1454–61.
endocytosis of I‐125 iodixanol in rat liver parenchymal, endothelial and
40. Fiete, D.J., Beranek, M.C., and Baenziger, J.U. A cysteine‐rich domain of
Kupffer cells. Cell Tissue Res, 2001;304(2):221–30.
the “mannose” receptor mediates GalNAc‐4‐SO4 binding. Proc Natl Acad
20. Juvet, L.K., Berg, T., and Gjøen, T. The expression of endosomal rab pro-
Sci U S A, 1998;95(5):2089–93.
teins correlates with endocytic rate in rat liver cells. Hepatology,
41. Asumendi, A., Alvarez, A., Martinez, I., Smedsrod, B., and Vidal‐
1997;25(5):1204–12.
Vanaclocha, F. Hepatic sinusoidal endothelium heterogeneity with respect
21. Magnusson, S. and Berg, T. Extremely rapid endocytosis mediated by the
to  mannose receptor activity is interleukin‐1 dependent. Hepatology,
mannose receptor of sinusoidal endothelial rat liver cells. Biochem J,
1996;23(6):1521–9.
1989;257(3):651–6.
42. Martinez, I., Sveinbjørnsson, B., Vidal‐Vanaclocha, F., Asumendi, A., and
22. Elvevold, K., Simon‐Santamaria, J., Hasvold, H., McCourt P., Smedsrød, B.,
Smedsrød, B. Differential cytokine‐mediated modulation of endocytosis in
and Sorensen, K.K. Liver sinusoidal endothelial cells depend on mannose
rat liver endothelial cells. Biochem Biophys Res Commun, 1995;212(1):
receptor‐mediated recruitment of lysosomal enzymes for normal degradation
235–41.
capacity. Hepatology, 2008;48(6):2007–15.
43. Liu, Y., Gardner, C.R., Laskin, J.D., and Laskin, D.L. Classical and alterna-
23. Knook, D.L. and Sleyster, E.C. Isolated parenchymal, Kupffer and endothe-
tive activation of rat hepatic sinusoidal endothelial cells by inflammatory
lial rat liver cells characterized by their lysosomal enzyme content. Biochem
stimuli. Exp Mol Pathol, 2013;94(1):160–7.
Biophys Res Commun, 1980;96(1):250–7.
432 THE LIVER:  REFERENCES
44. Mousavi, S.A., Sporstol, M., Fladeby, C., Kjeken, R., Barois, N., and Berg,
T. Receptor‐mediated endocytosis of immune complexes in rat liver
­sinusoidal endothelial cells is mediated by FcgammaRIIb2. Hepatology,
2007;46(3):871–84.
45. Pulford, K. and Souhami, R.L. The surface properties and antigen‐presenting
function of hepatic non‐parenchymal cells. Clin Exp Immunol,
1981;46(3):581–8.
46. Smedsrød, B., Eriksson, S., Fraser, J.R.E., Laurent, T.C., and Pertoft, H.
Properties of liver endothelial cells in primary monolayer cultures, in
Sinusoidal Liver Cells, (eds. D.L. Knook and E. Wisse), Elsevier, Amsterdam,
1982, pp. 263–70.
47. Smedsrød, B., Pertoft, H., Eggertsen, G., and Sundstrom, C. Functional and
morphological characterization of cultures of Kupffer cells and liver
endothelial cells prepared by means of density separation in Percoll, and
selective substrate adherence. Cell Tissue Res, 1985;241(3):639–49.
48. Skogh, T., Blomhoff, R., Eskild, W., and Berg, T. Hepatic uptake of circulat-
ing IgG immune complexes. Immunology, 1985;55(4):585–94.
49. Kosugi, I., Muro, H., Shirasawa, H., Hirano, M., Amashita, Y., and
Miyakawa, A. Effects of the subcutaneous injection of complete Freund’s
adjuvant on Fc receptor activity and IgG immune complex uptake in liver
sinusoidal endothelial cells, in Cells of the Hepatic Sinusoid, (eds. E. Wisse,
and D.L. Knook,), Kupffer Cell Foundation, Leiden, The Netherlands, 1993,
pp. 434–7.
50. Ganesan, L.P., Kim, J., Wu, Y. et  al. FcgammaRIIb on liver sinusoidal
endothelium clears small immune complexes. J Immunol, 2012;189(10):
4981–8.
51. Mouta Carreira, C., Nasser, S.M., di Tomaso, E. et  al. LYVE‐1 is not
restricted to the lymph vessels: expression in normal liver blood sinusoids
and downregulation in human liver cancer and cirrhosis. Cancer Res,
2001;61(22):8079–84.
52. Pohlmann, S., Soilleux, E.J., Baribaud, F. et  al. DC‐SIGNR, a DC‐SIGN
homologue expressed in endothelial cells, binds to human and simian immu-
nodeficiency viruses and activates infection in trans. Proc Natl Acad Sci U S
A, 2001;98(5):2670–5.
53. Zhao, D., Zhang, M., Wang, M. et al. Up‐regulation of Cbl‐b is associated
with LSECtin‐mediated inhibition of different CD4+ T‐cell subsets.
Immunobiology, 2013;218(4):602–8.
54. Gardner, J.P., Durso, R.J., Arrigale, R.R. et al. L‐SIGN (CD 209L) is a liver‐
specific capture receptor for hepatitis C virus. Proc Natl Acad Sci U S A,
2003;100(8):4498–503.
55. Li, Y., Hao, B., Kuai, X. et  al. C‐type lectin LSECtin interacts with DC‐
SIGNR and is involved in hepatitis C virus binding. Mol Cell Biochem,
2009;327(1–2):183–90.
56. Nonaka, H., Tanaka, M., Suzuki, K., and Miyajima, A. Development of
murine hepatic sinusoidal endothelial cells characterized by the expression
of hyaluronan receptors. Dev Dyn, 2007;236(8):2258–67.
57. Oie, C.I., Monkemoller, V., Hubner, W. et al. New ways of looking at very
small holes – using optical nanoscopy to visualize liver sinusoidal endothe-
lial cell fenestrations. Nanophotonics‐Berlin, 2018;7(3):575–96.
58. Braet, F., Taatjes, D.J., and Wisse, E. Probing the unseen structure and func-
tion of liver cells through atomic force microscopy. Semin Cell Dev Biol,
2018;73:13–30.
59. Di Martino, J.D., Mascalchi, P., Legros, P. et al. STED microscopy: A sim-
plified method for liver sinusoidal endothelial fenestrae analysis. Biol Cell,
2018;110(7):159–68.
60. Elvevold, K., Smedsrød, B., and Martinez, I. The liver sinusoidal endothelial
cell: a cell type of controversial and confusing identity. Am J Physiol
Gastrointest Liver Physiol, 2008;294(2):G391–400.
61. DeLeve, L.D. and Maretti‐Mira, A.C. Liver sinusoidal endothelial cell: an
update. Semin Liver Dis, 2017;37(4):377–87.
62. Le Couteur, D.G., Warren, A., Cogger, V.C. et al. Old age and the hepatic
sinusoid. Anat Rec (Hoboken), 2008;291(6):672–83.
63. Lalor, P.F., Lai, W.K., Curbishley, S.M., Shetty, S., and Adams, D.H. Human
hepatic sinusoidal endothelial cells can be distinguished by expression of
phenotypic markers related to their specialised functions in vivo. World
J Gastroenterol, 2006;12(34):5429–39.
64. Connolly, M.K., Bedrosian, A.S., Malhotra, A. et al. In hepatic fibrosis, liver
sinusoidal endothelial cells acquire enhanced immunogenicity. J Immunol,
2010;185(4):2200–8.
65. DeLeve, L.D., Wang, X., Hu, L., McCuskey, M.K., and McCuskey, R.S. Rat
liver sinusoidal endothelial cell phenotype is maintained by paracrine and
autocrine regulation. Am J Physiol Gastrointest Liver Physiol, 2004;287(4):
G757–63.
66. Ding, B.S., Nolan, D.J., Butler, J.M. et al. Inductive angiocrine signals from
sinusoidal endothelium are required for liver regeneration. Nature,
2010;468(7321):310–5.
67. Shahani, T., Covens, K., Lavend’homme, R. et al. Human liver sinusoidal
endothelial cells but not hepatocytes contain factor VIII. J Thromb Haemost,
2014;12(1):36–42.
68. Magnusson, S. and Berg, T. Endocytosis of ricin by rat liver cells in vivo and
in vitro is mainly mediated by mannose receptors on sinusoidal endothelial
cells. Biochem J, 1993;291(3):749–55.
69. Simon‐Santamaria, J., Rinaldo, C.H., Kardas, P. et  al. Efficient uptake of
blood‐borne BK and JC polyomavirus‐like particles in endothelial cells of
liver sinusoids and renal vasa recta. PLoS One, 2014;9(11):e111762.
70. Strauss, O., Phillips, A., Ruggiero, K., Bartlett, A., and Dunbar, P.R.
Immunofluorescence identifies distinct subsets of endothelial cells in the
human liver. Sci Rep, 2017;7:44356.
71. Yoshida, M., Nishikawa, Y., Omori, Y. et  al. Involvement of signaling of
VEGF and TGF‐beta in differentiation of sinusoidal endothelial cells during
culture of fetal rat liver cells. Cell Tissue Res, 2007;329(2):273–82.
72. March, S., Hui, E.E., Underhill, G.H., Khetani, S., and Bhatia, S.N.
Microenvironmental regulation of the sinusoidal endothelial cell phenotype
in vitro. Hepatology, 2009;50(3):920–8.
73. Ohmura, T., Enomoto, K., Satoh, H., Sawada, N., and Mori, M. Establishment
of a novel monoclonal antibody, SE‐1, which specifically reacts with rat
hepatic sinusoidal endothelial cells. J Histochem Cytochem, 1993;41(8):
1253–7.
74. Muro, H., Shirasawa, H., Takahashi, Y., Maeda, M., and Nakamura, S.
Localization of Fc receptors on liver sinusoidal endothelium. A histological
study by electron microscopy. Acta Pathol Jpn, 1988;38(3):291–301.
75. Kosugi, I., Muro, H., Shirasawa, H., and Ito, I. Endocytosis of soluble IgG
immune complex and its transport to lysosomes in hepatic sinusoidal
endothelial cells. J Hepatol,. 1992;16(1–2):106–14.
76. Xie, G., Wang, L., Wang, X., and DeLeve, L.D. Isolation of periportal, mid-
lobular, and centrilobular rat liver sinusoidal endothelial cells enables study
of zonated drug toxicity. Am J Physiol Gastrointest Liver Physiol,
2010;299(5):G1204–10.
77. Harb, R., Xie, G., Lutzko, C. et al. Bone marrow progenitor cells repair rat
hepatic sinusoidal endothelial cells after liver injury. Gastroenterology
2009;137(2):704–12.
78. McCloskey, T.W., Todaro, J.A., and Laskin, D.L. Lipopolysaccharide treat-
ment of rats alters antigen expression and oxidative metabolism in hepatic
macrophages and endothelial cells. Hepatology, 1992;16(1):191–203.
79. Wang, L., Wang, X., Xie, G., Wang, L., Hill, C.K., and DeLeve, L.D. Liver
sinusoidal endothelial cell progenitor cells promote liver regeneration in rats.
J Clin Invest, 2012;122(4):1567–73.
80. Wang, L., Wang, X., Wang, L. et  al. Hepatic vascular endothelial growth
factor regulates recruitment of rat liver sinusoidal endothelial cell progenitor
cells. Gastroenterology, 2012;143(6):1555–63.
81. Spolarics, Z., Lang, C.H., Bagby, G.J., and Spitzer, J.J. Glutamine and fatty
acid oxidation are the main sources of energy for Kupffer and endothelial
cells. Am J Physiol, 1991;261(2 Pt 1):G185–90.
82. Smedsrød, B. Cellular events in the uptake and degradation of hyaluronan.
Adv Drug Deliv Rev, 1991;7:265–78.
83. Martinez, I., Nedredal, G.I., Øie, C.I. et al. The influence of oxygen tension
on the structure and function of isolated liver sinusoidal endothelial cells.
Comp Hepatol, 2008;7:4.
84. Matzinger, P. Tolerance, danger, and the extended family. Annu Rev Immunol,
1994;12:991–1045.
85. Wu, J., Meng, Z., Jiang, M. et al. Toll‐like receptor‐induced innate immune
responses in non‐parenchymal liver cells are cell type‐specific. Immunology,
2010;129(3):363–74.
86. Uhrig, A., Banafsche, R., Kremer, M. et  al. Development and functional
­consequences of LPS tolerance in sinusoidal endothelial cells of the liver.
J Leukoc Biol, 2005;77(5):626–33.
87. Boaru, S.G., Borkham‐Kamphorst, E., Tihaa, L., Haas, U., and Weiskirchen,
R. Expression analysis of inflammasomes in experimental models of inflam-
matory and fibrotic liver disease. J Inflamm (Lond), 2012;9(1):49.
88. Triger, D.R., Cynamon, M.H., and Wright, R. Studies on hepatic uptake of
antigen. I. Comparison of inferior vena cava and portal vein routes of immu-
nization. Immunology, 1973;25(6):941–50.
 34:  The Liver Sinusoidal Endothelial Cell: Basic Biology and Pathobiology 433
89. Crispe, I.N., Giannandrea, M., Klein, I., John, B., Sampson, B., and
Wuensch, S. Cellular and molecular mechanisms of liver tolerance.
Immunol Rev, 2006;213:101–18.
90. Katz, S.C., Pillarisetty, V.G., Bleier, J.I., Shah, A.B., and DeMatteo, R.P.
Liver sinusoidal endothelial cells are insufficient to activate T cells.
J Immunol, 2004;173(1):230–5.
91. Limmer, A., Ohl, J., Kurts, C. et  al. Efficient presentation of exogenous
­antigen by liver endothelial cells to CD8+ T cells results in antigen‐specific
T‐cell tolerance. Nat Med, 2000;6(12):1348–54.
92. Thomson, A.W. and Knolle, P.A. Antigen‐presenting cell function in the
tolerogenic liver environment. Nat Rev Immunol, 2010;10(11):753–66.
93. Crispe, I.N. Liver antigen‐presenting cells. J Hepatol, 2011;54(2):357–65.
94. Virgin, H.W. The virome in mammalian physiology and disease. Cell,
2014;157(1):142–50.
95. Nguyen, S., Baker, K., Padman, B.S. et  al. Bacteriophage transcytosis
­provides a mechanism to cross epithelial cell layers. MBio, 2017;8(6).
96. Ganesan, L.P., Mohanty, S., Kim, J., Clark, K.R., Robinson, J.M., and
Anderson, C.L. Rapid and efficient clearance of blood‐borne virus by liver
sinusoidal endothelium. PLoS Pathog, 2011;7(9):e1002281.
97. Breiner, K.M., Schaller, H., and Knolle, P.A. Endothelial cell‐mediated
uptake of a hepatitis B virus: a new concept of liver targeting of hepato-
tropic microorganisms. Hepatology, 2001;34(4 Pt 1):803–8.
98. DeLeve, L.D. Liver sinusoidal endothelial cells and liver regeneration.
J Clin Invest, 2013;123(5):1861–6.
99. Shetty, S., Lalor, P.F., and Adams, D.H. Liver sinusoidal endothelial
cells – gatekeepers of hepatic immunity. Nat Rev Gastroenterol Hepatol,
2018;15(9):555–67.
100. Friedman, S.L. Mechanisms of hepatic fibrogenesis. Gastroenterology,
2008;134(6):1655–69.
101. Jarnagin, W.R., Rockey, D.C., Koteliansky, V.E., Wang, S.S., and Bissell,
D.M. Expression of variant fibronectins in wound healing: cellular source
and biological activity of the EIIIA segment in rat hepatic fibrogenesis.
J Cell Biol, 1994;127(6 Pt 2):2037–48.
102. Bilzer, M., Roggel, F., and Gerbes, A.L. Role of Kupffer cells in host
defense and liver disease. Liver Int, 2006;26(10):1175–86.
103. Xie, G., Wang, X., Wang, L. et al. Role of differentiation of liver sinusoidal
endothelial cells in progression and regression of hepatic fibrosis in rats.
Gastroenterology 2012;142(4):918–27 e6.
104. Lee, W.Y. and Kubes, P. Leukocyte adhesion in the liver: distinct adhesion
paradigm from other organs. J Hepatol, 2008;48(3):504–12.
105. Wong, J., Johnston, B., Lee, S.S. et al. A minimal role for selectins in the
recruitment of leukocytes into the inflamed liver microvasculature. J Clin
Invest, 1997;99(11):2782–90.
106. Zhou, B., Weigel, J.A., Fauss, L., and Weigel, P.H. Identification of the
hyaluronan receptor for endocytosis (HARE). J Biol Chem, 2000;275(48):
37733–41.
107. Tamaki, S., Ueno, T., Torimura, T., Sata, M., and Tanikawa, K. Evaluation
of hyaluronic acid binding ability of hepatic sinusoidal endothelial cells in
rats with liver cirrhosis. Gastroenterology, 1996;111(4):1049–57.
108. Ohtani, O. and Ohtani, Y. Lymph circulation in the liver. Anat Rec,
2008;291(6):643–52.
109. Naito, M. and Wisse, E. Filtration effect of endothelial fenestrations on
chylomicron transport in neonatal rat liver sinusoids. Cell Tissue Res,
1978;190(3):371–82.
110. Fraser, R., Cogger, V.C., Dobbs, B. et al. The liver sieve and atherosclero-
sis. Pathology, 2012;44(3):181–6.
111. Braet, F. and Wisse, E. Structural and functional aspects of liver sinusoidal
endothelial cell fenestrae: a review. Comp Hepatol, 2002;1(1):1.
112. DeLeve, L.D. Hepatic microvasculature in liver injury. Semin Liver Dis,
2007;27(4):390–400.
113. Fried, M.W., Duncan, A., Soroka, S. et al. Serum hyaluronic acid in patients
with veno‐occlusive disease following bone marrow transplantation. Bone
Marrow Transplant, 2001;27(6):635–9.
114. Copple, B.L., Roth, R.A., and Ganey, P.E. Anticoagulation and inhibition of
nitric oxide synthase influence hepatic hypoxia after monocrotaline expo-
sure. Toxicology, 2006;225(2–3):128–37.
115. Williams, A.M., Langley, P.G., Osei‐Hwediah, J., Wendon, J.A., and
Hughes, R.D. Hyaluronic acid and endothelial damage due to paracetamol‐
induced hepatotoxicity. Liver Int, 2003;23(2):110–5.
116. Park, Y.D., Yasui, M., Yoshimoto, T. et al. Changes in hemostatic parame-
ters in hepatic veno‐occlusive disease following bone marrow transplanta-
tion. Bone Marrow Transplant, 1997;19(9):915–20.
117. Godfrey, C., Desviat, L.R., Smedsrod, B. et  al. Delivery is key: lessons
learnt from developing splice‐switching antisense therapies. EMBO Mol
Med, 2017;9(5):545–57.
118. Frank, M.M., Lawley, T.J., Hamburger, M.I., and Brown, E.J. NIH confer-
ence: immunoglobulin G Fc receptor‐mediated clearance in autoimmune
diseases. Ann Intern Med, 1983;98(2):206–18.
119. Davies, K.A., Robson, M.G., Peters, A.M., Norsworthy, P., Nash, J.T., and
Walport, M.J. Defective Fc‐dependent processing of immune complexes in
patients with systemic lupus erythematosus. Arthritis Rheum, 2002;46(4):
1028–38.
120. Ahmed, S.S., Muro, H., Nishimura, M., Kosugi, I., Tsutsi, Y., and
Shirasawa, H. Fc receptors in liver sinusoidal endothelial cells in
NZB/W F1 lupus mice: a histological analysis using soluble immunoglobu-
lin G‐immune complexes and a monoclonal antibody (2.4G2). Hepatology,
1995;22(1):316–24.
121. Cogger, V.C., Warren, A., Fraser, R., Ngu, M., McLean, A.J., and Le
Couteur, D.G. Hepatic sinusoidal pseudocapillarization with aging in the
non‐human primate. Exp Gerontol, 2003;38(10):1101–7.
122. Hilmer, S.N., Cogger, V.C., Fraser, R., McLean, A.J., Sullivan, D., and
Le Couteur, D.G. Age‐related changes in the hepatic sinusoidal endothelium
impede lipoprotein transfer in the rat. Hepatology, 2005;42(6):1349–54.
123. Le Couteur, D.G., Cogger, V.C., Markus, A.M. et al. Pseudocapillarization
and associated energy limitation in the aged rat liver. Hepatology,
2001;33(3):537–43.
124. Le Couteur, D.G., Fraser, R., Cogger, V.C., and McLean, A.J. Hepatic pseu-
docapillarisation and atherosclerosis in ageing. Lancet, 2002;359(9317):
1612–5.
125. McLean, A.J., Cogger, V.C., Chong, G.C. et al. Age‐related pseudocapil-
larization of the human liver. J Pathol, 2003;200(1):112–7.
126. Warren, A., Bertolino, P., Cogger, V.C., McLean, A.J., Fraser, R., and
Le  Couteur, D.G. Hepatic pseudocapillarization in aged mice. Exp
Gerontol, 2005;40(10):807–12.
127. Ito, Y., Sørensen, K.K., Bethea, N.W. et  al. Age‐related changes in the
hepatic microcirculation in mice. Exp Gerontol, 2007;42(8):789–97.
128. Simon‐Santamaria, J., Malovic, I., Warren, A. et al. Age‐related changes in
scavenger receptor‐mediated endocytosis in rat liver sinusoidal endothelial
cells. J Gerontol A Biol Sci Med Sci, 2010;65(9):951–60.
129. Caperna, T.J. and Garvey, J.S. Antigen handling in aging. II. The role of
Kupffer and endothelial cells in antigen processing in Fischer 344 rats.
Mech Ageing Dev, 1982;20(3):205–21.
130. Heil, M.F., Dingman, A.D., and Garvey, J.S. Antigen handling in ageing.
III. Age‐related changes in antigen handling by liver parenchymal and non-
parenchymal cells. Mech Ageing Dev, 1984;26(2–3):327–40.
131. Brouwer, A., Barelds, R.J., and Knook, D.L. Age‐related changes in the
endocytic capacity of rat liver Kupffer and endothelial cells. Hepatology,
1985;5(3):362–6.
132. Smedsrød, B., Melkko, J., Araki, N., Sano, H., and Horiuchi, S. Advanced
glycation end products are eliminated by scavenger‐receptor‐mediated
endocytosis in hepatic sinusoidal Kupffer and endothelial cells. Biochem J,
1997;322(2):567–73.
133. Svistounov, D., Oteiza, A., Zykova, S.N. et al. Hepatic disposal of advanced
glycation end products during maturation and aging. Exp Gerontol,
2013;48(6):549–56.
134. Oteiza, A., Li, R., McCuskey, R.S., Smedsrød, B., and Sørensen, K.K.
Effects of oxidized low‐density lipoproteins on the hepatic microvascula-
ture. Am J Physiol Gastrointest Liver Physiol, 2011.
135. Cogger, V.C., Mohamad, M., Solon‐Biet, S.M. et al. Dietary macronutri-
ents and the aging liver sinusoidal endothelial cell. Am J Physiol Heart Circ
Physiol, 2016;310(9):H1064–70.
136. Franceschi, C., Bonafe, M., Valensin, S. et  al. Inflamm‐aging. An
evolutionary perspective on immunosenescence. Ann N Y Acad Sci,
­
2000;908:244–54.
137. Muller, A.M., Skrzynski, C., Nesslinger, M., Skipka, G., and Muller, K.M.
Correlation of age with in vivo expression of endothelial markers. Exp
Gerontol, 2002;37(5):713–9.
138. Jamieson, H.A., Hilmer, S.N., Cogger, V.C. et al. Caloric restriction reduces
age‐related pseudocapillarization of the hepatic sinusoid. Exp Gerontol,
2007;42(4):374–8.
139. Harris, E.N. and Weigel, P.H. The ligand‐binding profile of HARE: hyalu-
ronan and chondroitin sulfates A, C, and D bind to overlapping sites distinct
from the sites for heparin, acetylated low‐density lipoprotein, dermatan
sulfate, and CS‐E. Glycobiology, 2008;18(8):638–48.
434 THE LIVER:  REFERENCES
140. Smedsrød, B., Paulsson, M., and Johansson, S. Uptake and degradation in
vivo and in vitro of laminin and nidogen by rat liver cells. Biochem J,
1989;261(1):37–42.
141. Oie, C.I., Olsen, R., Smedsrød, B., and Hansen, J.B. Liver sinusoidal
endothelial cells are the principal site for elimination of unfractionated
heparin from the circulation. Am J Physiol Gastrointest Liver Physiol,
2008;294(2):G520–8.
142. Oynebraten, I., Hansen, B., Smedsrød, B., and Uhlin‐Hansen, L. Serglycin
secreted by leukocytes is efficiently eliminated from the circulation by sinusoi-
dal scavenger endothelial cells in the liver. J Leukoc Biol, 2000;67(2):183–8.
143. Melkko, J., Hellevik, T., Risteli, L., Risteli, J., and Smedsrød, B. Clearance
of NH2‐terminal propeptides of types I and III procollagen is a physiologi-
cal function of the scavenger receptor in liver endothelial cells. J Exp Med,
1994;179(2):405–12.
144. Smedsrød, B., Johansson, S., and Pertoft, H. Studies in vivo and in vitro on
the uptake and degradation of soluble collagen alpha 1(I) chains in rat liver
endothelial and Kupffer cells. Biochem J, 1985;228(2):415–24.
145. Niesen, T.E., Alpers, D.H., Stahl, P.D., and Rosenblum, J.L. Metabolism of
glycosylated human salivary amylase: in vivo plasma clearance by rat
hepatic endothelial cells and in vitro receptor mediated pinocytosis by rat
macrophages. J Leukoc Biol, 1984;36(3):307–20.
146. Van Berkel, T.J., De Rijke, Y.B., and Kruijt, J.K. Different fate in vivo of
oxidatively modified low density lipoprotein and acetylated low density
lipoprotein in rats. Recognition by various scavenger receptors on Kupffer
and endothelial liver cells. J Biol Chem, 1991;266(4):2282–9.
147. Hansen, B., Svistounov, D., Olsen, R., Nagai, R., Horiuchi, S., and Smedsrød,
B. Advanced glycation end products impair the scavenger function of rat
hepatic sinusoidal endothelial cells. Diabetologia, 2002;45(10):1379–88.
148. Eskild, W., Smedsrød, B., and Berg, T. Receptor mediated endocytosis of
formaldehyde treated albumin, yeast invertase and chondroitin sulfate in
suspensions of rat liver endothelial cells. Int J Biochem, 1986;18(7):
647–51.
149. Praaning‐van Dalen, D.P., de Leeuw, A.M., Brouwer, A., and Knook, D.L.
Rat liver endothelial cells have a greater capacity than Kupffer cells to
endocytose N‐acetylglucosamine‐ and mannose‐terminated glycoproteins.
Hepatology, 1987;7(4):672–9.
150. Smedsrød, B. and Seljelid, R. Fate of intravenously injected aminated
beta(1‐‐‐‐3) polyglucose derivatized with 125I‐tyraminyl cellobiose.
Immunopharmacol, 1991;21(3):149–58.
151. Blomhoff, R., Eskild, W., and Berg, T. Endocytosis of formaldehyde‐
treated serum albumin via scavenger pathway in liver endothelial cells.
Biochem J, 1984;218(1):81–6.
152. Nagelkerke, J.F., Barto, K.P., and van Berkel, T.J. In vivo and in vitro
uptake and degradation of acetylated low density lipoprotein by rat liver
endothelial, Kupffer, and parenchymal cells. J Biol Chem, 1983;258(20):
12221–7.
153. Adachi, H. and Tsujimoto, M. FEEL‐1, a novel scavenger receptor with in
vitro bacteria‐binding and angiogenesis‐modulating activities. J Biol Chem,
2002;277(37):34264–70.
154. Irjala, H., Alanen, K., Grenman, R., Heikkila, P., Joensuu, H., and Jalkanen,
S. Mannose receptor (MR) and common lymphatic endothelial and vascu-
lar endothelial receptor (CLEVER)‐1 direct the binding of cancer cells to
the lymph vessel endothelium. Cancer Res, 2003;63(15):4671–6.
155. Tamura, Y., Adachi, H., Osuga, J. et al. FEEL‐1 and FEEL‐2 are endocytic
receptors for advanced glycation end products. J Biol Chem,. 2003;278(15):
12613–7.
 Fenestrations in the Liver
35 Sinusoidal Endothelial Cell
Victoria C. Cogger, Nicholas J. Hunt, and David G. Le Couteur
Centre for Education and Research on Ageing, University of Sydney and Concord RG Hospital,
Sydney, Australia
INTRODUCTION
Liver sinusoidal endothelial cells (LSECs) occupy a strategic
position in the liver, separating blood in the sinusoid from the
extracellular space of Disse and surrounding sheets of hepato-
cytes. The cytoplasmic extensions of LSECs are very thin and
perforated with patent pores called fenestrations, which lack
diaphragms or underlying basal lamina. Fenestrations are a
portal for the direct transfer of substrates between sinusoidal
blood and extracellular fluid, and also permit circulating
immune cells, particularly T lymphocytes to interact with
hepatocytes [1].
HISTORICAL BACKGROUND
In the late nineteenth century, it was concluded from studies
investigating uptake in the liver after the injection of dyes into
the liver blood vessels, that there must be small channels linking
the hepatic capillaries and perivascular lymphatics [2]. In 1906,
Herring and Simpson performed a light microscopic examina-
tion of the cat liver and noted the “sinusoidal character of the
blood‐vessels and the incomplete nature of their endothelial
­lining” [3]. In the 1950s, Fawcett and others used electron
microscopy to detect pores or fenestrations in the LSEC [4].
Later, Wisse established the ultrastructure of fenestrations and
their arrangement in sieve plates [5] and contemporarily, Fraser
showed that fenestrations filter particulate substrates such as
lipoproteins on the basis of size [6]. Through calcium imaging
studies Arias [7] and Oda [8] went on to report that fenestrations
are dynamic structures that can be regulated by endogenous
and  exogenous agents such as serotonin, noradrenalin, and
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
neuropeptide Y [9] (Figure 35.1). Recently the development of
new imaging technologies (super‐resolution microscopy and
atomic force microscopy) has enhanced the study of fenestra-
tions, particularly by revealing their structure and behavior in
live cells.
Structure of fenestrations
Due to the predominantly venous blood supply hepatic sinu-
soids bridge afferent portal venules to exiting central hepatic
venules. Arterial blood, which is delivered by the hepatic
­arterioles, is mixed within the sinusoid providing the limited
oxygen required by the liver. The sinusoids have an average
diameter of approximately 5–10 μm and occupy between
10–30% of the total liver volume [10]. LSECs, which repre-
sent 2.5% of total liver volume and 15–20% of all liver cells
[7], are a highly specialized type of endothelial cell that lines
the wall of the hepatic sinusoid. The cytoplasmic extensions of
LSECs are very thin and perforated with fenestrations, which
are circular or polygonal pores approximately 50–200 nm in
diameter. The fenestrations are complete pores that connect
the sinusoidal and abluminal surfaces of the LSEC with no
associated diaphragm. There are approximately 3–20 fenestra-
tions per μm2 of endothelial surface and between 2–10% of the
surface of the LSEC is covered by fenestrations, which is
termed “porosity”. Fenestrations are scattered individually
across the endothelial surface or clustered into groups of ten or
more called liver sieve plates with between 60–75% of fenes-
trations to be found within sieve plates. There may be a zonal
gradient of fenestration diameter within the sinusoids with
smaller fenestrations in the periportal (zone 1) sinusoids and
greater porosity being found in pericentral (zone 3) sinusoids
[11, 12] (Figure 35.2).
436 THE LIVER:  HISTORICAL BACKGROUND

(a) (b)

(c)

Kupffer cell

Hepatic sinusoid Space of Disse

Lipoprotein
Agranular reticulum

Golgi complex

Figure 35.1  Early transmission electron micrographs of the rat liver sinusoid demonstrating fenestrations in the LSEC. (a) Bennett et al. (1959)
[91]. Reproduced with permission of the American Physiological Society. J, fenestrations; G, endothelial cell; S, space of Disse; L, sinusoidal
lumen. (b) Wisse (1970) [5]. Reproduced with permission of Elsevier. →, fenestrations; end, endothelial cell; DS, space of Disse; L, sinusoidal
lumen. (c) By 1968 the concept of fenestrations had entered texts (Bloom and Fawcett 1968) [92]. Reproduced with permission of Elsevier.
The diagram shows a fenestration and lipoproteins in the space of Disse, but also the incorrect historical assumption that Kupffer cells also line
the sinusoid.

The diameter of fenestrations has a Gaussian distribution,
which is skewed to the right by the presence of some larger
pores. Very small non‐perforating pores are called pits and
larger pores (greater than about 300 nm diameter, depending
upon the fixation and microscopy method) are called gaps. Gaps
are known to occur as a result of cellular injury, technical issues
related to high pressure perfusion either physiologically or
experimentally, hypoxia, fixation, or pathological states.
Fenestrations are also found in mesh‐like, labyrinthine struc-
tures reminiscent of vesiculo‐vacuolar organelles and pored
domes [13]. They have been observed in numerous vertebrate
species (man, rat, mouse, guinea pig, sheep, goat, rabbit, fowl,
monkey, baboon, bat, kitten, dog, turtle, goldfish) where they
have similar appearances, including sieve plate formation [12].
The diameter of fenestrations is smaller than the resolution of
standard light microscopy therefore most morphological studies
in the past have utilized transmission and scanning electron
microscopy. However, preparation of tissue for electron micros-
copy, particularly scanning electron microscopy generates arte-
facts including tissue shrinkage, which has led to underestimation
of the diameter of fenestrations by as much as a third. Another
important technical issue is that LSECs lose their fenestrations
within hours of isolation and culture, limiting the duration of
experiments that can be validly performed [14]. Moreover there
has been huge variability in specimen preparation and statistical
methods, which has led to published conclusions about fenestra-
tion morphology and response to interventions that can be
­difficult to interpret. Standardized methods for electron micros-
copy and reporting of fenestrations have been proposed but not
yet widely adopted [15]. At a minimum the following informa-
tion should be provided: the methods used to measure fenestra-
tions; the values for the number, frequency, diameter, and
porosity of fenestrations that were measured; and the diameter
limits used to define fenestrations.
Recent developments in microscopic techniques have been
applied to fenestrations [16, 17]. The first super‐resolution light
microscopy of fenestrations in fixed LSECs was undertaken
using structured illumination microscopy (SIM) in 2010 [18],
which yielded a fenestration diameter of 123 nm, and showed
the three‐dimensional structure of sieve plates including the
observation that the width of LSEC traversed by fenestrations is
about 165 nm. Following from this time numerous other tech-
niques have emerged that have provided significant advances in
our understanding of fenestration biology (Figure  35.3a).
Atomic force microscopy (AFM), is a technique that relies on
the scanning of a probe along the membrane of a fixed or live
cell allowing the observation of membrane topology. It was first
applied to LSECs in 2001 but at the time was unable to
 (a) 50 μm (b) 10 μm

(c) 1μm (d) 1μm

5 μm 1 μm
(e) (f)

Figure 35.2  Electron micrographs of LSECs and fenestrations. (a) Scanning electron microscope (EM) of vascular cast showing portal vein (PV)
branching into sinusoids (S) (b) Scanning EM of sinusoids (S) showing intercalated plates of hepatocytes (H). Fenestrations (→) are apparent in the
endothelial wall. (c) Scanning EM of sinusoidal endothelial lumen showing fenestrations clustered into sieve plates (SP). (d) Transmission EM
showing sinusoidal lumen, LSEC (E), stellate cell process (SC), and extravascular space of Disse containing microvilli. (e) Scanning EM of isolated
LSEC showing fenestrations distributed in the cytoplasm away from nucleus (Nuc) (f) Scanning EM of isolated LSEC showing fenestrations
­clustered in sieve plates and a network of fenestrations resembling a vesiculo‐vacuolar organelles (VVO). Magnification bars are shown.

1.0

0.8 Fenestration diameter

25 nm
50 nm
Sieving coefficient

0.6 75 nm
100 nm
150 nm

0.4

0.2

0.0
0 20 40 60 80 100 120 140
Lipoprotein diameter (nm)

Figure 35.3  Super resolution microscopy of the LSEC and their fenestrations. (a) Structured illumination microscopy (OMX Deltavision) of an
isolated rat LSEC. The membrane is stained with cell mask red. Fenestrations are readily visible with this method of light microscopy. (b) dSTORM
of an isolated rat LSEC, again fenestrations are clearly apparent using this method of observation. (c) and (d) 3‐dimensional structured illumination
microscopy (OMX Deltavision) reconstructed by rendering the surface of an isolated rat LSEC. The reduced thickness of the cell in regions of
fenestrations and sieves plates is evident. Acknowledgment to Mr. Hong Mao for undertaking the microscopy for images (a) and (b).
438 THE LIVER:  HISTORICAL BACKGROUND
distinguish fenestrations [19]; it has since been used success-
fully to study fixed and cultured living LSECs and has indicated
that fenestrations have a diameter of 140–220 nm and porosity
of about 4% [16, 20, 21]. Further, application of super‐resolu-
tion microscopy to LSECs has revealed important biophysical
properties of the LSEC membrane that facilitates the formation
of fenestrations. SIM analysis revealed that the cell height is
“pinched” in regions of fenestrations and sieve plates, this work
led to the understanding that fenestrations are present in cell
membrane regions that are devoid of classic lipid raft bound
lipids and proteins [22] and the formulation of the sieve‐raft
hypothesis for the regulation of fenestrations [23]. Another
super‐resolution method, direct stochastic optical reconstruc-
tion microscopy (dSTORM) has been used in isolated LSECs
and shown a fenestration diameter of 120 nm and identified the
close relationship between fenestrations and the cytoskeleton
(Figure  35.3b) [24]. The observations derived from the new
super resolution methods all confirm the presence of fenestra-
tions in living LSECs free of fixation artefacts, and have also
shown that fenestrations are dynamic structures responding over
minutes to various interventions such as antimycin A [20].
Additional super‐resolution techniques that involve single mol-
ecule localization such as PALM (photoactivated localization
microscopy) [25] and its related technique fPALM (fluorescent
photoactivated localization microscopy) [26] as well as STED
(stimulated emission depletion) [27] offer further promise in
uncovering unknown aspects of LSEC fenestration biology but
have yet to be applied to the task [17]. This is most likely due to
the limitations in the field caused by the absence of known
markers or proteins that define, label, or visualize fenestrations
directly.
Physiological role of fenestrations
Fenestrations facilitate the highly efficient bidirectional transfer
of substrates between the blood and hepatocytes, while prevent-
ing the access of blood cells and large particulate substrates
(such as platelets and chylomicrons) to the extracellular space.
The fenestrated endothelium acts as a filter, hence, is sometimes
referred to as “the liver sieve” [6, 11]. Fenestrations permit the
unimpeded flow of a wide range of substrates (plasma and
­substrates within plasma, plasma proteins including albumin,
smaller lipoproteins) into the extracellular space of Disse. The
narrowness of hepatic endothelial cells and the lack of basal
lamina and collagen in the space of Disse ensure that any other
permeability barriers to the diffusion of substrates between
blood and hepatocytes are minimized. Extensive investigations
of substrate transfer using multiple indicator dilution methods
have established that under normal conditions, there is no bar-
rier to the transfer of soluble substrates, albumin, and albumin‐
bound substrates at the LSEC; instead transfer into the space of
Disse is flow‐limited [28].
Both diameter and frequency of fenestrations determine
diffusive and convective transfer across the LSEC whereas
­ nism for hepatic insulin resistance and hyperinsulinemia.
permselectivity is determined only by fenestration diameter
[29]. The LSEC can be considered to be a typical low pressure
ultrafiltration system because ultrafiltration membranes are
­usually considered to have pores between 2–100 nm in diame-
ter. In ultrafiltration, the volume flux (J) is described by the
fR 2 P
Hagen–Poiseuille formula: J where f is the porosity of
8 l
the membrane, R diameter of the pores, ΔP the pressure gradient
across the membrane, η viscosity, and l the thickness of the
membrane. In conditions associated with loss of fenestrations
such as liver disease and aging, there is typically a reduction
of porosity of 30–50%, a reduction of fenestration diameter of
10%, and an increase in endothelial thickness of 50%. Changes
of this magnitude will be associated with a substantial reduction
in the flux of fluid and dissolved substrates across the LSEC (of
more than 50%). In addition, flow is greatest across the largest
pores, so permeability is sensitive to the fraction of large pores,
that is, the distribution of pore diameter that is present in the
LSEC (Figure 35.4).
Fenestrations influence the hepatic uptake of lipoproteins,
particularly chylomicron remnants [30, 31]. The first stage in
the metabolism of lipoproteins is the production of chylomi-
crons which are triglyceride‐rich lipoproteins formed in the
intestine from dietary lipids. The diameter of chylomicrons
range from 100–1000 nm and they are unable to pass through
fenestrations; moreover most chylomicrons bypass the liver via
the thoracic duct. Chylomicrons are metabolized to chylomi-
cron remnants by lipoprotein lipase present on the endothelium
of systemic capillaries. Chylomicron remnants are smaller
particles (30–80 nm) that have acquired apolipoprotein E
­
(apoE). Remnants pass through fenestrations and are seques-
tered within the space of Disse for receptor‐mediated uptake by
the low density lipoprotein receptor (LDL‐R) and lipoprotein‐
receptor related protein (LRP) into hepatocytes [31]. Electron
microscopy has demonstrated that chylomicrons are only found
in the sinusoid whereas remnants are also observed in the space
of Disse. There is d­ ifferential trapping by the liver of radiola-
beled lipoproteins of different sizes such that smaller particles
are trapped to a greater extent than those larger than 100 nm [6].
Similar results indicating exclusion on the basis of size have
been reported for large and small liposomes and colloidal gold
particles of different diameters [32, 33]. Reduction in the num-
ber or diameter of ­ fenestrations (“defenestration”) leads to
impaired clearance of  chylomicron remnants after meals and,
because remnants are still relatively rich in triglycerides, this is
manifested as postprandial hypertriglyceridemia [30, 31].
Defenestration associated with old age [34]; knockout of the
vascular endothelial growth factor (VEGF) receptor [35]; alco-
holic cirrhosis [30], poloxamer 407 [36], and knockout of the
plasmalemma vesicle‐associated protein (PLVAP) [37], is asso-
ciated with impaired lipoprotein uptake, hypertriglyceridemia
and/or increased ­circulating chylomicron remnants.
Fenestrations have been shown to be a portal for the transfer
of several other substrates. Defenestration associated with
poloxamer 407 and old age has been found to reduce the hepatic
uptake of some drugs (paracetamol, diazepam [38, 39]) and
insulin [40]. The effect of defenestration on insulin uptake in the
liver is particularly significant because it provides a new mecha-
Conversely, increased fenestrations generated by partial loss of
platelet‐derived growth factor‐beta (PDGF‐β) is associated with
increased hepatic insulin sensitivity and lower circulating levels
of insulin [41]. Fenestrations may be important in gene therapy
because they permit the transfecting virus, or other gene carrier,
 35:  Fenestrations in the Liver Sinusoidal Endothelial Cell 439

(a) (b)

(c) (d)

Figure 35.4  The effect of fenestration diameter on the sieving coefficient (the fraction of particles able to pass through ultrafiltration pores)
according to lipoprotein diameter [80]. Reproduced with permission of Elsevier.

to gain access to the liver cells, and interestingly many of the
strategies that improve uptake of gene therapy vectors such as
partial hepatectomy, liver ischemia, and cyclophosphamide
cause increased porosity and/or gap formation in the liver sinu-
soidal endothelium [42].
As well as their role in substrate transfer, fenestrations permit
interactions between cells in the sinusoidal lumen and the cell
membrane of underlying hepatocytes. LSECs express many
antigens important for interactions with leukocytes and lympho-
cytes and have a possible role in antigen presentation [43].
Naïve T cells interact directly with hepatocytes by inserting
filopodia through fenestrations (transendothelial hepatocyte
­ vasoactive cytokines and hormones, and local paracrine and
lymphocyte interactions, TEHLI) and this appears to be the first
step in the development of immunotolerance [1]. Activated lym-
phocytes and other leukocytes accessed the liver tissue in an
experimental model of autoimmune hepatitis via passage
through the fenestrations and conversely defenestration almost
fully abolished any hepatitis [44]. Cytotoxic T lymphocytes rec-
ognize hepatitis B viral antigens on hepatocytes by extending
cytoplasmic protrusions though fenestrations and then initiate
hepatocyte killing, a process that is impaired by defenestration
associated with hepatic fibrosis [45].
Fenestrations have several other functions. Contraction of
fenestrations might increase vascular resistance thereby influ-
encing hepatic blood flow and pressure [46]. For example, an
endothelin antagonist was found to dilate fenestrations and
cause a reduction in portal perfusion pressure of about 2.5 cm
H2O [47]. Fenestrations are also involved in the formation of
hepatic lymph. The space of Disse is continuous with lymphatic
vessels found in the portal triads. This morphology suggests that
plasma flows through the fenestrations, upstream along the
space of Disse, finally emptying into the lymphatic vessels
around the portal vein [28]. A hydrodynamic analysis of flow in
the hepatic sinusoids is consistent with the concept of retrograde
plasma flow along the space of Disse [48].
Regulation of fenestrations
Fenestrations are dynamic structures that change in frequency
and diameter in response to numerous stimuli in vitro. In vivo it
is likely that fenestrations open and close in response to various
stimuli such as inflammation, nutrition and fasting, circulating
autocrine factors [7] (Table 35.1). The embryonic development
of fenestrations depends upon PLVAP which is linked to
­diaphragm formation [37], and VEGF [35], while in adults, the
maintenance of fenestrations seems to depend upon pathways
regulating the actin cytoskeleton and its effects on lipid rafts
[22, 23]. There is often an inverse relationship between the
effects of an agent on fenestration diameter and frequency.
Lipid rafts
Recently a “sieve‐raft” hypothesis was proposed whereby
­fenestrations form in non‐raft membranes when the membrane
stabilizing effects of the cytoskeleton and membrane rafts are
diminished. Cell membranes are heterogeneous structures com-
posed of lipid rafts (liquid‐ordered microdomains, 10–200 nm)
and non‐raft, liquid‐disordered regions. Lipid rafts are regions
of cell membrane enriched in cholesterol, sphingolipids, and
proteins that provide a platform for many membrane proteins.
They are tethered to the cytoskeleton by a variety of proteins
440 THE LIVER:  HISTORICAL BACKGROUND

Table 35.1  Fenestrations have been reported in all species studied, is the primary stimulus for VEGF production, whereas
and a very wide range of species. Some of the many reports of VEGFR‐2 is found along the entire sinusoidal endothelium
different species are shown in Table 35.1 in order to show that [55]. Differentiation of LSECs and fenestrations require both
fenestrations are widespread, and quite similar, in animals and humans
VEGF and nitric oxide (NO) while VEGF acts via both NO‐
Porosity Diameter Frequency dependent and independent pathways [56]. Genetic reduction of
Species (area %) (nm) (per μm2) Citation
VEFG activity in the liver led to loss of fenestrations associated
Rat (zone 1) 9.6 73 ± 0.13 5.7 ± 0.1 [78] with impaired uptake of chylomicron remnants [35]. VEGF is
Rat (zone 3) 28.5 94 ± 0.11 10.2 ± 0.01 [78]
Rat (zone 1) 6.0 ± 0.2 111 ± 1 9.1 ± 0.3 [10] associated with increased intracellular calcium, and has marked
Rat (zone 3) 7.9 ± 0.3 105 ± 0.2 13.3 ± 0.5 [10] effects on the cytoskeleton [57] but unlike the agents studied by
Rat 4.1 ± 2.3 73 ± 1 2.7 ± 1.1 [66] Gatmaitan and Arias [58], these intracellular changes are associ-
Rat 12.0 ± 2.1 110 ± 7 12.4 ± 3.6 [79] ated with increased – rather than reduced – fenestrations.
Mice 4.1 ± 2.2 74 ± 4 — [80]
Rabbit 5.2 ± 0.9 60 ± 5 17.3 ± 3.8 [81] The Rho family of GTPases include Rho which regulates
Rabbit 4.0 ± 1.5 69 ± 8 12.7 ± 2.5 [79] F‐actin and the cytoskeleton, Rac which regulates lamellipodia,
Chicken 3.6 ± 1.6 99 ± 15 3.9 ± 0.9 [81] and Cdc42 which controls development of filopodia. The Rho
Chicken 2.2 ± 0.6 90 ± 18 2.9 ± 0.3 [79]
Rainbow trout — 123 — [82] pathway, via its effects on the cytoskeleton, is also involved in
Goldfish — 50–200 — [83] regulation fenestrations. Inhibition of the Rho pathway by C3‐
Dog 6.7 118 ± 2 7.2 [84] transferase caused reduction of myosin light chain phosphoryla-
Sheep — 60 ± 2 — [85] tion, loss and retraction of actin filaments, and led to increased
Baboon 2.6 ± 0.2 50 ± 1 12.1 ± 0.8 [86]
Baboon 4.2 ± 0.5 58 ± 1 9.4 ± 0.9 [87] porosity and the formation of large gaps. Similarly, the Rho‐
Baboon 1.8 82 3.3 [88] associated protein kinase (ROCK) inhibitor Y‐2342 increases
Human (zone 1) 7.6 — 19 [89] fenestrations [59]. On the other hand activating Rho with
Human (zone 3) 9.1 — 23.5 [89]
Human (zone 1) 3.4 ± 0.2 170 ± 12 9.8 ± 1.8 [90]
lysophosphatidic acid increased myosin light chain phospho-
Human (zone 3) 4.0 ± 0.4 160 ± 10 11.2 ± 2.6 [90] rylation and actin filaments and led to defenestration [60].
Endothelin‐1 (ET‐1) is a potent vasoconstrictor produced by
endothelial cells and an upstream factor of Rho kinase. ET‐1
increased intracellular calcium and decreased fenestration
including ezrin, radixin, moesin, and stabilin. Between the
diameter [61]. Antagonism of the endothelin receptor, ETA‐R
rafts  lie non‐raft, lipid‐disordered regions of cell membrane.
caused a marked increase in fenestration diameter associated
Super‐resolution microscopy has shown that liver sieve plates
with gap formation [47]. ET‐1 is increased in cirrhosis which is
are distributed between lipid rafts in the non‐raft regions.
associated with increased portal pressure and defenestration [62].
Depletion of lipid rafts (with 7‐ketocholesterol) increases fenes-
trations while depletion of non‐raft membrane (with Triton X)
reduces fenestrations [22, 23]. Nutritional factors
Acute fasting increases the diameter of fenestrations while
Cytoskeleton reducing their frequency [63]. Chronic reduction of food intake
Liver sieve plates are supported by the actin cytoskeleton [49, 50] over a lifetime (caloric restriction) reversed age‐related loss of
and a variety of actin‐based structures involved with the mainte- fenestrations [64]. In long‐term feeding experiments where
nance of fenestrations have been identified such as the fenestrae‐ mice were fed different ratios of macronutrients, it was found
associated cytoskeleton ring, sieve plate associated cytoskeleton, that both fenestration porosity and frequency were inversely
fenestrae forming center, and defenestration‐associated center. related to fat intake, while diameter was inversely associated
Agents that disrupt actin such as cytochalasin B, cytochalasin with protein or carbohydrate intake [65]. Given the role of the
D, misokinolide, and latrunculin increase the numbers of fenes- liver in metabolizing nutrients it is not surprising that fenestra-
trations, usually in the order of two‐fold [51, 52]. The effect of tions respond to diet. Overall it seems that lower amounts of
cytochalasin D on fenestrations is blocked by Triton X which food intake are associated with increased fenestrations.
depletes non‐raft membrane indicating that actin cytoskeleton
acts on fenestrations via its effect on lipid rafts [22].
Pathophysiology of fenestrations
Most other agents that have been shown to influence fenestra-
tions have direct or indirect action on the cytoskeleton. The role There are numerous reports of diseases and pathological
of calcium in regulating fenestrations through effects on the ­processes that influence fenestrations, including: primary liver
cytoskeleton was first reported by Gatmaitan and Arias [53]. disease (cirrhosis, fibrosis, steatosis, hepatitis, hepatic vascular
Several agents (serotonin, metoclopramide, propranolol, indo- diseases, and the sinusoidal obstruction syndrome, vena caval
methacin, calcium ionophore) were identified that reduced the obstruction), liver toxins (acetaminophen, oxidants, bacterial
diameter of fenestrations by about 20% in rat LSECs and were toxins), systemic disease (diabetes mellitus), and other liver
associated with an increase of intracellular calcium. processes (aging, partial hepatectomy, hypoxia, high pressure,
VEGF is a potent inducer of fenestrations. In the liver, hepat- ischemia reperfusion, and transplantation) (reviewed previously
ocytes produce VEGF which acts on liver endothelial cells via [12]). These changes have not usually been diagnostic but the
the receptors: VEGFR‐1 (Flt‐1) and VEGFR‐2 (KDR/Flk‐1) of overall trends are that: (i) acute toxic injury and acute medical
which VEGFR‐2 is the most important [54]. VEGF is expressed conditions are associated with loss of endothelial integrity
more highly in the pericentral regions reflecting hypoxia, which ­characterized by gap formation and (ii) subacute and chronic
 35:  Fenestrations in the Liver Sinusoidal Endothelial Cell 441
conditions have been associated with defenestration and reduced
porosity. Three important conditions are described below.
Aging and pseudocapillarization
Old age is associated with thickening and defenestration of the
LSEC, perisinusoidal fibrosis, and increased numbers of fat
engorged, non‐activated stellate cells [66, 67]. There also are
several immunohistochemical age‐related changes including
endothelial upregulation of von Willebrand’s factor and
ICAM‐1, and reduced expression of caveolin‐1. These changes
occur in the absence of light microscopic evidence of liver dis-
ease and have been termed age‐related pseudocapillarization.
Pseudocapillarization has been documented in mice, rats,
humans, baboons, and genetic mouse models of premature
aging. The age‐related loss of fenestrations is associated with
impaired transfer of lipoproteins [34], insulin [40], and some
drugs [38, 39], and therefore is a potential mechanism for age‐
related changes in circulating lipoproteins, insulin sensitivity,
and drug metabolism.
Cirrhosis and capillarization
The term “capillarization” was first used in 1963 by Schaffner
and Popper to describe the ultrastructural changes seen in the
sinusoidal endothelium in cirrhosis, including a thickened
endothelium with underlying basement membrane and loss of
fenestrations [68]. These findings in human cirrhosis and alco-
holic liver disease have been frequently observed [69] as well as
animal models of cirrhosis [62, 70, 71]. Because of the associa-
tion between alcohol consumption and cirrhosis, there have
been several studies of the acute and subacute effects on fenes-
trations. Results are inconsistent but overall, acute ethanol
causes dilatation of fenestrations [72]. Impaired transfer of
numerous substrates (albumin, lidocaine, propranolol, prazosin,
labetalol, diltiazem, IgM) across the capillarized sinusoidal
endothelium has been documented in cirrhotic livers [28].
Sinusoidal obstruction syndrome
The sinusoidal obstruction syndrome is the prototypic disease
of LSECs and probably the only recognized primary disease of
the LSEC [73]. The two main causes of this syndrome are dietary
pyrrolizidine alkaloids and chemo‐irradiation, especially asso-
ciated with bone marrow transplantation [74, 75]. The major
experimental model is induced by monocrotaline, a pyrrolizidine
alkaloid that is activated by cytochrome P450 and acts as an
actin disruptor. After administration of monocrotaline, there is
gap formation, endothelial swelling, and defenestration. This is
followed by extensive sinusoidal endothelial cell injury associ-
ated with dissection into the space of Disse and eventually
embolization of the sinusoidal endothelial fragments [73, 76].
Centrilobular necrosis ensues as a result of impaired perfusion
and clinically is associated with jaundice, hepatomegaly, and
ascites and a high mortality. Matrix metalloproteinases, particu-
larly MMP‐9 and MMP‐2 are key mediators of the syndrome,
but appear to influence the dehiscence of LSECs rather than
­fenestrations. On the other hand, the decrease in NO levels that
occurs early [74, 77] could directly influence fenestrations. Of
great significance is the observation that strategies that maintain
the integrity of the LSEC such as MMP inhibition and preserva-
tion of NO levels, totally abrogate any hepatocellular injury
[74]. The loss of LSEC morphology appears to be an initiating
step for some hepatotoxic agents.
CONCLUSIONS
Fenestrations are essential components of a healthy liver acting
as a conduit for many endogenous and exogenous substrates
between the blood and hepatocytes. As such presence of fenes-
trations in the liver sinusoid serve as an important biological
bellwether of overall liver health. Conditions such as aging,
fibrosis, and cirrhosis which are all associated with diminished
or total loss of liver function are all known to be associated with
loss of fenestrations and LSEC changes. The advent of exciting
new microscopic techniques such as super‐resolution micros-
copy has generated new knowledge about the biology of
­fenestrations. Further understanding of proteins and processes
associated with fenestration formation, maintenance, and loss
will provide essential insights into the physiology and patho-
physiology of the liver, as well as offering potential targets
for  therapies directed at treating liver conditions and diseases
associated with aging such as insulin resistance and diabetes.
REFERENCES
1. Warren, A., Le Couteur, D.G., Fraser, R., Bowen, D.G., McCaughan, G.W.,
and Bertolino, P. T lymphocytes interact with hepatocytes through
fenestrations in murine liver sinusoidal endothelial cells. Hepatology,
­
2006;44(5):1182–90.
2. Fraser, J.W. and Fraser, E.H. Preliminary note on intra and intercellular
­passages in the liver of the frog. J Anat Physiol, 1895;29:240–3.
3. Herring, P.T. and Simpson, S. On the relation of the liver cells to the blood‐
vessels and lymphatics. Proc Royal Soc Lond, 1906;78:455–97.
4. Fawcett, D.W. Observations on the cytology and electron microscopy of
hepatic cells. J Natl Cancer Inst, 1955;15 (5):1475–503.
5. Wisse, E. An electron microscopic study of the fenestrated endothelial lining
of rat liver sinusoids. J Ultrastruct Res, 1970;31:125–50.
6. Fraser, R., Bosanquet, A.G., and Day, W.A. Filtration of chylomicrons
by  the  liver may influence cholesterol metabolism and atherosclerosis.
Atherosclerosis, 1978;29:113–23.
7. Arias, I.M. The biology of hepatic endothelial cell fenestrae. Prog Liver Dis,
1990;9:11–26.
8. Oda, M., Tsukada, N., Komatsu, H., Kaneko, K., Nakamura, M., and
Tsuchiya, M. Electron microscopic localizations of actin, calmodulin and
calcium in the hepatic sinusoidal endothelium in the rat, in Cells of the
Hepatic Sinusoid, (eds. A. Kirn, D.L. Knook, E. Wisse), Kupffer Cell
Foundation, Leiden, 1986. pp. 511–2.
9. Oda, M., Han, J.Y., and Yokomori, H. Local regulators of hepatic sinusoidal
microcirculation: recent advances. Clin Hemorheol Microcirc, 2000;23
(2–4):85–94.
10. Wisse, E., De Zanger, R.B., Jacobs, R., and McCuskey, R.S. Scanning
­electron microscope observations on the structure of portal veins, sinusoids
and central veins in rat liver. Scan Electron Microsc, 1983(3):1441–52.
11. Wisse, W., De Zanger, R.B., Charels, K., Van Der Smissen, P., and McCuskey,
R.S. The liver sieve: considerations concerning the structure and function of
endothelial fenestrae, the sinusoidal wall and the space of Disse. Hepatology,
1985;5:683–92.
12. Cogger, V.C. and Le Couteur, D.G. Fenestrations in the liver sinusoidal
endothelial cell, in The Liver: Biology and Pathobiology, 5th edn, (eds. I.M.
Arias, A. Wolkoff, J.L. Boyer et  al.), John Wiley & Sons, Hoboken, NJ,
2009, pp. 387–404.
442 THE LIVER:  REFERENCES
13. Braet, F., Riches, J., Geerts, W., Jahn, K.A., Wisse, E., and Frederik, P.
Three‐dimensional organization of fenestrae labyrinths in liver sinusoidal
endothelial cells. Liver Int, 2009;29(4):603–13.
14. Martinez, I., Nedredal, G.I., Oie, C.I. et al. The influence of oxygen tension
on the structure and function of isolated liver sinusoidal endothelial cells.
Comp Hepatol, 2008;7:4.
15. Cogger, V.C., O’Reilly, J.N., Warren, A., and Le Couteur, D.G. A standard-
ized method for the analysis of liver sinusoidal endothelial cells and their
fenestrations by scanning electron microscopy. J Viz Exp, 2015(98):e52698.
16. Braet, F., Taatjes, D.J., and Wisse, E. Probing the unseen structure and
­function of liver cells through atomic force microscopy. Semin Cell Dev Biol,
2018;73:13–30.
17. Øie Cristina, I., Mönkemöller, V., Hübner, W. et al. New ways of looking at
very small holes  –  using optical nanoscopy to visualize liver sinusoidal
endothelial cell fenestrations. Nanophotonics, 2018:575.
18. Cogger, V.C., McNerney, G.P., Nyunt, T. et al. Three‐dimensional structured
illumination microscopy of liver sinusoidal endothelial cell fenestrations.
J Struct Biol, 2010;171(3):382–8.
19. Braet, F., de Zanger, R., Seynaeve, C., Baekeland, M., and Wisse, E. A com-
parative atomic force microscopy study on living skin fibroblasts and liver
endothelial cells. J Electron Microsc (Tokyo), 2001;50(4):283–90.
20. Zapotoczny, B., Szafranska, K., Owczarczyk, K., Kus, E., Chlopicki, S., and
Szymonski, M. Atomic force microscopy reveals the dynamic morphology
of fenestrations in live liver sinusoidal endothelial cells. Sci Rep, 2017;
7(1):7994.
21. Zapotoczny, B., Owczarczyk, K., Szafranska, K., Kus, E., Chlopicki, S., and
Szymonski, M. Morphology and force probing of primary murine liver sinu-
soidal endothelial cells. J Mol Recognit, 2017;30(7).
22. Svistounov, D., Warren, A., McNerney, G.P. et al. The relationship between
fenestrations, sieve plates and rafts in liver sinusoidal endothelial cells. PLoS
One, 2012;7(9):e46134.
23. Cogger, V.C., Roessner, U., Warren, A., Fraser, R., and Le Couteur, D.G. A
sieve‐raft hypothesis for the regulation of endothelial fenestrations. Comput
Struct Biotechnol J. 2013;8:e201308003.
24. Monkemoller, V., Schuttpelz, M., McCourt, P., Sorensen, K., Smedsrod, B.,
and Huser, T. Imaging fenestrations in liver sinusoidal endothelial cells by
optical localization microscopy. Phys Chem Chem Phys, 2014;16(24):12576–81.
25. Betzig, E. Single molecules, cells, and super‐resolution optics (Nobel lecture).
Angewandte Chemie (International edn in English). 2015;54(28):8034–53.
26. Hess, S.T., Girirajan, T.P., and Mason, M.D. Ultra‐high resolution imaging
by fluorescence photoactivation localization microscopy. Biophys J,
2006;91(11):4258–72.
27. Willig, K.I., Kellner, R.R., Medda, R., Hein, B., Jakobs, S., and Hell, S.W.
Nanoscale resolution in GFP‐based microscopy. Nat Methods,
2006;3(9):721–3.
28. Le Couteur, D.G., Fraser, R., Hilmer, S., Rivory, L.P., and McLean, A.J. The
hepatic sinusoid in aging and cirrhosis ‐ Effects on hepatic substrate disposi-
tion and drug clearance. Clin Pharmacokinet, 2005;44(2):187–200.
29. Henriksen, J.H., Horn, T., and Christoffersen, P. The blood‐lymph barrier in
the liver. A review based on morphological and functional concepts of nor-
mal and cirrhotic liver. Liver, 1984;4(4):221–32.
30. Fraser, R., Dobbs, B.R., and Rogers, G.W.T. Lipoproteins and the liver sieve:
the role of the fenestrated sinusoidal endothelium in lipoprotein metabolism,
atherosclerosis, and cirrhosis. Hepatology, 1995;21:863–74.
31. Fraser, R., Cogger, V.C., Dobbs, B. et al. The liver sieve and atherosclerosis.
Pathology, 2012;44:181–6.
32. Hardonk, M.J., Harms, G., and Koudstaal, J. Zonal heterogeneity of rat
hepatocytes in the in vivo uptake of 17 nm colloidal gold granules.
Histochemistry, 1985;83(5):473–7.
33. Romero, E.L., Morilla, M.J., Regts, J., Koning, G.A., and Scherphof, G.L.
On the mechanism of hepatic transendothelial passage of large liposomes.
FEBS Lett, 1999;448(1):193–6.
34. Hilmer, S.N., Cogger, V.C., Fraser, R., McLean, A.J., Sullivan, D., and Le
Couteur, D.G. Age‐related changes in the hepatic sinusoidal endothelium
impede lipoprotein transfer in the rat. Hepatology, 2005;42(6):1349–54.
35. Carpenter, B., Lin, Y., Stoll, S. et  al. VEGF is crucial for the hepatic
vascular  development required for lipoprotein uptake. Development,
­ Arias, I.M. Studies on fenestral contraction in rat liver endothelial cells in
2005;132(14):3293–303.
36. Cogger, V.C., Hilmer, S.N., Sullivan, D., Muller, M., Fraser, R., and Le
Couteur, D.G. Hyperlipidemia and surfactants: the liver sieve is a link.
Atherosclerosis, 2006;189(2):273–81.
37. Herrnberger, L., Hennig, R., Kremer, W. et  al. Formation of fenestrae in
murine liver sinusoids depends on plasmalemma vesicle‐associated protein
and is required for lipoprotein passage. PLoS One, 2014;9(12):e115005.
38. Mitchell, S.J., Huizer‐Pajkos, A., Cogger, V.C. et al. Age‐related pseudocap-
illarization of the liver sinusoidal endothelium impairs the hepatic clearance
of acetaminophen in rats. J Gerontol A Biol Sci Med Sci, 2011;66:400–8.
39. Mitchell, S.J., Huizer‐Pajkos, A., Cogger, V.C. et al. The influence of old age
and poloxamer‐407 on the hepatic disposition of diazepam in the isolated
perfused rat liver. Pharmacol, 2012;90(5–6):233–41.
40. Mohamad, M., Mitchell, S.J., Wu, L.E. et  al. Ultrastructure of the liver
microcirculation influences hepatic and systemic insulin activity and
provides a mechanism for age‐related insulin resistance. Aging Cell,
­
2016;15(4):706–15.
41. Raines, S.M., Richards, O.C., Schneider, L.R. et al. Loss of PDGF‐B activity
increases hepatic vascular permeability and enhances insulin sensitivity.
Am J Physiol Endocrinaol Metab, 2011;301(3):E517–26.
42. Jacobs, F., Wisse, E., and De Geest, B. The role of liver sinusoidal cells in
hepatocyte‐directed gene transfer. Am J Pathol, 2010;176(1):14–21.
43. Holz, L.E., Warren, A., Le Couteur, D.G., Bowen, D.G., and Bertolino, P.
CD8+ T cell tolerance following antigen recognition on hepatocytes.
J Autoimmun, 2010;34(1):15–22.
44. Warren, A., Bertolino, P., Benseler, V., Fraser, R., McCaughan G.W., and Le
Couteur, D.G. Marked changes of the hepatic sinusoid in a transgenic mouse
model of acute immune‐mediated hepatitis. J Hepatol, 2007;46:239–46.
45. Guidotti, L.G., Inverso, D., Sironi, L. et al. Immunosurveillance of the liver
by intravascular effector CD8(+) T cells. Cell, 2015;161(3):486–500.
46. McCuskey, R.S. Morphological mechanisms for regulating blood flow
through hepatic sinusoids. Liver, 2000;20(1):3–7.
47. Watanabe, N., Takashimizu, S., Nishizaki, Y., Kojima, S., Kagawa, T., and
Matsuzaki, S. An endothelin A receptor antagonist induces dilatation of
­sinusoidal endothelial fenestrae: implications for endothelin‐1 in hepatic
microcirculation. J Gastroenterol, 2007;42(9):775–82.
48. Popescu, D., Movileanu, L., Ion, S., and Flonta, M.L. Hydrodynamic effects
on the solute transport across endothelial pores and hepatocyte membranes.
Phys Med Biol, 2000;45(11):N157–65.
49. Van Der Smissen, P., Van Bossuyt, H., Charels, K., and Wisse, E. The struc-
ture and function of the cytoskeleton in sinusoidal endothelial cells, in Cells
of the Hepatic Sinusoid, (eds. A. Kirn, D.L. Knook, E. Wisse),The Kupffer
Cell Foundation, Rijswijk, 1986, pp. 517–22.
50. Braet, F., De Zanger, R., Baekeland, M., Crabbe, E., Van Der Smissen, P.,
and Wisse, E. Structure and dynamics of the fenestrae‐associated cytoskele-
ton of rat liver sinusoidal endothelial cells. Hepatology, 1995;21(1):180–9.
51. Steffan, A.M., Gendrault, J.L., and Kirn, A. Increase in the number of
fenestrae in mouse endothelial liver cells by altering the cytoskeleton with
cytochalasin B. Hepatology, 1987;7(6):1230–8.
52. Braet, F., De Zanger, R., Jans, D., Spector, I., and Wisse, E. Microfilament‐
disrupting agent latrunculin A induces and increased number of fenestrae in
rat liver sinusoidal endothelial cells: comparison with cytochalasin B.
Hepatology, 1996;24(3):627–35.
53. Gatmaitan, Z. and Arias, I.M. Hepatic endothelial cell fenestrae, in Cells of
the Hepatic Sinusoid, 4th edn, (eds. D.L. Knook and E. Wisse), KupfferCell
Foundation, Leiden, 1993. pp. 3–7.
54. Funyu, J., Mochida, S., Inao, M., Matsui, A., and Fujiwara, K. VEGF can act
as vascular permeability factor in the hepatic sinusoids through upregulation
of porosity of endothelial cells. Biochem Biophys Res Commun,
2001;280(2):481–5.
55. Cheluvappa, R., Hilmer, S.N., Kwun, S.Y., et al. The effect of old age on
liver oxygenation and the hepatic expression of VEGF and VEGFR2.
Exp Gerontol, 2007;42(10):1012–9.
56. Xie, G., Wang, X., Wang, L. et al. Role of differentiation of liver sinusoidal
endothelial cells in progression and regression of hepatic fibrosis in rats.
Gastroenterology, 2012;142(4):918–27.
57. Cogger, V.C., Arias, I.M., Warren, A. et  al. The response of fenestrations,
actin, and caveolin‐1 to vascular endothelial growth factor in SK Hep1 cells.
Am J Physiol Gastrointest Liver Physiol, 2008;295(1):G137–45.
58. Gatmaitan, Z., Varticovski, L., Ling, L., Mikkelsen, R., Steffan, A.M., and
culture. Am J Pathol, 1996;148(6):2027–41.
59. Venkatraman, L. and Tucker‐Kellogg, L. The CD47‐binding peptide of
thrombospondin‐1 induces defenestration of liver sinusoidal endothelial
cells. Liver Int, 2013;33(9):1386–97.
 35:  Fenestrations in the Liver Sinusoidal Endothelial Cell 443
60. Yokomori, H., Yoshimura, K., Funakoshi, S. et  al. Rho modulates hepatic
sinusoidal endothelial fenestrae via regulation of the actin cytoskeleton in rat
endothelial cells. Lab Invest, 2004;84(7):857–64.
61. Kamegaya, Y., Oda, M., Yokomori, H., and Ishii, H. Role of endothelin
receptors in endothelin‐1‐induced morphological changes of hepatic sinusoi-
dal endothelial fenestrae: morphometric evaluation with scanning electron
microscopy. Hepatol Res, 2002;22(2):89–101.
62. Takashimizu, S., Kojima, S., Nishizaki, Y. et  al. Effect of endothelin A
­receptor antagonist on hepatic hemodynamics in cirrhotic rats. Implications
for  endothelin‐1 in portal hypertension. Tokai J Exp Clin Med,
2011;36(2):37–43.
63. O’Reilly, J.N., Cogger, V.C., Fraser, R., and Le Couteur, D.G. The effect of
feeding and fasting on fenestrations in the liver sinusoidal endothelial cell.
Pathology, 2010;42(3):255–8.
64. Jamieson, H.A., Hilmer, S.N., Cogger, V.C., Warren, A., Cheluvappa, R., and
Abernethy, D.R., et al. Caloric restriction reduces age‐related pseudocapil-
larization of the hepatic sinusoid. Exp Gerontol, 2007;42(4):374–8.
65. Cogger, V.C., Mohamad, M., Solon‐Biet, S.M. et al. Dietary macronutrients
and the aging liver sinusoidal endothelial cell. Am J Physiol Heart Circ
Physiol, 2016;310(9):H1064–70.
66. Le Couteur, D.G., Cogger, V.C., Markus, A.M.A. et al. Pseudocapillarization
and associated energy limitation in the aged rat liver. Hepatology,
2001;33(3):537–43.
67. Le Couteur, D.G., Warren, A., Cogger, V.C. et al. Old age and the hepatic
sinusoid. Anat Rec, 2008;291(6):672–83.
68. Schaffner, F. and Popper, H. Capillarisation of hepatic sinusoids in man.
Gastroenterology, 1963;44:239–42.
69. Xu, B., Broome, U., Uzunel, M. et al. Capillarization of hepatic sinusoid by
liver endothelial cell‐reactive autoantibodies in patients with cirrhosis and
chronic hepatitis. Am J Pathol, 2003;163(4):1275–89.
70. Fraser, R., Rogers, G.W., Bowler, L.M., Day, W.A., Dobbs, B.R., and Baxter,
J.N. Defenestration and vitamin A status in a rat model of cirrhosis, in Cells
of the Hepatic Sinusoid, 2nd edn, (eds. , E. Wisse, D.L. Knook, and R.S.
McCuskey),Kupffer Cell Foundation, Leiden, 1991, pp. 195–8.
71. Mori, T., Okanoue, T., Sawa, Y., Hori, N., Ohta, M., and Kagawa, K.
Defenestration of the sinusoidal endothelial cell in a rat model of cirrhosis.
Hepatology, 1993;17(5):891–7.
72. Jacobs, F., Wisse, E., and De Geest, B. Early effect of a single intravenous
injection of ethanol on hepatic sinusoidal endothelial fenestrae in rabbits.
Comp Hepatol, 2009;8:4.
73. Vreuls, C.P., Driessen, A., Olde Damink, S.W. et al. Sinusoidal obstruction
syndrome (SOS): A light and electron microscopy study in human liver.
Micron, 2016;84:17–22.
74. DeLeve, L.D. Hepatic microvasculature in liver injury. Semin Liver Dis,
2007;27(4):390–400.
75. DeLeve, L.D., Shulman, H.M., and McDonald, G.B. Toxic injury to hepatic
sinusoids: sinusoidal obstruction syndrome (veno‐occlusive disease). Semin
Liver Dis, 2002;22(1):27–42.
76. DeLeve, L.D., Ito, Y., Bethea, N.W., McCuskey, M.K., Wang, X., and
McCuskey, R.S. Embolization by sinusoidal lining cells obstructs the
microcirculation in rat sinusoidal obstruction syndrome. Am J Physiol,
­
2003;284(6):G1045–52.
77. DeLeve, L.D., Wang, X., Kanel, G.C. et al. Decreased hepatic nitric oxide
production contributes to the development of rat sinusoidal obstruction
­syndrome. Hepatology, 2003;38(4):900–8.
78. Vidal‐Vanaclocha, F. and Barbera‐Guillem, E. Fenestration patterns
in  endothelial cells of rat liver sinusoids. J Ultrastruct Res,
1985;90(2):115–23.
79. Fraser, R., Heslop, V.R., Murray, F.E., and Day, W.A. Ultrastructural studies
of the portal transport of fat in chickens. Br J Exp Pathol, 1986;67:783–91.
80. Warren, A., Bertolino, P., Cogger, V.C., McLean, A.J., Fraser, R., and Le
Couteur, D.G. Hepatic pseudocapillarization in aged mice. Exp Gerontol,
2005;40(10):807–12.
81. Fraser, R., Day, W.A., and Fernando, N.S. Atherosclerosis and the liver sieve,
in Cells of the Hepatic Sinusoid, 1st edn, (eds. A. Kirn, D.L. Knook, and
E. Wisse), The Kupffer Cell Foundation, Rijswijk, 1986, pp. 317–22.
82. McCuskey, P.A., McCuskey, R.S., and Hinton, D.E. Electron microscopy of
the cells of the hepatic sinsusoids in rainbow trout, in Cells of the Hepatic
Sinusoid, 1st edn, (eds. A. Kirn, D.L. Knook, and E. Wisse), The Kupffer
Cell Foundation, Rijswijk, 1986, pp. 489–94.
83. Nopanitaya, W., Carson, J.L., Grisham, J.W., and Aghajanian, J.G. New
observations on the fine structure of the liver in goldfish (Carassius auratus).
Cell Tissue Res, 1979;196(2):249–61.
84. McCuskey, R.S., McCuskey, P.A., Mitchell, D.B., Dezanger, R.B., and
Wisse, E. Ultrastructure of the canine hepatic sinusoid, in Cells of the
Hepatic Sinusoid, 1st edn, (eds. A. Kirn, D.L. Knook, and E. Wisse),
The Kupffer Cell Foundation, Rijswijk, 1986, pp. 509–10.
85. Wright, P.L., Smith, K.F., Day, W.A., and Fraser, R. Hepatic sinusoidal
endothelium in sheep: an ultrastructural reinvestigation. Anat Rec,
1983;206(4):385–90.
86. Jamieson, H.A., Cogger, V.C., Twigg, S.M. et al. Alterations in liver sinusoi-
dal endothelium in a baboon model of type 1 diabetes. Diabetologia,
2007;50(9):1969–76.
87. Cogger, V.C., Warren, A., Fraser, R., Ngu, M., McLean, A.J., and Le Couteur,
D.G. Hepatic sinusoidal pseudocapillarization with aging in the non‐human
primate. Exp Gerontol, 2003;38(10):1101–7.
88. Mak, K.M. and Lieber, C.S. Alterations in endothelial fenestrations in liver
sinusoids of baboons fed alcohol: a scanning electron microscopic study.
HeHepatologypatol, 1984;4(3):386–91.
89. Horn, T., Henriksen, J.H., and Christoffersen, P. The sinusoidal lining cells
in “normal” human liver. A scanning electron microscopic investigation.
Liver, 1986;6(2):98–110.
90. Madarame, T., Nagaoka, T., Inomata, M. et  al. Basement membrane‐like
structure may play a role in the alteration of endothelial fenestration, in Cells
of the Hepatic Sinusoid, 3rd edn, (eds. E. Wisse, D.L. Knook, and R.S.
McCuskey), Kupffer Cell Foundation, Leiden, 1991, pp. 98–101.
91. Bennett, H.S., Luft, J.H., and Hampton, J.C. Morphological classification of
vertebrate blood capillaries. Am J Physiol, 1959;196:381–90.
92. Bloom, W. and Fawcett, D.W. A Textbook of Histology. 9th edn, Saunders,
Philadelphia, 1968.
 Stellate Cells and Fibrosis
36 Youngmin A. Lee1 and Scott L. Friedman2
1
2
Department of Surgery, Vanderbilt University Medical Center, Nashville, TN, USA
Division of Liver Diseases, Icahn School of Medicine at Mount Sinai, New York, NY, USA
INTRODUCTION
The identification and isolation of hepatic stellate cells (HSCs)
as the main fibrogenic cells in liver, as well as techniques to
establish cultured HSC lines have enabled significant progress
over the past two decades in understanding key events in the
transdifferentiation process of quiescent HSCs to fibrogenic
myofibroblasts (MFBs). Together with the implementation of
potent direct acting antivirals for the treatment of chronic viral
hepatitis, there is now clear evidence that fibrosis and even cir-
rhosis can regress. Advances in clarifying HSC biology and
responses to injury have led to the promise of effective antifi-
brotic therapies in the near future.
EPIDEMIOLOGY AND DISEASES
UNDERLYING HEPATIC FIBROSIS
Hepatic fibrosis due to chronic liver disease (CLD) represents a
major global health burden. It is estimated that over 840 million
people are affected by CLD, contributing to two million deaths
annually [1, 2]. Population‐based studies estimate the US preva-
lence of cirrhosis at 630 000 adults, of which 2/3 of patients are
unaware of having liver disease [3]. Chronic hepatitis B and C
are the most frequent underlying liver diseases and are also
globally endemic, with particularly high prevalence in Asia and
Sub‐Saharan Africa (up to 8%) [4], while alcoholic liver disease
is more prevalent in Europe and the United States (around 12%)
[5]. An epidemic rise in obesity in western countries has also led
to a massive increase in the prevalence of non‐alcoholic fatty
liver disease (NAFLD) (up to 46%) and non‐alcoholic steato-
hepatitis (NASH) (up to 16% in the United States). Other less
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
common diseases that can lead to cirrhosis include autoimmune
hepatitis, hemochromatosis, Wilson’s disease, and primary
and  secondary biliary cholangitis (see Table  36.1) [1, 2].
Traditionally, the development of antifibrotic therapies has been
focused on treatment of the underlying liver disease. With the
arrival of curative direct acting antivirals for chronic hepatitis C
and improved therapies for chronic hepatitis B virus (HBV)
infection, combined with preventive measures such as vaccina-
tion for HBV, the prevalences of these diseases are expected to
drop (Table  36.2). In contrast, lifestyle changes necessary for
improvement of obesity and metabolic syndrome‐associated
NAFLD and NASH are hard to implement; thus, development
of antifibrotic therapies has significantly intensified over the
past years for NASH, with a multitude of drugs now in the pipe-
line and in clinical trials.
HEPATIC FIBROSIS
The liver is a highly regenerative organ; however, repeated and
sustained epithelial injury leads to a wound‐healing response
with excessive deposition of extracellular matrix (ECM), even-
tually impairing the liver’s regenerative capacities and leading
to fibrosis. Fibrillar collagens I and III, which normally make up
about 3% of the liver by weight, are deposited by activated
HSCs, the main fibrogenic cell population in the liver. In cir-
rhotic liver, the content of fibrillar collagens I and III increases
three- to ten-fold and sulfated proteoglycans and ECM proteins
are deposited at the primary site of injury and within the suben-
dothelial space of Disse, at the interface of hepatocytes and
liver  sinusoidal endothelial cells where HSCs reside. While
the  molecular composition of the ECM in fibrotic diseases is
 36:  Stellate Cells and Fibrosis 445

Table 36.1  Causes of fibrotic liver diseases Table 36.3  Extracellular cell matrix (ECM) components and normal
Disease localization in liver. Adapted from [105]
ECM Normal
Alcoholic liver disease
Chronic infections Collagens  
Viruses Type I Portal tract/hepatic vein
Hepatitis B Type III Portal tract, space of Disse
Hepatitis C Type IV Portal tract basement, space of Disse
Hepatitis delta Type V Portal tract, space of Disse
Bacterial Type VI Portal tract, space of Disse
Brucellosis Glycoproteins  
Parasitic Laminin Portal tract basement membranes, space of Disse
Echinococcus Fibronectin Portal tract, space of Disse
Schistosomiasis Elastin Portal tract matrix
Autoimmune hepatitis Entactin Portal tract BM
Non‐alcoholic steatohepatitis Fibrillin Portal tract matrix
Biliary cirrhosis Proteoglycans  
Primary biliary cirrhosis Heparansulfate Portal tract basement membranes
Primary sclerosing cirrhosis
Autoimmune cholangiopathy
Obstructive cholestasis The space of Disse is usually comprised of delicate strands of
Inherited metabolic diseases
Alpha antitrypsin deficiency collagen III, IV, and VI, however, during injury fibrillar colla-
Cystic fibrosis gens, laminin and fibronectin are deposited within the space of
Hereditary hemochromatosis Disse, thus forming a major solute barrier that inhibits exchange
Wilson disease
Glycogen storage diseases
between hepatocytes and plasma (see Table 36.3). Liver sinusoi-
Fructosemia dal endothelial cells, which have fenestrae allowing for solute
Galactosemia exchange, lose these pores as excess matrix is deposited in a
Lysosomal storage diseases (e.g. Fabry’s, Gaucher’s diseases) process known as “capillarization”. In murine fibrosis models
Porphyrias
Urea cycle defects ECM content is increased to up to 25–40‐fold. All these altera-
Cryptogenic cirrhosis tions eventuate in advanced liver disease with cirrhosis with
Drug induced liver injury (e.g. MTX, α‐methyl DOPA amiodarone, portal hypertension and formation of arteriovenous shunts,
isoniazid, antibiotics)
Pediatric liver disorders
which are associated with impaired synthetic and metabolic
Congenital hepatic fibrosis function of the liver.
Biliary atresia The natural course of most chronic liver disease requires
Congenital biliary cysts ­decades for advanced fibrosis to develop, and its asymptomatic
Vascular diseases
Budd‐Chiari syndrome and painless course through most of that interval allows it to
Congestive hepatopathy (cardiac cirrhosis) remain undetected for up to decades. Once advanced fibrosis
Hereditary hemorrhagic telangiectasia develops, it portends a growing increased risk for hepatocellular
Sinusoidal obstruction syndrome (formerly “veno‐occlusive disease”) carcinoma (HCC), which currently has the fastest rising cancer
incidence in the world. Moreover, HCC is already the third lead-
Table 36.2  Current and estimated future global incidences and  ing cause of cancer‐related death worldwide. Curative therapies
prevalence of liver diseases. Adapted from [2] for HCC are only possible through ablative or surgical tech-
Incidence Prevalence Current estimation Future estimation niques when lesions are smaller and fewer in number, whereas
Disease (millions) (%) (millions) 2030 (millions) advanced HCCs are rarely curable. Although cirrhosis is the
HBV 4.5–6 3.6 240 120 single most important risk factor, apart from regular screening
HCV 3–4 2.5 170 85 and treatment of the underlying liver disease, no specific chemo-
ALD 16.6 4.5 N/A 19.3
NAFLD 13.6 5–8 570 16.2 preventive therapies are yet available (Chapter 16).
NASH 2.5 less than 4 145 3.8
HBV, hepatitis B virus; HCV, hepatitis C virus; ALD, alcoholic liver disease;
NAFLD, non‐alcoholic fatty liver disease; NASH, non‐alcoholic steatohepatitis.
CIRRHOSIS
similar across fibrosis of different organs and etiologies, fibrosis
patterns vary significantly, indicating disease‐specific mecha- The term “cirrhosis” was coined by the French physician Rene
nisms of fibrogenesis. In liver, patterns of fibrosis include: T. H. Lannaëc in 1812 after the Greek word “kirrhos” for the
(i) portal–central fibrotic septae and nodule formation (e.g. auto- tawny, yellow nodules associated with this disease [6] and was
immune hepatitis, hepatitis C), (ii), perisinusoidal/pericellular adapted internationally as a synonym for end‐stage liver dis-
fibrosis (also called “chicken wire” fibrosis) which is typical for ease. Cirrhosis is characterized by fibrous septae that are present
alcoholic liver disease and non‐alcoholic steatohepatitis, throughout the liver, subdividing the organ into parenchymal
(iii) biliary cirrhosis with portal–portal fibrosis and proliferation nodules, which may vary between micronodules (less than
of bile ducts, which is mainly observed in cholestatic liver dis- 3 mm) or macronodules (3 mm to several cm). As noted above,
eases and pediatric biliary atresia, (iv) centrilobular fibrosis the fibrous septae may be present as thin bands or in different
with development of central–central vein fibrotic septa as a con- patterns such as portal‐to‐portal, central‐to‐central, or also por-
sequence of chronic hepatic venous stasis (e.g. heart failure). tal‐to‐central patterns with portal tract‐based fibrosis, frequently
446 THE LIVER:  COMMON MECHANISMS OF LIVER INJURY CONTRIBUTING TO HEPATIC FIBROSIS
producing a more severe and irregular fibrosis pattern. While it
was once thought that cirrhosis is an irreversible, progressive
disease, evidence from clinical trials show that regression of
­cirrhosis is possible upon eradication of the underlying liver dis-
ease in a broad range of diseases from viral hepatitis B and C
[7–9], to hemochromatosis [10, 11], autoimmune hepatitis [12],
Wilson disease [13], and also in biliary fibrosis such as biliary
cholangitis [14]. Depending on the underlying liver disease and
the degree of fibrosis it may take years for fibrotic tissue to be
resorbed, with improved liver function and histology, however,
the liver’s architecture cannot be completely reconstituted and
incomplete septal cirrhosis may remain. Irreversible changes
which are evidence of a regressed cirrhosis are vascular aberra-
tions that include venous obliterative lesions likely resulting
from thrombosis and to a lesser extent portal venous outflow
obstruction. Loss of hepatocytes leads to collapse of liver paren-
chyma which may lead to narrowing of portal tracts and central
veins, allowing for the formation of fast channels in which
hepatic arterioles and portal veins directly drain into hepatic
veins, bypassing the perfusion of liver parenchyma. The forma-
tion of these portal–hepatic and arterio–venous shunts further
contributes to loss of both liver function and regenerative capac-
ity due to impaired access to nutrition and key substrates that
support hepatocyte homeostasis. Persistence of vascular changes
can be observed in fibrosis that has regressed, which may be a
contributing factor to sustained portal hypertension even after
improvement of liver function. Similarly, it remains unclear if
the loss of fenestrae in liver sinusoidal endothelia during fibro-
sis progress is reversible upon its cessation, which is a require-
ment to restore full metabolite exchange and secretory functions
of hepatocytes.
Alterations in liver structure and function that persist after
regression of fibrosis, which are not as easily detectable as vas-
cular aberrations, also include changes in the key fibrogenic cell
population, the HSCs. Animal models where fibrosis reverses
after cessation of hepatocellular injury demonstrate that acti-
vated and fibrogenic HSCs may revert to an inactivated pheno-
type, allowing for resorption and fibrosis improvement [15, 16].
However, upon renewed liver injury inactivated stellate cells
may reactivate faster and become more fibrogenic than truly
quiescent HSCs, which could explain why patients with ongo-
ing liver injury have acceleration of fibrosis progression as the
stage of fibrosis advances. Evaluation of liver specimens from
patients with hepatitis C‐related fibrosis, who have been fol-
lowed up after virological cure (SVR, sustained virological
response) demonstrate that elimination of the underlying liver
disease leads to a regression of fibrosis in the majority of
patients. However, a small but consistent fraction (1–14% of
patients, depending on the study) experience fibrosis progres-
sion with worsening of either fibrosis score or collagen propor-
tionate area based on liver biopsies (see overview in [17]). Liver
biopsies from patients with chronic HCV who achieved SVR
show return to normal lobular metabolic zonation as assessed by
staining for glutamine synthase (GS) and CYP2E1; however,
staining for alpha smooth muscle actin (αSMA), a marker of
myofibroblasts (MFB), persists in MFB before and after SVR.
Moreover, a significant fraction of patients (31%) display a
worsening of αSMA score after SVR [18], indicating persistent
remodeling of HSCs despite virologic cure. Enigmatic, but
fascinating observations suggest that HSCs might retain a
transgenerational memory of insult with decreased susceptibil-
ity to hepatotoxins and liver fibrosis in offspring (F1 and F2) of
rats (F0) that had been subjected to experimental liver fibrosis
[19]. In this seminal study, the authors detected decreased num-
bers of liver MFBs and increased expression of the antifibro-
genic mediator PPARγ, with decreased expression of
profibrogenic TGFβ in livers from F1 and F2 rats. Sperm from
rats were enriched in H2A.Z and H3K27me3 indicating epige-
netic regulation of fibrosis that requires further elucidation.
Cirrhosis is a highly dynamic process with constant meta-
bolic changes of the fibrotic tissue. This has been demonstrated
in patients with chronic hepatitis C in which liver function was
assessed by metabolic labeling with “heavy‐water”, a non‐­
radioactive but detectable labeling with deuterated water (2H2O),
which is incorporated into amino acids and proteins. Tandem
mass spectrometry measured high turnover rates of collagen
I and III and collagen‐associated proteins, indicating a high rate
of collagen remodeling even in advanced fibrosis [20].
These findings underscore that cirrhosis is a simplified term
for a range of stages of advanced liver disease, with varying prog-
noses and levels of reversibility. Although the term “cirrhosis”
had been used for more than 200 years, an international liver
pathology study group has proposed abandoning this term in
favor of a more refined diagnosis that incorporates the grade of
activity, features of progression and regression, and risk f­ actors
for malignancy, among others [21].
COMMON MECHANISMS OF LIVER
INJURY CONTRIBUTING TO HEPATIC
FIBROSIS
Oxidative and nitrosative stress, inflammation, angiogenesis,
and apoptosis are processes which are closely interrelated and
common to chronic liver diseases of different etiologies.
Oxidative stress
Reactive oxygen species (ROS) and reactive nitrogen species
(RNS) are generated during the consumption of oxygen or
are derived from nitric oxide (NO) and superoxide by inducible
nitric oxide (iNOS) and NADPH oxidases, respectively.
Endothelial nitric oxide, hepatocytic cytochrome P450
monooxygenases (e.g. CYP2E1), and NADPH oxidases (NOX)
in Kupffer cells and HSCs can contribute to generation of ROS/
RNS, which react indiscriminately with all major biologically
relevant macromolecules (DNA, lipids, proteins, cytoskeletal
proteins, mitochondria) leading to DNA strand breaks, damag-
ing and inactivating metabolic pathways that provoke cell death.
ROS/RNS inhibit parenchymal regeneration and enhance fibro-
genesis and inflammation by acting as chemotactic agents
towards HSCs and by directly activating HSCs by inducing
­proliferation, collagen I synthesis, and TGFβ production [22].
ROS directly activate Kupffer cells to produce proinflammatory
cytokines such as TNFα, and to amplify the inflammatory
response by stimulating resident and non‐resident liver cells to
produce proinflammatory and profibrogenic mediators [23].
 36:  Stellate Cells and Fibrosis 447
Other reactive oxygen species such as malondialdehyde (MDA)
and 4‐hydroxy‐2,3‐nonenal (HNE) can form protein adducts
which confer antigenic properties, thus inducing antibodies that
could promote immune‐mediated hepatocellular injury [24] in
patients with NASH and alcoholic steatohepatitis (ASH). Risk
factors which amplify ROS/RNS generation include endoge-
nous conditions such as obesity, insulin resistance, bile, and
exogenous drivers such as alcohol, drugs, pollutants, environ-
mental toxins, viruses, and ultraviolet light. An oral NOX1 and
NOX4 inhibitor, GKT137831 attenuates liver fibrosis and apop-
tosis in rodent models [25] and is currently being evaluated for
treatment of chronic liver disease.
Hypoxia/angiogenesis
Liver injury leads to activation of HSCs and subendothelial
ECM deposition, promoting capillarization of liver sinusoids
that increases resistance to blood flow and compromises oxygen
delivery. Loss in size and numbers of fenestrae in liver sinusoi-
dal endothelial cells impairs hepatocyte and mitochondrial
function and enhances inflammation. Hypoxia inducible factors
HIF1α and HIF2α increase transcription and synthesis of angio-
genic factors including VEGF, Ang‐1, and their receptors
VEGFR2 and Tie2, as well as PAI‐1, and adrenomedullin‐1
(ADM1) and ADM2. In addition to VEGF’s profibrogenic func-
tion toward HSCs it also promotes endothelial cell proliferation
[26–28]. PDGF, leptin (which may upregulate VEGF and
Ang‐1) and HGF also have proangiogenic functions [29, 30].
Apoptosis
Apoptosis, or programmed cell death has been implicated in many
chronic liver diseases. Apoptotic bodies are a strong profibrogenic
stimulus for HSCs in vivo and in vitro and is mediated in part by the
death receptor Fas and TRAIL [31–33]. The relationship between
fibrosis and apoptosis may be bidirectional, as fibrosis may
stimulate apoptosis in parenchymal cells due to induction of proa-
poptotic genes while stimulating survival in activated HSCs. Thus,
some therapeutic approaches that focus on inhibiting hepatocyte
apoptosis could also enhance HSC s­ urvival. Newer forms of pro-
grammed cell death may also stimulate fibrosis, for example,
necroptosis, which is activated by the necrosome consisting of
receptor‐interacting serine/threonine‐protein kinase1 (RIPK1) and
RIPK3 and mixed lineage kinase domain‐like protein (MLKL).
This pathway might be s­ pecifically important in NAFLD more
than in other types of liver injury (for review see [34]).
CELLULAR BASIS OF FIBROGENESIS
IN LIVER
Hepatic MFBs are a heterogenous population and mediate the
excessive deposition of ECM in fibrosis. They are contractile
and highly secretory cells that can be identified in part by their
expression of αSMA. Hepatic MFBs transdifferentiate from
quiescent resident mesenchymal cells, primarily stellate cells,
but also from periportal fibroblasts in biliary liver injuries asso-
ciated with portal fibrosis.
HSCs were first described by Kupffer in 1875 as “star‐
shaped” cells. Their function remained unknown for the next 75
years until Ito described lipid‐containing perisinusoidal cells
(lipocytes or Ito cells), but it was not until Wake confirmed that
these were the same cells that Kupffer had described. Quiescent
HSCs only represent 4–8% of cells in healthy liver and are char-
acterized by their storage of retinyl‐esters in cytoplasmic lipid
droplets, which allowed for their isolation and purification by
gradient centrifugation. This technique was crucial toward iden-
tifying HSCs as the main fibrogenic cells and to characterizing
their function. It is now clear that HSCs contribute to hepatic
development, regeneration, immune responses, angiogenesis,
and are the body’s main storage site of vitamin A, of which up
to 90% are stored in the HSCs [35].
In normal liver, HSCs maintain a quiescent, non‐prolifera-
tive phenotype. Upon liver injury, HSCs transdifferentiate to
highly secretory MFBs and acquire chemotactic, inflamma-
tory, and highly proliferative phenotype, losing their retinoid
droplets in the process. Activated HSCs produce most of the
matrix proteins of fibrotic liver, including fibrillar and non‐
fibrillar collagens, components of the organized sinusoidal
basement membrane (collagen IV, laminin and perlecan) and
proteoglycans [36–38]. Portal fibroblasts, which are located
around the portal vein at the sinusoidal inlet, may also contrib-
ute to the pool of MFBs in biliary fibrosis, although HSCs are
still the main source of MFBs [39–42]. Portal fibroblasts
likely maintain integrity of small bile ducts and interact with
cholangiocytes.
HSCs can be morphologically distinguished from portal
fibroblasts by their inclusion of vitamin A droplets and by the
expression of the following genes: desmin, glial fibrillary acidic
protein (GFAP) (see Figure  36.1), lecithin retinol acyltrans-
ferase (Lrat), Hand2, vimentin, PDGFRβ, cytoglobin, and ree-
lin, though there are some species‐dependent differences. For
example, human HSCs do not express desmin at all, while it is a
very reliable marker of mouse HSCs. Portal fibroblasts may be
differentiated from HSCs by their expression of ecto‐ATPase
nucleoside triphosphate diphosphohydrolase‐2 (NTPDase 2)
Figure 36.1  Hepatic stellate cells in normal liver. Immunohisto­
chemistry for glial fibrillary acidic protein (GFAP, dark red) in normal
mouse liver. Note their long processes and relatively small size com-
pared to hepatocytes.
448 THE LIVER:  PATHWAYS OF HSC ACTIVATION
Table 36.4  Markers of hepatic stellate cells (HSCs) and portal
fibroblasts
Hepatic stellate cells Portal myofibroblasts
GFAP Ecto‐ATPase nucleoside triphosphate
Synaptophysin diphosphohydrolase‐2 (NTPDase2)
CRBP1 Fibulin 2 P100
Desmin a2‐macroglobulin
Hand2 Synaptophysin
Reelin NCAM
fibulin‐2, elastin, thymocyte differentiation, Gremlin 1, meso-
thelin (Msln), and cellular retinol‐binding protein‐1 (Creb1)
[43–45]. Other differences from HSCs might include the com-
position of ECM each cell type secretes (see Table 36.4). ECM
deposited from HSC‐derived myofibroblasts is fibrillin‐1‐posi-
tive and elastin‐negative and thus differs from ECM from portal
fibroblasts‐derived MFBs, which is fibrillin‐1‐ positive and
elastin‐positive [46] though further studies are warranted but
currently limited due to fewer available tools for portal fibro-
blasts compared to HSCs. Single cell sequencing of mesenchy-
mal cell populations in liver will be required to definitively
characterize fibrogenic cell types in liver. Immortalized rat por-
tal MFB have been recently described [47], but there are no
mouse reporter lines developed to study portal fibroblasts in
vivo. In addition, isolation techniques and markers have species‐
specific differences for portal MFBs between rat and mouse
(personal observation).
The development of tools and murine models to study HSCs in
vivo has greatly enhanced the characterization of HSC biology,
including the availability of immortalized human (LX2, TWNT‐4)
and murine HSC lines [48, 49], transgenic reporters driving Cre
recombinase for stellate cell‐specific conditional deletion in liver
(lecithin retinol acetyltransferase, Lrat) [40], and murine models
to deplete HSCs in vivo [50, 51]. The growing capacity to com-
pare large datasets has enabled comprehensive comparisons of
gene expression between quiescent and activated HSCs, and
between HSCs and other liver cell populations, which have
helped identify HSC specific transcripts and potential therapeutic
targets [52]. Fluorescent‐­tagging of HSCs by bHLH transcription
factor (Tg:Hand2‐EGFP) in zebrafish allows for the characteriza-
tion of HSCs during liver development in this species, and has
further identified LSECs as pivotal to enable the migration and
localization of HSCs into the liver [53].
PATHWAYS OF HSC ACTIVATION
The process of HSC activation from quiescent vitamin A‐rich
cells into fibrogenic MFBs has been well characterized and has
yielded a plethora of antifibrotic targets being pursued in clini-
cal trials. HSC activation conceptually consists of two phases,
initiation and perpetuation which could be followed by either
apoptosis, senescence, or reversion to quiescence of HSCs if the
underlying insult has been resolved, thus contributing to resolu-
tion of fibrosis.
During initiation, quiescent HSCs become responsive to
stimuli of the injured liver and triggers of their environment by
upregulating growth factor receptors and modulation of growth
factor signaling leading to perpetuation of myofibroblast phe-
notypes. While HSCs are evenly and ubiquitously distributed
throughout the liver lobules in normal liver, during liver injury
HSCs become most activated in zones of the most severe liver
injury [54] and inflammation. Perpetuation of HSC activation is
characterized by events that amplify their activated phenotype
with distinctive features such as proliferation, contractility,
fibrogenesis, matrix degradation, chemotaxis, inflammatory
signaling, and loss of retinoids. These features are predomi-
nantly driven by specific sets of cytokines (Figure 36.2).
Transforming growth factor‐β (TGFβ)
TGFβ is a key mediator in hepatic fibrogenesis [55]. TGFβ is
synthesized in a pro‐form and remains bound to latent TGFβ‐
binding protein (LTBP) and latency associated peptide (LAP) in
a biologically inactive state within the ECM. Latent TGFβ may
be activated by matrix metalloproteinases (MMPs), plasmin,
plasminogen activators, avβ6 integrins, thrombospondins,
Kupffer cells, and other cell populations [56]. Mechanisms of
activating TGFβ may differ across different tissues and even in
different types of liver disease. In HSCs, TGFβ is profibrogenic
through its cell surface receptors, leading to downstream signal-
ing via Smad3 that induces collagen I, collagen, III and fibronec-
tin expression and inducing chemotaxis. TGFβ also induces
mitogen‐activated proteinase (MAPK) pathways (ERK, p38
MAPK, and JNK) independently of Smads [57, 58].
Because systemic inhibition of TGFβ promotes tumorigene-
sis and inflammation, liver‐ or cell type‐specific inhibition of
TGFβ would be ideal, since systemic neutralization of TGFβ
might decrease fibrosis but could also elevate the risk of hepato-
cellular or biliary cancers [59, 60].
Platelet‐derived growth factor B (PDGF‐B)
PDGF‐B is the most potent mitogen and chemoattractant for
HSCs. PDGF signals through the tyrosine kinase receptor β‐
PDGFR, which is rapidly upregulated during initiation in HSCs
in humans and in rodent studies [61, 62], thereby amplifying
PDGF-B signaling in HSCs. Major sources of PDGF-B are
HSCs themselves, macrophages, and platelets [61, 63]. In murine
models of liver fibrosis, conditional deletion of β-PDGFR on HSCs
led to decreased liver injury and fibrosis [64]. In preclinical
­studies, HSC specific delivery of erlotinib via nanoparticles that
bind to β‐PDGFR [65] has been proposed as a novel approach to
cancer chemoprevention based on animal studies. Additional
HSC mitogens include VEGF, basic fibrocyte growth factor
(bFGF), thrombin, TGFα, and keratinocyte growth factor [66].
Connective tissue growth factor (CTGF/CCN2)
CTGF is a promising antifibrotic target. CTGF is a profibro-
genic “CCN” protein which is mainly expressed by HSCs in
injured liver, thereby promoting fibrogenesis, adhesion, migra-
tion, and cell survival [67]. Inhibition of CTGF by human anti‐
CTGF antibodies (FG‐3019) are in clinical trials for pulmonary
fibrosis but the trial patients with hepatitis B was halted
because patients on antiviral therapies had significant fibrosis
improvement without the antibody therapy (ClinicalTrials.gov
ID #NCT01217632).
 36:  Stellate Cells and Fibrosis 449

Figure 36.2  Functions and features of hepatic stellate cells (HSCs) in healthy and diseased liver. Quiescent hepatic stellate cells activate upon
liver injury, entering from an initiation phase into perpetuation. HSCs transdifferentiate to myofibroblasts with specific phenotype changes includ-
ing proliferation, contractility, fibrogenesis, altered matrix degradation, chemotaxis, and inflammatory signaling. From [106]. Reproduced with
permission of BMJ Publishing Group Ltd.

Endothelin‐1 (ET1) preclinical models of liver injury (NASH and TAA‐induced

ET1 is a major regulator of HSC contractility. HSCs are located
in collagenous bands between nodules of fibrotic liver and may
constrict portal blood flow in individual sinusoids thus contrib-
uting to portal hypertension [68].
Vascular endothelial growth factor (VEGF)
VEGF is released from liver sinusoidal endothelial cells and
HSCs in injured liver. It is a potent mitogen and stimulates HSC
migration and collagen production. While angiogenesis is a
pathogenic process in advanced liver disease, it may also be a
requirement for liver regeneration [69, 70]. Thus, efforts to
block angiogenesis in liver disease should ideally be restricted
to tumor related angiogenesis and not angiogenesis that accom-
panies and supports regeneration.
liver fibrosis) [73]. Phase two clinical trials in patients with
NASH and fibrosis have demonstrated significant antifibrotic
activity compared to placebo treated patients after one year of
therapy [74], however the benefit did not extend to two years.
HSCs may also function as antigen presenting cells [75].
Tissue inhibitor of metalloproteinases (TIMP)
and metalloproteinases (MMPs)
Activated HSCs produce tissue inhibitors of metalloproteinases
(TIMPs) TIMP1 and TIMP2, of which TIMP1 is antiapoptotic
toward HSCs in part through induction of bcl‐2, thus promoting the
survival of fibrogenic cells [76]. MMP2 and MMP9, which are also
produced by HSCs, allow for disruption of normal hepatic matrix
that could hasten its replacement by fibrotic matrix [77–79].

Immunoregulation
HSCs produce several immunoregulatory chemokines including
macrophage inflammatory protein‐2 (MIP2) [71, 72] MCP‐1,
CCl2, RANTES, and CCR5 which activate and recruit mac-
rophages and other immune cells. The CC chemokines RANTES
(CCL5), MPC‐1, and CCL21 may also directly promote prolif-
eration and migration by HSCs. The dual CCR2/CCR5
chemokine receptor antagonist Cenicriviroc is antifibrotic in
METABOLIC REGULATION OF HSCS
FUNCTION AND ACTIVATION
Adipokines
Adipokine signaling has emerged as an important regulator of
HSC activation in the context of metabolic syndrome. Adipokines
including leptin and resistin secreted from adipose tissue
450 THE LIVER:  EXTRACELLULAR EVENTS THAT LEAD TO ACTIVATION OF HSCS

promote profibrogenic effects on liver. Leptin has been clearly Hepatocytes

associated with HSC fibrogenesis [80–82] increasing expression
of αSMA, collagen 1α1, and TGFβ in HSCs as well as proin-
flammatory and proangiogenic cytokines [83–85]. In contrast,
adiponectin may have protective effects by decreasing fatty acid
oxidation and inhibiting hepatic gluconeogenesis. In human
and murine models, lower adiponectin levels have been associ-
ated with increased severity of liver inflammation [83, 84]
Autophagy
HSCs are highly dependent on autophagy to generate energy sub-
strates that fuel cellular activation. This occurs by autophagy‐
mediated cleavage of retinyl esters within cytoplasmic droplets to
generate free fatty acids [86]. HSC‐specific genetic deletion of
autophagy related protein 7 (Atg7) in murine models of liver fibro-
sis and in cultured HSCs both lead to reduced fibrogenesis [86].
EXTRACELLULAR EVENTS THAT LEAD
TO ACTIVATION OF HSCS
HSCs have complex interactions with all liver cell types and
may be activated or inhibited by them in an etiology‐specific or
generalized manner (see Figure 36.3).
Disease‐specific mechanisms may lead to hepatocyte injury and
release of ROS, Hedgehog ligands, nucleotides, lipid peroxides,
and cytokines which contribute to HSC activation. Dead or
dying hepatocytes release damage‐associated molecular pat-
terns (DAMPS) which amplify an inflammatory response. IL‐33
[87] activates resident innate lymphoid cells (ILC2) to release
IL13, acting through type II IL‐4 receptor dependent signaling
via signal transducer and activator of transcription 6 (STAT6) to
activate HSCs [88] and promote fibrosis.
Kupffer cells and macrophages
Tissue resident macrophages, Kupffer cells, produce ROS and
TGFβ which not only promote transdifferentiation of HSCs into
MFBs, but TGFβ is also a potent profibrogenic cytokine, stimu-
lating collagen I production in HSCs via the Smad transcrip-
tional regulators and C/EBPβ‐dependent pathways [89]. TGFβ
also induces synthesis of TIMP1, which inhibits apoptosis in
HSCs. Kupffer cells release large amounts of TNFα, which acti-
vates HSCs to release other cytokines including NOX, PDGF,
MCP1, and PDGF. In mouse, monocytes can be differentiated
into at least two subgroups: M2A/Gr1hi likely promote profi-
brotic effects via TGFβ, PDGF, FGF2, Galectin 3, CCL2,
CCL18, and IGFBP5, whereas Ly6‐Clo macrophages (CD11bhi/

Figure 36.3  Extracellular stimuli of hepatic stellate cell (HSC) activation. HSCs may be inhibited or activated by hepatic cells including hepato-
cytes, liver sinusoidal endothelial cells, macrophages, and natural killer (NK) and natural killer T cells (NKT) cells through the release of cytokines,
growth factors, or signaling molecules. Green lines indicate promotion and red lines inhibition of HSC activation. From [106]. Reproduced with
permission of BMJ Publishing Group Ltd.
 36:  Stellate Cells and Fibrosis 451

F4/80int/Ly6Clo) promote fibrolytic effects via release of MMP9 Platelets

and MMP12; the human counterparts to these murine cell sub-
sets have not been definitively characterized.
Liver sinusoidal endothelial cells (LSEC)
LSECs are a key source of intrahepatic fibronectin A isoform,
which is crucial for initiation of HSCs and mitogenic and fibro-
genic stimulation by TGFβ [90, 91] enhancing HSC survival
[92]. A splice variant of fibronectin, fibronectin EDA produced
by LSECs, is primarily expressed during development and in
response to injury. While hepatocytes and HSCs may also
express fibronectin, LSECs are early responders with fibronectin
release [91, 93]. Depending on the context, LSECs have diver-
gent functions and may promote fibrosis by fibroblast growth
factor receptor 1 (FGFR1) and CXCR4 or proregenerative path-
ways which are regulated by CXCR7. These observations have
been tested in mouse models: deletion of CXCR7 in LSECS
impairs liver regeneration while deletion of either FGFR1 in
LSECS or CXCR4 promotes proregenerative pathways [94].
Platelets release growth factors such as PDGF and TGFβ, which
have potent mitogenic and profibrogenic properties, respec-
tively, toward HSCs. Platelet‐derived chemokine ligand 4
(CXCL4) stimulates proliferation, chemotaxis, and chemokine
expression of HSCs in vitro. CXCL4 knockout mice have sig-
nificantly reduced liver damage upon injury. Procoagulant fac-
tors (e.g. thrombin, factor V) are strongly positively associated
with fibrosis progression [95, 96].
Natural killer cells (NK)
NK cells are an important part of the liver’s innate immune sys-
tem. NK cells have context‐specific, divergent profibrotic and
antifibrogenic properties. NK cells may have inhibitory effects
on liver fibrosis by killing activated HSCs in a TRAIL−,
NKG2D− and granzyme‐dependent manner. Following release
of IFNγ, NK cells can induce cell arrest and apoptosis in HSC
via STAT1‐signaling and help clear senescent activated HSCs [97].

Figure 36.4  Strategies for antifibrotic approaches. (1) Control or cure of the underlying disease is still the most effective antifibrotic therapy. (2)
Inhibition of hepatic stellate cell (HSC) activation by modulation of receptor‐ligand interactions by either established or experimental drugs can
attenuate fibrosis development. (3) Inhibition of the most potent profibrogenic pathways in HSCs, for example, by blocking CTGF or inhibiting
activation of latent TGFβ. (4) Promoting resolution of fibrosis by enhancing apoptosis of HSCs or by enhancing matrix degradation by TIMP1
antagonists and preventing cross‐linking of collagen by LOX2 inhibition. FXR, farnesoid‐X‐receptor; PPAR, peroxisome proliferator‐activated
receptor; UDCA, ursodeoxycholic acid; SVR, sustained virological response; CB1, cannabinoid receptor type 1; ARB, angiotensin II receptor
blocker; ET‐1, endothelin 1; TGFβ, transforming growth factor β; CTGF, connective tissue growth factor; mAb, monoclonal antibody; NF‐κB,
nuclear factor kappa‐light‐chain‐enhancer of activated B cells; NK, natural killer; TIMP, tissue inhibitor of metalloproteinase; LOXL2, lysyl
oxidase 2. From [106]. Reproduced with permission of BMJ Publishing Group Ltd.
452 THE LIVER: REFERENCES
Natural killer T cells (NKT cells), a subpopulation of NK
cells also expressing the T‐cell receptor, likely have a dual role
in hepatic fibrogenesis by expressing both profibrogenic
cytokines (IL4, IL13) and the CXCL16 receptor CXCR6.
LSECs and macrophages, which express CXCL16, can influ-
ence NKT cell migration and promote NKT cell‐mediated
liver fibrogenesis [98, 99].
ANTIFIBROTIC STRATEGIES
The elucidation of fibrogenic pathways and determinants for
resolution has made immense strides within the past 15 years
changing our perception of cirrhosis and advanced liver disease.
Our understanding of underlying molecular mechanisms and key
cellular determinants for fibrosis regression have informed the
transition of many antifibrotic agents into clinical trials [100–
102] with the following strategies: (i) controlling and reducing
the underlying disease, (ii) inhibiting pathways and cytokines
critical for HSC activation, (iii) inhibiting most potent profibro-
genic pathways, and (iv) promoting resolution of fibrosis by
HSC clearance (apoptosis) or enhancing resolution of fibrosis
(see Figure 36.4) with a selection of drugs in clinical trials.
The identification of core fibrogenic pathways, which might be
shared across organ systems, for example, pulmonary fibrosis and
renal fibrosis has widened our arsenal of substances that might be
suitable for advanced liver diseases as well [103]. Big data analy-
ses with interrogation of large transcriptomic and genomic data-
sets also offer potential for discovering diagnostics and therapeutics
for liver disease [52, 104]. All these developments should pave the
way for a range of antifibrotic therapies in the coming years.
REFERENCES
1. Byass, P. The global burden of liver disease: a challenge for methods and for
public health. BMC Med, 2014;12:159.
2. Marcellin, P. and Kutala, B.K. Liver diseases: a major, neglected global pub-
lic health problem requiring urgent actions and large‐scale screening. Liver
Int, 2018;38(1):2–6.
3. Scaglione, S., Kliethermes, S., Cao, G. et al. The epidemiology of cirrhosis
in the United States: a population‐based study. J Clin Gastroenterol, 2015;
49:690–6.
4. Hope, V.D., Eramova, I., Capurro, D., and Donoghoe, M.C. Prevalence and
estimation of hepatitis B and C infections in the WHO European region: a
review of data focusing on the countries outside the European Union and the
European Free Trade Association. Epidemiol Infect, 2014;142:270–86.
5. European AFTSOL. EASL clinical practical guidelines: management of
alcoholic liver disease. J Hepatol, 2012;57:399–420.
6. Roguin, A. Rene Theophile Hyacinthe Laënnec (1781–1826): the man
behind the stethoscope. Clin Med Res, 2006;4:230–5.
7. Yuen, M.F. and Lai, C.L. Treatment of chronic hepatitis B: evolution over
two decades. J Gastroenterol Hepatol, 2011;26(1):138–43.
8. Ohkoshi, S., Hirono, H., Watanabe, K., Hasegawa, K., Kamimura, K., and
Yano, M. Natural regression of fibrosis in chronic hepatitis B. J Gastroenterol,
2016;22:5459–66.
9. Gonzalez, H.C. and Duarte‐Rojo, A. Virologic cure of hepatitis c: impact on
hepatic fibrosis and patient outcomes. Curr Gastroenterol Rep, 2016;18:32.
10. Powell, L.W. and Kerr, J.F. Reversal of “cirrhosis” in idiopathic haemochro-
matosis following long‐term intensive venesection therapy. Australas Ann
Med, 1970;19:54–7.
11. Blumberg, R.S., Chopra, S., Ibrahim, R. et al. Primary hepatocellular carcinoma
in idiopathic hemochromatosis after reversal of cirrhosis. Gastroenterology,
1988;95:1399–402.
12. Dufour, J.F., DeLellis, R., and Kaplan, M.M. Reversibility of hepatic fibrosis
in autoimmune hepatitis. Ann Intern Med, 1997;127:981–5.
13. Falkmer, S., Samuelson, G., and Sjölin, S. Penicillamine‐induced normaliza-
tion of clinical signs, and liver morphology and histochemistry in a case of
Wilson’s disease. Pediatrics, 1970;45:260–8.
14. Kaplan, M.M., DeLellis, R.A., and Wolfe, H.J. Sustained biochemical and
histologic remission of primary biliary cirrhosis in response to medical treat-
ment. Ann Intern Med, 1997;126:682–8.
15. Kisseleva, T., Cong, M., Paik, Y. et al. Myofibroblasts revert to an inactive
phenotype during regression of liver fibrosis. Proc Natl Acad Sci USA,
2012;109:9448–53.
16. Troeger, J.S., Mederacke, I., Gwak, G.Y. et al. Deactivation of hepatic stel-
late cells during liver fibrosis resolution in mice. Gastroenterology, 2012;
143:1073–83.e22.
17. Lee, Y.A. and Friedman, S.L. Reversal, maintenance or progression: What
happens to the liver after a virologic cure of hepatitis C? Antiviral Res, 2014;
107C:23–30.
18. D’Ambrosio, R., Aghemo, A., Rumi, M.G. et al. A morphometric and immu-
nohistochemical study to assess the benefit of a sustained virological response
in hepatitis C virus patients with cirrhosis. Hepatology, 2012;56:532–43.
19. Zeybel, M., Hardy, T., Wong, Y.K. et al. Multigenerational epigenetic adap-
tation of the hepatic wound‐healing response. Nat Med, 2012;18:1369–77.
20. Decaris, M.L., Emson, C.L., Li, K. et al. Turnover rates of hepatic collagen
and circulating collagen‐associated proteins in humans with chronic liver
disease. PLoS One, 2015;10:e0123311.
21. Hytiroglou, P., Snover, D.C., Alves, V. et al. Beyond “cirrhosis”: a proposal
from the International Liver Pathology Study Group. Am J Clin Pathol,
2012;137:5–9.
22. Browning, J.D. and Horton, J.D. Molecular mediators of hepatic steatosis
and liver injury. J Clin Invest, 2004;114:147–52.
23. Urtasun, R., Conde de la Rosa, L., and Nieto, N. Oxidative and nitrosative
stress and fibrogenic response. Clin Liver Dis, 2008;12:769–90.
24. Pessayre, D., Berson, A., Fromenty, B., and Mansouri, A. Mitochondria in
steatohepatitis. Semin Liver Dis, 2001;21:57–69.
25. Aoyama, T., Paik, Y.H., Watanabe, S. et al. Nicotinamide adenine dinucleo-
tide phosphate oxidase in experimental liver fibrosis: GKT137831 as a novel
potential therapeutic agent. Hepatology, 2012;56:2316–27.
26. Yoshiji, H., Kuriyama, S., Yoshii, J. et al. Vascular endothelial growth factor
and receptor interaction is a prerequisite for murine hepatic fibrogenesis.
Gut, 2003;52:1347–54.
27. Ankoma‐Sey, V., Matli, M., Chang, K.B. et  al. Coordinated induction of
VEGF receptors in mesenchymal cell types during rat hepatic wound heal-
ing. Oncogene, 1998;17:115–21.
28. Olaso, E., Salado, C., Egilegor, E. et al. Proangiogenic role of tumor‐acti-
vated hepatic stellate cells in experimental melanoma metastasis. Hepatology,
2003;37:674–85.
29. Semela, D., Das, A., Langer, D., Kang, N., Leof, E., and Shah, V. Platelet‐
derived growth factor signaling through ephrin‐b2 regulates hepatic vascular
structure and function. Gastroenterology, 2008;135:671–9.
30. Fernandez, M., Semela, D., Bruix, J., Colle, I., Pinzani, M., and Bosch, J.
Angiogenesis in liver disease. J Hepatol, 2009;50:604–20.
31. Canbay, A., Taimr, P., Torok, N., Higuchi, H., Friedman, S., and Gores, G.J.
Apoptotic body engulfment by a human stellate cell line is profibrogenic.
Lab Invest, 2003;83:655–63.
32. Canbay, A., Higuchi, H., Bronk, S.F., Taniai, M., Sebo, T.J., and Gores, G.J.
Fas enhances fibrogenesis in the bile duct ligated mouse: a link between
apoptosis and fibrosis. Gastroenterology, 2002;123:1323–30.
33. Canbay, A., Friedman, S., and Gores, G.J. Apoptosis: the nexus of liver
injury and fibrosis. Hepatology, 2004;39:273–8.
34. Schwabe, R.F. and Luedde, T. Apoptosis and necroptosis in the liver: a mat-
ter of life and death. Nat Rev Gastroenterol Hepatol, 2018;15:738–52.
35. Hendriks, H.F., Verhoofstad, W.A., Brouwer, A., de Leeuw, A.M., and
Knook, D.L. Perisinusoidal fat‐storing cells are the main vitamin A storage
sites in rat liver. Exp Cell Res, 1985;160:138–49.
36. Friedman, S.L. Molecular regulation of hepatic fibrosis, an integrated cellu-
lar response to tissue injury. J Biol Chem, 2000;275:2247–50.
37. Bataller, R. and Brenner, D.A. Liver fibrosis. J Clin Invest, 2005;115:209–18.
38. Schulze‐Krebs, A., Preimel, D., Popov, Y. et al. Hepatitis C virus‐replicating
hepatocytes induce fibrogenic activation of hepatic stellate cells. Gastroen­
terology, 2005;129:246–58.
39. Lua, I., Li, Y., Zagory, J.A. et al. Characterization of hepatic stellate cells,
portal fibroblasts, and mesothelial cells in normal and fibrotic livers.
J Hepatol, 2016;64:1137–46.
 36:  Stellate Cells and Fibrosis 453
40. Mederacke, I., Hsu, C.C., Troeger, J.S. et al. Fate tracing reveals hepatic stel-
late cells as dominant contributors to liver fibrosis independent of its aetiol-
ogy. Nat Commun, 2013;4:2823.
41. Iwaisako, K., Jiang, C., Zhang, M. et  al. Origin of myofibroblasts in the
fibrotic liver in mice. Proc Natl Acad Sci USA, 2014;111:E3297–305.
42. Lemoinne, S., Cadoret, A., El Mourabit, H., Thabut, D., and Housset, C.
Origins and functions of liver myofibroblasts. Biochim Biophys Acta, 2013;
1832:948–54.
43. Uchio, K., Tuchweber, B., Manabe, N., Gabbiani, G., Rosenbaum, J., and
Desmoulière, A. Cellular retinol‐binding protein‐1 expression and modula-
tion during in vivo and in vitro myofibroblastic differentiation of rat hepatic
stellate cells and portal fibroblasts. Lab Invest, 2002;82:619–28.
44. Dranoff, J.A. and Wells, R.G. Portal fibroblasts: underappreciated mediators
of biliary fibrosis. Hepatology, 2010;51:1438–44.
45. Karin, D., Koyama, Y., Brenner, D., and Kisseleva, T. The characteristics of
activated portal fibroblasts/myofibroblasts in liver fibrosis. Differentiation,
2016;92:84–92.
46. Lamireau, T., Dubuisson, L., Lepreux, S. et al. Abnormal hepatic expression of
fibrillin‐1 in children with cholestasis. Am J Surg Pathol, 2002;26:637–46.
47. Fausther, M., Goree, J.R., Lavoie, É.G., Graham, A.L., Sévigny, J., and
Dranoff, J.A. Establishment and characterization of rat portal myofibroblast
cell lines. PLoS One, 2015;10:e0121161.
48. Xu, L., Hui, A.Y., Albanis, E. et al. Human hepatic stellate cell lines, LX‐1
and LX‐2: new tools for analysis of hepatic fibrosis. Gut, 2005;54:142–51.
49. Shibata, N., Watanabe, T., Okitsu, T. et al. Establishment of an immortalized
human hepatic stellate cell line to develop antifibrotic therapies. Cell
Transplant, 2003;12:499–507.
50. Shen, K., Chang, W., Gao, X. et al. Depletion of activated hepatic stellate
cell correlates with severe liver damage and abnormal liver regeneration in
acetaminophen‐induced liver injury. Acta Biochim Biophys Sin (Shanghai),
2011;43:307–15.
51. Puche, J.E., Lee, Y.A., Jiao, J. et al. A novel murine model to deplete hepatic
stellate cells uncovers their role in amplifying liver damage in mice.
Hepatology, 2013;57:339–50.
52. Zhang, D.Y., Goossens, N., Guo, J. et al. A hepatic stellate cell gene expres-
sion signature associated with outcomes in hepatitis C cirrhosis and hepato-
cellular carcinoma after curative resection. Gut, 2016;65(10):1754–64.
53. Yin, C., Evason, K.J., Maher, J.J., and Stainier, D.Y. The bHLH transcription
factor Hand2 marks hepatic stellate cells in zebrafish: analysis of stellate cell
entry into the developing liver. Hepatology, 2012; 56(5):1958–70.
54. Enzan, H., Himeno, H., Iwamura, S. et al. Sequential changes in human Ito
cells and their relation to postnecrotic liver fibrosis in massive and submas-
sive hepatic necrosis. Virchows Arch, 1995;426:95–101.
55. Inagaki, Y. and Okazaki, I. Emerging insights into transforming growth fac-
tor beta Smad signal in hepatic fibrogenesis. Gut, 2007;56:284–92.
56. Hellerbrand, C., Stefanovic, B., Giordano, F., Burchardt, E.R., and Brenner,
D.A. The role of TGFbeta1 in initiating hepatic stellate cell activation in
vivo. J Hepatol, 1999;30:77–87.
57. Hanafusa, H., Ninomiya‐Tsuji, J., Masuyama, N. et al. Involvement of the
p38 mitogen‐activated protein kinase pathway in transforming growth fac-
tor‐beta‐induced gene expression. J Biol Chem, 1999;274:27161–7.
58. Engel, M.E., McDonnell, M.A., Law, B.K., and Moses, H.L. Interdependent
SMAD and JNK signaling in transforming growth factor‐beta‐mediated tran-
scription. J Biol Chem, 1999;274:37413–20.
59. Fan, X., Zhang, Q., Li, S. et al. Attenuation of CCl4‐induced hepatic fibrosis
in mice by vaccinating against TGF‐β1. PLoS One, 2013;8:e82190.
60. Ling, H., Roux, E., Hempel, D. et al. Transforming growth factor β neutrali-
zation ameliorates pre‐existing hepatic fibrosis and reduces cholangiocarci-
noma in thioacetamide‐treated rats. PLoS One, 2013;8:e54499.
61. Wong, L., Yamasaki, G., Johnson, R.J., and Friedman, S.L. Induction of
beta‐platelet‐derived growth factor receptor in rat hepatic lipocytes during
cellular activation in vivo and in culture. J Clin Invest, 1994;94:1563–9.
62. Pinzani, M. PDGF and signal transduction in hepatic stellate cells. Front
Biosci, 2002;7:d1720–6.
63. Pinzani, M., Milani, S., Grappone, C., Weber, F.L.J., Gentilini, P., and
Abboud, H.E. Expression of platelet‐derived growth factor in a model of
acute liver injury. Hepatology, 1994;19:701–7.
64. Kocabayoglu, P., Lade, A., Lee, Y.A. et al. beta‐PDGF receptor expressed by
hepatic stellate cells regulates fibrosis in murine liver injury, but not carcino-
genesis. J Hepatol, 2015;63(1):141–7.
65. Deshmukh, M., Nakagawa, S., Higashi, T. et al. Cell type‐specific pharma-
cological kinase inhibition for cancer chemoprevention. Nanomedicine,
2018;14:317–25.
66. Friedman, S.L. Hepatic stellate cells: protean, multifunctional, and enig-
matic cells of the liver. Physiol Rev, 2008;88:125–72.
67. Huang, G. and Brigstock, D.R. Regulation of hepatic stellate cells by con-
nective tissue growth factor. Front Biosci, 2012;17:2495–507.
68. Reynaert, H., Thompson, M.G., Thomas, T., and Geerts, A. Hepatic stellate
cells: role in microcirculation and pathophysiology of portal hypertension.
Gut, 2002;50:571–81.
69. Yang, L., Kwon, J., Popov, Y. et al. Vascular endothelial growth factor pro-
motes fibrosis resolution and repair in mice. Gastroenterology, 2014;146:
1339–50.
70. Kantari‐Mimoun, C., Castells, M., Klose, R. et al. Resolution of liver fibrosis
requires myeloid cell‐driven sinusoidal angiogenesis. Hepatology, 2015;61:
2042–55.
71. Sprenger, H., Kaufmann, A., Garn, H., Lahme, B., Gemsa, D., and Gressner,
A.M. Induction of neutrophil‐attracting chemokines in transforming rat
hepatic stellate cells. Gastroenterology, 1997;113:277–85.
72. Marra, F., Valente, A.J., Pinzani, M., and Abboud, H.E. Cultured human liver
fat‐storing cells produce monocyte chemotactic protein‐1. Regulation by
proinflammatory cytokines. J Clin Invest, 1993;92:1674–80.
73. Lefebvre, E., Moyle, G., Reshef, R. et  al. Antifibrotic effects of the dual
CCR2/CCR5 antagonist cenicriviroc in animal models of liver and kidney
fibrosis. PLoS One, 2016;11:e0158156.
74. Friedman, S.L., Ratziu, V., Harrison, S.A. et al. A randomized, placebo‐con-
trolled trial of cenicriviroc for treatment of nonalcoholic steatohepatitis with
fibrosis. Hepatology, 2018;67:1754–67.
75. Winau, F., Hegasy, G., Weiskirchen, R. et al. Ito cells are liver‐resident anti-
gen‐presenting cells for activating T cell responses. Immunity, 2007;26:
117–29.
76. Yoshiji, H., Kuriyama, S., Miyamoto, Y. et al. Tissue inhibitor of metallopro-
teinases‐1 promotes liver fibrosis development in a transgenic mouse model.
Hepatology, 2000;32:1248–54.
77. Benyon, R.C. and Arthur, M.J. Extracellular matrix degradation and the role
of hepatic stellate cells. Semin Liver Dis, 2001;21:373–84.
78. Iredale, J.P. Hepatic stellate cell behavior during resolution of liver injury.
Semin Liver Dis, 2001;21:427–36.
79. Arthur, M.J. Collagenases and liver fibrosis. J Hepatol, 1995;22:43–8.
80. Ikejima, K., Takei, Y., Honda, H. et al. Leptin receptor‐mediated signaling
regulates hepatic fibrogenesis and remodeling of extracellular matrix in the
rat. Gastroenterology, 2002;122:1399–410.
81. Leclercq, I.A., Farrell, G.C., Schriemer, R., and Robertson, G.R. Leptin is
essential for the hepatic fibrogenic response to chronic liver injury. J
Hepatol, 2002;37:206–13.
82. Saxena, N.K., Saliba, G., Floyd, J.J., and Anania, F.A. Leptin induces
increased alpha2(I) collagen gene expression in cultured rat hepatic stellate
cells. J Cell Biochem, 2003;89:311–20.
83. Coombes, J.D., Choi, S.S., Swiderska‐Syn, M. et al. Osteopontin is a proxi-
mal effector of leptin‐mediated non‐alcoholic steatohepatitis (NASH) fibro-
sis. Biochim Biophys Acta, 2016;1862:135–44.
84. Sahai, A., Malladi, P., Melin‐Aldana, H., Green, R.M., and Whitington, P.F.
Upregulation of osteopontin expression is involved in the development of
nonalcoholic steatohepatitis in a dietary murine model. Am J Physiol
Gastrointest Liver Physiol, 2004;287:G264–73.
85. Aleffi, S., Petrai, I., Bertolani, C. et al. Upregulation of proinflammatory and
proangiogenic cytokines by leptin in human hepatic stellate cells. Hepatology,
2005;42:1339–48.
86. Hernandez‐Gea, V., Ghiassi‐Nejad, Z., Rozenfeld, R. et  al. Autophagy
releases lipid that promotes fibrogenesis by activated hepatic stellate cells in
mice and in human tissues. Gastroenterology, 2012;142:938–46.
87. Arshad, M.I., Piquet‐Pellorce, C., and Samson, M. IL‐33 and HMGB1
alarmins: sensors of cellular death and their involvement in liver pathology.
Liver Int, 2012;32:1200–10.
88. McHedlidze, T., Waldner, M., Zopf, S. et  al. Interleukin‐33‐dependent
innate lymphoid cells mediate hepatic fibrosis. Immunity,
2013;39:357–71.
89. Garcia‐Trevijano, E.R., Iraburu, M.J., Fontana, L. et al. Transforming growth
factor beta1 induces the expression of alpha1(I) procollagen mRNA by a
hydrogen peroxide‐C/EBPbeta‐dependent mechanism in rat hepatic stellate
cells. Hepatology, 1999;29:960–70.
90. Serini, G., Bochaton‐Piallat, M.L., Ropraz, P. et al. The fibronectin domain
ED‐A is crucial for myofibroblastic phenotype induction by transforming
growth factor‐beta1. J Cell Biol, 1998;142:873–81.
91. Jarnagin, W.R., Rockey, D.C., Koteliansky, V.E., Wang, S.S., and Bissell,
D.M. Expression of variant fibronectins in wound healing: cellular source
454 THE LIVER: REFERENCES
and biological activity of the EIIIA segment in rat hepatic fibrogenesis.
J Cell Biol, 1994;127:2037–48.
92. Thiele, G.M., Duryee, M.J., Freeman, T.L. et al. Rat sinusoidal liver endothe-
lial cells (SECs) produce pro‐fibrotic factors in response to adducts formed
from the metabolites of ethanol. Biochem Pharmacol, 2005;70:1593–600.
93. George, J., Wang, S.S., Sevcsik, A.M. et al. Transforming growth factor‐beta
initiates wound repair in rat liver through induction of the EIIIA‐fibronectin
splice isoform. Am J Pathol, 2000;156:115–24.
94. Ding, B.S., Cao, Z., Lis, R. et al. Divergent angiocrine signals from vascular
niche balance liver regeneration and fibrosis. Nature, 2014;505:97–102.
95. Wright, M., Goldin, R., Hellier, S. et al. Factor V Leiden polymorphism and
the rate of fibrosis development in chronic hepatitis C virus infection. Gut,
2003;52:1206–10.
96. Papatheodoridis, G.V., Papakonstantinou, E., Andrioti, E. et al. Thrombotic
risk factors and extent of liver fibrosis in chronic viral hepatitis. Gut,
2003;52:404–9.
97. Krizhanovsky, V., Yon, M., Dickins, R.A. et al. Senescence of activated stel-
late cells limits liver fibrosis. Cell, 2008;134:657–67.
98. Liepelt, A., Wehr, A., Kohlhepp, M. et al. CXCR6 protects from inflamma-
tion and fibrosis in NEMOLPC‐KO mice. Biochim Biophys Acta Mol Basis Dis,
2018;1865(2):391–402.
 99. Wehr, A., Baeck, C., Heymann, F. et  al. Chemokine receptor CXCR6‐
dependent hepatic NK T Cell accumulation promotes inflammation and
liver fibrosis. J Immunol, 2013;190:5226–36.
100. Friedman, S.L. Hepatic fibrosis: emerging therapies. Dig Dis, 2015;33:
504–7.
101. Fallowfield, J.A. Future mechanistic strategies for tackling fibrosis—an
unmet need in liver disease. Clin Med (Lond), 2015;15(6):s83–7.
102. Friedman, S.L., Neuschwander‐Tetri, B.A., Rinella, M., and Sanyal, A.J.
Mechanisms of NAFLD development and therapeutic strategies. Nat Med,
2018;24:908–22.
103. Friedman, S.L., Sheppard, D., Duffield, J.S., and Violette, S. Therapy for
fibrotic diseases: nearing the starting line. Sci Transl Med, 2013;5:167sr1.
104. Wooden, B., Goossens, N., Hoshida, Y., and Friedman, S.L. Using big data
to discover diagnostics and therapeutics for gastrointestinal and liver dis-
eases. Gastroenterology, 2017;152:53–67.
105. Martinez‐Hernandez, A. and Amenta, P.S. The hepatic extracellular matrix.
I. Components and distribution in normal liver. (editorial). Virchows Arch A
Pathol Anat Histopathol, 1993;423(1):1.
106. Lee, Y.A., Wallace, M.C., and Friedman, S.L. Pathobiology of liver fibro-
sis: a translational success story. Gut, 2015;64:830–41.
PART THREE:
FUNCTIONS OF THE LIVER
SECTION A:
METABOLIC FUNCTIONS
 Non‐alcoholic Fatty Liver
37 Disease and Insulin
Resistance
Max C. Petersen1,3, Varman T. Samuel1,2, Kitt Falk Petersen1, and
Gerald I. Shulman1,3
1
Department of Internal Medicine, Section of Endocrinology, Yale University School of Medicine,
New Haven, CT, USA
2
Veterans Affairs Medical Center, West Haven, CT, USA
3
Department of Molecular and Cellular Physiology, Yale University School of Medicine, New Haven, CT, USA

INTRODUCTION myocyte and hepatocyte causes insulin resistance have been

In the wake of the obesity epidemic, the world has witnessed a
rising prevalence of associated conditions such as the metabolic
syndrome, type 2 diabetes mellitus (T2D), and non‐alcoholic
fatty liver disease (NAFLD). The International Diabetes
Federation estimated the 2017 prevalence of T2D at over 400
million adults worldwide [1]. Diabetes is a leading cause of
blindness in working adults, end‐stage renal failure, non‐traumatic
limb loss, and is a major risk factor for ischemic heart disease
[2]. Similarly, NAFLD is increasingly recognized as a major
threat to public health. NAFLD is now considered the most
common chronic liver disease, with estimated prevalence of
around 30% in the United States [3], and is a major risk factor
for the development of non‐alcoholic steatohepatitis (NASH),
cirrhosis, and liver‐related death [4]. As cirrhosis secondary to
viral hepatitis becomes rarer, NASH is poised to become the lead-
ing indication for liver transplantation in the United States [5].
Hepatic insulin resistance frequently accompanies NAFLD
[4]. Simply considered, resistance to insulin develops in the
peripheral tissues (muscle and fat) and in the liver. Insulin resist-
ance in the periphery impairs insulin‐stimulated glucose uptake
and muscle glycogen synthesis. In the liver, insulin resistance
diminishes the ability of insulin to suppress gluconeogenesis
and stimulate net hepatic glycogen synthesis. The distinct
mechanisms by which lipid accumulation within the skeletal
partially elucidated in recent years. This chapter will: (i) review
the current model for the pathogenesis of insulin resistance in
the skeletal muscle, (ii) discuss how insulin resistance in s­ keletal
muscle may promote the development of NAFLD, (iii) explore
the mechanisms whereby NAFLD may lead to hepatic insulin
resistance, and finally, (iv) review the role of insulin‐sensitizing
therapies in correcting hepatic insulin resistance.
MECHANISM OF LIPID‐INDUCED
SKELETAL MUSCLE INSULIN
RESISTANCE
A review of the mechanisms responsible for skeletal muscle
insulin resistance serves as an important reference point from
which to proceed to a discussion of hepatic insulin resistance. In
skeletal muscle, insulin binds to its receptor, activating insulin
receptor tyrosine kinase activity with subsequent phosphoryla-
tion and activation of insulin receptor substrates (e.g. IRS1).
When phosphorylated, IRS1 activates phosphatidylinositol‐3‐
kinase (PI3 kinase) and, through a cascade of downstream
­signaling effectors, ultimately increases the translocation of glu-
cose transporter 4 (GLUT4)‐containing vesicles to the plasma
membrane. A current concept of muscle insulin resistance links

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
460 THE LIVER:  MECHANISM OF LIPID‐INDUCED SKELETAL MUSCLE INSULIN RESISTANCE
intramyocellular lipid accumulation  –  specifically, diacylglyc-
erol (DAG) accumulation  –  with impaired insulin signaling
and  consequent impairments in insulin stimulation of glucose
transport [6].
This model has evolved over several decades. In the 1960s,
Randle et al. first formed a hypothesis linking increased lipids
to decreased glucose metabolism in muscle [7]. Randle ­reasoned
that increased delivery of fatty acids (FAs) should promote FA
oxidation and inhibit glucose oxidation. The mechanism was as
follows: lipid oxidation increases the intramitochondrial [acetyl‐
CoA]/[CoA] and [NADH]/[NAD+] ratios, which in turn leads to
the inactivation of pyruvate dehydrogenase (PDH), the enzyme
responsible for the conversion of pyruvate into acetyl‐CoA.
Additionally, accumulation of intracellular citrate inhibits phos-
phofructokinase (PFK), a key glycolytic enzyme. The block in
PFK would lead to accumulation of glucose 6‐phosphate (G6P),
which in turn inhibits hexokinase activity. This inhibition of
hexokinase activity would result in an increase in intracellular
glucose, decreasing the chemical driving force for glucose
uptake. In essence, Randle hypothesized that oxidation of fat
prevents oxidation of glucose by inhibiting key glycolytic
enzymes. This model, known as the glucose–fatty acid cycle,
provided a logical biochemical basis for oxidative substrate
switching based on local availability and was observed in
­multiple experimental settings during the 1980s and 1990s, par-
ticularly in humans and rats subjected to acute increases in
plasma FA concentrations [8, 9].
Intramyocellular lipid accumulation causes
muscle insulin resistance
The development of non‐invasive magnetic resonance spectros-
copy (MRS) methods to interrogate metabolism in vivo sparked
an interest in reassessing the link between intracellular lipids
and glucose metabolism. Using 1H MRS and the hyperinsuline-
mic–euglycemic clamp method in young, non‐diabetic, non‐
obese humans, intramyocellular lipid (IMCL) content was
found to be a strong predictor of insulin resistance [10, 11].
Rothman et al., using 13C and 31P MRS, studied the biochemical
mechanisms responsible by measuring insulin‐mediated
changes in muscle glycogen and G6P both in patients with T2D
[12] and in non‐diabetic, first‐degree relatives of patients with
T2D [13]. In contrast to Randle’s hypothesis, the accumulation
of G6P was diminished in both the diabetic subjects and the
first‐degree relatives of diabetic patients. Subsequent MRS
experiments in diabetic human subjects revealed that neither
intramyocellular G6P nor intramyocellular glucose rose to lev-
els remotely high enough for the glucose–fatty acid cycle to
account for the diabetic impairment in insulin‐stimulated
­muscle glycogen synthesis [14]. These data demonstrated that
muscle insulin resistance in human diabetes was due to impair-
ment of insulin stimulated glucose transport, rather than accu-
mulation of glycolytic intermediates. The data also indicated
that these changes were present in the insulin‐resistant muscle
prior to the development of overt T2D.
Although the Randle glucose–fatty acid cycle had been
observed in the setting of acute lipid administration, it could not
account for the impaired insulin‐mediated glucose metabolism
of insulin‐resistant human muscle. When lipid was infused for
more than three hours, a second mechanism for lipid inhibition
of glucose utilization in the skeletal muscle was apparent.
Roden et  al. infused normal insulin‐sensitive subjects with
either glycerol or Intralipid/heparin. As the triglyceride in the
Intralipid was hydrolyzed by lipoprotein lipase and released
from the endothelium by heparin, the plasma non‐esterified
fatty acid (NEFA) concentration increased. Using 13C MRS to
measure muscle glycogen content, insulin‐stimulated muscle
glycogen synthesis was then assessed during a euglycemic–
hyperinsulinemic clamp [15]. After 3 hours of Intralipid/heparin
infusion, the rate of insulin‐mediated glycogen synthesis began
to decrease. This decrease in glycogen synthesis was preceded
by a drop in the concentration of G6P, as assessed by 31P MRS.
This finding was inconsistent with the glucose–fatty acid cycle,
which would have predicted an accumulation of G6P. Instead,
these data suggested that the lipid infusion caused muscle insu-
lin resistance by inducing a defect in either glucose transport or
phosphorylation activity. To distinguish between these two pos-
sibilities, Dresner et al. studied normal subjects with a similar
glycerol vs. Intralipid/heparin infusion protocol and used a 13C
MRS technique to measure intramyocellular glucose concentra-
tions [16]. Compared to the glycerol infusion, the Intralipid/
heparin infusion was associated with lower concentrations of
intramyocellular glucose and a failure to increase G6P concen-
trations. These data suggested that the lipid infusion induced
muscle insulin resistance by impairing insulin‐stimulated glu-
cose transport activity, consistent with the defect observed in
diabetic humans [14]. Furthermore, in muscle biopsies obtained
from these subjects, the apparent defect in insulin‐stimulated
glucose uptake was associated with impaired activation of
IRS1‐associated PI3 kinase activity [16].
The generation of mice with constitutively active pyruvate
dehydrogenase (PDH), produced by deletion of pyruvate
­dehydrogenase kinases 2 and 4 (Pdk2−/−/Pdk4−/− mice), enabled
a particularly dramatic illustration of the unnecessity of the
­glucose–fatty acid cycle for lipid‐induced muscle insulin resist-
ance. With this genetic activation of PDH, the Pdk2−/−/Pdk4−/−
mice preferentially oxidized glucose in both fasting and fed
states. The Pdk2−/−/Pdk4−/− mice, unable to switch to fatty acid
oxidation, accumulated IMCL and displayed profound muscle
insulin resistance during hyperinsulinemic–euglycemic clamp
studies [17].
Taken together, these studies indicated that lipid‐induced
muscle insulin resistance occurs by a mechanism independent
of the Randle glucose–fatty acid cycle. In this mechanism,
intramyocellular lipids induce a defect in the insulin signaling
pathway that blunts insulin stimulation of myocellular glucose
transport.
How does lipid accumulate in muscle?
Simply stated, IMCL accumulates when myocellular lipid
uptake exceeds its utilization. On the supply side of the equation
are the increased concentration of plasma lipids often seen in
obesity and the metabolic syndrome. As described above,
increasing delivery of fat to the muscle via lipid infusion induces
insulin resistance in normal, insulin‐sensitive subjects. At the
level of the muscle, flux into the myocyte may be increased by
the action of lipoprotein lipase, which acts at the endothelium to
 37:  Non‐alcoholic Fatty Liver Disease and Insulin Resistance 461
release fatty acids from circulating triglycerides. Pollare et al.
measured lipoprotein lipase (LPL) activity and insulin sensitiv-
ity, as assessed by the hyperinsulinemic–euglycemic clamp in
insulin‐sensitive controls and various insulin‐resistant popula-
tions (obese/non‐diabetic, obese/diabetic, etc.) [18]. They found
that muscle LPL activity was inversely related to the glucose
infusion rate during the clamp and positively correlated to
­fasting plasma insulin. That is, the higher the LPL activity, the
more insulin‐resistant the subject. The valine breakdown prod-
uct 3‐hydroxyisobutyrate (3‐HIB) was recently shown to be a
paracrine positive regulator of myocellular FA uptake; produced
by muscle cells, it acts at the endothelium [19]. 3‐HIB supple-
mentation in mice increased IMCL and decreased muscle insu-
lin sensitivity, and muscle 3‐HIB levels were increased in
diabetic db/db mice and diabetic human subjects. The physio-
logic significance of 3‐HIB regulation of myocellular FA deliv-
ery is unclear, but it is worth noting that increased plasma levels
of branched‐chain amino acids, including valine, are associated
with insulin resistance [20]. These studies support the hypothe-
sis that increased delivery and uptake of FA promotes accumu-
lation of IMCL and muscle insulin resistance.
Balancing muscle lipid delivery is muscle lipid utilization,
specifically mitochondrial fat oxidation. Muscle mitochondrial
function has been shown to be impaired in two groups of patients
who are at risk for developing T2D: the offspring of patients
with T2D and the elderly [21, 22]. Mitochondrial function can
be assessed in vivo using a combination of 13C MRS with 13C‐
acetate infusion to quantify tricarboxylic acid (TCA) cycle flux
and 31P MRS to quantify ATP synthesis. These techniques were
used to study lean offspring of T2D patients, who already had
skeletal muscle lipid accumulation and insulin resistance, and
were at high risk for developing T2D later in life [22]. These
patients were found to have around 40% lower rates of basal
mitochondrial ATP synthesis [22]. The mechanistic basis for
decreased mitochondrial activity in aging is likely multifacto-
rial, but oxidative free radical damage to mitochondrial DNA
(mtDNA) is one potential contributor. In mice, targeted overex-
pression of the free radical scavenging enzyme catalase to the
mitochondrion prevented age‐associated mtDNA damage, asso-
ciated with preserved skeletal muscle mitochondrial oxidative
capacity and insulin sensitivity [23]. Interestingly, recent stud-
ies of the human genetics of T2D are consistent with a role for
mitochondrial function in T2D pathogenesis, although the tis-
sues implicated have been liver and adipose rather than skeletal
muscle. Loss of the murine homolog of N‐acetyltransferase 2
(NAT2), a polymorphism of which has been associated with
human insulin resistance independent of BMI, results in reduced
mitochondrial function and whole‐body energy expenditure,
predisposing to lipid‐induced insulin resistance in mice [24,
25]. Variants in SLC16A11, which account for around 20% of
the increase in T2D risk among Mexicans, alter mitochondrial
fatty acid oxidation to favor lipid storage in hepatocytes [26].
The development of insulin resistance is not limited to those
with an inherent susceptibility toward insulin resistance. Indeed,
as we age, most humans will manifest similar changes; T2D
prevalence increases steadily with age [1]. In hyperinsulinemic–
euglycemic clamp studies, elderly subjects have been found to
manifest both peripheral and hepatic insulin resistance [27, 28].
Using 13C and 31P MRS techniques, Petersen et al. showed that
age‐related decreases in mitochondrial oxidation and phospho-
rylation were associated with increases in intramyocellular tri-
glyceride content and decreases in insulin sensitivity [21]. These
observations, in both aging subjects and in the insulin‐resistant
offspring of T2D patients, suggest that impaired mitochondrial
function is a common mechanism which predisposes to myocel-
lular lipid accumulation and insulin resistance.
How does muscle fat accumulation cause
muscle insulin resistance?
The above discussion has presented the evidence associating
IMCL and muscle insulin resistance in humans. Mechanistic
insights linking accumulation of lipid metabolites to impair-
ments in insulin action have been gleaned from rodent studies,
and more recently examined in humans. DAG activates protein
kinase C (PKC) isoforms, and intramyocellular DAG accumula-
tion has been repeatedly associated with activation of protein
kinase C‐θ (PKCθ) (Figure 37.1a, b) [6]. This isoform belongs
to the “novel” class of PKCs that exhibit sn 1,2 DAG‐dependent,
Ca2+‐independent activation. Skeletal muscle DAG accumula-
tion has been shown to temporally precede the activation of
PKCθ and the development of muscle insulin resistance during
an acute lipid infusion [29]. Importantly, PKCθ knockout mice
were protected from lipid‐induced insulin resistance during
lipid infusion–glucose clamp studies [30], but the PKCθ targets
responsible for this phenotype remain incompletely defined.
IRS1 Ser1101 was identified as a PKCθ substrate [31], but mice
carrying a serine‐to‐alanine mutation at this residue are not pro-
tected from lipid‐induced insulin resistance [32].
Current data support a role for PKCθ activation in the patho-
genesis of lipid‐induced muscle insulin resistance in humans as
well. Increased myocellular DAG and PKCθ activity has been
observed in the muscles of patients with T2D compared to non-
diabetic controls [33, 34]. Muscle insulin resistance following
an oral or intravenous lipid challenge has also been associated
with intramyocellular DAG accumulation and PKCθ activation
[34, 35]. Overall, a substantial body of correlative data links the
DAG–PKCθ axis to lipid‐induced skeletal muscle insulin resist-
ance, but the mechanisms by which activated PKCθ impairs
insulin signaling remain incompletely defined.
PERIPHERAL INSULIN RESISTANCE
CONTRIBUTES TO THE DEVELOPMENT
OF HEPATIC STEATOSIS
There is a close association between the development of NAFLD
and peripheral insulin resistance. An attractive model posits that
muscle insulin resistance, by impairing postprandial muscle gly-
cogen synthesis [36], facilitates hepatic de novo lipogenesis and
resultant intrahepatic triglyceride (IHTG) accumulation. This
hypothesis fits data collected by several investigators. IHTG can
arise from either de novo lipogenesis or re‐esterification of FAs
released from adipose or absorbed after a meal. Donnelly et al.
measured the relative contributions of these pathways in subjects
with NAFLD [37]. Although re‐esterification of circulating
462 THE LIVER:  PERIPHERAL INSULIN RESISTANCE CONTRIBUTES TO THE DEVELOPMENT OF HEPATIC STEATOSIS

(a) Muscle (c) Liver

IRS2
IRS1
Insulin IR P P PI3K
Insulin IR P P PI3K
P P
P P AKT
AKT
Insulin
Sensitive G6PC
PCK1
GCK
GLUT4
storage
vesicle

Glucose
Glucose

Glucose
Glycogen Glucose Glycogen

(b) (d)

sn-1,2 DAG sn-1,2 DAG

PKCθ PKCε

Insulin IR Insulin IR pT1160

Insulin
Resistant
GLUT4 Insulin
storage
vesicle
Glycogen synthesis

EGP suppression

Figure 37.1  Pathogenesis of lipid‐induced insulin resistance. (a) Normal insulin action on glucose metabolism in skeletal muscle. In the myocyte,
insulin acts to stimulate the translocation of GLUT4‐containing vesicles to the plasma membrane, allowing for glucose entry into the myocyte.
(b) Lipid‐induced insulin resistance in the skeletal muscle involves a blockade in proximal insulin signaling by PKCθ, which is activated by the
bioactive lipid DAG. This results in impaired insulin‐stimulated GLUT4 translocation. (c) Normal insulin action on glucose metabolism in liver. In
the hepatocyte, insulin inhibits gluconeogenesis and activates glycogen synthesis, thereby decreasing hepatic glucose production. (d) Lipid‐induced
insulin resistance in the hepatocyte. In the steatotic liver, DAG‐activated PKCε inhibits insulin receptor tyrosine kinase activity by phosphorylation
of Thr1160, impairing all downstream insulin action.

NEFAs was the dominant pathway (accounting for over half of
IHTG synthesis), rates of de novo lipogenesis were severalfold
higher than those previously observed in lean control subjects
and accounted for about one‐fourth of IHTG synthesis [37].
Moreover, in these NAFLD patients, de novo lipogenesis was
persistently elevated without the normal oscillations seen during
fed and fasted states in control subjects [37]. Diraison et al. also
used a stable isotope approach to measure sources of IHTG and
observed an approximately threefold increase in rates of de novo
lipogenesis in subjects with NAFLD from normal controls [38].
Finally, in a study of adult subjects with the metabolic syndrome,
Lambert et  al. also observed a threefold increase in absolute
 37:  Non‐alcoholic Fatty Liver Disease and Insulin Resistance 463

de  novo lipogenic flux in subjects with NAFLD compared to
those without [39]. These reports strongly suggest that increased
de novo lipogenesis is a contributor to NAFLD.
However, is increased lipogenesis present even before
NAFLD has developed in subjects at risk, such as lean but insu-
lin‐resistant individuals? Petersen et al. measured lipogenesis in
lean, healthy, young individuals that were either insulin‐sensi-
tive or insulin‐resistant, as determined by oral glucose tolerance
test [40]. These subjects were matched for multiple factors,
including age, BMI, adiposity, blood pressure, and activity.
Fasting plasma glucose and insulin, while still within normal
limits, were significantly higher in the insulin‐resistant cohort.
In addition, plasma triglycerides were increased and plasma
HDL decreased in the insulin‐resistant subjects. At baseline,
hepatic triglyceride content, as assessed by 1H MRS, was nor-
mal (less than 1.0%) and similar between the two groups. After
a carbohydrate challenge, however, the insulin‐resistant group
had marked elevations in plasma insulin but decreased muscle
glycogen synthesis. Interestingly, the insulin‐resistant group
exhibited a significant increase in liver triglyceride content,
with a concordant twofold increase in de novo lipogenesis.
Thus, in these young insulin‐resistant individuals, the impaired
non‐oxidative disposal of glucose as muscle glycogen led to
increased de novo lipogenesis and greater hepatic triglyceride
content [40]. This suggests that a primary defect in muscle insu-
lin sensitivity enables increased hepatic de novo lipogenesis
from ingested carbohydrates, which over time could develop
into NAFLD (Figure 37.2).
De novo lipogenesis is an insulin‐regulated process; insulin
induces the SREBP‐1c transcriptional program that upregulates
multiple lipogenic genes. Yet, as described above, insulin
­resistant subjects commonly display increased rates of de novo
lipogenesis. For this reason, it has been proposed that the liver
exhibits “pathway‐selective hepatic insulin resistance”; resist-
ance to insulin’s effects on glucose metabolism but not to those
involving lipid metabolism [41]. However, upon closer exami-
nation, the supposed “paradox” of selective hepatic insulin
resistance reveals itself to be most likely the product of con-
founding. When decoupled from the carbohydrate excess that
often accompanies insulin resistance, hepatocyte insulin resist-
ance  –  produced by either brief high‐fat feeding or insulin
receptor knockdown  –  results in decreased rates of de novo
­lipogenesis as measured by stable isotope methods in rats [42].
This finding reveals that de novo lipogenesis can become resist-
ant to insulin, so why is de novo lipogenesis increased in insulin
resistant humans with NAFLD? Interestingly, insulin is not nec-
essary for induction of the de novo lipogenic program. Indeed,
de novo lipogenesis is regulated by multiple nutritional inputs,
including glucose and fructose (through the carbohydrate
response element binding protein [ChREBP]) and amino acids
(through mTORC1). In the setting of chronic nutritional, and
particularly glucose, excess that accompanies typical human
insulin resistance, it is quite conceivable that nutritional inputs
could bypass hepatic insulin resistance to induce the de novo
lipogenic program on their own.
Finally, it is worth emphasizing that the major pathway for
hepatic triglyceride synthesis is re‐esterification of circulating
FAs, as described above. Re‐esterification is a substrate‐­
dependent but insulin‐independent process [42]. So while
de novo lipogenesis may be a particularly important pathway for
the development of NAFLD, the quantitatively dominant source
of IHTG in NAFLD  –  re‐esterification  –  is not affected by

Insulin sensitive Insulin resistant

Hypertriglyceridemia

Glycogen De novo lipogenesis

Glucose

Glycogen Glycogen

Figure 37.2  Schematic of whole‐body energy distribution after high‐carbohydrate mixed meals in insulin‐sensitive and insulin‐resistant individu-
als. In insulin‐sensitive states, glucose is non‐oxidatively disposed as glycogen in liver and muscle. In insulin‐resistant states, entry of glucose into
the muscle is impaired, leading to increased delivery to the liver, where de novo lipogenesis ensues, leading to increased hepatic triglyceride accu-
mulation and secretion. Adapted from [40]. Reproduced with permission of the National Academy of Sciences.
464 THE LIVER:  MECHANISMS FOR HEPATIC INSULIN RESISTANCE IN NAFLD
hepatic insulin sensitivity or resistance. In light of current
knowledge, therefore, it does not seem necessary to invoke par-
adoxical pathway‐selective hepatic insulin resistance to explain
the development of NAFLD or increased rates of de novo lipo-
genesis in insulin‐resistant subjects.
MECHANISMS FOR HEPATIC INSULIN
RESISTANCE IN NAFLD
The close association between peripheral insulin resistance and
NAFLD can confound studies of NAFLD and hepatic insulin
resistance. Specifically, in studying the action of insulin in the
liver, it may become difficult to determine which changes are
due to peripheral insulin resistance and which are due solely to
hepatic steatosis. Thus, to understand how intrahepatic lipid
renders the liver resistant to insulin, it is instructive to consider
models in which hepatic steatosis is present without pronounced
adiposity and peripheral insulin resistance.
Liver‐specific LPL overexpressing mice
One model in which to study the mechanism for lipid‐induced
hepatic insulin resistance is the liver‐specific LPL overexpress-
ing mouse. LPL is the rate‐limiting enzyme for the hydrolysis
of triglyceride from triglyceride‐rich lipoproteins, such as chy-
lomicrons and very low‐density lipoproteins (VLDLs). By
crossing a transgenic mouse expressing human LPL under the
control of the apoA‐I promoter (A‐I LPL) with a heterozygote
LPL knockout (het KO) mouse, it was possible to generate a
mouse in which LPL was expressed primarily in the liver [43].
The resulting liver LPL overexpressing mouse displays IHTG
accumulation. Using the hyperinsulinemic–euglycemic clamp
technique, insulin‐sensitivity of this mouse was assessed [44].
While whole‐body glucose metabolism and insulin‐stimulated
muscle glucose uptake were normal, the ability of insulin to
suppress hepatic glucose production was markedly diminished,
with impaired hepatic insulin signaling to IRS2 and PI3 kinase.
This model illustrates that hepatic lipid accumulation can lead
specifically to hepatic insulin resistance through a block in the
insulin signaling cascade.
Lipoatrophic mice
Another mouse model of hepatic fat accumulation and hepatic
insulin resistance is the “fatless” lipoatrophic mouse. By
expressing a dominant‐negative protein, A‐ZIP/F, under the
control of the adipocyte‐specific AP‐1 promoter, Moitra et al.
generated mice that are devoid of white adipose tissue [45].
Without adipocytes, these mice store lipids in the muscle and
liver. When studied using the hyperinsulinemic–euglycemic
clamp, these fatless mice displayed profound peripheral and
hepatic insulin resistance [46]. Analysis of insulin signaling in
the liver revealed that insulin failed to increase IRS2‐associated
PI3 kinase activity in the fatless mice. Remarkably, this pheno-
type was rescued by transplantation of fat pads from wild‐type
littermates [46]. With restoration of adipocytes, tissue lipid
concentrations normalize, as do muscle and hepatic insulin
­ correlated with HOMA [52]. Most recently, Ter Horst et  al.
signaling. Thus, these results in the fatless mouse model also
support the hypothesis that hepatic fat accumulation leads to
insulin resistance through a proximal block in the insulin signal-
ing cascade.
Acute high‐fat feeding in normal rats
In normal male Sprague–Dawley (SD) rats, high‐fat feeding for
three days (3d HFF) results in the accumulation of IHTG, with-
out any significant change in body weight or muscle lipid con-
tent [47]. In 3d HFF rats, although plasma glucose was
unchanged, plasma insulin concentrations nearly doubled, sug-
gesting insulin resistance. When tissue‐specific insulin action
was assessed in 3d HFF rats using the hyperinsulinemic–eugly-
cemic clamp, there were no differences in insulin‐stimulated,
whole‐body glucose disposal or muscle‐ and adipose‐specific
glucose uptake. In contrast, the liver was completely insulin‐
resistant in the 3d HFF rat, with insulin suppressing hepatic glu-
cose production by only around 10%, compared with nearly
80% suppression of hepatic glucose production (HGP) in con-
trol chow‐fed rats. Interestingly, if hepatic fat accumulation was
prevented by promoting mitochondrial lipid oxidation with the
mitochondrial uncoupler 2,4‐dinitrophenol, hepatic insulin sen-
sitivity was preserved. In the 3d HFF rat model, IHTG accumu-
lation was associated with proximal defects in insulin signaling,
with impaired tyrosine phosphorylation of IRS1 and IRS2 by
the insulin receptor ultimately impairing the ability of insulin to
activate hepatic glycogen synthesis. Importantly, similar defects
in insulin‐stimulated hepatic glycogen synthesis and insulin‐
mediated suppression of hepatic glucose production have been
in patients with T2D when studied under careful conditions with
matched insulin concentrations [48].
To identify the mechanism by which IHTG accumulation
impairs insulin signaling, possible involvement of the DAG–PKC
axis was examined. A screen for PKC activation in 3d HFF liver
was performed. This revealed that hepatic steatosis and hepatic
insulin resistance were most strongly associated with the activa-
tion of PKCε, which, like PKCθ, is a DAG‐dependent, Ca2+‐inde-
pendent novel PKC isoform (Figure 37.1) [47]. If hepatic steatosis
was prevented with the use of 2,4, dinitrophenol, PKCε activation
was also prevented. This association between DAG accumula-
tion, PKCε activation, and hepatic insulin resistance has since
been observed in dozens of rodent models [49].
Four studies to date have examined the potential involvement
of the DAG–PKCε axis in hepatic insulin resistance in humans.
Kumashiro et  al. measured multiple putative mediators of
hepatic insulin resistance in liver biopsy samples from a cohort
of obese, non‐diabetic subjects [50]. Hepatic DAG content was
found to be the best predictor of insulin sensitivity as measured
by the homeostatic model assessment (HOMA), and was
strongly correlated with PKCε activation. Magkos et al. studied
obese non‐diabetic subjects using suppression of hepatic glu-
cose production during hyperinsulinemic–euglycemic clamp
studies as the readout of hepatic insulin sensitivity, and observed
that hepatic DAG, of several lipid species measured, was the
most strongly correlated with hepatic insulin resistance [51].
Luukkonen et  al. studied adults undergoing bariatric surgery
and reported that four out of five DAG species measured were
 37:  Non‐alcoholic Fatty Liver Disease and Insulin Resistance 465
measured insulin suppression of hepatic glucose production in
obese non‐diabetic subjects and noted that subjects with hepatic
insulin resistance displayed increased DAG in the cytosolic/
lipid droplet fraction and increased PKCε activation compared
to subjects with preserved hepatic insulin sensitivity [53].
The total hepatic DAG content does not always correlate
with the degree of hepatic insulin resistance in rodent models
[49]. The reason for this dissociation is not always apparent,
but a failure to measure individual DAG stereoisomers is likely
to explain at least some discrepant models. The placement of
two fatty acyl chains along the three‐carbon glycerol backbone
of DAG yields three possible stereoisomers: sn‐1,2, sn‐2,3,
and sn‐1,3 DAG. PKC enzymes are only activated by sn‐1,2
DAG [54]. Interestingly, adipose triglyceride lipase (ATGL, or
PNPLA2)  –  the major triglyceride lipase in the hepato-
cyte  –  preferentially generates sn‐1,3 DAG when working
alone and preferentially generates sn‐1,3 and sn‐2,3 DAG
when working with its cofactor CGI‐58 [55]. The sn‐1,2 DAG
moieties capable of activating PKC enzymes, including PKCε,
are ­ therefore primarily generated through hydrolysis of
­phosphatidylinositol‐4,5‐bisphosphate (PIP2) in the plasma
­membrane and through de novo glycerolipid synthesis in the
endoplasmic reticulum, not through lipolysis. Many rodent
models in which total hepatic DAG content has been dissoci-
ated from hepatic insulin resistance have genetic disruptions in
lipid synthesis or lipolysis, and the stereoisomer composition
of the DAG pool within those hepatocytes thus cannot be
assumed to approximate the normal state. This is a subject ripe
for future investigation.
PKCε links hepatic steatosis to hepatic insulin
resistance through insulin receptor kinase
inhibition
The specific role of PKCε in the pathogenesis of hepatic insulin
resistance was assessed using antisense oligonucleotides
(ASOs) against PKCε [56]. ASOs are synthetic, modified
­oligonucleotides that are taken up preferentially in liver, adipose
tissue, and kidney, though not in other key tissues such as mus-
cle, brain, or β‐cells. They offer the opportunity to specifically
inhibit any gene product of interest, enabling both target valida-
tion and potential therapeutic utility. After treatment with either
saline, a control (random sequence) ASO, or PKCε ASO, nor-
mal SD rats were subjected to the 3d HFF protocol and insulin
action was assessed using the hyperinsulinemic–euglycemic
clamp. Though hepatic lipid accumulation, and specifically
DAG accumulation, was equal in all groups, PKCε ASO treated
rats displayed improved hepatic insulin sensitivity and insulin
signaling compared to both control groups. Mechanistically,
this was associated with improved activation of the insulin
receptor itself. Whereas high‐fat feeding impaired liver insulin
receptor (INSR) activation compared to control chow‐fed rats,
INSR tyrosine kinase activity was preserved in HFF rats treated
with PKCε ASO. These studies suggested that PKCε mediated
lipid‐induced hepatic insulin resistance through inhibition of the
insulin receptor.
To identify the molecular mechanism of this inhibition, phos-
phopeptide mass spectrometry was employed and revealed
insulin receptor Thr1160 as a substrate of PKCε [57]. Structural
studies revealed that owing to its position in the activation loop
of the insulin receptor kinase domain, Thr1160 phosphorylation
was predicted to destabilize the active configuration of the
kinase, impairing INSR activity. These predictions were borne
out in studies of insulin receptors carrying the phosphomimetic
Thr1160Glu mutation, which were severely kinase‐defective.
To evaluate the significance of insulin receptor Thr1160 phos-
phorylation for lipid‐induced hepatic insulin resistance, mice
carrying an alanine mutation at the homologous residue
(InsrT1150A mice) were studied. Mice were studied after an 8–10d
HFF protocol, which like the 3d HFF protocol in rats achieved
liver‐specific lipid accumulation and insulin resistance in wild‐
type mice. The InsrT1150A mice developed hepatic lipid accumu-
lation, including DAG, and exhibited PKCε translocation on
this diet. However, the InsrT1150A mice were protected from
hepatic insulin resistance, displaying improved suppression of
hepatic glucose production, stimulation of net hepatic glycogen
synthesis, and activation of INSR activity compared to wild‐
type mice [57]. These studies demonstrated the necessity of
insulin receptor Thr1150 phosphorylation for high‐fat, diet‐
induced, hepatic insulin resistance in mice, and revealed a
molecular mechanism by which hepatic lipid accumulation
leads to impaired hepatic insulin signaling (Figure 37.1c, d).
INSR Thr1160 is conserved from fly to man. The identifica-
tion of a conserved molecular mechanism for lipid‐induced cel-
lular insulin resistance raises the question of its evolutionary
utility. One attractive possibility is that lipid‐induced hepatic
insulin resistance is an adaptation to starvation. Prolonged fast-
ing in the rat (48 hours) induces hepatic DAG and TAG accumu-
lation and activates PKCε [58]. Induction of hepatic insulin
resistance would impair glycogen storage upon refeeding,
­conserving glucose for essential tissues such as brain.
FASTING HYPERGLYCEMIA AND
GLUCONEOGENESIS
Fasting hyperglycemia is a cardinal feature of T2D. The devel-
opment of fasting hyperglycemia requires increased endoge-
nous glucose production (EGP), which is the sum of both
gluconeogenesis and net glycogenolysis. Resolving the relative
contributions of these two pathways, while not trivial, is critical
to understanding the pathogenesis of T2D. This section will
review current knowledge regarding the measurement, regula-
tion, and significance of gluconeogenesis in the development
of T2D.
Hepatic gluconeogenesis in fed
and fasted states
Gluconeogenesis contributes substantially to EGP in both the
fed and fasted states in the normal human. This was demon-
strated in several studies by employing 13C MRS measurements
of hepatic glycogen [59–61]. After a meal, liver glycogen
­concentration increases in a near linear fashion, peaking at
around 5 hours [59]. During this period of hepatic glycogen syn-
thesis, when ingested glucose is absorbed from the gut, EGP is
466 THE LIVER:  FASTING HYPERGLYCEMIA AND GLUCONEOGENESIS
normally suppressed. The rate of glycogenolysis can be meas-
ured using repeated 13C MRS measures of hepatic glycogen dur-
ing this time. By subtracting the rate of glycogenolysis from the
rate of EGP, the rate of gluconeogenesis is calculated. About 5
hours after a meal, hepatic glycogen is utilized in a linear rate up
to about 22 hours [60, 61]. During this period, hepatic glycogen-
olysis accounts for only 40–50% of total EGP. Gluconeogenesis
accounts for the remainder, or 50–60% of EGP during the early
part of a fast. Between 22 and 46 hours, net hepatic glycogen-
olysis still accounts for around 20% and gluconeogenesis around
80% of EGP [60]. With continued fasting up to 46–64 hours, net
hepatic glycogenolysis diminishes to less than 5%, while the
remaining 95% of EGP results from gluconeogenesis [60].
Gluconeogenesis is increased in T2DM
Magnusson et al., using the 13C MRS technique, determined the
relative contributions of gluconeogenesis in patients with T2D
[62]. Rates of glycogenolysis and EGP were determined in con-
trol and diabetic subjects fasted over 23 hours. The rate of EGP
was increased by around 25% in the diabetic subjects, and this
increase was entirely accounted for by an approximate 60%
increase in the contribution of gluconeogenesis to EGP [62].
These findings have been confirmed by other techniques as
well. Wajngot et al. used the 2H2O method to measure rates of
gluconeogenesis in normal and diabetic subjects between 15
and 22 hours of fasting [63]. They found that the percentage
contribution of gluconeogenesis increased from 57.6 ± 2.5% to
70.6 ± 2.6% in normal subjects, and from 62.5 ± 2.8% to 75.6 ±
3.1% in subjects with T2DM [63].
From insulin resistance to fasting
hyperglycemia
Hepatic steatosis alone will not increase hepatic gluconeogene-
sis. In rodents, hepatic steatosis, while associated with hepatic
insulin resistance, is insufficient for increased fasting EGP or
hyperglycemia [47, 56, 57]. Similarly, in humans the presence
of hepatic steatosis is associated with hepatic insulin resistance
[50, 51, 53], but not with increased rates of fasting EGP, as
is  seen in T2D [64]. Additional changes beyond the develop-
ment of NAFLD are clearly required to permit fasting
hyperglycemia.
Several models have been proposed to describe the progres-
sion from hepatic insulin resistance to overt T2D. A popular,
straightforward model holds that hyperglycemia develops only
after pancreatic β‐cell dysfunction leads to waning plasma
­insulin concentrations [65]. An alternative model, the Unger
“bi‐hormonal hypothesis” for the pathogenesis of hyperglyce-
mia, holds glucagon equally culpable as insulin [66]. Glucagon,
even at basal levels, potently sustains hepatic glucose
­production. In canines, with glucose and insulin held constant,
suppression of basal glucagon reverses net hepatic glucose pro-
duction, instead causing net hepatic glucose uptake [67]. In
humans, basal glucagon activates gluconeogenesis and blunts
insulin‐stimulated glycogen synthesis [68]. Moreover, patients
with diabetes have an inappropriately high glucagon level rela-
tive to their degree of glycemia [69–71]. Thus, in the setting of
impaired hepatic insulin action, secondary to hepatic insulin
resistance and β‐cell dysfunction, glucagon likely promotes glu-
coneogenesis and fasting hyperglycemia.
These two models have in common a central role for direct
hepatic effects of insulin and glucagon. Insulin regulates gluco-
neogenesis through chronic transcriptional mechanisms, while
glucagon utilizes both chronic transcriptional and acute post‐
translational modes of regulation [72]. Insulin can suppress
transcription by phosphorylating and inactivating key transcrip-
tion factors of the FOXO family, especially FOXO1. Insulin‐
stimulated phosphorylation of FOXO transcription factors by
AKT2 excludes them from the nucleus and prevents them from
joining with the transcriptional coactivator PPARγ coactivator
1α (PGC‐1α) to activate transcription of the gluconeogenic
enzymes phosphoenolpyruvate carboxykinase (PCK1) and glu-
cose‐6‐phosphatase (G6PC) [73]. The FOXO module appears to
be particularly critical for upregulating G6PC transcription
­during long periods of fasting (around 18 hours in mice) [74]. A
second major mode of transcriptional regulation of gluconeo-
genesis by insulin involves the cyclic AMP‐responsive element
binding protein (CREB) and its regulators CREB‐binding pro-
tein (CBP) and CREB‐regulated transcriptional coactivator 2
(CRTC2) [75]. The CREB‐CBP‐CRTC2 module upregulates
PCK1, G6PC, and PGC‐1α, and appears to be operative in the
earlier stages of a fast (around 0–6 hours in mice) [74]. Insulin
phosphorylates CRTC2, leading to its nuclear exclusion [76].
Thus, the overall direct effect of insulin on hepatic gluconeo-
genesis centers on deactivating transcription of gluconeogenic
genes, a slow process.
Glucagon also acts through transcriptional regulation, but in
addition acts quickly through phosphorylation of several key
enzymes of glucose metabolism. Glucagon increases adenylate
cyclase activity and activates protein kinase A (PKA). PKA then
phosphorylates liver‐type pyruvate kinase (L‐PK) and the
bifunctional enzyme phosphofructokinase‐2/fructose‐2,6‐bis-
phosphatase (PFK‐2/FBPase‐2). The net effect is to favor gluco-
neogenesis over glycolysis [72]. PKA also phosphorylates
CREB and leads to the dephosphorylation of CRTC2, activating
the CREB‐CBP‐CRTC2 transcriptional module to promote
­gluconeogenic gene expression [75].
As described above, insulin and glucagon exert powerful
effects on the expression of the key gluconeogenic genes PCK1
and G6PC. But what is the pathophysiological significance of
these effects? The hypothesis that increased transcription of glu-
coneogenic enzymes accounts for the increased gluconeogene-
sis was tested in two rodent models of T2D and in human
subjects with T2D [77]. The first T2D rat model was induced by
3d HFF in combination with nicotinamide and streptozotocin to
induce β‐cell dysfunction (STZ/HFF). In this model of T2D, the
rats had modest fasting hyperglycemia and increased rates of
EGP. However, there was no increase in the expression of PCK1
or G6PC [77]. The second T2D rat model sought to decouple 3d
HFF‐induced hepatic insulin resistance from compensatory
hyperinsulinemia by employing somatostatin to pause endo-
crine pancreatic secretion. Under these conditions, infusion of
insulin and glucagon into the portal vein at basal replacement
rates resulted in hyperglycemia in the HFF rats compared to
chow‐fed controls. Though there were clear increases in EGP
 37:  Non‐alcoholic Fatty Liver Disease and Insulin Resistance 467
and gluconeogenesis in HFF rats, PCK1 and G6PC expression
were equivalent in HFF liver compared to control liver [77].
Finally, PCK1 and G6PC mRNAs were quantitated from liver
biopsy samples obtained from normoglycemic (control) and
hyperglycemic untreated T2D patients undergoing bariatric sur-
gery. The expression of PCK1 and G6PC was not increased in
the T2D subjects [77]. Other studies have also observed no
increase in PCK1 and G6PC transcription in insulin resistant rat
liver [78, 79]. Interestingly, in G6PC knockout mice, re‐expression
of G6PC to only around 20% of wild‐type levels is sufficient to
rescue hypoglycemia, suggesting that modest oscillations in
gluconeogenic gene transcription may exert limited effects on
gluconeogenic flux [80]. In summary, these studies provide
­evidence against the hypothesis that increased transcription of
gluconeogenic enzymes is responsible for the increased gluco-
neogenesis and EGP observed in patients with T2D. Instead,
these data suggest that other mechanisms are responsible for
these abnormalities.
If transcriptional effects are insufficient to account for the
transition from hepatic insulin resistance to fasting hyperglyce-
mia, what other mechanisms might be at work? In this context,
it is useful to consider other regulators of EGP, especially sub-
strate availability and allosteric modulators of gluconeogenesis
[72]. The potently insulin‐regulated process of adipose tissue
lipolysis yields metabolites that act through both mecha-
nisms  –  substrate push and allosteric activation  –  to promote
gluconeogenesis [78]. Lipolysis yields glycerol, which is con-
verted to glucose at rates proportional to its availability [81]. It
also yields FAs, which upon β‐oxidation in the hepatocellular
mitochondrion generate acetyl‐CoA, a potent allosteric activa-
tor of the key gluconeogenic enzyme pyruvate carboxylase [78].
In fasted rats, insulin infusion suppressed lipolysis and therefore
decreased turnover of glycerol and FAs; this was accompanied
by reductions in hepatic acetyl‐CoA and EGP [78]. Remarkably,
this insulin suppression of EGP could be totally prevented by
co‐infusion of glycerol and acetate at rates calibrated to prevent
insulin‐induced suppression of glycerol turnover and hepatic
acetyl‐CoA concentration, respectively [78]. These studies were
consistent with multiple dog and human studies carried out in
the 1990s that pointed to the insulin concentration in the periph-
ery, rather than in the portal vein, as the major controller of EGP
[72]. Indeed, they indicated that, at least in the fasted rodent
dependent on gluconeogenesis for nearly all of EGP, hepatocel-
lular insulin signaling was dispensable for insulin suppression
of EGP. Multiple rodent models of severely disrupted hepatocel-
lular insulin signaling that retain the ability to suppress EGP
upon insulin administration have lent dramatic credence to this
hypothesis [78, 82, 83].
The realization that control of EGP has a large indirect com-
ponent raises two interesting questions: one, whether increased
lipolysis mediates the transition to fasting hyperglycemia that
marks overt T2D; and two, whether hepatocellular insulin resist-
ance even contributes to the increased EGP of T2D. These ques-
tions remain incompletely answered but merit some comment.
With regard to the first question, some evidence supports a role
of lipolysis in T2D pathophysiology. Poorly controlled diabetic
subjects (increased EGP with fasting glucose >250 mg dl−1)
exhibit higher plasma FA concentrations throughout the day
than nondiabetic controls [84]. Similarly, subjects with T2D dis-
play higher rates of glycerol turnover and gluconeogenesis from
glycerol than nondiabetic controls [85]. Importantly, increased
plasma FA concentrations are a risk factor for incident T2D
[86]. It is tempting to speculate that a key mechanism whereby
obesity‐induced adipose tissue inflammation and insulin resist-
ance promote T2D is by driving EGP. But studies directly
addressing the role of adipose lipolysis in the transition to overt
T2D are currently lacking. With regard to the second question,
what role hepatocellular insulin resistance and the DAG–PKCε–
INSR axis might play in fasting hyperglycemia, it is prudent to
note the limitations of the rodent studies that have emphasized
the importance of indirect insulin action for EGP suppression.
Specifically, the use of overnight‐fasted rodents in these studies
minimized the contribution of glycogenolysis to EGP and maxi-
mized the contribution of gluconeogenesis. However, as dis-
cussed above, glycogenolysis is a physiologically significant
contributor to EGP in humans during the first 24 hours of a fast
[60]. Hepatocellular insulin resistance potently affects hepatic
glycogen metabolism [57], and insulin‐stimulated hepatic gly-
cogen deposition is impaired in T2D [48]. Thus, for the two
major components of EGP, a reasonable simplification may be
to state that direct hepatocellular effects dominate insulin regu-
lation of hepatic glycogen metabolism, while indirect effects
dominate insulin regulation of gluconeogenesis (Figure 37.3).
INSULIN SENSITIZING AGENTS
AND HEPATIC INSULIN RESISTANCE
Metformin
Metformin, a biguanide, has been used to treat patients with T2D
for over 30 years and improves liver function test abnormalities,
at least temporarily, in some patients with NAFLD. Since hepatic
glucose production is the sum of gluconeogenesis and glycogen-
olysis, it is important to know which pathway is affected by met-
formin. Using [6,6−2H] glucose measures of EGP combined with
13
C MRS measures of net hepatic glycogenolysis and 2H2O
measures of gluconeogenesis, Hundal et al. studied nine diabetic
subjects before and after three months of metformin therapy and
compared the results with seven age‐, sex‐, and weight‐matched
controls [87]. At baseline, the diabetic subjects had increased
rates of EGP compared to non‐diabetic control subjects. Using
13
C MRS measures of net hepatic glycogenolysis and 2H2O
measures of gluconeogenesis, it was clear that increased hepatic
gluconeogenesis accounted for the increase in EGP, consistent
with previous studies [62]. After three months of metformin
­therapy, EGP was lowered in the diabetic subjects, due to a 36%
reduction in hepatic gluconeogenesis [87].
Though metformin clearly reduces fasting glycemia and it is
widely accepted that this occurs through decreased gluconeo-
genesis, the cellular mechanisms for the effect of metformin
remain debated. A recent study by Madiraju et al. is unique in
identifying both a specific molecular interaction for metformin
and a physiological basis for the rare metformin adverse effect
of lactic acidosis [88]. Madiraju et al. observed non‐competitive
468 THE LIVER:  INSULIN SENSITIZING AGENTS AND HEPATIC INSULIN RESISTANCE

Adipose insulin resistance

inflammation Hepatic insulin resistance
Insulin secretion

Hepatocellular insulin signaling

Adipose lipolysis

Acetyl CoA
Glycerol turnover

Gluconeogenesis
Net glycogenolysis

Hepatic glucose production

Figure 37.3  Effect of adipose and hepatic insulin resistance on hepatic glucose production. Hepatic glucose production is controlled through both
direct and indirect mechanisms. Direct effects are mediated through the hepatocellular insulin receptor and, in the acute setting, are mediated largely
through promotion of glycogen storage. Indirect effects include, but are not limited to, inhibition of adipocyte lipolysis which decreases whole‐body
glycerol and fatty acid turnover. This in turn decreases allosteric activation of gluconeogenesis by hepatic mitochondrial acetyl‐CoA and substrate‐
driven gluconeogenesis from glycerol. A prediction that follows from this model is that hepatocellular insulin resistance mainly affects the glycog-
enolytic component of hepatic glucose production, while adipocyte insulin resistance (often associated with inflammation) mainly  affects the
gluconeogenic component of hepatic glucose production.

inhibition of the redox shuttle enzyme mitochondrial glycer- Thiazolidinediones

ophosphate dehydrogenase (mGPD, or GPD2) by metformin
at physiologically relevant concentrations, an effect that is pre-
dicted to cause an increase in the cytoplasmic redox state and
a  reciprocal decrease in the mitochondrial redox state of the
hepatocyte. These redox effects were indeed observed with
­metformin treatment in rats, and resulted in inhibition of gluco-
neogenesis from redox‐dependent substrates (i.e. lactate and
glycerol) but not from redox‐independent substrates (i.e. pyru-
vate and alanine) in cultured hepatocytes. Rats treated with
­antisense oligonucleotides targeting mGPD, as well as mGPD
knockout mice, phenocopied metformin‐treated animals and did
not further suppress EGP upon metformin treatment. The role of
mGPD inhibition in metformin‐treated humans remains to be
investigated.
Other hypotheses for metformin action continue to be inves-
tigated as well. Metformin has been found to inhibit mitochon-
drial complex I [89], though this may occur only at
supraphysiologic concentrations [88]. Inhibition of the respira-
tory chain would be predicted to increase the cellular [AMP] :
[ATP] ratio, and indeed, metformin stimulates the AMP‐acti-
vated kinase (AMPK) [88]. However, AMPK knockout mice
remain responsive to metformin treatment [90] and pharmaco-
logical activation of AMPK to the same extent as metformin
was insufficient to phenocopy metformin’s suppression of EGP
[88]. Metformin‐induced increases in the [AMP] : [ATP] ratio
have also been hypothesized to suppress EGP by antagonizing
the cAMP‐mediated effects of glucagon [91]; however, the
physiological relevance of this mechanism was challenged by a
human study in which metformin did not inhibit glucagon‐­
stimulated increases in EGP [92].
The thiazolidinediones (TZDs) are PPARγ agonists that improve
insulin sensitivity and have been used in the treatment of T2D.
Both TZDs currently approved for use in the United States,
rosiglitazone and pioglitazone, have been shown to improve
hepatic insulin sensitivity in human studies [93, 94].
The ability of the TZDs to improve hepatic insulin action is
primarily due to their extrahepatic effects. PPARγ is mainly
expressed in adipocytes with lower expression in the liver and
skeletal muscle, though there are species differences in expres-
sion. Given this discordance between the site of PPARγ expres-
sion and the site of drug effects, it was hypothesized that TZDs
redistribute fat from the liver and muscle into the adipocyte
[95]. Mayerson et al. tested this hypothesis, giving rosiglitazone
to nine patients with T2D in order to assess changes in insulin
sensitivity (as assessed by the hyperinsulinemic–euglycemic
clamp) and tissue fat content (as measured by 1H MRS) [96].
Insulin‐mediated suppression of lipolysis in subcutaneous fat
was also assessed using microdialysis to measure glycerol
release. After three months of therapy, rosiglitazone therapy was
associated with around 40% decrease in hepatic triglyceride
content, around 40% increase in extramyocellular triglyceride
concentration, and improved suppression of adipocyte lipolysis.
Though there were no detectable decreases in intramyocellular
triglyceride in this study, rosiglitazone did improve insulin‐
mediated whole‐body glucose disposal. This apparent discon-
nect between intramyocellular triglyceride and peripheral
insulin action underscores the fact that intramyocellular triglyc-
eride is only a crude marker for the active metabolite (putatively
DAG) responsible for lipid‐induced insulin resistance [97].
 37:  Non‐alcoholic Fatty Liver Disease and Insulin Resistance 469
In  summary, TZDs exert their beneficial effects by shifting
intracellular lipid from the liver and muscle into the adipose
­tissue, via PPARγ‐mediated improvements in adipose insulin
sensitivity.
Weight loss
Weight loss reduces intrahepatic fat content and improves
hepatic insulin sensitivity in humans. Petersen et  al. studied
eight obese subjects before and after weight loss induced by a
hypocaloric, low‐fat diet [98]. Patients were kept on a diet until
their fasting plasma glucose levels had stabilized. This required
3–12 weeks, with an average weight loss of around 8 kg. This
only led to a modest change in BMI, with subjects dropping
from 30 to 28 kg m−2. Thus, even after weight loss, subjects
remained overweight. A comparison was made to lean, seden-
tary control subjects without diabetes, who had an average BMI
of 24 kg m−2. Weight loss did not alter IMCL content, which was
still elevated compared to the lean controls. However, intrahe-
patic triglyceride content dropped substantially from around
12% to 2%, approaching the less than 1% values seen in the lean
controls. Weight loss and the reduction in intrahepatic triglycer-
ide decreased the rate of fasting EGP and improved hepatic
insulin sensitivity (as assessed by hyperinsulinemic–euglyce-
mic clamp) to near normal. Thus, in this and other studies [99–
101], modest weight loss can potently correct hepatic steatosis
and restore hepatic insulin sensitivity. Finally, the DiRECT trial,
which compared an intensive weight loss intervention using
hypocaloric meal replacement shakes to control usual care,
extended these findings in overweight adults with type 2 diabe-
tes not on insulin [102]. A subset of this trial cohort underwent
measurements of liver fat content, and a large decrease (from
around 16% to 3%) in liver fat content was observed in the
weight loss cohort [103]. These findings underscore the revers-
ibility of NAFLD with weight loss.
As of this writing, no reports have determined whether exer-
cise, independent of weight loss, can improve hepatic steatosis
and hepatic insulin sensitivity. Several trials have tested exercise
regimens of varying intensity and duration and demonstrated
beneficial effects on intrahepatic triglyceride and metabolic
parameters, but also on body weight [104–106]. However, in
one cross‐sectional study, physical activity was inversely corre-
lated with intrahepatic triglyceride content, even after control-
ling for other key variables such as age, sex, BMI, HOMA, and
adiponectin [107]. It is tantalizing to hypothesize that exercise,
by reversing muscle insulin resistance [108], would allow more
glucose to be taken up into the muscle, with resultant decreases
in de novo lipogenesis and hepatic fat content. However, more
needs to be done to determine whether this hypothesis is valid,
and more importantly, whether such lifestyle interventions can
be implemented and adopted by an ever‐increasing at‐risk
population.
REFERENCES
1. International Diabetes Federation. IDF Diabetes Atlas, 2017, International
Diabetes Federation.
2. Gregg, E.W., Li, Y., Wang, J. et al. Changes in diabetes‐related complications
in the United States, 1990–2010. N Engl J Med, 2014;370(16):1514–23.
3. Le, M.H., Devaki, P., Ha, N.B. et al. Prevalence of non‐alcoholic fatty liver
disease and risk factors for advanced fibrosis and mortality in the United
States. PloS One, 2017;12(3):e0173499.
4. Tilg, H., Moschen, A.R., and Roden, M. NAFLD and diabetes mellitus.
Nat Rev Gastroenterol Hepatol, 2016;14(1):147.
5. Pais, R., Barritt, A.S., Calmus, Y. et  al. NAFLD and liver transplantation:
Current burden and expected challenges. J Hepatol, 2016;65(6):1245–57.
6. Shulman, G.I. Ectopic fat in insulin resistance, dyslipidemia, and cardio-
metabolic disease. N Engl J Med, 2014;371(12):1131–41.
7. Randle, P.J., Garland, P.B., Hales, C.N. et al. The glucose fatty‐acid cycle. Its
role in insulin sensitivity and the metabolic disturbances of diabetes mellitus.
Lancet, 1963;1(7285); 785–9.
8. Jucker, B.M., Rennings, A.J., Cline, G.W. et al. 13C and 31P NMR studies
on the effects of increased plasma free fatty acids on intramuscular glucose
metabolism in the awake rat. J Biol Chem, 1997;272(16):10464–73.
9. Randle, P.J. Regulatory interactions between lipids and carbohydrates: the glu-
cose fatty acid cycle after 35 years. Diabetes Metab Rev, 1998;14(4):263–83.
10. Krssak, M., Falk Petersen, K., Dresner, A. et al. Intramyocellular lipid con-
centrations are correlated with insulin sensitivity in humans: a 1H NMR
spectroscopy study. Diabetologia, 1999;42(1):113–6.
11. Perseghin, G., Scifo, P., De Cobelli, F. et  al. Intramyocellular triglyceride
content is a determinant of in vivo insulin resistance in humans: a 1H‐13C
nuclear magnetic resonance spectroscopy assessment in offspring of type 2
diabetic parents. Diabetes, 1999;48(8):1600–6.
12. Rothman, D.L., Shulman, R.G., and Shulman, G.I. 31P nuclear magnetic
resonance measurements of muscle glucose‐6‐phosphate. Evidence for
reduced insulin‐dependent muscle glucose transport or phosphorylation
activity in non‐insulin‐dependent diabetes mellitus. J Clin Invest,
1992;89(4):1069–75.
13. Rothman, D.L., Magnusson, I., Cline, G. et  al. Decreased muscle glucose
transport/phosphorylation is an early defect in the pathogenesis of non‐­
insulin‐dependent diabetes mellitus. Proc Natl Acad Sci USA,
1995;92(4):983–7.
14. Cline, G.W., Petersen, K.F., Krssak, M. et al. Impaired glucose transport as a
cause of decreased insulin‐stimulated muscle glycogen synthesis in type 2
diabetes. N Engl J Med, 1999;341(4):240–6.
15. Roden, M., Price, T.B., Perseghin, G. et al. Mechanism of free fatty acid‐
induced insulin resistance in humans. J Clin Invest, 1996;97(12):2859–65.
16. Dresner, A., Laurent, D., Marcucci, M. et  al. Effects of free fatty acids
on  glucose transport and IRS‐1‐associated phosphatidylinositol 3‐kinase
­activity. J Clin Invest, 1999;103(2):253–9.
17. Rahimi, Y., Camporez, J.‐P.G., Petersen, M.C. et  al. Genetic activation of
pyruvate dehydrogenase alters oxidative substrate selection to induce skeletal
muscle insulin resistance. Proc Natl Acad Sci USA, 2014;111(46):16508–13.
18. Pollare, T., Vessby, B., and Lithell, H. Lipoprotein lipase activity in skeletal
muscle is related to insulin sensitivity. Arterioscler Thromb J Vasc Biol,
1991;11(5):1192–203.
19. Jang, C., Oh, S.F., Wada, S. et al. A branched‐chain amino acid metabolite
drives vascular fatty acid transport and causes insulin resistance. Nat Med,
2016;22(4):421–6.
20. Lynch, C.J., and Adams, S.H. Branched‐chain amino acids in metabolic
­signalling and insulin resistance. Nat Rev Endocrinol, 2014;10(12):723–36.
21. Petersen, K.F., Befroy, D., Dufour, S. et al. Mitochondrial dysfunction in the
elderly: possible role in insulin resistance. Science, 2003;300 5622;, 1140–2.
22. Petersen, K.F., Dufour, S., Befroy, D. et al. Impaired mitochondrial activity
in the insulin‐resistant offspring of patients with type 2 diabetes. N Engl J
Med, 2004;350(7):664–71.
23. Lee, H.‐Y., Choi, C.S., Birkenfeld, A.L. et al. Targeted expression of catalase
to mitochondria prevents age‐associated reductions in mitochondrial
­function and insulin resistance. Cell Metab, 2010;12(6):668–74.
24. Camporez, J.P., Wang, Y., Faarkrog, K. et al. Mechanism by which arylamine
N‐acetyltransferase 1 ablation causes insulin resistance in mice. Proc Natl
Acad Sci USA, 2017;114(52):E11285–92.
25. Knowles, J.W., Xie, W., Zhang, Z. et  al. Identification and validation of
N‐acetyltransferase 2 as an insulin sensitivity gene. J Clin Invest,
2015;125(4):1739–51.
26. Rusu, V., Hoch, E., Mercader, J.M. et al. Type 2 Diabetes Variants Disrupt
Function of SLC16A11 through Two Distinct Mechanisms. Cell,
2017;170(1):199–212.
470 THE LIVER:  REFERENCES
27. Fink, R.I., Kolterman, O.G., Griffin, J. et  al. Mechanisms of insulin
­resistance in aging. J Clin Invest, 1983;71(6):1523–35.
28. Rowe, J.W., Minaker, K.L., Pallotta, J.A. et  al. Characterization of the
­insulin resistance of aging. J Clin Invest, 1983;71(6):1581–7.
29. Yu, C., Chen, Y., Cline, G.W. et al. Mechanism by which fatty acids inhibit
insulin activation of insulin receptor substrate‐1(IRS‐1)‐associated
phosphatidylinositol 3‐kinase activity in muscle. J Biol Chem,
­ 54. Rando, R.R. and Young, N. The stereospecific activation of protein kinase C.
2002;277(52):50230–6.
30. Kim, J.K., Fillmore, J.J., Sunshine, M.J. et  al. PKC‐theta knockout mice
are  protected from fat‐induced insulin resistance. J Clin Invest,
2004;114(6):823–7.
31. Li, Y., Soos, T.J., Li, X. et al. Protein kinase C θ inhibits insulin signaling
by phosphorylating IRS1 at Ser1101. J Biol Chem, 2004;279(44):45304–7.
32. Petersen, M.C., Camporez, J.P.G., and Shulman, G.I. IRS1 Ser1101
phosphorylation impairs insulin‐stimulated muscle glucose metabolism.
­ Invest, 2007;117(3):739–45.
Diabetes, 2017;66(1):A37.
33. Itani, S.I., Pories, W.J., Macdonald, K.G. et  al. Increased protein kinase
C  theta in skeletal muscle of diabetic patients. Metabolism,
2001;50(5):553–7.
34. Szendroedi, J., Yoshimura, T., Phielix, E. et  al. Role of diacylglycerol
­activation of PKCθ in lipid‐induced muscle insulin resistance in humans.
Proc Natl Acad Sci USA, 2014;111(26):9597–602.
35. Nowotny, B., Zahiragic, L., Krog, D. et al. Mechanisms underlying the onset
of oral lipid‐induced skeletal muscle insulin resistance in humans. Diabetes,
2013;62(7):2240–8.
36. Shulman, G.I., Rothman, D.L., Jue, T. et al. Quantitation of muscle glycogen
synthesis in normal subjects and subjects with non‐insulin‐dependent
­diabetes by 13C nuclear magnetic resonance spectroscopy. N Engl J Med,
1990;322(4):223–8.
37. Donnelly, K.L., Smith, C.I., Schwarzenberg, S.J. et al. Sources of fatty acids
stored in liver and secreted via lipoproteins in patients with nonalcoholic
fatty liver disease. J Clin Invest, 2005;115(5):1343–51.
38. Diraison, F., Moulin, P., and Beylot, M. Contribution of hepatic de novo
lipogenesis and reesterification of plasma non esterified fatty acids to plasma
triglyceride synthesis during non‐alcoholic fatty liver disease. Diabetes
Metab, 2003;29(5):478–85.
39. Lambert, J.E., Ramos‐Roman, M.A., Browning, J.D. et al. Increased de novo
lipogenesis is a distinct characteristic of individuals with nonalcoholic fatty
liver disease. Gastroenterology, 2014;146(3):726–35.
40. Petersen, K.F., Dufour, S., Savage, D.B. et al. The role of skeletal muscle
insulin resistance in the pathogenesis of the metabolic syndrome. Proc Natl
Acad Sci USA, 2007;104(31):12587–94.
41. Brown, M.S. and Goldstein, J.L. Selective versus total insulin resistance: a
pathogenic paradox. Cell Metab, 2008;7(2):95–6.
42. Vatner, D.F., Majumdar, S.K., Kumashiro, N. et  al. Insulin‐independent
regulation of hepatic triglyceride synthesis by fatty acids. Proc Natl Acad Sci
USA, 2015;112(4):1143–8.
43. Merkel, M., Weinstock, P.H., Chajek‐Shaul, T. et  al. Lipoprotein lipase
expression exclusively in liver. A mouse model for metabolism in the neona-
tal period and during cachexia. J Clin Invest, 1998;102(5):893–901.
44. Kim, J.K., Fillmore, J.J., Chen, Y. et  al. Tissue‐specific overexpression of
lipoprotein lipase causes tissue‐specific insulin resistance. Proc Natl Acad
Sci USA, 2001;98(13):7522–7.
45. Moitra, J., Mason, M.M., Olive, M. et al. Life without white fat: a transgenic
mouse. Genes Dev, 1998;12(20):3168–81.
46. Kim, J.K., Gavrilova, O., Chen, Y. et al. Mechanism of insulin resistance in
A‐ZIP/F‐1 fatless mice. J Biol Chem, 2000;275(12):8456–60.
47. Samuel, V.T., Liu, Z.‐X., Qu, X. et  al. Mechanism of hepatic insulin
resistance in non‐alcoholic fatty liver disease. J Biol Chem, 2004;
­ diabetes. Am J Physiol Endocrinol Metab, 2004;287(1):E55–62.
279(31):32345–53.
48. Krssak, M., Brehm, A., Bernroider, E. et  al. Alterations in postprandial
hepatic glycogen metabolism in type 2 diabetes. Diabetes, 2004;
53(12):3048–56.
49. Petersen, M.C., and Shulman, G.I. Roles of diacylglycerols and ceramides in
hepatic insulin resistance. Trends Pharmacol Sci, 2017;38(7):649–65.
50. Kumashiro, N., Erion, D.M., Zhang, D. et al. Cellular mechanism of insulin
resistance in nonalcoholic fatty liver disease. Proc Natl Acad Sci USA,
2011;108(39):16381–5.
51. Magkos, F., Su, X., Bradley, D. et al. Intrahepatic diacylglycerol content is
associated with hepatic insulin resistance in obese subjects. Gastroenterology,
2012;142(7):1444–6.
52. Luukkonen, P.K., Zhou, Y., Sädevirta, S. et al. Hepatic ceramides dissociate
steatosis and insulin resistance in patients with non‐alcoholic fatty liver
­disease. J Hepatol, 2016;64(5):1167–75.
53. Ter Horst, K.W., Gilijamse, P.W., Versteeg, R.I. et al. Hepatic Diacylglycerol‐
Associated Protein Kinase Cε Translocation Links Hepatic Steatosis to
Hepatic Insulin Resistance in Humans. Cell Rep, 2017;19(10):1997–2004.
Biochem Biophys Res Commun, 1984;122(2):818–23.
55. Eichmann, T.O., Kumari, M., Haas, J.T. et al. Studies on the substrate and
stereo/regioselectivity of adipose triglyceride lipase, hormone‐sensitive
lipase, and diacylglycerol‐O‐acyltransferases. J Biol Chem, 2012;
287(49):41446–57.
56. Samuel, V.T., Liu, Z.‐X., Wang, A. et al. Inhibition of protein kinase Cepsilon
prevents hepatic insulin resistance in nonalcoholic fatty liver disease. J Clin
57. Petersen, M.C., Madiraju, A.K., Gassaway, B.M. et  al. Insulin receptor
Thr1160 phosphorylation mediates lipid‐induced hepatic insulin resistance.
J Clin Invest, 2016;126(11):4361–71.
58. Perry, R.J., Wang, Y., Cline, G.W. et  al. Leptin mediates a glucose‐fatty
acid  cycle to maintain glucose homeostasis in starvation. Cell,
2018;172(1):234–48.
59. Taylor, R., Magnusson, I., Rothman, D.L. et al. Direct assessment of liver
glycogen storage by 13C nuclear magnetic resonance spectroscopy and
­regulation of glucose homeostasis after a mixed meal in normal subjects.
J  Clin Invest, 1996;97(1):126–32.
60. Rothman, D.L., Magnusson, I., Katz, L.D. et  al. Quantitation of hepatic
­glycogenolysis and gluconeogenesis in fasting humans with 13C NMR.
Science, 1991;254 (5031):573–6.
61. Petersen, K.F., Price, T., Cline, G.W. et  al. Contribution of net hepatic
­glycogenolysis to glucose production during the early postprandial period.
Am J Physiol, 1996;270(1 Pt 1):E186–91.
62. Magnusson, I., Rothman, D.L., Katz, L.D. et al. Increased rate of gluconeo-
genesis in type II diabetes mellitus. A 13C nuclear magnetic resonance study.
J Clin Invest, 1992;90(4):1323–7.
63. Wajngot, A., Chandramouli, V., Schumann, W.C. et al. Quantitative contribu-
tions of gluconeogenesis to glucose production during fasting in type 2
­diabetes mellitus. Metabolism, 2001;50(1):47–52.
64. Marchesini, G., Brizi, M., Bianchi, G. et al. Nonalcoholic fatty liver disease:
a feature of the metabolic syndrome. Diabetes, 2001;50(8):1844–50.
65. Kahn, S.E. The relative contributions of insulin resistance and beta‐cell
dysfunction to the pathophysiology of Type 2 diabetes. Diabetologia,
­
2003;46(1):3–19.
66. Lee, Y.H., Wang, M.‐Y., Yu, X.‐X. et al. Glucagon is the key factor in the
development of diabetes. Diabetologia, 2016;59(7):1372–5.
67. Shulman, G.I., Liljenquist, J.E., Williams, P.E. et al. Glucose disposal during
insulinopenia in somatostatin‐treated dogs. The roles of glucose and
­glucagon. J Clin Invest, 1978;62(2):487–91.
68. Roden, M., Perseghin, G., Petersen, K.F. et al. The roles of insulin and gluca-
gon in the regulation of hepatic glycogen synthesis and turnover in humans.
J Clin Invest, 1996;97(3):642–8.
69. Reaven, G.M., Chen, Y.D., Golay, A. et al. Documentation of hypergluca-
gonemia throughout the day in nonobese and obese patients with noninsulin‐
dependent diabetes mellitus. J Clin Endocrinol Metab, 1987;64(1):106–10.
70. Raskin, P. and Unger, R.H. Hyperglucagonemia and its suppression.
Importance in the metabolic control of diabetes. N Engl J Med,
1978;299(9):433–6.
71. Basu, R., Schwenk, W.F., and Rizza, R.A. Both fasting glucose production
and disappearance are abnormal in people with “mild” and “severe” type 2
72. Petersen, M.C., Vatner, D.F., and Shulman, G.I. Regulation of hepatic
glucose metabolism in health and disease. Nat Rev Endocrinol,
­
2017;13(10):572–87.
73. Accili, D. and Arden, K.C. FoxOs at the crossroads of cellular metabolism,
differentiation, and transformation. Cell, 2004;117(4):421–6.
74. Liu, Y., Dentin, R., Chen, D. et  al. A fasting inducible switch modu-
lates  gluconeogenesis via activator/coactivator exchange. Nature,
2008;456(7219):269–73.
75. Koo, S.‐H., Flechner, L., Qi, L. et al. The CREB coactivator TORC2 is a key
regulator of fasting glucose metabolism. Nature, 2005;437(7062):1109–11.
76. Dentin, R., Liu, Y., Koo, S.‐H. et al. Insulin modulates gluconeogenesis by
inhibition of the coactivator TORC2. Nature, 2007;449 7160;, 366–9.
 37:  Non‐alcoholic Fatty Liver Disease and Insulin Resistance 471
77. Samuel, V.T., Beddow, S.A., Iwasaki, T. et al. Fasting hyperglycemia is not
associated with increased expression of PEPCK or G6Pc in patients with
Type 2 Diabetes. Proc Natl Acad Sci, 2009;106(29):12121–6.
78. Perry, R.J., Camporez, J.‐P.G., Kursawe, R. et al. Hepatic acetyl CoA links
adipose tissue inflammation to hepatic insulin resistance and type 2 diabetes.
Cell, 2015;160(4):745–58.
79. Perry, R.J., Zhang, D., Zhang, X.‐M. et al. Controlled‐release mitochondrial
protonophore reverses diabetes and steatohepatitis in rats. Science,
2015;347(6227):1253–6.
80. Zingone, A., Hiraiwa, H., Pan, C.‐J. et  al. Correction of glycogen storage
disease type 1a in a mouse model by gene therapy. J Biol Chem,
2000;275(2):828–32.
81. Previs, S.F., Cline, G.W., and Shulman, G.I. A critical evaluation of mass
isotopomer distribution analysis of gluconeogenesis in vivo. Am J Physiol,
1999;277(1 Pt 1):E154–60.
82. Titchenell, P.M., Chu, Q., Monks, B.R. et  al. Hepatic insulin signalling
is  dispensable for suppression of glucose output by insulin in vivo.
Nat Commun, 2015;12(6):7078.
83. Buettner, C., Patel, R., Muse, E.D. et al. Severe impairment in liver insulin
signaling fails to alter hepatic insulin action in conscious mice. J Clin Invest,
2005;115(5):1306–13.
84. Reaven, G.M., Hollenbeck, C., Jeng, C.Y. et al. Measurement of plasma glu-
cose, free fatty acid, lactate, and insulin for 24 h in patients with NIDDM.
Diabetes, 1988;37(8):1020–4.
85. Puhakainen, I., Koivisto, V.A., and Yki‐Järvinen, H. Lipolysis and gluconeo-
genesis from glycerol are increased in patients with noninsulin‐dependent
diabetes mellitus. J Clin Endocrinol Metab, 1992;75(3):789–94.
86. Paolisso, G., Tataranni, P.A., Foley, J.E. et al. A high concentration of fasting
plasma non‐esterified fatty acids is a risk factor for the development of
NIDDM. Diabetologia, 1995;38(10):1213–17.
87. Hundal, R.S., Krssak, M., Dufour, S. et al. Mechanism by which metformin
reduces glucose production in type 2 diabetes. Diabetes, 2000;49(12):2063–9.
88. Madiraju, A.K., Erion, D.M., Rahimi, Y. et al. Metformin suppresses gluco-
neogenesis by inhibiting mitochondrial glycerophosphate dehydrogenase.
Nature, 2014;510(7506): 542–6.
89. El‐Mir, M.Y., Nogueira, V., Fontaine, E. et al. Dimethylbiguanide inhibits
cell respiration via an indirect effect targeted on the respiratory chain
­complex I. J Biol Chem, 2000;275(1):223–8.
90. Foretz, M., Hébrard, S., Leclerc, J. et  al. Metformin inhibits hepatic
­gluconeogenesis in mice independently of the LKB1/AMPK pathway via a
decrease in hepatic energy state. J Clin Invest, 2010;120(7):2355–69.
91. Miller, R.A., Chu, Q., Xie, J. et al. Biguanides suppress hepatic glucagon
signalling by decreasing production of cyclic AMP. Nature,
2013;494(7436):256–60.
92. Konopka, A.R., Esponda, R.R., Robinson, M.M. et al. Hyperglucagonemia
mitigates the effect of metformin on glucose production in prediabetes.
Cell Rep, 2016;15(7):1394–1400.
93. Tiikkainen, M., Häkkinen, A.‐M., Korsheninnikova, E. et  al. Effects of
rosiglitazone and metformin on liver fat content, hepatic insulin resistance,
insulin clearance, and gene expression in adipose tissue in patients with
type 2 diabetes. Diabetes 2004; 53(8):2169–76.
94. Basu, R., Basu, A., Chandramouli, V. et  al. Effects of pioglitazone and
metformin on NEFA‐induced insulin resistance in type 2 diabetes.
­
Diabetologia, 2008;51(11):2031–40.
95. Shulman, G.I. Cellular mechanisms of insulin resistance. J Clin Invest,
2000;106(2):171–6.
96. Mayerson, A.B., Hundal, R.S., Dufour, S. et al. The effects of rosiglitazone
on insulin sensitivity, lipolysis, and hepatic and skeletal muscle triglyceride
content in patients with type 2 diabetes. Diabetes, 2002;51(3):797–802.
97. Samuel, V.T. and Shulman, G.I. Nonalcoholic fatty liver disease as a nexus
of metabolic and hepatic diseases. Cell Metab, 2018;27(1):22–41.
98. Petersen, K.F., Dufour, S., Befroy, D. et  al. Reversal of nonalcoholic
hepatic steatosis, hepatic insulin resistance, and hyperglycemia by
­moderate weight reduction in patients with type 2 diabetes. Diabetes,
2005;54(3):603–8.
99. Klein, S., Mittendorfer, B., Eagon, J.C. et  al. Gastric bypass surgery
improves metabolic and hepatic abnormalities associated with nonalco-
holic fatty liver disease. Gastroenterology, 2006;130(6):1564–72.
100. Sato, F., Tamura, Y., Watada, H. et  al. Effects of diet‐induced moderate
weight reduction on intrahepatic and intramyocellular triglycerides and
glucose metabolism in obese subjects. J Clin Endocrinol Metab,
2007;92(8):3326–9.
101. Viljanen, A.P.M., Iozzo, P., Borra, R. et al. Effect of weight loss on liver
free fatty acid uptake and hepatic insulin resistance. J Clin Endocrinol
Metab, 2009;94(1):50–5.
102. Lean, M.E.J., Leslie, W.S., Barnes, A.C. et  al. Primary care‐led weight
management for remission of type 2 diabetes (DiRECT): an open‐label,
cluster‐randomized trial. Lancet, 2018;391(10120):541–51.
103. Taylor, R., Al‐Mrabeh, A., Zhyzhneuskaya, S. et al. Remission of human
type 2 diabetes requires decrease in liver and pancreas fat content but
is  dependent upon capacity for beta cell recovery. Cell Metab,
2018;28:547–56.
104. Oh, S., Shida, T., Yamagishi, K. et  al. Moderate to vigorous physical
­activity volume is an important factor for managing nonalcoholic fatty liver
disease: A retrospective study. Hepatology, 2015;61(4):1205–15.
105. Zhang, H.‐J., He, J., Pan, L.‐L. et  al. Effects of moderate and vigorous
exercise on nonalcoholic fatty liver disease: a randomized clinical trial.
JAMA Intern Med, 2016;176(8):1074–82.
106. Bacchi, E., Negri, C., Targher, G. et al. Both resistance training and aerobic
training reduce hepatic fat content in type 2 diabetic subjects with nonalco-
holic fatty liver disease (the RAED2 randomized trial). Hepatol Baltim Md,
2013;58(4):1287–95.
107. Perseghin, G., Lattuada, G., De Cobelli, F. et al. Habitual physical activity
is associated with intrahepatic fat content in humans. Diabetes Care,
2007;30(3):683–8.
108. Perseghin, G., Price, T.B., Petersen, K.F. et al. Increased glucose transport‐
phosphorylation and muscle glycogen synthesis after exercise training in
insulin‐resistant subjects. N Engl J Med, 1996;335(18):1357–62.
 AMPK: Central Regulator
38 of Glucose and Lipid
Metabolism and Target
of Type 2 Diabetes
Therapeutics
Daniel Garcia, Maria M. Mihaylova, and Reuben J. Shaw
Molecular and Cell Biology Laboratory, The Salk Institute for Biological Studies, La Jolla, CA, USA
AMPK STRUCTURE AND MECHANISM
OF ACTIVATION
A fundamental requirement of all cells is that they couple the
availability of nutrients to signals emanating from growth fac-
tors to drive proliferation only when nutrients are in sufficient
abundance to guarantee successful cell division. Even in non‐
dividing cells, nutrients in the environment supply the necessary
building blocks for cellular metabolism and survival, and fuel
the bioenergetic needs of the cell by providing substrates to pro-
duce intracellular ATP to be used for all cellular processes.
When nutrient levels fall, ATP levels fall and unless ATP‐con-
suming biosynthetic processes are curtailed, a critical shortage
of ATP will cause catastrophic cellular demise. Eukaryotic cells
all share a highly conserved metabolic checkpoint that acts as a
sensor of ATP levels in the cell, the AMP‐activated protein
kinase (AMPK). As intracellular ATP levels fall due to patho-
logical stresses such as glucose or oxygen shortages, osmotic
stress, or disruptions of glycolysis or mitochondrial oxidative
phosphorylation will all result in decreased ATP. Upon activa-
tion under low ATP conditions, AMPK acts a metabolic check-
point in the cell, suppressing ATP‐consuming biosynthetic
processes while stimulating ATP‐generating processes to repair
the initiating loss of ATP [1]. Upon activation, AMPK initiates
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
acute effects on metabolic enzymes as well as prolonged adap-
tions in glucose and lipid metabolism through modulation of
transcriptional programs of metabolic enzymes. In addition to
ubiquitous roles as an energy checkpoint, AMPK also plays
additional key roles in glucose and lipid metabolism in special-
ized metabolic tissues in mammals and higher eukaryotes such
as liver, muscle, and adipose [2]. Thus AMPK not only governs
cellular energetics, but indeed overall organismal bioenergetics
by coordinating the response between tissues to nutritional
input.
AMPK was first discovered as a mammalian protein kinase
that is activated by changes in intracellular adenosine nucleotide
levels [3]. However, it was not until later, that the yeast ortholog
SNF‐1 (sucrose non‐fermenting complex) was annotated from a
Saccharomyces cerevisiae mutant screen for cells that failed to
grow on non‐fermentable carbon sources or sucrose [4, 5].
The AMP‐activated protein kinase (AMPK) is an obligate
heterotrimeric kinase complex composed of a catalytic (α) sub-
unit and two regulatory (β and γ) subunits. AMPK is activated
under conditions of energy stress, when intracellular ATP levels
decline and intracellular AMP increases, as occurs during nutri-
ent deprivation or hypoxia [2]. Upon energy stress, AMP
directly binds to tandem repeats of cystathionine‐β‐synthase
(CBS) domains in the AMPK γ subunit. Binding of AMP is
 38:  CENTRAL REGULATOR OF GLUCOSE AND LIPID METABOLISM AND TARGET OF TYPE 2 DIABETES THERAPEUTICS 473

thought to prevent dephosphorylation of the critical activation
loop threonine in the α subunit [6]. The phosphorylation of the
activation loop threonine is absolutely required for AMPK acti-
vation. In mammals, there are seven mammalian genes encod-
ing each of the α, β, γ subunits, allowing for 12 distinct
heterotrimeric variations (see Figure 38.1). There are two genes
encoding catalytic subunits, α1 and α2, two regulatory β and
three γ subunits that participate in the heterotrimer. Of these, γ3
appears mostly skeletal muscle specific and α2 appears to be
most highly expressed in key metabolic tissues including mus-
cle and liver. The catalytic α1 and α2 subunits contain a kinase
domain within their N‐terminus, as well as a critical region for
binding β and γ subunits within their C‐terminus. All kinases
possess an activation loop which is often a target for upstream
kinases, creating a conformation change that allows substrate
assess to the catalytic pocket, and the activation loop of the
AMPK α subunits contains a single threonine (Thr172 in mam-
malian AMPKα) that is the key regulatory site whose phospho-
rylation is absolutely required across all species for AMPK
activation. It has been shown that α1 catalytic isoform is found
mainly in the cytoplasm, whereas the α2 isoform appears to be
nuclear in some cell types. In liver, phosphorylation of both the
α1 and α2 subunits accounts for half of the total AMPK activity
and there appears to be no preferential binding of the α1 and α2
subunits with the different β or γ subunits [7]. The crystal struc-
tures of S. cerevisiae Snf1 and the human α2 kinase domains
were first revealed (8, 9), but in the past decade our molecular
understanding of the regulation and function of AMPK has sig-
nificantly been advanced by the elucidation of the crystal struc-
tures of a variety of AMPK holoenzymes [10–16]. Although
some details differ, together these studies provide a detailed
view of the architecture of the AMPK complex. The structure of
the AMPK trimeric complex consists of three major segments or
“modules”: the catalytic module, the carbohydrate‐binding
module (CBM), and the nucleotide‐binding module (also called
“regulatory fragment”). The activation loop of the α‐subunit
resides at the interface between the catalytic and nucleotide‐
binding modules, in close proximity to the C‐terminus of the
β‐subunit and the CBS repeats of the γ‐subunit. This structural
arrangement ensures that phosphorylation and dephosphoryla-
tion of Thr172 is sensitive to conformational rearrangements
induced by nucleotide binding. The catalytic domain exhibits a
typical eukaryotic serine/threonine kinase domain structure
with a small N‐lobe and a large C‐lobe. The CBM directly con-
tacts the N‐lobe of the kinase domain and the interface between
these two modules forms a discrete pocket that was identified as
the binding site for many direct AMPK‐activating compounds.
It is speculated that natural metabolites might bind this site to
regulate AMPK, however, no such metabolite has been yet iden-
tified. The nucleotide‐binding module is made up mostly by the
γ‐subunit, which forms a flattened disk with the CBS repeats
symmetrically arranged around the disk, one in each quadrant.

Length Major Site Chromosome

T172 of Expression Location
βγ Binding
α1 Kinase domain α-CTD 550 ubiquitous 5p11-14

T172

liver, muscle
α2 Kinase domain α-CTD 552 1p31
widespread

glycogen
binding αγ Binding
β1 GBD 270 ubiquitous 12q24.1-.3

β2 GBD 272 ubiquitous? 1q21.1

Bateman domain 1 Bateman domain 2

AMP AMP AMP
γ1 CBS1 CBS2 CBS3 CBS4 331 ubiquitous 12q12-14

γ2 long γ2-NTD CBS1 CBS2 CBS3 CBS4 569 ubiquitous 7q36

γ2 short CBS1 CBS2 CBS3 CBS4 328 ubiquitous 7q36

γ3 long γ3-NTD CBS1 CBS2 CBS3 CBS4 489 skeletal muscle 2q35

γ3 short γ3-NTD CBS1 CBS2 CBS3 CBS4 464 skeletal muscle 2q35

Figure 38.1  Human AMPK subunit isoforms. Domain structure, expression pattern, and alternative splice isoforms of the two catalytic kinase (α)
isoforms, the two beta regulatory subunits which contain a glycogen‐binding domain (GBD), and the three genes encoding the gamma subunits,
which each contain 4 CBS domains which directly bind to AMP as drawn.
474 THE LIVER:  UPSTREAM REGULATORS OF AMPK: LKB1 AND CAMKK
Mechanistically, these crystallographic studies reveal the
molecular details of how adenine nucleotides and small molecule
activators activate AMPK. In the case of nucleotides, the crystal
structures show that when AMP is bound to site 3, the γ‐subunit
forms stable interactions with a few amino acids within the α‐
linker’s α‐RIM1 and α‐RIM2, which interact with the unoccupied
site 2 and the AMP molecule bound at site 3, respectively [12, 15,
16]. The binding of the α‐RIM motifs to the γ‐subunit restricts the
flexibility of the α‐linker, resulting in tighter association of the
catalytic and nucleotide‐binding modules, which physically pro-
tects Thr172 from dephosphorylation. Interestingly, the same
effect is proposed to occur when ADP binds site 3, raising the
possibility that in some contexts ADP might be the relevant
AMPK activating signal [14]. Moreover, the binding of the α‐
RIM motifs to the γ‐subunit shifts the autoinhibitory domain
(AID) in the α‐subunit away from the kinase domain when AMP
is bound, thus releasing the AID’s negative effects on the kinase
domain [10, 12, 13]. This rearrangement of the AID domain may
represent the molecular basis for the allosteric activation effect of
AMP. According to this model, the AID can shift between kinase
domain‐bound (inactive AMPK) and nucleotide‐module‐bound
states (active AMPK) depending on the nucleotide binding status.
In summary, the published crystal structures concur that binding
of AMP, especially at site 3, induces a conformational change that
is transmitted to the kinase domain by changes in the interaction
of the α‐RIM motifs and the AID with the nucleotide‐binding
module. These structural changes, which are opposed by ATP,
result in the allosteric activation of AMPK and a compaction of
the interface between the catalytic module and the nucleotide‐
binding module, which protects Thr172 from dephosphorylation.
However, it is not clear whether these structural rearrangements
also promote Thr172 phosphorylation.
On the other hand, activating compounds, such as A769662,
activate AMPK by a different mechanism. Binding of these com-
pounds, together with phosphorylation of serine 108 in the β‐sub-
unit, stabilizes the CBM and strengthens the interaction of the
CBM with the kinase domain [10, 13, 15]. Specifically, binding
of activating compounds induces the formation of an α‐helix in
the β‐subunit, termed the C‐interacting helix, which interacts with
the so‐called C‐helix of the kinase domain (a conserved helix
across multiple kinases which is important for ATP binding). This
conformational change results in a shift toward a closed, active
conformation of the kinase domain, protection from Thr172
dephosphorylation, and increased substrate affinity. Interestingly,
glycogen inhibits the CBM‐KD interaction and this may be the
mechanism by which glycogen inhibits AMPK [17].
STRESS, HORMONES,
AND THERAPEUTICS ACTIVATE AMPK
Animals, in their multi‐cell complexity and multi‐organ utiliza-
tion, have evolved intricate mechanisms to sense and initiate
immediate responses to energy requirement or nutrient depriva-
tion. Multiple studies have now shown that cellular stress caused
by starvation or exercise can activate AMPK. In single‐cell
eukaryotes as well as mammalian cell culture conditions,
AMPK is activated in response to nutrient depletion of glucose
or oxygen (hypoxia). It is now evident, that as complex organ-
isms developed, various circulating hormones gained function
to act as whole‐organism sensors and are capable of turning on
AMPK in response to metabolic stresses such as starvation. One
well‐known adipokine, adiponectin has been shown to activate
AMPK in liver, leading to fatty‐acid oxidation and decrease in
blood glucose levels, consistent with previous findings that adi-
ponectin is able to suppress hepatic glucose production [18]. In
addition to hormonal input from adiponectin, AMPK activity
has been shown to be modulated by leptin, resistin, ghrelin, epi-
nephrine, and cannabinoids [19]. Exercise is another metabolic
stress that has been shown to activate AMPK in response to
muscle contraction. Such activation may be due to increased
AMP to ATP ratios caused by movement and muscle contrac-
tion, correlating to studies in mice where electrical muscle stim-
ulation increased AMP levels and turned on AMPK [20].
Two different classes of drugs for treatment of diabetes mel-
litus type 2, biguanides such as metformin [21] and thiazolidin-
ediones (TZDs) such as rosiglitazone [22] and pioglitazone [23]
have been shown to activate hepatic AMPK, most likely through
perturbation of mitochondrial ATP output via mild inhibition of
complex I of the mitochondrial respiratory chain. Metformin is
a biguanide that has been structurally modified from galegine, a
naturally occurring compound found in the French lilac (Galega
officinalis). Although it was not until 1918 that biguanides were
discovered to have a blood glucose lowering effect, people have
been using the French lilac to ameliorate a variety of maladies
since the Middle Ages [24].
Bitter melon or Momordica charantia, like the French lilac,
has been used for hundreds of years in traditional Chinese medi-
cine to treat many different ailments and it was not until recently
that scientists isolated triterpenoid compounds from M. charan-
tia that are able to activate AMPK and facilitate fatty acid oxida-
tion and glucose utilization when administrated in mice [25]. In
a recent study, AMPK has been also been shown to become
active in response to resveratrol, a polyphenol that is found in
the skin of red grapes, certain nuts and berries, and has been
linked to longevity in model organisms such as yeast, nema-
todes, drosophila, fish, and mice. Further, it was shown that res-
veratrol could mimic the benefits of dietary restriction in mice
by perturbing fatty liver phenotype and increasing insulin sensi-
tivity in animals fed high calorie diet [26]. Although direct
mutations in AMPK have not been found in diabetic patents,
studies show that AMPK is implicated in various pathways
deregulated in such metabolic disorders and makes it an attrac-
tive therapeutic target in treatment of such diseases.
UPSTREAM REGULATORS OF AMPK:
LKB1 AND CAMKK
There was evidence for an upstream kinase activating AMPK as
early as 1978 [27] and in the years to follow scientists intently
labored to identify the kinase responsible for phosphorylation
and activation of catalytic subunits α1 or α2. However, it was not
until 1996 that an “AMP‐ activated protein kinase” (AMPKK)
was partially purified from rat liver and shown to phosphorylate
AMPK on Thr172 [28]. Quickly after the budding yeast genome
 38:  CENTRAL REGULATOR OF GLUCOSE AND LIPID METABOLISM AND TARGET OF TYPE 2 DIABETES THERAPEUTICS 475

was completed, three upstream kinases Sak1 (Pak1), Elm1, and DOWNSTREAM TARGETS I:
Tos3 were identified by whole genome screening methods and REGULATION OF ACUTE METABOLIC
were shown to act upstream of the yeast AMPK orthologue, the
SNF1 complex. When these kinases were genetically knocked
RESPONSE – ENZYME IN LIPOGENESIS
out in yeast, the results yielded the same phenotype as a snf1
In the liver, AMPK phosphorylates and regulates multiple
mutant, again placing them upstream of the SNF1 complex [29,
downstream targets involved in lipogenesis and lipid homeosta-
30]. In the human genome the closest related kinases to the yeast
sis (Figure 38.2). One of the first identified downstream targets
ones were found to be Calmodulin‐dependent protein kinases,
of AMPK was acetyl‐coenzyme A carboxylase (ACC) [41].
CaMKKα and CaMKKβ and the protein kinase LKB1. Several
ACC is an enzyme involved in the generation of fatty acid pre-
groups simultaneously showed that LKB1 [1, 31, 32] and
cursor malonyl‐CoA, a key metabolite in the regulation of
CAMKKβ [33–35] were indeed the upstream kinases acting on
energy homeostasis. Two genes encoding two different ACC
AMPK in mammals and were capable of phosphorylating both
isoforms are found in mammals – ACC1 and ACC2 – and they
AMPKα1 and AMPKα2 subunits on Thr172.
appear to have distinct tissue specificity. It has been shown that
Interestingly, LKB1 was first identified in humans as a serine/
ACC1 and ACC2 control the synthesis of two different pools of
threonine tumor suppressor kinase that is defective in the cancer
malonyl‐CoA production. ACC1 is thought to suppress the pro-
predisposing Peutz–Jeghers Syndrome [36]. In mammalian
duction of malonyl‐CoA used in fatty acid synthesis whereas
cells, LKB1 exists in a complex with two other proteins, STRAD
ACC2 stimulates fatty acid oxidation (reviewed in [42]).
(sterile‐20‐related adaptor) and MO25 (mouse protein‐25) [31]
AMPK inhibits both ACC1 and ACC2 through direct phos-
and when bound to these accessory proteins, it is stabilized and
phorylation of their homologous residues Ser79 in ACC1 and
constitutively activated. Recent data suggest that when AMP
Ser218 in ACC2. Downregulation of ACC1 activity leads to
nucleotides bind allosterically to the Bateman domains in the γ
reduced malonyl‐CoA levels and a decrease in lipogenesis.
subunit of AMPK complex, conformational changes occur that
AMPK phosphorylation of ACC2 inhibits its enzymatic activ-
protect the LKB1‐mediated phosphorylation Thr172 in the α
ity and decreases cellular levels of malonyl‐CoA levels, which
subunit [6], although effects on localization and complex forma-
leads to direct inhibition of mitochondrial fatty acid uptake
tion between LKB1 and AMPK following energy stress remain
and increased fatty acid oxidation and ATP production through
to be fully explored. In 2005, scientists showed that LKB1 is
carnitine palmitoyltransferase 1 (CPT‐1) (reviewed in [19]).
indeed the major upstream AMPK kinase in liver [37] and that
Interestingly, it has been also shown that calorie restriction can
genetic deletion of hepatic LKB1 almost completely reduces
increase AMPK activity causing a decrease of fatty acid syn-
hepatic AMPK activity. Lack of LKB1 in mouse liver rapidly
thesis or upregulation of fatty acid oxidation through inhibi-
leads to hyperglycemia and increased levels of gluconeogenic
tion of ACC. In a recent study, various AMPK activating
and lipogenic gene expression. It was also shown in these ani-
polyphenol compounds led to lowering of lipid accumulation
mals that activation of AMPK by the antidiabetic therapeutic
in HepG2 cells grown in high glucose and inhibited athero-
metformin is dependent on LKB1, and in the absence of hepatic
sclerosis in diabetic LDLR−/− mice treated with the various
LKB1, metformin is unable to lower blood glucose levels [37].
polyphenols [43].
However, it is important to note that AMPK is not the only sub-
In addition to reporting ACC as a downstream target of AMPK,
strate of LKB1. LKB1 similarly phosphorylates the activation
HMG‐CoA reductase was also found as one of the first key down-
loop of a family of 11 kinases all related to AMPK, also resulting
stream substrates [41, 44]. It had been already known for over ten
in their activation [38]. Importantly, of these 14 LKB1‐depend-
years at the point that HMG‐CoA reductase kinase (3‐hydroxy‐3‐
ent kinases, only AMPKα1 and AMPKα2 are activated under
methyl‐glutaryl‐CoA reductase or HMGR) activity was regulated
low ATP conditions, probably due to the fact that only they inter-
by an upstream kinase [27], however, that kinase had not yet been
act with AMPKγ which contains the AMP‐binding domains
discovered. Today, it is known that HMG‐CoA reductase is a rate‐
[39]. Little is currently known about what stimuli direct LKB1
limiting enzyme involved in the production of cholesterol and
toward any of these AMPK‐related kinases and current evidence
other isoprenoids, and more specifically functions in converting
suggests that LKB1 is constitutively active and these other
HMG‐CoA to mevalonic acid. By phosphorylating HMGR,
kinases may be regulated through phosphorylation at other sites
AMPK blocks anabolic or ATP consuming processes such as cho-
outside of their activation loops. Collectively, these findings map
lesterol synthesis in order to preserve intracellular ATP levels.
AMPK on the axis of a major tumor suppressor pathway and
Strikingly, the AMPK phosphorylation site in HMGR is con-
provide an interesting link between cancer and metabolism.
served throughout eukaryotes including plants.
In addition to LKB1, the calmodulin‐dependent protein kinases,
CaMKKα and CaMKKβ also regulate AMPK activity, though
only in response to calcium flux and not in response to changes in Downstream targets II: regulation of metabolic
AMP, which appears to work completely through LKB1 based on
adaptation: control of transcription
genetic knockout and RNAi studies. Genetic deletion of LKB1
dramatically reduces AMPK activation in liver suggesting little In response to changes in AMP : ATP ratios, AMPK can rapidly
role for CAMKK in this tissue, though in hypothalamic neurons regulate downstream targets via phosphorylation. However, in
controlling food intake, CAMKKβ (CAMKK2) appears to be the addition to these fast post‐translational modifications, AMPK
dominant upstream kinase for AMPK [40]. This is consistent with can also promote long‐term transcriptional changes and repro-
the fact that CaMKKα and CaMKKβ are most highly expressed in gram transcription of certain genes in response to the cellular
neurons, whereas LKB1 is more ubiquitously expressed. state. These effects are thought to be mediated by the direct
476 THE LIVER:  DOWNSTREAM TARGETS I: REGULATION OF ACUTE METABOLIC RESPONSE – ENZYME IN LIPOGENESIS

Exercise
Low Nutrients (glucose, O2) LKB1 adiponectin
Mitochondrial inhibitors: STRAD MO25 Ghrelin, cannabinoids
TZDs, biguanides
Resistin
AICAR (amp mimetic)
Resveratrol, polyphenols AMP CAMKKβ

Glycolysis inhibitors: 2-DG Ca2 +

P
Abbott A769662 γ
AMPKα
β

Glycolysis/ Glucose Uptake

iPFK2 Ser461 Transcriptional Control
TBC1D1 Ser237 of Glucose Metabolism

CRTC2 Ser171
Insulin Sensitivity FOXO3 Ser413
Fatty Acid Lipid Synthesis Protein Synthesis p300 Ser89
Oxidation AREBP Ser470
HMG CoR Ser872 TSC2 Ser1387
ACC2 Ser221 Raptor Ser792 HNF4a Ser313
ACC1 Ser79
IRS1 Ser794 Chrebp Ser568

Figure 38.2  The AMPK signaling pathway. AMPK is phosphorylated on its kinase activation loop threonine and activated by two distinct
upstream kinases in response to different stimuli. LKB1 activates AMPK in response to all stimuli that lower intracellular ATP and increase AMP.
CAMKKb activates AMPK in response to calcium flux in an AMP‐independent manner. AMPK is activated physiologically by exercise, low nutri-
ents such as lowered glucose or lowered oxygen, and hormones including ghrelin, leptin, adiponectin, and cannabinoids. Leptin is reported to
activate AMPK in peripheral tissues but inhibit AMPK in the central nervous system through poorly‐understood mechanisms. In addition, AMPK
is activated by agents that disrupt ATP production by inhibiting or poisoning the mitochondria, including uncouplers, or agents that inhibit glycoly-
sis such as the glucose analog 2‐deoxyglucose that as a competitive inhibitor for hexokinase. Additional pharmacological agents that activate
AMPK include resveratrol and related polyphenols as well as the cell‐permeable AMP‐mimetic AICAR and the first small molecule direct activator
Abbott A‐769662. Upon activation, AMPK serves to inhibit anabolic, ATP‐consuming biosynthetic processes such as protein, lipid, and glucose
synthesis, while upregulating catabolic processes to generate ATP production, including increased glycolysis, glucose uptake, and fatty acid oxida-
tion. AMPK modulates cell growth and insulin sensitivity through the mTOR signaling pathway. All current in best established in vivo direct AMPK
substrates and their AMPK phosphorylation sites are listed. All the sites listed conform to the identified optimal AMPK substrate motif.

phosphorylation of sequence‐specific transcription factors and
transcriptional coactivators by AMPK. Of the two best‐studied
transcriptional programs controlled by AMPK in liver, gluco-
neogenesis and lipogenesis, a number of direct substrates for
AMPK in gluconeogenesis have been identified.
Interestingly, two different coactivators which modulate
cyclic AMP‐responsive element binding protein (CREB)‐
dependent transcription in response to glucagon induced by
fasting are the histone acetyltransferase p300 and the CREB
coactivator CRTC2 (also known as TORC2 or transducer of
regulated CREB activity 2) [45–47]. CRTC2 is a critical rate‐
limiting transcriptional regulator of gluconeogenesis in mice
via effects on hepatic CREB targets including the PGC‐1α pro-
moter. When AMPK is active, it can phosphorylate CRTC2 on
residue Ser171, which in turn allows CRCT2 to associate to
14–3–3 proteins and be sequestered from the nucleus to the
cytoplasm, an event that shuts off transcription of gluconeo-
genic enzymes [48, 49]. If AMPK activity is decreased,
hypophosphorylated CRTC2 can conversely move into the
nucleus where it binds to transcriptional coactivator CREB and
promotes transcription of PGC1α and downstream gluconeo-
genic targets PEPCK and G6Pase (see Figure 38.3). Strikingly,
both p300 and CRTC2 have been shown to be phosphorylated
at the same regulatory sites by either AMPK or its related fam-
ily member SIK1 (salt‐inducible kinase 1), both of which are
LKB1‐dependent. In mice lacking hepatic LKB1, CRTC2 is
hypophosphorylated and predominantly nuclear compared to
wild‐type mice [37]. Furthermore, LKB1−/− mice in liver had
dramatic increases in fasting blood glucose levels, which were
greatly attenuated upon introduction of an shRNA reducing
CRTC2 levels in the liver, reinforcing the idea that gluconeo-
genesis in liver is controlled by LKB1‐dependant kinases and
that CRTC2 is a key downstream target of these kinases.
Consistent with these findings in the LKB1 liver‐specific
knockout mice, animals bearing a liver‐specific knockout of the
AMPKα2 isoform also exhibited elevated hepatic glucose out-
put, glucose intolerance, impaired leptin‐ and adiponectin‐reg-
ulated hepatic glucose production [50] (see Table 38.1). Future
studies will be needed to dissect the temporal and spatial regu-
lation of the contexts in which AMPKa2 or SIK1 control these
key modulators of CREB and gluconeogenesis. Interestingly,
SIK1 itself is a CREB target, providing a time‐delayed mecha-
nism to attenuate chronic CREB‐dependent transcription.
Phosphorylation of p300 and CRTC2 by AMPK and SIK
kinases may be key effectors of metformin in the control of
type 2 diabetes. How AMPK phosphorylation of p300 may
regulate its many other downstream interacting transcription
factors key to hepatic metabolism remains to be examined,
although mice lacking LKB1 or AMPK in liver provide excel-
lent tools for such future studies.
 38:  CENTRAL REGULATOR OF GLUCOSE AND LIPID METABOLISM AND TARGET OF TYPE 2 DIABETES THERAPEUTICS 477

14-3-3

LKB1 P

CRTC2
cytoplasm
P
P
SIK
14-3-3 AMPK nucleus
P
P
AMPK
p300

P PGC1α
P CRTC2 P P
ChREBP PEPCK
FASN CREB FoxO3 HNF4a AREBP
PGC-1α G6Pase

Inhibition of Inhibition of
Lipogenesis Gluconeogenesis

Figure 38.3  LKB1 and AMPK‐mediated control of metabolic transcriptional programs. AMPK and its related family members SIK1 and SIK2
(not shown) phosphorylate a common set of substrates including the CREB coactivator CRTC2 (CREB‐regulated transcriptional coactivator 2),
previously known as TORC2 (transducer of regulated CREB 2) and the histone acetyltransferase p300. Phosphorylation of CRTC2 creates a 14‐3‐3
docking site which then results in 14‐3‐3 mediated nuclear export of CRTC2. This cytoplasmic sequestration of CRTC2 by AMPK or SIK kinases
causes an inhibition of CREB‐dependent transcriptional targets including key mediators of gluconeogenesis like the PGC‐1α coactivator. In addi-
tion to suppressing PGC‐1a mRNA expression, AMPK also directly phosphorylates a handful of transcription factors (FOXO3, HNF4a, and
AREBP) that directly bind to the promoters of the two key gluconeogenic enzymes PEPCK and G6Pase. In addition to these gluconeogenesis regu-
lators, AMPK also is reported to phosphorylate Chrebp, a key lipogenic transcription factor that controls levels of fatty acid synthase (FASN),
acetyl‐CoA carboxylase 1 (ACC1), and L‐pyruvate kinase mRNA.

Table 38.1  Mouse models of AMPK/LKB1 function in liver It has been also shown that a subset of patients with MODY
Mouse model Metabolic phenotype Reference (maturity‐onset diabetes of the young) form of diabetes harbor
AMPKα1 knockout None [48]
mutations in HNF4α [53]. Interestingly, HNF4α liver specific
AMPKα2 knockout Glucose uptake in muscle [49] knockout mice do not develop hyperglycemia unlike MODY
Hyperglycemia, low insulin [48] patients, but do develop lipid accumulation, consistent with
AMPKα2 liver knockout Hyperglycemia [46] altered triglyceride levels in MODY patients [54]. These find-
Glucose intolerance
Hyperlipidemia ings further reinforce the possible role of AMPK in lipid and
LKB1 liver knockout Hyperglycemia [30] glucose homeostasis through modulation of downstream tran-
Glucose intolerance scriptional regulators such as HNF4α.
Hyperlipidemia
In addition to HNF4α, in a 2006 report AMPK was shown to
directly phosphorylate another transcription factor named the
In addition to direct phosphorylation of coactivators, AMPK AICAR response‐element binding protein (AREBP) that
also phosphorylates the hepatocyte nuclear factor 4 alpha directly binds the PEPCK promoter [55]. AMPK phosphoryla-
(HNF4α), which is a key transcription factor of the nuclear tion of AREBP on Ser470 reduces its ability to bind DNA and in
receptor superfamily which binds to the promoters of the two turn lowers expression of PEPCK. Thus, like HNF4α, AMPK
key gluconeogenic enzymes, phosphoenolpyruvate carboxykin- phosphorylation of AREBP controls its ability to bind to DNA
ase (PEPCK) and G6Pase. It is hypothesized that AMPK phos- and promote transcription of downstream target genes.
phorylation of HNF4α on residue Ser304 decreases the protein’s In addition to effects on gluconeogenesis, AMPK activation
stability by interfering with its ability to dimerize and bind to is also known to inhibit hepatic lipogenesis. While some of that
DNA and may further promote its degradation [51, 52]. In addi- is due to acute effects on lipogenic enzymes as previously dis-
tion to PEPCK and G6Pase, HNF4α controls gene expression of cussed, it is also known that AMPK activation leads to decreased
glucose transporter GLUT2 and glycolytic enzymes such as transcription of key hepatic lipogenic enzymes, including fatty
aldolase B and liver‐type pyruvate kinase (L‐PK), which are acid synthase (FASN) as well as ACC1. Two sequence–sequence
diminished when AMPK is activated by 5‐aminoimidazole‐4‐ transcription factors are known to coregulate many of these
carboxamide‐1‐β‐D‐ribofuranoside (AICAR) in hepatocytes [51]. lipogenic enzymes: the sterol‐responsive binding protein
478 THE LIVER:  DOWNSTREAM TARGETS I: REGULATION OF ACUTE METABOLIC RESPONSE – ENZYME IN LIPOGENESIS
(SREBP‐1c) and the carbohydrate response binding protein
(ChREBP). ChREBP is highly expressed in liver and has essen-
tial roles in glucose‐induced transcription of liver pyruvate
kinase (L‐PK), in addition to its effects on lipogenic enzyme
promoters. Phosphorylation by AMPK on Ser568 of ChREBP
[56], promotes decreased DNA binding which causes a decrease
in the transcription of glycolytic and lipogenic genes. It is
known now that glucose metabolism can be repressed by fatty
acids, which act as another source of energy when glucose needs
to be preserved. This effect has been named the fatty acid “spar-
ing effect” on glucose.
Like ChREBP, SREBP‐1 is also directly regulated by AMPK
[21, 57, 61]. Due to its rate‐limiting effects on lipogenic
enzyme expression, SREBP‐1 has been linked to insulin resist-
ance, dyslipidemia and type 2 diabetes [58, 59]. In one study,
treatment of rat hepatocytes with AMPK activators such as
metformin or AICAR led to suppressed SREBP‐1 mRNA
expression, and also lowered hepatic mRNA levels for
SREPB‐1 controlled genes FAS and S14 [21]. This suggests
that AMPK activation through metformin can inhibit the
expression of lipogenic genes. Indeed, metformin treatment or
overexpression of an activated allele of AMPK was found to be
sufficient to reduce triglyceride content in insulin resistant
HepG2 cells [60]. Mice lacking hepatic AMPK function due to
liver‐specific LKB1 deletion show elevated SREBP1 and
SREBP1 target genes resulting in lipid accumulation and
hepatic steatosis [37]. In 2011, Srebp1 was found to be directly
phosphorylated by AMPK and its highly related kinase SIK1,
both of which appear to suppress activation of full‐length
Srebp1a in the ER membrane, preventing appearance of
Srebp1c in the nucleus [61]. While the activity of the SREBP
proteins is primarily regulated through intracellular concentra-
tions of unsaturated fatty acids and cholesterol, AMPK phos-
phorylation of Ser372 on SREBP1 may also inhibit its activity
by preventing proteolytic processing and thus transcriptional
activity. Additional studies are needed however to truly sort out
the requirement and roles for AMPK‐dependent phosphoryla-
tion of Srebp1 in vivo. As seen from the above‐mentioned stud-
ies, AMPK plays a major role in the gluconeogenic
transcriptional program through p300, CRTC2, HNF4α, and
AREBP, as well as the lipogenic transcriptional program
through ChREBP and SREBP‐1.
Finally, studies have hinted that AMPK may play a more
global role in transcription through direct phosphorylation and
regulation of transcriptional repressors histone deacetylases
class IIa (HDACs IIa) enzymes [62] and transcriptional activa-
tor histone acetyltransferase p300 [45]. It has been shown that
AMPK can phosphorylate class IIa HDAC5 on residues Ser259
and Ser498 in human primary myotubes, promoting 14‐3‐3
binding and dissociation from DNA binding transcription part-
ner myocyte enhancer factor‐2 (MEF2), which in turn allows
expression of downstream target genes [62]. Although direct
regulation of class IIa HDACs by AMPK has not yet been impli-
cated in liver metabolism, this provides an attractive postulate
where AMPK can simultaneously control multiple downstream
transcriptional events in response to upstream metabolic
stresses. Such analysis will not happen without challenges,
since it has been shown that multiple upstream kinases are capa-
ble of phosphorylating the same critical residues of HDAC 5
[63]. Indeed, we previously demonstrated that AMPK‐related
kinases downstream of LKB1 can phosphorylate HDAC4,
HDAC5, and HDAC7 in mouse liver, which suppresses their
ability to enhance FOXO‐dependent gluconeogenesis [64].
Downstream targets III: regulation of cell
growth and insulin signaling via mTOR
Environmental factors cue cells to cease growing and dividing
when conditions are unfavorable. These mechanisms are con-
served from the smallest and simplest of eukaryotes to the most
complex ones. When nutrients are scarce, cellular energy sensor
AMPK gets activated and inhibits energy demanding processes
such as protein synthesis and cell growth. One of the ways
AMPK accomplishes that task is by negatively regulating the
mechanistic target of rapamycin (mTOR) pathway. mTOR is a
serine/threonine kinase highly conserved in all eukaryotes and a
central regulator of cell growth. Whereas AMPK is active under
nutrient‐poor conditions and inactive under nutrient‐rich condi-
tions, mTOR is activated in the inverse pattern. In higher eukar-
yotes, mTOR activation requires positive signals from both
nutrients (glucose, amino acids) and growth factors. mTOR, like
its budding yeast orthologs, is found in two biochemically and
functionally discrete signaling complexes [65]. In mammals,
the mTORC1 complex is composed of four known subunits:
raptor (regulatory associated protein of mTOR), PRAS40,
mLST8, and mTOR. The mTORC2 complex contains rictor
(rapamycin insensitive companion of mTOR), mSIN1, PRR5/
Protor, mLST8, and mTOR [66]. Signaling from mTOR com-
plex 1 (mTORC1) is nutrient‐sensitive, acutely inhibited by the
bacterial macrolide rapamycin, and controls cell growth, angio-
genesis, and metabolism. In contrast, mTORC2 is not sensitive
to nutrients, nor acutely inhibited by rapamycin, and its known
substrates include the hydrophobic motif phosphorylation sites
in AGC kinases including Akt and PKC family members.
Downstream of the AMPK‐ and rapamycin‐sensitive raptor‐
mTOR (mTORC1) complex are its two well‐characterized sub-
strates: 4EBP1 and the p70 ribosomal S6 kinase. Phosphorylation
of 4EBP‐1 by mTORC1 suppresses its ability to bind and inhibit
the translation initiation factor eIF4E. mTORC1 mediates phos-
phorylation of S6K at a Thr residue in a hydrophobic motif at
the C‐terminus of the kinase domain. A specific motif (TOS
motif) found in both 4EBP1 and S6K was shown to mediate
direct binding of these proteins to raptor allowing them to be
phosphorylated in the mTORC1 complex. Mechanistic details
of how mTORC1 regulates the assembly of translational initia-
tion complexes via a number of ordered phosphorylation events
were recently discovered [67]. mTORC1‐dependent translation
is known to control a number of specific cell growth regulators,
including cyclin D1, the HIF‐1α transcriptional factor and c‐
myc, which in turn promote processes including cell cycle pro-
gression, cell growth, glycolysis, and angiogenesis, all
contributing to enhanced tumorigenesis [66].
Upstream components of the mTORC1 complex were initially
discovered through classical cancer genetics. The TSC2 tumor
suppressor tuberin and its obligate binding partner hamartin
(TSC1), are mutated in a familial tumor syndrome called tuber-
ous sclerosis complex (TSC). TSC patients are predisposed to
widespread benign tumors termed hamartomas in kidney, lung,
 38:  CENTRAL REGULATOR OF GLUCOSE AND LIPID METABOLISM AND TARGET OF TYPE 2 DIABETES THERAPEUTICS 479

brain, and skin. Genetic studies in Drosophila and mammalian
cells identified the TSC tumor suppressors as critical upstream
inhibitors of the mTORC1 complex. TSC2 (also known as
tuberin) contains a GTPase‐activating protein (GAP) domain at
its C‐terminus that inactivates the small Ras‐like GTPase Rheb,
which has been shown to associate with and directly activate the
mTORC1 complex in vitro [68]. Loss of TSC1 or TSC2 therefore
leads to hyperactivation of mTORC1. Phosphorylation of TSC1
and TSC2 serves as an integration point for a wide variety of envi-
ronmental signals that regulate mTORC1 [69]. One of the key
activators of the mTORC1 pathway is PI3‐kinase, which plays a
key role in promoting cell growth and insulin‐mediated effects on
metabolism. PI3‐kinase activates the serine/threonine kinase Akt
which directly phosphorylates and inactivates both TSC2 and an
inhibitor of the mTORC1 complex named PRAS40 [68, 70].
In addition to these growth factor cues that activate mTORC1,
the complex is rapidly inactivated by a wide variety of cell
stresses, thereby ensuring that cells do not continue to grow
under unfavorable conditions. One of the unique aspects of the
mTORC1 complex is that unlike many of the aforementioned
growth factor activated kinases, it is dependent on nutrient avail-
ability for its kinase activity. Withdrawal of glucose, amino
acids, or oxygen leads to rapid suppression of mTORC1 activity
even in the presence of full growth factors [69]. Upon LKB1‐
and AMP‐dependent activation of AMPK by nutrient loss,
AMPK directly phosphorylates the TSC2 tumor suppressor on
conserved serine sites distinct from those targeted by other
kinases, which constitutes one mechanism by which glucose
and oxygen control mTORC1 activation [71–74]. Interestingly,
TSC2 orthologs are absent from lower eukaryotes such as S.
cerevisiae and Caenorhabditis elegans and mammalian cells
lacking TSC2 still remain partially sensitive to AMPK activa-
tion, indicating that there may be an alternative and more ancient
backup mechanism which allows AMPK to inhibit cell growth
and proliferation through the mTORC1 pathway. Collectively,
these observations prompted the discovery of yet another novel
mechanism of inhibition. The mTOR kinase exists in complex
consisting of mLST8/Gbl, PRAS40, and the scaffold protein
Raptor. In a recent study, it was shown that AMPK is able to
directly phosphorylate Raptor on two conserved residues Ser722
and Ser792, which in turn induces binding to 14‐3‐3 proteins
and inactivation the mTORC1 complex [75]. Taken together
with previous studies, these findings indicate that AMPK
directly phosphorylates both TSC2 and raptor to inhibit
mTORC1 activity by a dual‐pronged mechanism (see
Figure 38.4). Importantly, metformin treatment of mice leads to
robust phosphorylation of raptor Ser792 in murine liver, an
effect which is ablated in the LKB1‐liver specific knockout
mice [75]. LKB1‐liver specific knockout mice which lack
AMPK activity in liver exhibit hyperactivation of mTORC1
signaling in the liver including increased phosphorylation of
S6K1 and 4EBP1 under ad libitum fed conditions [37]. In addi-
tion, hormones that activate AMPK in liver including glucagon
[76] and adiponectin [77] have been reported to suppress
mTORC1 signaling. Very recently, our laboratory collaborating
with the lab of Brendan Manning demonstrated that AMPK is
genetically required in primary hepatocytes and in the livers of
mice for metformin to suppress mTORC1 signaling [78]. Using
liver‐specific knockouts of AMPKα1 and AMPKα2 we found
that metformin was completely unable to suppress mTORC1
after metformin treatment in mice. However, in cell culture, at
higher doses and at later timepoints, metformin induced sup-
pression of mTORC1 signaling suggestive of a transcriptional
response which may be due to activation of the mito ER stress
pathway and ATF4‐dependent activation of REDD1 mRNAs
and other mTORC1 inhibitor mRNAs [79].

Insulin R

IRS1/2

PI3K
mTOR rictor
LKB1 PIP3 PTEN
AMP FOXO
Akt
GSK3
metformin AMPK TSC2
TSC1
14

Rheb
-3
-3

-3
-3
14

raptor PRAS40
rapamycin mTOR

4EBP1 S6K
Srebp1
EIF4E S6

Figure 38.4  AMPK regulates the mechanistic target of rapamycin (mTOR) pathway to control cell growth and insulin sensitivity. AMPK directly
phosphorylates the TSC2 (tuberous sclerosis 2) tumor suppressor and the key mTOR scaffold subunit raptor to inhibit the activity of the mTOR‐rap-
tor kinase complex (mTORC1) toward its downstream substrates 4EBP1 and S6K1. PI3K‐activity activates mTORC1 by Akt‐dependent phospho-
rylation of TSC2 and the mTORC1 inhibitor PRAS40. Under conditions of nutrient excess, overstimulated mTORC1 and its substrate S6K both
directly phosphorylate the insulin receptor substrate 1 and 2 proteins, resulting in their degradation. Thus too much mTORC1 activity attenuates
insulin signaling, leading to cellular insulin resistance. AMPK reverses this resistance by inactivating mTORC1 and can also directly phosphorylate
IRS1 itself. mTORC1 has also been recently shown to play a role in the control of lipogenesis through regulation of SREBP1.
480 THE LIVER:  DOWNSTREAM TARGETS I: REGULATION OF ACUTE METABOLIC RESPONSE – ENZYME IN LIPOGENESIS
Beyond effects on cell growth, mTOR also has effects on
lipid metabolism that may be particularly important in liver.
One key regulator of lipogenesis is the aforementioned SREBP‐1
transcription factor. Recently, mTORC1 signaling was shown to
be required for nuclear accumulation of SREBP1 and the induc-
tion of SREBP1 target genes [80]. Consistent with previous
results with metformin, treatment with the AMPK activators
AICAR and 2DG, or the mTORC1 inhibitor rapamycin resulted
in suppression of nuclear SREBP1 accumulation [80]. In future
studies, it will be important to define how much of the lipid‐
reducing effects of AMPK are due to direct phosphorylation of
lipogenic enzymes such as acetyl‐CoA carboxylase (ACC), and
how much are due to effects on SREBP‐1 or Chrebp‐dependent
transcription through effects of AMPK on mTORC1, versus
direct phosphorylation of Srebp‐1 and Chrebp by AMPK.
One final aspect of liver physiology that may be under control of
the AMPK–mTOR axis is insulin signaling. A major route by
which excess nutrients downregulate insulin signaling leading to
cellular insulin resistance is through hyperactivation of the
mTORC1 complex. Excess glucose leads to hyperactivation of
mTOR via suppression of AMPK, and excess fats and excess
amino acids also act to hyperactivate mTORC1 [81]. The mTOR/
Raptor complex, along with its key downstream substrate S6K,
have been shown to directly phosphorylate the insulin receptor sub-
strate‐1 and 2 (IRS1 and IRS2), leading to their proteasome degra-
dation. The same is observed under conditions of hyperinsulinemia,
as insulin signaling itself also leads to increases in mTORC1 activ-
ity as described earlier. The net effect is a negative feedback loop
whereby too much mTOR/raptor activity leads to hyperphospho-
rylation of IRS1/IRS2 and suppression of PI3‐kinase and Akt sign-
aling downstream of the insulin receptor [81]. This nutrient induced
hyperactivation of mTORC1 and consequent downregulation of
Akt signaling is observed in a cultured cell systems and also in
peripheral tissues of mice on a high fat diet. Illustrating its impor-
tance downstream of mTORC1 in the IRS1/IRS2 inhibition, this
effect is lost in peripheral tissues from an S6K1−/− mouse [82].
As one of the key direct substrates of AMPK is raptor and
TSC2, AMPK activation leads to a inhibition of mTORC1 and
its phosphorylation of IRS1 Ser636/639 and negative feedback
loop on PI3‐kinase/Akt signaling. Exogenous LKB1 expression
in HEK293 cells and metformin treatment of human hepatocel-
lular carcinoma HepG2 cells can suppress phosphorylation of
IRS‐1 on Ser636/639 and induce Akt phosphorylation [83]. In
addition to suppressing phosphorylation of IRS‐1 by mTORC1,
AMPK has also been shown to directly phosphorylate IRS1
itself on S789, though the functional consequence of that phos-
phorylation event on IRS function remains uncertain [83, 84].
Taken altogether, these studies demonstrate a mechanism by
which AMPK activation can promote PI3‐kinase/Akt activity
while simultaneously reducing mTORC1 activity. This provides
cells with a negative feedback switch that integrates upstream
signals from both nutrients and growth factors and allows the
cells to maintain energy homeostasis and insulin sensitivity.
Downstream targets IV: autophagy
and mitophagy
Autophagy is a cellular process in which proteins, organelles,
and other macromolecules are delivered to the lysosomes for
degradation. It is a process used by cells both for normal turno-
ver and for the generation of nutrients in response to energy
shortages. AMPK potently promotes autophagy through several
mechanisms. AMPK phosphorylates and activates ULK1
(unc‐51‐like autophagy‐activating kinase 1), which triggers the
initiation of the autophagic cascade [85–87]. Importantly,
mTOR strongly suppresses autophagy, in part by directly phos-
phorylating and inhibiting ULK1 [86]. Accordingly, AMPK
promotes autophagy not only by direct activation of ULK1 but
also by negatively regulating mTORC1 and blocking its inhibi-
tory effect on ULK1. Thus, ULK1 is yet another node at which
AMPK and mTOR regulate a specific metabolic process in
opposing fashion. AMPK also stimulates autophagy initiation
by differential regulation of VPS34 (vacuolar protein sorting
34) containing complexes [88], which are important for the ini-
tiation and formation of autophagosomes. AMPK was reported
to directly phosphorylate and inhibit VPS34 in non‐autophagic
complexes that do not contain autophagy adaptor proteins,
while enhancing VPS34 activity in pro‐autophagic complexes
that contain Beclin‐1 by directly phosphorylating Beclin‐1 [88].
In this way, AMPK presumably suppresses nonessential vesicle
trafficking in favor of membrane trafficking into the autophagy
pathway during nutrient starvation. Given that both AMPK and
ULK1 have been reported to directly phosphorylate distinct
sites in both Beclin‐1 and Vps34, much remains to be clarified
about the temporal and spatial control of autophagy initiation in
response to different stresses. In addition, AMPK and ULK1
have also both been reported to phosphorylate and control the
localization of Atg9, a transmembrane protein involved in early
autophagosome formation [87, 89, 90].
AMPK has also recently been shown to promote autophagy
through transcriptional mechanisms, via regulation of Tfeb
(transcription factor EB), a master transcriptional regulator of
lysosomal genes and autophagy. Although no direct link
between AMPK and Tfeb has been reported, AMPK can acti-
vate Tfeb through inhibition of mTORC1, thus blocking the
ability of mTOR to phosphorylate and translocate Tfeb out of
the nucleus [91]. Furthermore, through phosphorylation and
activation of the transcription factor FOXO3a (Forkhead box
O3) [92], AMPK has been reported to increase the levels of
CARM1 (coactivator‐associated arginine methyltransferase 1),
an important cofactor for Tfeb transcription [93].
In addition to general autophagy, several lines of evidence
indicate that AMPK promotes mitophagy, the process of degra-
dation of defective mitochondria. Indeed, activation of ULK1
by AMPK was shown to be required for proper removal of dam-
aged mitochondria via mitophagy, though the details of how
ULK1 regulates mitophagy are not fully resolved [85]. A neces-
sary step preceding removal of damaged mitochondria is the
fragmentation of mitochondria in response to mitochondrial
insults, in order to separate and target damaged mitochondrial
fragments to turnover via the mitophagy pathway. This highly
conserved process is known as mitochondrial fission. Recently,
a novel mechanism was elucidated by which AMPK promotes
mitochondrial fission [94]. In this study, AMPK was demon-
strated to induce mitochondrial fission during energy stress
through direct phosphorylation of MFF (mitochondrial fission
factor), which then serves as a receptor for DRP1 (dynamin‐
related protein 1), the enzyme that catalyzes mitochondrial
 38:  CENTRAL REGULATOR OF GLUCOSE AND LIPID METABOLISM AND TARGET OF TYPE 2 DIABETES THERAPEUTICS 481
fission [94]. Once at the mitochondria, DRP1 splits damaged
mitochondria into smaller fragments that are presumably more
efficiently cleared by autophagosomes. Furthermore, AMPK
activates PGC1α (peroxisome proliferator‐activated receptor
gamma, coactivator 1α), a master regulator of mitochondrial
biogenesis, reportedly via direct phosphorylation of PGC1α
[95] but also by promoting NAD+‐dependent activation of
PGC1α by Sirt1 (sirtuin 1) [96]. Interestingly, Tfeb, similar to
its family member Tfe3, was recently reported to drive mito-
chondrial biogenesis as well [97, 98], which offers the possibil-
ity that activation of Tfeb, or Tfe3, might be yet another
mechanism by which AMPK can promote the regeneration of
mitochondria. In all, AMPK coordinates mitochondrial fission
and mitophagy in the acute response to mitochondrial insults,
and after sustained energy stress, AMPK promotes transcrip-
tional induction of mitochondrial biogenesis. In this fashion,
AMPK serves as a central mediator of mitochondrial quality,
ensuring metabolic efficiency in cells and tissues.
Downstream targets V: polarization
of hepatocytes
The predominant upstream activating kinase for AMPK, LKB1,
is a well‐conserved key regulator of cell polarity and trafficking,
and metabolism, due at least in part to its ability to phosphoryl-
ate and activate the MARK/Par1 subfamily of AMPK related
kinases [37]. In an intestinal cell line in culture, genetic activa-
tion of LKB1 induced apical–basolateral polarity in the absence
of cell–cell or cell–matrix cues [99], though evidence of which
of the 14 AMPK‐related kinase family were involved has not
been determined. Nonetheless, AMPK itself has been connected
to cell polarization, particularly in hepatocytes and its activation
enhanced bile canalicular formation, whereas inhibition resulted
in loss of polarity and mislocalization of apical transporters
[100–102]. In MDCK cells, AMPK regulates tight junction
assembly and disassembly in response to calcium depletion
[103–104]. LKB1‐AMPK activation phosphorylates the core
tight junction protein cingulin on Ser137 [105] while at the
same time stabilizing existing cell junctions to maintain cell
polarity through phosphorylation of Gα‐interacting vesicle‐
associated protein (GIV/Girdin) on Ser245 [106]. Interestingly,
GIV/Girdin was previously reported as an Akt substrate [107],
reinforcing the idea that AMPK and mTORC1/Akt converge on
a common small set of effector proteins (ULK1, MFF, Srebp1,
Foxo, Girdin) central to metabolism and growth [108].
Additional studies are needed to delineate the roles of AMPK
and mTORC1/Akt in hepatocyte polarity and liver zonation in
vivo.
Therapeutics and future perspectives
AMPK has received a lot of attention as a potential target for
treating diseases associated with metabolic perturbation. This
includes diabetes, obesity, and fatty liver diseases and also can-
cer, which is often associated with changes in metabolism. The
type 2 diabetes drug metformin had been used for decades when
a study showed that its mechanism of action involved the activa-
tion of AMPK in hepatocytes [21]. Mechanistically, metformin
induces energy stress through inhibition of complex I of the
respiratory chain in mitochondria [109]. This leads to a change in
the ATP‐to‐AMP ratio and canonical AMPK activation. There
has been extensive research to evaluate the contribution of
AMPK to the effect of metformin on circulating glucose and
lipids. AMPK phosphorylation of ACCs has been proposed as a
master contributor to the changes in lipid synthesis that are
induced by metformin, which in turn modulates insulin sensitiv-
ity and glucose uptake in muscle. One key piece of evidence in
favor of this hypothesis came from the generation of a mouse
knock‐in mutant lacking the AMPK site on both ACC1 and
ACC2 [110]. This mouse model revealed that these phosphoryla-
tion events mediate the insulin‐sensitizing effect of metformin,
thus establishing AMPK as a relevant target in the action of met-
formin. Despite controversy [111], AMPK is widely viewed as
an essential component of the action of metformin at physiologi-
cal concentrations [78, 112]. Given the role of the AMPK–ACC
pathway in regulating fatty acid synthesis, AMPK activation is
also an attractive treatment option for conditions associated with
increased fatty acid production, such as non‐alcoholic fatty liver
disease (NAFLD) [113]. In addition to diabetes, retrospective
studies have revealed that patients taking metformin had a
decreased occurrence of cancer [114]. However, whether direct
activation of AMPK would be sufficient to replicate the benefi-
cial effect of metformin was not known until recent advances in
the generation of potent and specific small‐molecule activators
of AMPK. Most notably, two recent studies have revealed that
direct AMPK activation indeed improves symptoms of type 2
diabetes in multiple animal models. One study revealed that a
pan‐specific AMPK activator improved diabetes symptoms in
several animal models, including rodents and monkeys [115]. In
another study, a direct AMPK activator increased glucose uptake
in muscle and reduced blood glucose in type 2 diabetes models
[116]. Interestingly, inactivation of AMPK in the liver had no
effect on the efficacy of the drug, whereas muscle‐specific dele-
tion of AMPK abolished the effect of the drug, establishing mus-
cle AMPK as a key therapeutic target in type 2 diabetes [116].
Although hepatic AMPK may not be as central as originally
hypothesized for metformin effects on glucose homeostasis in
type 2 diabetes, hepatic AMPK is still a very attractive target in
the emerging field of therapeutics for NASH and NAFLD [113].
Indeed in every system from C. elegans to mammals, all studies
concur that AMPK activation suppresses de novo lipogenesis and
lipid accumulation. While much has been learned, there is still
even more yet to be learned about this ancient energy sensor and
central metabolic regulator.
REFERENCES
1. Shaw, R.J., Kosmatka, M., Bardeesy, N. et al. The tumor suppressor LKB1
kinase directly activates AMP‐activated kinase and regulates apoptosis in
response to energy stress. Proc Natl Acad Sci USA, 2004;101:3329–35.
2. Kahn, B.B., Alquier, T., Carling, D., and Hardie, D.G. AMP‐activated protein
kinase: ancient energy gauge provides clues to modern understanding of
metabolism. Cell Metab, 2005;1:15–25.
3. Carlson, C.A. and Kim, K.H. Regulation of hepatic acetyl coenzyme A car-
bozylase by phosphorylation and dephosphorylation. J Biol Chem,
1973;248:378–80.
4. Mitchelhill, K.I. et al. Mammalian AMP‐activated protein kinase shares struc-
tural and functional homology with the catalytic domain of yeast Snf1 protein
kinase. J Biol Chem, 1994;269:2361–4.
482 THE LIVER:  REFERENCES
5. Woods, A., Munday, M.R., Scott, J., Yang, X., Carlson, M., and Carling, D. Yeast
SNF1 is functionally related to mammalian AMP‐activated protein kinase and
regulates acetyl‐CoA carboxylase in vivo. J Biol Chem, 1994;269:19509–15.
6. Sanders, M.J., Grondin, P.O., Hegarty, B.D., Snowden, M.A., and Carling,
D. Investigating the mechanism for AMP activation of the AMP‐activated
protein kinase cascade. Biochem J, 2007;403:139–48.
7. Cheung, P.C., Salt, I.P., Davies, S.P., Hardie, D.G., and Carling, D.
Characterization of AMP‐activated protein kinase gamma‐subunit isoforms
and their role in AMP binding. Biochem J, 2000;346(3):659–69.
8. Amodeo, G.A., Rudolph, M.J., and Tong, L. Crystal structure of the hetero-
trimer core of Saccharomyces cerevisiae AMPK homologue SNF1. Nature,
2007;449:492–5.
9. Xiao, B., Heath, R., Saiu, P. et al. Structural basis for AMP binding to mam-
malian AMP‐activated protein kinase. Nature, 2007;449:496–500.
10. Calabrese, M.F., Rajamohan, F., Harris, M.S. et al. Structural basis for AMPK
activation: natural and synthetic ligands regulate kinase activity from opposite
poles by different molecular mechanisms. Fold Des, 2014;22:1161–72.
11. Chen, L., Wang, J., Zhang, Y.‐Y. et al. AMP‐activated protein kinase under-
goes nucleotide‐dependent conformational changes. Nat Struct Mol Biol,
2012;19:716–8.
12. Chen, L., Xin, F.‐J., Wang, J. et al. Conserved regulatory elements in AMPK.
Nature, 2013;498:E8–10.
13. Li, X., Wang, L., Zhou, X.E. et al. Structural basis of AMPK regulation by
adenine nucleotides and glycogen. Cell Res, 2015.25:50–66.
14. Xiao, B., Sanders, M.J., Underwood, E. et  al. Structure of mammalian
AMPK and its regulation by ADP. Nature, 2011;472:230–3.
15. Xiao, B., Sanders, M.J., Carmena, D. et al. Structural basis of AMPK regula-
tion by small molecule activators. Nat Commun, 2013;4:3017.
16. Xin, F.‐J., Wang, J., Zhao, R.‐Q., Wang, Z.‐X., and Wu, J.‐W. Coordinated
regulation of AMPK activity by multiple elements in the α‐subunit. Cell Res,
2013;23:1237–40.
17. McBride, A., Ghilagaber, S., Nikolaev, A., and Hardie, D.G. The glycogen‐
binding domain on the AMPKβ subunit allows the kinase to act as a glycogen
sensor. Cell Metab, 2009;9:23–34.
18. Yamauchi, T., Kamon, J., Minokoshi, Y. et al. Adiponectin stimulates glu-
cose utilization and fatty-acid oxidation by activating AMP-activated protein
kinase. Nat Med, 2002;8:1288–95.
19. Hardie, D.G. AMP‐activated protein kinase as a drug target. Annu Rev
Pharmacol Toxicol, 2007;47:185–210.
20. Sakamoto, K., McCarthy, A., Smith, D. et al. Deficiency of LKB1 in skeletal
muscle prevents AMPK activation and glucose uptake during contraction.
EMBO J, 2005;24:1810–20.
21. Zhou, G., Myers, R., Li, Y. et al. Role of AMP‐activated protein kinase in
mechanism of metformin action. J Clin Invest, 2001;108:1167–74.
22. Fryer, L.G., Parbu‐Patel, A., and Carling, D. The anti‐diabetic drugs rosigli-
tazone and metformin stimulate AMP‐activated protein kinase through dis-
tinct signaling pathways. J Biol Chem, 2002;277:25226–32.
23. Saha, A.K., Avilucea, P.R., Ye, J.M., Assifi, M.M., Kraegen, E.W., and
Ruderman, N.B. Pioglitazone treatment activates AMP‐activated protein
kinase in rat liver and adipose tissue in vivo. Biochem Biophys Res Commun,
2004;314:580–5.
24. Witters, L.A. The blooming of the French lilac. J Clin Invest,
2001;108:1105–7.
25. Tan, M.J., Ye, J.M., Turner, N. et al. Antidiabetic activities of triterpenoids
isolated from bitter melon associated with activation of the AMPK pathway.
Chem Biol, 2008;15:263–73.
26. Baur, J.A., Pearson, K.J., Price, N.L. et al. Resveratrol improves health and
survival of mice on a high‐calorie diet. Nature, 2006;444:337–42.
27. Ingebritsen, T.S., Lee, H.S., Parker, R.A., and Gibson, D.M. Reversible modula-
tion of the activities of both liver microsomal hydroxymethylglutaryl coenzyme
A reductase and its inactivating enzyme. Evidence for regulation by phospho-
rylation‐dephosphorylation. Biochem Biophys Res Commun, 1978;81:1268–77.
28. Hawley, S.A., Davison, M., Woods, A. et al. Characterization of the AMP‐
activated protein kinase from rat liver and identification of threonine 172 as
the major site at which it phosphorylates AMP‐activated protein kinase. J
Biol Chem, 1996;271:27879–87.
29. Hong, S.P., Leiper, F.C., Woods, A., Carling, D., and Carlson, M. Activation
of yeast Snf1 and mammalian AMP‐activated protein kinase by upstream
kinases. Proc Natl Acad Sci USA, 2003;100:8839–43.
30. Sutherland, C.M., Hawley, S.A., McCartney, R.R. et  al. Elm1p is one of
three upstream kinases for the Saccharomyces cerevisiae SNF1 complex.
Curr Biol, 2003;13:1299–305.
31. Hawley, S.A., Boudeau, J., Reid, J.L. et al. Complexes between the LKB1
tumor suppressor, STRADalpha/beta and MO25alpha/beta are upstream
kinases in the AMP‐activated protein kinase cascade. J Biol, 2003;2:28.
32. Woods, A., Johnstone, S.R., Dickerson, K. et al. LKB1 is the upstream kinase
in the AMP‐activated protein kinase cascade. Curr Biol, 2003;13:2004–8.
33. Hawley, S.A., Pan, D.A., Mustard, K.J. et al. Calmodulin‐dependent protein
kinase kinase‐beta is an alternative upstream kinase for AMP‐activated pro-
tein kinase. Cell Metab, 2005;2:9–19.
34. Hurley, R.L., Anderson, K.A., Franzone, J.M., Kemp, B.E., Means, A.R.,
and Witters, L.A. The Ca2+/calmodulin‐dependent protein kinase kinases
are AMP‐activated protein kinase kinases. J Biol Chem, 2005;280:29060–6.
35. Woods, A., Dickerson, K., Heath, R. et al. C(Ca2+)/calmodulin‐dependent
protein kinase kinase‐beta acts upstream of AMP‐activated protein kinase in
mammalian cells. Cell Metab, 2005;2:21–33.
36. Hemminki, A., Markie, D., Tomlinson, I. et  al. A serine/threonine kinase
gene defective in Peutz‐Jeghers syndrome. Nature, 1998;391:184–7.
37. Shaw, R.J., Lamia, K.A., Vasquez, D. et al. The kinase LKB1 mediates glu-
cose homeostasis in liver and therapeutic effects of metformin. Science,
2005;310:1642–6.
38. Lizcano, J.M., Goransson, O., Toth, R. et al. LKB1 is a master kinase that
activates 13 kinases of the AMPK subfamily, including MARK/PAR‐1.
EMBO J, 2004;23:833–43.
39. Al‐Hakim, A.K., Goransson, O., Deak, M. et  al. 14‐3‐3 cooperates with
LKB1 to regulate the activity and localization of QSK and SIK. J Cell Sci,
2005;118:5661–73.
40. Anderson, K.A., Ribar, T.J., Lin, F. et al. Hypothalamic CaMKK2 contrib-
utes to the regulation of energy balance. Cell Metab, 2008;7:377–88.
41. Carling, D., Zammit, V.A., and Hardie, D.G. A common bicyclic protein
kinase cascade inactivates the regulatory enzymes of fatty acid and choles-
terol biosynthesis. FEBS Lett, 1987;223:217–22.
42. Saggerson, D. Malonyl‐CoA a key signaling molecule in mammalian cells.
Annu Rev Nutr, 2008;28:253–72.
43. Zang, M., Xu, S., Maitland‐Toolan, K.A. et al. Polyphenols stimulate AMP‐
activated protein kinase, lower lipids, and inhibit accelerated atherosclerosis
in diabetic LDL receptor‐deficient mice. Diabetes, 2006;55:2180–91.
44. Sato, R., Goldstein, J.L., and Brown, M.S. Replacement of serine‐871 of
hamster 3‐hydroxy‐3‐methylglutaryl‐CoA reductase prevents phosphoryla-
tion by AMP‐activated kinase and blocks inhibition of sterol synthesis
induced by ATP depletion. Proc Natl Acad Sci USA, 1993;90:9261–5.
45. Yang, W., Hong, Y.H., Shen, X.Q., Frankowski, C., Camp, H.S., and Leff, T.
Regulation of transcription by AMP‐activated protein kinase: phosphoryla-
tion of p300 blocks its interaction with nuclear receptors. J Biol Chem,
2001;276:38341–4.
46. Koo, S.H. et al. The CREB co‐activator TORC2 is a key regulator of fasting
glucose metabolism. Nature, 2005;437:1109–11.
47. Liu, Y., Dentin, R., Chen, D. et al. A fasting inducible switch modulates glu-
coneogenesis via activator/coactivator exchange. Nature, 2008;456:269–73.
48. Bittinger, M.A., McWhinnie, E., Meltzer, J. et  al. Activation of cAMP
response element‐mediated gene expression by regulated nuclear transport
of TORC proteins. Curr Biol, 2004;14:2156–61.
49. Screaton, R.A., Conkright, M.D., Katoh, Y. et  al. The CREB coactivator
TORC2 functions as a calcium‐ and cAMP‐sensitive coincidence detector.
Cell, 2004;119:61–74.
50. Andreelli, F., Foretz, M., Knauf, C. et al. Liver adenosine monophosphate‐
activated kinase‐alpha2 catalytic subunit is a key target for the control of
hepatic glucose production by adiponectin and leptin but not insulin.
Endocrinology, 2006 147, 243241.
51. Leclerc, I., Lenzner, C., Gourdon, L., Vaulont, S., Kahn, A., and Viollet, B.
Hepatocyte nuclear factor‐4a involved in type 1 maturity‐onset diabetes of the
young is a novel target of AMP‐activated protein kinase. Diabetes, 2001;50.
52. Hong, Y.H., Varanasi, U.S., Yang, W., and Leff, T. AMP‐activated protein
kinase regulates HNF4alpha transcriptional activity by inhibiting dimer for-
mation and decreasing protein stability. J Biol Chem, 2003;278:27495–501.
53. Doria, A., Patti, M.E., and Kahn, C.R The emerging genetic architecture of
type 2 diabetes. Cell Metab, 2008;8:186–200.
54. Hayhurst, G. P., Lee, Y.H., Lambert, G., Ward, J.M., and Gonzalez, F.J.
Hepatocyte nuclear factor 4a (nuclear receptor 2a1) is essential for mainte-
nance of hepatic gene expression and lipid homeostasis. Mol Cell Biol,
2001:1393–1403.
55. Inoue, E. and Yamauchi, J. AMP‐activated protein kinase regulates PEPCK
gene expression by direct phosphorylation of a novel zinc finger transcrip-
tion factor. Biochem Biophys Res Commun, 2006;351:793–9.
 38:  CENTRAL REGULATOR OF GLUCOSE AND LIPID METABOLISM AND TARGET OF TYPE 2 DIABETES THERAPEUTICS 483
56. Kawaguchi, T., Osatomi, K., Yamashita, H., Kabashima T., and Uyeda, K.
Mechanism of fatty acid “sparing” effect on glucose‐induced transcription:
regulation of carbohydrate‐responsive element‐binding protein by AMP‐
activated protein kinase. J Biol Chem, 2002;277:3829–35.
57. Foretz, M., Ancellin, N., Andreelli, F. et al. Short‐term overexpression of a
constitutively active form of AMP‐activated protein kinase in the liver leads
to mild hypoglycemia and fatty liver. Diabetes, 2005;54:1331–9.
58. Shimomura, I. et  al. Decreased IRS‐2 and increased SREBP‐1c lead to
mixed insulin resistance and sensitivity in livers of lipodystrophic and ob/ob
mice. Mol Cell, 2000;6:77–86.
59. Kakuma, T. et al. Leptin, troglitazone, and the expression of sterol regulatory
element binding proteins in liver and pancreatic islets. Proc Natl Acad Sci
USA, 2000;97:8536–41.
60. Zang, M., Zuccollo, A., Hou, X. et  al. AMP‐activated protein kinase is
required for the lipid‐lowering effect of metformin in insulin‐resistant human
HepG2 cells. J Biol Chem, 2004;279:47898–905.
61. Li, Y., Xu, S., Mihaylova, M. et  al. AMPK phosphorylates and inhibits
SREBP activity to attenuate hepatic steatosis and atherosclerosis in diet‐
induced insulin resistant mice. Cell Metab, 2011;13:376–88.
62. McGee, S.L., van Denderen, B.J., Howlett, K.F. et al. AMP‐activated protein
kinase regulates GLUT4 transcription by phosphorylating histone deacety-
lase 5. Diabetes, 2008;57:860–7.
63. Chang, S., Bezprozvannaya, S., Li, S., and Olson, E.N. An expression screen
reveals modulators of class II histone deacetylase phosphorylation. Proc
Natl Acad Sci USA, 2005;102:8120–5.
64. Mihaylova, M.M., Vasquez, D.S., Ravnskjaer, K. et al. Class IIa histone dea-
cetylases are hormone‐activated regulators of FOXO and mammalian glu-
cose homeostasis. Cell, 2011;145:607–21.
65. Wullschleger, S., Loewith, R., and Hall, M.N. TOR signaling in growth and
metabolism. Cell, 2006;124:471–84.
66. Guertin, D.A. and Sabatini, D.M. Defining the role of mTOR in cancer.
Cancer Cell, 2007;12:9–22.
67. Holz, M.K., Ballif, B.A., Gygi, S.P., and Blenis, J. mTOR and S6K1 mediate
assembly of the translation preinitiation complex through dynamic protein
interchange and ordered phosphorylation events. Cell 2005;123:569–80.
68. Sancak, Y., Thoreen, C.C., Peterson, T.R. et al. PRAS40 is an insulin‐regu-
lated inhibitor of the mTORC1 protein kinase. Mol Cell, 2007;25:903–15.
69. Shaw, R.J. and Cantley, L.C. Ras, PI(3)K and mTOR signalling controls
tumour cell growth. Nature, 2006;441:424–30.
70. Vander Haar, E., Lee, S.I., Bandhakavi, S., Griffin, T.J., and Kim, D.H.
Insulin signalling to mTOR mediated by the Akt/PKB substrate PRAS40.
Nat Cell Biol, 2007;9:316–23.
71. Inoki, K., Zhu, T., and Guan, K.L. TSC2 mediates cellular energy response
to control cell growth and survival. Cell, 2003;115:577–90.
72. Corradetti, M.N., Inoki, K., Bardeesy, N., DePinho, R.A., and Guan, K.L
Regulation of the TSC pathway by LKB1: evidence of a molecular link
between tuberous sclerosis complex and Peutz–Jeghers syndrome. Genes
Dev, 2004;18:1533–8.
73. Shaw, R.J., Bardeesy, N., Manning, B.D. et al. The LKB1 tumor suppressor
negatively regulates mTOR signaling. Cancer Cell, 2004;6:91–9.
74. Liu, L., Cash, T.P., Jones, R.G., Keith, B., Thompson, C.B., and Simon, M.C.
Hypoxia‐induced energy stress regulates mRNA translation and cell growth.
Mol Cell, 2006;21:521–31.
75. Gwinn, D.M., Shackelford, D.B., Egan, D.F. et al. AMPK phosphorylation
of raptor mediates a metabolic checkpoint. Mol Cell, 2008;30:214–26.
76. Kimball, S.R., Siegfried, B.A., and Jefferson, L.S. Glucagon represses sign-
aling through the mammalian target of rapamycin in rat liver by activating
AMP‐activated protein kinase. J Biol Chem, 2004;279:54103–9.
77. Wang, C., Mao, X., Wang, L. et al. Adiponectin sensitizes insulin signaling
by reducing p70 S6 kinase‐mediated serine phosphorylation of IRS‐1. J Biol
Chem, 2007;282:7991–6.
78. Howell, J.J., Hellberg, K., Turner, M. et  al. Metformin inhibits hepatic
mTORC1 signaling via dose‐dependent mechanisms involving AMPK and
the TSC complex. Cell Metab, 2017;25:463–71.
79. Kimball, S.R. and Jefferson, L.S. Induction of REDD1 gene expression in
the liver in response to endoplasmic reticulum stress is mediated through a
PERK eIF2α phosphorylation, ATF4‐dependent cascade. Biochem Biophys
Res Commun, 2012;427:485–9.
80. Porstmann, T., Santos, C.R., Griffiths, B. et al. SREBP activity is regulated
by mTORC1 and contributes to Akt‐dependent cell growth. Cell Metab,
2008;8:224–36.
81. Manning, B.D. Balancing Akt with S6K: implications for both metabolic
diseases and tumorigenesis. J Cell Biol, 2004;167:399–403.
82. Um, S.H., Frigerio, F., Watanabe, M. et  al. Absence of S6K1 protects
against age‐ and diet‐induced obesity while enhancing insulin sensitivity.
Nature, 2004;431:200–5.
83. Tzatsos, A. and Kandror, K.V. Nutrients suppress phosphatidylinositol 3‐
kinase/Akt signaling via raptor‐dependent mTOR‐mediated insulin recep-
tor substrate 1 phosphorylation. Mol Cell Biol, 2006;26:63–76.
84. Jakobsen, S.N., Hardie, D.G., Morrice, N., and Tornqvist, H.E 5’‐AMP‐
activated protein kinase phosphorylates IRS‐1 on Ser‐789 in mouse C2C12
myotubes in response to 5‐aminoimidazole‐4‐carboxamide riboside. J Biol
Chem, 2001;276:46912–6.
85. Egan, D.F., Shackelford, D.B., Mihaylova, M.M. et al. Phosphorylation of
ULK1 (hATG1) by AMP‐activated protein kinase connects energy sensing
to mitophagy. Science, 2011;331:456–61.
86. Kim, J., Kundu, M., Viollet, B., and Guan, K.‐L. AMPK and mTOR regu-
late autophagy through direct phosphorylation of Ulk1. Nat Cell Biol,
2011;13:132–41.
87. Mack, H.I.D., Zheng, B., Asara, J.M., and Thomas, S.M. AMPK‐dependent
phosphorylation of ULK1 regulates ATG9 localization. Autophagy,
2012;8:1197–1214.
88. Kim, J., Kim, Y.C., Fang, C. et al. Differential regulation of distinct Vps34 com-
plexes by AMPK in nutrient stress and autophagy. Cell, 2013;152:290–303.
89. Weerasekara, V.K., Panek, D.J., Broadbent, D.G. et  al. Metabolic‐stress‐
induced rearrangement of the 14–3‐3ζ interactome promotes autophagy via
a ULK1‐ and AMPK‐regulated 14–3‐3ζ interaction with phosphorylated
Atg9. Mol Cell Biol, 2014;34:4379–88.
90. Zhou, C., Ma, K., Gao, R. et al. Regulation of mATG9 trafficking by Src‐
and ULK1‐mediated phosphorylation in basal and starvation‐induced
autophagy. Cell Res, 2017;27:184–201.
91. Young, N.P., Kamireddy, A., Van Nostrand, J.L. et al. AMPK governs line-
age specification through Tfeb‐dependent regulation of lysosomes. Genes
Dev, 2016;30:535–52.
92. Greer, E.L., Oskoui, P.R., Banko, M.R. et al. The energy sensor AMP‐acti-
vated protein kinase directly regulates the mammalian FOXO3 transcrip-
tion factor. J Biol Chem, 2007;282:30107–19.
93. Shin, H.‐J.R., Kim, H., Oh, S. et al. AMPK–SKP2–CARM1 signalling cas-
cade in transcriptional regulation of autophagy. Nature, 2016;534:553–7.
94. Toyama, E.Q., Herzig, S., Courchet, J. et al. Metabolism. AMP‐activated
protein kinase mediates mitochondrial fission in response to energy stress.
Science, 2016;351:275–81.
95. Jäger, S., Handschin, C., St‐Pierre, J., and Spiegelman, B.M. AMP‐acti-
vated protein kinase (AMPK) action in skeletal muscle via direct phospho-
rylation of PGC‐1alpha. PNAS, 2007;104:12017–22.
96. Cantó, C., Gerhart‐Hines, Z., Feige, J.N. et  al. AMPK regulates energy
expenditure by modulating NAD+ metabolism and SIRT1 activity. Nature,
2009;458:1056–60.
97. Mansueto, G., Armani, A., Viscomi, C. et al. Transcription factor EB con-
trols metabolic flexibility during exercise. Cell Metab, 2017;25:182–96.
98. Wada, S., Neinast, M., Jang, C., et al. The tumor suppressor FLCN medi-
ates an alternate mTOR pathway to regulate browning of adipose tissue.
Genes Dev, 2016;30:2551–64.
99. Baas, A.F. et al. Complete polarization of single intestinal epithelial cells
upon activation of LKB1 by STRAD. Cell, 2004;116:457–66.
100. Fu, D., Wakabayashi, Y., Ido, Y., Lippincott‐Schwartz, J., and Arias, I.M.
Regulation of bile canalicular network formation and maintenance by
AMP‐activated protein kinase and LKB1. J Cell Sci, 2010;123:3294–302.
101. Fu, D., Wakabayashi, Y., Lippincott‐Schwartz, J., and Arias, I.M. Bile acid
stimulates hepatocyte polarization through a cAMP‐Epac‐ MEK‐LKB1‐
AMPK pathway. Proc Natl Acad Sci USA, 2011;108:1403–8.
102. Homolya, L., Fu, D., Sengupta, P. et al. LKB1/AMPK and PKA control
ABCB11 Trafficking and polarization in hepatocytes. PLoS One, 2014:9.
103. Zhang, L., Li, J., Young, L.H., and Caplan, M.J. AMP‐activated protein
kinase regulates the assembly of epithelial tight junctions. Proc Natl Acad
Sci USA, 2006;103:17272–7.
104. Zheng B. and Cantley, L.C. Regulation of epithelial tight junction assembly
and disassembly by AMP‐activated protein kinase. Proc Natl Acad Sci
USA, 2007;104:819–22.
105. Yano, T., Matsui, T., Tamura, A., Uji, M., and Tsukita, S. The association of
microtubules with tight junctions is promoted by cingulin phosphorylation
by AMPK. J Cell Biol, 2013;203:605–14.
484 THE LIVER:  REFERENCES
106. Aznar, N., Patel, A., Rohena, C.C. et al. AMP‐activated protein kinase forti-
fies epithelial tight junctions during energetic stress via its effector GIV/
Girdin. Elife, 2016;5:e20795.
107. Enomoto, A., Murakami, H., Asai, N. et al. Akt/PKB regulates actin organi-
zation and cell motility via Girdin/APE. Dev Cell, 2005;9:389–402.111.
108. Mihaylova, M.M. and Shaw, R.J. The AMPK signalling pathway coordinates
cell growth, autophagy, and metabolism. Nat Cell Biol, 2011;13:1016–23.
109. Owen, M.R., Doran, E., and Halestrap, A.P. Evidence that metformin exerts
its anti‐diabetic effects through inhibition of complex 1 of the mitochon-
drial respiratory chain. Biochem J, 2000;348(3):607–14.
110. Fullerton, M.D., Galic, S., Marcinko, K. et al. Single phosphorylation sites
in Acc1 and Acc2 regulate lipid homeostasis and the insulin‐sensitizing
effects of metformin. Nat Med, 2013;19:1649–54.
111. Foretz, M. et al. Metformin inhibits hepatic gluconeogenesis in mice inde-
pendently of the LKB1/AMPK pathway via a decrease in hepatic energy
state. J Clin Invest, 2010;120:2355–69.
112. An, H. and He, L. Current understanding of metformin effect on the control
of hyperglycemia in diabetes. J Endocrinol, 2016;228:R97–106.
113. Smith, B.K. Marcinko, K., Desjardins, E.M., Lally, J.S., Ford, R.J., and
Steinberg, G.R. Treatment of nonalcoholic fatty liver disease: role of
AMPK. Am J Physiol Endocrinol Metab, 2016;311:E730–40.
114. Quinn, B.J., Kitagawa, H., Memmott, R.M., Gills, J.J., and Dennis, P.A.
Repositioning metformin for cancer prevention and treatment. Trends
Endocrinol Metab, 2013;24:469–80.
115. Myers, R.W. et al. Systemic pan‐AMPK activator MK‐8722 improves glucose
homeostasis but induces cardiac hypertrophy. Science, 2017;357:507–11.
116. Cokorinos, E.C. et al. Activation of skeletal muscle AMPK promotes glu-
cose disposal and glucose lowering in non‐human primates and mice. Cell
Metab. 2017;25:1147–59.
 Insulin‐Mediated PI3K
39 and AKT Signaling
Hyokjoon Kwon1 and Jeffrey E. Pessin2,3
1
Department of Medicine, Division of Endocrinology, Metabolism and Nutrition, Rutgers‐Robert Wood
Johnson Medical School, New Brunswick, NJ, USA
2
Albert Einstein‐Mount Sinai Diabetes Research Center and the Fleischer Institute for Diabetes and
Metabolism, Albert Einstein College of Medicine, Bronx, NY, USA
3
Department of Medicine and Department of Molecular Pharmacology, Albert Einstein College of Medicine,
Bronx, NY, USA

INTRODUCTION CHARACTERISTICS OF PI3K AND AKT

Obesity is a pandemic in modern society and closely linked
with diverse metabolic diseases such as cardiovascular dis­
ease, type 2 diabetes mellitus (T2D) and non‐alcoholic fatty
liver disease (NAFLD). Thus, the cost of managing obesity
and related metabolic diseases is a burden on public health
care systems in modern society. T2D is a quickly growing
global metabolic disorder characterized by impaired insulin
secretion from pancreatic β cells and insulin resistance in
peripheral tissues such as liver, muscle, and adipose tissue.
Insulin resistance diminishes insulin‐stimulated glucose
uptake in muscle and glycogen synthesis in liver and also
impairs insulin‐mediated suppression of gluconeogenesis in
liver, resulting in hyperglycemia and hyperinsulinemia. As
insulin is a crucial endocrine hormone for modulating glucose
homeostasis, the molecular mechanism of insulin signaling in
glucose homeostasis has been a central theme in biomedical
research. Insulin signaling is initiated by the activation of the
insulin receptor (IR) through autophosphorylation of the
tyrosine residue in the IR, and then many signaling molecules
including insulin receptor substrate 1 and 2 (IRS1/2), phosph­
oinositide‐3‐kinase (PI3K) and AKT/protein kinase B (PKB)
are involved to modulate downstream signaling pathways. In
this chapter, we will focus on the role of PI3K and AKT/PKB
in insulin signaling related to pathophysiology in T2D,
NAFLD, and liver cancer.
PI3K and AKT/PKB are key signaling molecules, mediating
insulin signaling in diverse tissues including liver, muscle, and
adipose tissue. In general, the intracellular responses to extracel­
lular signals are mediated by small second‐messenger molecules
such as cAMP, Ca2+, and lipid molecules that are responsible for
transducing signals between the extracellular environment and
intracellular compartments. In this pathway, PI3K activation
induces the formation of insotitol‐3,4,5‐trisphosphate (PI(3,4,5)
P3) that serves as a second‐messenger to activate AKT in the
insulin signaling pathway, resulting in glucose disposal, glyco­
gen synthesis, and suppression of lipolysis. To understand the
physiologic role of these signaling molecules in insulin action
and glucose homeostasis, in the following sections we review the
biochemical characteristics of the PI3K and AKT kinases.
Phosphoinositide 3‐kinase (PI3K)
Phosphatidylinositol has free hydroxyl groups that are
­phosphorylated by diverse series of kinases including phosphoi­
nositide 3‐kinase (PI3K) for the generation of crucial second‐
messenger in cell signaling. Early studies demonstrated that the
cleavage of phosphatidylinositol‐4,5‐bisphosphate (PI(4,5)P2)
by a membrane bound phospholipase C (PLC) generated diacylg­
lycerol (DAG) and inositol‐1,4,5‐trisphosphate. Diacylglycerol
activates protein kinase C (PKC), and inositol‐1,4,5‐trisphosphate

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
486 THE LIVER:  CHARACTERISTICS OF PI3K AND AKT

(a) Class 1A
Catalytic p110α/β/δ ABD RBD C2 Helical Kinase

Regulatory p85α/β SH3 BH N-SH2 iSH2 C-SH2

p55α/p50α N-SH2 iSH2 C-SH2

p55γ N-SH2 iSH2 C-SH2

Class 1B
Catalytic p110γ RBD C2 Helical Kinase

Regulatory p101 GβγBD

p84

(b) T308 T450 S473

AKT1 1 PH Kinase Domain HD 480

T309 T451 S474

AKT2 1 PH Kinase Domain HD 481

T305 T447 S472

AKT3 1 PH Kinase Domain HD 479

Figure 39.1  Schematic structure of PI3K and AKT/PKB. (a) Domain structure of class I PI3K catalytic and regulatory subunits. ABD, adaptor‐
binding domain; RBD, Ras‐binding domain; SH2, Src homology‐2; BH, breakpoint cluster region homology; iSH, coiled‐coil inter‐SH2.
(b) Domain structure AKT isoforms. PH, pleckstrin homology domain; HD, hydrophobic domain.

induces Ca2+ influx from intracellular calcium stores. In contrast,
PI3K mediates phosphorylation of phosphoinositides at the
D3  position to generate various 3‐phosphorylated phosphoi­
nositides such as phosphatidyl‐3,5‐bisphosphate (PI(3,5)P2) and
phosphatidyl‐3,4,5‐triphosphate (PI(3,4,5)P3) in the absence of
phospholipase‐mediated cleavage. In insulin signaling, PI3K
generates plasma membrane‐bound PI(3,4,5)P3 to activate AKT
through both the binding of the AKT pleckstrin homology (PH)
domain to PI(3,4,5)P3 and AKT site‐specific phosphorylations
by the 3‐phosphoinositide‐dependent kinase 1 (PDK1) and by the
mechanistic target of rapamycin (mTOR) in complex 2 (mTORC2).
PI3K is classified into three classes (class I, class II, and
class III) according to structure and lipid substrate preferences
[1]. Class II PI3K including PI3K‐C2α, PI3K‐C2β and PI3K‐
C2γ was discovered in mammals on the basis of sequence
homology with class I and III PI3K instead of functional con­
text. Thus, their functional role is unclear although class II
PI3K is constitutively associated with intracellular membranes.
Class II PI3K consists of two subclasses α and β and contains
a C‐terminal C2 domain as observed in PKC molecules for
phospholipid binding. Vps34 was originally identified in
Saccharomyces cerevisiae for the endosomal sorting and is the
only member of class III PI3K. Vps34 generates a constitutive
heterodimer with Vps15/p150 to locate on the intracellular
membrane, and the biological function of Vps34 in mammals
relates to the regulation of vesicle trafficking such as autophagy,
endocytosis, and phagocytosis. In mammals, class I PI3K is
present in all cell types and mediates insulin signaling. Class I
PI3K consists of a heterodimer with a catalytic subunit (110–
120 kDa) and regulatory subunit, and phosphatidylinositol‐4,
5‐bisphosphate (PI(4,5)P2) is a preferred substrate in vivo.
Catalytic subunits contain C‐terminal catalytic and phosphati­
dylinositol kinase domains, N‐terminal adaptor‐binding
domain (ABD), and Ras‐binding domain (RBD) (Figure 39.1a).
Low concentrations (nanomolar) of Wortmannin irreversibly
inhibits the catalytic subunit of class I PI3K by Schiff base for­
mation with a lysine in the C‐terminal kinase domain [2].
Regulatory subunits have C‐terminal Src homology‐2 (SH2)
domains that are separated by an inter‐SH2 (iSH2) region to
provide a binding site for the ABD of the catalytic subunit. The
SH2 domain is a module of about 100 amino acids to bind
phosphotyrosine‐containing motifs. Regulatory subunits also
have N‐terminal proline‐binding SH3 and a breakpoint cluster
region homology domain (BH). Class I PI3K catalytic p110
subunits are divided into a class IA group such as p110α,
p110β, and p110δ, which bind the regulatory p85 type subunit
 39:  Insulin‐Mediated PI3K and AKT Signaling 487
and into a class IB p110γ to interact with regulatory p101 and
p84. The p110α and p110β are expressed ubiquitously, whereas
p110γ and p110δ are expressed in immune cells. Each catalytic
subunit generates a dimer with a regulatory subunit to modulate
the catalytic activity and subcellular localization. The p110α
catalytic subunit generates a heterodimer with one of five dif­
ferent regulatory subunits including p85α, p85δ, p55α, p55γ,
and p50α. In the p85α/p110α heterodimer, the ABD of p110α
catalytic subunit interacts with coiled‐coil inter‐SH2 (iSH)
domain of the p85α regulatory subunit to maintain a stable het­
erodimer with a low kinase activity state [3]. Thus, binding of
the SH2 domain in a p85α regulatory subunit with phosphoty­
rosine residues such as phosphotyrosine in IRS1/2 releases
inhibitory interactions between p85α and p110α, resulting in
the activation of PI3K in insulin signaling [4, 5].
AKT/protein kinase B
The v‐AKT oncogene known as AKT8 was identified from
transforming retrovirus in a spontaneous thymoma of AKR
mice, and then cellular homologue serine and threonine kinase
AKT (roughly 57 kDa), also termed as protein kinase B (PKB),
was cloned and characterized [6–9]. Human AKT has three
isoforms (AKT1/PKBα, AKT2/PKBβ, AKT3/PKBγ), and
each isoform is encoded from separate human genes: AKT1/
PKBα in Chr14 q32.33, AKT2/PKBβ in Chr19 q13.2, and
AKT3/PKBγ in Chr1 q44 (Figure 39.1b). For functional stud­
ies, murine models demonstrate that each AKT isoform has
distinct and overlapping signaling functions. AKT1 deficient
mice show reduced body size, AKT2 deficiency impairs glu­
cose homeostasis, and AKT3 deficient mice show diminished
brain size. In contrast, all AKT isoforms contribute to phos­
phorylation of GSK3α and GSK3β in 3T3‐L1 adipocytes and
regulation of forkhead box O1 (FOXO1), TSC2, and GSK3β
in H157 cells.
AKT kinases belong to a class of AMP/GMP kinase and protein
kinase C (AGC) kinases. AKT has a PH domain, kinase domain,
and hydrophobic domain (HD) with PH and kinase domains con­
nected with an α‐helical linker. The PH domain found as a major
phosphorylation site by PKC in pleckstrin proteins at its N‐termi­
nus interacts with PI(3,4,5)P3 to localize AKT into the plasma
membrane and then to interact with 3‐phosphoinositide‐dependent
kinase 1 (PDK1) that also has a PH domain to be enriched in
plasma membrane. The PH domain is a 100–120 amino acids motif
with seven anti‐parallel β‐sheets forming a hydrophobic pocket that
interacts with the C‐terminal amphipathic helix. The PH domain is
a primary lipid binding site although it is also involved in protein–
protein interactions. As the PH domain has different phosphoi­
nositide binding specificity, PH domains of dynamin bind to PI(4)
P1 and PI(4,5)P2, whereas PH domains of AKT and PDK1 bind to
PI(3,4)P2 and PI(3,4,5)P3 [10, 11]. The kinase domain of AKT
shares high homology with other AGC kinase members, and the
kinase domain of three AKT kinase isoforms displays around 87%
sequence homology. HD is a characteristic feature of AGC kinases
including protein kinase A (PKA), protein kinase C (PKC), and
PDK1 and contains a F‐X‐X‐F/Y‐S/T‐Y/F motif where X is any
amino acid. Upon biosynthesis of AKT, the nascent AKT poly­
peptide is phosphorylated at the ribosome by mTORC2 at the
C‐terminal turn motif [12], enhancing the protein stability in
cytosol. Indeed, mTORC2‐induced phosphorylation of Thr‐450 in
the AKT1 turn motif localizes AKT1 to the cytosol as an inactive
conformation through the interaction between PH and kinase
domains [13]. AKT activity is regulated through the phospho­
rylation in kinase domain at Thr‐308 by PDK1 and in the hydro­
phobic domain at Ser‐473 by mTORC2, necessary for the full
induction of Akt substrate kinase activity [14].
INSULIN‐MEDIATED PI3K AND AKT
SIGNALING AND INSULIN RESISTANCE
Glucose homeostasis is maintained by delicate regulation of the
pancreatic exocrine and endocrine systems. The pancreatic exo­
crine cells composed of acinar and ductal cells secrete digestive
enzymes into the duodenum for nutrient digestion. In contrast,
the pancreatic endocrine cells in the islets of Langerhans secrete
endocrine hormones into the blood circulation to regulate nutri­
ent metabolism. Pancreatic islets include several endocrine cells
such as α cells (glucagon production), β cells (insulin produc­
tion), δ cells (somatostatin production), PP cells (pancreatic
polypeptide production), and ε cells (ghrelin production) for
specific endocrine functions. The secretion of glucagon and
insulin is tightly regulated in response to circulating glucose
levels to maintain normoglycemia during fasting and feeding,
respectively. In peripheral tissues, insulin stimulates glucose
uptake (skeletal muscle and adipose tissue) and glycogen syn­
thesis (skeletal muscle and liver) and inhibits gluconeogenesis
and glycogenolysis (liver). Insulin also increases lipogenesis in
hepatocytes and adipocytes and diminishes lipolysis in adipo­
cytes to reduce circulating free fatty acid and glycerol [15].
Thus, impaired regulation of insulin secretion and IR‐mediated
signaling pathway results in T2D.
Insulin‐mediated PI3K and AKT signaling
Identification and characterization of the IR initiated a huge
effort to understand the molecular mechanisms of insulin action
in glucose homeostasis. At the molecular level, insulin binds to
the cell surface IR to initiate intracellular signaling cascades
that ultimately results in specific cellular biological responses
[16]. The IR is a transmembrane tyrosine kinase receptor
encoded by Chr19 p13.2 in humans and has two isoforms, IR‐A
and IR‐B, generated by alternative splicing of exon11. The IR‐A
isoform excludes exon 11 and is predominantly expressed in
fetal tissues and the brain with high affinity for insulin and insu­
lin‐like growth factor 2 (IGF‐2), whereas IR‐B includes exon 11
and is highly expressed in liver. To interact with insulin, the
insulin receptor consists of two α subunits and two β subunits
that are bound by a disulfide link into an α2β2 heterotetrameric
complex. The α and β subunits are derived from a single large
precursor by one or more proteolytic cleavages. The extracellu­
lar α subunits directly bind insulin that allows for a transmem­
brane conformational change that activates the intracellular
tyrosine kinase domain of the IR β subunits, resulting in intra­
molecular transphosphorylation of the β subunit to phosphoryl­
ate its adjacent partner on specific tyrosine residues [17, 18].
Autophosphorylated tyrosine residues in the tyrosine kinase
488 THE LIVER:  INSULIN‐MEDIATED PI3K AND AKT SIGNALING AND INSULIN RESISTANCE

domain of β subunits recruit receptor substrates such as insulin
receptor substrate (IRS) that is a scaffold for organizing the
proximal signaling complex. Phosphotyrosine residues in acti­
vated IRs interact with IRS through its phosphotyrosine binding
domain (PTB) [19]. IRS also has several tyrosine residues,
which are phosphorylated by activated IR tyrosine kinase activ­
ity, for the interaction with the SH2 domain of the regulatory
subunit of PI3K [5]. IRS isoforms (IRS1‐IRS4) show differen­
tial function in physiology. IRS‐1 knockout mice have growth
retardation and insulin resistance particularly in muscle, albeit
with normal whole body glucose tolerance [20], whereas IRS‐2
knockout mice show impaired insulin action in liver and pancre­
atic β cell loss, resulting in T2D [21]. IRS1/2 phosphorylated on
specific tyrosine residues activates two major signaling path­
ways; (i) the PI3K‐AKT/PKB pathway and (ii) Ras/mitogen‐
activated protein kinase (MAPK) pathway. In addition, there are
inhibitory molecules for insulin signaling such as protein tyros­
ine phosphatase 1B (PTP1B), suppressor of cytokine signaling
(SOCS), and growth factor receptor bound protein 10 (Grb10)
that suppress insulin signaling by inducing IR dephosphoryla­
tion, physical blocking substrate phosphorylation, and degrada­
tion of the IR and/or IRS substrates. Metabolites such as DAG
also mediate the inhibition of insulin signaling (Figure 39.2).
Insulin‐mediated PI3K activation
PI3K‐AKT/PKB signaling pathway modulates most metabolic
functions of insulin. Activation of PI3K by extracellular stimu­
lation including insulin induces the activation of AKT. In the
basal state, p85α/p110α heterodimeric PI3K has interactions
between ABD of the p110α catalytic subunit and coiled‐coil
iSH domain of p85α regulatory subunit to maintain stable PI3K
with a low kinase activity state. However, insulin‐induced phos­
phorylation on tyrosine residue in IRS1/2 mediates interaction
between phosphotyrosine of IRS1/2 and the SH2 domain in
p85α regulatory subunit to release inhibitory interactions
between p85α and p110α [5, 22]. Thus, PI3K accesses the
­membrane to produce phosphatidylinositol‐3,4,5‐triphosphate
(PI(3,4,5)P3) from PI(4,5)P2 at the plasma membrane. Although
p110α or p110β knockout mice are embryonic lethal, liver spe­
cific p110β deficient mice showed little effect in insulin signal­
ing. However, liver specific p110α deficient mice have glucose
intolerance and insulin resistance, suggesting that p110α is criti­
cal to mediate insulin signaling in hepatocytes [23]. Deletion of
the p85 regulatory subunits in heterozygous deletion of p85α,
p85β, or p50α/p55α showed enhanced insulin sensitivity as
­regulatory subunits are in excess concentrations to catalytic
subunits, thereby competing with IRS for the formation of

Figure 39.2  Insulin signaling and insulin resistance. IRS1/2 phosphorylated on specific tyrosine residues activates the phosphoinositide 3‐kinase
(PI3K)‐AKT/protein kinase B (PKB) pathway and Ras/mitogen‐activated protein kinase (MAPK) pathway. PI3K‐AKT signaling pathway regulates
metabolic processes such as glycogen synthesis (muscle and liver), glucose uptake (muscle and adipocytes), protein synthesis (muscle and liver),
and gluconeogenesis (liver). Inflammatory signals such as TNF‐α, saturated free fatty acid, IL‐6, LPS, and diacylglycerol (DAG) activate inhibitory
molecules such as suppressor of cytokine signaling (SOCS) and cJun N‐terminal kinase (JNK) to suppress insulin signaling, resulting in insulin
resistance.
 39:  Insulin‐Mediated PI3K and AKT Signaling 489
p85α/p110α heterodimer [24]. PI(3,4,5)P3 generated by insulin‐
induced activation of PI3K is metabolized rapidly into PI(4,5)P2
by lipid phosphatase including tumor suppressor phosphatase
and tensin homologue (PTEN) and SH2‐containing inositol
5′‐phosphatase‐2 (SHIP2) to terminate proximal signaling [25,
26]. PTEN encoded in chromosome 10q23 was identified as a
tumor suppressor that is inactivated in diverse tumors including
endometrial, prostate, and mammary carcinomas.
Insulin‐mediated AKT activation
PI(3,4,5)P3 generated by PI3K triggers the activation of 3‐phos­
phoinositide‐dependent protein kinase 1 (PDK1) that is respon­
sible for the phosphorylation and activation of AKT. PDK1
contains two domains, an N‐terminal kinase domain and a C‐
terminal PH domain. Autophosphorylation of Ser‐241 in kinase
domain by PDK1 is required for kinase activity, and the small
PH domain binds to membrane bound PI(3,4,5)P3 and PI(3,4)P2.
As AKT also has a PH domain to interact with membrane‐bound
PI(3,4,5)P3, AKT is subsequently recruited from the cytosol to
the plasma membrane through binding of its PH domain to
PI(3,4,5)P3, resulting in a conformational change that separates
the PH and kinase domains from inactive conformation and
exposes two key regulatory residues whose phosphorylations
are required for maximal activation of AKT kinase. The acti­
vated PKD1 closely localized with AKT through PI(3,4,5)P3
binding phosphorylates Thr‐308 in the activation loop of AKT
[13]. AKT is also phosphorylated on Ser‐473 by the mTORC2,
and this dual phosphorylation results in full activation of AKT
kinase activity [14]. Deletion of mTORC2‐specific subunits
such as Rictor or Sin1 abrogates AKT phosphorylation at both
the C‐terminal turn motif (Thr‐450 of AKT1) and HD (Ser‐473
of AKT1), resulting in the impaired proximal signaling of AKT
[14, 27]. Interestingly, the AKT isoforms show differential func­
tions according to the distinctive expression profiles in tissues.
AKT1 and AKT2 are broadly expressed with AKT2 being more
closely linked to metabolic processes. AKT1 deficient mice
have growth retardation without defects in metabolism. In
contrast, AKT2 deficient mice show insulin resistance due to
interrupted insulin signaling [28].
Proximal signaling of insulin‐induced
AKT activation
Activated AKT/PKB regulates diverse insulin‐mediated meta­
bolic pathways such as glucose transport, glycogen synthesis,
gluconeogenesis, protein synthesis, and cell growth. Several
AKT substrates have been identified by the AKT consensus
motif (R‐X‐R‐X‐X‐S/T‐B) where X is any amino acid, and B
represents bulky hydrophobic amino acids [29]. AKT promotes
cell growth and proliferation through the suppression of p27 to
activate cyclin D1 along with the activation of MAPK signaling
pathway (Figure 39.2). AKT phosphorylates AKT substrate of
160 kDa (AS160) to activate the Rab family of small GTPases
that initiate the translocation of the glucose transporter 4
(GLUT4), resulting in glucose uptake in muscle and adipocytes.
AKT also suppresses glycogen synthase kinase‐3 (GSK3) via
phosphorylation on Ser‐21 or Ser‐9 to activate glycogen syn­
thase in muscle and liver [30]. AKT phosphorylates FOXO1
that induces FOXO1 association with nuclear 14‐3‐3 proteins to
exclude FOXO1 from the nucleus to the cytosol in an inactive
state. In the liver, this suppresses gluconeogenic gene expres­
sion and thereby inhibits hepatic glucose production in feeding.
AKT phosphorylates tuberous sclerosis complex 1 and 2
(TSC1/2), which releases the inhibition of Ras homologue
enriched in brain (Rheb) for the activation of mTORC1 complex
[14], that in turn enhances protein synthesis through the acti­
vation of eukaryotic translation initiation factor 4E binding
protein‐1 (4E‐BP) and p70 ribosomal protein S6 kinase 1
(p70S6K1). Furthermore, mTORC1 activation induces
SREBP1c activation for the expression of lipogenic genes
including FAS and ACC to induce lipogenesis in liver [31].
Fasting‐induced β‐adrenergic receptor signaling activates PKA
to phosphorylate perilipin1 (PLIN1) and hormone sensitive
lipase (HSL), facilitating lipolysis in adipocytes. In feeding,
however, insulin‐induced AKT activation mediates phospho­
rylation of phosphodieterase‐3B (PDE‐3B) to decrease cAMP,
resulting in the suppression of PKA activity and HSL activity
in adipocytes [32]. Thus, insulin‐induced suppression of
­adipose tissue lipolysis mediates acute suppression of gluconeo­
genesis in liver as suppressed lipolysis reduces acetyl‐CoA
­levels in liver [33, 34].
Insulin resistance in liver
The molecular mechanisms accounting for insulin resistance are
still unclear although there is substantial progress in our under­
standing of insulin signaling. Insulin resistance is the integral
result of alterations in insulin secretion in pancreatic β cells, IR
expression, ligand binding, and downstream IR signaling,
resulting in diverse metabolic disease such as T2D and cardio­
vascular and NAFLD diseases. Liver is one of major tissues for
regulating glucose homeostasis in response to pancreatic hor­
mones such as insulin and glucagon produced by pancreatic β
and α cells, respectively. Insulin‐mediated PI3K and AKT acti­
vation regulate hepatic glucose and lipid metabolism. In insulin
sensitive individuals, regular meals increase the blood glucose
that in turn releases pancreatic insulin to activate insulin signal­
ing and glucose uptake through IR and GLUT4 in muscle and
adipose tissue, respectively. In the liver, PI3K and AKT activa­
tion enhance glycogen synthesis and de novo lipogenesis (DNL)
and also suppresses gluconeogenesis through FOXO inactiva­
tion (Figure  39.3a). Thus, hepatic insulin signaling may sup­
press gluconeogenesis through transcriptional suppression of
gluconeogenic genes such as PCK1 and G6PC for long‐term
regulation. However, the ability of insulin to acutely suppress
gluconeogenesis is mostly mediated by an inhibition of adipose
tissue lipolysis [34]. First, insulin reduces hepatic glucose pro­
duction within 30 minutes but does not reduce gluconeogenic
protein levels [35]. Second, insulin suppresses adipose tissue
lipolysis to regulate hepatic gluconeogenesis as acetyl‐CoA
(Ac‐CoA) generated by adipose tissue lipolysis allosterically
activates pyruvate carboxylase (PyC) activity for hepatic gluco­
neogenesis. Thus, suppressed adipose tissue lipolysis lowers
hepatic acetyl‐CoA and glycerol contents to reduce PyC activity
and glucose production [33, 34, 36]. In contrast, in insulin
resistance status, glycogenolysis and gluconeogenesis are
enhanced to produce hepatic glucose due to dysfunctional
490 THE LIVER:  INSULIN‐MEDIATED PI3K AND AKT SIGNALING AND INSULIN RESISTANCE

(a)

(b)

Figure 39.3  Insulin resistance in liver. (a) In insulin sensitive individuals, insulin derived from pancreatic β cells activates AKT to enhance gly­
cogen synthesis, de novo lipogenesis, and suppress gluconeogenesis through FOXO1 degradation. (b) In insulin resistance, impaired insulin signal­
ing increases glycogenolysis and gluconeogenesis, and flux of free fatty acids and glycerol from adipose tissue and intestine results in glucose
production and fat accumulation in liver. DNL, de novo lipogenesis; GS, glycogen synthase; GP, glycogen phosphorylase; PyC, pyruvate carboxy­
lase; Ac‐CoA, acetyl‐CoA; FA‐CoA, fatty acyl‐CoA; TAG, triacylglycerol; CM, chylomicron; VLDL, very low density lipoprotein.

insulin signaling (Figure  39.3b). Impaired activation of PI3K
and AKT induces activation of glycogen phosphorylase (GP) to
induce glycogenolysis along with suppressed glycogen syn­
thase (GS) activity. Impaired AKT‐mediated phosphorylation of
FOXO1 also generates nuclear active FOXO1 to mediate hepatic
gluconeogenesis. Flux of free fatty acids and glycerol from adi­
pose tissue and intestine supplies substrate for glucose produc­
tion and fat accumulation in liver, resulting in hyperglycemia
and hyperlipidemia.
Hepatic insulin resistance is reproducibly correlated with
increased hepatic triglyceride content as shown in NAFLD.
High levels of DAG generated by incomplete synthesis to tri­
glyceride or the breakdown of triglyceride to DAG have been
proposed to inhibit insulin signaling through protein kinase
Cε (PKCε) activation, which phosphorylates IR at Thr‐1160
(Thr‐1150 in mouse) to suppress tyrosine kinase activity of IR
[37–41]. Thus, intervention of accumulation of triglyceride in
liver results in the reversal of hepatic insulin resistance in
humans and rodent models with NAFLD. In this regard, adipose
triglyceride lipase (ATGL) deficient mice that have reduced
ability to convert triglyceride to DAG show enhanced glucose
tolerance and insulin sensitivity [42]. More recently, an alterna­
tive model of increased ceramide levels has also been shown to
be associated with insulin resistance [43]. Decreased IRS pro­
tein levels also contribute to insulin resistance in rodents and
humans although complete molecular understanding of the
mechanisms of reduced IRS levels are still under investigation
[44]. SOCS1/3 increased by obesity‐induced inflammatory
cytokines such as TNF‐α and IL‐6 enhances the degradation of
IRS1/2 through E3 ubiquitin ligase activation, resulting in insu­
lin resistance [45, 46]. IRS phosphorylation on serine residues is
another mechanism that induces insulin resistance as the IRS
proteins contain several serine residues that are phosphorylated
by kinases such as extracellular signal regulated kinase (ERK),
cJun N‐terminal kinase (JNK), protein kinase Cζ (PKCζ), and
p70S6K [47]. The phosphorylation of IRS on Ser‐307 is a
­typical inhibitory signal to suppress insulin signaling as Ser‐307
locates in the phosphotyrosine‐binding (PTB) domain of IRS [48].
 39:  Insulin‐Mediated PI3K and AKT Signaling 491
Thus, increased TNF‐α, saturated free fatty acids, and endoplas­
mic reticulum (ER) stress in obese individuals activate JNK and
inhibitor of nuclear factor κB kinase β (IKKβ) to phosphorylate
Ser‐307 of IRS. In addition, ERK activated by insulin also phos­
phorylates IRS1 on Ser‐612 to attenuate AKT activation [49].
PI3K AND AKT IN GLUCOGENOGENESIS
AND LIPOGENESIS
The secretion of insulin and glucagon are tightly regulated to
modulate gluconeogenesis and lipogenesis in hepatocytes,
and dysregulation of the proximal signaling pathways of these
ligands results in the hyperglycemia and hyperlipidemia seen in
metabolic diseases. In fasting, glucagon and catecholamines
regulate gluconeogenesis through the activation of cAMP‐
dependent PKA. Activated PKA phosphorylates CREB‐Ser133
to induce the expression of gluconeogenic genes such as Pck1
and G6pc and also phosphorylates liver pyruvate kinase (L‐PK)
to suppress glycolysis. Glucagon decreases the level of fruc­
tose‐2,6‐bisphosphate, an allosteric activator of phosphofruc­
tokinase and an inhibitor of fructose‐1,6‐bisphosphatase to
suppress glycolysis via a phosphorylation of phosphofructoki­
nase 2 (PFK2) [50]. Glucagon‐induced activation of PKA regu­
lates gene expression mediated by the CREB‐CBP‐CRTC2
complex, and CRTC2 is a key regulator in CREB‐CBP‐CRTC2
complex regulation. Glucagon‐induced PKA activation sup­
presses constitutively active serine/threonine kinase SIK2 to
decrease CRTC2‐Thr171 phosphorylation and also phospho­
rylates inositol‐1,4,5‐triphsophate receptors (IP3R) to increase
Ca2+ influx for CRTC2‐specific phosphatase calcineurin acti­
vation, resulting in the formation of the CREB‐CBP‐CRTC2
complex in the nucleus to induce gluconeogenic target gene
expression [51, 52]. In contrast, insulin suppresses the secretion
of glucagon from pancreatic α‐cells, and phosphorylated
CRTC2 through insulin signaling interacts with 14‐3‐3 protein
to be sequestered in cytoplasm, thereby suppressing the forma­
tion of the CREB‐CBP‐CRTC2 complex [53]. Thus, insulin
regulates glucose homeostasis after feeding through the activa­
tion of PI3K and AKT to suppress gluconeogenesis and enhance
lipogenesis in hepatocytes.
PI3K and AKT‐mediated modulation
in gluconeogenesis
Gluconeogenesis in liver contributes approximately half of the
hepatic glucose production in humans during overnight fasting
and is the primary mechanism responsible for the increased
fasting glucose levels in T2D patients. Gluconeogenesis is
­ pathogenesis of T2D. Regulation of DNL is mediated by tran­
­regulated by complex mechanisms: (i) availability of gluconeo­
genic substrate such as lactate, alanine and glycerol, (ii) meta­
bolic intermediate‐induced allosteric regulation in metabolic
enzymes, and (iii) the balance of hormones such as insulin,
glucagon, and catecholamines. Glycerol and non‐esterified fatty
acids (NEFA) generated by lipolysis are involved in the regula­
tion of gluconeogenesis. Glycerol, one of major substrates
in  gluconeogenesis, is converted to glycerol‐3‐phosphate by
glycerol kinase and then dihydroxyacetone‐3‐phosphate to
participate glucose production in liver. Although acetyl‐CoA
generated by β‐oxidation of NEFA in mitochondria does not
contribute to the substrates for glucose production, acetyl‐CoA
allosterically activates pyruvate carboxylase to enhance gluco­
neogenesis [33]. Thus, insulin‐mediated suppression of gluca­
gon secretion and adipose tissue lipolysis is important in the
regulation of hepatic gluconeogenesis. Experimentally, infusion
of insulin in fasted rats decreased the concentration of glycerol
and acetyl‐CoA along with the suppression of hepatic glucose
production. However, when infusion of glycerol and acetate was
added to the infusion of insulin, hepatic glucose production was
restored as increased glycerol and acetyl‐CoA enhanced gluco­
neogenesis [33–35]. These results suggest that lipolysis is also
involved in the regulation of gluconeogenesis.
FOXO1 and transcriptional coactivator peroxisome prolif­
erator‐activated receptor‐γ coactivator‐1α (PGC‐1α) enhance
the expression of gluconeogenic genes to mediate gluconeo­
genesis in liver. To suppress gluconeogenesis in the post­
prandial state, insulin‐dependent PI3K and AKT activation
phosphorylates FOXO1 (Thr‐24, Ser‐253, and Ser‐316 for
murine FoxO1) to exclude FOXO1 from the nucleus that is
then subsequently ubiquitinated and degraded by proteasome,
resulting in the suppression of gluconeogenesis [54, 55].
However, the role of insulin in gluconeogenesis through
FOXO1‐PGC1α is still unclear as liver specific IR and FOXO1
gene double knockout mice show normal glucose homeostasis
although liver specific IR knockout mice have glucose intoler­
ance and insulin resistance [56, 57]. In addition, T2D patients
and insulin resistant high‐fat diet fed rodent models do not
show differential expression of gluconeogenic genes including
PCK1 and G6PC [58]. In contrast to the high‐fat diet, fructose
fed rodents display increased G6pc expression that is the tran­
scriptional target of both FOXO1 (primarily a gluconeogenic
gene transcriptional activator) and ChREBPβ (primarily a
lipogenic gene transcriptional regulator) [59]. Since insulin
inhibits G6pc expression through suppression of FOXO1, and
fructose is a poor inducer of insulin secretion, it is hypothe­
sized that one mechanism for enhanced gluconeogenesis by
fructose is due to an imbalance between FOXO1 and ChREBPβ
regulation.
PI3K and AKT‐mediated modulation
in lipogenesis
Glucose from excess dietary carbohydrate undergoes glycoly­
sis in liver and is eventually converted into fatty acids through
DNL and then esterified to triglyceride for very low density
lipoprotein (VLDL) secretion into blood circulation. DNL is
abnormally increased in NAFLD and closely linked to the
scriptional regulation of enzymes such as fatty acid synthase
(FAS) and acetyl‐CoA carboxylase (ACC) for fatty acid syn­
thesis and allosteric regulation of ACC. The transcriptional
regulation of critical enzymes in DNL is modulated by two
major transcriptional regulators including sterol regulatory ele­
ment binding protein 1c (SREBP1c) and carbohydrate response
element binding protein (ChREBP) that are activated by
enhanced insulin signaling and glucose concentration, respec­
tively [60, 61]. Thus, insulin‐induced PI3K and AKT activation
492 THE LIVER:  PI3K AND AKT IN LIVER DISEASES: NAFLD AND CANCER
closely regulate the DNL in liver. mTORC1 activated by AKT
suppresses Lipin1, a phosphatidic acid phosphatase, to increase
phosphorylation of nascent SREBP1c in ER, resulting in the
activation of SREBP1c to induce the expression of fatty acid
synthesis enzymes such as FAS and ACC in liver [62, 63]. In
contrast, ChREBP regulation is mediated by glucose imported
by insulin independent GLUT2. Thus, glucose metabolites
such as glucose‐6‐phosphate and fructose‐2,6‐bisphosphate
and dephosphorylation of Ser‐196 in ChREBP are proposed to
regulate ChREBP activity [64, 65]. ACC is also one of the key
regulators modulating DNL in liver as ACC converts acetyl‐
CoA to malonyl‐CoA to provide monomer for fatty acid syn­
thesis in DNL. ACC has two isoforms: ACC1 in adipose tissue,
mammary gland, and liver and ACC2 in skeletal and cardiac
muscle. ACC1 activity is regulated by several different levels.
First, ACC expression is enhanced by insulin signaling through
SREBP1c activation along with ChREBP and LXRs [61].
Second, ACC1 exists as a low activity dimer, and allosteric
activators such as citrate and glutamate stimulate polymeriza­
tion of ACC1 to generate high activity polymers to enhance
DNL [66], whereas malonyl‐CoA and fatty acyl‐CoA inhibits
ACC1 polymerization as a feedback inhibition mechanism.
Third, dephosphorylation and phosphorylation of ACC1 by
insulin and glucagon are also important for the regulation of
ACC1 activity although the molecular mechanism is still
unclear [67–69]. Furthermore, malonyl CoA, a product of
ACC, inhibits carnitine : palmitoyl‐CoA transferase‐1 (CPT1)
to modulate the entry of long‐chain fatty acyl‐CoA into mito­
chondria for β‐oxidation. Thus, ACC suppression is important
to reduce lipid stores in muscle and liver and hence to enhance
insulin sensitivity.
PI3K AND AKT IN LIVER DISEASES:
NAFLD AND CANCER
NAFLD is a common hepatic disorder characterized by fat
accumulation in the liver in the absence of excessive alcohol
intake. Steatosis, which has benign fat accumulation within the
cytoplasm of at least 5% of the hepatocytes in the liver, pro­
gresses to non‐alcoholic steatohepatitis (NASH), fibrosis, cir­
rhosis, and liver cancer. For the diagnosis of NASH, hepatocyte
ballooning and inflammation appear with steatosis. However,
NASH is not necessary for liver fibrosis, and in general NASH
and fibrosis frequently occur together. Liver fibrosis character­
ized by the deposition of excess extracellular matrix (ECM)
such as collagen by activated hepatic stellate cells progresses to
cirrhosis, which demonstrates the loss of hepatic structure, por­
tal hypertension, and the formation of regenerative nodules.
Cirrhosis is irreversible and induces liver organ failure and
hepatocellular carcinoma (HCC). NAFLD is closely linked with
hepatic and adipose tissue insulin resistance, showing that about
50% reduction in glucose disposal and impaired insulin‐medi­
ated suppression in endogenous gluconeogenesis, and approxi­
mately 80% of the T2D patients demonstrates NAFLD [70, 71].
Thus, understanding the molecular mechanisms linking these
pathophysiologic symptoms is one of major targets of current
biomedical research.
PI3K and AKT signaling in non‐alcoholic
fatty liver disease
The development of NAFLD is an early step in the pathogenesis
of T2D, and most obese T2D patients have NAFLD. As lipid
accumulation induces insulin resistance in liver, weight loss
leads to the resolution of NAFLD and normalization of fasting
glucose level [72]. Thus, hyperinsulinemia associated with
hepatic insulin resistance is a hallmark feature of NAFLD, espe­
cially in humans with T2D. Diverse tissue specific IR knockout
mice demonstrate that muscle specific IR knockout mice or
muscle/adipose tissue specific IR knockout mice have normal
glucose levels. However, liver specific IR knockout mice result
in hyperglycemia with peripheral insulin resistance and NASH,
suggesting that liver insulin resistance is critical in pathogenesis
of NAFLD [73]. Peripheral insulin resistance induces lipolysis
in adipose tissue to increase free fatty acid flux into the liver,
and liver insulin resistance induces DNL through SREBP1c
activation, resulting in fat accumulation in liver.
The liver‐specific overexpression in mice of a mutated cata­
lytic subunit p110α (Pik3ca) found in human HCC results in
severe liver steatosis in six months and liver tumors in a year
[74]. In contrast liver specific Pik3ca deficient mice show sup­
pressed liver steatosis with a high‐fat diet [75], suggesting that
PI3K is involved in NAFLD. Unlike Pik3ca, Pik3cb (encoding
for p110β) deficiency mice show normal hepatic lipid levels,
indicating that Pik3ca is specifically involved in NAFLD and
HCC development. PTEN dephosphorylates PI(3,4,5)P3 to sup­
press PI3K‐AKT signaling pathways, and PTEN deficient mice
demonstrate early onset of hepatic microvesicular steatosis due
to increased DNL and free fatty acid uptake that progresses into
NASH, liver fibrosis, and cancer [76]. As Pik3ca deficient mice
show steatosis without NASH and liver fibrosis, PTEN deficient
mice are useful rodent models to recapitulate human NAFLD
pathogenesis. AKT, a downstream PI3K signaling molecule, is
also involved in NAFLD development. Adenoviral overexpres­
sion of AKT significantly induces micro and macrovesicular
steatosis and NASH in 12 weeks and then results in HCC [77].
High‐fat diet‐induced hepatic fibrosis enhances deposition of
ECM, resulting in interactions between the ECM and AKT
through the integrin‐linked protein kinase (ILK) [78]. Thus,
liver‐specific ILK deletion shows improved HFD‐induced liver
steatosis and insulin resistance.
PI3K and AKT signaling in cancer
Platelet‐derived growth factor (PDGF) stimulates PI3K to gener­
ate PI(3,4)P2 and PI(3,4,5)P3 in smooth muscle, suggesting that
PI3K is important for cellular response to growth factors and
tumorigenesis [79], and the PI3K pathway is one of the most fre­
quently activated signaling pathways in human cancer. Thus,
PI3K mutations are commonly found in a variety of cancers, and
development of PI3K inhibitors is currently being conducted to
treat various cancers. Systemic identification of somatic muta­
tions in cancer genomes uncovers diverse mutations in human
cancer to help understand the molecular mechanisms of tumori­
genesis and develop targeted therapeutics. Class I PI3K, espe­
cially p110α/p85α heterodimer, activation is closely linked with
tumorigenesis [80, 81]. Most of the reported mutations in p110α
 39:  Insulin‐Mediated PI3K and AKT Signaling 493
encoded by the PIK3CA cluster are conserved in the region for
the helical (exon 9, E542K and E545K) and kinase domains
(exon20, H1047R) of p110α [33]. These mutations in PIK3CA
constitutively activate p110α kinase activity without upstream
stimuli by growth factors, promoting uncontrolled proliferative
signaling through the constitutive activation of AKT, S6K, and
4E‐BP in tumorigenesis. PIK3CA mutations in the helical and
kinase domains also show a distinct pattern in gender and tissue
specificity. In colorectal cancer, PIK3CA mutations are more fre­
quently found in women, and helical domain mutations (exon 9)
demonstrate more effects than kinase domain mutations (exon
20). The mutations of PIK3CB encoding for the p110β subunit on
E633K in the helical domain are linked with increased kinase
activity and enhanced cell proliferation. Like the p110α H1047R
mutation, the p110β E633K mutation enhances membrane tar­
geting to activate downstream signaling. Increased expression
of p110γ has been reported in chronic myeloid leukemia, invasive
breast carcinoma, and pancreatic ductal adenocarcinoma.
Recently, somatic mutations in PIK3R1, which encodes p85α, are
also found in cancers, and these mutations cluster in the iSH2
region (Figure 39.1a) releasing p110α to enhance kinase activity,
resulting in the phosphorylation of AKT on S473 [82]. Tumor
suppressor PTEN suppresses PI3K and AKT by dephosphoryla­
tion of PI(3,4,5)P3 and PI(3,4)P2, thus suppressing tumorigenesis.
In addition to the alteration in PI3Ks, mutations of other signaling
molecules in PI3K/AKT/mTORC axis such as PTEN, AKT,
TSC1/,2 and mTOR are also present. PTEN is frequently mutated
in human cancers including HCC (about 5% of HCC), and
decreased PTEN levels are observed in human hepatic steatosis
[83, 84]. Thus, liver‐specific deletion of the PTEN gene results in
hyperplastic, fatty degeneration of the liver and increased cell
proliferation [76]. Thus, the class I PI3K, p100α, appears to be an
ideal target for drug development, whereas little is known about
the genetic modification in class II and III PI3K in cancer.
AKT regulates cell proliferation and survival and is one of the
most activated downstream effectors of PI3K activation in tumo­
rigenesis. AKT phosphorylates GSK3β to prevent degradation of
cyclin D1 and activates the mTORC pathway to enhance the
translation of cyclin D1 and D3, resulting in cell proliferation.
AKT also prevents programmed cell death through the phospho­
rylation of Bcl‐2‐associated death promoter (BAD) to suppress
cytochrome c release from mitochondria and caspase activation
by phosphorylation of pro‐caspase 9. AKT promotes additional
processes such as angiogenesis by increased nitric oxide produc­
tion and metastasis through the increased secretion of matrix
metalloproteinases and epithelial‐mesenchymal transition
(EMT) [85]. Thus, AKT is closely linked with tumorigenesis.
Indeed, the AKT activation is observed in human cancers through
the amplification, overexpression, or point mutation of the AKT
kinase genes. AKT1 amplification causes the resistance to cis­
platinum treatment in gastric carcinomas. Somatic mutations of
AKT1 on E17K causes localization of AKT1 at the plasma mem­
brane to increase phosphorylation on Ser‐473 and Thr‐308,
resulting in leukemia in the mouse model [86]. In HCC, enhanced
expression of AKT2 protein was detected in about 40% tumors,
whereas AKT1 expression was similar in all cases.
As the PI3K/AKT/mTORC signaling axis is a major determi­
nant of tumorigenesis, several molecules for suppressing this
pathway have been developed and evaluated in clinical trials
[87]. Pan‐PI3K inhibitors such as buparlisib (BKM120) and
XL147 suppress all p110 isoforms, whereas isoform‐specific
PI3K inhibitors are developed to minimize the side‐effects of
pan‐PI3K inhibitors. p110α‐specific alpelisib and MLN1117
and p110β‐specific taselisib (GDC‐0032) are more effective to
tumors with PIK3CA mutation. In contrast, p110α‐specific
GSK2636771 is more effective in PTEN‐deficient tumors. AKT
inhibitors such as AZD‐5363, GDC‐0068, and MK‐2206 are in
clinical trials. MK‐2206, an allosteric pan‐AKT inhibitor, is in
clinical trial to treat cancers such as breast and colorectal can­
cers. MK‐2206 suppresses phosphorylation of AKT Thr‐308
and Ser‐473, resulting in the suppression of cell proliferation in
cell lines, which has PIK3CA activating mutations, PTEN inac­
tivation, and amplification of AKT [88]. As MK‐2206 treatment
demonstrates limited antitumor activity in phase II clinical trial,
combined treatment with chemotherapeutics and small mole­
cule inhibitors are in trials. Although rapalogs, the rapamycin
analogs to suppress mTOR, are developed as immunosuppres­
sants, they also suppress cell proliferation and angiogenesis to
treat cancer. Temsirolimus is the first rapalog that is approved
by the Food and Drug Administration (FDA) for the treatment
of renal cell carcinoma as it arrests the cell cycle in the G1 phase
and suppresses angiogenesis by reducing vascular endothelial
growth factor (VEGF) synthesis.
SUMMARY
Numerous studies over the past decade have begun to unravel
the complex molecular details and specific signaling pathways
that regulate normal liver function with respect to hepatic glu­
cose and fatty acid metabolism. These efforts have also begun to
elucidate the primary basis for dysregulated liver metabolism
that occurs during states of insulin resistance and T2D. We now
know that elevated fatty acid levels can serve as a significant
contributor to hepatic glucose production. When coupled with
liver insulin resistance this combination becomes a strong driver
of unrestrained gluconeogenesis and hyperglycemia. On the
other hand, hyperinsulinemia associated with insulin resistance
activates the transcriptional network that drives lipogenic gene
expression and DNL. When coupled with other dietary factors,
fructose further enhances DNL through ChREBP that at the
same time exacerbates hepatic glucose production through tran­
scriptional activation of G6PC. Several challenges remain but
further understanding of the network and interplay between
liver gluconeogenic and lipogenic signaling should provide new
avenues to develop specific therapeutic approaches for the treat­
ment of metabolic liver disease.
REFERENCES
1. Vanhaesebroeck, B., Guillermet‐Guibert, J., Graupera, M., and Bilanges, B.
The emerging mechanisms of isoform‐specific PI3K signalling. Nat Rev Mol
Cell Biol, 2010;11(5):329–41.
2. Wymann, M.P., Bulgarelli‐Leva, G., Zvelebil, M.J. et  al. Wortmannin
­inactivates phosphoinositide 3‐kinase by covalent modification of Lys‐802,
a  residue involved in the phosphate transfer reaction. Mol Cell Biol,
1996;16(4):1722–33.
494 THE LIVER:  REFERENCES
3. Backer, J.M. The regulation of class IA PI 3‐kinases by inter‐subunit interac­
tions. Curr Top Microbiol Immunol, 2010;346:87–114.
4. Carpenter, C.L., Auger, K.R., Chanudhuri, M. et  al. Phosphoinositide 3‐
kinase is activated by phosphopeptides that bind to the SH2 domains of the
85‐kDa subunit. J Biol Chem, 1993;268(13):9478–83.
5. Myers, M.G., Jr., Backer, J.M., Sun, X.J. et al. IRS‐1 activates phosphati­
dylinositol 3’‐kinase by associating with src homology 2 domains of p85.
Proc Natl Acad Sci USA, 1992;89(21):10350–4.
6. Bellacosa, A., Testa, J.R., Staal, S.P., and Tsichlis, P.N. A retroviral onco­
gene, akt, encoding a serine‐threonine kinase containing an SH2‐like region.
Science, 1991;254(5029):274–7.
7. Coffer, P.J. and Woodgett, J.R. Molecular cloning and characterisation of a
novel putative protein‐serine kinase related to the cAMP‐dependent and pro­
tein kinase C families. Eur J Biochem, 1991;201(2):475–81.
8. Jones, P.F., Jakubowicz, T., Pitossi, F.J., Maurer, F., and Hemmings, B.A.
Molecular cloning and identification of a serine/threonine protein kinase of the
second‐messenger subfamily. Proc Natl Acad Sci USA, 1991;88(10):4171–5.
9. Staal, S.P. Molecular cloning of the akt oncogene and its human homologues
AKT1 and AKT2: amplification of AKT1 in a primary human gastric adeno­
carcinoma. Proc Natl Acad Sci USA, 1987;84(14):5034–7.
10. Klarlund, J.K., Guilherme A., Holik, J.J., Virbasius, J.V., Chawla, A., and
Czech, M.P. Signaling by phosphoinositide‐3,4,5‐trisphosphate through pro­
teins containing pleckstrin and Sec7 homology domains. Science,
1997;275(5308):1927–30.
11. Stephens, L., Anderson, K., Stokoe, D. et al. Protein kinase B kinases that
mediate phosphatidylinositol 3,4,5‐trisphosphate‐dependent activation of
protein kinase B. Science, 1998;279(5351):710–4.
12. Keranen, L.M., Dutil, E.M., and Newton, A.C. Protein kinase C is regulated
in vivo by three functionally distinct phosphorylations. Curr Biol,
1995;5(12):1394–403.
13. Calleja, V., Alcor, D., Laguerre, M. et al. Intramolecular and intermolecular
interactions of protein kinase B define its activation in vivo. PLoS Biol,
2007;5(4):e95.
14. Sarbassov, D.D., Guertin, D.A., Ali, S.M., and Sabatini, D.M.
Phosphorylation and regulation of Akt/PKB by the rictor‐mTOR complex.
Science, 2005;307(5712):1098–101.
15. Pessin, J.E. and Saltiel, A.R. Signaling pathways in insulin action: molecular
targets of insulin resistance. J Clin Invest, 2000;106(2):165–9.
16. Taniguchi, C.M., Emanuelli, B., and Kahn, C.R. Critical nodes in signalling
pathways: insights into insulin action. Nat Rev Mol Cell Biol, 2006;7(2):
85–96.
17. Ullrich, A., Bell, J.R., Chen, E.Y. et al. Human insulin receptor and its rela­
tionship to the tyrosine kinase family of oncogenes. Nature,
1985;313(6005):756–61.
18. Ebina, Y., Ellis, L., Jarnagin, K. et al. The human insulin receptor cDNA: the
structural basis for hormone‐activated transmembrane signalling. Cell,
1985;40(4):747–58.
19. Voliovitch, H., Schindler, D.G., Hadari, Y.R., Taylor, S.I., Accili, D., and
Zick, Y. Tyrosine phosphorylation of insulin receptor substrate‐1 in vivo
depends upon the presence of its pleckstrin homology region. J Biol Chem,
1995;270(30):18083–7.
20. Araki, E., Lipes, M.A., Patti, M.E. et al. Alternative pathway of insulin sig­
nalling in mice with targeted disruption of the IRS‐1 gene. Nature,
1994;372(6502):186–90.
21. Withers, D.J., Gutierrez, J.S., Towery, H. et al. Disruption of IRS‐2 causes
type 2 diabetes in mice. Nature, 1998;391(6670):900–4.
22. Zhang, X., Vadas, O., Perisic, O. et al. Structure of lipid kinase p110beta/
p85beta elucidates an unusual SH2‐domain‐mediated inhibitory mechanism.
Mol Cell, 2011;41(5):567–78.
23. Sopasakis, V.R., Liu, P., Suzuki, R. et  al. Specific roles of the p110alpha
isoform of phosphatidylinsositol 3‐kinase in hepatic insulin signaling and
metabolic regulation. Cell Metab, 2010;11(3):220–30.
24. Terauchi, Y., Tsuji, Y., Satoh, S. et al. Increased insulin sensitivity and hypo­
glycaemia in mice lacking the p85 alpha subunit of phosphoinositide 3‐
kinase. Nat Genet, 1999;21(2):230–5.
25. Wijesekara, N., Konrad, D., Eweida, M. et al. Muscle‐specific Pten deletion
protects against insulin resistance and diabetes. Mol Cell Biol,
2005;25(3):1135–45.
26. Sleeman, M.W., Wortley, K.E., Lai, K.M. et  al. Absence of the lipid
phosphatase SHIP2 confers resistance to dietary obesity. Nat Med,
­ 51. Koo, S.H., Flechner, L., Qi, L. et al. The CREB coactivator TORC2 is a key
2005;11(2):199–205.
27. Jacinto, E., Facchinetti, V., Liu, D. et al. SIN1/MIP1 maintains rictor‐mTOR
complex integrity and regulates Akt phosphorylation and substrate specific­
ity. Cell, 2006;127(1):125–37.
28. Cho, H., Mu, J., Kim, J.K., Thorvaldsen, J.L. et al. Insulin resistance and a
diabetes mellitus‐like syndrome in mice lacking the protein kinase Akt2
(PKB beta). Science, 2001;292(5522):1728–31.
29. Manning, B.D. and Cantley, L.C. AKT/PKB signaling: navigating down­
stream. Cell, 2007;129(7):1261–74.
30. Cross, D.A., Alessi, D.R., Cohen, P., Andjelkovich, M., and Hemmings, B.A.
Inhibition of glycogen synthase kinase‐3 by insulin mediated by protein
kinase B. Nature, 1995;378(6559):785–9.
31. Saxton, R.A. and Sabatini, D.M. mTOR signaling in growth, metabolism,
and disease. Cell, 2017;169(2):361–71.
32. Czech, M.P., Tencerova, M., Pedersen, D.J., and Aouadi, M. Insulin
­signalling mechanisms for triacylglycerol storage. Diabetologia, 2013;56(5):
949–64.
33. Perry, R.J., Camporez, J.P., Kursawe, R. et al. Hepatic acetyl CoA links adi­
pose tissue inflammation to hepatic insulin resistance and type 2 diabetes.
Cell, 2015;160(4):745–58.
34. Rebrin, K., Steil, G.M., Mittelman, S.D., and Bergman, R.N. Causal linkage
between insulin suppression of lipolysis and suppression of liver glucose
output in dogs. J Clin Invest, 1996;98(3):741–9.
35. Ramnanan, C.J., Edgerton, D.S., Rivera, N. et al. Molecular characterization
of insulin‐mediated suppression of hepatic glucose production in vivo.
Diabetes, 2010;59(6):1302–11.
36. Perry, R.J., Zhang, X.M., Zhang, D. et  al. Leptin reverses diabetes by
­suppression of the hypothalamic‐pituitary‐adrenal axis. Nat Med, 2014;20(7):
759–63.
37. Petersen, M.C., Madiraju, A.K., Gassaway, B.M. et  al. Insulin receptor
Thr1160 phosphorylation mediates lipid‐induced hepatic insulin resistance.
J Clin Invest, 2016;126(11):4361–71.
38. Chin, J.E., Liu, F., and Roth, R.A. Activation of protein kinase C alpha
inhibits insulin‐stimulated tyrosine phosphorylation of insulin receptor
­
­substrate‐1. Mol Endocrinol, 1994;8(1):51–8.
39. Badin, P.M., Vila, I.K., Louche, K. et al. High‐fat diet‐mediated lipotoxicity
and insulin resistance is related to impaired lipase expression in mouse skel­
etal muscle. Endocrinology, 2013;154(4):1444–53.
40. Griffin, M.E., Marcucci, M.J. et al. Free fatty acid‐induced insulin resistance
is associated with activation of protein kinase C theta and alterations in the
insulin signaling cascade. Diabetes, 1999;48(6):1270–4.
41. Jornayvaz, F.R. and Shulman, G.I. Diacylglycerol activation of protein
kinase Cepsilon and hepatic insulin resistance. Cell Metab, 2012;15(5):
574–84.
42. Haemmerle, G., Lass, A., Zimmermann, R. et  al. Defective lipolysis and
altered energy metabolism in mice lacking adipose triglyceride lipase.
Science, 2006;312(5774):734–7.
43. Chavez, J.A. and Summers, S.A. A ceramide‐centric view of insulin resist­
ance. Cell Metab, 2012;15(5):585–94.
44. Shimomura, I., Matsuda, M., Hammer, R.E., Bashmakov, Y., Brown, M.S.,
and Goldstein, J.L. Decreased IRS‐2 and increased SREBP‐1c lead to mixed
insulin resistance and sensitivity in livers of lipodystrophic and ob/ob mice.
Mol Cell, 2000;6(1):77–86.
45. Rui, L., Yuan, M., Frantz, D., Shoelson, S., and White, M.F. SOCS‐1 and
SOCS‐3 block insulin signaling by ubiquitin‐mediated degradation of IRS1
and IRS2. J Biol Chem, 2002;277(44):42394–8.
46. Kwon, H. and Pessin, J.E. Adipokines mediate inflammation and insulin
resistance. Front Endocrinol, 2013;4:71.
47. Boura‐Halfon, S. and Zick, Y. Phosphorylation of IRS proteins, insulin
action, and insulin resistance. Am J Physiol Endocrinol Metab, 2009;296(4):
E581–91.
48. Hirosumi, J., Tuncman, G., Chang, L. et al. A central role for JNK in obesity
and insulin resistance. Nature, 2002;420(6913):333–6.
49. Bard‐Chapeau, E.A., Hevener, A.L., Long, S., Zhang, E.E., Olefsky, J.M.,
and Feng, G.S. Deletion of Gab1 in the liver leads to enhanced glucose toler­
ance and improved hepatic insulin action. Nat Med, 2005;11(5):567–71.
50. Rider, M.H., Bertrand, L., Vertommen, D., Michels, P.A., Rousseau, G.G.,
and Hue, L. 6‐phosphofructo‐2‐kinase/fructose‐2,6‐bisphosphatase: head‐
to‐head with a bifunctional enzyme that controls glycolysis. Biochem J,
2004;381(Pt 3):561–79.
regulator of fasting glucose metabolism. Nature, 2005;437(7062):1109–11.
 39:  Insulin‐Mediated PI3K and AKT Signaling 495
52. Wang, Y., Li G., Goode, J. et  al. Inositol‐1,4,5‐trisphosphate receptor
­regulates hepatic gluconeogenesis in fasting and diabetes. Nature, 2012;
485(7396):128–32.
53. Dentin, R., Liu, Y., Koo, S.H. et al. Insulin modulates gluconeogenesis by
inhibition of the coactivator TORC2. Nature, 2007;449(7160):366–9.
54. Brunet, A., Bonni, A., Zigmond, M.J. et al. Akt promotes cell survival by
phosphorylating and inhibiting a Forkhead transcription factor. Cell,
1999;96(6):857–68.
55. Tzivion, G., Dobson, M., and Ramakrishnan, G. FoxO transcription
­factors;  regulation by AKT and 14‐3‐3 proteins. Biochim Biophys Acta,
2011;1813(11):1938–45.
56. O’Sullivan, I., Zhang, W., Wasserman, D.H. et al. FoxO1 integrates direct
and indirect effects of insulin on hepatic glucose production and glucose
utilization. Nat Commun, 2015;6:7079.
57. Titchenell, P.M., Chu, Q., Monks, B.R., and Birnbaum, M.J. Hepatic insulin
signalling is dispensable for suppression of glucose output by insulin in vivo.
Nat Commun, 2015;6:7078.
58. Samuel, V.T., Beddow, S.A., Iwasaki, T. et al. Fasting hyperglycemia is not
associated with increased expression of PEPCK or G6Pc in patients with
Type 2 diabetes. Proc Natl Acad Sci USA, 2009;106(29):12121–6.
59. Kim, M.S., Krawczyk, S.A., Doridot, L. et al. ChREBP regulates fructose‐
induced glucose production independently of insulin signaling. J Clin Invest,
2016;126(11):4372–86.
60. Solinas, G., Boren, J., and Dulloo, A.G. De novo lipogenesis in metabolic
homeostasis: more friend than foe? Mol Metab, 2015;4(5):367–77.
61. Wang, Y., Viscarra, J., Kim, S.J., and Sul, H.S. Transcriptional regulation of
hepatic lipogenesis. Nat Rev Mol Cell Biol, 2015;16(11):678–89.
62. Hegarty, B.D., Bobard, A., Hainault, I., Ferre, P., Bossard, P., and Foufelle, F.
Distinct roles of insulin and liver X receptor in the induction and cleavage
of  sterol regulatory element‐binding protein‐1c. Proc Natl Acad Sci USA,
2005;102(3):791–6.
63. Peterson, T.R., Sengupta, S.S., Harris, T.E. et al. mTOR complex 1 regulates
lipin 1 localization to control the SREBP pathway. Cell, 2011;146(3):
408–20.
64. Dentin, R., Tomas‐Cobos, L., Foufelle, F. et al. Glucose 6‐phosphate, rather
than xylulose 5‐phosphate, is required for the activation of ChREBP in
response to glucose in the liver. J Hepatol, 2012;56(1):199–209.
65. Uyeda, K. and Repa, J.J. Carbohydrate response element binding protein,
ChREBP, a transcription factor coupling hepatic glucose utilization and lipid
synthesis. Cell Metab, 2006;4(2):107–10.
66. Brownsey, R.W., Boone, A.N., Elliott, J.E., Kulpa, J.E., and Lee, W.M.
Regulation of acetyl‐CoA carboxylase. Biochem Soc Trans,
2006;34(2):223–7.
67. Peng, I.C., Chen, Z., Sun, W. et al. Glucagon regulates ACC activity in adi­
pocytes through the CAMKKbeta/AMPK pathway. Am J Physiol Endocrinol
Metab, 2012;302(12):E1560–8.
68. Witters, L.A., Watts, T.D., Daniels, D.L., and Evans, J.L. Insulin stimulates
the dephosphorylation and activation of acetyl‐CoA carboxylase. Proc Natl
Acad Sci USA, 1988;85(15):5473–7.
69. Zammit, V.A. Role of insulin in hepatic fatty acid partitioning: emerging
concepts. Biochem, J., 1996;314(1):1–14.
70. Lazo, M. and Clark, J.M. The epidemiology of nonalcoholic fatty liver dis­
ease: a global perspective. Semin Liver Dis, 2008;28(4):339–50.
71. Marchesini, G., Brizi, M., Bianchi, G. et al. Nonalcoholic fatty liver disease:
a feature of the metabolic syndrome. Diabetes, 2001;50(8):1844–50.
72. Henry, R.R., Scheaffer, L., and Olefsky, J.M. Glycemic effects of intensive
caloric restriction and isocaloric refeeding in noninsulin‐dependent diabetes
mellitus. J Clin Endocrinol Metab, 1985;61(5):917–25.
73. Biddinger, S.B. and Kahn, C.R. From mice to men: insights into the insulin
resistance syndromes. Annu Rev Physiol, 2006;68:123–58.
74. Kudo, Y., Tanaka, Y., Tateishi, K. et al. Altered composition of fatty acids
exacerbates hepatotumorigenesis during activation of the phosphatidylinosi­
tol 3‐kinase pathway. J Hepatol, 2011;55(6):1400–8.
75. Chattopadhyay, M., Selinger, E.S., Ballou, L.M., and Lin, R.Z. Ablation of
PI3K p110‐alpha prevents high‐fat diet‐induced liver steatosis. Diabetes,
2011;60(5):1483–92.
76. Horie, Y., Suzuki, A., Kataoka, E. et al. Hepatocyte‐specific Pten deficiency
results in steatohepatitis and hepatocellular carcinomas. J Clin Invest,
2004;113(12):1774–83.
77. Calvisi, D.F., Wang, C., Ho, C. et al. Increased lipogenesis, induced by AKT‐
mTORC1‐RPS6 signaling, promotes development of human hepatocellular
carcinoma. Gastroenterology, 2011;140(3):1071–83.
78. Williams, A.S., Trefts, E., Lantier, L. et al. Integrin‐linked kinase is neces­
sary for the development of diet‐induced hepatic insulin resistance. Diabetes,
2017;66(2):325–34.
79. Auger, K.R., Serunian, L.A., Soltoff, S.P., Libby, P., and Cantley, L.C.
PDGF‐dependent tyrosine phosphorylation stimulates production of novel
polyphosphoinositides in intact cells. Cell, 1989;57(1):167–75.
80. Yuan, T.L. and Cantley, L.C. PI3K pathway alterations in cancer: variations
on a theme. Oncogene, 2008;27(41):5497–510.
81. Samuels, Y., Wang, Z., Bardelli, A. et al. High frequency of mutations of the
PIK3CA gene in human cancers. Science, 2004;304(5670):554.
82. Jaiswal, B.S., Janakiraman, V., Kljavin, N.M. et  al. Somatic mutations in
p85alpha promote tumorigenesis through class IA PI3K activation. Cancer
Cell, 2009;16(6):463–74.
83. Xu, Z., Hu, J., Cao, H. et al. Loss of Pten synergizes with c‐Met to promote
hepatocellular carcinoma development via mTORC2 pathway. Exp Mol
Med, 2018;50(1):e417.
84. Whittaker, S., Marais, R., and Zhu, A.X. The role of signaling pathways in
the development and treatment of hepatocellular carcinoma. Oncogene,
2010;29(36):4989–5005.
85. Grille, S.J., Bellacosa, A., Upson, J. et  al. The protein kinase Akt induces
epithelial mesenchymal transition and promotes enhanced motility and inva­
siveness of squamous cell carcinoma lines. Cancer Res, 2003;63(9):2172–8.
86. Carpten, J.D., Faber, A.L., Horn, C. et  al. A transforming mutation in the
pleckstrin homology domain of AKT1 in cancer. Nature, 2007;448(7152):
439–44.
87. Mayer, I.A. and Arteaga, C.L. The PI3K/AKT pathway as a target for cancer
treatment. Annu Rev Med, 2016;67:11–28.
88. Sangai, T., Akcakanat, A., Chen, H. et  al. Biomarkers of response to
Akt  inhibitor MK‐2206 in breast cancer. Clin Cancer Res, 2012;18(20):
5816–28.
 Ca2+ Signaling in the Liver
40 Mateus T. Guerra1, M. Fatima Leite2, and Michael H. Nathanson1
Departments of Medicine and Cell Biology, Yale University School of Medicine, New Haven, CT, USA
1
Department of Physiology and Biophysics, UFMG, Belo Horizonte, Brazil
2
MECHANISMS OF Ca2+ SIGNALING
Hormone receptors and initiation
of Ca2+ signals
Hormones, neurotransmitters, and growth factors initiate cyto-
solic Ca2+ (Cai2+) signaling through a variety of mechanisms [1].
Most commonly, these factors trigger intracellular signaling
cascades by binding to either G protein‐coupled receptors
(GPCRs) or receptor tyrosine kinases (RTKs) on the surface of
liver cells. Several GPCRs in the liver have been particularly
well studied. These include the V1a vasopressin receptor, the α1B
adrenergic receptor, several subtypes of the P2Y class of
purinergic receptors, and the angiotensin receptors. RTKs for
several types of growth factors are expressed in liver as well,
such as insulin, epidermal growth factor (EGF), and hepatocyte
growth factor (HGF). Upon activation, each of these receptors
initiates signaling events that lead to an increase in cytosolic
and/or nuclear Ca2+. The binding of Ca2+ mobilizing hormones
and growth factors to their specific plasma membrane receptors
activates phospholipase C (PLC), which is membrane‐associ-
ated. α, β, and γ subtypes of PLC are recognized [2]. Many iso-
forms of these subtypes have been identified; PLCβ1 and PLCβ2
are the isoforms activated by G proteins, while PLCγ is acti-
vated by RTKs [2]. Activation of PLC by GPCRs hydrolyzes the
pool of phospholipid phosphatidylinositol‐4,5‐bisphosphate
(PIP2) within the plasma membrane. The hydrolysis of PIP2
by  PLC results in the formation of diacylglycerol (DAG) and
inositol 1,4,5 trisphosphate (InsP3). DAG remains at the plasma
membrane to activate protein kinase C (PKC), while InsP3 dif-
fuses into the cytosol to release Ca2+ from intracellular stores via
its interaction with the InsP3 receptor (InsP3R). Activation of
PLC by RTKs was thought to similarly act on the plasma
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
membrane pool of PIP2 [1], but several lines of evidence now
suggest it instead hydrolyzes nuclear PIP2, which generates
InsP3 in the nucleus to increase free Ca2+ within the nucleo-
plasm [3]. The sequence of events linking hormone receptors
and RTKs to Cai2+ signaling is summarized in Figure 40.1.
Inositol 1,4,5‐trisphosphate receptor
The InsP3R is an InsP3‐gated Ca2+ channel located on the endo-
plasmic reticulum (ER), although there may also be active chan-
nels at the plasma membrane. In hepatocytes, increases in Cai2+
are initiated by binding of InsP3 to the InsP3R. Full length
sequences for three distinct InsP3R genes have been determined
and knockout mice have been generated for each of these three
isoforms [4]. These isoforms share considerable sequence
homology, but each subtype is expressed and regulated in a dis-
tinct fashion. There also are isoform‐specific differences in tis-
sue expression and subcellular distribution, suggesting that the
isoforms serve distinct roles in Cai2+ signaling. InsP3Rs are
homotetramers consisting of 313, 307, or 304 KDa subunits,
corresponding to the type I, II, or III isoforms, respectively, and
heterotetramers can form as well [5]. The InsP3R has six mem-
brane spanning domains, oriented so that the N‐terminus of the
protein is in the cytoplasm. The receptor has an uneven bell
shape as deduced by cryo‐electron microscopy with the bulkier
N‐terminus directed toward the cytoplasm and the narrower end
facing the ER lumen.
Deletion analysis studies of the mouse InsP3R1 have revealed
three functional regions within the InsP3R: an N‐terminal
InsP3‐binding domain, a Ca2+ channel‐forming C‐terminal
domain, and a regulatory domain flanked by the InsP3‐binding
domain and the channel region. Several phosphorylation sites
are found along the InsP3R’s amino acid sequence. Moreover,
 40: Ca2+ SIGNALING IN THE LIVER 497

Growth factor
Hormone

PLC
PLC γ
β Receptor
InsP3 tyrosine kinase
G-
protein
G-protein
coupled receptor
Ca2+

InsP3R

Endoplasmic
reticulum

To nucleus

Ca2+

Figure 40.1  Mechanisms of hormone‐ and growth factor‐induced Ca2+ signaling in hepatocytes. Upon binding to its specific G protein‐coupled
plasma membrane receptor, a hormone induces phospholipase C (PLC)‐beta to hydrolyze phosphatidylinositol 4,5 bisphosphate (PIP2) to form
diacylglycerol (DAG) and inositol 1,4,5‐trisphosphate (InsP3). Alternatively, upon binding to its specific receptor tyrosine kinase (RTK), a growth
hormone induces PLC‐gamma to hydrolyze PIP2 to form DAG and InsP3. InsP3 then binds to its tetrameric receptor (InsP3R) in the ER, which acts
as a Ca2+ channel to allow Ca2+ to enter the cytosol. RTKs such as c‐Met and the insulin receptor may also translocate to the nucleus in hepatocytes
to activate nuclear PLC and increase Ca2+ within the nucleoplasm.

the regulatory domain interacts with several protein partners
responsible for modulating channel activity. The C‐terminus
also interacts with protein partners, which influences the subcel-
lular localization of the receptor [6].
The InsP3 binding domain includes multiple sequences scat-
tered throughout the N‐terminal region and key residues respon-
sible for the interaction between InsP3 and its receptor have been
identified by site‐directed mutagenesis and X‐ray crystallography
[7]. Upon InsP3 binding, the receptor undergoes a conformational
change that opens the Ca2+ channel, so that Ca2+ in the ER is
released into the cytosol. Although each of the three InsP3R iso-
forms acts as an InsP3‐gated Ca2+ channel, the isoforms are not
uniformly sensitive to InsP3. The relative order of affinity is type
II > type I > type III. The Kd for InsP3R2 is 27 nM, which is twice
as great as the affinity of InsP3R1, and ten times the affinity of
InsP3R3. InsP3 is absolutely required for Ca2+ release via the
InsP3R, but the concentration of Ca2+ in the cytosol modulates the
open probability of the Ca2+ channel [8]. This dependence of the
InsP3R on the cytosolic Ca2+ concentration is important for
organizing the spatial and temporal pattern of Cai2+ signals.
The activity of InsP3R Ca2+ channels is highly dependent on
post‐translational modifications such as phosphorylation, binding
of protein cofactors [6], and O‐linked β‐N‐acetylglucosamine
glycosylation (O‐GlcNAcylation) [9], which operates by
reversible addition of N‐acetylglucosamine monosaccharide to
serine/threonine residues. It is speculated that O‐GlcNAcylation
of InsP3R in liver may be particularly important in states of high
nutrient availability such as NAFLD or diabetes. InsP3R activity
can also be modulated by protein degradation through the protea-
some pathway or by selective proteolysis, providing yet another
level of regulation of this Ca2+ release channel.
Many cell types express more than one InsP3R isoform.
Hepatocytes normally express only InsP3R1 and InsP3R2.
InsP3R3 is not detected in hepatocytes, but is the predominant
isoform in bile duct epithelia [10]. Moreover, InsP3R2 is most
concentrated in the apical region of hepatocytes while InsP3R1
is dispersed throughout the cell (Figure 40.2a) [11]. In contrast,
in bile duct epithelia InsP3R3 is most concentrated in the apical
region, while InsP3R1 and InsP3R2 are dispersed throughout
the cell (Figure 40.2b) [12]. The apical localization of InsP3R2
in hepatocytes depends upon lipid rafts in the canalicular mem-
brane [13], but the mechanism by which InsP3Rs associate with
lipid rafts is not yet known.
Ryanodine receptor
The other major class of intracellular Ca2+ release channels is
the ryanodine receptor (RyR). Like InsP3R, RyR has three
498 THE LIVER:  MECHANISMS OF Ca2+ SIGNALING

(a) (b)

Figure 40.2  InsP3 receptors (InsP3Rs) are concentrated in the apical region of hepatocytes and cholangiocytes. (a) Confocal immunofluorescence
image of rat liver shows the distribution of the InsP3R2 (green) in hepatocytes. Cells also are labeled with the actin stain phalloidin (red), which
outlines individual hepatocytes and labels their apical region most intensely. InsP3R2 colocalizes with periapical actin (arrowheads), and thus is
most concentrated in the apical region. (b) Confocal immunofluorescence image of a bile duct from a biopsy of a normal human liver shows the
distribution of InsP3R3 (green). Cholangiocytes also are labeled to reveal CFTR (red), which resides on their apical membrane. The InsP3R3 is
most concentrated in the apical region, just beneath CFTR. Reproduced from [105] with permission of Elsevier.

family members (RyR1, RyR2, and RyR3). These channels
contribute to Cai2+ signaling by releasing Ca2+ from the lumen of
the sarcoplasmic/endoplasmic reticulum into the cytosol [1].
RyRs autocatalytically release Ca2+ via Ca2+‐induced Ca2+
release (CICR). Rat hepatocytes express only a truncated RyR1
isoform that by itself is not able to elicit Ca2+ release but can
increase the frequency of InsP3‐induced Ca2+ oscillations [14].
These findings suggest there is a novel RyR‐like protein which
contributes to Ca2+ signaling in hepatocytes.
Nicotinic acid adenine dinucleotide phosphate (NAADP) is a
second messenger that mediates release of Ca2+ from intracel-
lular acidic compartments such as lysosomes and secretory
granules. The role of NAADP in hepatocyte Cai2+ signaling is
not clear, although there is in vitro evidence demonstrating
NAADP‐dependent Ca2+ release from reconstituted hepatocyte
lysosomes and microsomes [15].
Mitochondria
Mitochondria are best known for the metabolic and respiratory
role they play in cells. However, mitochondria strongly influ-
ence Cai2+ signaling as well, by taking up Ca2+ from and releas-
ing it back into the cytosol. Mitochondria have their own Ca2+
transport machinery, involving Ca2+ influx through a uniporter,
and Ca2+ efflux via both a Na+ exchanger and a H+ exchanger
[16]. The uniporter is driven by the potential gradient across the
mitochondrial membrane, while the Ca2+ efflux mechanisms
are  active transport systems. The CCDC109A gene encodes
the  mitochondrial calcium uniporter (MCU) protein, and Ca2+
influx in mitochondria occurs via a macromolecular complex on
the inner mitochondrial matrix composed of pore‐forming subu-
nits (MCU, MCUb, and essential MCU regulator [EMRE]),
plus regulatory proteins MICU1/2 that modulate the activity of
the uniporter according to the concentration of cytosolic Ca2+
in  the vicinity of the uniporter [16]. Pathological Ca2+ efflux
from mitochondria also can occur through the permeability
transition pore (PTP). Formation of this pore results in a sudden,
marked increase in the permeability of the mitochondrial inner
membrane to ions and small molecules. Irreversible formation
of the PTP can dissipate the potential gradient across the mito-
chondrial membrane, leading to mitochondrial swelling and
irreversible cell injury, including apoptosis and necrosis. Indeed,
liver regeneration after partial hepatectomy is halted due to mas-
sive necrotic cell death in mice lacking MICU1 in hepatocytes,
which undergo sustained elevations in mitochondrial Ca2+ [17].
Mitochondria, like the ER, are densely distributed in hepato-
cytes. There are regions of close proximity between these two
organelles (Figure 40.3), and InsP3R1 clustered in these regions
are largely responsible for mitochondrial Ca2+ signaling in
hepatocytes [18].
The discovery of interactions between apoptotic proteins,
InsP3R, and mitochondrial proteins helped uncover an essential
interplay between ER and mitochondrial Ca2+ signaling in the
process of programmed cell death. The current view is that anti‐
apoptotic proteins belonging to the Bcl‐2 family such as Bcl‐XL
and Bcl‐2 induce ER Ca2+ leak through direct interaction with
InsP3R1. This Ca2+ leak then reduces the ER Ca2+ available to be
released into the cytosol and consequently, decreases the amount
of Ca2+ taken up by surrounding mitochondria. Because mito-
chondria are less likely to become overloaded with Ca2+, PTP
formation is hampered and there is diminished cellular sensitiv-
ity to apoptotic stimuli [19]. Accordingly, buffering of mito-
chondrial matrix Ca2+ prevents apoptotic death of hepatocytes
and accelerates liver regeneration after partial hepatectomy
[20]. Mcl‐1 is another anti‐apoptotic protein that acts in part
by diminishing mitochondrial Ca2+ signals. Unlike other Bcl‐2
family members, Mcl‐1 is expressed in mitochondria and
­inhibits mitochondrial Ca2+ signals directly [21]. Mcl‐1 is the
principal anti‐apoptotic protein in cholangiocytes, and its over-
expression may promote the development of cholangiocarci-
noma. In contrast, overexpression of the pro‐apoptotic proteins
Bax and Bak lead to ER Ca2+ overload and greater susceptibility
to Ca2+‐mediated apoptosis.
Mitochondria can sequester significant amounts of Ca2+ from
the cytosol and mitochondrial Ca2+ can closely parallel the
­cytosolic Ca2+ increase induced by receptor activation. This
 40: Ca2+ SIGNALING IN THE LIVER 499

(a) (b)

Figure 40.3  Mitochondria and endoplasmic reticulum form regions that are in close proximity in hepatocytes. (a) Histologically normal human liver
biopsy specimen stained with markers for an ER protein (PDI, green) and a mitochondrial protein (Tom‐22, red) show the close association between
these two organelles in hepatocytes. Regions in white represent areas of colocalization, scale bar = 20 μM. (b) EM micrograph of a mouse hepatocyte
showing regions of close (less than 40 nm) contact between the ER (green) and mitochondria (magenta). Mitochondrial Ca2+ signals are derived from
Ca2+ that is released from InsP3Rs that cluster in these regions. Reproduced from [18] with permission of John Wiley and Sons.

functional effect is consistent with electron microscopy data
that show a close proximity between mitochondria and ER and
a dynamic physical interaction between these two organelles
[22]. There is also functional evidence for microdomains where
mitochondria and InsP3Rs are in close apposition, so that mito-
chondria take up a significant fraction of the Ca2+ released
by the InsP3R [23]. In hepatocytes, InsP3R1 is the isoform in
proximity to mitochondria and is responsible for mitochondrial
Ca2+ signals [18].
Ca2+ uptake by mitochondria affects multiple factors in cell
metabolism, including the mitochondrial proton motive force,
electron transport, and the activity of dehydrogenases associ-
ated with the TCA cycle, adenine nucleotide translocase, the
F1‐ATPase, and PDH phosphorylase. Slow or small Cai2+ eleva-
tions are not transmitted effectively into mitochondria, and
are less able to activate mitochondrial metabolism. In contrast,
Cai2+ oscillations trigger mitochondrial Ca2+ oscillations and
sustained NAD(P)H formation [24]. Thus, the frequency rather
than the amplitude of Cai2+ oscillations regulates mitochondrial
metabolism [24].
in turn to act as a second messenger that regulates multiple cell
functions simultaneously.
Ca2+ signaling was initially studied with fluorescent dyes and
the bioluminescent protein aequorin, but genetically encoded
fluorescent Ca2+ indicators (GECI) are now more widely used.
These GECI are based on modified versions of the circularly
permuted (cp) green fluorescent protein (GFP) and other fluo-
rescent proteins such as blue fluorescent protein (BFP) and
mApple red fluorescent protein (RFP) [26]. These cp proteins
are fused with the Ca2+ binding protein calmodulin (CaM) and a
calmodulin (CaM)–binding region of chicken myosin light
chain kinase (M13). Ca2+ binding induces a conformational
change that results in increased fluorescence. Additional refine-
ments have yielded ratiometric GECI and sensors with different
affinities for Ca2+. Other GECI have been developed that are
based on the principle of fluorescence resonance energy transfer
(FRET). These indicators provide the advantage of better signal
to noise ratio and photostability as compared to the aequorin
reporters and yet retain the ability to be expressed in diverse tis-
sues and subcellular locations.

Ca2+ signaling patterns in hepatocytes

ORGANIZATION OF Ca SIGNALS 2+
Detection of Ca signals in hepatocytes
2+
Our understanding of the complexity of Cai2+ signals has evolved
dramatically, largely due to two technical advances: (i) develop-
ment of Ca2+‐sensitive fluorescent dyes and proteins has permit-
ted Cai2+ to be monitored continuously in live cells, and (ii)
improvements in fluorescence imaging techniques allows Cai2+
to be detected not only in cell populations but in single cells and
in distinct subcellular regions of individual cells [25]. As obser-
vations move from cell populations to single cells to subcellular
regions, the complexity of Cai2+ signaling patterns increases and
the time scale of signaling events decreases (Figure 40.4). It is
now appreciated that Cai2+ signaling is regulated at the subcel-
lular level, and that this level of regulation is necessary for Cai2+
Peptide hormones such as vasopressin or angiotensin generally
induce biphasic increases in Cai2+ in populations of isolated
hepatocytes (Figure 40.4). These increases typically consist of
two components. The first component is the rapid peak, then fall
in Cai2+ which takes place over a period of seconds. This is due
to release of Ca2+ from InsP3‐sensitive stores and occurs even in
Ca2+‐free medium. The second component is the sustained pla-
teau in Cai2+, which follows the rapid peak. This occurs only in
the presence of extracellular Ca2+ and is due mostly to influx
of Ca2+ to replenish depleted intracellular stores. Although this
population response is highly reproducible, Ca2+ signaling pat-
terns vary markedly among single hepatocytes. Different stimuli
evoke distinct responses, and additional variation occurs among
hepatocytes stimulated under identical conditions [27]. The
range of signaling patterns seen among single hepatocytes
500 THE LIVER:  ORGANIZATION OF Ca2+ SIGNALS

(a) phenylephrine‐induced oscillations as well. Differences in the

700 frequency of Cai2+ oscillations regulate gene transcription
600 in  some cell systems [29], although this has not yet been
Cytosolic Ca2+ (nM)

500 ­demonstrated in hepatocytes.

It is currently thought that Cai2+ oscillations do not depend
400
upon InsP3 oscillations. Instead, oscillations are thought to
300 result from the bell‐shaped dependence of the open probability
200 of the InsP3R on Cai2+. Extracellular Ca2+ contributes to Cai2+
100 oscillations in hepatocytes, because Cai2+ oscillations gradually
dissipate in Ca2+‐free medium. Extracellular Ca2+ thus serves to
0
−50 0 50 100 150 200 250 maintain internal Ca2+ stores, which are the primary source of
Time (sec) Ca2+ for oscillations in hepatocytes. The mechanism by which
(b) Ca2+ release from the ER triggers Ca2+ influx through the plasma
250
200
membrane has been described in many cell types including
Fluorescence intensity

150 hepatocytes [30]. The process of store‐operated Ca2+ entry

200 (SOCE) involves the interaction of the integral ER membrane
100
(pixel value)

50
0 100 200 300
protein stromal interacting molecule 1 (STIM1) and its binding
150 Time (sec) partner on the plasma membrane, the ORAI calcium release‐
activated calcium modulator (ORAI) channels [31]. STIM1
100 works as a Ca2+ sensor in the ER; once ER Ca2+ is reduced due
to InsP3‐dependent activation or pharmacological inhibition of
50 the sarco/endoplasmic reticulum Ca2+ ATPase (SERCA) pump,
0 10 20 30 40 STIM1 forms homodimers that interact with and gate ORAI
Relative time (sec) channels on the plasma membrane, leading to influx of Ca2+ into
(c)
250 the cytosol. This Ca2+ influx ultimately assists in replenishing
ER Ca2+ stores.
Fluorescence intensity

200
Increases in Cai2+ in hepatocytes typically begin near the
apical membrane, where InsP3R2 is most concentrated [32].
(pixel value)

Both vasopressin‐ and phenylephrine‐induced Cai2+ waves

150
100
50
0.0 0.5 1.0
Relative time (sec)
Figure 40.4  Different views of cytosolic Ca2+ signaling in hepato-
cytes. In each case, isolated rat hepatocytes were stimulated with the
α1B‐adrenergic agonist phenylephrine. (a) In a population of hepato-
cytes, a single transient peak is observed, followed by a sustained eleva-
tion. (b) In a single hepatocyte, a series of repetitive peaks (oscillations)
are observed. (c) In different regions of the same hepatocyte, the
increase in Ca2+ occurs at different times. This represents a Ca2+ wave.
Notice that these Ca2+ signaling events occur over progressively
shorter  time intervals as the level of focus moves from populations
to  single cells to subcellular regions. Reproduced from [25] with
­permission of Elsevier.
includes single transient or sustained Cai2+ increases and repeti-
tive Cai2+ spikes (i.e. oscillations). For example, lower concen-
trations of vasopressin induce Cai2+ oscillations, while higher
concentrations induce sustained increases in Cai2+ [28], while
stimulation of hepatocytes with phenylephrine typically evokes
Cai2+ oscillations, but the oscillation frequency is dose‐dependent
[28]. The duration of individual Cai2+ spikes also depends upon
the agonist. For example, Cai2+ spikes induced by phenylephrine
are short (around seven seconds) compared to the duration of
spikes induced by vasopressin (around ten seconds) or angio-
tensin (around fifteen seconds). The frequency of vasopressin‐
induced Cai2+ oscillations tends to be greater than that of
began there, then spread in a non‐diminishing fashion across the
cell. Immunofluorescence studies in rat liver, isolated rat hepat-
ocyte couplets, and hepatocytes in collagen sandwich culture
established that InsP3R2 is concentrated subapically [33, 34]
and that this localization depends on intact lipid rafts (special-
1.5
ized membrane patches rich in cholesterol) [13] (Figure 40.2).
Examination of Cai2+ signaling in polarized preparations of
isolated hepatocytes has shown that this is the region where
Cai2+ waves originate, and that this depends on local clustering
of InsP3R2 [13, 32].
Ca2+ signals in cholangiocytes begin in the apical region, just
as they do in hepatocytes, even though InsP3R3 rather than
InsP3R2 is concentrated near their apical surface, where the
Ca2+ waves originate [12].
Spread of Ca2+ signals among hepatocytes
Cai2+ signals occur asynchronously among isolated hepatocytes.
For example, the lag time between stimulation with vasopressin
and initiation of Cai2+ signaling varies among isolated hepato-
cytes by up to several seconds, while the frequency of Ca2+
oscillations can vary by up to 50% among isolated hepatocytes
stimulated with phenylephrine [27]. However, hepatocytes that
communicate via gap junctions coordinate their Cai2+ signals.
For example, stimulation of isolated hepatocyte couplets with
vasopressin induces a single Cai2+ wave that crosses both of the
cells, while stimulation with phenylephrine induces Cai2+ oscil-
lations that are synchronized in the two cells [27]. In the isolated
perfused rat liver, Cai2+ signaling displays an even higher level
of organization, because vasopressin induces Cai2+ waves that
 40: Ca2+ SIGNALING IN THE LIVER 501

(a)
0 sec. 4 sec. 30 sec.

(b)

1.5

[Ca2+]i (F/F0)
1

0 100 200
Time (s)

Figure 40.5  Organization of Ca2+ signals in the intact liver. (a) Serial confocal images of a vasopressin‐induced Ca2+ wave in the isolated perfused
rat liver demonstrates that the increase in Ca2+ begins in the pericentral region, then spreads as a wave toward the periportal regions. Images were
obtained at baseline and after 4 and 30 seconds of stimulation with vasopressin (20 nM). Reproduced from [35] with permission of the American
Physiological Society. (b) Ca2+ signals in individual hepatocytes within an isolated perfused rat liver stimulated with vasopressin. Tracing shows
Ca2+ oscillations detected in three hepatocytes sequentially arranged along the hepatic lobule. The distance between the first (red tracing) and third
(green tracing) cell is 40 μm. The phase difference between the cells is due to the time needed for a Ca2+ wave to spread across the hepatic lobule.
Reproduced from [106] with permission of Elsevier.

cross the entire lobule [35] (Figure 40.5). Cai2+ waves cross indi-
vidual hepatocytes at the same speed, regardless of whether the
hepatocytes are isolated or within the liver. Vasopressin‐induced
Cai2+ waves cross the hepatic lobule in a pericentral‐to‐peripor-
tal direction [36], presumably directed by the V1a vasopressin
receptor gradient present from the pericentral to periportal
region [36]. In contrast, ATP induces Cai2+ signals in a random
fashion across the hepatic lobule, consistent with the lack of a
P2Y receptor gradient across the lobule [37]. Thus, sophisti-
cated patterns of Cai2+ signaling are induced in the intact liver,
and these patterns are agonist‐specific. This may permit differ-
ent Ca2+ agonists to have distinct effects in liver even though
Cai2+ signals induced by these agonists appear similar in isolated
hepatocytes.
The basis for organization of Cai2+ signals in liver has been
studied in multicellular systems of hepatocytes. Studies in iso-
lated rat hepatocyte couplets demonstrate that hepatocytes
communicate via gap junctions, and that both Ca2+ and InsP3
can cross these gap junctions [38]. Hormone‐induced Cai2+ sign-
aling is highly coordinated in such couplets, and this coordi-
nation depends upon gap junction conductance as well [27].
Hepatocytes express two gap junction isoforms, connexin 32
(Cx32) and connexin 26 (Cx26) [39]. Expression of both iso-
forms is dramatically reduced after bile duct ligation, and coor-
dination of Cai2+ signals is impaired under this condition as well
[39]. Furthermore, cell‐to‐cell spread of InsP3 and Cai2+ waves
is markedly impaired in hepatocytes isolated from Cx32 knock-
out mice [40]. Moreover, the expression of Cx32 or Cx43 in a
liver cell line, where intercellular Ca2+ signals are not normally
observed, causes propagation of Ca2+ from cell to cell [41].
Thus, gap junctions play an essential role in organizing Cai2+
signals among adjacent hepatocytes. This also is important for
associated physiological responses. For example, bile secretion
in isolated perfused rat liver is reduced by pharmacological
inhibition of connexin function [42]. Another physiological pro-
cess that requires gap junction communication among hepato-
cytes is glucose output [43] and this is impaired in Cx32
knockout mice stimulated with either glucagon or norepineph-
rine [44]. Cx32 function also is involved in drug‐induced liver
injury, because mice defective in Cx32 are partially protected
from acetaminophen‐induced hepatocyte cell death and liver
failure [45].
Organization of cell‐to‐cell Cai2+ waves depends on other fac-
tors in addition to gap junctions. For example, increases in
InsP3 are required in each cell across which a Ca2+ wave spreads
[46]. Moreover, neither InsP3 nor Ca2+ alone is sufficient to sup-
port the spread of a Cai2+ wave across a hepatocyte [47]. The
presence of agonist binding to its specific receptor at the surface
of the cell also is required to support the spread of Cai2+ waves.
This was demonstrated in experiments in which one or both
cells of a hepatocyte couplet were microperfused with norepi-
nephrine. Stimulation of individual cells evoked Cai2+ oscilla-
tions only in the perfused cell, and perfusion of an entire couplet
was necessary to evoke Cai2+ oscillations in both cells [47].
Thus, the presence of hormone at each cell ensures that suffi-
cient levels of intracellular messengers are generated in order to
reach a level of excitability necessary for supporting the propa-
gation of a Cai2+ wave.
Other studies have focused on the mechanism by which inter-
cellular Cai2+ waves become oriented within the hepatic acinus.
Vasopressin‐induced waves begin in the region of the central
venule [36], while ATP‐induced Cai2+ waves begin in a
502 THE LIVER:  ORGANIZATION OF Ca2+ SIGNALS
seemingly random pattern across the hepatic acinus [35].
Studies in isolated hepatocytes similarly show that pericentral
hepatocytes are more sensitive to vasopressin but not to ATP
[37]. Studies in isolated hepatocyte couplets and triplets stimu-
lated with vasopressin or norepinephrine show that one of the
cells generally has increased sensitivity to a particular hormone
and acts as a pacemaker to drive Cai2+ oscillations in neighbor-
ing cells. This increased sensitivity is likely due to increased
expression of hormone receptor rather than differences in down-
stream signaling components such as G proteins or InsP3R [48].
Cells with increased expression of hormone receptor produce
higher concentrations of InsP3, so they respond sooner than
other cells stimulated with the same concentration of hormone.
As a result, cells with the greatest level of hormone receptor
expression act as pacemakers, and different cells can act as the
pacemaker for different hormones. Thus, the pattern of Cai2+
waves and oscillations in the intact liver depends upon multiple
factors, which include: (i) the establishment of pacemaker cells
by virtue of increased expression of hormone receptors, (ii)
simultaneous stimulation of both pacemaker and non‐pace-
maker cells, and (iii) communication of second messengers
among these cells via gap junctions [49].
The liver also possesses paracrine mechanisms for generating
and regulating Cai2+ signals. Hepatocytes and bile duct cells
each secrete ATP [50, 51], and both cell types express P2Y ATP
receptors [52, 53]. Because P2Y receptors are GPCRs that link
to InsP3‐mediated Cai2+ signaling, secretion of ATP by hepato-
cytes stimulates Cai2+ signaling in neighboring hepatocytes and
bile duct cells [50]. This paracrine signaling mechanism thus
permits increases in Cai2+ to spread among neighboring cells
independent of communication via gap junctions. Hepatocytes
and bile duct cells express P2X receptors as well [54], which are
plasma membrane ATP‐gated Ca2+ channels, but the physiologi-
cal role of these receptors in liver is less clear. Cai2+ signaling in
liver also can be modified rather than initiated by paracrine
pathways. For example, bradykinin does not mobilize Cai2+ in
isolated hepatocytes, yet in the intact liver it modifies the propa-
gation of Cai2+ waves induced by vasopressin [55], likely via
nitric oxide (NO) release from endothelial cells, which diffuses
to hepatocytes where it stimulates generation of cGMP.
Nuclear Ca2+ signaling
The nucleus is separated from the cytosol by the nuclear enve-
lope, which is a specialized region of the ER. Like the ER, the
nuclear envelope is able to store and release Ca2+. The nuclear
envelope accumulates Ca2+ via a Ca2+‐ATPase pump [56], and
releases Ca2+ via channels that are sensitive to InsP3, cADPR, and
NAADP [57, 58]. These Ca2+ storage pumps and release channels
have distinct distributions within the nuclear envelope. The Ca2+‐
ATPase pump is located only in the outer leaflets of the envelope,
while the InsP3R is located only in the inner membrane, and
cADPR‐sensitive channels are present on both sides of the nuclear
envelope. Both InsP3Rs and RyRs are localized along invagina-
tions of the nuclear envelope, denoted the nucleoplasmic reticu-
lum, which work as a regulatory Ca2+ domain within the nucleus
[59]. The importance of nuclear Ca2+ is also highlighted by find-
ings demonstrating that different transcription factors are directly
or indirectly dependent on Ca2+ in the nucleus [60].
Evidence suggests that InsP3 releases Ca2+ directly from the
nuclear envelope into the nucleus. InsP3 releases 45Ca2+ into iso-
lated hepatocyte nuclei, and InsP3 increases free nuclear Ca2+ even
if the nucleus is surrounded by a Ca2+ chelator. In addition, extra-
cellular ATP preferentially activates nuclear Ca2+ release in HepG2
cells, via an InsP3R‐dependent mechanism [61]. The nuclear
envelope possesses the machinery necessary to produce InsP3,
including PIP2 and PLC [62], and this machinery may be activated
selectively through RTK pathways. In one study, IGF‐1 and integ-
rins caused PIP2 breakdown in the nucleus but not at the plasma
membrane, while activation of G protein‐linked receptors caused
breakdown of PIP2 in the cytosol but not the nucleus. Similarly,
activation of the HGF receptor c‐Met in a liver cell line caused
PIP2 breakdown in the nucleus resulting in nuclear Ca2+ signals.
Moreover, activation of this highly localized cascade was depend-
ent on translocation of the activated receptor to the nucleus [3].
Nuclear Ca2+ signaling also occurs by spread of Cai2+ signals
into the nucleus. The nuclear envelope contains pores that are per-
meable to molecules up to 60 KDa in size. In the absence of a
gating mechanism, a pore this size would allow rapid equilibra-
tion of Ca2+ between the nucleus and cytosol. Under certain cir-
cumstances, free diffusion of Ca2+ through the nuclear pore
indeed occurs. However, a nuclear‐cytosolic Ca2+ gradient has
been demonstrated in a number of cell types [63], suggesting that
the permeability of nuclear pores can be regulated. Moreover,
electrophysiological studies suggest that Ca2+ permeability
through nuclear pores is restricted. Atomic force microscopy
studies similarly suggest that nuclear pore permeability is regu-
lated, and that depletion of Ca2+ within the nuclear envelope
closes the pores. Other work using fluorescent dyes or aequorin
also demonstrates that depletion of Ca2+ attenuates the permeabil-
ity of the pores to intermediate‐sized molecules that lack a nuclear
localization sequence. Studies monitoring diffusion of photoacti-
vatable GFP across the nuclear envelope suggest that permeabil-
ity of the nuclear pore may be regulated by cytosolic Ca2+ rather
than by Ca2+ stores within the nuclear envelope [64]. EF‐hand
Ca2+ binding motifs are present in proteins of the nuclear pore, so
it is possible that these function as Ca2+‐gating sensors for the
pore. There is additional evidence that nuclear Ca2+ sometimes
passively follows Cai2+ [65]. For example, stimulation of hepato-
cytes with vasopressin results in a Ca2+ wave that appears to
spread from the cytosol into the nucleus. Moreover, a mathemati-
cal analysis of Ca2+ waves in hepatocytes stimulated with vaso-
pressin suggests that nuclear Ca2+ signals can be described simply
by diffusion of Ca2+ inward from the nuclear envelope.
Nuclear Ca2+ also can contribute to Ca2+ signals in the cyto-
sol. For example, localized increases of Ca2+ in the cytosol (Ca2+
puffs) can spread across the cell by diffusing across the nucleus.
Ca2+ puffs are highly transient and localized Cai2+ signals that
result from the coordinated opening of small clusters of InsP3Rs.
Puffs can be triggered by a subthreshold concentration of ago-
nist and the resulting Cai2+ signal rapidly dissipates by diffusion
in the cytosol and sequestration of Ca2+ into intracellular stores.
However, the range of diffusion of Ca2+ in the nucleus can be
much greater than in cytosol, so Ca2+ puffs generated near the
nuclear envelope can spread into and across the nucleus in order
to spread to other, more distant regions of the cytosol [65].
Thus, the nucleus may function as a tunnel that helps distribute
Ca2+ to the cytosol.
 40: Ca2+ SIGNALING IN THE LIVER 503
Ca2+ signaling in bile duct cells
Cai2+ signaling has been examined to a lesser extent in bile duct
epithelia than in hepatocytes. ATP and UTP both increase Cai2+
in the Mz‐ChA‐1 cholangiocarcinoma cell line, a model for bile
duct epithelium. ATP and UTP also increase Cai2+ in primary cul-
tures of rat bile duct epithelia, and acetylcholine (ACh) increases
Cai2+ in these cells via M3 muscarinic receptors [52]. As in
hepatocytes and other epithelia, the range of patterns of agonist‐
induced Cai2+ signals include sustained and transient Cai2+
increases and Cai2+ oscillations. Cai2+ spikes induced by ACh are
longer in duration and lower in frequency than those induced by
ATP [52]. Cai2+ signaling is mediated by InsP3 in bile duct cells,
because increases in Cai2+ are blocked by InsP3R antagonists. As
discussed above, all three InsP3R isoforms are expressed in these
cells [12]; Ca2+ waves begin in their apical region, where InsP3R3
is concentrated, and then spread basolaterally via the types I and
II InsP3Rs. Bile duct cells are coupled via connexin 43 (Cx43)
gap junctions, and expression of Cx43 synchronizes their Ca2+
oscillations. Moreover, Cx43 permeability is under hormonal
control, and activation of either protein kinase A or C decreases
permeability and impairs intercellular communication [66].
EFFECTS OF Ca2+ SIGNALS IN HEALTH
AND DISEASE
Ca2+ regulates a wide range of functions in liver. The following
sections are meant to provide illustrative examples of the differ-
ent ways in which Ca2+ regulates normal and abnormal liver
function, rather than to provide an exhaustive list of Ca2+‐medi-
ated functions.
Energy metabolism
Storage and release of glucose was among the first functions of
the liver shown to be regulated by Cai2+. Synthesis of glycogen
is regulated by glycogen synthase, while phosphorylase is the
rate‐limiting enzyme for glycogenolysis. Both enzymes are
regulated by phosphorylation and dephosphorylation, and an
increase in Cai2+ is one of the most important signals for regulat-
ing these events [67]. For example, hormones such as vasopres-
sin and angiotensin increase InsP3 in hepatocytes, which
mobilize Ca2+, leading to phosphorylation and activation of gly-
cogen phosphorylase, and then glycogenolysis. Similarly, extra-
cellular nucleotides activate glycogen phosphorylase and thus
stimulate glycogenolysis by binding to P2Y nucleotide recep-
tors [68]. Glucagon and beta‐adrenergic agonists also stimulate
glycogen phosphorylase activity in liver, but through an alterna-
tive, cAMP‐dependent pathway. Ca2+‐mobilizing bile acids such
as UDCA, TLCA, and LCA activate phosphorylase to the same
extent as hormones such as vasopressin. These bile acids acti-
vate phosphorylase through a Ca2+‐dependent but InsP3‐inde-
pendent mechanism, consistent with the observation that they
increase Cai2+ in an InsP3‐independent fashion [69].
Gluconeogenic enzymes are preferentially located in the per-
iportal region [70], although other factors also may be involved
in regional differences in glycogenolytic capacity. ATP
mobilizes glucose mainly from the periportal zone, while nor-
epinephrine and vasopressin preferentially release glucose from
pericentral hepatocytes. This may in part reflect the fact that
pericentral hepatocytes are more sensitive than periportal hepat-
ocytes to vasopressin and norepinephrine [37]. When hepato-
cytes are not uniformly sensitive to a particular hormone, then
intercellular communication via gap junctions enhances glucose
release. For example, glucose release is impaired in perfused
livers from Cx32‐deficient mice upon stimulation with either
norepinephrine or glucagon [44]. Similarly, vasopressin‐ or
glucagon‐induced glucose release is impaired in perfused rat
livers treated with the gap junction blocker 18α‐glycyrrhetinic
acid (αGA) [42]. However, glucose release is not altered in
αGA‐treated livers stimulated with dibutyryl cAMP or 2,5‐
di(tert‐butyl)‐1,4‐benzohydroquinone (tBuBHQ), both of which
stimulate glucose release in a receptor‐independent fashion
[42]. Hormone‐induced glucose release also is impaired if gap
junctions are blocked in isolated rat hepatocytes, or if hepato-
cytes are dispersed. Therefore, hepatocytes may contribute dif-
ferently to glucose metabolism across the hepatic lobule,
although there is some integration of metabolic activity via gap
junctions. This integration of metabolic function is particularly
important in times of stress, because fasting induces hypoglyce-
mia in Cx32‐deficient but not wild‐type mice, and because
endotoxin‐induced hypoglycemia is exacerbated in the knock-
out mice as well [40]. Similarly, liver glucose production upon
stimulation with glucagon and norepinephrine is reduced in
Cx32 knockout mice [44]. Glucose output by the liver is also
regulated by Ca2+‐calmodulin kinase II gamma (CamKIIγ), a
downstream effector of intracellular Ca2+ signaling [71]. In a
physiological context, this kinase is activated by InsP3R‐
dependent Ca2+ release during fasting and promotes transloca-
tion of the transcription factor FoxO1 to the nucleus of
hepatocytes. This translocation in turn controls a transcriptional
program that potentiates glycogenolysis and gluconeogenesis
and leads to increased glucose output by the liver. This mecha-
nism is potentiated in experimental obesity and contributes to
the high serum levels of glucose found in obese mice.
Accordingly, genetic deletion of CamKIIγ or deletion of FoxO1
in mice are associated with lower blood concentrations of glu-
cose. A second Ca2+‐dependent transcription program that mod-
ulates glucose production by the liver is operated by the CREB
regulated transcription coactivator 2 (CRTC2) [72]. Here,
glucagon activates cAMP formation, PKA activation, and phos-
phorylation of InsP3R in hepatocytes. This phosphorylation
renders the receptor more susceptible to activation by InsP3 and
thus Ca2+ is more readily released in the cytosol. Free Ca2+ can
then bind to calcineurin which dephosphorylates CRTC2 allow-
ing its translocation to the nucleus. The final result is the activa-
tion of genes that promote glucose production. Termination of
this mechanism is mainly by insulin‐dependent activation of
Akt. In diabetes, insulin resistance thus ensures prolonged acti-
vation of the CRTC2 and sustained glucose output by the liver.
CREB also increases transcription and thus expression of
InsP3R2 [73]. Hepatic InsP3R2 expression also increases dur-
ing fasting, and it has been postulated that this occurs via gluca-
gon‐ and cAMP‐mediated activation of CREB [73].
Ca2+ signaling also plays a role in lipid metabolism and the
pathogenesis of non‐alcoholic fatty liver disease (NAFLD).
504 THE LIVER:  EFFECTS OF Ca2+ SIGNALS IN HEALTH AND DISEASE
Relevant studies mainly describe changes in expression of Ca2+
handling proteins, such as intracellular channels and pumps, and
their effect on lipid metabolism. Studies in obese mice showed
that expression of Serca2b is specifically downregulated in the
liver [74]. The physiological function of this sarco(endo)plasmic
reticulum Ca2+‐ATPase is to maintain high levels of Ca2+ within
the ER, which are required for proper protein folding and secre-
tion. This is achieved by active transport of free Ca2+ from the
cytosol to the ER cisternae. In experimental obesity, the down-
regulation of Serca2b leads to reduced ER Ca2+ content and ER
stress, which collectively activates glucose production and trig-
gers a lipogenic gene expression program. Conversely, re‐expres-
sion of Serca2b in obese mice restores ER protein folding and
improves both glucose and fatty acid metabolism.
InsP3Rs have also been implicated in lipid processing and
droplet formation in hepatocytes. Liver specific InsP3R1 knock-
out (LSKO1) mice are resistant to the development of diet‐
induced liver steatosis and this effect is due to impaired Ca2+
transfer from ER to mitochondria [18]. Similarly, mice fed a high‐
fat diet (HFD) or genetically obese mice display a reorganization
of ER‐mitochondria contact sites [75]. These contact regions are
known as mitochondrial‐associated membranes (MAMs) and
facilitate the exchange of phospholipids and transmission of Ca2+
between the two organelles. In obesity, the overall percentage of
MAMs is increased, leading to InsP3R1‐dependent Ca2+ overload
in mitochondria and formation of reactive oxygen species (ROS)
and mitochondrial dysfunction. Defects in SOCE also occur in
experimental fatty liver disease. In obese mice, STIM1 transloca-
tion to the vicinity of the plasma membrane is impaired, which
results in markedly diminished SOCE. These alterations in SOCE
in turn are associated with lipid droplet formation and lipid toxic-
ity in hepatocytes [76]. Overall, pathological lipid accumulation
in hepatocytes and alterations in the expression of Ca2+ signaling
proteins are interrelated, and there is evidence that this occurs in
human disease as well. In particular, there is a progressive
increase in the colocalization between ER and mitochondria in
hepatocytes from normal liver to simple steatosis to non‐alcoholic
steatohepatitis (NASH). There also is increased expression of
InsP3R1, which is responsible for mitochondrial Ca2+ signals, in
liver biopsies from patients with NASH [18].
Bile flow
Cai2+ regulates fluid and electrolyte secretion in many types of
epithelia [25]. This is mediated in part by polarized Cai2+ waves,
which also occur in hepatocytes [77]. Cai2+ has multiple effects
on bile flow. Early studies in isolated perfused rat liver showed
that the net effect of Cai2+ on bile acid‐independent bile flow is
inhibitory [78], but more recent studies demonstrate that both
bile acid‐independent and bile acid‐dependent flow are potenti-
ated by Ca2+ signaling in hepatocytes [33, 34]. First, the mem-
brane localization and activity of the main apical bile‐acid
independent solute transporter, the multidrug resistance‐associ-
ated protein 2 (Mrp2) is potentiated by Ca2+ agonists in in vitro
systems. Moreover, bile flow is decreased in InsP3R2 knockout
animals suggesting that Ca2+ release near the apical membrane
is important for bile secretion [33]. Total internal reflection flu-
orescence (TIRF) microscopy studies furthermore suggest that
periapical Ca2+ signals induce Mrp2‐containing vesicles to fuse
with the plasma membrane [33]. Bile‐salt dependent flow,
which relies on the activity of the bile salt export pump (Bsep),
is similarly decreased in rat hepatocytes with reduced expres-
sion of InsP3R2 or in cells treated with an intracellular Ca2+
buffering agent [34]. Moreover, disruption of lipid rafts in the
apical membrane induces InsP3R2 to migrate away from the
periapical region, impairs Ca2+ signals, and reduces secretion
[13, 34]. Finally, InsP3R2 expression is decreased and the
remaining receptor moves from the apical region in two separate
models of cholestasis induced by estrogen and endotoxin [34].
Collectively, these studies suggest that subapical Ca2+ signals in
hepatocytes potentiate secretion of bile solutes.
Bile acids such as lithocholic acid (LCA), taurolithocholic acid
(TLCA), taurodeoxycholic acid and taurochenodeoxycholic, as
well as the therapeutic bile acids ursodeoxycholic acid (UDCA)
and tauroursodeoxycholic acid (TUDCA) increase intracellular
Ca2+ in hepatocytes [69, 79]. LCA and TLCA are cholestatic, but
UDCA and TUDCA are choleretic. In fact, TUDCA can reverse
the cholestatic effect of TLCA, and this depends on InsP3R2 [33].
UDCA and TUDCA also induce hepatocytes to secrete ATP,
which may relate to their therapeutic effect [80].
Cell proliferation
Ca2+ signals are associated with progression through the cell
cycle [81]. In sea urchin embryos, there are two prominent cell
cycle‐related Ca2+ signals. The first Ca2+ transient occurs just
prior to entry into mitosis, and the second one occurs during the
metaphase–anaphase transition. Intracellular injection of Ca2+
chelators such as BAPTA or an InsP3R antagonist such as hepa-
rin abolish these Ca2+ signals and prevent entry into mitosis.
Introduction of InsP3 or Ca2+ in the cytosol has the opposite
effect, to accelerate entry into mitosis.
Ca2+ transients are also observed during cell cycle progres-
sion in somatic cells, although the relationship between these
Ca2+ signals and progression through the cell cycle is less estab-
lished. In Swiss 3T3 cells, serum withdrawal suppresses these
Ca2+ transients but does not affect progression though mitosis.
However, mitosis is blocked by specific inhibition of Ca2+ tran-
sients through injection of BAPTA plus incubation in calcium‐
free media. Moreover, photorelease of caged Ca2+ induces
premature entry into mitosis.
Downstream targets of Ca2+ have also been implicated in cell
cycle progression. Calcineurin is essential for Xenopus laevis
embryonic development [82]. In addition, pharmacological
inhibition of Ca2+/calmodulin kinase II (CaMKII) arrests cells at
the G2/M transition. Moreover, calmodulin overexpression
accelerates the cell cycle in mouse C127 cells and its downregu-
lation extends the cell cycle.
Heterologous expression of the Ca2+ binding protein parval-
bumin has also been used to study Ca2+ signaling in the regula-
tion of the cell cycle. This protein is normally expressed in
skeletal muscle and neurons; in myocytes it modulates relaxa-
tion of fast twitch muscle fibers due to its Ca2+ buffering capac-
ity. The first report using this protein as a molecular tool showed
that buffering Ca2+ slowed progression through the cell cycle in
mouse C127 cells. More recently, parvalbumin variants targeted
to the nucleus or the cytoplasm [60] were used to investigate the
relative importance of Ca2+ signals in each of these cellular
 40: Ca2+ SIGNALING IN THE LIVER 505
compartments for regulation of the cell cycle in a liver cell line.
It was found that nucleoplasmic rather than cytosolic Ca2+ is
essential for cellular proliferation, and is necessary in particular
for progression through early prophase [83]. Recent findings
suggest that HGF and insulin, two potent growth factors in liver,
selectively form InsP3 in the nucleus to initiate nuclear Ca2+ sig-
nals [3, 84]. These findings suggest that certain growth factors
may stimulate proliferation of hepatocytes by selectively induc-
ing Ca2+ signals in the nucleus.
A common model of hepatocyte proliferation in vivo is liver
regeneration after partial hepatectomy. Following 70% hepatec-
tomy, the liver undergoes a coordinated regenerative response
involving all hepatic cell types, but this is punctuated by marked
proliferation of hepatocytes. In rats, 24 hours after partial hepatec-
tomy there is a decrease in InsP3R2 expression accompanied by a
decrease in the frequency of Ca2+ oscillations. This decrease in
receptor expression normalizes within 4 days and is then associ-
ated with an increase in the frequency of Ca2+ oscillations. This
remodeling of the Ca2+ signaling machinery is thought to be essen-
tial for the initial regenerative response [85]. However, complete
loss of InsP3R2 results in delayed liver regeneration and also
impaired Ca2+ signaling in the hepatocyte nucleus [86]. Fatty liver
causes a c‐Jun mediated decrease in InsP3R2 expression in hepat-
ocytes [86], which may contribute in part to the impaired liver
regeneration seen in patients with NAFLD. At least three growth
factors, HGF, EGF, and insulin, mobilize Cai2+ in hepatocytes and
contribute to liver regeneration [87–89]. Liver regeneration after
partial hepatectomy is significantly delayed in rats expressing par-
valbumin in the cytosol of hepatocytes [90]. Buffering of InsP3 in
the nucleus of hepatocytes also delays liver regeneration [89], sug-
gesting that RTK ‐mediated, InsP3‐dependent Ca2+ signals in the
nucleus are essential for proper hepatocyte proliferation. On the
other hand, liver regeneration is enhanced by expression of parval-
bumin in mitochondria, because this mitigates apoptosis that is
occurring concomitant with proliferation [20]. Therefore, the rate
of liver regeneration is determined by a careful balance between
Ca2+ signals in the cytosol, nucleus, and mitochondria.
Ca2+‐dependent processes have been associated with progres-
sion of hepatocellular carcinoma (HCC). For instance, expression
of Ca2+/calmodulin‐dependent kinase kinase II (CamKKII) is
upregulated in liver cancer and its experimental downregulation
or pharmacological inhibition reduce tumor growth in a mouse
model of HCC [91]. Also, overexpression of hepatitis B virus X
protein (HBx), a known oncogenic viral protein, promotes liver
cancer cell invasion and metastasis through a Ca2+‐mediated
increase in secretion of HMGB1 [92]. The role of intracellular
Ca2+ channels and downstream targets of Ca2+ signaling in the
pathogenesis of liver cancer is a topic of active investigation.
Ductular secretion
Cai2+ directly regulates fluid and electrolyte secretion in bile
duct epithelia. Cholangiocytes express the apical, Ca2+‐activated
Cl− channel TMEM16a [93], which is the primary mechanism
for Ca2+‐stimulated secretion in these cells. The other principal
mechanism for regulating cholangiocyte secretion is via the api-
cal, cAMP‐activated, cystic fibrosis transmembrane conductance
regulator (CFTR) Cl− channel [51]. Together, these two mecha-
nisms create the Cl− gradient responsible for driving anion
exchanger 2 (AE2)‐mediated HCO3−/Cl− exchange that is respon-
sible for secreting bicarbonate into the bile. Ca2+ potentiates ade-
nylyl cyclase activity and cAMP formation in cholangiocytes via
a calcineurin‐dependent pathway [94] and indirectly through
cAMP formation induced by SOCE [95]. This crosstalk between
the Ca2+ and cAMP signaling pathways has been considered a
secondary, indirect mechanism for Ca2+‐stimulated secretion.
However, cAMP‐ and CFTR‐mediated secretion may rely on Ca2+
in a more direct way as well [96]. Activation of CFTR in cholan-
giocytes is linked to exocytosis of vesicles rich in ATP [97], which
releases ATP into the bile, resulting in stimulation of apical P2Y
receptors [51, 96]. Activation of these P2Y receptors links to
GPCRs and InsP3 formation and then Ca2+ release from apical,
InsP3R3, which in turn drives Ca2+‐dependent Cl− channels and
then HCO3− secretion [96]. Conversely, neurotransmitters such
as acetylcholine stimulate basolateral M3 muscarinic receptors
on cholangiocytes, leading to Ca2+ release from InsP3R1 and
InsP3R2. This Ca2+ similarly activates apical Ca2+‐dependent Cl−
channels, which ultimately stimulate biliary HCO3− secretion.
Peptide hormones such as secretin stimulate basolateral secretin
receptors, which induce the formation of cAMP, so this pathway
serves to activate CFTR leading to paracrine, ATP‐mediated
HCO3− secretion. This putative universal role of InsP3R‐mediated
Ca2+ signaling in ductular secretion is in agreement with the
observation that loss of InsP3R3 is a common molecular event
in cholestatic disorders including primary biliary cholangitis
(PBC), primary sclerosing cholangitis (PSC), extrahepatic biliary
obstruction, and biliary atresia [98]. Factors that contribute to this
specific loss of InsP3R3 in biliary diseases include the transcrip-
tion factors Nrf2 (nuclear factor erythroid 2‐related factor 2) and
NF‐κB, plus the microRNA miR‐506. Nrf2 is activated during
conditions of oxidative stress and specifically downregulates
InsP3R3 in a human cholangiocyte cell line. Accordingly, it is
translocated to the nucleus of bile duct cells in a range of ductular
diseases including PBC and PSC [99]. The NF‐κB pathway is
activated by endotoxin‐induced stimulation of TLR4, and this
mechanism has been implicated in loss of InsP3R3 from cholan-
giocytes in patients with sepsis or alcoholic hepatitis [100]. The
specific loss of InsP3R3 in PBC might also be mediated in part by
miR‐506, because this microRNA decreases InsP3R3 expression
in a human cholangiocyte cell line and is increased in bile ducts
in liver biopsies from PBC patients [101].
Cholangiocytes express primary cilia on their apical mem-
branes, and this sensory organelle also links to Ca2+ signaling
and secretion. For example, primary cilia can sense and trans-
duce mechanical stimuli from within the bile duct lumen, so that
increases in bile flow activate both cAMP production and Ca2+
release [102]. Cholangiocyte cilia also sense increases in osmo-
larity within the bile duct lumen, and this links to activation of
TRPV4 channels, which allow entry of extracellular Ca2+ into
the cytoplasm [103]. The role of cilia and ciliary Ca2+ signaling
in cholestatic disorders remains an area of investigation.
The GPCR for bile acids (TGR5) also is expressed at the api-
cal membrane of cholangiocytes and it is enriched in cilia of
rodent and human bile duct cells [104]. TGR5 activation cou-
ples to the formation of cAMP in cholangiocytes, so pharmaco-
logical agonists of this receptor are currently under study as
choleretic agents in a variety of acquired and genetic cholestatic
disorders.
506 THE LIVER:  REFERENCES
ACKNOWLEDGMENTS
This work was supported by NIH grants DK57751, DK34989,
DK112797, and DK114041, and by grants from CNPq and
FAPEMIG.
REFERENCES
1. Berridge, M.J., Bootman, M.D., and Roderick, H.L. Calcium signalling:
dynamics, homeostasis and remodelling. Nat Rev Mol Cell Biol,
2003;4(7):517–29.
2. Rhee, S.G. and Choi, K.D. Regulation of inositol phospholipid‐specific
phospholipase C isozymes. J Biol Chem, 1992;267:12393–6.
3. Gomes, D.A., Rodrigues, M.A., Leite, M.F. et al. c‐Met must translocate to
the nucleus to initiate calcium signals. J Biol Chem, 2008;283(7):4344–51.
4. Futatsugi, A., Nakamura, T., Yamada, M.K. et  al. IP3 receptor types 2
and  3  mediate exocrine secretion underlying energy metabolism. Science,
2005;309(5744):2232–4.
5. Chandrasekhar, R., Alzayady, K.J., Wagner, L.E., 2nd, and Yule, D.I. Unique
regulatory properties of heterotetrameric inositol 1,4,5‐trisphosphate recep-
tors revealed by studying concatenated receptor constructs. J Biol Chem,
2016;291(10):4846–60.
6. Choe, C.U. and Ehrlich, B.E. The inositol 1,4,5‐trisphosphate receptor
(IP3R) and its regulators: sometimes good and sometimes bad teamwork. Sci
STKE, 2006;2006(363):re15.
7. Bosanac, I., Alattia, J.R., Mal, T.K. et al. Structure of the inositol 1,4,5‐tris-
phosphate receptor binding core in complex with its ligand. Nature,
2002;420(6916):696–700.
8. Hagar, R.E., Burgstahler, A.D., Nathanson, M.H., and Ehrlich, B.E. Type III
InsP3 receptor channel stays open in the presence of increased calcium.
Nature, 1998;396(6706):81–4.
9. Bimboese, P., Gibson, C.J., Schmidt, S., Xiang, W., and Ehrlich, B.E.
Isoform‐specific regulation of the inositol 1,4,5‐trisphosphate receptor by
O‐linked glycosylation. J Biol Chem, 2011;286(18):15688–97.
10. Dufour J‐F., Luthi, M., Forestier, M., and Magnino, F. Expression of inositol
1,4,5‐trisphosphate receptor isoforms in rat cirrhosis. Hepatology, 1999;30:
1018–26.
11. Hirata, K., Pusl, T., O’Neill A.F., Dranoff, J.A., and Nathanson, M.H. The
type II inositol 1,4,5‐trisphosphate receptor can trigger Ca2+ waves in rat
hepatocytes. Gastroenterology, 2002;122(4):1088–100.
12. Hirata, K., Dufour, J.F., Shibao, K. et al. Regulation of Ca(2+) signaling in
rat bile duct epithelia by inositol 1,4,5‐trisphosphate receptor isoforms.
Hepatology, 2002;36(2):284–96.
13. Nagata, J., Guerra, M.T., Shugrue, C.A., Gomes, D.A., Nagata, N., and
Nathanson, M.H. Lipid rafts establish calcium waves in hepatocytes.
Gastroenterology, 2007;133(1):256–67.
14. Pierobon, N., Renard‐Rooney, D.C., Gaspers, L.D., and Thomas, A.P.
Ryanodine receptors in liver. J Biol Chem, 2006;281(45):34086–95.
15. Zhang, F. and Li, P.L. Reconstitution and characterization of a nicotinic acid
adenine dinucleotide phosphate (NAADP)‐sensitive Ca2+ release channel
from liver lysosomes of rats. J Biol Chem, 2007;282(35):25259–69.
16. De Stefani, D., Rizzuto, R., and Pozzan, T. Enjoy the trip: calcium in mito-
chondria back and forth. Annu Rev Biochem, 2016;85:161–92.
17. Antony, A.N., Paillard, M., Moffat, C. et al. MICU1 regulation of mitochon-
drial Ca(2+) uptake dictates survival and tissue regeneration. Nat Comm,
2016;7:10955.
18. Feriod, C.N., Oliveira, A.G., Guerra, M.T. et al. Hepatic inositol 1,4,5 tris-
phosphate receptor type 1 mediates fatty liver. Hepatol Commun,
2017;1(1):23–35.
19. White, C., Li, C., Yang, J. et al. The endoplasmic reticulum gateway to apopto-
sis by Bcl‐X(L) modulation of the InsP3R. Nat Cell Biol, 2005;7(10):1021–8.
20. Guerra, M.T., Fonseca, E.A., Melo, F.M. et al. Mitochondrial calcium regu-
lates rat liver regeneration through the modulation of apoptosis. Hepatology,
2011;54(1):296–306.
21. Minagawa, N., Kruglov, E.A., Dranoff, J.A., Robert, M.E., Gores, G.J., and
Nathanson, M.H. The anti‐apoptotic protein Mcl‐1 inhibits mitochondrial
Ca2+ signals. J Biol Chem, 2005;280(39):33637–44.
22. Csordas, G., Renken, C., Varnai, P. et al. Structural and functional features
and significance of the physical linkage between ER and mitochondria. J
Cell Biol, 2006;174(7):915–21.
23. Hajn¢czky, G., Hager, R., and Thomas, A.P. Mitochondria suppress local
feedback activation of inositol 1,4,5‐trisphosphate receptors by Ca2+. J Biol
Chem, 1999;274:14157–62.
24. Robb‐Gaspers, L.D., Burnett, P., Rutter, G.A., Denton, R.M., Rizzuto, R.,
and Thomas, A.P. Integrating cytosolic calcium signals into mitochondrial
metabolic responses. EMBO J, 1998;17:4987–5000.
25. Nathanson, M.H. Cellular and subcellular calcium signaling in gastrointesti-
nal epithelium. Gastroenterology, 1994;106:1349–64.
26. Zhao, Y., Araki, S., Wu, J. et al. An expanded palette of genetically encoded
Ca(2)(+) indicators. Science, 2011;333(6051):1888–91.
27. Nathanson, M.H. and Burgstahler, A.D. Coordination of hormone‐induced
calcium signals in isolated rat hepatocyte couplets: demonstration with con-
focal microscopy. Mol Biol Cell, 1992;3:113–21.
28. Rooney, T.A., Sass, E.J., and Thomas, A.P. Characterization of cytosolic cal-
cium oscillations induced by phenylephrine and vasopressin in single Fura‐2
loaded hepatocytes. J Biol Chem, 1989;264:17131–41.
29. Dolmetsch, R.E., Xu, K., and Lewis, R.S. Calcium oscillations increase the
efficiency and specificity of gene expression. Nature, 1998;392(6679):933–6.
30. Jones, B.F., Boyles, R.R., Hwang, S.Y., Bird, G.S., and Putney, J.W. Calcium
influx mechanisms underlying calcium oscillations in rat hepatocytes.
Hepatology, 2008;48(4):1273–81.
31. Feske, S., Skolnik, E.Y., and Prakriya, M. Ion channels and transporters in
lymphocyte function and immunity. Nat Rev Immunol, 2012;12(7):532–47.
32. Hernandez, E., Leite, M.F., Guerra, M.T. et  al. The spatial distribution of
inositol 1,4,5‐trisphosphate receptor isoforms shapes Ca2+ waves. J Biol
Chem, 2007;282(13):10057–67.
33. Cruz, L.N., Guerra, M.T., Kruglov, E. et al. Regulation of multidrug resist-
ance‐associated protein 2 by calcium signaling in mouse liver. Hepatology,
2010;52(1):327–37.
34. Kruglov, E.A., Gautam, S., Guerra, M.T., and Nathanson, M.H. Type 2 ino-
sitol 1,4,5‐trisphosphate receptor modulates bile salt export pump activity in
rat hepatocytes. Hepatology, 2011;54(5):1790–9.
35. Motoyama, K., Karl, I.E., Flye, M.W., Osborne, D.F., and Hotchkiss, R.S.
Effect of Ca2+ agonists in the perfused liver: determination via laser scan-
ning confocal microscopy. Am J Physiol Regul Integr Comp Physiol,
1999;276:R575–85.
36. Nathanson, M.H., Burgstahler, A.D., Mennone, A., Fallon, M.B., Gonzalez,
C.B., and Saez, J.C. Ca2+ waves are organized among hepatocytes in the
intact organ. Am J Physiol, 1995;269(1 Pt 1):G167–71.
37. Tordjmann, T., Berthon, B., Combettes, L., and Claret, M. The location of
hepatocytes in the rat liver acinus determines their sensitivity to calcium‐
mobilizing hormones. Gastroenterology, 1996;111:1343–52.
38. Saez, J.C., Connor, J.A., Spray, D.C., and Bennett, M.V.L. Hepatocyte gap
junctions are permeable to the second messenger, inositol 1,4,5‐triphosphate,
and to calcium ions. Proc Natl Acad Sci USA, 1989;86:2708–12.
39. Fallon, M.B., Nathanson, M.H., Mennone, A., Sáez J.C., Burgstahler, A.D.,
and Anderson, J.M. Altered expression and function of hepatocyte gap junc-
tions after common bile duct ligation in the rat. Am J Physiol Cell Physiol,
1995;268:C1186–94.
40. Correa, P.R., Guerra, M.T., Leite, M.F., Spray, D.C., and Nathanson, M.H.
Endotoxin unmasks the role of gap junctions in the liver. Biochem Biophys
Res Commun, 2004;322(3):718–26.
41. Leite, M.F., Hirata, K., Pusl, T. et al. Molecular basis for pacemaker cells in
epithelia. J Biol Chem, 2002;277:16313–23.
42. Nathanson, M.H., Rios‐Velez, L., Burgstahler, A.D., and Mennone, A.
Communication via gap junctions modulates bile secretion in the isolated
perfused rat liver. Gastroenterology, 1999;116:1176–83.
43. Bartlett, P.J., Gaspers, L.D., Pierobon, N., and Thomas, A.P. Calcium‐
dependent regulation of glucose homeostasis in the liver. Cell Calcium,
2014;55(6):306–16.
44. Stumpel, F., Ott, T., Willecke, K., and Jungermann, K. Connexin 32 gap
junctions enhance stimulation of glucose output by glucagon and noradrena-
line in mouse liver. Hepatology, 1998;28:1616–20.
45. Patel, S.J., Milwid, J.M., King, K.R. et al. Gap junction inhibition prevents
drug‐induced liver toxicity and fulminant hepatic failure. Nat Biotechnol,
2012;30(2):179–83.
46. Boitano, S., Dirksen, E.R., and Sanderson, M.J. Intercellular propagation of
calcium waves mediated by inositol trisphosphate. Science, 1992;258:292–5.
 40: Ca2+ SIGNALING IN THE LIVER 507
47. Tordjmann, T., Berthon, B., Claret, M., and Combettes, L. Coordinated
­intercellular calcium waves induced by noradrenaline in rat hepatocytes:
dual control by gap junction permeability and agonist. EMBO J, 1997;16:
5398–407.
48. Tordjmann, T., Berthon, B., Jacquemin, E. et al. Receptor‐oriented intercel-
lular calcium waves evoked by vasopressin in rat hepatocytes. EMBO J,
1998;17:4695–703.
49. Burgstahler, A.D. and Nathanson, M.H. Coordination of calcium waves among
hepatocytes: teamwork gets the job done. Hepatology, 1998;27:634–5.
50. Schlosser, S.F., Burgstahler, A.D., and Nathanson, M.H. Isolated rat hepato-
cytes can signal to other hepatocytes and bile duct cells by release of nucleo-
tides. Proc Natl Acad Sci USA, 1996;93:9948–53.
51. Fiorotto, R., Spirli, C., Fabris, L., Cadamuro, M., Okolicsanyi, L., and
Strazzabosco, M. Ursodeoxycholic acid stimulates cholangiocyte fluid
secretion in mice via CFTR‐dependent ATP secretion. Gastroenterology,
2007;133(5):1603–13.
52. Nathanson, M.H., Burgstahler, A.D., Mennone, A., and Boyer, J.L.
Characterization of cytosolic Ca2+ signaling in rat bile duct epithelia. Am J
Physiol Gastrointest Liver Physiol, 1996;271:G86–96.
53. Kitamura, T., Brauneis, U., Gatmaitan, Z., and Arias, I.M. Extracellular ATP,
intracellular calcium and canalicular contraction in rat hepatocyte doublets.
Hepatology, 1991;14:640–7.
54. Gonzales, E., Prigent, S., Abou‐Lovergne, A. et al. Rat hepatocytes express
functional P2X receptors. FEBS Lett, 2007;581(17):3260–6.
55. Patel, S., Robb‐Gaspers, L.D., Stellato, K.A., Shon, M., and Thomas, A.P.
Coordination of calcium signalling by endothelial‐derived nitric oxide in the
intact liver. Nat Cell Biol, 1999;1:467–71.
56. Lanini, L., Bachs, O., and Carafoli, E. The calcium pump of the liver nuclear
membrane is identical to that of endoplasmic reticulum. J Biol Chem,
1992;267(16):11548–52.
57. Gerasimenko, J.V., Maruyama, Y., Yano, K. et al. NAADP mobilizes Ca2+
from a thapsigargin‐sensitive store in the nuclear envelope by activating
ryanodine receptors. J Cell Biol, 2003;163(2):271–82.
58. Gerasimenko, O.V., Gerasimenko, J.V., Tepikin, A.V., and Petersen, O.H.
ATP‐dependent accumulation and inositol trisphosphate‐ or cyclic ADP‐
ribose‐mediated release of Ca2+ from the nuclear envelope. Cell,
1995;80:439–44.
59. Malhas, A., Goulbourne, C., and Vaux, D.J. The nucleoplasmic reticulum:
form and function. Trends Cell Biol, 2011;21(6):362–73.
60. Pusl, T., Wu, J.J., Zimmerman, T.L. et al. Epidermal growth factor‐mediated
activation of the ETS domain transcription factor Elk‐1 requires nuclear cal-
cium. J Biol Chem, 2002;277(30):27517–27.
61. Leite, M.F., Thrower, E.C., Echevarria, W. et al. Nuclear and cytosolic cal-
cium are regulated independently. Proc Natl Acad Sci USA,
2003;100(5):2975–80.
62. Divecha, N., Rhee S‐G., Letcher, A.J., and Irvine, R.F. Phosphoinositide sig-
nalling enzymes in rat liver nuclei: phosphoinositidase C isoform ·1 is spe-
cifically, but not predominantly, located in the nucleus. Biochem J,
1993;289:617–20.
63. Waybill, M.M., Yelamarty, R.V., Zhang, Y. et al. Nuclear calcium gradients
in cultured rat hepatocytes. Am J Physiol Endocrinol Metab,
1990;261:E49–57.
64. O’Brien E.M., Gomes, D.A., Sehgal, S., and Nathanson, M.H. Hormonal
regulation of nuclear permeability. J Biol Chem, 2007;282(6):4210–7.
65. Lipp, P., Thomas, D., Berridge, M.J., and Bootman, M.D. Nuclear calcium
signalling by individual cytoplasmic calcium puffs. EMBO J,
1997;16:7166–73.
66. Bode, H.P., Wang, L., Cassio, D. et  al. Expression and regulation of gap
junctions in rat cholangiocytes. Hepatology, 2002;36(3):631–40.
67. Blackmore, P.F., Strickland, W.G., Bocckino, S.B., and Exton, J.H.
Mechanism of hepatic glycogen synthase in activation induced by Ca2+‐
mobilizing hormones. Biochem J, 1986;237:235–42.
68. Keppens, S. and DeWulf, H. Characterization of the liver P2‐purinoceptor
involved in the activation of glycogen phosphorylase. Biochem J,
1986;240:367–71.
69. Combettes, L., Dumont, M., Berthon, B., Erlinger, S., and Claret, M. Release
of calcium from the endoplasmic reticulum by bile acids in rat liver cells. J
Biol Chem, 1988;263:2299–303.
70. Halpern, K.B., Shenhav, R., Matcovitch‐Natan, O. et al. Single‐cell spatial
reconstruction reveals global division of labour in the mammalian liver.
Nature, 2017;542(7641):352–6.
71. Ozcan, L., Wong, C.C., Li, G. et  al. Calcium signaling through CaMKII
regulates hepatic glucose production in fasting and obesity. Cell Metabol,
2012;15(5):739–51.
72. Wang, Y., Li, G., Goode, J. et al. Inositol‐1,4,5‐trisphosphate receptor regu-
lates hepatic gluconeogenesis in fasting and diabetes. Nature, 2012;485(7396):
128–32.
73. Kruglov, E., Ananthanarayanan, M., Sousa, P., Weerachayaphorn, J., Guerra,
M.T., and Nathanson, M.H. Type 2 inositol trisphosphate receptor gene
expression in hepatocytes is regulated by cyclic AMP. Biochem Biophys Res
Commun, 2017;486(3):659–64.
74. Park, S.W., Zhou, Y., Lee, J., Lee, J., and Ozcan, U. Sarco(endo)plasmic
reticulum Ca2+‐ATPase 2b is a major regulator of endoplasmic reticulum
stress and glucose homeostasis in obesity. Proc Natl Acad Sci USA,
2010;107(45):19320–5.
75. Arruda, A.P., Pers, B.M., Parlakgul, G., Guney, E., Inouye, K., and
Hotamisligil, G.S. Chronic enrichment of hepatic endoplasmic reticulum‐
mitochondria contact leads to mitochondrial dysfunction in obesity. Nat
Med, 2014;20(12):1427–35.
76. Arruda, A.P., Pers, B.M., Parlakgul, G. et al. Defective STIM‐mediated store
operated Ca(2+) entry in hepatocytes leads to metabolic dysfunction in
­obesity. Elife, 2017;6.
77. Nathanson, M.H., Burgstahler, A.D., and Fallon, M.B. Multi‐step mecha-
nism of polarized Ca2+ wave patterns in hepatocytes. Am J Physiol
Gastrointest Liver Physiol, 1994;267:G338–49.
78. Nathanson, M.H., Gautam, A., Bruck, R., Isales, C.M., and Boyer, J.L.
Effects of Ca2+ agonists on cytosolic Ca2+ in isolated hepatocytes and
on  bile secretion in the isolated perfused rat liver. Hepatology,
1992;15:107–16.
79. Beuers, U., Nathanson, M.H., and Boyer, J.L. Effects of tauroursodeoxy-
cholic acid on cytosolic Ca2+ signals in isolated rat hepatocytes.
Gastroenterology, 1993;104:604–12.
80. Nathanson, M.H., Burgstahler, A.D., Masyuk, A., and Larusso, N.F.
Stimulation of ATP secretion in the liver by therapeutic bile acids.
Biochem, J., 2001;358(1):1–5.
81. Poenie, M., Alderton, J., Steinhardt, R., and Tsien, R. Calcium rises abruptly
and briefly throughout the cell at the onset of anaphase. Science, 1986;233(4766):
886–9.
82. Nishiyama, T., Yoshizaki, N., Kishimoto, T., and Ohsumi, K. Transient acti-
vation of calcineurin is essential to initiate embryonic development in
Xenopus laevis. Nature, 2007;449(7160):341–5.
83. Rodrigues, M.A., Gomes, D.A., Leite, M.F. et al. Nucleoplasmic calcium is
required for cell proliferation. J Biol Chem, 2007;282(23):17061–8.
84. Rodrigues, M.A., Gomes, D.A., Andrade, V.A., Leite, M.F., and Nathanson,
M.H. Insulin induces calcium signals in the nucleus of rat hepatocytes.
Hepatology, 2008;48(5)1621–31.
85. Nicou, A., Serriere, V., Hilly, M. et al. Remodelling of calcium signalling
during liver regeneration in the rat. J Hepatol, 2007;46(2):247–56.
86. Khamphaya, T., Chukijrungroat, N., Saengsirisuwan, V. et al. Nonalcoholic
fatty liver disease impairs expression of the type II inositol 1,4,5‐trisphos-
phate receptor. Hepatology, 2018;67(2):560–74.
87. Paranjpe, S., Bowen, W.C., Bell, A.W., Nejak‐Bowen, K., Luo, J.H., and
Michalopoulos, G.K. Cell cycle effects resulting from inhibition of hepato-
cyte growth factor and its receptor c‐Met in regenerating rat livers by RNA
interference. Hepatology, 2007;45(6):1471–7.
88. Rodrigues, M.A., Gomes, D.A., Andrade, V.A., Leite, M.F., and Nathanson,
M.H. Insulin induces calcium signals in the nucleus of rat hepatocytes.
Hepatology, 2008.
89. Amaya, M.J., Oliveira, A.G., Guimaraes, E.S. et  al. The insulin receptor
translocates to the nucleus to regulate cell proliferation in liver. Hepatology,
2014;59(1):274–83.
90. Lagoudakis, L., Garcin, I., Julien, B. et al. Cytosolic calcium regulates liver
regeneration in the rat. Hepatology, 2010;52(2):602–11.
91. Lin, F., Marcelo, K.L., Rajapakshe, K. et al. The camKK2/camKIV relay
is  an essential regulator of hepatic cancer. Hepatology, 2015;62(2):
505–20.
92. Chen, S., Dong, Z., Yang, P. et al. Hepatitis B virus X protein stimulates high
mobility group box 1 secretion and enhances hepatocellular carcinoma
metastasis. Cancer Lett, 2017;394:22–32.
93. Li, Q., Dutta, A., Kresge, C., Bugde, A., and Feranchak, A.P. Bile acids stim-
ulate cholangiocyte fluid secretion by activation of transmembrane member
16A Cl(‐) channels. Hepatology, 2018;68(1):187–99.
508 THE LIVER:  REFERENCES
94. Alvaro, D., Alpini, G., Jezequel, A.M. et al. Role and mechanisms of action
of acetylcholine in the regulation of rat cholangiocyte secretory function. J
Clin Invest, 1997;100:1349–62.
95. Spirli, C., Locatelli, L., Fiorotto, R. et  al. Altered store operated calcium
entry increases cyclic 3’,5’‐adenosine monophosphate production and extra-
cellular signal‐regulated kinases 1 and 2 phosphorylation in polycystin‐2‐
defective cholangiocytes. Hepatology, 2012;55(3):856–68.
96. Minagawa, N., Nagata, J., Shibao, K. et al. Cyclic AMP regulates b­ icarbonate
secretion in cholangiocytes through release of ATP into bile. Gastroenterology,
2007;133(5):1592–602.
97. Sathe, M.N., Woo, K., Kresge, C. et al. Regulation of purinergic signaling in
biliary epithelial cells by exocytosis of SLC17A9‐dependent ATP‐enriched
vesicles. J Biol Chem, 2011;286(28):25363–76.
98. Shibao, K., Hirata, K., Robert, M.E., and Nathanson, M.H. Loss of inositol
1,4,5‐trisphosphate receptors from bile duct epithelia is a common event in
cholestasis. Gastroenterology, 2003;125(4):1175–87.
99. Weerachayaphorn, J., Amaya, M.J., Spirli, C. et  al. Nuclear factor, eryth-
roid  2‐like 2 regulates expression of type 3 inositol 1,4,5‐trisphosphate
receptor and calcium signaling in cholangiocytes. Gastroenterology,
2015;149(1):211–22.
100. Franca, A., Filho, A., Guerra, M.T. et  al. Effects of endotoxin on type 3
inositol 1,4,5‐trisphosphate receptor in human cholangiocytes. Hepatology,
2019;69(2):817–830.
101. Ananthanarayanan, M., Banales, J.M., Guerra, M.T. et  al. Post‐transla-
tional regulation of the type III inositol 1,4,5‐trisphosphate receptor by
miRNA‐506. J Biol Chem, 2015;290(1):184–96.
102. Masyuk, A.I., Masyuk, T.V., Splinter, P.L., Huang, B.Q., Stroope, A.J.,
and  LaRusso, N.F. Cholangiocyte cilia detect changes in luminal fluid
flow  and transmit them into intracellular Ca2+ and cAMP signaling.
Gastroenterology, 2006;131(3):911–20.
103. Gradilone, S.A., Masyuk, A.I., Splinter, P.L. et  al. Cholangiocyte cilia
express TRPV4 and detect changes in luminal tonicity inducing bicarbo-
nate secretion. Proc Natl Acad Sci USA, 2007;104(48):19138–43.
104. Keitel, V., Ullmer, C., and Haussinger, D. The membrane‐bound bile acid
receptor TGR5 (Gpbar‐1) is localized in the primary cilium of cholangio-
cytes. Biol Chem, 2010;391(7):785–9.
105. Pusl, T. and Nathanson, M.H. The role of inositol 1,4,5‐trisphosphate
receptors in the regulation of bile secretion in health and disease. Biochem
Biophys Res Commun, 2004;322(4):1318–25.
106. Gaspers, L.D. and Thomas, A.P. Calcium signaling in liver. Cell Calcium,
2005;38(3–4):329–42.
 Clinical Genomics
41 of NAFLD
Frank Lammert
Department of Medicine II, Saarland University Medical Center, Homburg, Germany
INTRODUCTION
To date fatty liver is a common liver disease worldwide and is to
be included in the differential diagnosis of elevated liver
enzymes, in particular in the setting of indicators of the meta-
bolic syndrome, such as central obesity, dyslipidemia, hyperten-
sion, or increased fasting plasma glucose. The prevalence of
non‐alcoholic fatty liver disease (NAFLD) depends on the defi-
nition, since liver fat content represents a continuous parameter.
In liver biopsies, steatosis (NAFL) is diagnosed when hepatic
steatosis exceeds 5%, and non‐alcoholic steatohepatitis (NASH)
is defined as the presence of at least 5% hepatic steatosis plus
inflammation with hepatocyte injury [1]. In the Dallas heart
study (a multi‐ethnic population‐based probability sample of
Dallas County residents), the upper 95th percentile of liver fat
measured by proton magnetic spectroscopy (1H‐MRS) in
healthy subjects was 5.6%, corresponding to 15% histological
liver fat, and according to this definition, 31% of the cohort
­displayed hepatic steatosis [2]. When measured by magnetic
resonance imaging‐based techniques such as the proton density
fat fraction (PDFF), 5% steatosis corresponds to a PDFF of 6.0
to 6.4% [3].
The presence of cirrhosis with current or previous histopatho-
logical evidence of steatosis or NASH defines NASH‐cirrhosis.
These definitions of fatty liver diseases imply that there is a gradi-
ent from common benign NAFL to defined subgroups of patients
with more advanced disease such as NASH or cirrhosis. Given
the lack of reliable non‐invasive markers, a liver biopsy is still
formally required to detect or exclude the presence of NASH [4].
Modeling the epidemic of NAFLD until 2030 in a Markov
model based on historical and projected changes in adult preva-
lence of obesity and type 2 diabetes mellitus demonstrates an
exponential increase in burden of NAFLD [5]. In the United
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
States, prevalent NAFLD cases are forecasted to increase 21%
from 83 million in 2015 to 101 million 2030, while prevalent
NASH cases will increase 63% from 17 million to 27 million
cases. The overall NAFLD prevalence among the population
aged greater than or equal to 15 years is projected at 34% in
2030. In 2015, approximately 20% of NAFLD cases were clas-
sified as NASH, increasing to 27% by 2030. Incidence of
decompensated cirrhosis increases 170% to 105 000 cases by
2030, while incidence of hepatocellular cancer (HCC) will
increase by 140% to 12 000 cases. Liver deaths will increase
180% to an estimated 78 000 deaths in 2030, and during 2015–
2030, there are projected to be nearly 800 000 excess liver
deaths [5].
Family studies show that first degree relatives of patients with
NAFLD display a markedly higher risk to develop NAFLD than
the general population, independent of obesity. Recently,
­magnetic resonance imaging was used to quantify hepatic stea-
tosis (PDFF) and liver fibrosis in 60 pairs of twins residing in
Southern California [6]. The presence of hepatic steatosis and
the stage of liver fibrosis correlated between monozygotic but
not dizygotic twins. In multivariate models adjusted for age,
sex, and ethnicity, the heritability of hepatic steatosis (based on
PDFF) and liver fibrosis was 50%.
PNPLA3 (ADIPONUTRIN) VARIANT
AND FATTY LIVER DISEASE
Adiponutrin, the enzyme encoded by the PNPLA3 gene (aka
calcium‐independent phospholipase A2ε), is a 481‐amino acid
member of the patatin‐like phospholipase domain‐containing
family (PNPLA). This domain was originally discovered in
510 THE LIVER:  PNPLA3 (ADIPONUTRIN) VARIANT AND FATTY LIVER DISEASE
lipid hydrolases of potato and named after the most abundant
protein of the potato tuber, patatin. However, since several fam-
ily members are not phospholipases, a more appropriate gene
designation has been called for [7]. PNPLA3 is expressed
­predominantly in the liver, retina, skin, and adipose tissue [8]. In
a series of seminal genetic studies, the common non‐synony-
mous variant p.I148M (c.444C>G, rs738409) of the PNPLA3
gene has emerged as the key genetic determinant of fatty liver
disease in pediatric and adult patients [9] (Table  41.1).
Figure  41.1 summarizes the geographical distribution of
PNPLA3 p.I148M genotypes in patients with NAFLD. Of note,
the risk allele frequency is 25% in Europeans [10]. The
­prevalence differs among ethnic groups, and these differences
generally parallel those for NASH and its sequelae. The variant
is most prevalent in Hispanics (49%), is less common in non‐
Hispanic Europeans (23%), and is least common in African
Americans (17%) [2, 11].
The first genome‐wide association study (GWAS) in the
­population‐based Dallas heart study comprising 2111 individu-
als with different ethnic backgrounds demonstrated the p.I148M
variant to be associated (P = 5.9 × 10−10) with increased liver fat
content, as determined by 1H‐magnetic resonance spectrometry
[2]. There was a stepwise increase of hepatic triglyceride ­content
with the number of p.I148M risk alleles. The association was
most prominent in Hispanic patients, who are in general at a
greater risk of developing fatty liver as compared to European
and African Americans.
Of note, the PNPLA3 variant p.I148M is also associated with
serum activities of liver enzymes. The analyses of two large
population‐based cohorts demonstrated that PNPLA3 polymor-
phisms are associated with serum alanine aminotransferase
(ALT) and γ‐glutamyl transpeptidase activities in healthy indi-
viduals [12, 13]. The genetic association between the PNPLA3
mutation p.I148M and fatty liver disease was subsequently
­replicated in many studies [14–17].
Further studies showed that the PNPLA3 variant not only
increases the odds of developing fatty liver itself but it also
determines the degree of hepatic injury and the full spectrum of
histopathological consequences of NAFLD [18]. Valenti et al.
[19, 20] reported that the PNPLA3 variant was not only associ-
ated with hepatic steatosis but with NAFLD severity as deter-
mined by liver biopsy, in particular the presence of NASH,
steatosis grade greater than S1, and fibrosis stage greater than
F1, independent of age, body mass index (BMI), or diabetes.
Another detailed analysis of histopathological severity of
NAFLD was performed by Rotman et al. [21], demonstrating
Table 41.1  Key examples of liver diseases associated with variant PNPLA3
Disease
Non‐alcoholic fatty liver disease
Alcoholic liver cirrhosis
Liver fibrosis
HBV steatosis
HCV steatosis
Alcohol and HCV cirrhosis
HCV cirrhosis
Hepatocellular cancer
HBV, hepatitis B virus; HCV, hepatitis C virus; PNPLA3, patatin‐like phospholipase domain‐containing protein 3.
that the PNPLA3 variant is associated with steatosis, portal and
lobular inflammation, NAFLD activity score (NAS), and fibro-
sis. The first meta‐analysis by Sookoian and Pirola [9] included
data from 16 studies and concluded that the PNPLA3 p.I148M
variant is associated with fatty liver (odds ratio [OR] for
homozygous carriers 3.3 and heterozygous carriers 1.9), NASH
(OR 3.3 and 2.7, respectively), necroinflammation (OR 3.2 and
2.6, respectively), and fibrosis (OR 3.3 and 2.1, respectively);
the association with necroinflammation and fibrosis was inde-
pendent of the severity of steatosis. The recent meta‐analysis by
Xu et al. [22] confirmed these results, with subgroup and sensi-
tivity analyses showing that the results were neither influenced
by ethnicities nor the age of subjects or the source of controls.
Using two histopathologically characterized cohorts encom-
passing steatosis, steatohepatitis, fibrosis, and cirrhosis
(N = 1074), Liu et al. [23] confirmed that the PNPLA3 variant is
associated with advanced fibrosis/cirrhosis, which is independ-
ent of age, BMI, and type 2 diabetes as confounders (Table 41.2).
Availing of transient elastography to quantify liver fibrosis in
899 patients with chronic liver diseases, an association between
the PNPLA3 mutation and enhanced liver stiffness was identi-
fied in a wide spectrum of viral and nonviral chronic liver
­diseases [24]. Sensitivity analysis showed that the association
was present across a broad range of stiffness values (12–40 kPa)
[24], indicating that the variant affects not only fibrogenesis but
also cirrhosis severity.
Non‐obese NAFLD is defined as NAFLD that develops in
patients with a BMI less than 25 kg m−2 [25]. In patients with
non‐obese NAFLD, the PNPLA3 p.I148M allele is more
­frequent than in other NAFLD patients [11, 26–28] and inde-
pendently associated with both NASH and fibrosis stage greater
than or equal to F2 [29].
Carriers of the PNPLA3 mutation who are obese [30] or pre-
sent with NASH [31] are predisposed to HCC development. In
multivariate analysis adjusted for age, sex, BMI, diabetes, and
cirrhosis, carriage of each copy of the PNPLA3 risk allele con-
ferred an additive risk for HCC (OR 2.3), with homozygotes
exhibiting a fivefold increased risk over the wild‐type genotype.
When compared to the UK general population, the risk‐effect
was more pronounced (OR 12.2) [31]. Further analysis of the
data found the positive predictive value of p.I148M to be weak,
but the negative predictive value of the absence of p.I148M is
strong (97%). Hence, prospective studies are warranted to
­validate the clinical utility of PNPLA3 genotyping to select the
patients who are least likely to develop HCC and therefore least
likely to benefit from surveillance [32]. In a series of newly
Study Year Reference
Romeo et al. 2008 [2]
Tian et al. 2010 [98]
Stickel et al. 2011 [99]
Krawczyk et al. 2011 [24]
Vigano et al. 2013 [100]
Cai et al. 2011 [101]
Müller et al. 2011 [102]
Valenti and Fargion 2011 [103]
Nischalke et al. 2011 [104]
 41:  Clinical Genomics of NAFLD 511

Figure 41.1  Worldwide estimated prevalence of NAFLD and distribution of PNPLA3 genotypes. PNPLA3 genotype is presented as minor risk
allele frequency (light blue section of the pie chart). Reproduced with permission of Springer Nature from [105].

Table 41.2  Multivariate analysis of association between PNPLA3 and TM6SF2 genotypes and fibrosis stages F0–1 (mild) versus F2–4
(advanced)
Discovery cohort (n = 349) Validation cohort (n = 725) Combined cohort (n = 1047)
Variables OR (95% CI) P‐value OR (95% CI) P‐value OR (95% CI) P‐value
TM6SF2 genotype 2.94 (1.76–4.89) 3.44 × 10−5 1.46 (1.03–2.09) 0.0362 1.88 (1.41–2.5) 1.63 × 10−5
PNPLA3 genotype 1.57 (1.21–2.19) 0.0086 1.32 (1.05–1.66) 0.0183 1.40 (1.16–1.69) 4.84 × 10−4
Age 1.03 (1.01–1.06) 0.0045 1.02 (1.01–1.04) 0.0041 1.03 (1.01–1.04) 1.57 × 10−5
Gender (female) 0.94 (0.57–1.56) 0.8297 1.81 (1.30–2.50) 4.50 × 10−4 1.43 (1.09–1.89) 0.0096
BMI 1.05 (1.00–1.10) 0.0368 1.03 (1.01–1.05) 9.80 × 10−4 1.04 (1.02–1.05) 3.78 × 10−5
T2DM 2.39 (1.49–3.84) 0.0003 2.73 (1.93–3.88) 1.68 × 10−8 2.57 (1.95–3.39) 1.78 × ×10−11
BMI, body mass index; CI, confidence interval; OR, odds ratio; T2DM, type 2 diabetes mellitus.
Additive model including age, gender, BMI, T2DM, and PNPLA3 rs738409 genotype as covariates. Discovery/validation/combined cohorts: stage F0–1 (mild) n = 198/439/637,
stage F2–4 (advanced) n = 151/286/437.

diagnosed HCC cases  [33], homozygosity for the genotype
p.I148M was even an independent risk factor for death; HCC
patients with this PNPLA3 genotype displayed reduced median
survival (16.8 months) in comparison to carriers of the wild‐
type allele (25.9 months).
Overall the above studies document that the PNPLA3 mutation
increases the risk of developing severe hepatic fat accumulation,
progressive inflammation, advanced fibrosis, and HCC across dif-
ferent ethnicities worldwide (Figure  41.2). Quantitative analyses
showed that adiposity was shown to amplify the genetic risk con-
ferred by PNPLA3 across the full spectrum of NAFLD, ranging
from increased hepatic triglyceride content to cirrhosis (Figure 41.3)
[34]. Additional studies are needed to determine the cost effective-
ness and utility of multifactorial risk stratification that incorporates
PNPLA3 genotype and other risk factors for HCC.
Variant PNPLA3 and metabolic traits
In general, patients with fatty liver disease often present with
dyslipidemia and insulin resistance [35]. However, the associ-
ation of PNPLA3 with metabolic traits in humans is more com-
plex. Several studies did not detect a relationship between the
p.I148M polymorphism and serum glucose or lipid concentra-
tions [17, 19, 36, 37]. In contrast, a Danish analysis of more
than 4000 individuals with normal glucose tolerance demon-
strated an association of the risk allele with increased fasting
glucose levels, but the same allele was associated with lower
serum concentrations of triglycerides and cholesterol in
patients with impaired glucose intolerance [38]. Lower fasting
triglyceride levels were also observed in severely overweight
patients carrying the risk variant [39]. In line with these results,
512 THE LIVER:  PNPLA3 (ADIPONUTRIN) VARIANT AND FATTY LIVER DISEASE

Figure 41.2  Model illustrating the regulation of PNPLA3 gene expression in liver and the link between the p.I148M variant and progressive liver
disease. The PNPLA3 variant drives the full phenotypic spectrum from steatosis to steatohepatitis, fibrosis, cirrhosis, and HCC. Note that ChREBP
controls Pnpla3 mRNA levels in mouse hepatocytes, whereas the specific ChREBP response element is missing in the human PNPLA3 promoter.
Abbreviations: ChREBP, carbohydrate response element binding protein; HCC, hepatocellular cancer; SREBP1c, sterol regulatory element binding
protein 1c. Reproduced with permission of Elsevier from [106].

(a) (b)
25 M variant
II M variant
IM 10 II
20 MM IM
MM
OR (95%CI)
HTGC (%)

15
5
10

2
5
1
<25 25–30 30–35 >35
0 BMI (kg/m2)
<25 25–30 30–35 >35 M variant
BMI (kg/m2) n total II 24,787 22,159 7,034 2,026
IM 14,528 13,079 4,156 1,232
n
MM 2,086 1,851 608 173
PNPLA3 II 360 526 396 342 24
n events II 77 67
M variant IM 90 309 199 171 34 10
IM 48 69
MM 33 60 57 31 7 8
MM 20
16 4

Figure 41.3  (a) Hepatic triglyceride content (HTGC) measured by magnetic resonance spectroscopy, stratified PNPLA3 genotype, and BMI in
the Dallas heart study. The triglyceride increasing effect of the variant is amplified by increasing obesity (Pinteraction = 4 × 10−5). Data are presented
as median ± interquartile range. (b) Risk of cirrhosis by PNPLA3 genotype and BMI in the Copenhagen cohort. The risk‐increasing effect of the
variant is amplified by increasing obesity (Pinteraction = 0.026). Data are OR ± 95% confidence interval (CI). Reproduced with permission of Springer
Nature from [34].
 41:  Clinical Genomics of NAFLD 513

Krawczyk et  al. [10] and other groups [40, 41] identified a
possible association between higher fasting glucose levels and
the PNPLA3 p.I148M mutation. These results contradict the
general paradigm that insulin resistance represents the main
driver of common NAFLD. Indeed, a dissociation between the
presence of fatty liver and insulin resistance appears to be pre-
sent among carriers of the PNPLA3 risk variant [36]. Specific
circulating triacylglycerol signatures were observed in carriers
of the variant, which resembled those of obese subjects with
NAFLD [42]. PNPLA3‐associated NAFLD is associated with
a relative deficiency of triacylglycerols, supporting the idea
that the variant impedes intrahepatocellular lipolysis rather
than stimulating lipid synthesis. NAFLD in obese patients is
associated with multiple changes in triacylglycerols, which
can be attributed to obesity and insulin resistance but not
increased liver fat content per se.
Functional analyses of the PNPLA3 variant
The close similarity to adipose triglyceride lipase and the
presence of typical structural motifs (α–β–α sandwich struc-
ture, consensus serine lipase GXSXG motif, catalytic dyad
S‐D) indicate a lipase function of PNPLA3 [7]. Alternatively,
lysophosphatidic acyltransferase activity has been proposed
[43]. Furthermore, Pirazzi et  al. [44] demonstrated that
PNPLA3 has retinyl‐palmitate lipase activity in hepatic stel-
late cells. The expression of human PNPLA3 is induced by
carbohydrates and fatty acids via the sterol response element
binding protein 1c (SREBP1c, Figure  41.3) as transcription
factor [43, 45–47].
Livers from NAFLD patients are characterized by the
increased number and size of lipid droplets within hepatocytes.
Of note, PNPLA3 is predominantly localized in the endoplas-
mic reticulum (ER) and on the surface of lipid droplets [48]
(Figure  41.4). The rs738409 variant of PNPLA3 results in an
isoleucine to methionine substitution at amino acid residue 148.
Structural modeling indicates that this substitution occludes
access of substrates to the catalytic dyad, resulting in loss of
function [49, 50]. In carriers of the p.I148M mutation, the vari-
ant PNPLA3 evades ubiquitylation and proteosomal degrada-
tion but accumulates on lipid droplets, which increase in number
and median size and display impaired triglyceride mobilization
[49, 51–53].
Studies performed in transgenic mice overexpressing variant
PNPLA3148M but not wild‐type PNPLA3 in liver demonstrated
that these animals develop steatosis due to triacylglycerol accu-
mulation as well as several alterations of hepatic lipid metabo-
lism, including increased synthesis of fatty acids and

Figure 41.4  (a) PNPLA3 is induced with increased free fatty acid availability during hyperinsulinemia and expressed in the ER and on the surface
of lipid droplets. It is rapidly ubiquitylated and degraded upon fasting. (b) PNPLA3 p.I148M on lipid droplets escapes degradation and its accumu-
lation favors triglyceride accumulation and steatohepatitis. (c) Reduced expression of PNPLA3 p.I148M due to the copresence of the p.E434K vari-
ant reduces liver damage. Wild‐type PNPLA3 (p.148I) with intact protective enzymatic activity is illustrated as green curves, and variant PNPLA3
p.148M causing lipid droplet formation as red curves. The number is proportional to protein levels. Red arrows indicate damaging pathways, green
arrows protective pathways, and dashed lines denote suppressed pathways. Abbreviations: TAG, triglyceride; Ub, ubiquitin. Reproduced with per-
mission of John Wiley and Sons from [48].
514 THE LIVER:  TM6SF2 AS SECOND NAFLD RISK GENE
triacylglycerol, impaired hydrolysis of triacylglycerol, and
depletion of triacylglycerol long‐chain polyunsaturated fatty
acids [54]. To determine whether the mutant p.148M allele
causes fat accumulation in the liver when expressed at physio-
logical levels, Pnpla3148M knockin gene were generated, which
displayed normal levels of hepatic fat on chow diet, but when
challenged with a high‐sucrose diet the liver fat content
increased two‐ to threefold compared to wild‐type controls
without changes in glucose homeostasis. The increased liver fat
in the knockin mice was accompanied by a fortyfold increase in
PNPLA3 on hepatic lipid droplets with no increase in hepatic
Pnpla3 mRNA [55].
Genotype‐guided treatment of NAFLD
The in vitro and in vivo studies indicate that suppression of
mutant PNPLA3 would have beneficial effects in NAFLD
and represent a novel therapeutic target for NAFLD. Until
then, lifestyle modification represents the cornerstone of
NAFLD treatment. Although to date large prospective stud-
ies concerning long‐term effects of the common PNPLA3
p.I148M variant on liver phenotypes are lacking, the first
pilot studies indicate that weight loss has positive effects in
carriers of the risk allele [56, 57]. To explore whether the
PNPLA3 variant modulates the effects of weight loss on liver
fat and insulin sensitivity, eight homozygous carriers were
placed on a hypocaloric low‐carbohydrate diet for six days
[57]. Liver fat content (measured by PDFF), whole‐body
insulin sensitivity of glucose metabolism (euglycemic clamp
technique), and lipolysis ([2H5]glycerol infusion) were meas-
ured before and after the diet. At baseline, fasting serum
insulin and C‐peptide concentrations were lower in mutation
carriers. However after a mean weight loss of 3.1 kg, liver fat
decreased by 45% in patients with PNPLA3148M but by 18%
only in controls. The PNPLA3 variant also modulated the
changes in metabolic profile and intrahepatic triglycerides
(IHTG, measured by proton magnetic resonance spectros-
copy) in 154 NAFLD patients [58]. The presence of the
mutation and BMI were independently associated with
greater reduction in IHTG (genotype II: 3.7 ± 5.2%, IM: 6.5
± 3.6%, MM: 11.3 ± 8.8%). Although the PNPLA3 risk allele
confers a higher risk of NAFLD, these patients are more sen-
sitive to the beneficial effects of lifestyle modification. In
the  same line, obese patients carrying the prosteatotic
PNPLA3148M allele lose more weight and liver fat one year
after bariatric surgery, as compared to carriers of PNPLA3
wild‐type alleles [59]. The PNPLA3 genotype and the initial
grade of steatosis were independent predictors of improve-
ment of steatosis after surgery.
In the WELCOME trial, 103 patients with NAFLD were
randomized to ω3 fatty acids or placebo for 15–18 months in
a randomized controlled trial. Adjusting for baseline meas-
urement and covariates, the PNPLA3 p.I148M variant was
independently associated with percentage of docosahexae-
noic acid enrichment and end of study liver fat percentage, but
did not influence the change in serum triglyceride concentra-
tions [60]. The PNPLA3 p.I148M variant has also been
reported to reduce the beneficial effects of statins on steato-
hepatitis [61].
TM6SF2 AS SECOND NAFLD RISK GENE
A GWAS with a denser panel of exonic polymorphisms in three
independent populations (N > 80 000) identified the TM6SF2
variant p.E167K (c.449C>T, rs58542926) to confer susceptibil-
ity to NAFLD in addition to PNPLA3 [23]. The TM6SF2 variant
is markedly less frequent than the PNPLA3 polymorphism in all
ethnic groups (3–7%). TM6SF2 encodes a protein of 351 amino
acids with 7–10 predicted transmembrane domains. TM6SF2 is
localized in the ER and the Golgi complex of hepatocytes and
enterocytes (Figure  41.5), which synthesize apolipoprotein B‐
containing lipoproteins [52]. In contrast to PNPLA3, the expres-
sion of TM6SF2 is not altered by dietary challenges [52]. The
polymorphism is associated with higher liver fat content but
lower serum levels of total and low‐density lipoprotein (LDL)
cholesterol, and triglycerides. The TM6SF2 p.E167K variant
does not exert its effect on hepatic fat content or circulating lipid
profiles by causing hepatic insulin resistance [62]. In patients
with histologically proven NAFLD, an association of this vari-
ant with the degree of steatosis, necroinflammation, ballooning,
and advanced fibrosis was observed, after adjustment for age,
sex, BMI, and diabetes [23, 63] (Table 41.2). Accordingly, the
TM6SF2 variant was independently associated with advanced
fibrosis, when assessed noninvasively by transient elastography
in 890 individuals [64].
TM6SF2 protein expression was decreased markedly in
liver  of NAFLD patients, compared to controls [65]. siRNA
knockdown of Tm6sf2 in mice decreases very‐low‐density
­lipoprotein (VLDL) secretion and increases cellular triglycer-
ide concentration and lipid droplet content, whereas overex-
pression reduces liver cell steatosis [66, 67]. Chronic
inactivation of Tm6sf2 in mice is associated with hepatic stea-
tosis and inflammation as well as decreased plasma levels of
total and LDL cholesterol, thus recapitulating the phenotypes
observed in humans [52, 68]. These observations indicate that
TM6SF2 activity is required for normal VLDL assembly and
that carriers of the TM6SF2 p.E167K variant have fatty liver as
a result of reduced VLDL lipidation (Figure 41.5). As a result,
these patients display lower circulating lipids and reduced risk
of developing carotid plaques and cardiovascular events [63].
In addition, the effect of insulin on glucose production and
lipolysis is better in patients carrying the TM6SF2 p.E167K
variant. Hence, they display a distinct subtype of NAFLD,
which is characterized by preserved insulin sensitivity with
regard to lipolysis, hepatic glucose production, and lack of
hypertriglyceridemia despite clearly increased hepatic fat
­content [62].
The dual and opposite role of the TM6SF2 variant in pro-
tecting against cardiovascular disease and conferring risk for
NAFLD was confirmed in a meta‐analysis of 10 studies.
Pooled estimates of random effects in more than 100 000 indi-
viduals showed that homozygous or heterozygous carriers of
the minor T allele are protected from cardiovascular disease,
showing lower levels of total and LDL cholesterol as well as
triglycerides, while hepatic lipid fat content is about 2% higher
[69]. Accordingly, there is a moderate overall effect on the risk
of NAFLD (OR 2.1). Hence, the variant confers protection
against cardiovascular disease at the expense of an increased
risk of NAFLD.
 41:  Clinical Genomics of NAFLD 515

Figure 41.5  Role of TM6SF2 in VLDL lipidation. VLDL synthesis is initiated in the ER with the co‐translational addition of phospholipids to
apolipoprotein (Apo) B. The addition of triglycerides to the particle also begins in the ER, a process that requires MTTP. The partially lipidated
VLDL particle is packaged into vesicles and exported to the Golgi, where they appear to undergo further bulk phase lipidation. TM6SF2 promotes
this bulk phase lipidation, either by transporting neutral lipids from lipid droplets to the particle by transferring lipid to MTTP (➀), to neutral lipid
droplets in the ER lumen (➁), or directly to the nascent VLDL particle (➂); alternatively, TM6SF2 could participate in the transfer of lipid to the
particle en route to or within the Golgi complex (➃ and ➄). Abbreviations: ER, endoplasmic reticulum; LD, lipid droplet; MTTP, microsomal tri-
glyceride transfer protein; PL, phospholipid; PLTP, phospholipid transfer protein; TG, triglycerides; VLDL, very‐low‐density lipoprotein.
Reproduced with permission of Smagris from [52].

Interestingly, the PNPLA3 and TM6SF2 variants are not only
associated with clinical phenotypes but also with health services
utilization in the general population [70]. In 3759 participants
from the study of health in Pomerania (SHIP), homozygous carri-
ers of the PNPLA3 risk allele had an increased odds of hospitali-
zation (OR 1.5) as compared to major allele homozygous subjects,
and carriers of the TM6SF2 risk allele had higher outpatient utili-
zation (+68%) and inpatient days than major allele homozygous
subjects. These findings highlight the strong genetic effects
that can even be detected in health economic analysis. Recently
the PNPLA3 p.I148M polymorphism was studied in four large
cohorts with extensive health information (23andMe, UK
biobank, FINRISK, CHOP) for association with 1683 binary end-
points in up to 700 000 individuals [71]. This study design has
been termed phenome‐wide association study (PheWAS), which
is an unbiased approach to test for associations between genetic
variants and a wide range of phenotypes in large population
cohorts [72]. Of note, the PNPLA3 variant, which predisposes to
fatty liver disease, was associated with diabetes and lower serum
cholesterol levels, and an inverse association was also detected
for gallstones, acne, and gout. Beyond that, the analysis also indi-
cated that carriers of the risk allele are prone to develop drug‐
induced liver injury (OR 1.5) when treated with non‐steroidal
anti‐inflammatory drugs (NSAID) such as ibuprofen or aspirin.
These associations were replicated in the UK biobank cohort and
remained when adjusting for elevated liver tests [71].
MBOAT7: THE THIRD MAN
After a GWAS had reported that the rs641738 C>T variant linked
to the 3’ untranslated region of the gene encoding membrane‐
bound O‐acyltransferase domain‐containing 7 gene (MBOAT7,
aka lysophosphatidylinositol‐acyltransferase 1, LPIAT1) increases
the risk for alcoholic cirrhosis [73], 2736 participants from the
Dallas heart study who had undergone proton magnetic resonance
spectroscopy to measure IHTG and 1149 European individuals
from the liver biopsy cross‐sectional cohort were genotyped for
rs641738 [74]. This variant, which encodes p.G17E in the trans-
membrane channel‐like 4 (TMC4), is associated with suppression
of MBOAT7 at mRNA and protein levels. It was also associated
with increased hepatic fat content, more severe liver damage, and
increased stage of fibrosis compared with subjects without the
variant. MBOAT7 transfers polyunsaturated fatty acids (PUFA)
such as arachidonic acid to lysophospholipids. The MBOAT7
rs641738 risk allele was associated with altered plasma phosphati-
dylinositol (PI) species, consistent with decreased MBOAT7 func-
tion. Metabolic profiling indicates changes of PI side‐chain
remodeling, in particular a lack of transfer of arachidonoyl‐CoA to
lyso‐PI (Lands cycle) [75]. Interestingly, PNPLA3 promotes the
transfer of PUFA, especially arachidonic acid, from triglycerides
to phospholipids in hepatic lipid droplets [76]. These findings
together indicate a potential role of arachidonic acid incorporation
in phospholipids during NAFLD pathogenesis.
516 THE LIVER:  PROTECTIVE GENE VARIATION
ADDITIONAL RISK GENES
GWAS and their meta‐analyses have identified additional vari-
ants that might be associated with NAFLD. For example,
Speliotes and coworkers [17, 77] reported that variants associ-
ated at genome‐wide significant levels in or near the glucoki-
nase regulatory protein (GCKR), lysophospholipase‐like 1
(LYPLAL1) and protein phosphatase 1‐regulatory subunit 3b
(PPP1R3B) might be associated with liver fat content and/or
histopathological NAFLD phenotypes. The common missense
loss‐of‐function GCKR mutation p.P446L (c.1337T>C,
rs1260326) reduces the ability of GCKR to inhibit glucokinase
in response to fructose‐6‐phosphate, thereby increasing glu-
cokinase activity and hepatic glucose uptake. The unrestricted
hepatic glycolysis reduces fasting glucose and insulin levels but
increases the production of malonyl‐CoA, which in turn favors
hepatic fat accumulation by serving as a substrate for de novo
lipogenesis and by disruption of mitochondrial fatty acid β‐oxi-
dation [78, 79]. The variants at these loci exhibit distinct pat-
terns of association with serum lipids as well as glycemic and
anthropometric traits, which also differ with ethnicity. Again
NAFLD‐associated variants are not uniformly associated with
abnormalities in serum lipids or glycemic and anthropometric
traits, suggesting genetic heterogeneity in the pathways influ-
encing these traits.
PROTECTIVE GENE VARIATION
Abul‐Husn et  al. [80] reported that a truncated variant of
hydroxysteroid 17β dehydrogenase 13 (HSD17B13) is associ-
ated with a reduced risk of chronic liver disease and of progres-
sion from steatosis to steatohepatitis. It has to be noted that the
term “protection” represents the flipside of “susceptibility”,
hence it depends on the population frequencies of the alleles,
with the minor allele defining the direction of the effect (pre-
disposition vs. protection). Using exome sequence data and
electronic health records from 46 544 participants in the
DiscovEHR human genetics study, gene variants that are asso-
ciated with serum activities of aminotransferases were identi-
fied. Replicated variants were evaluated for association with
clinical diagnoses of chronic liver disease in DiscovEHR study
participants as well as two independent cohorts (N = 37 173)
and with histopathological severity of liver disease in 2391
human liver biopsies. In homozygous and heterozygous carri-
ers of the variant, the risks of NAFLD and NASH‐cirrhosis
were reduced by 30% and 17% and 49% and 26%, respectively.
Of note, the HSD17B13 variant ameliorated liver injury associ-
ated with the PNPLA3 p.I148M risk allele. In a second detailed
histopathological study of 356 patients with biopsy‐proven dis-
ease the variant protected against lobular inflammation, bal-
looning degeneration, and fibrosis [81].
Hydroxysteroid17β dehydrogenases (HSD17B) form an
enzyme family characterized by their ability to catalyze reac-
tions in steroid and lipid metabolism. Recently Ma et al. [82]
reported that HSD17B13 is a hepatic retinol dehydrogenase that
is targeted to lipid droplets. Mice deficient in Hsd17b13 develop
hepatic steatosis, while serum steroid concentrations are normal
[83]. In line with these changes, the expression of key enzymes
in fatty acid synthesis, such as fatty acid synthase, acetyl‐CoA
carboxylase 1, and stearoyl‐CoA desaturase, is increased in liv-
ers from Hsd17b3 knockout mice, while glucose tolerance does
not differ [83].
Studies in pediatric cohorts
NAFLD is becoming an emerging health problem in pediatric
populations. In line with the studies in adult patients, an associa-
tion between fatty liver and the PNPLA3 mutation p.I148M was
observed in children with NAFLD, too. In a study by Valenti
et al. [20], the PNPLA3 variant was associated with the steatosis
grade in 149 children with biopsy‐proven NAFLD. In this cohort,
all 23 children who were homozygous carriers of the PNPLA3
mutation were diagnosed with NASH [20]. The association
between the PNPLA3 mutation and pediatric NAFLD has also
been detected in African American [84] and Chinese children
[85]. Interestingly, children carrying the PNPLA3 risk allele
seem to be predisposed to the early development of NAFLD
[21]. The analysis of 6–12‐year‐old Mexican children showed
that already at this age the PNPLA3 mutation is associated with
increased serum ALT activities [86], and we detected the same
association in a cohort of German children aged 5–9 years [87].
The TM6SF2 variant p.E167K was also shown to be associated
with hepatic steatosis and increased aminotransferase activities
but lower serum cholesterol and triglyceride concentrations in
obese children [88, 89]. In multiethnic cohorts of obese children
and adolescents [90, 91], the MBOAT7 risk allele was associated
with hepatic fat content (as determined by magnetic resonance
imaging), fasting insulin and plasma glucose levels, and reduced
whole‐body insulin sensitivity, independent of age, sex, and BMI
z‐score. Moreover, these studies detected joint effects among
TM6SF2 p.E167K, PNPLA3 p.I148M, and the MBOAT7 rs626283
and GCKR rs1260326 polymorphisms in determining IHTG. The
four polymorphisms combined explain about 20% of the hepatic
fat fraction in Caucasian obese children.
A practical consideration is that physical activity and weight
reduction might substantially improve the liver status in chil-
dren carrying the risk variants and therefore rescue their delete-
rious liver phenotypes [56]. This possibility points to the need of
early detection of children carrying risk mutations who may
require more careful follow‐up and tailored therapies aiming to
intercept NAFLD progression.
Combined effects of gene variants
Mancina et  al. [74] analyzed the combined genetic effects of
PNPLA3, TM6SF2, and MBOAT7 effects on NAFLD histopathol-
ogy (Table 41.3). They did not observe an interaction, instead the
three gene variants appear to act in an additive fashion, with a
stepwise increase in mean hepatic fat content per each additional
risk allele (Figure 41.6). The combined effects of the PNPLA3,
TM6SF2, and MBOAT7 variants on NAFLD severity were stud-
ied in a German multicenter biopsy‐based study that recruited
515 patients with NAFLD [92]. In the multivariate model, the
PNPLA3 and TM6SF2 variants were associated with steatosis,
and fibrosis stages were affected by the PNPLA3 and MBOAT7
polymorphism. Of note, a significant increase of serum AST
 41:  Clinical Genomics of NAFLD 517

Table 41.3  Population attributable fractionsa of NAFLD genes

Steatosis Fibrosis
(confidence intervals) NASH F2–F4
PNPLA3 23% (11–36) 19% (12–26) 18% (10–26)
TM6SF2 4% (0–14) 4% (1–8) 3% (0–7)
MBOAT7 15% (3–28) 7% (0–14) 11% (2–23)
a
 The contribution of a risk factor to a disease or a death is quantified using PAF. PAF is the
proportional reduction in population disease or mortality that would occur if exposure to a risk
factor were reduced to an alternative ideal exposure scenario.

(b) P < 0.0001

500
300
100
100

AST [U/I]
(a)
20 50
P-trend < .0001
Mean hepatic TG%

15 0

le

s
le

le

le

le

le
le
le

le

le

le

le
al
al

al

al

al

al
k
k

ris

k
ris

ris

ris

ris

ris
10

1
0

5
Number of risk alleles

5 P = 0.08
400
300
200
0 150
0 1 2 3 4 5
ALT [U/I]

n= 627 1066 747 248 41 6 100

Number of risk alleles

50

s
s
s

le

s
le
le
le

le

le
n = 515
le

le
le
le

le

le
al

al
al
al

al

al
k

k
k
k

ris

k
ris
ris
ris

ris

ris
1

4
3
0

5
Number of risk alleles

Figure 41.6  Combined genetic effects in NAFLD. (a) Association between number of PNPLA3, TM6SF2, and MBOAT7 risk alleles and hepatic
triglyceride (TG) content in the Dallas heart study. The graph shows mean hepatic TG content by the number of risk alleles (error bars: SEM). From
[74]. (b) Association between number of PNPLA3, TM6SF2, and MBOAT7 risk alleles and liver function tests in the NAFLD CSG study. The graphs
illustrate median aspartate aminotransferase (AST) and alanine aminotransferase (ALT) activities by the number of risk alleles (error bars: range).
Analyses were performed using trend tests. The following frequencies of carriers of risk alleles were detected: 0 risk alleles, N = 56; one risk allele,
N = 142; two risk alleles, N = 170; three risk alleles, N = 117; four risk alleles, N = 27; five risk alleles, N = 3. Reproduced with permission of
Elsevier from [92].

activities was associated with the increment of risk alleles of
either of the genotypes, and similar trends existed for ALT and
γ‐glutamyl transpeptidase levels. The same observation was made
in Korea: the PNPLA3 and TM6SF2 variants, but not the MBOAT7
variant were associated with both NASH and significant fibrosis
(≥ F2), even after adjustment for insulin resistance [93].
MONOGENIC DISEASES CAUSING NAFLD
Monogenic diseases that are associated with NAFLD are rare
and usually result in extrahepatic manifestations that dominate
over hepatic steatosis. However, they provide insight into the
pathogenesis of hepatic steatosis and can be considered in rare
cases when the etiology is unclear. Examples include abetali-
poproteinemia (microsomal triglyceride transfer protein,
MTTP), hypobetalipoproteinemia (apolipoprotein B, APOB),
citrullinemia type 2 (solute carrier SLC25A13), Wilson disease
(ATPase ATP7B), neutral lipid storage disease (PNPLA2), and
cholesteryl ester storage disease (lysosomal acid lipase, LIPA).
Rare inherited mitochondriopathies have also been associated
with NASH and point to the role of impaired fatty acid oxida-
tion in the pathogenesis of NASH. Genetic testing is helpful
where a rare monogenic inherited disease of lipid metabolism
or mitochondrial function is suspected. More recently exome
518 THE LIVER:  REFERENCES
and whole genome sequencing is applied to identify novel gene
variants causing liver diseases, including severe types of
NAFLD.
PNPLA3‐ AND TM6SF2‐ASSOCIATED
STEATOHEPATITIS AND FUTURE
DIRECTIONS
PNPLA3 (and TM6SF2 plus MBOAT7) genotyping may be used
as novel non‐invasive indicator for an increased risk of progres-
sive fatty liver disease and could be included in the clinical deci-
sion‐making in patients with NAFLD. Because in patients who
carry the PNPLA3 risk variant increased lipid content and inflam-
mation in liver can be driven predominantly by PNPLA3 together
with environmental risk factors, the name PNPLA3‐NAFLD or
PASH (i.e. PNPLA3‐associated steatohepatitis) might be used as
a novel gene‐based liver disease entity [94, 95]. PNPLA3‐
NAFLD/PASH represents an example of how to reclassify dis-
ease according to molecular pathways and pathophysiological
changes in the era of personalized medicine [96]. Carriers of the
genetic risk factors could benefit from a more systematic, early,
and careful surveillance of complications of progressive fatty
liver disease, including HCC in the absence of cirrhosis [97].
REFERENCES
  1. Chalasani, N., Younossi, Z., Lavine, J.E. et al. The diagnosis and management
of nonalcoholic fatty liver disease: practice guidance from the American
Association for the Study of Liver Diseases. Hepatology, 2018;67(1):328–57.
  2. Romeo, S., Kozlitina, J., Xing, C. et al. Genetic variation in PNPLA3 confers
susceptibility to nonalcoholic fatty liver disease. Nat Genet, 2008;40(12):
1461–5.
  3. Petaja, E.M. and Yki‐Järvinen, H. Definitions of normal liver fat and the
association of insulin sensitivity with acquired and genetic NAFLD‐A sys-
tematic review. Int J Mol Sci, 2016;17(5).
  4. Verhaegh, P., Bavalia, R., Winkens, B. et al. Noninvasive tests do not accurately
differentiate nonalcoholic steatohepatitis from simple steatosis: a systematic
review and meta‐analysis. Clin Gastroenterol Hepatol, 2018;16(6):837–61.
  5. Estes, C., Razavi, H., Loomba, R. et al. Modeling the epidemic of nonalco-
holic fatty liver disease demonstrates an exponential increase in burden of
disease. Hepatology, 2018;67(1):123–33.
  6. Loomba, R., Schork, N., Chen, C.H. et al. Heritability of hepatic fibrosis and
steatosis based on a prospective twin study. Gastroenterology, 2015;149(7):
1784–93.
  7. Zechner, R., Zimmermann, R., Eichmann, T.O. et al. Fat signals – lipases
and lipolysis in lipid metabolism and signaling. Cell Metab,
2012;15(3):279–91.
  8. Huang, Y., He, S., Li, J.Z. et al. A feed‐forward loop amplifies nutritional
regulation of PNPLA3. Proc Natl Acad Sci USA, 2010;107(17):7892–7.
  9. Sookoian, S. and Pirola, C.J. Meta‐analysis of the influence of I148M variant
of patatin‐like phospholipase domain containing 3 gene (PNPLA3) on the
susceptibility and histological severity of nonalcoholic fatty liver disease.
Hepatology, 2011;53(6):1883–94.
10. Krawczyk, M., Grünhage, F., Mahler, M. et  al. The common adiponutrin
variant p.I148M does not confer gallstone risk but affects fasting glucose and
triglyceride levels. J Physiol Pharmacol, 2011;62(3):369–75.
11. Diehl, A.M. and Day, C. Nonalcoholic steatohepatitis. N Engl J Med,
2018;378(8):781.
12. Chambers, J.C., Zhang, W., Sehmi, J. et al. Genome‐wide association study
identifies loci influencing concentrations of liver enzymes in plasma. Nat
Genet, 2011;43(11):1131–8.
13. Yuan, X., Waterworth, D., Perry, J.R. et al. Population‐based genome‐wide
association studies reveal six loci influencing plasma levels of liver enzymes.
Am J Hum Genet, 2008;83(4):520–8.
14. Kawaguchi, T., Sumida, Y., Umemura, A. et al. Genetic polymorphisms of
the human PNPLA3 gene are strongly associated with severity of non‐alco-
holic fatty liver disease in Japanese. PLoS One, 2012;7(6), e38322.
15. Kitamoto, T., Kitamoto, A., Yoneda, M. et al. Genome‐wide scan revealed
that polymorphisms in the PNPLA3, SAMM50, and PARVB genes are asso-
ciated with development and progression of nonalcoholic fatty liver disease
in Japan. Hum Genet, 2013;132(7):783–92.
16. Kotronen, A., Johansson, L.E., Johansson, L.M. et al. A common variant in
PNPLA3, which encodes adiponutrin, is associated with liver fat content in
humans. Diabetologia, 2009;52(6):1056–60.
17. Speliotes, E.K., Yerges‐Armstrong, L.M., Wu, J. et al. Genome‐wide asso-
ciation analysis identifies variants associated with nonalcoholic fatty liver
disease that have distinct effects on metabolic traits. PLoS Genet, 2011;7(3),
e1001324.
18. Sookoian, S., Castano, G.O., Burgueno, A.L. et al. A nonsynonymous gene
variant in the adiponutrin gene is associated with nonalcoholic fatty liver
disease severity. J Lipid Res, 2009;50(10):2111–6.
19. Valenti, L., Al‐Serri, A., Daly, A.K. et al. Homozygosity for the patatin‐like
phospholipase‐3/adiponutrin I148M polymorphism influences liver fibrosis
in patients with nonalcoholic fatty liver disease. Hepatology, 2010;51(4):
1209–17.
20. Pelusi, S., Cespiati, A., Rametta, R. et al. Prevalence and risk factors of sig-
nificant fibrosis in patients with nonalcoholic fatty liver without steatohepa-
titis. Clin Gastroenterol Hepatol, 2019;epub.
21. Rotman, Y., Koh, C., Zmuda, J.M. et al. The association of genetic variability
in patatin‐like phospholipase domain‐containing protein 3 (PNPLA3) with
histological severity of nonalcoholic fatty liver disease. Hepatology,
2010;52(3):894–903.
22. Xu, R., Tao, A., Zhang, S. et al. Association between patatin‐like phospholi-
pase domain containing 3 gene (PNPLA3) polymorphisms and nonalcoholic
fatty liver disease: a HuGE review and meta‐analysis. Sci Rep, 2015;5, 9284.
23. Liu, Y.L., Reeves, H.L., Burt, A.D. et  al. TM6SF2 rs58542926 influences
hepatic fibrosis progression in patients with non‐alcoholic fatty liver disease.
Nat Commun, 2014;5, 4309.
24. Krawczyk, M., Grünhage, F., Zimmer, V. et al. Variant adiponutrin (PNPLA3)
represents a common fibrosis risk gene: non‐invasive elastography‐based
study in chronic liver disease. J Hepatol, 2011;55(2):299–306.
25. Kim, D. and Kim, W.R. Nonobese fatty liver disease. Clin Gastroenterol
Hepatol, 2017;15(4):474–85.
26. Wei, J.L., Leung, J.C., Loong, T.C. et al. Prevalence and severity of nonalco-
holic fatty liver disease in non‐obese patients: a population study using
proton‐magnetic resonance spectroscopy. Am J Gastroenterol, 2015;110(9):
1306–14.
27. Leung, J.C., Loong, T.C., Wei, J.L. et al. Histological severity and clinical
outcomes of nonalcoholic fatty liver disease in nonobese patients.
Hepatology, 2017;65(1):54–64.
28. Krawczyk, M., Bantel, H., Rau, M. et  al. Could inherited predisposition
drive non‐obese fatty liver disease? Results from German tertiary referral
centers. J Hum Genet, 2018;63(5):621–6.
29. Fracanzani, A.L., Petta, S., Lombardi, R. et  al. Liver and cardiovascular
damage in patients with lean nonalcoholic fatty liver disease, and association
with visceral obesity. Clin Gastroenterol Hepatol, 2017;15(10):1604–11.
30. Burza, M.A., Pirazzi, C., Maglio, C. et  al. PNPLA3 I148M (rs738409)
genetic variant is associated with hepatocellular carcinoma in obese indi-
viduals. Dig Liver Dis, 2012;44(12):1037–41.
31. Liu, Y.L., Patman, G.L., Leathart, J.B. et  al. Carriage of the PNPLA3
rs738409 C >G polymorphism confers an increased risk of non‐alcoholic
fatty liver disease associated hepatocellular carcinoma. J Hepatol,
2014;61(1):75–81.
32. Anstee, Q.M., Seth, D., Day, C.P. Genetic factors that affect risk of
alcoholic  and nonalcoholic fatty liver disease. Gastroenterology,
­
2016;150(8):1728–44.
33. Hassan, M.M., Kaseb, A., Etzel, C.J. et al. Genetic variation in the PNPLA3
gene and hepatocellular carcinoma in USA: risk and prognosis prediction.
Mol Carcinog, 2013;52(1), E139–47.
34. Stender, S., Kozlitina, J., Nordestgaard, B.G. et al. Adiposity amplifies the
genetic risk of fatty liver disease conferred by multiple loci. Nat Genet,
2017;49(6):842–7.
 41:  Clinical Genomics of NAFLD 519
35. Palasciano, G., Moschetta, A., Palmieri, V.O. et al. Non‐alcoholic fatty liver
disease in the metabolic syndrome. Curr Pharm Des, 2007;13(21):2193–8.
36. Kantartzis, K., Peter, A., Machicao, F. et al. Dissociation between fatty liver
and insulin resistance in humans carrying a variant of the patatin‐like phos-
pholipase 3 gene. Diabetes, 2009;58(11):2616–23.
37. Speliotes, E.K., Butler, J.L., Palmer, C.D. et al. PNPLA3 variants specifi-
cally confer increased risk for histologic nonalcoholic fatty liver disease but
not metabolic disease. Hepatology, 2010;52(3):904–12.
38. Krarup, N.T., Grarup, N., Banasik, K. et al. The PNPLA3 rs738409 G‐allele
associates with reduced fasting serum triglyceride and serum cholesterol in
Danes with impaired glucose regulation. PLoS One, 2012;7(7), e40376.
39. Stojkovic, I.A., Ericson, U., Rukh, G. et al. The PNPLA3 Ile148Met inter-
acts with overweight and dietary intakes on fasting triglyceride levels. Genes
Nutr, 2014;9(2):388.
40. Palmer, C.N., Maglio, C., Pirazzi, C. et al. Paradoxical lower serum triglyc-
eride levels and higher type 2 diabetes mellitus susceptibility in obese indi-
viduals with the PNPLA3 148M variant. PLoS One, 2012;7(6), e39362.
41. Rembeck, K., Maglio, C., Lagging, M. et al. PNPLA 3 I148M genetic vari-
ant associates with insulin resistance and baseline viral load in HCV geno-
type 2 but not in genotype 3 infection. BMC Med Genet, 2012;13, 82.
42. Hyysalo, J., Gopalacharyulu, P., Bian, H. et  al. Circulating triacylglycerol
signatures in nonalcoholic fatty liver disease associated with the I148M vari-
ant in PNPLA3 and with obesity. Diabetes, 2014;63(1):312–22.
43. Kumari, M., Schoiswohl, G., Chitraju, C. et al. Adiponutrin functions as a
nutritionally regulated lysophosphatidic acid acyltransferase. Cell Metab,
2012;15(5):691–702.
44. Pirazzi, C., Valenti, L., Motta, B.M. et  al. PNPLA3 has retinyl‐palmitate
lipase activity in human hepatic stellate cells. Hum Mol Genet,
2014;23(15):4077–85.
45. Huang, Y., Cohen, J.C., Hobbs, H.H. Expression and characterization of a
PNPLA3 protein isoform (I148M) associated with nonalcoholic fatty liver
disease. J Biol Chem, 2011;286(43):37085–93.
46. Dubuquoy, C., Robichon, C., Lasnier, F. et al. Distinct regulation of adiponu-
trin/PNPLA3 gene expression by the transcription factors ChREBP and
SREBP1c in mouse and human hepatocytes. J Hepatol, 2011;55(1):145–53.
47. Perttila, J., Huaman‐Samanez, C., Caron, S. et al. PNPLA3 is regulated by
glucose in human hepatocytes, and its I148M mutant slows down triglycer-
ide hydrolysis. Am J Physiol Endocrinol Metab, 2012;302(9), E1063–9.
48. Valenti, L. and Dongiovanni, P. Mutant PNPLA3 I148M protein as pharma-
cological target for liver disease. Hepatology, 2017;66(4):1026–8.
49. He, S., McPhaul, C., Li, J.Z. et al. A sequence variation (I148M) in PNPLA3
associated with nonalcoholic fatty liver disease disrupts triglyceride hydroly-
sis. J Biol Chem, 2010;285(9):6706–15.
50. Xin, Y.N., Zhao, Y., Lin, Z.H. et  al. Molecular dynamics simulation of
PNPLA3 I148M polymorphism reveals reduced substrate access to the cata-
lytic cavity. Proteins, 2013;81(3):406–14.
51. Chamoun, Z., Vacca, F., Parton, R.G. et al. PNPLA3/adiponutrin functions in
lipid droplet formation. Biol Cell, 2013;105(5):219–33.
52. Smagris, E., Gilyard, S., BasuRay, S. et al. Inactivation of Tm6sf2, a gene
defective in fatty liver disease, impairs lipidation but not secretion of very
low density lipoproteins. J Biol Chem, 2016;291(20):10659–76.
53. BasuRay, S., Smagris, E., Cohen, J.C. et al. The PNPLA3 variant associated
with fatty liver disease (I148M) accumulates on lipid droplets by evading
ubiquitylation. Hepatology, 2017;66(4):1111–24.
54. Li, J.Z., Huang, Y., Karaman, R. et al. Chronic overexpression of PNPLA3I148M
in mouse liver causes hepatic steatosis. J Clin Invest, 2012;122(11):4130–44.
55. Smagris, E., BasuRay, S., Li, J. et al. Pnpla3I148M knockin mice accumu-
late PNPLA3 on lipid droplets and develop hepatic steatosis. Hepatology,
2015;61(1):108–18.
56. Marzuillo, P., Grandone, A., Perrone, L. et al. Weight loss allows the dissec-
tion of the interaction between abdominal fat and PNPLA3 (adiponutrin) in
the liver damage of obese children. J Hepatol, 2013;59(5):1143–4.
57. Sevastianova, K., Kotronen, A., Gastaldelli, A. et  al. Genetic variation in
PNPLA3 (adiponutrin) confers sensitivity to weight loss‐induced decrease in
liver fat in humans. Am J Clin Nutr, 2011;94(1):104–11.
58. Shen, J., Wong, G.L., Chan, H.L. et al. PNPLA3 gene polymorphism and
response to lifestyle modification in patients with nonalcoholic fatty liver
disease. J Gastroenterol Hepatol, 2015;30(1):139–46.
59. Krawczyk, M., Jimenez‐Aguero, R., Alustiza, J.M. et al. PNPLA3 p.I148M
variant is associated with greater reduction of liver fat content after bariatric
surgery. Surg Obes Relat Dis, 2016;12(10):1838–46.
60. Scorletti, E., West, A.L., Bhatia, L. et al. Treating liver fat and serum triglyc-
eride levels in NAFLD, effects of PNPLA3 and TM6SF2 genotypes: Results
from the WELCOME trial. J Hepatol, 2015;63(6):1476–83.
61. Dongiovanni, P., Petta, S., Mannisto, V. et al. Statin use and non‐alcoholic
steatohepatitis in at risk individuals. J Hepatol, 2015;63(3):705–12.
62. Zhou, Y., Llaurado, G., Oresic, M. et al. Circulating triacylglycerol signa-
tures and insulin sensitivity in NAFLD associated with the E167K variant in
TM6SF2. J Hepatol, 2015;62(3):657–63.
63. Dongiovanni, P., Petta, S., Maglio, C. et al. Transmembrane 6 superfamily
member 2 gene variant disentangles nonalcoholic steatohepatitis from car-
diovascular disease. Hepatology, 2015;61(2):506–14.
64. Petta, S., Di Marco, V., Pipitone, R.M. et al. Prevalence and severity of non-
alcoholic fatty liver disease by transient elastography: genetic and metabolic
risk factors in a general population. Liver Int, 2018;38(11):2060–8.
65. Sookoian, S., Castano, G.O., Scian, R. et al. Genetic variation in transmem-
brane 6 superfamily member 2 and the risk of nonalcoholic fatty liver disease
and histological disease severity. Hepatology, 2015;61(2):515–25.
66. Kozlitina, J., Smagris, E., Stender, S. et al. Exome‐wide association study
identifies a TM6SF2 variant that confers susceptibility to nonalcoholic fatty
liver disease. Nat Genet, 2014;46(4):352–6.
67. Mahdessian, H., Taxiarchis, A., Popov, S. et al. TM6SF2 is a regulator of
liver fat metabolism influencing triglyceride secretion and hepatic lipid
droplet content. Proc Natl Acad Sci USA, 2014;111(24):8913–8.
68. Fan, Y., Lu, H., Guo, Y. et al. Hepatic transmembrane 6 superfamily member
2 regulates cholesterol metabolism in mice. Gastroenterology,
2016;150(5):1208–18.
69. Pirola, C.J. and Sookoian, S. The dual and opposite role of the TM6SF2‐
rs58542926 variant in protecting against cardiovascular disease and confer-
ring risk for nonalcoholic fatty liver: a meta‐analysis. Hepatology,
2015;62(6):1742–56.
70. Kopp, J., Flessa, S., Lieb, W. et al. Association of PNPLA3 rs738409 and
TM6SF2 rs58542926 with health services utilization in a population‐based
study. BMC Health Serv Res, 2016;16, 41.
71. Diogo, D., Tian, C., Franklin, C.S. et al. Phenome‐wide association studies
across large population cohorts support drug target validation. Nat Commun,
2018;9(1):4285.
72. Bush, W.S., Oetjens, M.T., Crawford, D.C. Unravelling the human genome‐
phenome relationship using phenome‐wide association studies. Nat Rev
Genet, 2016;17(3):129–45.
73. Buch, S., Stickel, F., Trepo, E. et al. A genome‐wide association study con-
firms PNPLA3 and identifies TM6SF2 and MBOAT7 as risk loci for alcohol‐
related cirrhosis. Nat Genet, 2015;47(12):1443–8.
74. Mancina, R.M., Dongiovanni, P., Petta, S. et al. The MBOAT7‐TMC4 vari-
ant rs641738 increases risk of nonalcoholic fatty liver disease in individuals
of european descent. Gastroenterology, 2016;150(5):1219–30.
75. Wang, B. and Tontonoz, P. Phospholipid remodeling in physiology and dis-
ease. Annu Rev Physiol, 2018;81:165–88.
76. Mitsche, M.A., Hobbs, H.H., and Cohen, J.C. Patatin‐like phospholipase
domain‐containing protein 3 promotes transfer of essential fatty acids from
triglycerides to phospholipids in hepatic lipid droplets. J Biol Chem,
2018;293(18):6958–68.
77. Hernaez, R., McLean, J., Lazo, M. et al. Association between variants in or
near PNPLA3, GCKR, and PPP1R3B with ultrasound‐defined steatosis
based on data from the third National Health and Nutrition Examination
Survey. Clin Gastroenterol Hepatol, 2013;11(9):1183–90.
78. Beer, N.L., Tribble, N.D., McCulloch L.J. et al. The P446L variant in GCKR
associated with fasting plasma glucose and triglyceride levels exerts its
effect  through increased glucokinase activity in liver. Hum Mol Genet,
2009;18(21):4081–8.
79. Eslam, M., Valenti, L., and Romeo, S. Genetics and epigenetics of NAFLD
and NASH: clinical impact. J Hepatol, 2018;68(2):268–79.
80. Abul‐Husn, N.S., Cheng, X., Li, A.H. et al. A protein‐truncating HSD17B13
variant and protection from chronic liver disease. N Engl J Med,
2018;378(12):1096–106.
81. Pirola, C.J., Garaycoechea, M., Flichman, D. et al. Splice variant rs72613567
prevents worst histologic outcomes in patients with nonalcoholic fatty liver
disease. J Lipid Res, 2019;60(1):176–85.
82. Ma, Y., Belyaeva, O.V., Brown, P.M. et  al. 17‐Beta hydroxysteroid
dehydrogenase 13 is a hepatic retinol dehydrogenase associated with
­
histological features of nonalcoholic fatty liver disease. Hepatology,
­
2019;69(4):1504–19.
520 THE LIVER:  REFERENCES
83. Adam, M., Heikela, H., Sobolewski, C. et al. Hydroxysteroid (17beta) dehy-
drogenase 13 deficiency triggers hepatic steatosis and inflammation in mice.
FASEB J, 2018;32(6):3434–47.
84. Santoro, N., Kursawe, R., D’Adamo E. et al. A common variant in the pata-
tin‐like phospholipase 3 gene (PNPLA3) is associated with fatty liver disease
in obese children and adolescents. Hepatology, 2010;52(4):1281–90.
85. Lin, Y.C., Chang, P.F., Hu, F.C. et al. A common variant in the PNPLA3 gene
is a risk factor for non‐alcoholic fatty liver disease in obese Taiwanese chil-
dren. J Pediatr, 2011;158(5):740–4.
86. Larrieta‐Carrasco, E., Leon‐Mimila, P., Villarreal‐Molina, T. et  al.
Association of the I148M/PNPLA3 variant with elevated alanine transami-
nase levels in normal‐weight and overweight/obese Mexican children. Gene,
2013;520(2):185–8.
87. Krawczyk, M., Liebe, R., Maier, I.B., et  al. The frequent adiponutrin
(PNPLA3) variant p.Ile148Met is associated with early liver injury:
analysis of a German pediatric cohort. Gastroenterol Res Pract,
2015;205079.
88. Grandone, A., Cozzolino, D., Marzuillo, P. et al. TM6SF2 Glu167Lys poly-
morphism is associated with low levels of LDL‐cholesterol and increased
liver injury in obese children. Pediatr Obes; 2016;11(2):115–9.
89. Goffredo, M., Caprio, S., Feldstein, A.E. et al. Role of TM6SF2 rs58542926
in the pathogenesis of nonalcoholic pediatric fatty liver disease: a multieth-
nic study. Hepatology, 2016;63(1):117–25.
90. Di Sessa, A., Umano, G.R., Cirillo, G. et al. The membrane‐bound O‐acyl-
transferase7 rs641738 variant in pediatric nonalcoholic fatty liver fisease. J
Pediatr Gastroenterol Nutr, 2018;67(1):69–74.
91. Umano, G.R., Caprio, S., Di Sessa, A. et  al. The rs626283 variant in the
MBOAT7 gene is associated with insulin resistance and fatty liver in
Caucasian obese youth. Am J Gastroenterol, 2018;113(3):376–83.
92. Krawczyk, M., Rau, M., Schattenberg, J.M. et al. Combined effects of the
PNPLA3 rs738409, TM6SF2 rs58542926, and MBOAT7 rs641738 variants
on NAFLD severity: a multicenter biopsy‐based study. J Lipid Res,
2017;58(1):247–55.
93. Koo, B.K., Joo, S.K., Kim, D. et  al. Additive effects of PNPLA3 and
TM6SF2 on the histological severity of non‐alcoholic fatty liver disease. J
Gastroenterol Hepatol, 2018;33(6):1277–85.
  94. Krawczyk, M., Portincasa, P., Lammert, F. PNPLA3‐associated steatohepa-
titis: toward a gene‐based classification of fatty liver disease. Semin Liver
Dis, 2013;33(4):369–79.
  95. Yki‐Järvinen, H. Non‐alcoholic fatty liver disease as a cause and a consequence
of metabolic syndrome. Lancet Diabetes Endocrinol, 2014;2(11):901–10.
  96. Mirnezami, R., Nicholson, J., Darzi, A. Preparing for precision medicine. N
Engl J Med, 2012;366(6):489–91.
  97. Valenti, L., Dongiovanni, P., Ginanni Corradini, S. et al. PNPLA3 I148M
variant and hepatocellular carcinoma: a common genetic variant for a rare
disease. Dig Liver Dis, 2013;45(8):619–24.
  98. Tian, C., Stokowski, R.P., Kershenobich, D. et  al. Variant in PNPLA3 is
associated with alcoholic liver disease. Nat Genet, 2010;42(1):21–3.
  99. Stickel, F., Buch, S., Lau, K. et al. Genetic variation in the PNPLA3 gene is
associated with alcoholic liver injury in caucasians. Hepatology,
2011;53(1):86–95.
100. Vigano, M., Valenti, L., Lampertico, P. et  al. Patatin‐like phospholipase
domain‐containing 3 I148M affects liver steatosis in patients with chronic
hepatitis B. Hepatology, 2013;58(4):1245–52.
101. Cai, T., Dufour, J.F., Müllhaupt, B. et  al. Viral genotype‐specific role of
PNPLA3, PPARG, MTTP, and IL28B in hepatitis C virus‐associated stea-
tosis. J Hepatol, 2011;55(3):529–35.
102. Müller, T., Buch, S., Berg, T. et al. Distinct, alcohol‐modulated effects of
PNPLA3 genotype on progression of chronic hepatitis C. J Hepatol,
2011;55(3):732–3.
103. Valenti, L. and Fargion, S. Patatin‐like phospholipase domain containing‐3
Ile148Met and fibrosis progression after liver transplantation. Hepatology,
2011;54(4):1484.
104. Nischalke, H.D., Berger, C., Luda, C. et al. The PNPLA3 rs738409 148M/M
genotype is a risk factor for liver cancer in alcoholic cirrhosis but shows no
or weak association in hepatitis C cirrhosis. PLoS One, 2011;6(11):e27087.
105. Younossi, Z., Anstee, Q.M., Marietti, M. et al. Global burden of NAFLD
and NASH: trends, predictions, risk factors and prevention. Nat Rev
Gastroenterol Hepatol, 2018;15(1):11–20.
106. Dubuquoy, C., Burnol, A.F., Moldes, M. PNPLA3, a genetic marker of pro-
gressive liver disease, still hiding its metabolic function? Clin Res Hepatol
Gastroenterol, 2013;37(1):30–5.
SECTION B:
LIVER GROWTH
AND REGENERATION
 Stem Cell‐Fueled Maturational
42 Lineages in Hepatic and
Pancreatic Organogenesis
Wencheng Zhang1, Amanda Allen1, Eliane Wauthier1, Xianwen Yi2,
Homayoun Hani1, Praveen Sethupathy3,10, David Gerber2, Vincenzo
Cardinale5, Guido Carpino6,7, Juan Dominguez‐Bendala8,9, Giacomo
Lanzoni9, Domenico Alvaro5,8,*, Eugenio Gaudio7,*, and Lola Reid1,4,*
1 
Department of Cell Biology and Physiology, University of North Carolina School of Medicine,
Chapel Hill, NC, USA
2 
Department of Surgery, University of North Carolina School of Medicine, Chapel Hill, NC, USA
3 
Department of Genetics, University of North Carolina School of Medicine, Chapel Hill, NC, USA
4 
Program in Molecular Biology and Biotechnology, University of North Carolina School of Medicine,
Chapel Hill, NC, USA
5 
Department of Medico‐Surgical Sciences and Biotechnologies, Sapienza University of Rome, Latina, Italy
6 
Department of Movement, Human and Health Sciences, Division of Health Sciences, University of Rome
“Foro Italico”, Rome, Italy
7 
Department of Anatomical, Histological, Forensic Medicine and Orthopedics Sciences, Sapienza University
of Rome, Rome, Italy
8 
Department of Medicine and Medical Specialties, Sapienza University of Rome, Rome, Italy
9 
Diabetes Research Institute, Miller School of Medicine, University of Miami, Miami, FL, USA
10 
Department of Biomedical Sciences, Cornell University, Ithaca, NY, USA

INTRODUCTION
Liver, biliary tree, and pancreas are mid‐gut endodermal organs
central to handling glycogen and lipid metabolism, detoxifica-
tion of xenobiotics, processing of nutrients for optimal utiliza-
tion, regulation of energy needs, and synthesis of diverse factors
ranging from coagulation proteins to carrier proteins (e.g.
alpha‐fetoprotein (AFP), albumin, transferrin) [1]. The integrity
* L.M. Reid was the only one who wrote most of the chapter with the qualifier
that the last segment, that on plasticity was written by Domenico Alvaro and Lola
Reid and with input also from Eugenio Gaudio. Amanda Allen, an artist and
animator, did the figures. All of the other authors have done the experiments
resulting in the discoveries that are summarized in the review. The senior authors
in the management of the experiments comprise G. Lanzoni and J. Dominguez‐
Bendala at the Diabetes Research Institute, D. Alvaro and E. Gaudio at Sapienza
Medical Center, P. Sethupathy at Cornell, and L. Reid at UNC.
of the body depends heavily on liver, biliary tree, and pancreatic
functions, and failure in any of them results in rapid death.
The focus of this review is on current knowledge of the newly
discovered network of multiple stem/progenitor cell populations
present primarily in the biliary tree and giving rise to maturational
lineages of cells forming liver and pancreas throughout life. In addi-
tion, the chapter summarizes information regarding key paracrine
signals from lineage‐dependent epithelial–mesenchymal cell inter-
actions in the maturational processes and found useful for ex vivo
maintenance and differentiation of hepato/biliary and pancreatic
cells at particular lineage stages. In prior reviews and previous edi-
tions of this book are given more details on hepatic stem cells and
hepatoblasts, on early studies on the biliary tree stem cells, and on
facets of extracellular matrix chemistry and biology, key require-
ments for ex vivo maintenance of the cells [1–9].
For the sake of brevity, we will not discuss findings regarding
the lineage restriction of embryonic stem (ES) cells or induced

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
524 THE LIVER:  THE BILIARY TREE AS A ROOT SYSTEM FOR HEPATIC/PANCREATIC ORGANOGENESIS
pluripotent stem (iPS) cells to an hepatic or pancreatic fate.
Information from these studies is provided in representative
recent articles and reviews [10–13].
This chapter focuses primarily on studies in human tissues
and with some references to similar investigations in mice, rats,
and/or pigs. The phenomena associated with stem/progenitor
cell maturational lineages in hepatic and pancreatic organogen-
esis are evident in all mammals studied (mice, rats, pigs, and
humans) but with a few distinctive variations in some species as
noted below.
The stem cell or progenitor cell populations are indicated by
an acronym which is preceded by a small letter indicating the
species: m, murine; r, rat; p, porcine; h, human.
EMBRYONIC DEVELOPMENT
During early development, definitive endoderm derives from
embryonic stem cells through the effects of a number of key
transcription factors including Goosecoid, MIXL1, SMAD2/3,
SOX7, and SOX17 (a transcription factor essential for differen-
tiation of liver) [14]. Endoderm subsequently segregates into
foregut (lung, thyroid), midgut (pancreas, biliary tree, and liver),
and both foregut and hindgut (intestine) through the effects of
specific mixes of transcription factors. Those dictating the mid-
gut organs include SOX9 (transcription factor associated with
endodermal tissues), SOX17, FOXA1/FOXA2 (forkhead box
a1/2), Onecut2/OC‐2, and others [15–18]. The liver, biliary tree,
and pancreas derive from midgut endoderm established at the
gastrulation stage of early embryonic development [19]. Among
the other organs of endodermal origin, endogenous adult stem
cells have been identified in most, including the small and large
intestines [20], the stomach [21], and the lungs [22, 23].
The pancreas is distinct in that lineage tracing experiments
indicate that there are only very rare stem cells postnatally [24–
26]. Clarification of this paradox has emerged with the many
studies on the biliary tree: the stem cells for the pancreas are
not  in the pancreas proper but rather within the biliary tree,
especially within the peribiliary glands (PBGs) of the hepato‐
pancreatic common duct. These pancreatic stem cells are
anatomically linked and give rise to committed progenitors
­ points in the biliary tree, in the cystic duct connecting the gall-
located in pancreatic duct glands (PDGs) within the pancreas
proper [27, 28].
During fetal development, the formation of the liver and pan-
creas occurs with outgrowths on either side of the duodenum
that extend and ramify into a branching biliary tree structure
that, at its ends, is influenced by the cardiac mesenchyme to
form liver and by the retroperitoneal mesenchymal to form pan-
creas [28]. The gallbladder is a major branch point connected to
the common duct via the cystic duct [29]. The pancreas derives
from two separate anlage: the dorsal pancreas connected to the
duodenum via the accessory duct. The ventral pancreas begins
as a branch from the common duct and nearer to the duodenum.
The formation of the intestine involves a twisting motion that
swings the ventral pancreas anlage to the other side where it
subsequently merges with the dorsal pancreas anlage to form
the complete pancreatic organ. The liver cannot swing to the
opposite side, given its size and its connections into the cardiac
mesenchyme, connections associated with rapid vascularization
of the forming liver tissue. This results in the liver and the
­ventral pancreas sharing the hepato‐pancreatic common duct
connecting them to the duodenum.
THE BILIARY TREE AS A ROOT SYSTEM
FOR HEPATIC/PANCREATIC
ORGANOGENESIS (GENERAL ANATOMY
AND IN SITU STUDIES)
Newly discovered within the last decade is that hepatic and pan-
creatic organogenesis are ongoing throughout life “fueled” by a
network of stem cells within the biliary tree, the ramifying ducts
connecting the liver and the pancreas to the duodenum [5, 9, 27,
30–33] (Figure 42.1). The biliary tree, long assumed to be the
conduits managing removal of bile from the liver and digestive
enzymes from the pancreas, has proven also to be a “root”
­system, a reservoir of stem cells giving rise to maturational
­lineages of cells contributing to hepatic and pancreatic organo-
genesis and to regenerative processes in these organs. The most
primitive of the stem cells are located within extramural peribil-
iary glands (PBGs, stem cell niches for biliary tree stem cells)
tethered to the outside of large ducts (i.e. in humans: greater
than 300 μm) [34]. Their functions are not known but are
hypothesized to be “seeds” for organogenesis given that they
sprout into ducts with regenerative demands (Reid and associ-
ates, unpublished observations).
The network of intramural (within the duct walls) niches
begins with the Brunner’s glands, submucosal glands found in the
duodenum [35, 36] (Cardinale et  al. personal communication).
These are located between the major papilla (the papilla of Vater),
the entranceway to the hepato‐pancreatic duct, and the minor
papilla, the port connecting the duodenum to the dorsal pancreatic
duct. The Brunner’s glands are not found elsewhere within the
intestinal tract. Indeed, they are used to define the transition from
the duodenum to the beginning of the small intestine.
The intramural network of stem cells extends throughout the
biliary tree in the PBGs found in high numbers behind branch
bladder to the common duct, in the large intrahepatic bile ducts,
and in the hepato‐pancreatic common duct [9, 30, 34, 37]. The
gallbladder does not have PBGs, but its stem cells are found in
crypts at the base of villi in the gallbladder [29]. The second
largest numbers of PBGs are found within the large, intrahe-
patic bile ducts and connect anatomically to the canals of
Hering and thence to the sinusoidal plates of parenchymal cells
within liver acini [34]. The largest numbers of PBGs are found
within the hepato‐pancreatic common duct and connect to the
PDGs within the ventral pancreas [27, 34]. Although not yet
well defined, they are found also in the accessory duct connect-
ing to the dorsal pancreas.
The network of stem cells transition to niches of bipotent,
transit amplifying cells, cells with considerable proliferative
potential but questionable self‐replicative ability. These com-
prise hepatoblasts, located next to the canals of Hering [38–41]
and to pancreatic ductal progenitors in the PDGs [42–44].
 42:  Stem Cell‐Fueled Maturational Lineages in Hepatic and Pancreatic Organogenesis 525

Figure 42.1  The anatomy of the connections between duodenum, biliary tree, gallbladder, liver, and pancreas in humans. Most of the anatomi-
cal features are self‐evident from the figure. The connections of the biliary tree to the duodenum are via two ports, the ampulla of Vater for the
hepato‐pancreatic common duct, and the minor duodenal papilla for the accessory duct (also called the duct of Santorini) connecting to the
dorsal pancreas.

These, in turn give rise to unipotent committed progenitors that
link to mature hepatic or pancreatic cell populations.
LINEAGE‐DEPENDENT EPITHELIAL–
MESENCHYMAL CELL PARTNERSHIPS
The cells throughout the biliary tree, liver, and pancreas are in
epithelial–mesenchymal cell partnerships that undergo coordi-
nate maturation resulting in lineage‐stage‐dependent paracrine
signaling influencing the stage‐dependent phenotypic traits as
summarized in more detail in prior reviews [1, 45, 46]. In brief,
the beginning stages are comprised of epithelial stem cells
­partnered with angioblasts (CD117+, prominin 1 [CD133]+,
vascular endothelial cell growth factor [VEGF]‐receptor+, Van
Willebrand Factor+, CD31-) [47]. These give rise to cellular
descendants that mature coordinately and then split into two
branches: epithelia partnered with endothelia (hepatocytes, islets)
and epithelia partnered with stellate cells and their descendants,
stroma and myofibroblasts (cholangiocytes, acinar cells). The
endothelial cell precursors (CD133+, VEGF‐receptor+, Van
Willebrand Factor+, CD31+) and the stellate cell precursors
(CD146+, intercellular adhesion molecule [ICAM]‐1+,
desmin+, alpha‐smooth muscle actin [ASMA]+, glial fibrillary
acidic protein [GFAP]-) give rise to descendants with distinct
phenotypic traits [1] The lineage‐dependent traits comprise
morphology, cell size, DNA content (ploidy), growth potential,
antigenic profile, gene expression, and lineage‐dependent par-
acrine signaling by the synergistic effects of extracellular matrix
components, matrix‐bound signals (growth factors and cytokines
linked primarily to glycosaminoglycan [GAG] chains bound to
core proteins in proteoglycans) and soluble signals present in
blood and interstitial fluid. The net sum of the activities of cells
at the sequential maturational lineage stages yields that for the
composite tissue.
The phenomena associated with stem cell networks in
hepatic/pancreatic organogenesis are common to all mammals.
However, there are some species‐associated distinctions.
Examples include the species‐specific variations with respect to
ploidy profiles in the liver (note: these have not been analyzed
yet in pancreas). The lineages in all mammals begin in fetal and
neonatal tissues with entirely diploid cells that transition to
increasing proportions of polyploid cells in adulthood and espe-
cially in geriatric hosts. Within the liver acinus, diploid cells are
always found periportally; the polyploid cells are always found
pericentrally. What varies is the proportion of the cells of the
liver plates that are polyploid. In humans, by around 20 years of
age, the shift is to a minor fraction of tetraploid cells, all in zone
three of the liver acinus; in rats, by four weeks of age, it transi-
tions to predominantly tetraploid cells (80%), all in zone two
and with a small fraction (10%) of octoploid cells in zone three;
and in mice the shift occurs by three to four weeks of age and
results in 97% polyploid cells comprised mostly of tetraploid
and octoploid in a part of zone one, all of zone two, and with a
minor fraction of cells that are 16N and 32N in zone three. The
exceptions to this ploidy profile are the newly discovered dip-
loid parenchymal cells linked on their lateral borders to endothe-
lia encompassing the central vein; these constitute a reservoir of
committed (unipotent) hepatocytic progenitors that replace
apoptotic (senescing) hepatocytes [48].
Another notable variation is that pigs and oxen do not have an
hepato‐pancreatic common duct, the location of the stem cell
niches for the ventral pancreas for all other mammals. Ongoing
studies are exploring possible locations for these stem cell
niches for the ventral pancreas in pigs.
526 THE LIVER:  RADIAL AXIS OF MATURATION
MICROENVIRONMENT OF THE STEM
CELL AND PROGENITOR CELL NICHES
Stem and progenitor cells reside in discrete locations called niches,
with unique microenvironments that are poorly defined for the
hepato/biliary/pancreatic network (Table  42.1 and Figure  42.2).
The niches include PBGs in the extrahepatic and intrahepatic
­biliary tree [4, 27, 30, 34] and in the cystic duct connecting the
gallbladder to the biliary tree [29]; the ductal plates in fetal and
neonatal livers and transitioning into the canals of Hering in pedi-
atric and adult livers [49–53]; and the PDGs found throughout the
pancreas but with an especial concentration in the head of the pan-
creas [28, 54–56]. The gallbladder does not have PBGs, but does
have crypts at the base of villi and contains stem cells similar to
late stage biliary tree stem cells (BTSCs). Collectively, these
niches form a network that is continuous throughout the biliary
tree and anatomically connects directly through to the canals of
Hering to the sinusoidal plates within the liver and through to the
PDGs, the reservoirs of committed progenitors within the pan-
creas, and then to the acinar cells and islets.
As noted above, a key paradigm to the tissue organization is the
epithelial–mesenchymal cell partnerships. It can be mimicked in
vitro by use of feeder cells of the relevant mesenchymal type or by
use of defined mixes of matrix components and soluble signals [47].
Matrix and soluble signals in the stem cell and progenitor cell niches
have been only partially defined [40, 41, 47, 57, 58]. That which is
known is summarized in Table 42.1 and in Figure 42.2. The known
matrix chemistry of these stem/progenitor cell niches is domi-
nated by hyaluronans (HA) [59], non‐sulfated glycosaminoglycans
(GAGs), and minimally sulfated forms of chondroitin sulfate prote-
oglycans (CS‐PGs) [60]. In association with the transit amplifying
cells, there are also fetal collagens (e.g. type IV), and fetal adhesion
molecules (e.g. fetal laminins) along with more sulfated proteogly-
cans (e.g. heparan sulfate proteoglycans [HS‐PGs] and CS‐PGs)
[61, 62]. Receptors for one or another of the known hyaluronan
receptors (e.g. CD44) are common features of the stem cells [63].
With maturation to adult cell types, there is a branching to separate
lineages (hepatocytes, islets) associated with endothelia versus chol-
angiocytes and acinar cells associated with stellate cells. The matrix
chemistry changes parallel the split: those in epithelial–endothelial
relations contain matrix chemistry dominated by network collagens
(type IV, type VI), by forms of laminin, and by HS‐PGS with
increasing sulfation resulting in the expression of heparin proteogly-
cans (HP‐PGs) associated with the most mature cells (polyploid
hepatocytes, mature islets). The branch of the epithelia‐stellate cells
gives rise to cholangiocytes and acinar cells and is associated with a
matrix chemistry comprised of fibrillar collagens (e.g. type III, type
I), adhesion molecules (forms of fibronectins, entactin), and with
CS‐PGs that transition in late stages of cells to highly sulfated forms,
dermatan sulfate proteoglycans (DS‐PGs).
PERIBILIARY GLANDS
The PBGs occur throughout the biliary tree as intramural glands,
found within the bile duct walls, and as extramural glands that
are tethered by extensions to the large bile ducts [64]. They
occur in highest frequencies at the branching points of the
b­iliary tree, with the greatest numbers being present in the
hepato‐pancreatic common duct and the large intrahepatic
bile ducts [9, 34]. Other than findings from the pioneering stud-
ies of Nakanuma and associates [64–66], almost nothing used to
be known of the possible roles of the PBGs until the recent stud-
ies within the last decade analyzing them as reservoirs, crypts,
containing stem cell populations. Each PBG contains a ring of
cells at its perimeter and is replete with mucous production and
periodic acid‐Schiff (PAS)‐positive material in its center. The
cells in the ring are phenotypically quite homogeneous in
the PBGs in some sites (e.g. near the fibromuscular layers of the
hepato‐pancreatic common duct and the large intrahepatic bile
ducts), but are quite heterogeneous in other sites (e.g. neared to
the lumens of all ducts, especially the cystic duct, hilum, com-
mon duct). The pattern of the variations implicate maturational
lineages for which there are two axes [27, 34].
RADIAL AXIS OF MATURATION
A radial axis of maturation is evident within the duct walls (the
intramural PBGs), beginning with the cells in PBGs near the fibro-
muscular layer (center of the duct walls) and ending with cells at
the ducts’ lumens (Figure 42.3). The most primitive of the stem
cells, those in the PBGs at the fibromuscular layer, have no mark-
ers of adult liver or pancreas but instead have moderately high lev-
els of expression of pluripotency genes (e.g. SOX2 [a transcription
factor that is essential for maintaining self‐renewal, or pluripo-
tency in embryonic and determined stem cells], SALL4 [Sal‐like
protein 4 found to be important for self‐replication of stem cells],
BMI-1, NANOG, KLF4/5, octamer‐binding transcription factor 4
[OCT4, also known as POU5F1 {POU domain, class 5, transcrip-
tion factor 1}, a gene expressed by stem cells]); co‐expression of
transcription factors for both liver and pancreas (e.g. SOX17,
PDX1 [pancreatic and duodenal homeobox 1, a transcription fac-
tor critical for pancreatic development]); express classic markers
indicative of active proliferation such as Ki‐67; have high levels of
CD44 (hyaluronan receptors); and express sodium iodide sym-
porter (NIS). NIS is hypothesized to transport anionic intermedi-
ates needed in the synthesis of hyaluronans [67].
As one progresses toward the lumen, the cells in the PBGs
lose expression of the pluripotency traits, diminish the evidence
for proliferation, and acquire intermediate markers associated
with stemness such as LGR5 and epithelial cell adhesion mole-
cule (EpCAM). At the lumens, all traits of stemness have faded
and been replaced with hepatic traits in the vicinity of the car-
diac mesenchyme or pancreatic traits for ducts in the vicinity of
the retroperitoneal mesenchyme.
We hypothesize that a new nomenclature will be required
for the PBGs and PDGs, since they have traits that are less
those of glands and instead have features of crypts. The phe-
nomenon parallels that observed in the intestinal crypts con-
taining stem cells that mature in a radial axis along the villi to
yield adult intestinal cells (e.g. enterocytes and goblet cells).
Temporarily we will abstain from converting to a new nomen-
clature to enable further studies to be done that should help
with establishing a logical new nomenclature for the compo-
nents of the network.
 Table 42.1  Representative biomarkers of major lineage stages of precursors in organogenesis of liver and pancreas

Periportal region of PBGs in large intrahepatic PBGs in extrahepatic PBGs in hepato‐

liver acini Canals of Hering bile ducts biliary tree pancreatic common duct PDGs in pancreas
Lineage stage of cells Hepatoblasts, hepatic HpSCs Stage 1–3 BTSCs Pancreatic ductal progenitors
committed progenitors HpSCs PSCs
Endodermal transcription Sox9 SOX 9, SOX 9, SOX 17, PDX1 in Stages 1 and 2 BTSCs SOX 9, PDX1
factors SOX 17 SOX 9 and SOX 17 in Stage 3 SOX9 and PDX in
Stage 3
Pluripotency genes Low to none Intermediate level of Moderate to strong expression in stages 1 and 2 BTSCs; Intermediate level of None
expression intermediate level of expression in stage 3 BTSCs expression
OCT4, SOX2, NANOG, SALL4, BMi‐1, KLF4/KLF5
Cell adhesion molecules EpCAM, EpCAM, NCAM NCAM in stage 1 and 2 BTSCs, EpCAM and NCAM in stage 3 BTSCs EpCAM, ICAM‐1
ICAM‐1
Other stem cell markers Weak CXCR4 CD133, LGR5 CXCR4 in stages 1 and 2 BTSCs, CD133 in all of the stages of BTSCS; LGR5 in CXCR4, CD133, CD24
CD133, LGR5 stages 2 and 3 BTSCs
Hedgehog proteins Weak Indian and Sonic Strong Indian and Not yet studied Weak Sonic
Sonic
Matrix proteins Laminin, type IV collagen Laminin 5, type III Not yet studied Islets: network collagens; acinar
collagen cells: fibrillar collagens
GAGs/PGS HA, CD44, syndecans HA, CD44, minimally HA, CD44, others not yet studied Islets: syndecans and glypicans;
(HS‐PGs and CS‐PGs) sulfated CS‐PGs acinar cells: CS‐PGs, DS‐PGs
Multidrug resistance MDRI‐negative; MDR1 moderate, Not yet studied None
genes ABCG2‐moderate ABCG2 very strong
Hepatic traits Albumin +, Albumin +/‐ Negative for hepatic traits
AFP +++, P450A7, AFP‐, P450A7‐
Glycogen
Pancreatic traits Negative for pancreatic traits ISL1, PROX1, NGN3, MAFA, MUC6,
NeuroD, PAX4 Nkxx6.1/NKx6.2, Ptf1a, Glut2
HA, hyaluronans; HS‐PGs, heparan sulfate proteoglycan; CS‐PGs, chondroitin sulfate proteoglycan; DS‐PGs, dermatan sulfate proteoglycan; syndecans, proteoglycans with a transmembrane core protein;
glypicans, proteoglycans linked to plasma membrane by PI linkages; MDR1, multidrug resistance genes; HpSCs, hepatic stem cells; HBs, hepatoblasts; BTSCs, biliary tree stem cells; PSCs, pancreatic
stem cells; PBGs, peribiliary glands; PDGs, pancreatic duct glands.
All of the known stages of the BTSCs are in the large intrahepatic bile ducts including in the extrahepatic biliary tree and in the hepato‐pancreatic common duct. The precursors to these are in the
extramural peribiliary glands and in the Brunner’s glands in the submucosa of the duodenum.

0004435634.INDD 527 11/5/2019 6:46:31 PM

528 THE LIVER:  RADIAL AXIS OF MATURATION

A. Stem Cell Niches relevant to Hepatic and Pancreatic Organogenesis

Stem Cells = self-replicate; multipotent; all express moderate levels of pluripotency genes; all express one or more of the isoforms of
CD44 (hyaluronan receptors) and the most primitive ones found also express NIS (sodium iodide symporter). NIS is hypothesized
to be a transporter of anionic intermediates in the hyaluronan biosynthesis pathway.

Determined Endodermal Stem Cell subpopulations identified to date:

Extramural Peribiliary Gland Stem Cells: extremely primitive. With regenerative demands, produce ductules.
Intramural Stem Cell Populations:
Brunner’s Glands’ Stem Cells --submucosa of the duodenum; extremely primitive
Biliary Tree Stem Cells (BTSCs). Three phenotypically distinct stages within bile duct walls
Stages: 1) NIS+, Negative for LGR5 and EpCAM 2) NIS+, LGR5+ EpCAM- 3) NIS-, LGR5+, EpCAM+
Gallbladder stem cells --similar to stage 3-BTSCs
Hepatic stem cells --primarily within canals of Hering and in PBGs of biliary tree
Pancreatic stem cells --primarily within PBGs of hepato-pancreatic common duct
Note: Only committed progenitors found within pancreas proper (so, no true stem cells within pancreas)

Epithelial-mesenchymal Cell Partnerships. All are comprised of epithelial stem cells coupled with mesenchymal cell partners;
there is lineage-stage-specific paracrine signaling and coordinate maturation. Stem Cell stages= epithelial stem cells partnered with
angioblasts

Stem Cell Niche Microenvironment = known components comprise hyaluronans, non-sulfated glycosaminoglycans, plus minimally
sulfated chondroitin sulfate proteoglycans (no heparan sulfate proteoglycans).

B. Intermediates In Hepatic and Pancreatic Organogenesis

Transit Amplifying Cells = highly proliferative but debatable if they self-replicate. Weak to negligible levels of pluripotency genes.

Hepatoblasts --bipotent (yield hepatocytes and cholangiocytes); signature feature= alpha-fetoprotein (AFP); located adjacent
to canals of Hering; anatomically connected through the hepatic stem cells to network of stem/progenitor cells
within peribiliary glands (PBGs) of the large intrahepatic bile ducts.

Pancreatic Ductal Progenitor Cells – bipotent (yield acinar cells and islets); found within pancreatic duct
glands (PDGs) within the pancreas proper; are anatomically connected to the network of stem cells within PBGs of
the hepato-pancreatic common duct of the biliary tree.

Committed Progenitor Cells = do not self-replicate; highly proliferative; unipotent; no expression of pluripotency
genes. Anatomically linked to the mature cells

Epithelial-mesenchymal cell partnerships:

Hepatocytic or islet progenitors and descendants coupled to endothelial cell lineage stages
Cholangiocytic or acinar progenitors and descendants coupled to stellate cell lineage stages; late
lineage stages are associated with myofibroblasts

Niche Microenvironment = known components include hyaluronans + minimally sulfated sulfated chondroitin
sulfate proteoglycans and heparan sulfate proteoglycans + laminin + type IV collagen

Figure 42.2  The stem/progenitor cell niches in the biliary tree, liver, and pancreas. (a) A chart shows the known connections between the niches
throughout the biliary tree and within the organs. (b) Summarizes the features of the stem cell niches. (c) A summary of the transit amplifying cells.

The maturational progression of the cells is always within the

Proximal‐to‐distal axis in maturation
A proximal‐to‐distal axis in the maturation occurs beginning at
the duodenum with the stem cells in the Brunner’s glands; con-
nects into the stem cells within the PBGs within the hepato‐­
pancreatic common duct; and then transitions into the pancreas
to connect with the PDGs. The branch for the liver is via the
common duct and once past the cystic duct leading to the gall-
bladder connects into the PBGs of the large intrahepatic bile
ducts. These give rise to links with the canals of Hering that
transition into the sinusoidal plates of the liver acini.
radial axes in the duct walls, but that in or near the liver is within
the domain of the cardiac mesenchyme and results in mature
cells of an hepatic fate. The radial axes within ducts near the
pancreas are in the domain of the retroperitoneal mesenchyme
and result in mature cells of pancreatic fates. Those in between
liver and pancreas yield cells with mature bile duct markers. The
implications are that isolated BTSCs, especially the most primi-
tive of them, should be able ex vivo to give rise both to liver and
pancreas if provided the requisite paracrine signals from either
cardiac mesenchyme versus retroperitoneal mesenchyme. This
 42:  Stem Cell‐Fueled Maturational Lineages in Hepatic and Pancreatic Organogenesis 529

C. Hepato/Pancreatic Stem/Progenitor Cell Niches

Extramural PBGs Brunner’s Glands
Tethered to outside of large Bile Ducts (Submucosa of the Duodenum)
Very primitive Endodermal Stem Cells BGSCs, CPs

Common Duct Intramural PBGs Hepato-Pancreatic Common Duct

(Extrahepatic Biliary Tree)
BTSCs, HpSCs, PSCs, CPs

PBGs (Large Intrahepatic Bile Ducts) PBGs (Cystic Duct) PBGs (Hepato-pancreatic
BTSCs, HpSCs, HBs, CPs BTSCs, CPs Common Duct)
BTSCs, PSCs, CPs

Ductal Plates Crypts-bottoms of villi

Pancreatic Duct Glands
Canals of Hering (Gallbladder) (Pancreas)
(Liver) Stage 3-BTSCs, CPs Pancreatic Ductal Progenitors (CPs)
HpSCs, HBs, CPs

Hepatocytes and Cholangiocytes Cholangiocytes Islets and Acinar Cells

BGSCs = Brunner’s glands stem cells
BTSCs = biliary tree stem cells (3 stages)
PBG= Peribiliary Gland HpSCs = hepatic stem cells; HBs = Hepatoblasts
Extramural PBGs = tethered to duct wall PSCs = pancreatic stem cells
Intramural PBGs = inside the duct wall CPs = committed progenitors (unipotent)

Figure 42.2  (Continued)

(a)

Figure 42.3  The radial axis of maturation in (a) liver, (b) pancreas, and (c) gallbladder. The peribiliary glands (PBGs) near to the fibromuscular layer
within the bile duct walls are crypts containing the most primitive of the stem cells, ones that can give rise both to liver and to pancreas. There are hints
that the cells move along the walls of the PBGs and progress upwards towards the lumens of the ducts. With that progression, the phenotypic traits
transition from those for stem cells to those for mature cells. Some of the traits for the cells in the radial axis maturation are noted. The gallbladder does
not have PBGs, but does have late stage BTSCs located at the base of villi in the gallbladder. These move upwards toward the tops of the villi.

is indeed what has been found [27, 30]. It also implicates the
potential for using the BTSCs in cell therapies for both liver and
for pancreas, an hypothesis that is currently under investigation.
The network provides a biological framework for ongoing
organogenesis of liver, biliary tree and pancreas throughout life.
These phenomena parallel the well‐described, intestinal mat-
urational lineage system. The radial axis of maturation in the
intestine progresses from stem cells in the crypts to fully dif-
ferentiated cells at the tops of the villi. The proximal‐to‐distal
axis follows the length of the intestine and results in distinct
530 THE LIVER:  EX VIVO STUDIES ON THE STEM/PROGENITORS
(b)
(c)
Figure 42.3  (Continued)
mature cells depending on whether the radial axis is located in
the esophagus, stomach, duodenum, small or large intestine.
EX VIVO STUDIES ON THE STEM/
PROGENITORS
The in situ and anatomical findings summarized above are com-
plemented by ex vivo (monolayer and organoid) cultures of the
various lineage stages including those of Brunner’s glands
(Cardinale et al. personal communication), of two of the BTSC
stages [30], of the hepatic stem cells (HpSCs) and hepatoblasts
(HBs) [1, 39, 40, 68], pancreatic ductal progenitors and islets
[28], and adult hepatocytes and adult cholangiocytes [47, 69].
The culture conditions required for the different lineage
stages are distinct. The stem cell stages require conditions with
soluble forms of hyaluronans for organoids and with hyaluronan
hydrogel substrata for monolayers or for embedded organoids;
these hydrogel must be very soft, under 100 Pa to maintain
stemness traits [68]. The medium required for the stem cells is
serum‐free, devoid of growth factors and cytokines, and tailored
for the cells at this stage. One of the best ones was that estab-
lished by Hiroshi Kubota [38]. It is comprised of any rich basal
medium with low calcium (around 0.3 mM), no copper, sele-
nium (10−10 M), zinc (10−12 M), insulin (around 5 μg mL−1),
transferrin/Fe (around 5 μg mL−1), high density lipoprotein
(around 10 μg mL−1), and a defined mixture of purified free fatty
acids bound to highly purified albumin. Mature cells do not
­survive in Kubota’s medium, only the stem cells from both
 42:  Stem Cell‐Fueled Maturational Lineages in Hepatic and Pancreatic Organogenesis 531
epithelial and mesenchymal cell lineages. However, more rapid
expansion of the mesenchymal partners of the epithelial stem
cells, such as the angioblasts, and conditions permissive for self‐
replication of both are facilitated by the addition of leukemia
inhibitory factor (LIF) and by VEGF [30, 38, 41, 47, 69]. Thus,
the conditions co‐select for any of the true stem cell subpopula-
tions and their mesenchymal cell partners, angioblasts and their
descendants, precursors of stellate cells, or endothelia [38, 47]
Under these conditions, we observed two major types of
BTSC colonies in cultures that correlate with stage two and
stage three BTSCs from the in situ studies; we have yet to iden-
tify conditions for the stage one BTSCs. The stage two category
consist of cells that undulate (“dancing cells”), are very motile,
and initially do not express EpCAM (CD326) but acquire it at
the edges (the perimeters) of the colonies, corresponding to
slight cellular differentiation [30]. These are precursors to stage
three BTSCs that show uniform expression of EpCAM from the
outset and display a carpet‐like appearance with cells of uni-
form morphology [30]. The stage three BTSCs found through-
out the biliary tree are also the same or similar to the stem cell
subpopulation found in the gallbladder [29].
The expansion potential of the cells in culture in Kubota’s
medium is considerable: two to three cells can grow to colonies
of more than 500 000 cells in around eight weeks [30]. The cells
retain a stable stem cell phenotype (i.e. self‐renew) throughout
months of culture, and may be subcultured (“passaged”). Initially
cells show a typical division time of about one to two days, but
within a week, they slow to a division every two to three days. At
eight weeks the colonies contain cells in the centers that are mor-
phologically uniform, small (7–9 μm) and express high levels of
stem cell markers. Cells at the edges of the large colonies are
slightly larger (around 10–12 μm), have weak expression of
EpCAM and expression of markers intermediate in the differen-
tiation pathways, indicating potential loss of stemness and transi-
tion to more mature progenitors.
Using three‐dimensional (3D) hyaluronan hydrogels contain-
ing appropriate signaling molecules, the BTSCs can be induced
to differentiate to hepatocytes, cholangiocytes, or pancreatic
neo‐islets [30]. We have not succeeded yet to lineage restrict the
BTSCs to acinar cells. The differentiation is achieved by embed-
ding the BTSCs in specific mixes of extracellular matrix
­components (hyaluronans and type I collagen for bile ducts;
hyaluronans and type IV collagen and laminin for hepatocytes
or islets) and providing a serum‐free, hormonally defined
medium (HDM) tailored for each specific transit amplifying
cells or mature cell type.
The HDM are prepared by supplementing Kubota’s medium
with copper (10−12 M), higher calcium (0.6 mM), bFGF (basic
fibroblast growth factor, 10 ng mL−1) and then adding a unique
set of hormones and growth factors for hepatoblasts (EGF,
hepatocyte growth factor [HGF], T3, soluble [or substrata]
forms of type IV collagen and laminin), hepatocytes (substrata
of type IV collagen and laminin plus glucagon, galactose, T3,
oncostatin M, HGF, EGF, glucocorticoids), cholangiocytes
(substrata of type I and III collagen + HGF, EGF, VEGF, gluco-
corticoids), versus for pancreatic islets (soluble [or substrata]
forms of type IV collagen and laminin) plus B27, ascorbic acid,
cyclopamine, retinoic acid, HGF, and, after four days, replace-
ment of bFGF with Exendin‐4] [30].
The gene expression profiles of cells in 3D hydrogels com-
plement the morphological observations. For example, cells
cultured under conditions for hepatocytes produce albumin,
transferrin, and P450s. Cells in conditions for cholangiocytes
express anion exchanger 2 (AE2), cystic fibrosis transmem-
brane conductancy regulator (CFTR), gamma glutamyl trans-
peptidase (GGT), and secretin receptor. Cells in conditions
for pancreatic islets express transcription factor PDX1 and
the hormones glucagon, somatostatin, and insulin. Specific
staining for human C‐peptide confirms de novo synthesis of
proinsulin, and its secretion can be regulated in response to
the level of glucose. In vivo studies in which the cultured cells
are transplanted provide further evidence for the multipotency
of the BTSCs for hepatic, biliary tree, and pancreatic fates
[4, 27, 30, 70].
To confirm endocrine pancreatic differentiation, pre‐
induced neo‐islet structures were implanted into mouse fat
pads, and the animals were treated with a toxin (streptozo-
tocin) at a dose sufficient to destroy their own pancreatic beta
cells, but not human beta cells. Those mice transplanted with
the human neo‐islets showed significant resistance to hyper-
glycemia compared to controls that did not receive cell
­therapy. The presence of functional beta‐like cells derived
from the biliary tree stem cells produced serum‐levels of
human C‐peptide, that was regulated appropriately in response
to a glucose challenge [30]. Further studies have confirmed
and expanded upon these initial findings, leading us to con-
clude that the hepato‐pancreatic common duct is the major
reservoir of stem cells giving rise to committed progenitors
found in pancreatic duct glands, and thence to pancreatic
islets throughout life [27].
HEPATIC STEM CELLS
Those familiar with the myth of Prometheus will recall that the
liver possesses a remarkable capacity for regeneration [2, 3, 71].
Yet liver diseases, potentially leading to organ failure due to
hepatitis viruses, alcohol consumption, diet and metabolic dis-
orders, and other causes, constitute a major medical burden
[72–74].
Cell‐based therapies and tissue engineering represent possi-
ble approaches to address these needs [1, 75–78]. Sourcing of
cells for such applications is a significant challenge. In some
countries it is possible to obtain fetal tissues. In others neonatal
or adult tissues can be used. Given the newly discovered source
of stem cell populations in the biliary tree and precursors for
both liver and pancreas, the need for liver or pancreatic tissue as
a source is reduced. There is likely to be a transition to biliary
tree tissue as a primary source for clinical programs, and it is
considerably easier to obtain.
The role(s) of hepatic stem cells in the normal maintenance
of the liver and in regeneration from various insults remains a
subject of active research and debate [52, 53, 71, 78–85].
Although there is now general acceptance that hepatic stem
cells exist postnatally, their relevance is compared and debated
with that of the plasticity of the postnatal parenchymal cells
[86]. Further discussion of this is given below.
532 THE LIVER:  HEPATIC STEM CELL ISOLATION AND EXPANSION
HEPATIC STEM CELL ISOLATION
AND EXPANSION
Information on the findings on HpSCs is important, given that
human hepatic stem cells (hHpSCs) have already been used in
clinical trials in India and in Italy for patients with diverse liver
diseases and conditions [70, 87–89]. The isolation of hHpSCs
from fetal, neonatal, and adult human livers was achieved by
immunoselection with a monoclonal antibody for the surface
marker EpCAM [41]. These cells constitute approximately one
percent (0.5–1.5%) of the total liver population from early
childhood onwards. Unlike mature hepatocytes, they survive
extended periods of ischemia, allowing collection even several
days after cardiac arrest [90]. The hHpSCs express additional
surface markers often found on stem/progenitor cells, such as
CD133 (prominin), CD56 (neural cell adhesion molecule,
NCAM), and CD44 (the hyaluronan receptor); they also express
characteristic endodermal transcription factors SOX9, SOX17,
and HES1. They are small (diameter 7–9 μm, which is less than
half that of mature parenchymal cells) and express weak or neg-
ligible levels of adult liver‐specific functions such as albumin,
cytochrome P450s, and transferrin. The hHpSCs display far
greater capacity to proliferate in culture than hepatocytes or
cholangiocytes, and can continue to expand for months with a
doubling time of 36–40 hours. The colonies that form look
remarkably similar to those of embryonic stem (ES) cells or
induced pluripotent stem (iPS) cells [5, 1, 47, 91].
The hHpSCs serve as immediate precursors of human hepa-
toblasts (hHBs). The hHBs are readily distinguished from hHp-
SCs by the expression of AFP and intercellular adhesion
molecule‐1 (ICAM‐1), for which the hHpSCs are entirely nega-
tive [41, 47, 51]. The hHBs, in turn, are precursors of committed
unipotent progenitors for hepatocytes and cholangiocytes. When
injected into livers of immune‐deficient mice, the ­hHpSCs give
rise to cells expressing characteristic human liver and bile duct
proteins, especially after the host’s liver has been damaged by
treatment with carbon tetrachloride [41].
Whereas there has been limited success to achieve ex vivo
expansion of hematopoietic stem cells, the hHpSCs proliferate
for a sustained period in Kubota’s medium [38] which, as stated
previously, contains no additional growth factor or cytokine.
Pathways important for hHpSC survival in vivo, such as
Hedgehog (Hh) signaling [40], are activated through autocrine
loops. The expanded hHpSCs maintain a stable marker pheno-
type and also express the enzyme telomerase. The telomerase
mRNA and the protein encoded are localized to the nucleus
of the hHpSCs and the hHBs and telomeric enzymatic activity
correlate well with both the mRNA and protein levels [92].
However, later lineage stages (committed progenitors to late
­lineage stage mature cells) have no evidence of synthesis of tel-
omerase but have large amounts of telomerase protein localized
cytoplasmically. Telomeric enzymatic activity does not corre-
late with total telomerase protein levels; we hypothesize that it
correlates with the nuclear levels of the protein. Therefore, we
further hypothesize that regenerative demands will result in
small amounts of the cytoplasmic reserves of telomerase relo-
cating to the nucleus. If we are correct, the enzymatic activity
levels should correlate with the amount of telomerase (protein)
in the nucleus [92].
More recently, we have observed that Sal‐like protein 4
(SALL4, found to be important for self‐replication of stem cells) is
strongly expressed in the hBTSCs and hHpSCs but not in commit-
ted progenitors of either liver or pancreas [93]. SALL4 is a mem-
ber of a family of zinc finger transcription factors and a regulator
of embryogenesis, organogenesis, and pluripotency. It can elicit
reprogramming of somatic cells and is a marker of stem cells.
A crucial prerequisite for successful expansion of hHpSCs is
to mimic an appropriate microenvironment. When selected for
growth in vitro on tissue culture plastic and in Kubota’s medium,
the hHpSCs grow as colonies with feeders of angioblasts
(CD117+, VEGF‐receptor+, CD133+, Von Willebrand Factor+)
[41, 47, 58]; the feeders can be replaced with soft forms of hya-
luronans (under 100 Pa) and type III collagen [47, 68, 91, 94].
The cells expand for months under these conditions. By con-
trast, hHBs survive for only about a week under these same con-
ditions, but they can survive if they are co‐cultured with stellate
cell precursors (CD146+, alpha‐smooth muscle actin+, desmin+,
VCAM+, ICAM‐1+, GFAP‐negative) or feeders of mesenchy-
mal stem cells (MSCs). The stellate [95] feeder cells (or feeders
of MSCs) can be replaced with hyaluronans, type IV collagen,
and/or laminin [47, 96, 97]. The medium and matrix conditions
described above allow for flow cytometrically‐purified hHpSCs
or hHBs to survive and proliferate in culture and without the
need for feeders. Both type III collagen and hyaluronans are
constituents of the normal liver stem cell niche [47, 57, 91].
A systematic study of hHpSC behavior in 3D cultures using
hyaluronan hydrogels of differing stiffness indicated that rigid-
ity of the microenvironment is an important parameter in regu-
lating maintenance of stemness versus differentiation to more
restricted progenitors [68]. This was studied previously in dif-
ferentiation of progenitors for bone and other hard tissues, but
this report is the first for internal organs such as the liver.
The hHpSCs, like human ES cells, grow in tight colonies.
Dissociating either type of stem cells has proven to be an impor-
tant practical problem for their efficient expansion ex vivo and
for cryopreservation [98]. When treated enzymatically to gener-
ate a single cell suspension, both of these stem cell types undergo
a high level of cell death. Ding’s laboratory screened for chemi-
cals that would enable ES cells to survive enzymatic dissociation
and remain pluripotent. They identified two compounds, a
2,4‐disubstituted thiazole (Thiazovivin) and a 2,4‐disubstituted
pyrimidine (Tyrintegin), that met these criteria [99]. They found
that Thiazovivin inhibits the Rho‐associated kinase (ROCK), a
key component of the pathway that controls cytoskeleton remod-
eling, and a likely regulator of cell–ECM and cell–cell interac-
tions. Tyrintegin enhances attachment of dissociated ES cells to
ECM and stabilizes E‐cadherin. The investigators concluded that
ES cell interactions in the normal niche generate signals essen-
tial to survival, and that small molecules modulating those sig-
nals can maintain viability of dissociated cells.
Interestingly, we have observed that hyaluronans, a normal
component of stem cell niches, can protect hHpSCs during dis-
sociation and cryopreservation [98]. Thus, hyaluronans, a natural
molecule, can mediate the needed protection of the cells as well
as the artificial ones described above. The addition of hyaluro-
nans was found to protect cell adhesion mechanisms including
the hyaluronan receptor, E‐cadherin, and certain integrins, mark-
ers shared by many other stem cell subpopulations [100].
 42:  Stem Cell‐Fueled Maturational Lineages in Hepatic and Pancreatic Organogenesis 533
THE NEED FOR GRAFTING STRATEGIES
IN TRANSPLANTATION OF CELLS
FROM SOLID ORGANS
Cell therapies attempted to date are usually by delivery of the
cells by a vascular route. This is logical for hemopoietic cell
types. Long‐lived hemopoietic cells have evolved to be able to
flip between splice forms of matrix molecules (e.g. fibronec-
tins) from ones with no cell binding domains (resulting in cells
floating freely in blood or interstitial fluid) versus ones with
cell binding domains (resulting in attachment  –  a process
referred to as “homing”). By contrast, transplantation of cells
from solid organs involves cell types in which their attachment
proteins always have cell binding domains and so they rapidly
(within seconds) aggregate. When delivered to a target organ/
tissue by a vascular route, the aggregates cause an embolus
essential for engraftment but if too large resulting potentially in
lethal consequences for the host. Moreover, even when suc-
cessful, there is inefficient engraftment (typically around
10–20%) and with the remainder of the cells either dying or
dispersing to ectopic sites [94].
Our studies and those conducted by many others have found
that infusion of mature hepatocytes achieves only around 20%
engraftment if injected into the portal vein of the liver [73, 101,
102]. Stem cells are even more challenging, with approximately
only 3% of the cells engrafting if administered via the portal
vein (or via the spleen that connects directly to the portal vein).
This can be improved to around 20–25% engraftment in the
liver if the hHpSCs are injected into the hepatic artery [89]. The
remaining majority of the hHpSCs either die or engraft in
ectopic sites, most commonly the lung. Cells that lodge in the
vascular beds of ectopic sites can survive for months [100], a
finding of unknown significance at this time, but of potential
clinical concerns.
We have devised grafting strategies for transplantation of
hHpSCs embedded into a mix of soluble signals and extracel-
lular matrix biomaterials (e.g. hyaluronans) found in stem cell
niches [100]. The hHpSCs maintain a stable stem cell pheno-
type under the graft conditions. The grafts were transplanted by
injection grafting into the livers of immuno‐compromised
murine hosts, with and without carbon tetrachloride treatment,
to assess the effects of quiescent versus injured liver conditions.
Grafted cells remained localized to the livers, resulting in a
larger bolus of engrafted cells in the host livers under quiescent
conditions and demonstrated more rapid expansion upon liver
injury. We therefore have proposed grafting as a preferred strat-
egy for cell therapies for solid organs such as liver [94, 100].
Ongoing studies are resulting in assessment of other forms of
grafting (e.g. patch grafting) that enable transplantation of large
numbers of cells.
DIFFERENTIATION
The pharmacology of stem cell differentiation also must encom-
pass both soluble signals (i.e. conventional biologics and/or
drugs) and matrix components corresponding to the cells’ 3D
microenvironments. Cytokines and other soluble factors neces-
sary for liver development and for the maintenance of differenti-
ated hepatocytes have been known for some time [46, 103, 104].
However, the specific and efficiently directed differentiation of
stem or progenitor cells to fully mature hepatocytes and cholan-
giocytes ex vivo has remained a difficult challenge. This, in fact,
is a general problem in much of stem cell biology, whether start-
ing with lineage‐restricted adult stem cells or pluripotent ES and
iPS cells. The most promising strategies are to make use of com-
plex extracellular matrix scaffolds, effectively solid‐state sign-
aling apparatuses that can guide the differentiation of the cells.
EXTRACELLULAR MATRIX SCAFFOLDS
In recent years, complex matrix scaffolds are being utilized for
optimal differentiation of cells [5–8]. However, all of the scaf-
fold types reported are limited and inefficient in their effects due
to their methods for isolation, ones resulting in the loss of criti-
cal matrix components such as the proteoglycans. The only
known method by which to isolate a matrix scaffold with reten-
tion of these critical factors is one developed by the Reid lab in
partnership with collagen chemists [105]. They designed a
method tailored to the known solubility constants of given col-
lagens using the strategy to isolate a matrix complex with a salt
buffer at a concentration to keep insoluble all of the known
types of collagens in a given tissue. The insoluble complex so
isolated was termed “biomatrices” [106]. Frozen sections or
pulverized liver biomatrices used as cell culture substrata ena-
bled the long‐term survival of highly functional hepatocytes, far
beyond what could be achieved on plastic or with simple type I
collagen gels. Recently, we have revisited the method and estab-
lished an improved protocol, one involving perfusion strategies
and an improved delipidation method along with the high
salt  strategies, to prepare decellularized organs/tissues called
“­biomatrix scaffolds” [5]. They are tissue‐specific but mini-
mally (if at all) species‐specific, and they potently induce cell
differentiation [5]. The biomatrix scaffolds contain more than
98% of the collagens and known collagen‐bound matrix compo-
nents, including most of the fibronectins, laminins, nidogen,
entactin, elastin, and so on, and essentially all the proteoglycans
(PGs). They retain physiological levels of the known cytokines
and growth factors found in the tissue. Mature parenchymal
cells plated on biomatrix scaffolds in a serum‐free HDM
remained stable for many weeks and continued to express liver‐
specific functions equivalent to that achieved by freshly isolated
cells [107, 108].
LIVER REGENERATION
The renowned regenerative capacity of the liver has inspired
countless studies on mechanisms associated with the process
[71]. We will not summarize that considerable literature but
refer the readers to some recent reviews [3, 109, 110]. Here we
will note only the known responses of the stem cells and pro-
genitors in two distinct forms of liver regeneration: that after
534 THE LIVER:  PLASTICITY ISSUES AND REGENERATIVE PHENOMENA
partial hepatectomy, and that after selective loss of cells in aci-
nar zone three (the pericentral zone). (We assume that a parallel
process occurs in pancreatic regeneration, though it has been
studied in far less detail.)
A key to understanding the responses is recognition of “feed-
back loop signals,” factors produced by the most mature liver
cells, those in zone three of the liver acinus, and secreted into
the bile. The bile flows from pericentral zone to periportal zone
and then into the biliary tree and finally into the gut. The signal-
ing molecules include bile acids and salts that affect differentia-
tion [111]; acetylcholinesterase [112], which is produced by
mature hepatocytes and serves to inactivate acetylcholine pro-
duced by periportal cells [113, 114]; and heparins, which are
produced by mature hepatocytes [115] (J. Esko, A. Cadwallader,
and L. Reid, unpublished observations) and are relevant in
­control of stem cells and of tissue‐specific gene expression
[116, 117]. In addition the flow of the bile mechanically affects
­primary cilia on periportal cells and thereby influences signal
transduction processes mediated by these organelles [118–120].
In the presence of feedback signals, the stem cells remain in a
quiescent state. Diminution or loss of these signals results in
dis‐inhibition of the stem/progenitor cell compartments. This
leads to hyperplasia of the stem cells and other early lineage
stage cells. Factors that may release the stem cell compartment
from the normal feedback signaling control loops include
viruses, toxins, or radiation that selectively kill cells in zone
three, the pericentral zone of the acinus. The hyperplasia transi-
tions into differentiation of the cells. The resulting fully mature
cells produce bile, and the restoration or enhancement of feed-
back loop signals then inactivates the proliferative response.
Regeneration of the liver after partial hepatectomy is distinct
from that described above [71, 110, 121]. The tissue remaining
after surgical removal of a portion of the liver (e.g. two‐thirds of
its mass) continues to have feedback loop signals, and the early
lineage stage cells remain competent to respond to these signals.
The depletion below threshold levels of various liver functions
and secreted products triggers DNA synthesis as a wave across
the liver plates [110]. However, the DNA synthesis in most of the
cells of the liver (especially those in zones two and three) is not
accompanied by cytokinesis [122]. So these cells increase their
level of ploidy and demonstrate hypertrophic growth [123]. The
polyploidy triggers an increased rate of apoptosis resulting in
turnover of the liver. With the loss of the apoptotic cells, there is a
low level of proliferation of the stem cells and early lineage stage
cells to replace those cells eliminated during apoptotic processes.
In mammalian species examined, this turnover occurs in weeks.
PLASTICITY ISSUES
AND REGENERATIVE PHENOMENA
Kopp et al. [86] and others [3, 124] have hypothesized that plastic-
ity is dominant or even the sole mechanism mediating regenerative
responses for liver and pancreas. It is an hypothesis emanating
from discoveries that somatic cells can be reprogrammed by artifi-
cial means to dedifferentiate or to transdifferentiate to other cell
types by transfection of cells with multiple transcription factors
such as those identified by Takahashi and Yamanaka [125].
Although cellular reprogramming is achievable under conditions
ex vivo or in extreme artificial conditions in vivo such as hosts with
suicidal transgenes, it has yet to be demonstrated in transplanted
adult cells in common diseases. By contrast, there is evidence that
stem/progenitors give rise to maturational lineages of cells sup-
porting tissue regeneration [2, 32].
Stem cell functions of BTSCs and HpSCs are manifested
both by their genetic signatures and by their clonogenic, self‐
replicative capacity ex vivo for months when under specific
serum‐free, wholly defined conditions. In addition, these cell
populations are multipotent, the other trait required for proof of
stemness, as revealed by their ability to lineage restrict in vitro
or in vivo to hepatocytes, cholangiocytes, or islets depending on
their microenvironment before or after culture expansion [30,
41]. This is consistent with the data of Dorrell et al. [126] who
showed phenotypic and functional similarity of organoid‐initi-
ating cells in mouse pancreas and liver.
Another issue contributing to misunderstandings is that liver
regenerative phenomena in murine versus human tissues can be
distinct [2, 32]. In long‐lasting chronic human liver diseases, a
severe and progressive impairment of hepatocyte proliferative
capabilities is common. Indeed, specific insults exhaust hepato-
cyte proliferation, induce cellular senescence, and/or arrest the
hepatocyte cell cycle [32]. In human pathologic tissue, HpSCs
activate to produce the so‐called ductular reactions that give rise
to nascent hepatocytes repopulating cirrhotic livers through for-
mation of hepatocyte buds [2, 32]. By contrast, murine models
of liver injury typically do not result in a severe blockade of
hepatocyte proliferation [2]. In a novel mouse model [32] in
which apoptosis, necrosis, and senescence are induced in nearly
all hepatocytes, HpSC activation proved crucial for survival and
functional liver reconstitution.
Evidence promoting plasticity in the liver is based on studies in
a single publication by Tarlow et al. [128] These studies made use
of a highly artificial model of liver regeneration: a murine model
of Type I tyrosinemia, caused by a shortage of the enzyme fuma-
rylacetoacetate hydrolase (FAH). Moreover, Tarlow et al. [128]
subjected the FAH mice to a second insult: the administration of
DDC (2′‐3′‐dideoxycytidine). DDC causes alterations of the bil-
iary tree mimicking the spectra of primary sclerosing cholangitis;
the consequent secondary cholestasis heavily impacts the hepatic
lineages, leading to the hyperactivation of ductular reactions
shown in Tarlow et  al. [128]. The findings in such an extreme
model should not be used to make generic statements about stem/
progenitors under either normal or ­typical disease conditions.
Additional evidence used to promote the concept of plasticity
is based on experimental findings that cholangiocarcinomas can
originate from dedifferentiated hepatocytes. This provocative
assumption is based on observations by genetic lineage tracing
studies that cholangiocarcinomas can arise from cells expressing
albumin or transthyretin [129], markers erroneously ascribed
only to mature hepatocytes. Albumin and transthyretin are
expressed also in HpSCs and HBs subpopulations [30]. Similarly,
low levels of insulin are expressed in BTSC subpopulations in
the hepato‐pancreatic common duct and in multipotent progeni-
tors within the PDGs in the pancreas [32, 130]. Therefore, claims
that new beta cells derive exclusively from pre‐existing beta cells
and based on insulin expression [24] ignore the fact that insulin
lineage tracing also labels stem/progenitors.
 42:  Stem Cell‐Fueled Maturational Lineages in Hepatic and Pancreatic Organogenesis 535
The claim that mature liver cells are capable of extensive,
complete cell division is also not true except when these cells are
transplanted into livers with the FAH mutation [131] or ones
expressing suicide transgenes [132], all of them having highly
artificial microenvironments that, theoretically, could be result-
ing in reprogramming of the cells. Transplantation of mature
cells into the livers of normal animals results in ­negligible cell
division. Transplantation into hosts with classic disease or regen-
eration conditions (e.g. carbon tetrachloride or partial hepatec-
tomy) results in transient, moderate amounts of cell division.
These findings in experimental systems are in line with
those from clinical trials of mature hepatocyte transplantation
into patients with diverse liver dysfunctions or of islets into
patients with diabetes [32]. Transplantation of adult hepato-
cytes provides effects lasting only a few months. Transplanted
islets tend to be insufficient to maintain long‐term glucose
homeostasis and exhibit limited proliferation in vivo. By con-
trast, early findings of stem cell therapies for liver diseases
indicate that transplantation with BTSCs or HpSCs into the
livers of patients with diverse liver dysfunctions results in
long‐term effects over years with a steady improvement in
liver functions [31].
In summary, reports from clinical trials using stem cells
­versus mature cells for cell therapies offer the most substantive
and incontrovertible evidence for rejection of the claim that
plasticity alone mediates regenerative responses in liver and
pancreas. We prefer the assumption that mechanisms for stem/
progenitors and their maturational lineages, along with minor
contributions from epigenetic phenomena, contribute to tissue
turnover and repair.
ACKNOWLEDGMENTS
Financial support
UNC School of Medicine (Chapel Hill, NC). Funding derived
from Vesta Therapeutics (Bethesda, MD), a wholly owned
subsidiary of Toucan Capital Investments, and from the
­  3. Itoh, T. Stem/progenitor cells in liver regeneration. Review. Hepatology.
Fibrolamellar Carcinoma Foundation (Greenwich, CT).
Additional support was provided through discounted rates for
core services via federal funding of the cores: a microscopy ser-
vices laboratory in pathology and laboratory medicine core
facility grant (NIH P30DK34987), core director Victoria
Madden, PhD; a histology core funded by the Center for
Gastrointestinal and Biliary Disease Biology (CGIBD) via an
NIDDK Grant (DK34987); the Lineberger Cancer Center grant
(NCI grant #CA016086); the Carolina Center for Genome
Sciences (Katherine Hoadley, director), and the UNC Center for
Bioinformatics (Hemant Kelkar, director).
Diabetes Research Institute (Miami, FL). Studies were funded
by grants from NIH, the Juvenile Diabetes Research Foundation,
ADA, and the Diabetes Research Institute Foundation.
Sapienza University Medical Center (Rome, Italy). Professor
Gaudio was supported by research project grant from the
University “Sapienza” of Rome and FIRB grant
#RBAP10Z7FS_001 and by  PRIN grant #2009X84L84_001.
Professor Alvaro was supported by FIRB grant #RBAP10Z7FS_
004 and by PRIN grant #2009X84L84_002. The study was also
supported by Consorzio Interuniversitario Trapianti d’Organo,
Rome, Italy.
Cornell University (Ithaca, NY). Dr. Sethupathy is funded in part
by a UNC genetics and molecular biology curriculum T32 training
grant (T32‐GM‐007092–41); a grant (SRA‐60486) from Vesta
Therapeutics (awarded to L.M.R); a grant (A11–0552) from Vesta
Therapeutics (awarded to P.S); and a grant (A16–0311) from the
Fibrolamellar Cancer Foundation (awarded to P.S.).
Intellectual property
Findings from the studies summarized in this review have been
included in patent applications belonging to one or more institu-
tions including the following: University of North Carolina
(UNC) at Chapel Hill, Sapienza University in Rome, Italy, and
the Diabetes Research Institute (DRI) of the  University of
Miami, Florida. The IP has been licensed to Vesta Therapeutics
(Bethesda, MD) for clinical uses and to  PhoenixSongs
Biologicals (PSB, Branford, CT) for non‐clinical, commercial
uses. None of the authors have equity or a position in Vesta, and
none are paid consultants to the company. By contrast, LMR is
one of the founders, is the ­scientific director, and does hold an
equity position in PSB; to  date she has received no salary or
consulting fees for these efforts. Other than LMR’s connections
with PSB, the authors declare no conflicts of interest. This
review is derivative of  and updated from a book chapter in a
book on stem cells and edited by Stewart Sell [45].
REFERENCES
  1. Turner, R., Lozoya, O., Wang, Y.F. et al. Hepatic stem cells and maturational
liver lineage biology. Hepatology, 2011;53:1035–45.
  2. Miyajima, A., Tanaka, M., and Itoh, T. Stem/progenitor cells in liver development,
homeostasis, regeneration, and reprogramming. Cell Stem Cell,
2014;14(5):561–74.
2016;64(2):663–8.
 4. Carpino, G., Renzi, A., Franchitto, A. et  al. Stem/progenitor cell niches
involved in hepatic and biliary regeneration. Review. Stem Cell International,
2017:3658013.
  5. Wang, Y., Cui, C., Miguez, P. et al. Lineage restriction of hepatic stem cells
to mature fates is made efficient by tissue‐specific biomatrix scaffolds.
Hepatology, 2011;53(1):293–305.
  6. Badylak, S.F., Taylor, D., and Uygun, K. Whole‐organ tissue engineering:
decellularization and recellularization of three‐dimensional matrix scaffolds.
Review. Ann Rev Biomed Eng, 2011;13:27–53.
  7. Baptista, P.M., Siddiqui, M.M., Lozier, G. et  al. The use of whole organ
decellularization for the generation of a vascularized liver organoid.
Hepatology, 2011;53(2):604–17.
  8. Uygun, B., Soto‐Gutierrez, A., Yagi, H. et al. Organ re‐engineering through
development of a transplantatble recellularized liver graft using decellular-
ized liver matrix. Nat Med, 2010;16(7):814–20.
  9. Cardinale, V., Wang, Y., Gaudio, E. et  al. The biliary tree: a reservoir of
multipotent stem cells. Nat Rev Gastroenterol Hepatol, 2012;9:231–40.
10. Zhang, D., Jiang, W., Liu, M. et al. Highly efficient differentiation of human
ES cells and iPS cells into mature pancreatic insulin‐producing cells. Cell
Res, 2009;19(4):429–38.
11. Polo, J.M., Anderssen, E., Walsh, R.M. et al. A molecular roadmap of repro-
gramming somatic cells into iPS cells. Cell, 2012;151(7):1617–32.
536 THE LIVER:  REFERENCES
12. Wang, Y., Qin, J., Wang, S. et al. Conversion of human gastric epithelial cells
to multipotent endodermal progenitors using defined small molecules. Cell
Stem Cell, 2016;19(4):449–61.
13. Rezania, A., Bruin, J.E., Riedel, M.J. et al. Maturation of human embryonic
stem cell‐derived pancreatic progenitors into functional islets capable of
treating pre‐existing diabetes in mice. Diabetes, 2012;61:2016–29.
14. Zorn, A.M. and Wells, J.M. Molecular basis of vertebrate endoderm develop-
ment. Int Rev Cytol, 2007;259:49–111.
15. McLin V.A., Zorn, A.M. Organogenesis: making pancreas from liver. Curr
Biol, 2003;13(3):R96–8.
16. Sinner, D., Rankin, S., Lee, M., and Zorn, A.M. SOX 17 and beta‐catenin
cooperate to regulate the transcription of endodermal genes. Development,
2004;131(13):3069–80.
17. Zaret, K. Developmental competence of the gut endoderm: genetic potentia-
tion by GATA and HNF3/fork head proteins. Dev Biol, 1999;209(1):1–10.
18. Wandzioch, E. and Zaret, K.S. Dynamic signaling network for the specifica-
tion of embryonic pancreas and liver progenitors. Science, 2009;324(5935):
1707–10.
19. Tremblay, K.D. and Zaret, K.S. Distinct populations of endoderm cells con-
verge to generate the embryonic liver bud and ventral foregut tissues. Dev
Biol, 2005;280(1):87–99.
20. Barker, N., van de Wetering, M., and Clevers, H. The intestinal stem cell.
Genes Dev, 2008;22:1856–64.
21. Barker, N., Huch, M., Kujala, P. et al. Lgr5(+ve) stem cells drive self‐renewal
in the stomach and build long‐lived gastric units in vitro. Cell Stem Cell,
2010;6:25–36.
22. Lange, A.W., Keiser, A.R., Wells, J.M., Zorn, A.M., and Whitsett, J.A.
SOX17 promotes cell cycle progression and inhibits TGF‐beta/Smad3 sign-
aling to initiate progenitor cell behavior in the respiratory epithelium. PLoS
One, 2009;4(5):e5711.
23. Snyder, J.C., Teisanu, R.M., and Stripp, B.R. Endogenous lung stem cells
and contribution to disease. J Pathol, 2009;217:254–64.
24. Dor, Y., Brown, J., Martinez, O.I., and Melton, D.A. Adult pancreatic beta‐
cells are formed by self‐duplication rather than stem‐cell differentiation.
Nature, 2004;429(6987):41–6.
25. Houbracken, I. and Bouwens, L. The quest for tissue stem cells in the pan-
creas and other organs, and their application in beta‐cell replacement. Rev
Diabet Stud, 2010;7:112–23.
26. Xu, X., D’Hoker, J., Stange, G. et  al. Beta cells can be generated from
endogenous progenitors in injured adult mouse pancreas. Cell, 2008;
132(2):197–207.
27. Wang, Y., Lanzoni, G., Carpino, G. Biliary tree stem cells, precursors to pan-
creatic committed progenitors: evidence for life‐long pancreatic organogen-
esis. Stem Cells, 2013;31(9):1966–79.
28. Carpino, G., Renzi, A., Cardinale, V. et  al. Progenitor cell niches in the
human pancreatic duct system and associated pancreatic duct glands: an ana-
tomical and immunophenotyping study. J Anat, 2016;228(3):474–86.
29. Carpino, G., Cardinale, V., Gentile, R. et al. Evidence for multipotent endo-
dermal stem/progenitor cell populations in human gallbladder. J Hepatology,
2014;60(6):1194–2020.
30. Cardinale, V., Wang, Y., Carpino, G. Multipotent stem cells in the extrahe-
patic biliary tree give rise to hepatocytes, bile ducts and pancreatic islets.
Hepatology, 2011;54(6):2159–72.
31. Lanzoni, G., Cui, C., Oikawa, T. et al. Clinical programs of stem cell thera-
pies for liver and pancreas. Stem Cells, 2013;31(10):2047–60.
32. Lanzoni, G., Cardinale, V., Carpino, G. The hepatic, biliary and pancreatic
network of stem/progenitor cells niches in humans: a new reference frame
for disease and regeneration. Hepatology, 2016;64(1):277–86.
33. Alvaro, D. and Gaudio, E. Liver capsule. Biliary tree stem cell subpopula-
tions. Hepatology, 2016;64(2)644.
34. Carpino, G., Cardinale, V., Onori, P. et al. Biliary tree stem/progenitor cells
in glands of extrahepatic and intraheptic bile ducts: an anatomical in situ
study yielding evidence of maturational lineages. J Anat, 2012;220(2):
186–99.
35. Leeson, T.S. and Leeson, C.R. The fine structure of Brunner’s glands in man.
J Anat, 1968;103(2):263–76.
36. Krause, W.J. Brunner’s glands: a structural, histochemical and pathological
profile. Prog Histochem Cytochem, 2000;35(4):259–367.
37. Semeraro, R., Carpino, G., Cardinale, V. et al. Multipotent stem/progenitor
cells in the human foetal biliary tree. J Hepatol, 2012;220(2):186–99.
38. Kubota, H. and Reid, L.M. Clonogenic hepatoblasts, common precursors
for hepatocytic and biliary lineages, are lacking classical major histocom-
patibility complex class I antigens. Proc Natl Acad Sci (USA), 2000;97(22):
12132–7.
39. Schmelzer, E., Wauthier, E., and Reid, L.M. Phenotypes of pluripotent
human hepatic progenitors. Stem Cell, 2006;24(8):1852–8.
40. Sicklick, J.K., Li, Y.X., Melhem, A. et  al. Hedgehog signaling maintains
resident hepatic progenitors throughout life. Am J Physiol Gastrointest Liver
Physiol, 2006;290(5):G859–70.
41. Schmelzer, E., Zhang, L., Bruce, A. et al. Human hepatic stem cells from
fetal and postnatal donors. J Exp Med, 2007;204(8):1973–87.
42. Bonner‐Weir, S. and Sharma, A. Pancreatic stem cells. J Pathol, 2002;
197(4):519–26.
43. Lysy, P.A., Weir, G.C., and Bonner‐Weir, S. Concise review: pancreas regen-
eration: recent advances and perspectives. Stem Cells Translat Med, 2012;
1(2):150–9.
44. Jiang, W., Sui, X., Zhang, D. et al. CD24: a novel surface marker for PDX1‐
positive pancreatic progenitors derived from human embryonic stem cells.
Stem Cells, 2011;29(4):609–17.
45. Furth, M.E., Wang, Y., Cardinale, V. et al. Stem cell populations giving rise
to liver, biliary tree and pancreas, in, The Stem Cells Handbook, 2nd edn, (ed.
S. Sell), Springer Science Publishers, New York, 2013, pp. 75–126.
46. Macdonald, J.M., Xu, A., Kubota, H. et  al. Liver cell culture and lineage
biology, in Methods of Tissue Engineering, (eds., A. Atala and R.P. Lanza),
Academic Press, London, 2002, pp. 151–202.
47. Wang, Y., Yao, H., Barbier, C. et al. Paracrine signals from mesenchymal cell
populations govern the expansion and differentiation of human hepatic stem
cells to adult liver fates. Hepatology, 2010;52(4):1443–54.
48. Nusse, R. Wnt signaling and stem cell control. Cell Res, 2008;18:523–7.
49. Roskams, T. and Desmet, V. Embryology of extra‐ and intrahepatic bile
ducts, the ductal plate. Anat Rec, 2008;291(6):628–35.
50. Carpentier, R., Español‐Suñer, R., van Hul, N. et al. Embryonic ductal plate
cells give rise to cholangiocytes, periportal hepatocytes, and adult liver pro-
genitor cells. Gastroenterology, 2011;141(4):1432–8.
51. Zhang, L., Theise, N., Chua, M., and Reid, L.M. The stem cell niche of
human livers: symmetry between development and regeneration. Hepatology,
2008;48(5):1598–607.
52. Saxena, R. and Theise, N. Canals of Hering: recent insights and current
knowledge. Semin Liver Dis,2004;24(1):43–8.
53. Theise, N.D., Saxena, R., Portmann, B.C. et al. The canals of Hering and
hepatic stem cells in humans. Hepatology, 1999;30(6):1425–33.
54. Strobel, O., Rosow, D.E., Rahaklin, E.Y. J. et al. Pancreatic duct glands are
distinct ductal compartments that react to chronic injury and mediate Shh‐
induced metaplasia. Gastroenterology, 2010;138:1166–77.
55. Kushner, J.A., Weir, G.C., and Bonner‐Weir, S. Ductal origin hypothesis of
pancreatic regeneration under attack. Cell Metab, 2010;11(1):2–3.
56. Bonner‐Weir, S., Tosch, E., Inada, A. et al. The pancreatic ductal epithelium
serves as a potential pool of progenitor cells. Pediatric Diabetes, 2004;
5:16–22.
57. McClelland, R., Wauthier, E., Uronis, J., and Reid, L.M. Gradient in extra-
cellular matrix chemistry from periportal to pericentral zones: regulation of
hepatic progenitors. Tissue Eng,2008;14(1):59–70.
58. Kubota, H., Yao, H., and Reid, L.M. Identification and characterization of
vitamin A‐storing cells in fetal liver. Stem Cell, 2007;25:2339–49.
59. Lesley, J., Hascall, V.C., Tammi, M., and Hyman, R. Hyaluronan binding by
cell surface CD44. J Biol Chem, 2000;275(35):26967–75.
60. Hayes, A., Tudor, D., Nowell, M., Caterson, B., and Hughes, C. Chondroitin
sulfate sulfation motifs as putative biomarkers for isolation of articular carti-
lage progenitor cells. J Histochem Cytochem, 2007;56:125–38.
61. McClelland R., Wauthier E., Uronis J, et al. Gradients in the liver’s extracel-
lular matrix chemistry from periportal to pericentral zones: influence on
human hepatic progenitors. Tissue Eng Part A, 2008;14(1):59–70.
62. Zern, M. and Reid, Z. Extracellular Matrix Chemistry and Biology,
Academic Press, New York, 1993.
63. Aruffo, A., Stamenkovic, I., Melnick, M., Underhill, C.B., and Seed, B.
CD44 is the principal cell surface receptor for hyaluronate. Cell, 1990;
61:1303–13.
64. Nakanuma, Y., Hoso, M., Sanzen, T., and Sasaki, I.M. Microstructure and
development of the normal and pathologic biliary tract in humans, including
blood supply. A review. Microsc Res Tech, 1997;15(38):552–70.
 42:  Stem Cell‐Fueled Maturational Lineages in Hepatic and Pancreatic Organogenesis 537
65. Nakanuma, Y., Katayanagi, K., Terada, T., and Saito, K. Intrahepatic peribil-
iary glands of humans. I. Anatomy, development and presumed functions. A
review. J Gastroenterol Hepatol, 1994;9(1):75–9.
66. Nakanuma, Y., Sasaki, M., Terada, T., and Harada, K. Intrahepatic peribiliary
glands of humans. II. Pathological spectrum. Review. J Gastroenterol
Hepatol, 1994;9(1):80–6.
67. Portulano, C., Paroder‐Belenitsky, M., and Carrasco, N. The Na+/I‐ sym-
porter (NIS): mechanism and medical impact. Endocrine Rev,2014;35(1):
106–49.
68. Lozoya, O.A., Wauthier, E., Turner, R. et  al. Regulation of hepatic stem/
progenitor phenotype by microenvironment stiffness in hydrogel models of
the human liver stem cell niche. Biomaterials, 2011;32(30):7389–402.
69. Wauthier, E., McClelland, R., Turner, W. et al. Hepatic stem cells and hepa-
toblasts: identification, isolation and ex vivo maintenance. Methods Cell
Biol, 2008;86:137–225.
70. Cardinale, V., Carpino, G., Gentile, R. et al. Transplantation of human fetal
biliary tree stem/progenitor cells into two patients with advanced liver cir-
rhosis. BMC Gastroenterology, 2014;14:204.
71. Michalopoulos, G.K. Liver regeneration: alternative epithelial pathways. Int
J Biochem Cell Biol, 2011;43:173–9.
72. Cohen, D.E. and Melton, D. Turning straw into gold: directing cell fate for
regenerative medicine. Nat Rev Genet, 2011;12:243–52.
73. Puppi, J., Strom, S.J., Hughes, R.D. et  al. Improving the techniques for
human hepatocyte transplantation: report from a consensus meeting in
London. Cell Transplant, 2012;21(1)1–10.
74. Washburn, M.L., Bility, M.T., Zhang, L. et al. A humanized mouse model to
study hepatitis C virus infection, immune response, and liver disease.
Gastroenterology, 2011;140(4):1334–44.
75. Fukumitsu, K., Yagi, H., and Soto‐Gutierrez, A. Bioengineering in organ
transplantation: targeting the liver. Transplant Proc,2011;43:2137–8.
76. Gerlach, J.C. Bioreactors for extracorporeal liver support. Cell Transplant,
2006;15(1):S91–103.
77. Russo, F.P. and Parola, M. Stem and progenitor cells in liver regeneration and
repair. Cytotherapy, 2011;13:135–44.
78. Parveen, N., Aleem, A.K., Habeeb, M.A., and Habibullah, C.M. An update
on hepatic stem cells: bench to bedside. Curr Pharm Bioltechnol,
2011;12(2):226–30.
79. Duncan, A., Hickey, R.D., Paulk, N.K. et  al. Ploidy reductions in murine
fusion‐derived hepatocytes. PloS Genet, 2009;5(2):e1000385.
80. Tanaka, M., Itoh, T., Tanimizu, N., and Miyajima, A. Liver stem/progenitor
cells: their characteristics and regulatory mechanisms. J Biochem, 2011;
149(3):231–9.
81. Formin, M.E., Tai, L.K., Bárcena, A., and Muench, M.O. Coexpression of
CD14 and CD326 discriminate hepatic precursors in the human fetal liver.
Stem Cells, 2011;20(7):1247–57.
82. Thorgeirsson, S., Factor, V., and Grisham, J. Early activation and expansion
of hepatic stem cells, in Handbook of Stem Cells, 2nd edn, (eds. R. Lanza, H.
Blau, D.A. Melton et al.) Elsevier, New York, 2004, pp. 497–512.
83. Navarro‐Alvarez, N., Soto‐Gutierrez, A., and Kobayashi, N. Hepatic stem
cells and liver development. Method Mol Biol, 2010;640:181–236.
84. Vessey, C.J. and de la Hall, P.M. Hepatic stem cells: a review. Pathology,
2001;33(2):130–41.
85. Reid, L.M., Fiorino, A.S., Sigal, S.H., Brill, S., and Holst, P.A. Extracellular
matrix gradients in the space of Disse: relevance to liver biology. Hepatology,
1992;15(6):1198–203.
86. Kopp, J.L., Grompe, M., and Sanders, M. Stem cells versus plasticity in liver
and pancreas regeneration. Nat Cell Biol, 2016;18(3):238–45.
87. Khan, A.A., Parveen, N., Mahaboob, V.S. et al. Management of hyperbili-
rubenemia in biliary atresia by hepatic progenitor cell transplantation
through hepatic artery: a case report. Transplant Proc, 2008;40(4):1153–5.
88. Khan, A.A., Parveen, N., Mahaboob, V.S. et al. Treatment of Crigler‐Najjar
syndrome type 1 by hepatic progenitor cell therapy: a simple procedure for
hyperbilirubinemia. Transplant Proc, 2008;40(4):1148–50.
89. Khan, A.A., Shaik, M.V., Parveen, N. et al. Human fetal liver‐derived stem
cell transplantation as supportive modality in the management of end‐stage
decompensated liver cirrhosis. Cell Transplant, 2010;19(4):409–18.
90. Stachelscheid, H., Urbaniak, T., Ring, A., Spengler, B., Gerlach, J.C., and
Zeilinger, K. Isolation and characterization of adult human liver progenitors
from ischemic liver tissue derived from therapeutic hepatectomies. Tissue
Eng Part C Methods, 2009;15(7):1633–43.
 91. McClelland, R., Wauthier, E., Zhang, L. et  al. Ex vivo conditions for
self‐replication of human hepatic stem cells. Tissue Eng Part C Methods,
2008;14(4):314–51.
  92. Schmelzer, E. and Reid, L.M. Telomerase activity in human hepatic stem
cells, hepatoblasts and hepatocytes from neonatal, pediatric, adult and geri-
atric donors. Eur J Hepatol Gastroenterol, 2009;21(10):1191–8.
  93. Oikawa, T., Kamiya, A., Zeniya, M. et  al. Sal‐like protein 4 (SALL4), a
stem cell biomarker in liver cancers. Hepatology, 2012;57(4):1469–83.
  94. Turner, R., Gerber, D., and Reid, L.M. Transplantation of cells from solid
organs requires grafting protocols. Transplantation, 2010;90:807–10.
  95. Hodgman, C.D. Handbook of Chemistry and Physics, 37th ed, Chemical
Rubber Publishing Co., 1955–1956, pp. 3156.
  96. Turner, W.S., Schmelzer, E., McClelland, R., Wauthier, E., Chen, W., and
Reid, L.M. Human hepatoblast phenotype maintained by hyaluronan
hydrogels. J Biomed Mater Res, 2007;82(1):156–68.
  97. Turner, W.S., Seagle, C., Galanko, J. et  al. Metabolomic footprinting of
human hepatic stem cells and hepatoblasts cultured in engineered hyaluro-
nan‐matrix hydrogel scaffolds. Stem Cell, 2008;26:1547–55.
  98. Turner, R., Mendel, G., Wauthier, E., Barbier, C., and Reid, L.M. Hyaluronan‐
supplemented buffers preserve adhesion mechanisms facilitating cryopreser-
vation of human hepatic stem/progenitor cells. Cell Transplant, 2012;21(10):
2257–66.
  99. Li, W. and Ding, S. Small molecules that modulate embryonic stem cell fate
and somatic cell reprogramming. Trends Pharmacol Sci,2010;31(1):36–45.
100. Turner, R., Wauthier, E., Lozoya, O. et  al. Successful transplantation of
human hepatic stem cells with restricted localization to liver using hyaluro-
nan grafts. Hepatology, 2013;57:775–84.
101. Weber, A., Mahieu‐Caputo, D., Michelle Hadchoue, M., and Franco, D.
Hepatocyte transplantation: studies in preclinical models. J Inherit Metab
Dis, 2006;29(2–3):436–41.
102. Weber, A., Groyer‐Picard, M.T., Franco, D., and Dagher, I. Hepatocyte
transplantation in animal models. Liver Transplant, 2009;15(1):7–14.
103. Kinoshita, T. and Miyajima, A. Cytokine regulation of liver development.
Biochim Biophys Acta, 2002;1592(3):303–12.
104. Kamiya, A., Kinoshita, T., Ito, Y. et al. Fetal liver development requires a
paracrine action of oncostatin M through the gp130 signal transducer.
EMBO J, 1999;18(8):2127–36.
105. Rojkind, M., Gatmaitan, Z., Mackensen, S., Giambrone, M.A., Ponce, P., and
Reid, L.M. Connective tissue biomatrix: its isolation and utilization for long‐
term cultures of normal rat hepatocytes. J Cell Biol, 1980;87(1):255–63.
106. Reid, L.M., Gaitmaitan, Z., Arias, I., Ponce, P., and Rojkind, M. Long‐term
cultures of normal rat hepatocytes on liver biomatrix. Ann N T Acad Sci,
1980;349:70–6.
107. Purushothaman, A., Hurst, D.R., Pisano, C., Mizumoto, S., Sugahara, K.,and
Sanderson, R.D. Heparanase‐mediated loss of nuclear syndecan‐1 enhances
histone acetyltransferase (HAT) activity to promote expression of genes that
drive an aggressive tumor phenotype. J Biol Chem, 2011;286(35):30377–83.
108. Capila, I. and Linhardt, R.J. Heparin ± protein interactions. Angewandte
Chemie Int Ed, 2002;41:390–412.
109. Itoh, T. and Miyajima, A. Liver regeneration by stem/progenitor cells.
Hepatology, 2014;58(4):1617–26.
110. Michalopoulos, G.K. and DeFrances, M.C. Liver regeneration. Review.
Science, 1997;276(5309):60–6.
111. Chiang, J.Y. Bile acid regulation of gene expression: roles of nuclear hor-
mone receptors. Endocrine Rev, 2002;23(4):443–63.
112. Perelman, A. and Brandan, E. Different membrane‐bound forms of acetyl-
cholinesterase are present at the cell surface of hepatocytes. Eur J Biochem,
1989;182(1):203–7.
113. Alvaro, D., Alpini, G., Jezequel, A.M. et al. Role and mechanisms of action
of acetylcholine in the regulation of rat cholangiocyte secretory functions.
J Clin Invest, 1997;100(6):1349–62.
114. LeSage, E.G., Alvaro, D., Benedetti, A. et  al. Cholinergic system modu-
lates growth, apoptosis, and secretion of cholangiocytes from bile duct‐
ligated rats. Gastroenterology, 1999;117(1):191–9.
115. Vongchan, P., Warda, M., Toyoda, H., Toida, T., Marks, R.M., and Linhardt,
R.J. Structural characterization of human liver heparan sulfate. Biochim
Biophys Acta, 2005;1721(1–3):1–8.
116. Fujita, M., Spray, D.C., Choi, H. et al. Glycosaminoglycans and proteogly-
cans induce gap junction expression and restore transcription of tissue‐­
specific mRNAs in primary liver cultures. Hepatology, 1987;7(1):1S–9S.
538 THE LIVER:  REFERENCES
117. Spray, D.C., Fujita, M., Saez, J.C. et al. Proteoglycans and glycosamino-
glycans induce gap junction synthesis and function in primary liver cul-
tures. J Cell Biol, 1987;105(1):541–51.
118. Masyuk, A.I., Masyuk, T.V., and LaRusso, N.F. Cholangiocyte primary
cilia in liver health and disease. Dev Dynam, 2008;237:2007–12.
119. Huang, B.Q., Masyuk, T.V., Muff, M.A., Tietz, P.S., Masyuk, A.I., and
Larusso, N.F. Isolation and characterization of cholangiocyte primary cilia.
Am J Physiol Gastrointest Liver Physiol, 2006;291(3):G500–9.
120. Cervantes, S., Lau, J., Cano, D.A., Borromeo‐Austin, C., and Hebrok, M.
Primary cilia regulate Gli/Hedgehog activation in pancreas. Proc Nat Acad
Sci USA, 2010;107(22):10109–14.
121. Michalopoulos, G.K. and Appasamy, R. Metabolism of HGF‐SF and its
role in liver regeneration. Review. EXS, 1993;65:275–83.
122. Liu, H., Di Cunto, F., Imarisio, S., and Reid, L.M. Citron kinase is a cell
cycle‐dependent, nuclear protein required for G2/M transition of hepato-
cytes. J Biol Chem, 2003;278(4):2541–8.
123. Sigal, S.H., Rajvanshi, P., Gorla, G.R. et al. Partial hepatectomy‐induced
polyploidy attenuates hepatocyte replication and activates cell aging
events. Am J Physiol Gastrointest Liver Physiol, 1999;276(5):G1260–72.
124. Huch, M. and Dollé, L. The plastic cellular states of liver cells: are EpCAM
and Lgr5 fit for purpose? Review. Hepatology, 2016:64(2);652–62.
125. Takahashi, K. and Yamanaka, S. A developmental framework for induced
pluripotency. Review. Development, 2015;142(19):3274–85.
126. Dorrell, C., Tarlow, B., Wang, Y. et  al. The organoid‐initiating cells in
mouse pancreas and liver are phenotypically and functionally similar. Stem
Cell Res, 2014;13(2):275–83.
127. Lu, W.Y., Bird, T.G., Boulter, L. et al. Hepatic progenitor cells of biliary
origin with liver repopulation capacity. Nat Cell Biol, 2015;17(8):971–83.
128. Tarlow, B.D., Pelz, C., Naugler, W.E. et al. Bipotential adult liver progeni-
tors are derived from chronically injured mature hepatocytes. Cell Stem
Cell, 2014;15(5):605–18.
129. Fan, B., Malato, Y., Calvisi, D.F. et al. Cholangiocarcinomas can originate
from hepatocytes in mice. J Clin Invest, 2012;122(8):2911–5.
130. Smukler, S.R., Arntfield, M.E., Razavi, R. et  al. The adult mouse and
human pancreas contain rare multipotent stem cells that express insulin.
Cell Stem Cell, 2011;8:281–93.
131. Grompe, M. and Strom, S. Mice with human livers. Gastroenterology,
2013;145(6):1209–14.
132. Rhim, J.A., Sandgren, E.P., Palmiter, R.D., and Brinster, R.L. Complete
reconstitution of mouse liver with xenogeneic hepatocytes. Proc Nat Acad
Sci USA, 1995;92(11):4942–6.
 Developmental Morphogens
43 and Adult Liver Repair
Mariana Verdelho Machado1 and Anna Mae Diehl2
1
Gastroenterology and Hepatology Department, Hospital de Santa Maria, CHLN, Lisbon, and Faculty of
Medicine, Lisbon University, Lisbon, Portugal
2
School of Medicine, Duke University, Durham, NC, USA
HEDGEHOG SIGNALING
Pathway overview
The canonical Hh pathway is a conserved, highly complex
signaling cascade with four fundamental components: (i) the
ligand Hedgehog, (ii) the receptor Patched (Patch), (iii) the
­signal transducer Smoothened (Smo), and (iv) the effector
transcription factor, Gli (Figure 43.1). Canonical Hh signaling
occurs along the primary cilium (PC) with components of the Hh
pathway concentrating in PC [1] and a complex PC trafficking
system regulating the interaction of Hh pathway components to
enhance, or block, the Hh‐initiated signal [2].
Hh is a protein produced as a 45 kDa precursor that under-
goes proteolytic processing in the endoplasmic reticulum (ER)
[3] and subsequent lipid modification to acquire cholesterol and
palmitoyl groups [4, 5]. Hh is secreted into the extracellular
space, diffusing away from the ligand‐producing cell to bind to
other cells whereby it determines their fate according to the con-
centration and duration of exposure [6]. Extracellular matrix
proteins, such as proteoglycans, modulate the diffusion of Hh
through the extracellular space and thus, regulate the concentra-
tion of Hh to which target cells are exposed [7]. Mammals have
three different Hh proteins: Sonic (Shh), Indian (Ihh), and
Desert (Dhh) hedgehog. The three ligands similarly activate the
Hh pathway in Hh‐responsive cells, however their expression is
differently regulated. While Shh and Ihh are widely expressed,
Dhh is thought to be expressed mainly in the nervous system
and testis [8].
Patch, the Hh receptor, is a protein with 12 transmembrane
domains. When Hh ligands are absent, Patch localizes to the PC
and constitutively inhibits the Hh pathway by blocking Smo, the
signal transducer protein, from being activated and entering
the  PC. When Hh ligand binds to Patch, these inhibitory
actions of Patch are relieved and Smo becomes active. Hh‐Patch
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
interactions regulate Smo activity by controlling cholesterol
modification of Smo. Smo is directly activated by binding cho-
lesterol. Patch suppresses this cholesterol modification of Smo
in the absence of Hh ligands, and this inhibition is relieved when
Hh binds to Patch [9]. The Hh–Patch complex is subsequently
internalized and degraded [10]. Three Hh co‐receptors, CAM‐
related downregulated by oncogene (Cdo), brother of Cdo
(Boc), and growth‐arrest‐specific (GAS)‐2, potentiate Hh
­signaling by enhancing Hh‐Ptch interaction [11]. Conversely,
Hhip, a soluble Hh receptor, inhibits Hh signaling by preventing
Hh‐Patch binding [12].
Smo is a transmembrane G‐protein coupled receptor that
mediates activation of Gli transcription factors in Hh‐responsive
cells. Gli proteins promote transcription of several genes
important in the regenerative/repair process, including vascular
endothelial growth factors, angiopoietin‐1 and ‐2, snail, twist‐2,
α‐smooth muscle actin, vimentin, nanog, sox2 and sox9 [8].
In the absence of Hh, Smo activity is repressed by Patch, and
Gli binds to fused kinase (Fu), suppressor of fused (Sufu) and
Costal‐2, to form a suppressor protein complex which prevents
Gli from entering the nucleus [13]. Arrested in the cytoplasm by
the suppressor protein complex, Gli is sequentially phosphoryl-
ated by protein kinase A (PKA), glycogen synthase kinase‐3
(GSK3), and calmodulin kinase‐1 (CK1). Phosphorylated Gli
then binds to β‐transducin repeat containing protein (βTrCp)
and the Gli‐βTrCp complex is targeted to the proteasome where
Gli can be either degraded entirely or processed to generate a
truncated transcription repressor (Gli‐R) [14]. When Hh binds to
Patch, Smo is de‐repressed and activated Smo dissociates Gli
from the suppressor protein complex, preventing Gli phospho-
rylation and subsequent degradation. This enables full‐length Gli
to move to the nucleus where it acts as a transcription factor.
Mammals have three known Gli proteins: Gli1, ‐2 and ‐3.
Gli1 does not undergo proteosomal degradation and hence,
remains untruncated and always promotes transcription. Gli1 is
540 THE LIVER:  HEDGEHOG SIGNALING

PC PC

Hh
Smo
Patch
Plasma Plasma
membrane membrane

CKS
PKA
GKA
Phosphorylation CKS
PKA
Ubiquitination Gli
GKA
Smo Gli
SCF-β-
TrCP
Partial Processing
Hh-Patch
Proteasome complex

Nucleus Nucleus
STOP
Gli-R PATHWAY Gli-A PATHWAY
OFF ON

Figure 43.1  The Hedgehog signaling pathway. (a) Pathway off: Patched (Patch) blocks the entry of Smoothened (Smo) into the primary cilium
(PC), which represses Smo activity and allows sequential phosphorylation of Gli by several kinases: protein kinase A (PKA), glycogen synthase‐3β
(GSK3β) and casein kinase‐1 (CK1). Phosphorylated Gli undergoes ubiquitination by Skp‐Cullin‐F‐box (SCF) protein/β‐transducing repeat con-
taining protein (TrCP), which primes Gli to limited degradation in the proteasome. Truncated Gli (Gli‐R) is a repressor of gene transcription. (b)
Pathway on: Hedgehog binds to Patch and removes it from the PC, which allows Smo entering into the PC and Smo activation. Active Smo blocks
phosphorylation and subsequent degradation of Gli. Full length Gli translocates to the nucleus and promotes transcription of several target genes.

an important target gene for Gli2 [15]. Full‐length Gli2 accumu-
lates when Smo is activated because activated Smo protects Gli2
from proteosomal degradation. When Smo is inactive, both Gli2
and Gli3 are targeted to the proteasome; Gli2 is usually fully
degraded but Gli3 is frequently partially processed to a trun-
cated form that represses transcription [16]. Hence, Gli3 can act
as either a transcriptional repressor (when Smo is inactive) or as
an activator of transcription (when activated Smo protects it
from proteosomal degradation). In contrast, Gli1 and Gli2 act
predominantly as transcription promoters. Emerging evidence
suggests that there may be a “Gli code” whereby the different
Gli factors interact to control their expression. According to this
model, Gli3 and Gli1 promote each other’s expression when
Gli2 is low but become mutually antagonistic when Gli2 accu-
mulates [17].
Besides the canonical Hh pathway, there are also two known
types of non‐canonical Hh signaling. Type 1 non‐canonical Hh
signaling depends on Patch but is Smo‐independent. In the
absence of Hh, Patch has direct pro‐apoptotic and antiprolifera-
tive effects, by activating caspase‐3 [18] and preventing nuclear
localization of cyclin D [19], respectively. Both effects of Patch
are lost when Hh binds to Patch. Type 2 noncanonical Hh sign-
aling depends on Smo but it does not require PC [20]. This non-
canonical signaling depends on the Gαi activity of Smo that
directly regulates metabolism (e.g. it promotes a Warburg‐like
effect promoting glycolysis in muscle, adipose tissue, and
myofibroblasts [21, 22]), proliferation, calcium flux, and migra-
tion (in myofibroblasts and endothelial cells [23, 24]).
Additionally, Gli signaling can occur in the absence of Hh via a
process that also appears to be Patch and Smo‐independent, as
demonstrated by evidence that Gli induction is a direct down-
stream consequence of transforming growth factor (TGF) beta
and RAS signaling [25, 26].
Liver development
The exact contribution of the Hh pathway to liver embryogene-
sis remains unknown although it is clear that the pathway is cru-
cial for differentiation of the liver bud from the ventral foregut
endoderm [27]. Later in development, activation and repression
of Hh signaling control hepatoblast fate, with activation stimu-
lating a proliferative undifferentiated phenotype and repression
permitting hepatoblasts to differentiate into mature hepatocytes
[28]. Hedgehog is also crucial for the proper development of the
muscle layer and functioning of the gallbladder and disruption
of the Hh pathway at that point in development abrogates the
protective effect of the gallbladder against the toxic effects of
bile acids on the biliary system, which may contribute to the
development of biliary atresia [29].
Liver neoplasia
Primary liver cancers
Hedgehog pathway activation has been demonstrated in various
types of primary liver cancer, including hepatocellular carci-
noma [30, 31], cholangiocarcinoma [32], and fibrolamellar
hepatocellular carcinoma [33]. In all of these cancers, the level
of pathway activity correlates with worse clinical outcomes and
reduced cancer‐free, and overall, survival [31]. Relatively
uncommon genetic mechanisms, and more prevalent epigenetic
mechanisms, contribute to hedgehog dysregulation in liver can-
cers. For example, liver cancer cell growth is increased both by
certain mutations of Smo that activate it constitutively [30], and
by hypermethylation of Hh signaling inhibitors which epige-
netically suppresses expression of factors that typically con-
strain pathway activity [34]. Dysregulated Hh signaling has
been demonstrated in diverse cell types of liver cancers,
 43:  Developmental Morphogens and Adult Liver Repair 541
including the malignant epithelial cells themselves, cancer‐
associated fibroblasts (CAFs), and tumor‐associated mac-
rophages (TAMs). Hedgehog signaling in CAFs promotes
fibrogenesis and generation of factors that maintain a supportive
niche for cancer stem‐like cells, responses that likely foster can-
cer growth [35]. Similarly, Hh signaling in TAMs promotes
immune tolerance that permits cancer cells to escape immune
surveillance mechanisms [36]. On the other hand, Hh signaling
also promotes vasculogenesis and this might enhance chemo-
therapy bioavailability [37, 38]. These opposing actions of Hh
make it difficult to predict if treatments that inhibit Hh signaling
might be beneficial in primary liver cancers. Enthusiasm has
also been dampened by inconsistent efficacy and relatively
prevalent toxicities observed when Hedgehog was inhibited in
other cancers [39]. Clinicaltrials.gov lists ongoing trials of
hedgehog inhibitors in hepatocellular carcinoma.
Hepatic adenomas
A subtype of hepatic adenomas was recently identified in which
a chromosomal deletion results in constitutive activation of Gli1
in hepatocytes. Because Gli1 is both a downstream target and a
proximal effector of the Hh signaling pathway, neoplasia is
attributed to dysregulated Hedgehog signaling. Hedgehog sign-
aling drives vasculogenesis during development and these ade-
nomas are particularly prone to clinically‐significant hemorrhage
[40]. Interestingly, the Hedgehog‐high adenoma subclass is also
marked by high expression of arginosuccinate synthase 1, a urea
cycle enzyme [41]. This finding supports emerging evidence
that Hh signaling activity in hepatocytes may regulate metabolic
zonation in healthy adult liver [42]. According to this model, Hh
activity is higher in periportal hepatocytes (which use the urea
cycle to detoxify ammonia) than in perivenous hepatocytes
(which exhibit negligible urea cycle activity) [41]. Conversely,
Wnt signaling is known to be higher in zone three hepatocytes
(which synthesize glutamine to scavenge ammonia) than in
zone one hepatocytes (which do not) [43]. Differential Hh and
Wnt activity also occurs along the intestinal crypt–villus axis. In
that tissue, Hh and Wnt appear to be mutually antagonistic, with
Hh restricting propagation of Wnt signals and thereby restrict-
ing the stem cell compartment to the crypt [44].
Liver repair
The adult liver is designed to resist injury. Hence, hepatocyte
attrition is mainly due to natural senescence and liver mass is
easily maintained by a small subpopulation of hepatocytes that
are more proliferative than the vast majority of adult hepato-
cytes, which are devoted to performing various liver‐specific
functions [45]. In contrast, liver injuries that kill hepatocytes
dramatically increase the demand for hepatocyte replacement
and this triggers multifaceted wound healing responses that aim
to restore functional hepatocyte mass [46]. Upregulation of Hh
signaling has been demonstrated in injured livers, regardless of
injury etiology [47]. Further, the level of pathway activity typi-
cally parallels the severity of liver damage and fibrosis [48].
These observations suggest that Hh critically regulates how
adult livers respond to injury. The mechanisms involved are
summarized below.
Hepatocytes produce Hedgehog ligands
One mechanism that may explain how liver injury is tightly
coupled to Hh pathway activation was suggested by the discov-
ery that adult hepatocytes produce and release large amounts of
Shh and Ihh ligands when subjected to challenges that promote
ER stress and apoptosis [49, 50]. In such circumstances,
­biologically‐active Hh ligands appear to localize in hepatocyte‐
derived exosomes and microparticles [51]. Once released into
the hepatic microenvironment, bile, and blood, these mem-
brane‐bound Hh ligands are capable of activating Hh signaling
in local and distant Hedgehog‐responsive cells and hence,
function as hepatocyte‐derived hormones [51]. In injured liv-
ers, increased exposure to Hh ligands triggers changes in pro-
liferation, viability, and differentiation of various types of local
target cells to orchestrate reconstruction of adult liver tissue via
regenerative mechanisms that resemble those involved in fetal
liver development [47]. The impact of liver‐derived Hh ligands
on Hh signaling in extrahepatic tissues is less studied but might
be significant because Hh signaling suppresses adipogenesis
in  white adipose depots [52], promotes vasculogenesis [53],
and modulates immune function [54, 55], and advanced liver
disease is characterized by cachexia, vascular remodeling, and
immune dysfunction.
Hepatic stellate cells both produce and respond
to Hedgehog ligands
The major fibrogenic cell type in liver, the hepatic stellate cell
(HSC), produces and is highly responsive to, Hedgehog ligands
[56]. Coupled with the fact that HSCs live in the space of Disse
immediately adjacent to hepatocytes and sinusoidal endothelial
cells (other liver‐resident cell types that produce and/or respond
to Hh ligands), HSC are positioned to be a key node in Hh‐
regulated regenerative responses. HSCs are also capable of
­noncanonical Hh signaling because they express various G
­protein‐coupled receptors that regulate Smo independently of
Hedgehog–Patched interaction [22]. In addition, they can acti-
vate Gli2 via morphogen‐driven mechanisms that do not require
Smo [57]. Such signaling flexibility supports the concept that
net Hh pathway activity must be tightly regulated in HSC
in  order to assure optimal liver repair. Indeed, Hh pathway
activation is necessary for quiescent HSC to become and
remain myofibroblasts and in mice, liver fibrosis is inhibited by
various approaches that inhibit Smo and/or suppress Gli1/2
activity [56, 58, 59]. Conversely, simply overexpressing Shh
ligand in hepatocytes is sufficient to induce progressive liver
fibrosis in mice [60].
Liver sinusoidal cells (LSECs) respond to Hh ligands
Liver sinusoidal cells (LSECs) respond to Hh ligands with a
repertoire of behaviors that results in blood vessel formation
and growth during liver injury. In LSECs, increased Hh
­signaling induces loss of fenestrae (capillarization) [61] and
may promote portal hypertension/portal–systemic shunting
[62]. Liver endothelial cells can also produce Hh ligands
which modulate liver repair via autocrine and paracrine
mechanisms [61].
542 THE LIVER:  NOTCH SIGNALING
The innate immune system is also highly
Hedgehog‐responsive
T cells and natural killer T (NKT) cells require Hh signaling to
remain viable [54]. Hedgehog signaling also promotes the syn-
thesis of certain chemotactic factors for NKT cells (CXCL16),
and macrophages (CCL2) [63, 64]. Further, Hedgehog‐sensitive
mechanisms modulate immune polarization of T cells, NKT
cells, and macrophages: Hh signaling typically promotes Th2/
M2 polarization that enhances immune tolerance [55]. Tolerance
is mediated, in part, via increased production of cytokines that
also induce fibrogenesis (e.g. IL‐4, IL‐13, TGFβ) by re‐enforc-
ing Hh signaling in hepatic stellate cells [54].
Ductal cells produce and respond to Hh ligands
Paracrine Hh signaling between ductal cells and neighboring
myofibroblasts induces a migratory, less mature, more prolifer-
ative phenotype in ductal cells and re‐enforces the fibrogenic
phenotype of the myofibroblasts [65]. Together with the afore-
mentioned effects on Hh responsive immune cells, these actions
promote the ductular reaction, a fibroproliferative inflammatory
response to severe acute or chronic liver injury [66].
Liver metabolism
The Hh pathway is known to regulate metabolism in adipocytes
[67], fibroblasts [22, 68], stem cells [69], immune cells [70],
and many cancer cells [35]. Pathway activity generally stimu-
lates glycolysis [22] and inhibits lipogenesis [71] but the mech-
anisms involved are multi‐faceted and not fully understood. For
example, Smo can directly activate AMP kinase, a master regu-
lator of cellular energy balance [21]. Hedgehog signaling may
also control mitochondrial mass [72] and has been shown to
interact with other pathways that modulate metabolism, such as
Notch [73] and Wnt [44]. Hence, contextual factors are likely to
influence exactly how Hh impacts metabolism in any given cell.
Until recently, Hh signaling was not thought to be directly
involved in hepatocyte metabolism. However, emerging evi-
dence suggests that Hh may be a critical regulator of hepatic
lipid metabolism [17]. This activity appears to be exquisitely
controlled by circadian forces [17] and might have systemic, as
well as local, implications for tissue health. This realization is
triggering research to delineate the mechanisms that control Hh
signaling in hepatocytes.
Healthy hepatocytes may produce Hh ligands, although this
has been difficult to demonstrate by immunostaining intact liver
tissue. Failure to visualize Hh ligands in such cells could reflect
the fact that healthy hepatocytes generally produce low levels of
Hh ligands and/or release them efficiently. The latter possibility
is supported by two lines of recent evidence. First, Hh ligands
can traffic along cytonemes, providing a mechanism to shuttle
ligands from ligand‐producing hepatocytes to immediately
adjacent ligand‐receiving cells without releasing ligands into
the extracellular space [74]. Second, VLDL and other lipopro-
tein particles harbor biologically‐active Hh ligands in healthy
adults [75]. The association of Hh ligands with cholesterol‐
containing lipoproteins is particularly intriguing because cho-
lesterol can bind to the extracellular domain of Smo and activate
it directly [76], while Hh ligands activate Smo only after
engaging their receptor, Patched [77]. Interestingly, Smo is
directly inhibited by certain other lipoprotein‐associated lipids
[78] and thus, the healthy liver may modulate Hh signaling by
varying lipoprotein particle composition. Hedgehog ligand pro-
ducing cells that traffic into and out of the liver (e.g. immune
cells, bone marrow‐derived mononuclear cells) might be another
source of Hh ligands for healthy hepatocytes. Hedgehog ligands
might also be provided by subpopulations of liver‐resident cells
that are capable of generating Hh ligands (e.g. HSCs, LSECs,
ductal cells). Signaling downstream of Hedgehog–Patched
could also be modulated by non‐canonical Hh signaling that
operates independently of Patched or Smo, affording hepato-
cytes multiple methods to titrate pathway activity. Indeed, Hh
signaling is likely to be tightly regulated in most cells because
pathway activity critically modulates DNA methylation and
chromatin remodeling by regulating one carbon metabolism and
redox state to exert epigenetic control of cell fate decisions [79].
Liver diseases with hedgehog dysregulation
Hedgehog pathway activity increases with the severity of liver
damage and fibrosis in multiple human liver diseases, including
non‐alcoholic fatty liver disease (NAFLD) [80], alcoholic liver
disease [81], viral hepatitis [82], schistosomiasis [55], primary
biliary cholangitis (PBC) [83], primary sclerosing cholangitis
(PSC) [84], and biliary atresia [85]. Further, liver damage and
fibrosis were improved by inhibiting Hh signaling in animal
models of some of these conditions (e.g. NAFLD [86, 87],
schistosomiasis [88], PSC [56], biliary atresia [89]). To date,
clinical trials of Hh inhibitors have not been done in any of these
diseases. The major impediment seems to be the risk of adverse
events caused by inhibiting Hh activity in extrahepatic tissues.
Sustained inhibition of Hh might also have negative impacts on
the liver itself because genetic defects that globally impair Smo
activation promote the metabolic syndrome [90] and hepatic
steatosis [91] in humans and inhibiting Smo blocks liver regen-
eration after partial hepatectomy in mice [92]. The aggregate
data suggest that future therapeutic approaches to target exces-
sive Hh signaling must bring pathway activity back to physio-
logical levels, rather than silencing signaling completely.
NOTCH SIGNALING
Pathway overview
The Notch signaling pathway is fundamentally different from
the Hh pathway in that it requires cell‐to‐cell contact and the
interaction of a transmembrane receptor in the Notch target cell
with a transmembrane ligand in the ligand‐producing cell [93].
Both pathways ultimately activate transcription factors. The
canonical Notch pathway has three fundamental components:
(i) the ligands (delta‐like ligands and jagged), (ii) the receptor
(Notch), (iii) the effector (transcription factor CBF‐1/Su(H)/
LAG1 [CSL] and coactivator master‐mind‐like [MAML]) [94]
(Figure 43.2).
Both ligands and receptors undergo complex post‐translational
modifications before reaching the cell membrane. Ligands are
 43:  Developmental Morphogens and Adult Liver Repair 543

Notch ligand
expressing cell

Jag
Dll
Notch

Notch receptor
expressing cell γ-secretase

NICD

NICD

NICD
Nucleus
Hes
RBP-Jk Hey

Figure 43.2  The Notch signaling pathway. Transmembrane ligands (Jagged or Dll) in one cell, bind to transmembrane receptor (Notch) in another
cell, and exposes Notch to cleavage by γ‐secretases, resulting in the release of Notch intracellular domain (NICD). NICD enters the nucleus and
binds to the transcription factor RBP‐Jk, promoting the expression of several target genes such as Hes and Hey.

expressed by the afferent cell as transmembrane proteins from
the Delta/Serrate/LAG‐2 (DSL) family. Mammals have three
Delta‐like ligands (Dll1, Dll3, and Dll4) and two Jagged ligands
(Jag1 and Jag2) [94]. Before reaching maturation, ligands
undergo a cycle of endocytosis and recycling to the cell surface
via a process that involves ubiquitylation by the E3 ubiquitin
ligases, Neuralized and Mindbomb [93]. Interactions between
ligand and receptor can occur between two cells (in trans) lead-
ing to receptor activation, or in the same cell (in cis), resulting
in signaling inhibition [95]. There are four receptor paralogs
(Notch1–4), with different signal strength and thus, ligands
may also have different effects depending on which receptor
they engage [96]. The extracellular domain of Notch is com-
posed of epidermal growth factor (EGF)‐like repeats and a
negative regulatory region (NRR). The Notch receptor is modi-
fied by O‐glycans: O‐fucose adducts generated by Pofut
enzyme are required for Notch function; O‐glucose modifica-
tion by Rumi enzymes enhances Notch cleavage; Fringe pro-
teins elongate O‐fucose by addition of GlcNac, and O‐xylose
adducts. All of these glycations modulate ligand‐receptor
interaction [93]. Other post‐translational modifications also
­ Unbound CSL acts as a transcription repressor by binding sev-
modulate cell signaling. For example, methylation by methyl-
transferase CARM1 enhances Notch activation [97]; Notch
hydroxylation by protein factor inhibiting HIF‐1 modulates the
hypoxia response [98]. Acetylation of Notch by PCAF and
p300 increases Notch stability, and deacetylation by SIRT1
reduces Notch stability [99]; ubiquitylation targets Notch to
rapid proteasome degradation; phosphorylation of Notch1
mediated by GSK3β enhances stability of Notch1, whereas it
diminishes activity of Notch2 [100].
The NRR region in Notch prevents cleavage by proteases
[101]. The interaction of ligand with Notch receptor changes the
structure of the receptor, exposing cleavage sites to the sequen-
tial action of ADAM metalloproteases and the γ‐secretase com-
plex [102]. This results in the release of Notch intracellular
domain (NICD), which translocates into the nucleus due to its
nuclear localization sequence. NICD can translocate directly to
the nucleus or traffic there indirectly via the endosome compart-
ment. The latter process is highly regulated; for example, Numb
negatively regulates Notch through modulation of NICD endo-
somal trafficking [103]. In contrast, atypical protein kinase C
amplifies Notch signaling by enhancing NICD relocalization
from late endosomes to the nucleus [104]. Besides translocating
into the nucleus, endosomal Notch can also recycle to the cell
membrane or undergo lysosomal degradation [105].
In the nucleus NICD interacts with the DNA‐binding protein
CSL (also known as DNA‐binding recombination signal‐bind-
ing protein Jκ [RBP‐Jκ]), and together recruit MAML [106].
eral corepressors. In contrast, the complex NICD/CSL/MAML
act as a transcription factor, binding to activating cofactors [93].
Classical target genes from the Notch pathway are hairy
enhancer of split (Hes) and hairy related (Hey) families, but
Notch regulates the expression of components of many impor-
tant pathways in liver disease, such as hepatocyte nuclear ­factors
(HNF), sox9, Hh, Wnt, PDGF, TGF, and VEGF [93].
544 THE LIVER:  NOTCH SIGNALING
Notch signaling can also occur via three noncanonical routes:
type 1, ligand independent; type 2, CSL‐independent, and type
3, Notch independent. Regarding the type 1 noncanonical path-
way, other ligands have been shown to activate Notch (for
example MAGP1,2 and DLK1) but the functional consequences
are not fully understood [107, 108]. On the other hand, types 2
and 3 noncanonical signaling seem likely to have functional
effects: CSL‐independent Notch‐dependent signaling upregu-
lates IL‐6 [109] and some viral proteins directly activate CSL
activation independently of Notch [110].
Liver development
Notch is crucial during liver development because it coordinates
biliary differentiation and morphogenesis of intra‐ and extrahe-
patic bile ducts, as well as the gallbladder [111]. Mutations in
Jag1 [112], or less frequently Notch2 [113], cause Alagille syn-
drome, an autosomal disease characterized by intrahepatic bile
duct paucity and cholestasis, as well as heart, ocular, and verte-
bral defects. Mice heterozygous for mutations in Jag1 or Notch2
have an Alagille‐like phenotype [114, 115]. Severe cases of
extrahepatic biliary atresia have also been associated with mis-
sense mutations in Jag1 [116].
Large extrahepatic bile ducts derive from branching of the
primitive gut‐derived diverticulum, whereas small intrahepatic
bile ducts derive from tubular structures within the ductal plate.
The ductal plate is a layer of hepatoblasts surrounding the portal
vein branches in a ring‐like array with ductal commitment, as
opposed to hepatoblasts distant to the portal vein, which differ-
entiate into hepatocytes [117]. Interestingly, some studies sug-
gest that hepatoblasts express the Notch2 receptor and
cholangiocyte differentiation is driven by interaction with Jag1‐
expressing periductal and periportal myofibroblasts [118].
Notch induces transcription factors critical for biliary specifica-
tion (including Hes1, HNF1β and Sox9) and downregulates the
expression of HNF1α and HNF4 [119]. The role of Notch in
biliary differentiation is highly regulated and dependent on an
intricate network of intercommunicating signaling pathways
such as TGFβ [120], Wnt [121], and Hippo [122].
Liver neoplasia
The effect of the Notch pathway in liver carcinogenesis is
diverse and receptor‐dependent. Mouse models of liver cancer
suggest that Notch1 and Notch2 promote hepatocellular carci-
noma [123, 124]. However, while Notch2 also seems to promote
the development of cholangiocarcinoma, Notch1 seems to
inhibit it [123]. The Notch oncogenic effect was linked with
upregulation of insulin growth factor 2 and Sox9 [124].
Interestingly, cancer associated fibroblasts may promote car-
cinogenesis by expressing Notch ligands that act on neighbor
tumor cells [125]. However, the role of Notch in hepatocellular
tumorigenesis is complex. For example, other authors demon-
strated a p53‐dependent proapoptotic TRAIL‐sensitizing and
chemotherapy‐sensitizing effect of Notch activation in tumor
cells in vitro [126].
Regarding human hepatocellular carcinoma, one third of
patients present a genetic signature of Notch activation, with
upregulation of Jag1, α‐secretase ADAM17, Notch effectors
(e.g. MAML1), and target genes (e.g. Sox9, Spp1, and Hey1) [124].
Although the Notch genetic signature does not correlate with
overall prognosis, it might help select patients who would
respond to Notch‐targeted therapies, since cell lines that share
Notch activation signature respond to Notch inhibition/inacti-
vation with decreased proliferation [124]. A subset of hepato-
cellular carcinomas also showed increased components of
Notch pathway expression at the protein level [126]. Those
tumors are enriched with CK19+ and Sox9+ cells, markers of
biliary epithelial cells and progenitor cells with bi‐phenotypic
potential and/or stem characteristics. Furthermore, tumor
expression of Notch components correlates with reduced patient
survival [127, 128] and Notch polymorphisms in tumor cells
correlate with increased susceptibility to tumor recurrence post‐
surgery [128]. Preclinical studies also suggest that Notch can be
a target for cancer therapy, since inhibition of different Notch
receptors or Jag1 in the mice after hepatocellular tumorigenesis
had occurred, inhibited proliferation, blocked epithelial‐to‐
mesenchymal transition, and resulted in reduced tumor burden
and metastization [123].
The Notch pathway also seems to play a key role in the patho-
genesis of intrahepatic cholangiocarcinoma (ICC) given that the
Notch pathway is a top overexpressed pathway [129], and Jag1
overexpression has been demonstrated in virtually all ICC [130].
ICC can derive from transformation of cholangiocytes [131] or
from the transdifferentiation of fully mature hepatocytes [132],
both processes being dependent on Notch activation [133].
Interestingly, Jag1‐expressing Kupffer cells stimulate hepatocyte
to cholangiocyte transition and hence, cholangiocarcinogenesis
[134]. In human ICC, the overexpression of different Notch
receptors has been shown to correlate with tumor burden, tumor
aggressiveness, and reduced overall survival [135]. There is also
increasing evidence linking Notch deregulation with extrahe-
patic cholangiocarcinoma and gallbladder cancer [136, 137].
Liver repair
The Notch pathway is crucial to liver regeneration/repair, regu-
lating biliary repair, progenitor cell‐mediated liver repair,
­vascular remodeling, and fibrogenesis. All Notch receptors are
expressed in the liver [138]. Notch1 and ‐2 predominate in chol-
angiocytes and hepatic progenitor cells (HPC) and increase after
biliary damage, whereas Notch3 and ‐4 expression increase
after endothelial cell damage [139]. Interestingly, quiescent
HSC express Notch1 but activated myofibroblasts downregulate
Notch1 and the Notch‐inhibitor Numb, while upregulating
Notch2 and ‐3 [73, 140]. The main Notch ligands expressed in
the liver are Jag1 on HPC, biliary cells and HSC, and Dll4 on
endothelial cells [139].
In the partial hepatectomy animal model of liver regenera-
tion, the Notch pathway is upregulated and promotes biliary and
vascular regeneration. After partial hepatectomy, Notch‐RBP/Jk
activation on HPC stimulates these cells to become cholangio-
cytes and blocks their hepatocyte differentiation [141]. This
regulation of cell‐fate resembles embryonic morphogenesis,
being similarly mediated by RBP/Jk downregulation of YAP (a
factor that is known to promote hepatocyte regeneration [142]),
suppression of HNF1α and HNF4 (factors that promote hepato-
cytic differentiation), and upregulation of factors that promote
 43:  Developmental Morphogens and Adult Liver Repair 545
biliary differentiation (e.g. HNF1β and Sox9) [143]. The Notch
pathway also promotes hepatocyte regeneration, although not
directly through activation of Notch signaling in hepatocytes
[144]. Rather, activation of Notch on endothelial cells induces
their de‐differentiation, proliferation, vascular remodeling, and
release of hepatocyte trophic factors such as Wnt2 and hepatic
growth factor (HGF) [145], and profibrogenic factors leading to
HSC activation [146]. Furthermore, Notch signaling promotes
the homing of bone‐marrow derived endothelial progenitor
cells, which are also major sources of trophic factors such as
HGF that are essential for hepatocyte proliferation and restora-
tion of liver weight [146, 147].
Regarding wound‐healing responses, HSC respond to biliary
damage with upregulation of Jag1, which promotes Notch sign-
aling on HPC and biliary specification to cholangiocytes [148].
On the other hand, after hepatocyte injury, macrophages engulf
hepatocyte debris and upregulate Wnt3, which induces Numb to
block Notch signaling on HPC and promote specification to
hepatocytes [148]. Evidence also suggests that during biliary
injury, hepatocytes can transdifferentiate into primitive ductular
cells and regenerate the biliary epithelium via a Notch‐depend-
ent process [149].
The Notch pathway promotes fibrogenesis in liver repair, act-
ing synergically with the Hh pathway [73]. Notch promotes Shh
signaling through the regulation of transport in and out the PC
[150], and Shh promotes Notch signaling through direct upregu-
lation of Hes1 and Jag2 [151]. Both in vitro and in vivo studies
showed that the Notch pathway drives an epithelial to mesen-
chymal‐like transition that transdifferentiates HSC into myofi-
broblasts, promoting fibrogenesis [73, 152]. In patients with
chronic liver disease, activation of Notch correlates with the
severity of fibrosis, regardless of etiology of liver disease [153].
Lastly, Notch also enhances inflammatory responses and M1
polarization on macrophages [153].
Liver metabolism
There is growing evidence that Notch signaling reconfigures cel-
lular metabolic pathways [154], with consequences for the meta-
bolic syndrome, as well as cancer biology. The effect of Notch
on adipocyte metabolism/function seems to be vary according to
the receptor. Activation of Notch pathway generally blocks
expansion of the adipose tissue, rendering it less able to cope
with energy surplus. Notch also impairs adipocyte function lead-
ing to insulin resistance, decreased fatty acid oxidation, spill out
of fatty acids from adipocytes, and ectopic fat accumulation in
hepatocytes [155, 156]. Paradoxically, Notch‐1 activation pro-
motes adipogenesis through upregulation of peroxisome prolif-
erator‐activated receptor (PPAR)‐δ and ‐γ. Interestingly, Notch‐1
also associates with an adipogenic PPARγ dependent quiescent
phenotype in HSC [73]. Both obesity/energy surplus, two condi-
tions associated with increased free serum fatty acids, induce
Notch pathway activity in hepatocytes [157]. This results in
increased hepatic glucose production due to NCID enhancement
of forkhead box protein O1 (FOXO1), with consequent tran-
scriptional upregulation of key enzymes of gluconeogenesis (e.g.
glucose 6‐phosphatase and phosphoenolpyruvate carboxykin-
ase‐1). Simultaneously, Notch activation promotes mTORC1/
raptor pathway of lipogenesis, upregulating expression of
SREBP‐1 regulated lipogenic genes (e.g. acetyl‐CoA carboxy-
lase‐1 and fatty acids synthase). Additionally, Notch inhibits
fatty acid oxidation. The aggregate outcomes of increased Notch
activity in hepatocytes promote insulin resistance and hepatic
steatosis [157]. The Notch pathway can also change metabolism
to be cancer cell‐like, modulating mitochondrial function and
deregulating glutamine catabolism such that cell growth becomes
independent of exogenous glutamine [154].
Liver diseases with notch dysregulation
The role of Notch dysregulation on human disease was first
described in diseases associated with ductal plate malforma-
tions such as Alagille syndrome and biliary atresia. However,
recent evidence has linked Notch to other chronic liver diseases.
Adult primary cholangiopathies differ with regards to Notch
signaling. For example, PBC has more striking Notch upregula-
tion and less Wnt activation than PSC. The ductular reaction
also differs in these two diseases: it tends to be exuberant earlier
in the course of PBC and associates with increased expression
of ductal markers (Sox9 and K‐19) but reduced expression of
hepatocyte markers [158]. NAFLD animal models show that
Notch signaling upregulation contributes to obesity‐induced
hepatic steatosis [157]. In human steatohepatitis, Notch path-
way activation not only correlates with severity of steatosis, but
also with severity of liver disease (necroinflammatory activity
and aminotransferases levels) [159]. Further, Notch‐3 expres-
sion associates with ductular reaction and Notch‐4 with sinusoi-
dal neovessel proliferation and Kupffer cell activation [160].
SUMMARY
Hedgehog and Notch are among the morphogenic signaling
pathways that are typically inactivated once liver development
finishes. Both pathways reactivate in adulthood in response to
liver injury so that healthy liver tissue can be regenerated.
Because these pathways have pleiotropic actions that control cell
fate, they are tightly regulated. Dysfunction of these regulatory
mechanisms can lead to either insufficient, or excessive, pathway
activity – both result in progressive liver damage. For example,
hypoactivation of Notch signaling can impair biliary regenera-
tion and thereby promote progressive ductopenia and/or ductal
sclerosis, as occurs in certain primary cholangiopathies.
Conversely, hyperactivation of the Hh pathway promotes pro-
gressive liver fibrosis in many chronic liver diseases, as well as
aggressive cancer biology in different types of primary liver can-
cer. Hence, both Hedgehog and Notch are emerging as targets for
biomarker and therapeutic development in adult liver diseases.
REFERENCES
  1. Roy, S. Cilia and Hedgehog: when and how was their marriage solemnized?
Differentiation, 2012;83(2):S43–8.
  2. Liu, A., Wang, B., and Niswander, L.A. Mouse intraflagellar transport pro-
teins regulate both the activator and repressor functions of Gli transcription
factors. Development, 2005;132(13):3103–11.
546 THE LIVER:  REFERENCES
3. Chen, Y., Sasai, N., Ma, G. et al. Sonic Hedgehog dependent phospho-
rylation by CK1alpha and GRK2 is required for ciliary accumulation and
activation of smoothened. PLoS Biol, 2011;9(6):e1001083.
4. Porter, J.A., Young, K.E., and Beachy, P.A. Cholesterol modification of hedge-
hog signaling proteins in animal development. Science, 1996;274(5285):255–9.
5. Pepinsky, R.B., Zeng, C., Wen, D. et  al. Identification of a palmitic acid‐
modified form of human Sonic hedgehog. J Biol Chem, 1998;273(22):
14037–45.
6. Briscoe, J. and Therond, P.P. The mechanisms of Hedgehog signalling
and  its roles in development and disease. Nat Rev Mol Cell Biol,
2013;14(7):416–29.
7. Ayers, K.L., Gallet, A., Staccini‐Lavenant, L., and Therond, P.P. The long‐
range activity of Hedgehog is regulated in the apical extracellular space by
the glypican Dally and the hydrolase Notum. Dev Cell, 2010;18(4):605–20.
8. Merchant, J.L. and Saqui‐Salces, M. Inhibition of Hedgehog signaling in the
gastrointestinal tract: targeting the cancer microenvironment. Cancer Treat
Rev, 2014;40(1):12–21.
9. Xiao, X., Tang, J.J., Peng, C. et al. Cholesterol Modification of Smoothened
is required for Hedgehog signaling. Mol Cell,Cell, 2017;66(1):154–62.
10. Denef, N., Neubuser, D., Perez, L., and Cohen, S.M. Hedgehog induces
opposite changes in turnover and subcellular localization of patched and
smoothened. Cell, 2000;102(4):521–31.
11. Izzi, L., Levesque, M., Morin, S. et  al. Boc and Gas1 each form distinct
Shh receptor complexes with Ptch1 and are required for Shh‐mediated cell
proliferation. Dev Cell, 2011;20(6):788–801.
12. Chuang, P.T. and McMahon, A.P. Vertebrate Hedgehog signalling modulated
by induction of a Hedgehog‐binding protein. Nature, 1999;397(6720):617–21.
13. Teperino, R., Aberger, F., Esterbauer, H., Riobo, N., and Pospisilik, J.A.
Canonical and non‐canonical Hedgehog signalling and the control of metab-
olism. Semin Cell Dev Biol, 2014.;33:81–92.
14. Wang, B. and Li, Y. Evidence for the direct involvement of {beta}TrCP in
Gli3 protein processing. Proc Nat Acad Sci USA, 2006;103(1):33–8.
15. Ikram, M.S., Neill, G.W., Regl, G., Eichberger, T., Frischauf, A.M.,
Aberger, F. et al. GLI2 is expressed in normal human epidermis and BCC
and induces GLI1 expression by binding to its promoter. J Invest Dermatol,
2004;122(6):1503–9.
16. Pan, Y. and Wang, B. A novel protein‐processing domain in Gli2 and Gli3
differentially blocks complete protein degradation by the proteasome. J Biol
Chem, 2007;282(15):10846–52.
17. Matz‐Soja, M., Rennert, C., Schonefeld, K. et al. Hedgehog signaling is a
potent regulator of liver lipid metabolism and reveals a GLI‐code associated
with steatosis. Elife, 2016;5.
18. Chinchilla, P., Xiao, L., Kazanietz, M.G., and Riobo, N.A. Hedgehog pro-
teins activate pro‐angiogenic responses in endothelial cells through non‐
canonical signaling pathways. Cell Cycle, 2010;9(3):570–79.
19. Barnes, E.A., Kong, M., Ollendorff, V., and Donoghue, D.J. Patched1
interacts with cyclin B1 to regulate cell cycle progression. EMBO J,
­ adult mouse hepatocytes. Mol Biosyst, 2015;11(8):2190–7.
2001;20(9):2214–23.
20. Yuan, X., Cao, J., He, X. et  al. Ciliary IFT80 balances canonical versus
non‐canonical hedgehog signalling for osteoblast differentiation. Nat
Commun, 2016;7:11024.
21. Teperino, R., Amann, S., Bayer, M. et  al. Hedgehog partial agonism
drives  Warburg‐like metabolism in muscle and brown fat. Cell,
2012;151(2):414–26.
22. Chen, Y., Choi, S.S., Michelotti, G.A. et al. Hedgehog controls hepatic stel-
late cell fate by regulating metabolism. Gastroenterology, 2012;143(5):1319–
29 e1–11.
23. Polizio, A.H., Chinchilla, P., Chen, X., Manning, D.R., and Riobo, N.A.
Sonic Hedgehog activates the GTPases Rac1 and RhoA in a Gli‐independent
manner through coupling of smoothened to Gi proteins. Sci Signal,
2011;4(200):pt7.
24. Bijlsma, M.F., Borensztajn, K.S., Roelink, H., Peppelenbosch, M.P., and
Spek, C.A. Sonic hedgehog induces transcription‐independent cytoskeletal
rearrangement and migration regulated by arachidonate metabolites. Cell
Signal, 2007;19(12):2596–604.
25. Dennler, S., Andre, J., Verrecchia, F., and Mauviel, A. Cloning of the human
GLI2 Promoter: transcriptional activation by transforming growth factor‐beta
via SMAD3/beta‐catenin cooperation. J Biol Chem, 2009;284(46):31523–31.
26. Nolan‐Stevaux, O., Lau, J., Truitt, M.L. et  al. GLI1 is regulated through
Smoothened‐independent mechanisms in neoplastic pancreatic ducts and medi-
ates PDAC cell survival and transformation. Genes Dev, 2009;23(1):24–36.
27. Zhao, R. and Duncan, S.A. Embryonic development of the liver. Hepatology,
2005;41(5):956–67.
28. Hirose, Y., Itoh, T., and Miyajima, A. Hedgehog signal activation coordinates
proliferation and differentiation of fetal liver progenitor cells. Exp Cell Res,
2009;315(15):2648–57.
29. Higashiyama, H., Ozawa, A., Sumitomo, H. et al. Embryonic cholecystitis
and defective gallbladder contraction in the Sox17‐haploinsufficient mouse
model of biliary atresia. Development, 2017;144(10):1906–17.
30. Sicklick, J.K., Li, Y.X., Jayaraman, A. et al. Dysregulation of the Hedgehog
pathway in human hepatocarcinogenesis. Carcinogenesis, 2006;27(4):748–57.
31. Della Corte, C.M., Viscardi, G., Papaccio, F. et  al. Implication of the
Hedgehog pathway in hepatocellular carcinoma. World J Gastroenterol,
2017;23(24):4330–40.
32. Jinawath, A., Akiyama, Y., Sripa, B., and Yuasa, Y. Dual blockade of the
Hedgehog and ERK1/2 pathways coordinately decreases proliferation and
survival of cholangiocarcinoma cells. J Cancer Res Clin Oncol,2007;
133(4):271–8.
33. Li, Y.C., Deng, Y.H., Guo, Z.H. et al. Prognostic value of hedgehog signal
component expressions in hepatoblastoma patients. Eur J Med Res,
2010;15(11):468–74.
34. Tada, M., Kanai, F., Tanaka, Y. et al. Down‐regulation of hedgehog‐interact-
ing protein through genetic and epigenetic alterations in human hepatocel-
lular carcinoma. Clin Cancer Res, 2008;14(12):3768–76.
35. Chan, I.S., Guy, C.D., Chen, Y. et  al. Paracrine Hedgehog signaling
drives  metabolic changes in hepatocellular carcinoma. Cancer Res,
2012;72(24):6344–50.
36. Wan, S., Zhao, E., Kryczek, I. et al. Tumor‐associated macrophages produce
interleukin 6 and signal via STAT3 to promote expansion of human hepato-
cellular carcinoma stem cells. Gastroenterology, 2014;147(6):1393–404.
37. Li, W., Miao, S., Miao, M. et al. Hedgehog signaling activation in hepatic
stellate cells promotes angiogenesis and vascular mimicry in hepatocellular
carcinoma. Cancer Invest. 2016;34(9):424–30.
38. Olive, K.P., Jacobetz, M.A., Davidson, C.J. et  al. Inhibition of Hedgehog
signaling enhances delivery of chemotherapy in a mouse model of pancreatic
cancer. Science, 2009;324(5933):1457–61.
39. Gan, G.N. and Jimeno, A. Emerging from their burrow: Hedgehog pathway
inhibitors for cancer. Expert Opin Investig Drugs, 2016;25(10):1153–66.
40. Nault, J.C., Couchy, G., Balabaud, C. et al. Molecular classification of hepa-
tocellular adenoma associates with risk factors, bleeding, and malignant
transformation. Gastroenterology, 2017;152(4):880–94.
41. Nault, J.C., Couchy, G., Caruso, S. et al. Argininosuccinate synthase 1 and
periportal gene expression in sonic hedgehog hepatocellular adenomas.
Hepatology, 2018;68(3):964–76.
42. Schmidt‐Heck, W., Matz‐Soja, M., Aleithe, S., Marbach, E., Guthke, R., and
Gebhardt, R. Fuzzy modeling reveals a dynamic self‐sustaining network of
the GLI transcription factors controlling important metabolic regulators in
43. Gebhardt, R. and Coffer, P.J. Hepatic autophagy is differentially regulated in
periportal and pericentral zones ‐ a general mechanism relevant for other
tissues? Cell Commun Signal, 2013;11(1):21.
44. van den Brink, G.R., Bleuming, S.A., Hardwick, J.C. et al. Indian Hedgehog
is an antagonist of Wnt signaling in colonic epithelial cell differentiation. Nat
Genet, 2004;36(3):277–82.
45. Lin, S., Nascimento, E.M., Gajera, C.R. et  al. Distributed hepatocytes
expressing telomerase repopulate the liver in homeostasis and injury. Nature,
2018;556(7700):244–8.
46. Machado, M.V. and Diehl, A.M. Liver renewal: detecting misrepair and opti-
mizing regeneration. Mayo Clin Proc, 2014;89(1):120–30.
47. Machado, M.V. and Diehl, A.M. Hedgehog signalling in liver pathophysiol-
ogy. J Hepatol, 2018;68(3):550–62.
48. Fleig, S.V., Choi, S.S., Yang, L. et al. Hepatic accumulation of Hedgehog‐
reactive progenitors increases with severity of fatty liver damage in mice.
Lab Invest, 2007;87(12):1227–39.
49. Rangwala, F., Guy, C.D., Lu, J. et al. Increased production of sonic hedge-
hog by ballooned hepatocytes. J Pathol, 2011;224(3):401–10.
50. Machado, M.V., Michelotti, G.A., Pereira, T.A. et al. Reduced lipoapoptosis,
hedgehog pathway activation and fibrosis in caspase‐2 deficient mice with
non‐alcoholic steatohepatitis. Gut, 2015;64(7):1148–57.
51. Witek, R.P., Yang, L., Liu, R. et al. Liver cell‐derived microparticles activate
hedgehog signaling and alter gene expression in hepatic endothelial cells.
Gastroenterology, 2009;136(1):320–30 e2.
 43:  Developmental Morphogens and Adult Liver Repair 547
52. Kopinke, D., Roberson, E.C., and Reiter, J.F. Ciliary Hedgehog signaling
restricts injury‐induced adipogenesis. Cell, 2017;170(2):340–51.
53. Byrd, N. and Grabel, L. Hedgehog signaling in murine vasculogenesis and
angiogenesis. Trends Cardiovasc Med, 2004;14(8):308–13.
54. Syn, W.K., Witek, R.P., Curbishley, S.M. et al. Role for hedgehog pathway
in regulating growth and function of invariant NKT cells. Eur J Immunol,
2009;39(7):1879–92.
55. Pereira, T.A., Xie, G., Choi, S.S. et  al. Macrophage‐derived Hedgehog
ligands promotes fibrogenic and angiogenic responses in human schistoso-
miasis mansoni. Liver Int, 2013;33(1):149–61.
56. Michelotti, G.A., Xie, G., Swiderska, M. et al. Smoothened is a master regu-
lator of adult liver repair. J Clin Invest, 2013;123(6):2380–94.
57. Fingas, C.D., Bronk, S.F., Werneburg, N.W. et  al. Myofibroblast‐derived
PDGF‐BB promotes Hedgehog survival signaling in cholangiocarcinoma
cells. Hepatology, 2011;54(6):2076–88.
58. Kumar, V., Mundra, V., and Mahato, R.I. Nanomedicines of Hedgehog inhib-
itor and PPAR‐gamma agonist for treating liver fibrosis. Pharm Res,
2014;31(5):1158–69.
59. El‐Agroudy, N.N., El‐Naga, R.N., El‐Razeq, R.A., and El‐Demerdash, E.
Forskolin, a hedgehog signalling inhibitor, attenuates carbon tetrachloride‐
induced liver fibrosis in rats. Br J Pharmacol, 2016;173(22):3248–60.
60. Chung, S.I., Moon, H., Ju, H.L. et al. Hepatic expression of Sonic Hedgehog
induces liver fibrosis and promotes hepatocarcinogenesis in a transgenic
mouse model. J Hepatol, 2016;64(3):618–27.
61. Xie, G., Choi, S.S., Syn, W.K. et  al. Hedgehog signalling regulates liver
sinusoidal endothelial cell capillarisation. Gut, 2013;62(2):299–309.
62. Uschner, F.E., Ranabhat, G., Choi, S.S. et al. Statins activate the canonical
hedgehog‐signaling and aggravate non‐cirrhotic portal hypertension, but
inhibit the non‐canonical hedgehog signaling and cirrhotic portal hyperten-
sion. Sci Rep, 2015;5:14573.
63. Omenetti, A., Syn, W.K., Jung, Y. et al. Repair‐related activation of hedge-
hog signaling promotes cholangiocyte chemokine production. Hepatology,
2009;50(2):518–27.
64. Xie, J. The hedgehog’s trick for escaping immunosurveillance: The molecu-
lar mechanisms driving myeloid‐derived suppressor cell recruitment in
hedgehog signaling‐dependent tumors. Oncoimmunology, 2014;3:e29180.
65. Omenetti, A., Porrello, A., Jung, Y. et al. Hedgehog signaling regulates epi-
thelial‐mesenchymal transition during biliary fibrosis in rodents and humans.
J Clin Invest, 2008;118(10):3331–42.
66. Omenetti, A., Popov, Y., Jung, Y. et  al. The hedgehog pathway regulates
remodelling responses to biliary obstruction in rats. Gut, 2008;57(9):
1275–82.
67. Fleury, A., Hoch, L., Martinez, M.C. et al. Hedgehog associated to micropar-
ticles inhibits adipocyte differentiation via a non‐canonical pathway. Sci
Rep, 2016;6:23479.
68. Du, K., Hyun, J., Premont, R.T. et  al. Hedgehog‐YAP signaling pathway
regulates glutaminolysis to control activation of hepatic stellate cells.
Gastroenterology, 2018;154(5):1465–79.
69. Fu, X., Zhu, M.J., Dodson, M.V., and Du, M. AMP‐activated protein kinase
stimulates Warburg‐like glycolysis and activation of satellite cells during
muscle regeneration. J Biol Chem, 2015;290(44):26445–56.
70. Song, M., Han, L., Chen, F.F. et al. Adipocyte‐derived exosomes carrying
Sonic Hedgehog mediate M1 macrophage polarization‐induced insulin
resistance via Ptch and PI3K pathways. Cell Physiol Biochem, 2018;48(4):
1416–32.
71. Bhatia, B., Hsieh, M., Kenney, A.M., and Nahle, Z. Mitogenic Sonic hedge-
hog signaling drives E2F1‐dependent lipogenesis in progenitor cells and
medulloblastoma. Oncogene, 2011;30(4):410–22.
72. Yao, P.J., Manor, U., Petralia, R.S. et al. Sonic hedgehog pathway activation
increases mitochondrial abundance and activity in hippocampal neurons.
Mol Biol Cell, 2017;28(3):387–95.
73. Xie, G., Karaca, G., Swiderska‐Syn, M. et  al. Cross‐talk between Notch
and  Hedgehog regulates hepatic stellate cell fate in mice. Hepatology,
2013;58(5):1801–13.
74. Gonzalez‐Mendez, L., Seijo‐Barandiaran, I., and Guerrero, I. Cytoneme‐
mediated cell‐cell contacts for Hedgehog reception. Elife, 2017;6.
75. Panakova, D., Sprong, H., Marois, E., Thiele, C., and Eaton, S. Lipoprotein
particles are required for Hedgehog and Wingless signalling. Nature,
2005;435(7038):58–65.
76. Huang, P., Nedelcu, D., Watanabe, M. et al. Cellular cholesterol directly acti-
vates Smoothened in Hedgehog signaling. Cell, 2016;166(5):1176–87 e14.
77. Murone, M., Rosenthal, A., and de Sauvage, F.J. Sonic hedgehog signaling
by the patched‐smoothened receptor complex. Curr Biol, 1999;9(2):76–84.
78. Jiang, X.L., Chen, T., and Zhang, X. Activation of sonic hedgehog signaling
attenuates oxidized low‐density lipoprotein‐stimulated brain microvascular
endothelial cells dysfunction in vitro. Int J Clin Exp Pathol, 2015;8(10):
12820–8.
79. Jeon, S. and Seong, R.H. Anteroposterior limb skeletal patterning requires
the bifunctional action of SWI/SNF chromatin remodeling complex in
Hedgehog pathway. PLoS Genet, 2016;12(3):e1005915.
80. Guy, C.D., Suzuki, A., Zdanowicz, M. et al. Hedgehog pathway activation
parallels histologic severity of injury and fibrosis in human nonalcoholic
fatty liver disease. Hepatology, 2012;55(6):1711–21.
81. Jung, Y., Brown, K.D., Witek, R.P. et  al. Accumulation of hedgehog‐
responsive progenitors parallels alcoholic liver disease severity in mice and
humans. Gastroenterology, 2008;134(5):1532–43.
82. Pereira Tde, A., Witek, R.P., Syn, W.K. et al. Viral factors induce Hedgehog
pathway activation in humans with viral hepatitis, cirrhosis, and hepatocel-
lular carcinoma. Lab Invest, 2010;90(12):1690–703.
83. Jung, Y., McCall, S.J., Li, Y.X., and Diehl, A.M. Bile ductules and stromal
cells express hedgehog ligands and/or hedgehog target genes in primary
biliary cirrhosis. Hepatology, 2007;45(5):1091–6.
84. Carpino, G., Cardinale, V., Renzi, A. et al. Activation of biliary tree stem
cells within peribiliary glands in primary sclerosing cholangitis. J Hepatol,
2015;63(5):1220–8.
85. Jung, H.Y., Jing, J., Lee, K.B., and Jang, J.J. Sonic hedgehog (SHH) and
glioblastoma‐2 (Gli‐2) expressions are associated with poor jaundice‐free
survival in biliary atresia. J Pediatr Surg, 2015;50(3):371–6.
86. Syn, W.K., Jung, Y., Omenetti, A. et al. Hedgehog‐mediated epithelial‐to‐
mesenchymal transition and fibrogenic repair in nonalcoholic fatty liver
disease. Gastroenterology, 2009;137(4):1478–88 e8.
87. Hirsova, P., Ibrahim, S.H., Bronk, S.F., Yagita, H., and Gores, G.J.
Vismodegib suppresses TRAIL‐mediated liver injury in a mouse model of
nonalcoholic steatohepatitis. PloS One, 2013;8(7):e70599.
88. de Almeida Pereira, T., Borthwick, L., Xie, G. et  al. Crosstalk between
IL13 and Hedgehog pathways contributes to Schistosomiasis mansoni
fibrosis. J Hepatol, 2016;62(2):S204–S.
89. Tang, V., Cofer, Z.C., Cui, S., Sapp, V., Loomes, K.M., and Matthews, R.P.
Loss of a candidate biliary atresia susceptibility gene, add3a, causes biliary
developmental defects in zebrafish. J Pediatr Gastroenterol Nutr, 2016;63(5):
524–30.
90. Ali, O., Cerjak, D., Kent, J.W., Jr. et al. An epigenetic map of age‐associated
autosomal loci in northern European families at high risk for the metabolic
syndrome. Clin Epigenetics, 2015;7:12.
91. Sacoto, M.J.G., Martinez, A.F., Abe, Y. et al. Human germline Hedgehog
pathway mutations predispose to fatty liver. J Hepatol, 2017;67(4):809–17.
92. Swiderska‐Syn, M., Syn, W.K., Xie, G. et al. Myofibroblastic cells function
as progenitors to regenerate murine livers after partial hepatectomy. Gut,
2014;63(8):1333–44.
93. Siebel, C. and Lendahl, U. Notch Signaling in development, tissue homeo-
stasis, and disease. Physiol Rev, 2017;97(4):1235–94.
94. Meurette, O. and Mehlen, P. Notch signaling in the tumor microenviron-
ment. Cancer Cell, 2018;34(4):536–48.
95. Jacobsen, T.L., Brennan, K., Arias, A.M., and Muskavitch, M.A. Cis‐inter-
actions between Delta and Notch modulate neurogenic signalling in
Drosophila. Development, 1998;125(22):4531–40.
96. Van de Walle, I., Waegemans, E., De Medts, J. et  al. Specific Notch
receptor‐ligand interactions control human TCR‐alphabeta/gammadelta
­
development by inducing differential Notch signal strength. J Exp Med,
2013;210(4):683–97.
97. Hein, K., Mittler, G., Cizelsky, W. et al. Site‐specific methylation of Notch1
controls the amplitude and duration of the Notch1 response. Sci Signal,
2015;8(369):ra30.
98. Coleman, M.L., McDonough, M.A., Hewitson, K.S. et  al. Asparaginyl
hydroxylation of the Notch ankyrin repeat domain by factor inhibiting
hypoxia‐inducible factor. J Biol Chem, 2007;282(33):24027–38.
99. Guarani, V., Deflorian, G., Franco, C.A. et  al. Acetylation‐dependent
regulation of endothelial Notch signalling by the SIRT1 deacetylase.
­
Nature, 2011;473(7346):234–8.
100. Foltz, D.R., Santiago, M.C., Berechid, B.E., and Nye, J.S. Glycogen syn-
thase kinase‐3beta modulates notch signaling and stability. Curr Biol,
2002;12(12):1006–11.
548 THE LIVER:  REFERENCES
101. Kovall, R.A., Gebelein, B., Sprinzak, D., and Kopan, R. The canonical
Notch signaling pathway: structural and biochemical insights into shape,
sugar, and force. Dev Cell, 2017;41(3):228–41.
102. Gordon, W.R., Zimmerman, B., He, L. et al. Mechanical allostery: evidence
for a force requirement in the proteolytic activation of Notch. Dev Cell,
2015;33(6):729–36.
103. Couturier, L., Mazouni, K., and Schweisguth, F. Numb localizes at
endosomes and controls the endosomal sorting of notch after asymmetric
division in Drosophila. Curr Biol, 2013;23(7):588–93.
104. Sjoqvist, M., Antfolk, D., Ferraris, S. et al. PKCzeta regulates Notch recep-
tor routing and activity in a Notch signaling‐dependent manner. Cell Res,
2014;24(4):433–50.
105. Leitch, C.C., Lodh, S., Prieto‐Echague, V., Badano, J.L., and Zaghloul,
N.A. Basal body proteins regulate Notch signaling through endosomal traf-
ficking. J Cell Sci, 2014;127(11):2407–19.
106. Nam, Y., Weng, A.P., Aster, J.C., and Blacklow, S.C. Structural require-
ments for assembly of the CS.L.intracellular Notch1.Mastermind‐like 1
transcriptional activation complex. J Biol Chem, 2003;278(23):21232–9.
107. Miyamoto, A., Lau, R., Hein, P.W., Shipley, J.M., and Weinmaster, G.
Microfibrillar proteins MAGP‐1 and MAGP‐2 induce Notch1 extracellular
domain dissociation and receptor activation. J Biol Chem, 2006;281(15):
10089–97.
108. Traustadottir, G.A., Jensen, C.H., Thomassen, M. et al. Evidence of non‐
canonical NOTCH signaling: delta‐like 1 homolog (DLK1) directly inter-
acts with the NOTCH1 receptor in mammals. Cell Signal, 2016;28(4):
246–54.
109. Jin, S., Mutvei, A.P., Chivukula, I.V. et al. Non‐canonical Notch signaling
activates IL‐6/JAK/STAT signaling in breast tumor cells and is controlled
by p53 and IKKalpha/IKKbeta. Oncogene. 2013;32(41):4892–902.
110. Zimber‐Strobl, U. and Strobl, L.J. EBNA2 and Notch signalling in Epstein‐
Barr virus mediated immortalization of B lymphocytes. Semin Cancer Biol,
2001;11(6):423–34.
111. Zong, Y., Panikkar, A., Xu, J. et al. Notch signaling controls liver develop-
ment by regulating biliary differentiation. Development, 2009;136(10):
1727–39.
112. Oda, T., Elkahloun, A.G., Pike, B.L. et al. Mutations in the human Jagged1
gene are responsible for Alagille syndrome. Nat Genet, 1997;16(3):235–42.
113. McDaniell, R., Warthen, D.M., Sanchez‐Lara, P.A. et al. NOTCH2 muta-
tions cause Alagille syndrome, a heterogeneous disorder of the notch sign-
aling pathway. Am J Hum Genet, 2006;79(1):169–73.
114. McCright, B., Lozier, J., and Gridley, T. A mouse model of Alagille syn-
drome: Notch2 as a genetic modifier of Jag1 haploinsufficiency.
Development, 2002;129(4):1075–82.
115. Andersson, E.R., Chivukula, I.V., Hankeova, S. et  al. Mouse model of
Alagille syndrome and mechanisms of Jagged1 missense mutations.
Gastroenterology, 2018;154(4):1080–95.
116. Kohsaka, T., Yuan, Z.R., Guo, S.X. et al. The significance of human jagged
1 mutations detected in severe cases of extrahepatic biliary atresia.
Hepatology, 2002;36(4 Pt 1):904–12.
117. Lemaigre, F. and Zaret, K.S. Liver development update: new embryo mod-
els, cell lineage control, and morphogenesis. Curr Opin Genet Dev,
2004;14(5):582–90.
118. Furubo, S., Sato, Y., Harada, K., and Nakanuma, Y. Roles of myofibroblasts
and notch and hedgehog signaling pathways in the formation of intrahe-
patic bile duct lesions in polycystic kidney rats. Pediatr Dev Pathol,
2013;16(3):177–90.
119. Strazzabosco, M. and Fabris, L. Notch signaling in hepatocellular carci-
noma: guilty in association! Gastroenterology, 2012;143(6):1430–4.
120. Wang, W., Feng, Y., Aimaiti, Y., Jin, X., Mao, X., and Li, D. TGFbeta
signaling controls intrahepatic bile duct development may through
­ livers of mice. Gastroenterology, 2012;142(4):967–77 e2.
regulating the Jagged1‐Notch‐Sox9 signaling axis. J Cell Physiol,
­ 145. Ding, B.S., Nolan, D.J., Butler, J.M. et  al. Inductive angiocrine signals
2018;233(8):5780–91.
121. So, J., Khaliq, M., Evason, K. et  al. Wnt/beta‐catenin signaling controls
intrahepatic biliary network formation in zebrafish by regulating notch
activity. Hepatology, 2018;67(6):2352–66.
122. Wu, N., Nguyen, Q., Wan, Y. et al. The Hippo signaling functions through
the Notch signaling to regulate intrahepatic bile duct development in mam-
mals. Lab Invest, 2017;97(7):843–53.
123. Huntzicker, E.G., Hotzel, K., Choy, L. et al. Differential effects of targeting
Notch receptors in a mouse model of liver cancer. Hepatology,
2015;61(3):942–52.
124. Villanueva, A., Alsinet, C., Yanger, K. et al. Notch signaling is activated in
human hepatocellular carcinoma and induces tumor formation in mice.
Gastroenterology, 2012;143(6):1660–9.
125. Xiong, S., Wang, R., Chen, Q. et al. Cancer‐associated fibroblasts promote
stem cell‐like properties of hepatocellular carcinoma cells through IL‐6/
STAT3/Notch signaling. Am J Cancer Res, 2018;8(2):302–16.
126. Giovannini, C., Gramantieri, L., Chieco, P. et  al. Selective ablation of
Notch3 in HCC enhances doxorubicin’s death promoting effect by a p53
dependent mechanism. J Hepatol, 2009;50(5):969–79.
127. Wu, T., Jiao, M., Jing, L. et al. Prognostic value of Notch‐1 expression in hepa-
tocellular carcinoma: a meta‐analysis. Onco Targets Ther, 2015;8:3105–14.
128. Yu, T., Han, C., Zhu, G. et al. Prognostic value of Notch receptors in post-
surgical patients with hepatitis B virus‐related hepatocellular carcinoma.
Cancer Med, 2017;6(7):1587–600.
129. Xue, T.C., Zhang, B.H., Ye, S.L., and Ren, Z.G. Differentially expressed
gene profiles of intrahepatic cholangiocarcinoma, hepatocellular carci-
noma, and combined hepatocellular‐cholangiocarcinoma by integrated
microarray analysis. Tumour Biol, 2015;36(8):5891–9.
130. Che, L., Fan, B., Pilo, M.G. et al. Jagged 1 is a major Notch ligand along
cholangiocarcinoma development in mice and humans. Oncogenesis,
2016;5(12):e274.
131. Guest, R.V., Boulter, L., Kendall, T.J. et al. Cell lineage tracing reveals a
biliary origin of intrahepatic cholangiocarcinoma. Cancer Res, 2014;
74(4):1005–10.
132. Fan, B., Malato, Y., Calvisi, D.F. et al. Cholangiocarcinomas can originate
from hepatocytes in mice. J Clin Invest, 2012;122(8):2911–5.
133. Wang, J., Dong, M., Xu, Z. et al. Notch2 controls hepatocyte‐derived chol-
angiocarcinoma formation in mice. Oncogene, 2018;37(24):3229–42.
134. Terada, M., Horisawa, K., Miura, S. et al. Kupffer cells induce Notch‐medi-
ated hepatocyte conversion in a common mouse model of intrahepatic chol-
angiocarcinoma. Sci Rep, 2016;6:34691.
135. Wu, W.R., Shi, X.D., Zhang, R. et al. Clinicopathological significance of
aberrant Notch receptors in intrahepatic cholangiocarcinoma. Int J Clin
Exp Pathol, 2014;7(6):3272–9.
136. Aoki, S., Mizuma, M., Takahashi, Y. et  al. Aberrant activation of Notch
signaling in extrahepatic cholangiocarcinoma: clinicopathological features
and therapeutic potential for cancer stem cell‐like properties. BMC Cancer,
2016;16(1):854.
137. Yoon, H.A., Noh, M.H., Kim, B.G. et al. Clinicopathological significance
of altered Notch signaling in extrahepatic cholangiocarcinoma and gall-
bladder carcinoma. World J Gastroenterol, 2011;17(35):4023–30.
138. Nijjar, S.S., Crosby, H.A., Wallace, L., Hubscher, S.G., and Strain, A.J. Notch
receptor expression in adult human liver: a possible role in bile duct forma-
tion and hepatic neovascularization. Hepatology, 2001;34(6):1184–92.
139. Morell, C.M., Fiorotto, R., Fabris, L., and Strazzabosco, M. Notch signal-
ling beyond liver development: emerging concepts in liver repair and onco-
genesis. Clin Res Hepatol Gastroenterol, 2013;37(5):447–54.
140. Sawitza, I., Kordes, C., Reister, S., and Haussinger, D. The niche of stellate
cells within rat liver. Hepatology, 2009;50(5):1617–24.
141. Lu, J., Zhou, Y., Hu, T. et al. Notch signaling coordinates progenitor cell‐
mediated biliary regeneration following partial hepatectomy. Sci Rep,
2016;6:22754.
142. Yimlamai, D., Christodoulou, C., Galli, G.G. et al. Hippo pathway activity
influences liver cell fate. Cell, 2014;157(6):1324–38.
143. Fabris, L., Cadamuro, M., Guido, M. et al. Analysis of liver repair mecha-
nisms in Alagille syndrome and biliary atresia reveals a role for notch sign-
aling. Am J Pathol, 2007;171(2):641–53.
144. Dill, M.T., Rothweiler, S., Djonov, V. et al. Disruption of Notch1 induces
vascular remodeling, intussusceptive angiogenesis, and angiosarcomas in
from sinusoidal endothelium are required for liver regeneration. Nature,
2010;468(7321):310–5.
146. DeLeve, L.D. Liver sinusoidal endothelial cells and liver regeneration. J
Clin Invest, 2013;123(5):1861–6.
147. Wang, L., Wang, Y.C., Hu, X.B. et al. Notch‐RBP‐J signaling regulates the
mobilization and function of endothelial progenitor cells by dynamic mod-
ulation of CXCR4 expression in mice. PloS One, 2009;4(10):e7572.
148. Boulter, L., Govaere, O., Bird, T.G. et al. Macrophage‐derived Wnt opposes
Notch signaling to specify hepatic progenitor cell fate in chronic liver dis-
ease. Nat Med, 2012;18(4):572–9.
 43:  Developmental Morphogens and Adult Liver Repair 549
149. Jeliazkova, P., Jors, S., Lee, M. et  al. Canonical Notch2 signaling deter-
mines biliary cell fates of embryonic hepatoblasts and adult hepatocytes
independent of Hes1. Hepatology, 2013;57(6):2469–79.
150. Ingram, W.J., McCue, K.I., Tran, T.H., Hallahan, A.R., and Wainwright,
B.J. Sonic Hedgehog regulates Hes1 through a novel mechanism that is
independent of canonical Notch pathway signalling. Oncogene,
2008;27(10):1489–500.
151. Stasiulewicz, M., Gray, S.D., Mastromina, I. et  al. A conserved role
for  Notch signaling in priming the cellular response to Shh through
ciliary  localisation of the key Shh transducer Smo. Development,
2015;142(13):2291–303.
152. Chen, Y., Zheng, S., Qi, D. et al. Inhibition of Notch signaling by a gamma‐
secretase inhibitor attenuates hepatic fibrosis in rats. PloS One,
2012;7(10):e46512.
153. He, F., Guo, F.C., Li, Z. et al. Myeloid‐specific disruption of recombina-
tion signal binding protein Jkappa ameliorates hepatic fibrosis by attenu-
ating inflammation through cylindromatosis in mice. Hepatology,
2015;61(1):303–14.
154. Basak, N.P., Roy, A., and Banerjee, S. Alteration of mitochondrial pro-
teome due to activation of Notch1 signaling pathway. J Biol Chem,
2014;289(11):7320–34.
155. Chartoumpekis, D.V., Palliyaguru, D.L., Wakabayashi, N. et  al. Notch
intracellular domain overexpression in adipocytes confers lipodystrophy in
mice. Mol Metab, 2015;4(7):543–50.
156. Song, N.J., Yun, U.J., Yang, S. et al. Notch1 deficiency decreases hepatic
lipid accumulation by induction of fatty acid oxidation. Sci Rep,
2016;6:19377.
157. Pajvani, U.B., Qiang, L., Kangsamaksin, T., Kitajewski, J., Ginsberg, H.N.,
and Accili, D. Inhibition of Notch uncouples Akt activation from
hepatic  lipid accumulation by decreasing mTorc1 stability. Nat Med,
2013;19(8):1054–60.
158. Carpino, G., Cardinale, V., Folseraas, T. et al. Hepatic stem/progenitor cell
activation differs between primary sclerosing and primary biliary cholangi-
tis. Am J Pathol, 2018;188(3):627–39.
159. Valenti, L., Mendoza, R.M., Rametta, R. et al. Hepatic notch signaling cor-
relates with insulin resistance and nonalcoholic fatty liver disease. Diabetes,
2013;62(12):4052–62.
160. Liew, P.L., Wang, W., Lee, Y.C., Huang, M.T., and Lee, W.J. Roles
of  hepatic progenitor cells activation, ductular reaction proliferation
and  Notch signaling in morbid obesity. Hepatogastroenterology, 2012;
59(118):1921–7.
 Liver Repopulation by Cell
44 Transplantation and the
Role of Stem Cells in Liver
Biology
David A. Shafritz1 and Markus Grompe2
1
Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY, USA
2
Papé Pediatric Research Institute, Oregon Health Sciences University, Portland, OR, USA
INTRODUCTION
The major impetus for trying to reconstitute the liver by cell
transplantation is based on the very high regenerative capacity
of this organ. The very first experiments were done in the early
1900s, when liver fragments were transplanted into the anterior
chamber of the eye, but the transplanted liver tissue degenerated
rapidly and disappeared within a few days [1]. The first success-
ful report of liver cell transplantation was in the 1970s, when
isolated hepatocytes were transplanted to the liver, leading to
transient reduction in serum bilirubin in the Gunn rat model for
Crigler–Najjar syndrome, type 1 [2]. More recently, substantial
progress has been made defining the cell types that can repopu-
late the liver and restore function, and the host liver conditions
under which effective repopulation can be achieved. Both adult
hepatocytes and hepatic stem‐like progenitor cells (stem/pro-
genitor cells), as well as stem and progenitor cells of non‐hepatic
origin, have been explored for transplantation into the liver.
Studies concerning hepatocyte regeneration during both homeo-
stasis and injury, including cell plasticity and interconversion
between hepatocytes and bile duct epithelial cells and the cur-
rent status of hepatic cell transplantation in humans will be cov-
ered, as well as expectations for the future.
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
HEPATOCYTE TRANSPLANTATION:
RATIONALE AND EARLY STUDIES
Currently, orthotopic liver transplantation (OLT) is the only
effective method available to cure acquired and inherited hepatic
disorders, especially when these diseases reach their end stages
[3, 4]. However, the number of patients who are fortunate
enough to receive a liver transplant is limited by donor organ
availability. Liver transplantation is also expensive, carries sig-
nificant morbidity and mortality, and requires long‐term
immunosuppression.
Hepatocyte transplantation has been shown to be therapeutic
in animal models of acute liver failure [5] and limited human
clinical studies have been promising [6]. Inherited liver disorders
are also candidates for cell therapy because they are caused by
loss of expression or dysfunction of a single hepatocyte‐specific
gene (i.e. monogenic). Replacement of these “diseased” hepato-
cytes by normal hepatocytes should be therapeutic. Specific dis-
ease examples include Crigler–Najjar syndrome, type 1,
ornithine transcarbamylase deficiency and other urea cycle dis-
orders, familial hypercholesterolemia, factors VII and IX defi-
ciency, and phenylketonuria (diseases in which there is no
underlying liver injury or damage) and other diseases such as
 44:  Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 551
Wilson’s disease, α1‐antitrypsin deficiency, and hemochromato-
sis (diseases in which there is extensive liver injury). In inherited
monogenic liver disorders, autologous cell transplantation may
be possible, in which the patient’s cells are genetically manipu-
lated ex vivo and then transplanted back into his/her liver without
the need for immunosuppression (and indeed this has been done)
[7]. However, only a small percentage of the total hepatocellular
mass (around 1–2% maximum) can be replaced by simple hepat-
ocyte transplantation without causing portal hypertension and/or
hepatic infarction. Therefore, in most instances, it will be neces-
sary to selectively expand the therapeutic cells in vivo after their
engraftment. The therapeutic threshold will be different for each
disease, but generally cell replacement levels of at least 5–10%
will be needed for most disorders.
Attempts to increase the proportion of transplanted hepato-
cytes in the liver simply by stimulating liver regeneration (e.g.
through partial hepatectomy (PH) or carbon tetrachloride (CCl4)‐
induced hepatic necrosis) have generally failed to show signifi-
cant benefit. This is not surprising, considering that hepatocytes
need to undergo only one or two rounds of cell division to replace
the mass of liver removed by two‐thirds PH [8, 9]. Performing
repeated PH or CCl4 administration does not significantly
increase repopulation by donor hepatocytes [10], since both
transplanted and host hepatocytes respond similarly to this
regenerative stimulus. Repeated cell transplantations have also
been performed, [11, 12] but this also did not significantly
increase the efficacy of liver replacement by transplanted cells.
CLINICAL TRIALS OF HEPATOCYTE
TRANSPLANTATION
To date, allogeneic hepatocyte transplantation in humans has
had limited clinical success. In 1998, a child with Crigler–Najjar
syndrome type I, suffering from severe hyperbilirubinemia, was
infused with 7.5 × 109 allogeneic donor hepatocytes via a portal
vein catheter [13]. This resulted in a significant reduction of
serum bilirubin. However, after two and a half years serum bili-
rubin progressively increased to pretreatment levels, although
bilirubin glucuronides were still detectable in the bile (J. Roy
Chowdhury, personal communication). Studies have also been
conducted in patients with chronic liver disease and cirrhosis,
again with only marginal, if any, success [14].
Hepatocyte transplantation has also been used in conjunction
with ex vivo retroviral gene therapy in five patients with a defect
in the low‐density lipoprotein (LDL) receptor [7]. In this study,
the proportion of liver cell mass replaced was estimated to be
around 1%. The procedure resulted in a very modest decrease in
plasma cholesterol in several of the patients.
The worldwide experience with human hepatocyte transplan-
tation has been updated recently [15, 16]. Usually, suspensions
of adult hepatocytes (autographs or allographs) were utilized in
a single injection, although some patients received multiple
infusions. Many patients had inherited metabolic disorders,
including familial hypercholesterolemia, Crigler–Najjar
­syndrome type 1, factor VII deficiency, urea cycle defects, gly-
cogen storage diseases, and others [14, 16]. Initial improvement
of metabolic function was frequently observed. Some patients
subsequently received a liver transplant, so that the long‐term
effects of hepatocyte transplantation could not be assessed.
However, in the remaining patients, the function of transplanted
hepatocytes was not sustained. Whether this resulted from rejec-
tion of transplanted cells, their lack of proliferation, or eventual
apoptosis has not been determined. In no patient was expansion
of transplanted hepatocytes or progressive improvement of met-
abolic function over time documented.
Hepatocyte transplantation therapy for fulminant hepatic fail-
ure as a bridge to orthotopic liver transplantation has had some
encouraging results [5, 6]. The numbers of patients treated by
this modality are low, the diseases are of varying etiologies and
therefore it is not possible to draw definitive conclusions regard-
ing the efficacy of this intervention. Data from several pilot
studies have been reported, many with reductions in both blood
ammonia and encephalopathy and successful organ transplant
[5]. In several cases, the patients recovered without the need for
a transplant, but others died.
At this stage it can be said that allogeneic hepatocyte trans-
plantation in humans is safe and that the available data suggest
partial efficacy for both genetic and acquired hepatic disorders.
BASIC REQUIREMENTS FOR LIVER
REGENERATION
In the normal adult liver, hepatocytes are quiescent, turning over
very slowly (only two to three times per year). However, follow-
ing surgical reduction of liver mass or extensive acute toxic liver
injury, hepatocytes rapidly enter the cell cycle and proliferate to
restore liver mass. During liver regeneration, 70–90% of the
residual mature hepatocytes engage in DNA synthesis and
undergo cell division [9]. Liver regeneration is a highly organ-
ized, complex process involving growth factors and cytokines,
transcription factors, cell signaling pathways, and expression of
cell cycle regulatory genes [17]; (see Chapter 45). From many
studies, it has been concluded that the proliferative activity of
adult hepatocytes is sufficient to regenerate the liver following
two‐thirds PH without participation of stem cells [18].
LIVER SIZE CONTROL AND “HIPPO
SIGNALING PATHWAY”
For many years, it has been known that the liver size (mass) is
proportional to total body weight, ranging from 3 to 5% in differ-
ent mammalian species (see Chapter  45). After two‐thirds PH,
cellular proliferation begins within 12–18 hours and the liver size
returns to normal within 1–2 weeks in rodents and after 4–12
months in humans. When an undersized liver is transplanted, it
grows to the expected full size for the host, and when an oversized
liver is transplanted, its size is reduced to the expected mass com-
pared with total body weight (see Chapter  45). However, until
very recently, little was known concerning how this process is
controlled. In 2007, Pan and colleagues [19] showed that mam-
malian genes comparable to those in the Drosophila Hippo kinase
signaling cascade, that regulates wing mass during development,
552 THE LIVER:  ANIMAL MODELS FOR THERAPEUTIC LIVER REPOPULATION
control hepatocyte proliferation. When YAP, the mammalian
counterpart to Yorki, the last gene in the Drosophila Hippo kinase
cascade, is overexpressed in a transgenic mouse model, hepato-
cyte proliferation becomes unchecked and there is massive liver
hyperplasia, leading to hepatic carcinogenesis. When YAP hyper-
expression is turned off or blocked within four to six weeks after
birth, liver size returns to normal [19].
YAP is synthesized in the cytoplasm and translocated to the
nucleus, where it functions as a transcriptional coactivator of
TEADs, p73, RUNX2, and other genes. Control of YAP func-
tion is achieved through phosphorylation by upstream kinases
in the Hippo signaling pathway, Mst1/2 that phosphorylates
Lats1/2, and pLats1/2 that phosphorylates YAP at amino acid
S127. pYAP phosphorylated at S127 is retained in the cyto-
plasm, where it is subsequently degraded [19, 21]. YAP is a
proliferative gene with around 500 known downstream targets,
including many that effect cell growth and the cell cycle. YAP
also induces two anti‐apoptotic genes, Birc2 and Birc5, so that
it also increases survival of cells in which YAP nuclear func-
tion is enhanced (i.e. by reduced phosphorylation of YAP at
S127). Most studies reporting on the tumorigenic properties of
YAP in hepatocytes have been performed in transgenic mice or
in hepatocytes expressing a non‐phosphorylatable mutant of
YAP (YAP S127A), in which YAP expression/function is con-
stitutive and under control of a strong promoter (i.e. the tet
operon) [20].Therefore, it unclear whether the YAP is inher-
ently oncogenic or whether this reported phenotype is the
result of aberrant or uncontrolled YAP hyperexpression/
function.
ANIMAL MODELS FOR THERAPEUTIC
LIVER REPOPULATION
For many years, it was thought that mature hepatocytes could
undergo only two or three divisions, after which they become
terminally differentiated and incapable of further prolifera-
tion. More recently, however, it has been shown in several
rodent model systems that extensive modifications of the liver
microenvironment permit hepatocytes to retain high prolifera-
tive capability and to effectively repopulate the liver. Under
these conditions, the level of cell replacement can be 90% and
higher, providing proof of principle for therapeutic liver
repopulation by transplanted cells. Originally, Sandgren et al.
developed a transgenic mouse model in which a protease,
urokinase plasminogen activator (uPA), is expressed exclu-
sively in hepatocytes under control of the albumin (Alb) pro-
moter [22]. In this model, tissue protease activity caused
continuous and extensive liver injury and sub‐fulminant liver
failure, leading to death of the mice at four to six weeks of
age. However, some mice survived and, in these there were
scattered nodules of normal liver tissue distributed throughout
the hepatic parenchyma (Figure  44.1a). This occurred by
deletion or inactivation of the uPA transgene from individual
hepatocytes, which then clonally expanded into large clusters
that replaced damaged tissue. These findings prompted
the  investigators to transplant normal hepatocytes (contain-
ing a β‐galactosidase marker gene) into uPA transgenic mice,
after  which they observed extensive liver repopulation
(Figure  44.1b). They estimated that each donor hepatocyte
engrafted into the albumin uPA host liver, underwent 12–14
cell divisions [23].
Another mouse model for liver repopulation was generated
by targeted disruption of the last gene in tyrosine catabolism,
fumarylacetoacetate hydrolase (Fah) [24]. Deletion of Fah leads
to accumulation of upstream intermediates in tyrosine catabo-
lism, some of which (namely, fumarylacetoacetate) are toxic
and cause extensive and continuous liver injury. The Fah null
mouse represents an animal model for the human metabolic dis-
order hereditary tyrosinemia, type 1 (HT1), which causes exten-
sive liver injury, hepatocellular carcinoma, and death at an early
age in affected individuals. Administration of 2‐(2‐nitro‐4‐trif-
luoromethylbenzoyl) cyclohexane‐1,3‐dione (NTBC), a phar-
macological inhibitor of tyrosine catabolism upstream of
homogentisic acid, prevents accumulation of fumarylacetoace-
tate and is successful in treating patients with HT1 [25]. NTBC
also allows Fah null mice to survive and liver failure occurs in
these mice only when NTBC is discontinued.
After transplanting syngeneic wild‐type (wt) hepatocytes
into Fah null mice maintained on NTBC, only scattered small
clusters of transplanted hepatocytes are detected (Figure 44.1c
left panel). However, if NTBC treatment is discontinued
shortly after cell transplantation, liver injury resumes and
transplanted cells proliferate extensively, forming large clus-
ters within three weeks (middle panel) and replacing most of
the liver mass within six weeks (right panel). Fah null mice
with repopulated livers remain healthy, have normal liver
function tests and show a relatively normal liver structure for
many months after wt hepatocyte transplantation [24]. These
studies provide proof of principle that liver repopulation can
effectively cure a monogenic liver disease, namely, the mouse
equivalent to HT1.
Transplanted repopulating hepatocytes integrate into the
host hepatic architecture and express all functions necessary
for normal health of the animals. In the Fah null mouse, not
only do transplanted wt hepatocytes replace Fah null hepato-
cytes, but the transplanted cells can also be serially transplanted
through up to twelve consecutive Fah null mice, while retain-
ing full ability to proliferate and replace host hepatocytes [26,
27]. In these studies, it was calculated that each serially trans-
planted hepatocyte underwent an average of at least 69 cell
divisions. Thus, murine hepatocytes exhibit essentially infinite
capacity to proliferate and restore liver function under circum-
stances in which there is both massive and continuous liver
injury and the transplanted hepatocytes have a significant
selective advantage for survival compared with host hepato-
cytes. Therefore, under the specialized circumstances existing
in the liver of Fah null mice, hepatocytes exhibit many proper-
ties of stem cells, except for the ability to differentiate into
more than one lineage, in this case into hepatocytes but not into
bile duct epithelial (BDE) cells.
More recently, several additional murine liver repopulation
models have been developed (reviewed in [28]). In AFC8 [29]
and TK‐NOG [30] transgenic mice selection can be induced by
administration of a small molecule drug. Wild‐type hepatocytes
can also repopulate the livers of mice expressing a mutant form
of human alpha1‐antitrypsin [31]. Similarly, transgenic mice
 44:  Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 553

Figure 44.1  Liver regeneration and hepatocyte transplantation in the uPA transgenic mouse. (a) Wt mouse liver (left), regenerating uPA trans-
genic mouse liver by deletion of uPA transgene (middle), uPA transgenic liver (right). (b) LacZ transgenic mouse liver (left), uPA transgenic mouse
liver (middle), uPA transgenic mouse liver repopulated with LacZ hepatocytes (right). (c) Repopulation of the Fah null mouse liver by wt hepato-
cytes under NTBC administration. Immunohistochemistry with FAH staining (dark), 400× magnification: (left panel) 2 days, (middle) 3 weeks, and
(right) 6 weeks after cell transplantation.

overexpressing the p53 regulator mdm2 can be repopulated by
transplanted donor cells [32].
In all of these genetic‐based liver repopulation models, suc-
cessful liver repopulation by adult hepatocytes required two
experimental conditions: (i) the liver was under massive and
continuous liver injury and (ii) the transplanted hepatocytes
had a strong cell‐autonomous selective advantage for survival
in the host liver. In the Fah model, it has been shown that this
selective advantage is based on p21‐dependent cell cycle arrest
in host hepatocytes, resulting from p21 induction caused by
DNA damage [33]. Thus, an alternative strategy to obtain a
high level of liver repopulation by transplanted hepatocytes is
to block proliferation of endogenous hepatocytes using exoge-
nous DNA‐damaging agents and then transplant normal hepat-
ocytes in conjunction with a liver proliferative stimulus. The
first method described to achieve this effect was to treat rats
with retrorsine, a plant alkaloid that is taken up and metabo-
lized selectively by hepatocytes to produce a DNA alkylating
agent that cross‐links cellular DNA and disrupts hepatocyte
division [34]. When retrorsine or a related compound, mono-
crotaline, is administered to rats or mice, there is a long‐lived
inhibition of hepatocyte proliferation. However, the basic meta-
bolic functions of DNA damaged hepatocytes are maintained
and the animals survive. Two to four weeks after retrorsine
administration, the animals are subjected to two‐thirds PH or
CCl4 administration in conjunction with transplantation of
hepatocytes from normal animals. This leads to a brisk regen-
erative response specifically by transplanted hepatocytes and
there is near total liver repopulation in three to six months [34].
Another method to achieve effective liver repopulation by
transplanted hepatocytes is to induce DNA damage with
­selective irradiation of parts of the liver in conjunction with
hepatocyte transplantation and either two‐thirds PH, CCl4
administration, or ischemic liver injury [35, 36]. In X‐irradi-
ated rodents, administration of HGF has been used to replace
PH as a liver regenerative stimulus [35]. As with retrorsine or
monocrotaline treatment of the host liver, transplanted hepato-
cytes have a proliferative advantage over host hepatocytes in
irradiated liver lobes. Recently, this approach is beginning to be
explored in humans [37]. However, definitive data demonstrat-
ing selective liver repopulation in humans that underwent such
liver conditioning have not yet been published.
554 THE LIVER:  STEM CELLS: THEIR ORIGIN, PROPERTIES, AND TRANSPLANTATION
XENOREPOPULATION MODELS
Extensive repopulation of the liver by transplanted hepato-
cytes can also be achieved with human hepatocytes. Since
human cells are rejected by other animals, this requires the
use of immune‐deficient recipients or pharmacological
immune suppression. Small rodents capable of harboring
human hepatocytes are not only of interest as a test bed for
cell therapy, but also for a multitude of preclinical pharma-
ceutical research applications, including models of drug
metabolism, infectious diseases (including malaria, hepatitis
B and C, gene therapy, and toxicology) [28]. Three murine
models capable of supporting extensive repopulation of the
mouse liver with human hepatocytes have been reported
(reviewed in [28]). All three models have been developed
commercially and are being used by the pharmaceutical
industry as well as academia in preclinical experimentation.
In 2001, Dandri et al. first showed that immune‐deficient uPA
transgenic mice could be engrafted with human hepatocytes
and used as a model for hepatitis B [38]. Subsequently, this
model was further developed to permit extensive repopula-
tion, reaching levels as high as 90% human cells [39]. Fah
knockout mice can also be repopulated extensively with
human cells from a variety of sources when they are crossed
onto severely immune deficient backgrounds (Figure  44.2)
[40]. This model has the advantage that the hepatic injury is
not constitutive but can be titrated by NTBC. Most recently, a
third xenorepopulation platform has emerged, the TK‐NOG
mouse [30]. All of these models are of potential use to assess
the therapeutic potential of transplanted cells, be they primary
human hepatocytes or stem cell derivatives.
STEM CELLS: THEIR ORIGIN,
PROPERTIES, AND TRANSPLANTATION
Given the promising results seen in experimental cell trans-
plantation models, the source of cells that could be used clini-
cally has been of considerable interest. Historically,
hepatocytes isolated from adult donors were used for trans-
plantation. However, in the clinical setting the availability of
such cells from cadaveric donors is even more constrained
than whole organs for orthotopic transplantation, because
Figure 44.2  Repopulation of the Fah−/−/Rag2−/−/IL2rg−/− mouse liver by primary human hepatocytes at six weeks after cell transplantation, shown
at (a) low and (b) high magnification.
improved surgical techniques and perioperative management
have vastly increased the percentage of harvested livers that
can be used directly for transplantation. Therefore, considera-
ble attention has been focused on the potential use of a renew-
able and expandable cell source, such as stem cells, to
reconstitute liver mass. Several kinds of stem cells have been
explored as potential hepatocyte precursors for transplanta-
tion, including fetal liver stem cells, hepatocytes derived from
pluripotent stem cells, and adult liver stem/progenitor cells.
While the existence of a bipotential stem cell capable of gen-
erating both cholangiocytes and hepatocytes in fetal life – the
hepatoblast – is well accepted, the notion of a true adult liver
stem cell is currently controversial. Much of the existing lit-
erature on liver stem/progenitor cells can be explained by cell
plasticity [41]. The current state of this field will be presented
in the following sections. First, the literature on hepatic stem
cells will be reviewed. This will then be followed by a discus-
sion of cell plasticity as a possible alternative mechanism for
the published data, and recent studies using genetically modi-
fied hepatocytes to increase their proliferative/repopulation
potential [21].
During embryogenesis, the earliest stem cells originate from
the inner cell mass of the blastocyst and are pluripotent (capable
of differentiating into all the cell types in mammalian species);
these cells are generally referred to as embryonic stem (ES)
cells [42]. These cells give rise to somatic stem cells that subse-
quently differentiate into multipotent tissue‐specific stem cells
[42–44]. The latter give rise to lineage‐committed progenitor
cells that proliferate and differentiate into mature phenotypes
that ultimately become the somatic, tissue‐specific, working
cells in different organs (Figure 44.3). Through studies of cell
transplantation to reconstitute the hematopoietic system and
studies of cell turnover in other tissues undergoing rapid and
continuous regeneration, such as skin and intestinal epithelium,
stem cells have been shown to exhibit four essential properties:
(i) they have the capacity to maintain themselves (self‐renew),
while at the same time they generate progeny that differentiate
into mature cellular phenotypes; this is referred to as asymmet-
ric cell division [45, 46]; (ii) they are multipotent, that is, capa-
ble of producing differentiated cells in at least two lineages; (iii)
their progeny are stable, reconstitute organ mass, and remain
functional in the tissue for a long time; and (iv) by virtue of their
capability of self‐renewal, stem cells can be transplanted seri-
ally through successive hosts.
 44:  Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 555
Figure 44.3  Stem cells and tissue differentiation.
LIVER REPOPULATION BY FETAL LIVER
STEM/PROGENITOR CELLS
During days 11–15 of embryonic development a bipotent liver
epithelial cell, the hepatoblast, can be found in rats and mice (see
Chapter  2). These hepatoblasts co‐express markers of adult
hepatocytes (albumin) and cholangiocytes (CK19) as well as
AFP. The ultimate test for a putative stem cell is to demonstrate
its ability to self‐renew in vivo and to repopulate functionally a
tissue or organ, long‐term. Sandhu et  al. [47] reported 5–10%
repopulation of DPPIV− mutant F344 rat liver by transplanting
wt ED14 fetal liver epithelial cells in conjunction with two‐thirds
PH. Liver repopulation by transplanted cells increased progres-
sively over six months, and the bulk of repopulating clusters con-
tained both hepatocytes and mature bile duct cells. The
transplanted cells were integrated into the host parenchyma and
formed hybrid bile canaliculi with host hepatocytes. Thus, trans-
planted rat ED14 fetal liver epithelial cells exhibited three major
properties of liver stem cells: (i) extensive proliferation, (ii) bipo-
tency, and (iii) long‐term repopulation in vivo [47]. Liver repopu-
lation by transplanted rat fetal liver cells was achieved in a
non‐selective host liver environment but required PH to initiate
the process. This is consistent with studies in hematopoietic stem
cells, in which hematopoietic reconstitution does not occur
unless there is near total ablation of the host bone marrow.
In ED14 rat fetal liver, there are three distinct populations of
epithelial cells, those positive for AFP and Alb but negative for
CK‐19, those positive for AFP, Alb, and CK‐19, and those posi-
tive for CK‐19 but negative for AFP and Alb [48]. The number
of AFP+/Alb+/CK‐19+ cells decreased dramatically at ED16,
after which liver repopulation potential of rat fetal liver cells
also decreased significantly [47]. The level of liver repopulation
by ED14 fetal liver cells under non‐selective conditions (i.e. in
a normal liver) can also be increased to 20–25% by simply
increasing the number of ED14 fetal liver cells transplanted
(Figure 44.4) [49]. Repopulation continues to increase for up to
1 year, reaching an average of around 30% for the total liver, and
remains stable for the life of the animal [50]. This represents a
several thousand‐fold amplification of transplanted fetal liver
epithelial cells in the host organ. Both hepatic parenchymal
cords and mature bile ducts are formed by transplanted fetal
liver cells, and the progeny of the transplanted cells express nor-
mal levels of hepatocytic and cholangiocytic genes in the
respective cell types [49, 51]. Since serial transplantation has
not yet been demonstrated with fetal liver epithelial cells, they
are referred to as fetal liver stem/progenitor cells (FLSPCs),
rather than stem cells.
The mechanism of liver repopulation by rat FLSPCs has
been shown to be cell competition between the transplanted
cells and host hepatocytes [49], a process that was originally
556 THE LIVER:  THE LIVER STEM CELL “NICHE”
Figure 44.4  Repopulation of the normal adult rat liver by fetal liver stem/progenitor cells. (a) Two examples of whole rat liver sections repopu-
lated by ED14 fetal liver cells. (b) Selected areas of repopulated liver at higher magnification showing both hepatocytes and mature bile duct
structures.
described in Drosophila during wing development [52, 53].
These cells have been cryopreserved with full ability to repopu-
late the normal adult liver after thawing [54] and rat FLSPCs
have been enriched to 95% purity by selection with immuno-
magnetic beads [51].
STEM CELLS IN THE ADULT LIVER
Although many studies claim to have identified, isolated, and
purified hepatic epithelial stem cells from the adult liver, much
of the evidence is derived from in vitro culture of clonogenic
epithelial cells [55–60]. Whether these cells represent bona fide
in vivo bipotential stem cells remains uncertain. In the mouse,
all of the reported liver “stem cell” markers are also found in
cholangiocytes and therefore the existence of true stem cells
distinct from cholangiocytes in the adult mouse liver still
remains to be established. Currently, the published data regard-
ing the emergence of bipotential progenitors can all be
explained by cell plasticity, that is, fate conversion of hepato-
cytes to cholangiocytes and vice versa [41]. In the rat, however,
classic experiments are suggestive of a specialized adult liver
stem cell. Unlike in the mouse, AFP becomes expressed in cells
of ductal origin upon liver injury [61]. In conjunction with [3H]
thymidine pulse labeling and chase, AFP expression was shown
to decrease as [3H] labeled cells differentiated into basophilic
hepatocytes, which were strongly positive for albumin [62].
Thus, AFP can be considered a specific marker for stem/pro-
genitor cells in rats.
THE LIVER STEM CELL “NICHE”
If bipotential liver stem cells do exist, the question remains as to
where they reside [61]? The original idea of a stem cell “niche”
evolved from the concept that stem cells reside in tissues within
an “inductive microenvironment” that directs their differentia-
tion [63, 64]. More recently, the stem cell niche has been
described further as “a specific location in a tissue where stem
cells can reside for an indefinite period and produce progeny
cells, while self‐renewing” [65]. Local stromal cells and other
extracellular environmental factors attract stem cells to these
niches and affect their behavior (i.e. gene expression program,
proliferation, and/or differentiation), and such niches have been
identified in the bone marrow, brain, skin, and intestinal mucosa
[66–70].
At present, the most likely candidate for a liver stem cell
niche is the canal of Hering (Figure 44.5). The canals of Hering
were originally identified more than 100 years ago as luminal
channels linking the hepatocyte canalicular system to the biliary
system [71]. These channels contain small undifferentiated epi-
thelial cells [72, 73] that are in direct physical continuity with
hepatocytes at one membrane boundary and bile duct cells at
another boundary to form duct‐like structures, enclosing a
lumen (i.e. the canal; Figure  44.5b). In most models of liver
structure, the canals of Hering are depicted as very limited
structures confined to the portal space. However, studies con-
ducted in humans have demonstrated that the canals of Hering
are much more elaborate than previously thought and extend far
into the hepatic parenchyma [73, 74]. These structures contain
epithelial cells with dual expression of bile ductular and fetal
 44:  Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 557
Figure 44.5  (a) Canal of Hering as the proposed liver stem cell “niche”. “Oval cells” within the canal are the precursors of both hepatocytes and
bile ducts. (b) Electron micrograph of the canal of Hering. The canal of Hering (*) is bounded by two mature hepatocytes (Hc) and one undifferenti-
ated epithelial cell (Ec).
hepatocytic markers (AFP, HepPar1, CK‐19) and were thus con-
sidered to represent “facultative hepatic stem cells” [73, 75].
“OVAL CELLS” AS HEPATOCYTE
PROGENITORS
The term “oval cells” was coined by Farber [76] to describe non‐
parenchymal cells in the periportal region that were ­present after
treatment of rats with carcinogens: ethionine, α‐­acetylaminofluorene
(2‐AAF), and 3‐methyl‐4‐diethylaminobenzene. Other methods
to induce proliferation of “oval cells” are to treat rats with
d‐galactosamine [77,78], or allyl alcohol [79]. In each of these
models, cells are induced in the adult liver that have a small,
oval shaped, pale‐stained blue nucleus and very scant, lightly
basophilic cytoplasm. Farber did not believe that “oval cells”
were hepatocyte progenitors, but Thorgeirsson and co‐workers
[61] demonstrated that “oval cells” induced to ­proliferate in the
periportal region after treatment of rats with 2‐AAF followed by
two‐thirds PH subsequently differentiated into distinct clusters
of basophilic hepatocytes. This was demonstrated by pulse labe-
ling of the liver with [3H]thymidine and following the progres-
sion of label from periportal “oval cells” to hepatocytic clusters
in the mid‐parenchyma in conjunction with the kinetic pattern
of expression of bile ductular (CK‐7 and CK‐19) and hepato-
cytic (α‐fetoprotein [AFP] and Alb) markers over time [61, 62].
Other indirect evidence suggesting that “oval cells” are hepatic
progenitors is their expression of c‐kit [80], CD34 [81], flt3
receptor [82], and LIF [83], all known to be expressed in hemat-
opoietic stem cells or their immediate derivatives.
Studies in the rat 2‐AAF/PH model have demonstrated that
proliferating “oval cells” are indeed located in the canals of
Hering [84]. Thorgeirsson and colleagues [85] performed
extensive immunohistochemical and ultrastructural studies
in which they demonstrated that “oval cells,” induced to pro-
liferate by 2‐AAF/PH, are derived from undifferentiated cells
in the canals of Hering, after which they pass through discon-
tinuities in the laminar basement membrane of the ductal
limiting plate and join together with stellate cells as they
enter the hepatic parenchyma, proliferate, and differentiate
into hepatocytes.
Oval cell induction methods were also reported for mice,
including a choline‐deficient (CD)/ethionine‐substituted diet [86,
87], treatment with dipin [88], or 3,5‐diethoxycarbonyl‐1,
4‐dihydrocollidine (DDC) [89]. However, their utility has been
controversial, because multiple groups have shown that “oval
cells” emerging in these models generally lack bipotentiality [90].
A role for “oval cells” in normal liver physiology and cell
turnover has not been established. However, as indicated above,
“oval cells” are induced to proliferate when liver injury is super-
imposed on circumstances in which hepatocyte proliferation is
impaired. In the rat, these cells exhibit many features of pro-
genitor cells, dividing rapidly and appearing to differentiate into
both hepatocytes and BDE cells.
Attempts to establish specific markers for “oval cells” to dis-
tinguish them from mature hepatocytes and BDE cells, and to
determine their lineage origin (mesoderm or endoderm), have
led to conflicting findings. All investigators agree that “oval
cells” express common liver epithelial progenitor cell markers,
such as AFP and Alb for hepatocyte progenitors and CK‐19 (and
OV6 in the rat) for bile duct progenitor cells. They were initially
thought to express hematopoietic stem cell markers, c‐kit,
CD34, and Thy 1 [80, 81, 84, 92], but subsequent studies
reported that both fetal liver progenitor cells and “oval cells” are
negative for these markers [75, 93, 94–96].
If “oval cells” are indeed stem cells or hepatic progenitor cells,
they should be able to restore liver mass after their transplantation.
About 30 years ago, Faris and Hixson [97] reported that “oval
cells” isolated from the liver of rats fed a CD diet, treated with 2‐
AAF, and transplanted into the liver of secondary hosts produced
“colonies” or clusters of cells with an hepatocytic phenotype in
recipients that had also been subjected to the CD diet but not in
recipients that had received a normal diet. However, the level of
liver repopulation by transplanted CD/2‐AAF “oval cells” was not
determined. “Oval cells” isolated from the liver of rats treated with
d‐galactose (d‐Gal) also proliferate and differentiate into hepato-
cytes after their transplantation into rats undergoing two‐thirds PH
[98]. Duct‐like epithelial cells isolated from the atrophic pancreas
of rats undergoing treatment with a copper chelating agent (trien)
also proliferate modestly and differentiate into hepatocytes after
transplantation into normal rat liver [99]. Isolated pancreatic cells
from normal mice also repopulate the liver of Fah null mice [100].
“Oval cells” isolated from the liver of DDC‐fed mice also
558 THE LIVER: PLASTICITY, TRANSDIFFERENTIATION, AND CLONOGENIC SUBSETS OF MATURE EPITHELIAL CELLS
repopulate the liver of Fah null mice, albeit with less efficiency
than with mature hepatocytes [101]. Similarly, “oval cells” from
GFP transgenic mice maintained on a DDC diet repopulate the
liver of wt mice treated with monocrotaline in conjunction with
PH [102]. Other studies also showed effective repopulation of the
liver by purified “oval cells” in both rats treated with retrorsine
[75] and in Fah null mice [60], but not in animals with a normal
liver. Numerous studies have reported the isolation of “oval cell”
lines from mice and rats, and also from humans, that are clonal,
bipotent, and exhibit other stem and progenitor cell characteristics
in vitro and in vivo (for a review see [18]). This has provided valu-
able information concerning the basic biological properties of
these cell lines, but, in general, in vivo repopulation by trans-
planted “oval cell” lines has been very low.
HUMAN “OVAL CELLS” AND
STEM CELLS
A human counterpart to “oval cell” activation has been described
in liver tissue obtained from patients with extensive chronic
liver injury or submassive hepatic necrosis, that is, the so‐called
“ductular reaction” (for a detailed description, see [103]).
“Ductular reactions” are comprised of collections of cells in
ductular arrays with the morphological appearance and immu-
nohistochemical markers comparable to those found in rodent
“oval cells.” These cells are present primarily in the portal tracts
with extension into the parenchyma and express both hepato-
cytic and bile ductular markers, and also certain neuroendocrine
genes [103–106]. Using double and triple label immunohisto-
chemistry, Zhou et  al. [107] have shown that “ductular reac-
tions” are bipolar structures with cells at one pole exhibiting
hepatocytic morphology and gene expression (HepPar1 or
HepPar1/NCAM) and cells at the other pole exhibiting biliary
morphology and gene expression (CK‐19 or CK‐19/NCAM),
with undifferentiated epithelial cells in the center expressing
only NCAM. Cells with similar morphological and immunohis-
tochemical properties have also been identified in the human
fetal liver beginning at four weeks gestation [108].
A number of investigators have isolated, cultured, and/or pas-
saged human fetal liver epithelial cells with bipotent properties,
and several of these studies have demonstrated their differentia-
tion into hepatocytes after transplantation into SCID or nude
mice [109–111]. Schmelzer et al. [55] have identified two popu-
lations of clonogenic hepatic epithelial cells from human fetal,
neonatal, and pediatric liver that exhibit stem cell properties.
While these cells are progenitor‐like in vitro and can extensively
expand, their ability to efficiently give rise to functional hepato-
cytes either in vitro or upon transplantation remains unproven.
In the intestine, a self‐renewing classic stem cells resides at
the bottom of intestinal crypts and can be defined by expression
of the wnt‐target gene Lgr5 [112]. Clevers and co‐workers devel-
oped tissue culture conditions which permit the clonal expansion
of these stem cells and were able to obtain all mature intestinal
cell types in “organoids” derived from single Lgr5+ cells [113].
They applied similar tissue culture conditions to other endoder-
mal tissues, including liver, stomach, and pancreas and were able
to grow immortal Lgr5+ liver organoids from both mice and
humans [56, 57]. The cell of origin has a ductal phenotype and
liver organoids are extensively cholangiocytic in phenotype.
Nonetheless, they can be induced to express hepatocyte markers
in culture [56, 57]. Upon transplantation, they can produce bona
fide hepatocytes, but the frequency is very low. The future poten-
tial of clonally derived Lgr5+ cells is discussed in Chapter 77.
PLASTICITY, TRANSDIFFERENTIATION,
AND CLONOGENIC SUBSETS
OF MATURE EPITHELIAL CELLS
Much of the data on liver injury responses in the historic litera-
ture has been interpreted as either proliferation of existing
mature epithelial cells (hepatocytes or cholangiocytes) or acti-
vation of facultative stem cells. More recently, however, the
phenomenon of plasticity and transdifferentiation between
ductal epithelial cells and hepatocytes has become more appre-
ciated and experimentally proven. Several studies showed con-
clusively that proliferating ductal cells, previously considered to
be derived from facultative stem cells, were in fact not the
source of parenchymal regeneration in several classical models
of oval cell injury in the mouse [90, 114–116]. To the contrary,
it has been shown clearly that mature hepatocytes can efficiently
convert into cholangiocyte‐like cells [115] and even form a
functional biliary tree [117]. Hepatocyte‐derived ducts can
extensively proliferate and retain the ability to reconvert to their
cell of origin, the hepatocytes, once the injury subsides [115].
The ability of hepatocytes to differentiate into BDE cells has
also been well demonstrated in several rat models under injury
conditions [118, 119]. These studies have taken advantage of
recently developed lineage tracing techniques to identify the
origin of cells in liver tissue under different physiologic and
pathophysiologic states (see Chapter 85).
Taken together these findings raise the question, whether the
bipotential stem/progenitor cells and oval cells previously
reported may actually have also been transdifferentiated hepato-
cytes, not stem cells of ductal origin. Recent papers have taken
advantage of this plasticity of hepatocytes. While mature hepat-
ocytes cannot be efficiently expanded in culture, transdifferenti-
ated hepatocytes can be grown extensively [120]. Excitingly,
these cells retain the ability to redifferentiate into mature hepat-
ocytes and function in transplantation.
While non‐hepatocytes are clearly not the source of new
hepatocytes in most of the classic mouse oval cell literature,
newer injury models which involve activation of p21 and senes-
cence in mature hepatocytes have shown that cholangiocytes
can become hepatocytes [32, 121, 122]. Thus, transdifferentia-
tion has now been shown to work in both directions. If cholan-
giocytes can produce hepatocytes and vice versa, the question
arises whether typical stem/progenitor cells exist in the adult
liver or whether all regeneration can be explained by plasticity
of mature epithelial cells.
An important question arising from the concept of plastic-
ity is whether all mature epithelial cells are equally prolifera-
tive and plastic. Recently, several labs have demonstrated
heterogeneity within hepatocytes in the adult mouse [123,
124], once again giving rise to claims of having discovered
 44:  Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 559
“the” liver stem cell. In addition, significant differences in the
proliferative ability of cholangiocyte subsets have also been
found [125].
LIMITED LIVER REPOPULATION BY
EXTRAHEPATIC AND EMBRYONIC
STEM CELLS
Various studies have reported that cells released from the bone
marrow (BM) into the circulation, migrate to the liver and dif-
ferentiate into hepatocytes. However, the extent to which this
occurs and the mechanism(s) involved remain highly controver-
sial (for reviews, see [126–128]). Estimates of liver repopula-
tion by hematopoietic cells vary widely, ranging from less than
0. 01 to 40%. Originally, Petersen et al. reported that BM stem
cells from DPPIV+ F344 rats transplanted into sublethally irra-
diated DPPIV− F344 rats repopulate the BM after which they
migrate to the liver and “transdifferentiate” into hepatocytes
through the liver “oval cell” progenitor pathway [129]. This
mechanism was generally accepted until studies by Wang et al.
[101] using lacZ marking showed that BM cells did not enter the
“oval cell” pool in wt mice treated with DDC or contribute to
liver repopulation by “oval cells” in secondary Fah−/− mouse
recipients. Menthena et  al. [130] also showed in rats that
DPPIV+ BM cells transplanted into DPPIV− rats contributed less
than 1% to “oval cells” derived from three different model sys-
tems: (i) 2‐AAF/PH, (ii) retrorsine/PH, and (iii) d‐Gal induced
liver injury.
In Fah−/− mice as well as other model systems, it has been
demonstrated that cell fusion and reprogramming, rather than
transdifferentiation, is the mechanism by which hematopoi-
etic cells acquire an hepatocytic phenotype. Initial studies in
cell culture showed that BM and neuronal cells can fuse with
ES cells [131, 132]. Wang et al. [133] and Vassilopoulos et al.
[134] subsequently showed that hematopoietic stem cells fuse
with hepatocytes in Fah null mice to produce cells expressing
the deficient enzyme, which then expand massively to restore
liver mass and function [133, 134]. Fusion also occurs
between hematopoietic cells and neurons or muscle cells
[135, 136] and it has been shown that myelomonocytic cells
can fuse with hepatocytes [137, 138] or muscle cells [139] to
produce somatic hybrids expressing genes from both parental
cell types.
Other studies reported that fusion does not appear to be
required for BM‐derived cells to differentiate into hepatocytes
[140–142]. Unfractionated or CD34+ enriched cells from human
cord blood [143–146], multipotent adult progenitor cells
(MAPCs) [147, 148], or mesenchymal stem cells [149–153]
have been transplanted into the liver of immunodeficient mice.
These transplanted cells express a differentiated hepatocytic
phenotype [143, 154], but liver repopulation was once again
very low. Several studies have reported that mesenchymal stem
cells, isolated from adipose tissue and differentiated in culture
along the hepatocytic lineage, can also engraft in the liver paren-
chyma and contribute to liver regeneration [155, 156]. One
study [156] reported large repopulation clusters with hepato-
cyte‐differentiated mesenchymal stem cells, but this required
retrorsine treatment. These studies are promising, but the ability
of mesenchymal stem cells to repopulate the adult liver under
more normal, clinically viable circumstances has not been
established.
PLURIPOTENT STEM CELLS
(ES and iPSC)
Because of their extensive proliferative capacity, pluripotent
stem cells are an attractive potential source of transplantable
hepatocytes. Not only can these cells divide indefinitely, but
they also retain the ability to differentiate into multiple different
mature cell types [42]. Until 2007, ES cells were the sole source
of pluripotent human cells and are ethically controversial. Now,
however, pluripotent cells can be derived from direct genetic
reprogramming of somatic cell types, such as dermal fibroblasts
[157, 158]. Pluripotent stem cells (both ES and iPSC) cells in
culture can be induced along the endodermal and hepatocytic
lineages by addition of specific cytokines and growth factors
[159–166]. The first step typically involves the induction of
definitive endoderm using activin A. Numerous slightly differ-
ent step‐wise differentiation protocols have been developed by
different labs with the aim of generating mature hepatocytes as
the final product [159, 166]. Historically, cells produced in this
fashion expressed typical markers such as Alb, but usually also
express AFP. More recent protocols yield cells that have silenced
AFP and express multiple mature hepatocyte markers. Despite
this, genome‐wide expression analysis demonstrates that these
hepatocyte‐like cells are not fully mature and are significantly
different from primary human hepatocytes. Such cells can be
transplanted into the liver of immune deficient recipients with
differentiation into cells expressing mature hepatocyte [165,
166] and cholangiocyte markers [166]. Nonetheless, the level of
liver repopulation obtained with hepatocyte‐differentiated,
pluripotent stem cells is generally low and functional tests, such
as blood albumin measurements, fail to demonstrate equiva-
lence to primary hepatocytes [167]. The repopulation level is
somewhat higher when the cells are transplanted into MUP‐
uPA/SCID mice [166]. To date, all pluripotent stem cell differ-
entiation protocols generate only “hepatocyte‐like” cells, but
not fully functional, mature, and transplantable equivalents of
hepatocytes isolated from adult liver. More mature “hepatocyte‐
like” cells have been selected using surface markers, such as the
asialoglycoprotein receptor for purification [167], but even this
approach does not yield fully mature donor cells. In the future,
it is hoped that conditions will be developed in which lineage‐
specified pluripotent derivatives will be therapeutically equiva-
lent to primary cells.
DIRECT REPROGRAMMING
Since it is possible to generate hepatocyte‐like cells from der-
mal fibroblasts via a pluripotent intermediate stage, several
groups have reported direct reprogramming of somatic cells into
hepatocyte‐like cells, terming them iHeps, induced hepatocytes
560 THE LIVER: FUTURE HORIZONS
[168–172]. The concept of direct reprogramming using a cock-
tail of cell‐type specific transcription factors was first developed
for neurons [173]. For the liver, a combination Hnf4α plus
Foxa1, Foxa2, or Foxa3 was successfully used to directly con-
vert fibroblasts into transplantable hepatocyte‐like cells [171].
Soon thereafter, the generation of iHeps from human cells using
a similar combination of factors (FOXA3, HNF1α, and HNF4α)
was also reported [170]. While transplantation success has been
reasonably good with human iHeps in rodents and other ani-
mals, the cells in the repopulated liver are not fully differenti-
ated [174]. Nonetheless, these cells may be functional enough to
be of use in an extracorporeal bioartificial liver device (BAL)
(Lijian Hui, personal communication) and further developments
may eventually render them useful for in vivo human transplan-
tation in the future.
OTHER ROLES FOR BM CELLS IN
LIVER REGENERATION
Several studies have reported that injections of BM‐derived
stem cells can restore liver function during chronic liver injury
by enhancing the degradation of liver fibrosis in mice [175,
176]. This has been associated with induction of metalloprotein-
ases, especially MMPs 2, 9, and 13 [177]. It has been reported
that BM‐derived endothelial progenitor cells (EPCs), injected
into the spleen during liver injury, engraft in the liver, form new
blood vessels, and secrete growth factors, such as HGF, TGFα,
EGF, and VEGF, stimulating liver regeneration and improving
survival of animals with massive liver injury [178]. Thus, the
role of BM stem cells in liver regeneration may be supportive in
generating new parenchymal mass and, under some circum-
stances, in ameliorating hepatic fibrosis.
USE OF EX VIVO GENETICALLY
MODIFIED HEPATOCYTES
TO REPOPULATE THE LIVER
IN A NORMAL HEPATIC
MICROENVIRONMENT
As noted previously, bipotent fetal hepatoblasts can repopulate
the liver under normal physiologic circumstances. Because of
ethical concerns, use of adult hepatocytes for this purpose would
be preferred. Thus far, however, significant liver repopulation
by adult hepatocytes has required extensive and ongoing genetic,
physical, or chemical injury to the host liver. A substantial num-
ber of monogenic liver disorders in which there is no underlying
or ongoing liver injury could be treated by transplantation of
normal adult hepatocytes, if the cells to be transplanted were
first genetically modified to provide a proliferative and/or sur-
vival advantage compared to host hepatocytes.
Shafritz and coworkers determined that Yap1 is hyperex-
pressed in ED14 rat fetal liver cells, as well as the anti‐apop-
totic BIRC5 (survivin) gene [21]. Based on these findings, they
transduced normal adult hepatocytes with an hYap gene linked
to a modified estrogen receptor gene (ERT2), whose function
could be controlled be tamoxifen. Thus, hYapERT2 expression/
function and liver repopulation would be regulated by tamox-
ifen administration [21]. The hYapERT2 sequence was incor-
porated into a lentivirus vector which produced stable Yap
expression after transduction of normal adult hepatocytes.
Liver repopulation in the DPPIV‐ Fischer (F) 344 rat transplan-
tation system produced 8–10% liver repopulation after six
months on continuous tamoxifen feeding (Figure  44.6).
Repopulation was tightly controlled by tamoxifen administra-
tion, transplanted hepatocytes expanded into large clusters of
morphologically normal hepatocytes and integrated into host
hepatic parenchymal plates with no evidence for dysplasia,
dedifferentiation or tumorigenesis.
A major concern in using Yap to stimulate proliferation of
cells is potential tumor risk. A very recent study by Peterson
et  al. [180] showed increased repopulation by transplanted
hYapERT2 transduced hepatocytes on continuous tamoxifen
feeding for one year in the DPPIV model system with 10–20%
repopulation, no evidence for tumorigenesis, and a cell com-
petition mechanism for repopulation, as previously found with
ED14 fetal liver cells [49]. Using hYapERT2 transduced
Wistar RHA rat hepatocytes transplanted into the hyperbiliru-
binemic Gunn rat liver, there was comparable liver repopula-
tion with a progressive 70–80% reduction in serum bilirubin
on continuous tamoxifen feeding for six months. If tamoxifen
feeding was initiated six months after hepatocyte transplanta-
tion, serum bilirubin levels dropped progressively over the
next six months with the same slope as observed when tamox-
ifen administration was initiated at the time of cell transplanta-
tion. If DPPIV− rats transplanted with hYapERT2 transduced
wt hepatocytes were taken off tamoxifen after six months of
liver repopulation, the liver repopulation level did not decrease
after an additional six months on a normal diet; thus, trans-
planted hepatocytes remained in the liver long‐term. These
studies provide hope that transplantation of genetically modi-
fied hepatocytes will provide a therapeutic avenue to treat
selected liver diseases.
FUTURE HORIZONS
Many questions regarding the basic biology of stem cells in the
liver remain unresolved, including the all‐important question of
whether hepatic stem cells exist at all in the adult liver. Most
data suggest that it is possible to generate new hepatocytes
from cells other than hepatocytes themselves and that there is
an intrahepatic source of these precursors. Their precise nature
and location, however, remains unclear, as do the molecular
mechanisms that lead to their activation. Much of this uncer-
tainly is related to the lack of markers that would permit the
dissection of the complex cellular heterogeneity of the liver.
However, cell surface markers, and also genetic lineage tracing
tools, are now available to address these issues. Thus, biologi-
cal studies of antigenically defined and genetically marked
cells are now feasible and are likely to solve some of these
dilemmas. In addition, pluripotent stem cells now afford the
opportunity for “developmental biology in a dish,” making
 44:  Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 561

Figure 44.6  (a) Hepatocyte transplantation protocol. (b) Detection of lentivirus TTR‐hYapERT2 transduced hepatocytes in DPPIV− host liver by
DPPIV enzyme histochemistry. Clusters of DPPIV+ cells are clearly visible at three months after cell transplantation in tamoxifen fed rats (b1). In
the absence of tamoxifen feeding, clusters of transplanted cells are not apparent (b2). At six months after transplantation of lenti TTR‐hYapERT2
transduced hepatocytes, clusters of DPPIV+ cells are much larger in tamoxifen fed rats, and in some instances comprise whole liver lobules (b3). In
the absence of tamoxifen feeding, transplanted DPPIV+ cells are still not visible (b4). (c) Quantification of liver repopulation by scanning digital
images was performed in DPPIV− rats transplanted with lenti TTR‐hYapERT2 transduced hepatocytes, lenti TTR‐GFP transduced hepatocytes, or
non‐transduced hepatocytes, all maintained on tamoxifen feeding (+Tam) or on rats transplanted with lenti TTR‐hYapERT2 transduced hepatocytes
maintained on normal chow (−Tam).

attractive models to study molecular mechanisms of hepatic
lineage development.
Although substantial progress has been made during the past
10–15 years concerning the possibility of liver repopulation by
transplanted cells, much still needs to be learned. Factors gov-
erning engraftment of transplanted cells into the liver and their
homing to the correct niche, factors regulating proliferation and
differentiation of transplanted cells into specific phenotypes
required for organ function, and specific host conditions under
which effective liver replacement can be achieved, all need to be
determined.
The best starting point for therapeutic liver repopulation will
probably be a genetic disorder with ongoing liver injury, which
will hopefully induce or augment proliferation of transplanted
cells in the host liver. A good example of such a condition is
Wilson’s disease, in which transplanted cells might also have a
modest selective advantage [179], since they will not store high
levels of copper. Another example is α1‐antitrypsin deficiency,
in which a mutated form of α1‐antitrypsin is not secreted from
the cell and causes liver injury but much less severe than in uPA
transgenic mice. Most encouraging is the recent success in
repopulating the liver with normal hepatocytes in a mouse
model of a1‐antitrypsin deficiency [31].
To date, very few patients have been transplanted with fetal
liver cells. Studies in rats have shown that these cells have the
capacity to proliferate in the host, replace hepatic mass with
functional hepatocytes, and maintain differentiated hepatic
function, long‐term [50]. This requires a liver proliferative
stimulus or liver injury at the time the cells are transplanted,
but no selective advantage other than their ability to replace
host hepatocytes by cell competition. Use of stem cells from
pediatric or adult cadaveric liver or from other sources, such as
bone marrow, cord blood, ES cells, or iPS cells, is a possibil-
ity, and also cultured human fetal cells or adult hepatocytes
modified to favor engraftment and proliferation in the host, but
this will require substantial additional research. Cell lines have
been established, including ES cells, iPS cells, fetal liver cells,
and “oval (progenitor) cells” that also exhibit stem cell proper-
ties and differentiate into hepatocytes and/or bile ducts in vitro
and in vivo. However, these cell lines have shown only limited
repopulation of the normal liver at the current state of the art.
In order to advance the field of liver cell therapy further, it will
be necessary to find conditions under which cells and cell lines
derived from ES cells, iPS cells, fetal liver, or adult liver can
be expanded in culture and successfully repopulate the liver
under conditions that will be clinically acceptable. Use of
newly emerging three dimensional or “organoid” culture con-
ditions to maintain and expand hepatic derived cells or cells in
the hepatic lineage will help to advance this field (see
Chapter 77). Although we are not yet there, restoration of liver
function by therapeutic cell transplantation holds great prom-
ise for the future.
562 THE LIVER:  REFERENCES
ACKNOWLEDGMENTS
The authors would like to thank Anna Caponigro for assistance
in preparing the manuscript and Dr Irmin Sternlieb, a former
colleague at Albert Einstein College of Medicine, who provided
the electron micrograph illustrating a canal of Hering to D.A.S.
REFERENCES
  1. Boeck, J. and Popper, H. Ueber Lebertransplantation in die Vorderkammer
des Auges. Virchow Arch Path Anat, 1937;299:219–234.
  2. Matas, A.J., Sutherland, D.E., Steffes, M.W. et al. Hepatocellular transplan-
tation for metabolic deficiencies: decrease of plasms bilirubin in Gunn rats.
Science, 1976;192:892–4.
  3. Grompe, M. Liver repopulation for the treatment of metabolic diseases. J
Inherit Metab Dis, 2001;24:231–44.
  4. Oishi, K., Arnon, R., Wasserstein, M.P. et al. Liver transplantation for pediat-
ric inherited metabolic disorders: Considerations for indications, complica-
tions, and perioperative management. Pediatr Transplant, 2016;20:756–69.
  5. Strom, S.C., Fisher, R.A., Thompson, M.T. et al. Hepatocyte transplantation
as a bridge to orthotopic liver transplantation in terminal liver failure.
Transplantation, 1997;63:559–69.
  6. Bilir, B.M., Guinette, D., Karrer, F. et al. Hepatocyte transplantation in acute
liver failure. Liver Transpl, 2000;6:32–40.
  7. Grossman, M., Rader, D.J., Muller, D.W. et al. A pilot study of ex vivo gene
therapy for homozygous familial hypercholesterolaemia. Nat Med,
1995;1:1148–54.
  8. Bucher, N.L. and Swaffield, M.N. The rate of incorporation of labeled thy-
midine into the deoxyribonucleic acid of regenerating rat liver in relation to
the amount of liver excised. Cancer Res, 1964;24:1611–25.
  9. Grisham, J.W. A morphologic study of deoxyribonucleic acid synthesis and
cell proliferation in regenerating rat liver; autoradiography with thymidine‐
H3. Cancer Res, 1962;22:842–9.
10. Rajvanshi, P., Kerr, A., Bhargava, K.K. et al. Studies of liver repopulation
using the dipeptidyl peptidase IV‐deficient rat and other rodent recipients:
cell size and structure relationships regulate capacity for increased trans-
planted hepatocyte mass in the liver lobule. Hepatology, 1996;23:482–96.
11. Rajvanshi, P., Kerr, A., Bhargava, K.K. et al. Efficacy and safety of repeated
hepatocyte transplantation for significant liver repopulation in rodents.
Gastroenterology, 1996;111:1092–1102.
12. Rozga, J., Holzman, M., Moscioni, A.D. et al. Repeated intraportal hepato-
cyte transplantation in analbuminemic rats. Cell Transplant,
1995;4:237–43.
13. Fox, I.J., Chowdhury, J.R., Kaufman, S.S. et al. Treatment of the Crigler–
Najjar syndrome type I with hepatocyte transplantation. N Engl J Med,
1998;338:1422–6.
14. Strom, S.C., Chowdhury, J.R., Fox, I.J. Hepatocyte transplantation for the
treatment of human disease. Semin Liver Dis, 1999;19:39–48.
15. Iansante, V., Chandrashekran, A. and Dhawan, A. Cell‐based liver therapies:
past, present and future. Philos Trans R Soc Lond B Biol Sci, 2018;373.
16. Squires, J.E., Soltys, K.A., McKiernan, P. et al. Clinical hepatocyte trans-
plantation: what is next? Curr Transplant Rep, 2017;4:280–289.
17. Michalopoulos, G.K. Advances in liver regeneration. Expert Rev
Gastroenterol Hepatol, 2014;8:897–907.
18. Miyajima, A., Tanaka, M., and Itoh, T. Stem/progenitor cells in liver devel-
opment, homeostasis, regeneration, and reprogramming. Cell Stem Cell,
2014;14:561–74.
19. Dong, J., Feldmann, G., Huang, J. et al. Elucidation of a universal size‐con-
trol mechanism in Drosophila and mammals. Cell, 2007;130:1120–33.
20. Camargo, F.D., Gokhale, S., Johnnidis, J.B. et al. YAP1 increases organ size
and expands undifferentiated progenitor cells. Curr Biol, 2007;17:2054–60.
21. Yovchev, M., Jaber, F.L., Lu, Z. et al. experimental model for successful liver
cell therapy by lenti TTR‐YapERT2 transduced hepatocytes with tamoxifen
control of Yap subcellular location. Sci Rep, 2016;6:19275.
22. Sandgren, E.P., Palmiter, R.D., Heckel, J.L. et al. Complete hepatic regen-
eration after somatic deletion of an albumin‐plasminogen activator transgene.
Cell, 1991;66:245–56.
23. Rhim, J.A., Sandgren, E.P., Degen, J.L. et  al. Replacement of diseased
mouse liver by hepatic cell transplantation. Science, 1994;263:1149–52.
24. Overturf, K., Al‐Dhalimy, M., Tanguay, R. et al. Hepatocytes corrected by
gene therapy are selected in vivo in a murine model of hereditary tyrosinae-
mia type I. Nat Genet, 1996;12:266–73.
25. Lindstedt, S., Holme, E., Lock, E.A. et al. Treatment of hereditary tyrosinae-
mia type I by inhibition of 4‐hydroxyphenylpyruvate dioxygenase. Lancet,
1992;340:813–7.
26. Overturf, K., al‐Dhalimy, M., Ou, C.N. et al. Serial transplantation reveals
the stem‐cell‐like regenerative potential of adult mouse hepatocytes. Am J
Pathol, 1997;151:1273–80.
27. Wang, M.J., Chen, F., Li, J.X. et al. Reversal of hepatocyte senescence after
continuous in vivo cell proliferation. Hepatology, 2014;60:349–61.
28. Grompe, M. and Strom, S. Mice with human livers. Gastroenterology,
2013;145:1209–14.
29. Washburn, M.L., Bility, M.T., Zhang, L. et al. A humanized mouse model to
study hepatitis C virus infection, immune response, and liver disease.
Gastroenterology, 2011;140:1334–44.
30. Hasegawa, M., Kawai, K., Mitsui, T. et  al. The reconstituted “humanized
liver” in TK‐NOG mice is mature and functional. Biochem Biophys Res
Commun, 2011;405:405–10.
31. Ding, J., Yannam, G.R., Roy‐Chowdhury, N. et  al. Spontaneous hepatic
repopulation in transgenic mice expressing mutant human alpha1‐antitrypsin
by wild‐type donor hepatocytes. J Clin Invest, 2011;121:1930–4.
32. Lu, W.Y., Bird, T.G., Boulter, L. et  al. Hepatic progenitor cells of biliary
origin with liver repopulation capacity. Nat Cell Biol, 2015;17:971–83.
33. Willenbring, H., Sharma, A.D., Vogel, A. et al. Loss of p21 permits carcino-
genesis from chronically damaged liver and kidney epithelial cells despite
unchecked apoptosis. Cancer Cell, 2008;14:59–67.
34. Laconi, E., Oren, R., Mukhopadhyay, D.K. et al. Long‐term, near‐total liver
replacement by transplantion of isolated hepatocytes in rats treated with ret-
rorsine. Am J Pathol, 1998;153:319–29.
35. Guha, C., Sharma, A., Gupta, S. et  al. Amelioration of radiation‐induced
liver damage in partially hepatectomized rats by hepatocyte transplantation.
Cancer Res, 1999;59:5871–4.
36. Malhi, H., Gorla, G.R., Irani, A.N. et  al. Cell transplantation after
­oxidative hepatic preconditioning with radiation and ischemia‐reperfusion
leads  to extensive liver repopulation. Proc Natl Acad Sci USA,
2002;99:13114–9.
37. Soltys, K.A., Setoyama, K., Tafaleng, E.N. et  al. Host conditioning and
rejection monitoring in hepatocyte transplantation in humans. J Hepatol,
2017;66:987–1000.
38. Dandri, M., Burda, M.R., Torok, E. et al. Repopulation of mouse liver with
human hepatocytes and in vivo infection with hepatitis B virus. Hepatology,
2001;33:981–8.
39. Tateno, C., Yoshizane, Y., Saito, N. et al. Near completely humanized liver in
mice shows human‐type metabolic responses to drugs. Am J Pathol,
2004;165:901–12.
40. Azuma, H., Paulk, N., Ranade, A. et al. Robust expansion of human hepato-
cytes in Fah‐/‐/Rag2‐/‐/Il2rg‐/‐ mice. Nat Biotechnol, 2007;25:903–10.
41. Kopp, J.L., Grompe, M., and Sander, M. Stem cells versus plasticity in liver
and pancreas regeneration. Nat Cell Biol, 2016;18:238–45.
42. Thomson, J.A., Itskovitz‐Eldor, J., Shapiro, S.S. et al. Embryonic stem cell
lines derived from human blastocysts. Science, 1998;282:1145–7.
43. Fuchs, E. and Segre, J.A. Stem cells: a new lease on life. Cell,
2000;100:143–55.
44. Weissman, I.L. Stem cells: units of development, units of regeneration, and
units in evolution. Cell, 2000;100:157–68.
45. Marshman, E., Booth, C., and Potten, C.S. The intestinal epithelial stem cell.
Bioessays, 2002;24:91–8.
46. Lechler, T. and Fuchs, E. Asymmetric cell divisions promote stratification
and differentiation of mammalian skin. Nature, 2005;437:275–80.
47. Sandhu, J.S., Petkov, P.M., Dabeva, M.D. et  al. Stem cell properties and
repopulation of the rat liver by fetal liver epithelial progenitor cells. Am J
Pathol, 2001;159:1323–34.
48. Dabeva, M.D., Petkov, P.M., Sandhu, J. et al. Proliferation and differentia-
tion of fetal liver epithelial progenitor cells after transplantation into adult rat
liver. Am J Pathol, 2000;156:2017–31.
49. Oertel, M., Menthena, A., Dabeva, M.D. et al. Cell competition leads to a
high level of normal liver reconstitution by transplanted fetal liver stem/pro-
genitor cells. Gastroenterology, 2006;130:507–20.
 44:  Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 563
50. Shafritz, D.A. and Oertel, M. Model systems and experimental conditions
that lead to effective repopulation of the liver by transplanted cells. Int J
Biochem Cell Biol, 2011;43:198–213.
51. Oertel, M., Menthena, A., Chen, Y.Q. et al. Purification of fetal liver stem/
progenitor cells containing all the repopulation potential for normal adult rat
liver. Gastroenterology, 2008;134:823–32.
52. Moreno, E. and Basler, K. dMyc transforms cells into super‐competitors.
Cell, 2004;117:117–29.
53. de la Cova, C., Abril, M., Bellosta, P. et al. Drosophila myc regulates organ
size by inducing cell competition. Cell, 2004;117:107–16.
54. Oertel, M., Menthena, A., Chen, Y.Q. et al. Properties of cryopreserved fetal
liver stem/progenitor cells that exhibit long‐term repopulation of the normal
rat liver. Stem Cells, 2006;24:2244–51.
55. Schmelzer, E., Wauthier, E., and Reid, L.M. The phenotypes of pluripotent
human hepatic progenitors. Stem Cells, 2006;24:1852–8.
56. Huch, M., Dorrell, C., Boj, S.F. et al. In vitro expansion of single Lgr5+ liver
stem cells induced by Wnt‐driven regeneration. Nature, 2013:1–6.
57. Huch, M., Gehart, H., van Boxtel, R. et al. Long‐term culture of genome‐sta-
ble bipotent stem cells from adult human liver. Cell, 2015;160:299–312.
58. Dorrell, C., Erker, L., Schug, J. et al. Prospective isolation of a bipotential clo-
nogenic liver progenitor cell in adult mice. Genes Dev, 2011;25:1193–203.
59. Shin, S., Walton, G., Aoki, R. et al. Foxl1‐Cre‐marked adult hepatic progeni-
tors have clonogenic and bilineage differentiation potential. Genes Dev,
2011;25:1185–92.
60. Suzuki, A., Sekiya, S., Onishi, M. et al. Flow cytometric isolation and clonal
identification of self‐renewing bipotent hepatic progenitor cells in adult
mouse liver. Hepatology, 2008;48:1964–78.
61. Evarts, R.P., Nagy, P., Marsden, E. et  al. A precursor‐product relationship
exists between oval cells and hepatocytes in rat liver. Carcinogenesis,
1987;8:1737–40.
62. Evarts, R.P., Nagy, P., Nakatsukasa, H. et  al. In vivo differentiation of rat
liver oval cells into hepatocytes. Cancer Res, 1989;49:1541–7.
63. Schofield, R. The relationship between the spleen colony‐forming cell and
the haemopoietic stem cell. Blood Cells, 1978;4:7–25.
64. Trentin, J.J. Determination of bone marrow stem cell differentiation by stro-
mal hemopoietic inductive microenvironments (HIM). Am J Pathol,
1971;65:621–8.
65. Ohlstein, B., Kai, T., Decotto, E. et al. The stem cell niche: theme and varia-
tions. Curr Opin Cell Biol, 2004;16:693–9.
66. Lapidot, T., Dar, A., and Kollet, O. How do stem cells find their way home?
Blood, 2005;106:1901–10.
67. Christiano, A.M. Epithelial stem cells: stepping out of their niche. Cell,
2004;118:530–2.
68. Williams, E.D., Lowes, A.P., Williams, D. et al. A stem cell niche theory of
intestinal crypt maintenance based on a study of somatic mutation in colonic
mucosa. Am J Pathol, 1992;141:773–6.
69. Tumbar, T., Guasch, G., Greco, V. et  al. Defining the epithelial stem cell
niche in skin. Science, 2004;303:359–63.
70. Parati, E.A., Pozzi, S., Ottolina, A. et al. Neural stem cells: an overview. J
Endocrinol Invest, 2004;27:64–7.
71. Hering, E. Ueber den Bau der Wirbeltierleber. Sitzungsberichte der
Kaiserlichen Akademie der Wissenschaften. Mathematisch‐
Naturwissenschaftliche Klasse 1866;Bd. 54:335–341.
72. Steiner, J.W. and Carruthers, J.S. Studies on the fine structure of the terminal
branches of the biliary tree: II. Observations of pathologically altered bile
canaliculi. Am J Pathol, 1961;39:41–63.
73. Theise, N.D., Saxena, R., Portmann, B.C. et al. The canals of Hering and
hepatic stem cells in humans. Hepatology, 1999;30:1425–33.
74. Saxena, R., Theise, N.D., and Crawford, J.M. Microanatomy of the human
liver‐exploring the hidden interfaces. Hepatology, 1999;30:1339–46.
75. Yovchev, M.I., Grozdanov, P.N., Joseph, B. et al. Novel hepatic progenitor
cell surface markers in the adult rat liver. Hepatology, 2007;45:139–49.
76. Farber, E. Similarities in the sequence of early histological changes induced
in the liver of the rat by ethionine, 2‐acetylamino‐fluorene, and 3’‐methyl‐4‐
dimethylaminoazobenzene. Cancer Res, 1956;16:142–8.
77. Lemire, J.M., Shiojiri, N., and Fausto, N. Oval cell proliferation and the ori-
gin of small hepatocytes in liver injury induced by D‐galactosamine. Am J
Pathol, 1991;139:535–52.
78. Dabeva, M.D. and Shafritz, D.A. Activation, proliferation, and differentia-
tion of progenitor cells into hepatocytes in the D‐galactosamine model of
liver regeneration. Am J Pathol, 1993;143:1606–20.
  79. Yin, L., Lynch, D., and Sell, S. Participation of different cell types in the
restitutive response of the rat liver to periportal injury induced by allyl alco-
hol. J Hepatol, 1999;31:497–507.
  80. Fujio, K., Evarts, R.P., Hu, Z. et al. Expression of stem cell factor and its
receptor, c‐kit, during liver regeneration from putative stem cells in adult
rat. Lab Invest, 1994;70:511–6.
  81. Omori, N., Omori, M., Evarts, R.P. et al. Partial cloning of rat CD34 cDNA
and expression during stem cell‐dependent liver regeneration in the adult
rat. Hepatology, 1997;26:720–7.
  82. Omori, M., Omori, N., Evarts, R.P. et al. Coexpression of flt‐3 ligand/flt‐3
and SCF/c‐kit signal transduction system in bile‐duct‐ligated SI and W
mice. Am J Pathol, 1997;150:1179–87.
  83. Omori, N., Evarts, R.P., Omori, M. et al. Expression of leukemia inhibitory
factor and its receptor during liver regeneration in the adult rat. Lab Invest,
1996;75:15–24.
  84. Crosby, H.A., Kelly, D.A., and Strain, A.J. Human hepatic stem‐like cells
isolated using c‐kit or CD34 can differentiate into biliary epithelium.
Gastroenterology, 2001;120:534–44.
  85. Paku, S., Schnur, J., Nagy, P. et al. Origin and structural evolution of the
early proliferating oval cells in rat liver. Am J Pathol, 2001;158:1313–23.
  86. Sells, M.A., Katyal, S.L., Shinozuka, H. et al. Isolation of oval cells and
transitional cells from the livers of rats fed the carcinogen DL‐ethionine. J
Natl Cancer Inst, 1981;66:355–62.
  87. Akhurst, B., Croager, E.J., Farley‐Roche, C.A. et al. A modified choline‐
deficient, ethionine‐supplemented diet protocol effectively induces oval
cells in mouse liver. Hepatology, 2001;34:519–22.
  88. Factor, V.M., Radaeva, S.A., Thorgeirsson, S.S. Origin and fate of oval cells
in dipin‐induced hepatocarcinogenesis in the mouse. Am J Pathol,
1994;145:409–22.
 89. Preisegger, K.H., Factor, V.M., Fuchsbichler, A. et  al. Atypical ductular
proliferation and its inhibition by transforming growth factor beta1 in the
3,5‐diethoxycarbonyl‐1,4‐dihydrocollidine mouse model for chronic alco-
holic liver disease. Lab Invest, 1999;79:103–9.
  90. Yanger, K., Knigin, D., Zong, Y. et al. Adult hepatocytes are generated by
self‐duplication rather than stem cell differentiation. Cell Stem Cell,
2014;15:340–9.
  91. Wright, N., Samuelson, L., Walkup, M.H. et al. Enrichment of a bipotent
hepatic progenitor cell from naive adult liver tissue. Biochem Biophys Res
Commun, 2008;366:367–72.
  92. Petersen, B.E., Goff, J.P., Greenberger, J.S. et al. Hepatic oval cells express the
hematopoietic stem cell marker Thy‐1 in the rat. Hepatology, 1998;27:433–45.
  93. Nierhoff, D., Ogawa, A., Oertel, M. et al. Purification and characterization
of mouse fetal liver epithelial cells with high in vivo repopulation capacity.
Hepatology, 2005;42:130–9.
 94. Suzuki, A., Zheng, Y., Kondo, R. et  al. Flow‐cytometric separation and
enrichment of hepatic progenitor cells in the developing mouse liver.
Hepatology, 2000;32:1230–9.
  95. Dezso, K., Jelnes, P., Laszlo, V. et al. Thy‐1 is expressed in hepatic myofi-
broblasts and not oval cells in stem cell‐mediated liver regeneration. Am J
Pathol, 2007;171:1529–37.
 96. Tanimizu, N., Nishikawa, M., Saito, H. et  al. Isolation of hepatoblasts
based on the expression of Dlk/Pref‐1. J Cell Sci, 2003;116:1775–86.
  97. Faris, R.A. and Hixson, D.C. Selective proliferation of chemically altered
rat liver epithelial cells following hepatic transplantation. Transplantation,
1989;48:87–92.
  98. Dabeva, M.D., Hwang, S.G., Vasa, S.R. et al. Differentiation of pancreatic
epithelial progenitor cells into hepatocytes following transplantation into
rat liver. Proc Natl Acad Sci USA, 1997;94:7356–61.
 99. Dabeva, M.D., Hurston, E., and Shafritz, D.A. Transcription factor and
liver‐specific mRNA expression in facultative epithelial progenitor cells of
liver and pancreas. Am J Pathol, 1995;147:1633–48.
100. Wang, X., Al‐Dhalimy, M., Lagasse, E. et al. Liver repopulation and cor-
rection of metabolic liver disease by transplanted adult mouse pancreatic
cells. Am J Pathol, 2001;158:571–9.
101. Wang, X., Foster, M., Al‐Dhalimy, M. et al. The origin and liver repopulating
capacity of murine oval cells. Proc Natl Acad Sci USA, 2003;100(1):11881–8.
102. Song, S., Witek, R.P., Lu, Y. et al. Ex vivo transduced liver progenitor cells
as a platform for gene therapy in mice. Hepatology, 2004;40:918–24.
103. Roskams, T., van den Oord, J.J., De Vos, R. et al. Neuroendocrine features
of reactive bile ductules in cholestatic liver disease. Am J Pathol,
1990;137:1019–25.
564 THE LIVER:  REFERENCES
104. Demetris, A.J., Seaberg, E.C., Wennerberg, A. et al. Ductular reaction after
submassive necrosis in humans. Special emphasis on analysis of ductular
hepatocytes. Am J Pathol, 1996;149:439–48.
105. Roskams, T., De Vos, R., Van Eyken, P. et al. Hepatic OV‐6 expression in
human liver disease and rat experiments: evidence for hepatic progenitor
cells in man. J Hepatol, 1998;29:455–63.
106. Roskams, T.A., Theise, N.D., Balabaud, C. et al. Nomenclature of the finer
branches of the biliary tree: canals, ductules, and ductular reactions in
human livers. Hepatology, 2004;39:1739–45.
107. Zhou, H., Rogler, L.E., Teperman, L. et al. Identification of hepatocytic and
bile ductular cell lineages and candidate stem cells in bipolar ductular reac-
tions in cirrhotic human liver. Hepatology, 2007;45:716–24.
108. Haruna, Y., Saito, K., Spaulding, S. et al. Identification of bipotential pro-
genitor cells in human liver development. Hepatology, 1996;23:476–81.
109. Malhi, H., Irani, A.N., Gagandeep, S. et al. Isolation of human progenitor
liver epithelial cells with extensive replication capacity and differentiation
into mature hepatocytes. J Cell Sci, 2002;115:2679–88.
110. Lazaro, C.A., Croager, E.J., Mitchell, C. et al. Establishment, characteriza-
tion, and long‐term maintenance of cultures of human fetal hepatocytes.
Hepatology, 2003;38:1095–106.
111. Mahieu‐Caputo, D., Allain, J.E., Branger, J. et al. Repopulation of athymic
mouse liver by cryopreserved early human fetal hepatoblasts. Hum Gene
Ther, 2004;15:1219–28.
112. Barker, N., van Es, J.H., Kuipers, J. et  al. Identification of stem cells in
small intestine and colon by marker gene Lgr5. Nature, 2007;449:1003–7.
113. Sato, T., Vries, R.G., Snippert, H.J. et al. Single Lgr5 stem cells build crypt‐
villus structures in vitro without a mesenchymal niche. Nature,
2009;459:262–5.
114. Schaub, J.R., Malato, Y., Gormond, C. et al. Evidence against a stem cell
origin of new hepatocytes in a common mouse model of chronic liver
injury. Cell Rep, 2014;8:933–9.
115. Tarlow, B.D., Pelz, C., Naugler, W.E. et al. Bipotential adult liver progeni-
tors are derived from chronically injured mature hepatocytes. Cell Stem
Cell, 2014;15:605–18.
116. Grompe, M. Liver stem cells, where art thou? Cell Stem Cell,
2014;15:257–8.
117. Schaub, J.R., Huppert, K.A., Kurial, S.N.T. et al. De novo formation of the
biliary system by TGFbeta‐mediated hepatocyte transdifferentiation.
Nature, 2018;557:247–51.
118. Michalopoulos, G.K., Barua, L., and Bowen, W.C. Transdifferentiation of
rat hepatocytes into biliary cells after bile duct ligation and toxic biliary
injury. Hepatology, 2005;41:535–44.
119. Yovchev, M.I., Locker, J., and Oertel, M. Biliary fibrosis drives liver repop-
ulation and phenotype transition of transplanted hepatocytes. J Hepatol,
2016;64:1348–57.
120. Katsuda, T., Kawamata, M., Hagiwara, K. et al. Conversion of terminally
committed hepatocytes to culturable bipotent progenitor cells with regen-
erative capacity. Cell Stem Cell, 2017;20:41–55.
121. Raven, A., Lu, W.Y., Man, T.Y. et al. Cholangiocytes act as facultative liver
stem cells during impaired hepatocyte regeneration. Nature, 2017;
547:350–4.
122. Deng, X., Zhang, X., Li, W. et al. Chronic liver injury induces conversion
of biliary epithelial cells into hepatocytes. Cell Stem Cell, 2018;23:
114–22.
123. Wang, B., Zhao, L., and Fish, M. et al. Self‐renewing diploid Axin2(+) cells
fuel homeostatic renewal of the liver. Nature, 2015;524:180–5.
124. Lin, S., Nascimento, E.M., Gajera, C.R. et  al. Distributed hepatocytes
expressing telomerase repopulate the liver in homeostasis and injury.
Nature, 2018;556:244–8.
125. Li, B., Dorrell, C., Canaday, P.S. et al. Adult mouse liver contains two dis-
tinct populations of cholangiocytes. Stem Cell Reports, 2017;9:478–89.
126. Goodell, M.A. Stem‐cell “plasticity”: befuddled by the muddle. Curr Opin
Hematol, 2003;10:208–13.
127. Wagers, A.J. and Weissman, I.L. Plasticity of adult stem cells. Cell,
2004;116:639–48.
128. Thorgeirsson, S.S. and Grisham, J.W. Hematopoietic cells as hepatocyte
stem cells: a critical review of the evidence. Hepatology, 2006;43:2–8.
129. Petersen, B.E., Bowen, W.C., Patrene, K.D. et al. Bone marrow as a poten-
tial source of hepatic oval cells. Science, 1999;284:1168–70.
130. Menthena, A., Deb, N., Oertel, M. et al. Bone marrow progenitors are not the
source of expanding oval cells in injured liver. Stem Cells, 2004;22:1049–61.
131. Terada, N., Hamazaki, T., Oka, M. et al. Bone marrow cells adopt the phe-
notype of other cells by spontaneous cell fusion. Nature, 2002;416:542–5.
132. Ying, Q.L., Nichols, J., Evans, E.P. et al. Changing potency by spontaneous
fusion. Nature, 2002;416:545–8.
133. Wang, X., Willenbring, H., Akkari, Y. et  al. Cell fusion is the principal
source of bone‐marrow‐derived hepatocytes. Nature, 2003;422:897–901.
134. Vassilopoulos, G., Wang, P.R., and Russell, D.W. Transplanted bone mar-
row regenerates liver by cell fusion. Nature, 2003;422:901–4.
135. Alvarez‐Dolado, M., Pardal, R., Garcia‐Verdugo, J.M. et  al. Fusion of
bone‐marrow‐derived cells with Purkinje neurons, cardiomyocytes and
hepatocytes. Nature, 2003;425:968–73.
136. Weimann, J.M., Johansson, C.B., Trejo, A. et al. Stable reprogrammed het-
erokaryons form spontaneously in Purkinje neurons after bone marrow
transplant. Nat Cell Biol, 2003;5:959–66.
137. Camargo, F.D., Finegold, M., and Goodell, M.A. Hematopoietic myelo-
monocytic cells are the major source of hepatocyte fusion partners. J Clin
Invest, 2004;113:1266–70.
138. Willenbring, H., Bailey, A.S., Foster, M. et al. Myelomonocytic cells are
sufficient for therapeutic cell fusion in liver. Nat Med, 2004;10:744–8.
139. Camargo, F.D., Green, R., Capetanaki, Y. et al. Single hematopoietic stem
cells generate skeletal muscle through myeloid intermediates. Nat Med,
2003;9:1520–7.
140. Newsome, P.N., Johannessen, I., Boyle, S. et al. Human cord blood‐derived
cells can differentiate into hepatocytes in the mouse liver with no evidence
of cellular fusion. Gastroenterology, 2003;124:1891–900.
141. Harris, R.G., Herzog, E.L., Bruscia, E.M. et al. Lack of a fusion require-
ment for development of bone marrow‐derived epithelia. Science,
2004;305:90–3.
142. Jang, Y.Y., Collector, M.I., Baylin, S.B. et al. Hematopoietic stem cells con-
vert into liver cells within days without fusion. Nat Cell Biol, 2004;
6:532–9.
143. Danet, G.H., Luongo, J.L., Butler, G. et al. C1qRp defines a new human
stem cell population with hematopoietic and hepatic potential. Proc Natl
Acad Sci USA, 2002;99:10441–5.
144. Wang, X., Ge, S., McNamara, G. et al. Albumin‐expressing hepatocyte‐like
cells develop in the livers of immune‐deficient mice that received trans-
plants of highly purified human hematopoietic stem cells. Blood,
2003;101:4201–8.
145. Kakinuma, S., Tanaka, Y., Chinzei, R. et al. Human umbilical cord blood as
a source of transplantable hepatic progenitor cells. Stem Cells, 2003;
21:217–27.
146. Kollet, O., Shivtiel, S., Chen, Y.Q. et  al. HGF, SDF‐1, and MMP‐9 are
involved in stress‐induced human CD34+ stem cell recruitment to the liver.
J Clin Invest, 2003;112:160–9.
147. Jiang, Y., Jahagirdar, B.N., Reinhardt, R.L. et al. Pluripotency of mesen-
chymal stem cells derived from adult marrow. Nature, 2002;418:41–9.
148. Schwartz, R.E., Reyes, M., Koodie, L. et al. Multipotent adult progenitor
cells from bone marrow differentiate into functional hepatocyte‐like cells. J
Clin Invest, 2002;109:1291–302.
149. Lee, O.K., Kuo, T.K., Chen, W.M. et al. Isolation of multipotent mesenchy-
mal stem cells from umbilical cord blood. Blood, 2004;103:1669–75.
150. Anjos‐Afonso, F., Siapati, E.K., and Bonnet, D. In vivo contribution of
murine mesenchymal stem cells into multiple cell‐types under minimal
damage conditions. J Cell Sci, 2004;117:5655–64.
151. Kogler, G., Sensken, S., Airey, J.A. et al. A new human somatic stem cell
from placental cord blood with intrinsic pluripotent differentiation poten-
tial. J Exp Med, 2004;200:123–35.
152. Lee, K.D., Kuo, T.K., Whang‐Peng, J. et al. In vitro hepatic differentiation
of human mesenchymal stem cells. Hepatology, 2004;40:1275–84.
153. Aurich, I., Mueller, L.P., Aurich, H. et al. Functional integration of hepato-
cytes derived from human mesenchymal stem cells into mouse livers. Gut,
2007;56:405–15.
154. Sato, Y., Araki, H., Kato, J. et al. Human mesenchymal stem cells xeno-
grafted directly to rat liver are differentiated into human hepatocytes with-
out fusion. Blood, 2005;106:756–63.
155. Banas, A., Teratani, T., Yamamoto, Y. et al. Adipose tissue‐derived mesen-
chymal stem cells as a source of human hepatocytes. Hepatology,
2007;46:219–28.
156. Sgodda, M., Aurich, H., Kleist, S. et al. Hepatocyte differentiation of mes-
enchymal stem cells from rat peritoneal adipose tissue in vitro and in vivo.
Exp Cell Res, 2007;313:2875–86.
 44:  Liver Repopulation by Cell Transplantation and the Role of Stem Cells in Liver Biology 565
157. Takahashi, K., Tanabe, K., Ohnuki, M. et al. Induction of pluripotent stem
cells from adult human fibroblasts by defined factors. Cell,
2007;131:861–72.
158. Yu, J., Vodyanik, M.A., Smuga‐Otto, K. et al. Induced pluripotent stem cell
lines derived from human somatic cells. Science, 2007;318:1917–20.
159. Hamazaki, T., Iiboshi, Y., Oka, M. et al. Hepatic maturation in differentiat-
ing embryonic stem cells in vitro. FEBS Lett, 2001;497:15–9.
160. Jones, E.A., Tosh, D., Wilson, D.I. et al. Hepatic differentiation of murine
embryonic stem cells. Exp Cell Res, 2002;272:15–22.
161. Yamada, T., Yoshikawa, M., Kanda, S. et  al. In vitro differentiation of
embryonic stem cells into hepatocyte‐like cells identified by cellular uptake
of indocyanine green. Stem Cells, 2002;20:146–54.
162. Yamamoto, H., Quinn, G., Asari, A. et  al. Differentiation of embryonic
stem cells into hepatocytes: biological functions and therapeutic applica-
tion. Hepatology, 2003;37:983–93.
163. Rambhatla, L., Chiu, C.P., Kundu, P. et al. Generation of hepatocyte‐like
cells from human embryonic stem cells. Cell Transplant, 2003;12:1–11.
164. Kubo, A., Shinozaki, K., Shannon, J.M. et al. Development of definitive endo-
derm from embryonic stem cells in culture. Development, 2004;131:1651–62.
165. Gouon‐Evans, V., Boussemart, L., Gadue, P. et al. BMP‐4 is required for
hepatic specification of mouse embryonic stem cell‐derived definitive
endoderm. Nat Biotechnol, 2006;24:1402–11.
166. Heo, J., Factor, V.M., Uren, T. et  al. Hepatic precursors derived from
murine embryonic stem cells contribute to regeneration of injured liver.
Hepatology, 2006;44:1478–86.
167. Basma, H., Soto‐Gutierrez, A., Yannam, G.R. et  al. Differentiation and
transplantation of human embryonic stem cell‐derived hepatocytes.
Gastroenterology, 2009;136:990–9.
168. Suzuki, A. Artificial induction and disease‐related conversion of the hepatic
fate. Curr Opin Genet Dev, 2013;23:579–84.
169. Du, Y., Wang, J., Jia, J. et al. Human hepatocytes with drug metabolic func-
tion induced from fibroblasts by lineage reprogramming. Cell Stem Cell,
2014;14:394–403.
170. Huang, P., Zhang, L., Gao, Y. et  al. Direct reprogramming of human
fibroblasts to functional and expandable hepatocytes. Cell Stem Cell,
­
2014;14:370–84.
171. Sekiya, S. and Suzuki, A. Direct conversion of mouse fibroblasts to
­hepatocyte‐like cells by defined factors. Nature, 2011;475:390–3.
172. Zhu, S., Rezvani, M., Harbell, J. et  al. Mouse liver repopulation with
­hepatocytes generated from human fibroblasts. Nature, 2014;508:93–7.
173. Vierbuchen, T., Ostermeier, A., Pang, Z.P. et  al. Direct conversion of
­fibroblasts to functional neurons by defined factors. Nature, 2010;463:
1035–41.
174. Roy‐Chowdhury, N., Wang, X., Guha, C. et  al. Hepatocyte‐like cells
derived from induced pluripotent stem cells. Hepatol Int, 2017;11:54–69.
175. Sakaida, I., Terai, S., Yamamoto, N. et al. Transplantion of bone marrow
cells reduces CCl4‐induced liver fibrosis in mice. Hepatology,
2004;40:1304–11.
176. Ueno, T., Nakamura, T., Torimura, T. et  al. Angiogenic cell therapy for
hepatic fibrosis. Med Mol Morphol, 2006;39:16–21.
177. Higashiyama, R., Inagaki, Y., Hong, Y.Y. et al. Bone marrow‐derived cells
express matrix metalloproteinases and contribute to regression of liver
fibrosis in mice. Hepatology, 2007;45:213–22.
178. Taniguchi, E., Kin, M., Torimura, T. et  al. Endothelial progenitor cell
transplantation improves the survival following liver injury in mice.
­
Gastroenterology, 2006;130:521–31.
179. Yoshida, Y., Tokusashi, Y., Lee, G.H. et al. Intrahepatic transplantation of
normal hepatocytes prevents Wilson’s disease in Long‐Evans cinnamon
rats. Gastroenterology, 1996;111:1654–60.
180. Peterson, E.A., Polgar, Z., Devakanmalai, G.S. et al. Genes and pathways
promoting long‐term liver repopulation by ex vivo hYAP‐ERT2 transduced
hepatocytes and treatment of jaundice in Gunn rats. Hepatol Commun,
2018;3:129–46.
 Liver Regeneration
45 George K. Michalopoulos
Department of Pathology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA
INTRODUCTION
Liver is the largest organ of the body. It functions as a portal of
entry and metabolic processing for all substances absorbed
through the gastrointestinal tract. It is the major biochemical
transducer for body homeostasis, triaging the absorbed sub-
stances from the gut for storage, elimination, transformation
into other types of organic chemicals, or packaged transport to
the bloodstream for whole body utilization. To the extent that
muscles work mostly on fatty acids and brain primarily on glu-
cose, liver controls the availability of both the nutrients. The key
role of liver as the “provider” for metabolism of all body tissues
is perhaps the reason for an established tight relationship
between liver weight and body weight. Changes in body status
(cachexia, puberty, pregnancy, chronic disease, etc.) typically
alter the algorithm of the “liver to body weight ratio” (LBWR).
Whatever the operative algorithm of LBWR may be for a given
body at any given time, the LBWR is tightly maintained, by
mechanisms not fully understood. The rates of proliferation of
hepatic cells under normal conditions are typically very low (for
hepatocytes, less than 0.2%), but they exceed zero. Recent stud-
ies have demonstrated that under the placid calmness of the
hepatic capsule, there are slow proliferative events of cells from
different zones of the lobule, aiming to keep liver cell numbers
and organ weight steady. The precise mechanisms controlling
these processes are not fully understood, but for operative
reasons, we need to accept the existence of a set of controls
comprising a “hepatostat”, a set of processes that ascertain
maintenance of liver weight to where it needs to be [1]. Such
demands for maintenance of steady weight exist in endocrine
glands controlled by the pituitary, and the feedbacks controlling
those homeostatic events are better understood. Other organs
(pancreas, kidneys, intestine, lungs, etc.) will increase slightly
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
in mass if a big portion of the organ (one kidney, one lung, part
of pancreas) is lost, but the residual organ tissue will not attain
the exact total weight that existed prior to the tissue loss. This is
not the case for liver! If a major portion of liver tissue is acutely
lost, liver will enter into a regenerative process so that the total
hepatostat‐driven LWBR is precisely reestablished in the status
quo ante. This is the process of liver regeneration (LR) that will
be majorly examined in this chapter.
In a clinical disease setting, regeneration is manifested in con-
ditions leading to severe loss of hepatocytes. Chronic loss of
hepatocytes is seen in infectious diseases (e.g. HCV and HBV
viruses), chronic toxic conditions (e.g. alcohol, metabolic dis-
eases, non‐alcoholic steatohepatitis (NASH), storage diseases,
hemochromatosis, alpha‐1 antitrypsin deficiency), ischemia‐
reperfusion injury (most often following liver transplantation),
or chronic immune attacks (autoimmune hepatitis). Chronic loss
of hepatocytes is accompanied by compensatory proliferation of
the surviving hepatocytes. As we will address below in this chap-
ter, this occurs in potentially genotoxic environments and the
chronic compensatory proliferation of the surviving hepatocytes
may lead to development of neoplasia. Acute loss of hepatocytes
is rarer, often caused by acute ingestion of toxins (e.g. attempted
suicide by acetaminophen), trauma, or a short course of an acute
hepatitis. In all these conditions, death of hepatocytes and com-
pensatory proliferation proceed in tandem with inflammatory
processes which aim to remove dead hepatocytes and provide
cytokines for tissue repair. Overall, the regenerative processes in
acute loss of hepatocytes operate very efficiently. Studies of
patients with Dubin–Johnson (pigment similar to melanin accu-
mulating in normal hepatocytes) have shown that approximately
90% of hepatocytes die in an acute hepatitis, but they are restored
in proper numbers by the end of the disease [2]. It is highly likely
that in a clinical setting, the inflammatory processes and the
 45:  Liver Regeneration 567
regenerative events cooperate synergistically. It is difficult,
however, to design experiments that would cleanly dissect the
regulatory signals regulating only the compensatory prolifera-
tion of hepatocytes, from the comingled inflammatory events.
Experimental models in rodents associated with acute adminis-
tration of a liver toxin (typically CCl4 or acetaminophen) are
more relevant to the ­clinical setting.
The study of signals exclusively associated with regulation of
hepatic regeneration, not affected by inflammatory processes,
has been carried mostly in rodent livers. Each rodent liver is
composed of several lobes, and each lobe is supported by its
own vasculature (arterial and portal vein) and separate branch of
the common bile duct. An easy surgical procedure, first
described by Higgins and Anderson in 1933 [3], allows clean
removal of the two larger lobes, comprising around two‐thirds
of the liver. The non‐resected small lobes remain intact, with no
damage or associated inflammatory response. In rats, this pro-
cedure allows removal of about 68% of the liver mass, whereas
in mice, due to the presence of a gallbladder, typically about
60% of liver mass can be easily removed. The procedure is
known as “2/3 partial hepatectomy” (PHx) and the results
obtained by using this procedure will be the basis for discussion
of regulatory signals associated with LR. At the end of the
regenerative process, the non‐resected lobes become larger and,
in aggregate, restore the hepatic mass that was present prior to
PHx. The regenerated liver has a different shape; it now has
fewer lobes (two or three, based on anatomic conventions), and
the resected lobes are not restored.
SIGNALS REGULATING LR
Signals related to LR have been identified by triggering prolifera-
tion in hepatocyte cultures, delay of regeneration associated with
signal ablation in genomically modified mice, or response of
intact liver in vivo by injecting specific signaling substances [4,
5]. Many of these signals appear not only in the regenerating liver,
but also in the peripheral blood. Table  45.1 shows a list of the
most studied mitogenic signals associated with LR. Of these sig-
nals, epidermal growth factor receptor (EGFR) and associated
ligands (epidermal growth factor [EGF], transforming growth
factor [TGFα], heparin binding epidermal growth factor
[HBEGF], amphiregulin) and hepatocyte growth factor (HGF)
and its receptor (MET) are capable of inducing proliferation of
hepatocytes in serum‐free cultures [6]. They also induce hepato-
cyte DNA synthesis when injected into the portal vein of normal
rodent livers. They should be viewed as “direct mitogens”. Others
show their relevance by delays in LR following elimination of the
signaling molecule or its cognate receptor. They are not directly
mitogenic for hepatocytes in culture or in vivo. They should be
viewed as “indirect mitogens”. Some more recently discovered
signals (e.g. Wnt and Hedgehog) have not been tested in vivo or
in vitro but they do have strong effects on LR and homeostasis. In
this chapter, we will discuss these signals with emphasis on their
effects on LR. It should be noted that LR always finds pathways
to bypass elimination of any single mitogenic signal, direct or
indirect, and proceeds to completion by following alternate path-
ways overcoming the signaling obstacle.
Table 45.1  Mitogenic signals associated with liver regeneration after
partial hepatectomy
Complete mitogens
1. Mitogenic in hepatocyte cultures in chemically defined (serum‐free)
media
2. Cause liver enlargement and hepatocyte DNA synthesis when injected
into whole animals:
Hepatocyte Growth Factor (HGF) and receptor (c‐Met)
Ligands of the EGF R (EGF, TGFα, HB‐EGF, amphiregulin)
Combined elimination of these two signaling receptors abolishes liver
regeneration
Auxiliary mitogens
1. Ablation of their signaling pathways causes delay but does not abolish
liver regeneration.
2. They are not mitogenic in hepatocyte cultures and when injected in
vivo do not cause hepatocyte DNA synthesis and liver enlargement
Norepinephrine and the α1 adrenergic receptor
TNF and TNFR1
IL6
Notch and Jagged (recombinant Jagged causes DNA synthesis in
hepatocyte cultures)
VEGF and receptors I and II
Bile acids
Serotonin
Complement proteins
Leptin
Insulin
PPAR gamma
Very important and yet to be defined
Elimination of these signaling molecules does not abolish liver
regeneration. Their effects by exogenous administration to intact animals
or hepatocyte cultures have not been fully tested
Hedgehog
Wnt/beta‐catenin
Signals controlling Hippo pathway and Yap
Hepatocyte growth factor (HGF) and its
receptor (c‐Met)
HGF was first isolated on the basis of its capacity to induce
DNA synthesis in primary cultures of hepatocytes [7].
Subsequent cloning and sequencing revealed an unusual struc-
ture, composed of four Kringle domains and a pseudo‐protease
domain [8, 9]. There is no active protease function in HGF how-
ever, despite its otherwise strong homology to plasminogen. It is
synthesized as a single polypeptide, cleaved to its active heter-
odimeric form in an Arg‐Val‐Val site, distal to the Kringle
domains (identical to plasminogen). Urokinase plasminogen
activator (uPA), free or bound to its receptor (uPAR) directly
activates HGF; tissue plasminogen activator (TPA) can also
activate HGF, but not as effectively [10]. HGF can also be acti-
vated by a soluble protein with high homology to Factor XII,
known as HGF activator (HGFA). Hepsin and matriptase have
also been shown to activate HGF. In normal liver, HGF is pro-
duced by hepatic stellate cells and it is deposited in large
amounts in the extracellular matrix (ECM), mostly in the peri-
portal region of the lobule, bound to glycosaminoglycans and
specific collagens [11]. HGF production by stellate cells is con-
trolled by the neurotrophin receptor p75NTR [12]. At later
stages of LR, it is also produced by sinusoidal endothelial cells
and bone marrow endothelial progenitor cells homing to the
liver [13, 14]. HGF is produced in most peripheral tissues by
mesenchymal cells; it is also produced by specific populations
of neurons and glial cells in both brain (hippocampus, frontal
568 THE LIVER:  SIGNALS REGULATING LR
and parietal cortex) and spinal cord [15] and also in the
­placental syncytiotrophoblast [16]. The HGF receptor, MET, is
the product of the gene c‐Met. It is also synthesized as a long
polypeptide and activated into its dimeric form at the Golgi
apparatus. It resides as a transmembrane protein in the plasma
membrane of most epithelial cells, selective groups of neurons
and glial cells, as well macrophages and endothelial cells.
Both hepatocytes and cholangiocytes express MET. Activation
of MET is typically the result of ligation with active HGF. This
results in formation of HGF–MET dimers, based on the dimeri-
zation domain present on HGF. The formation of the HGF–
MET dimers is associated with cross‐phosphorylation of
tyrosine amino acids at sites 1234 and 1235. This is followed
by  phosphorylation of tyrosine in position 1349, which then
becomes a docking site for attachment of other proteins (includ-
ing GAB1, GRB2), responsible for a multiplicity of downstream
signals, which eventually cause activation of PIPK3, AKT (pro-
tein kinase B), and mammalian target of rapamycin (mTOR),
rearrangement of cell polarity and enhancement of motility and/
or stimulation of cell proliferation [17]. The “signalosome” for-
mation around Gab1 is already assembled within 30 minutes
after PHx [18]. Hepatocyte MET appears always activated in the
liver, presumably due to availability of HGF in abundance in the
hepatic ECM [19]. Incomplete forms of HGF containing only
one or two Kringles (NK1, NK2) can also ligate MET, but since
they do not contain the dimerization domain of complete HGF,
dimers can only form and MET activation occurs only in the
presence of glycosaminoglycans.
Epidermal growth factor receptor (EGFR)
and associated ligands
EGFR is expressed in both hepatocytes and cholangiocytes. It is
a member of the ErbB family of receptors. ERB1 (EGFR) and
ERB3 are expressed in adult hepatocytes and cholangiocytes,
whereas fetal hepatocytes express the above as well as ErbB2
(Her2/Neu) [20]. ERB3 does not have a kinase domain and
ErbB4 is not expressed in liver. In contrast to HGF/MET, EGFR
and ERB3 can form heterodimers with themselves or with each
other without being ligated. There are multiple ligands of EGFR.
The ones studied in liver and in the context of LR, primarily
include EGF, TGFα, amphiregulin, and HB‐EGF. EGF is pro-
duced in secretions of exocrine glands, including salivary
glands. Brunner’s glands, with histology similar to salivary
glands, exist in the submucosa of the duodenum. They secrete
EGF in the intestinal lumen. A portion of the secreted EGF is
absorbed intact from the lumen of the duodenum and trans-
ported to the liver via the portal circulation [21]. Thus, hepato-
cytes are constantly exposed to EGF, and EGFR is found
activated in normal liver [19]. EGF secretion from Brunner’s
glands is enhanced by norepinephrine [22]. TGFα is minimally
expressed in resting normal liver but its expression is maximally
enhanced following PHx [23]. It is produced as a transmem-
brane protein and the mature form is the extracellular domain,
released by proteolytic action by TACE/ADAM17 protease.
Germline elimination of TGFα does not affect liver generation,
presumably due to complementary effects by the other EGFR
ligands involved in the process. Amphiregulin, a downstream
target of Yap protein, is also minimally expressed in normal
liver, but its expression in hepatocytes rapidly increases after
PHx. Germline elimination of amphiregulin delays liver genera-
tion [24]. Heparin binding EGF (HB‐EGF) is produced by mac-
rophages and endothelial cells but not in hepatocytes [25]. It is
also produced as a transmembrane protein, and the extracellular
domain is released by the action of metalloproteinases as the
mature form of HB‐EGF. It rises rapidly after PHx, prior to the
appearance of TGFα. Germline elimination of HB‐EGF delays
LR and the progression of the steps through the cell cycle [26].
In the normal liver, EGF is the major interacting ligand with
EGFR. TGFα, amphiregulin, and HB‐EGF are rapidly mobi-
lized after PHx and they have complementary functions, with
neither one individually being critical for regeneration, but hav-
ing effects related both to hepatocyte proliferation as well as to
proliferation of cholangiocytes and endothelial cells, all of
which abundantly express EGFR.
Fibroblast growth factors (FGF)
and their receptors
FGF is a large family of signaling proteins, composed of 23 mem-
bers with diverse structures and cells or tissues of origin. Their
common feature is their capacity to activate some or all of four
receptors (FGFR1–FGR4). Mature hepatocytes express only
FGFR4 [27], with the other receptors present in non‐epithelial
hepatic cells. FGF1/2 are expressed by hepatocytes during LR
[5]. They are slightly mitogenic in hepatocyte cultures (less than
20% of the effect of HGF or EGF). It is not clear whether, during
LR, FGF1/2 are exerting autocrine or paracrine effects, especially
on endothelial cells expressing FGFR1/2. In hepatocytes, FGFR4
is co‐expressed and functioning in conjunction with βKlotho.
Elimination of either βKlotho or FGFR4 s had no effect on hepat-
ocyte proliferation but it disrupted hepatic cholesterol, bile acid,
and lipid metabolism [28]. In another study, however, knockdown
of FGFR4 was associated with increased necrosis of non‐paren-
chymal cells and 25% mortality after PHx; liver weight was
restored by hepatocyte hypertrophy [29]. The role of FGFR4 is
crucial in hepatocyte biology. It is the receptor ligated by FGF15
(human FGF19), produced in the intestine following stimulation
by bile acids binding to the intestinal farnesoid X receptor (FXR)
[30]. Functioning as an endocrine signal, FGF15 enters through
the portal circulation and regulates cholesterol metabolism and
suppresses hepatocyte bile acid synthesis via FGFR4, by down-
regulating CYP7a1, the rate limiting enzyme for bile acid synthe-
sis. Elimination of FGF15 caused massive increase of hepatic bile
acids and increased mortality after PHx with delayed LR [31].
FGF21 is another endocrine signal, often viewed as a “hepa-
tokine”, produced by hepatocytes. Its synthesis is controlled by
PPAR’α and PGC‐1a. It is also produced in testes, pancreas, and
adipose tissue. It has insulin‐mimetic effects and exerts general
metabolic effects in a variety of tissues, including adipose tissue
[32]. Stellate cells both produce and respond to FGF; evidence
suggests that FGF regulate synthesis of ECM components by stel-
late cells [33]. All evidence suggests that FGF and their receptors
are crucial for regulation of hepatic function, proliferation of non‐
parenchymal cells, and regulation of key metabolic functions in
normal and regenerating liver.
 45:  Liver Regeneration 569
Transforming growth factor beta (TGFβ)
There are three different forms of TGFβ, all present in the liver.
TGFβ receptor is the same for all three forms, composed of
three components (TGFβr I, II, and III), present in all hepatic
cell types. TGFβR II is the site in which TGFβ binds directly.
In liver, TGFβ1 has been the most studied and it is the one to be
referred to in this segment. It is produced by stellate cells and
Kupffer cell macrophages. In normal liver, TGFβ is bound to its
receptor and also to decorin, a GPI‐linked protein on plasma
membrane. Following PHx, TGFβ is gradually removed from
hepatocytes as regeneration proceeds from periportal to peri-
central areas of the lobule [34]. It is also massively increasing
in the plasma, probably released from decorin as a result of the
early remodeling of ECM following PHx (see below, section on
Urokinase plasminogen activator (uPA) and ECM remodeling)
[4]. Circulating alpha‐2 macroglobulin binds TGFβ and trans-
ports it to hepatocytes where it is inactivated [4, 5, 35]. TGFβ is
mitoinhibitory in hepatocyte cultures, and its removal from
liver soon after PHx would be a rational expectation, in order to
allow hepatocytes to proliferate. Administration of exogenous
TGFβ immediately after PHx delays initiation of hepatocyte
DNA synthesis. Surprisingly, however, TGFβ expression
increases at about 3 hours after PHx and reaches a plateau at 72
hours, at the same time as hepatocytes are at their maximal
state of proliferation [36]. The successful completion of hepat-
ocyte proliferation as TGFβ levels rise, is accomplished by the
simultaneous decrease in expression of hepatocyte TGFβ
receptors at the time of hepatocyte proliferation [37]. The con-
temporaneous rise in plasma norepinephrine likely also con-
tributes to the hepatocyte “resistance” to TGFβ; as evidenced
by studies in hepatocyte cultures, norepinephrine attenuates the
effect of TGFβ and suppresses its mitoinhibitory activity [38].
Studies in hepatic organoid cultures show that the increase in
TGFβ expression after PHx is linked to activation of EGFR and
MET [39]. TGFβ plays an important role at the later stages of
regeneration, though not by affecting termination of liver
regeneration (see below, section on Termination of liver regen-
eration). TGFβ is important in angiogenesis, inducing forma-
tion of tubules by endothelial cells [35]. It also likely stimulates
production of ECM by stellate cells, a phenomenon that occurs
at the later stages of LR, restoring ECM after its remodeling
immediately following PHx [40]. All evidence suggests that
TGFβ expression during LR is a constructive event and impor-
tant for the completion of LR by restoration of intact tissue
histology. This is also bolstered by the finding that loss of
beta‐2 spectrin, a component of the TGFβ signaling pathway, is
associated with suppressed and delayed liver regeneration [41].
Also of interest, is a previous study in which normal rats were
injected with a dominant negative construct against TGFβ
receptor II. This, unexpectedly, resulted in initiation of DNA
synthesis in the normal liver. The results suggest that under
normal conditions, hepatocytes are under opposite “tonic”
influences between ambient mitogens HGF, EGF, and so on,
and the mitoinhibitory effect of TGFβ, with the combined
effect resulting in maintaining hepatocytes in G0 state and in
a stable state of differentiation; acute removal of TGFβ signal-
ing creates an imbalance driving hepatocytes toward DNA
­synthesis [42].
Interleukin 6 (IL6)
IL6 has been studied extensively in hepatocytes, as the major
ligand triggering the “acute phase response”, a rapid increase
and secretion of many proteins associated with innate immunity.
Circulating IL6 binds to a soluble receptor, and the complex
binds to a hepatocyte receptor known as gp130, which is shared
with oncostatin M, LIF, CNTF, and so on. This trimeric com-
plex dimerizes to form a hexameric complex and is subsequently
phosphorylated in Tyr residues, becoming a docking site for
activation of JAK tyrosine kinases and STAT transcription fac-
tors. In the context of LR, IL6 is the major regulator of activa-
tion of STAT3 transcription factor. Germline elimination of IL6
is associated with delayed regeneration due to delayed STAT3
activation. LR however proceeds and completes normally, as
other signals (MET, EGFR) can also phosphorylate and activate
STAT3 [43–45]. IL6 is produced by Kupffer cells, but also by
hepatocytes [46]. It rises in plasma after PHx, following the rise
in circulating tumor necrosis factor (TNF), a major stimulus for
IL6 secretion [47].
Tumor necrosis factor alpha (TNF)
TNF is produced primarily by macrophages and binds to two
receptors (TNFR1, TNFR2). Direct administration of TNF to
normal animals leads to liver damage. TNF, however, rapidly
rises in the plasma after PHx, with no apparent damage to the
liver [48]. Studies by Fausto et al. have shown that germline
elimination of either TNFR1 or TNFR2 leads to delayed LR
and decreased production of IL6 [47]. Perhaps the most
important function of TNF in LR is the activation of the tran-
scription factor NFκB, which is delayed in TNFR knockout
(KO) mice [49]. TNF also regulates expression of TACE/
ADAMS17, the plasma membrane protease associated with
secretion of the mature (extracellular) form of TGF [50]. It is
of interest that either TNFR1 or FAS activation can induce
complete liver failure, independent of each other, and yet,
similar to TNFR, germ line elimination of FAS also leads to
delayed LR [51].
Norepinephrine (NE)
This neurotransmitter is produced at the synaptic endings and,
to a lesser extent, by adrenal medulla. Stellate cells also produce
and respond to NE [52]. Addition of NE in serum‐free hepato-
cyte cultures dramatically enhances the mitogenic effects of
EGF and HGF [53]. In addition, it downregulates the mitoin-
hibitory effects of TGFβ [38]. In “balanced” concentrations of
mitogen EGF and mito‐inhibitor TGFβ, addition of NE leads to
hepatocyte DNA synthesis [38]. NE acts by binding to the
G‐protein coupled alpha‐1 adrenergic receptor (A1AR). Its
effects on hepatocyte proliferation are mediated, at least in part,
by A1AR inducing phosphorylation of EGFR and activation of
STAT3 via Src kinase [54]. NE rises rapidly in plasma after PHx
and exerts the above effects directly on hepatocytes at the time
when both EGFR and MET are becoming increasingly activated
[55]. NE, however, has effects on LR from extrahepatic sites.
NE increases availability of both EGF and HGF to the regener-
ating liver by directly enhancing secretion of EGF from
570 THE LIVER:  SIGNALS REGULATING LR
Brunner’s duodenal glands [22] and production of HGF by mes-
enchymal cells [56]. Given the high levels of NE in plasma after
PHx, it may be the mediating factor for the observed increased
in HGF production in extrahepatic sites (lungs, kidney) after
PHx [57]. Administration of prazosin, a specific A1AR inhibi-
tor, has long‐lasting (three days) effects of suppression of hepat-
ocyte DNA synthesis after PHx. Similar effects were seen after
surgical sympathectomy of the liver prior to PHx [55].
Bile acids
Bile acids increase in circulating blood after PHx and depletion
of bile acids leads to decreased regeneration [58]. The rise of
bile acids in plasma, however, occurs several hours after PHx,
suggesting that bile acids act after LR has already been initiated.
Bile acids bind to the transcription factor FXR, resulting in a
feedback suppression of bile acid synthesis by downregulation
of CYP7a1, the first enzyme involved in bile acid synthesis.
Mice with genetic elimination of FXR have defective regenera-
tion [59]. Despite the evidence of suppressed regeneration fol-
lowing bile acid depletion or germline elimination of FXR, the
mediating pathways are not clear. It was originally thought that
the effects of bile acids on LR were mediated by the production
of FGF15 in the intestine (mediated by bile acids binding to
intestinal FXR). As mentioned above, FGF15 acts as an endo-
crine FGF and regulates many aspects of lipid metabolism,
including synthesis of bile acids, via receptor FGR4. Germline
elimination of FGF15, however, has minimal effects on LR.
FXR elimination and its effects on LR are probably mediated by
the uncontrolled biosynthesis of bile acids observed in the FXR
KO mice, associated with cholestatic damage of hepatic tissue.
In FXR KO mice, bile acid synthesis is not inhibited neither via
FXR, nor by FGF15 acting on FGFR4, because synthesis of
FGF15 is also suppressed in FXR KO mice. There is recent evi-
dence that bile acids may play a direct role in cholangiocyte
proliferation via the G‐protein coupled receptor TGR5 [60].
Serotonin
Thrombocytopenic mice have deficient LR and this is partially
corrected by administration of serotonin [61]. Administration
of serotonin to normal liver or hepatocytes in culture does not
have direct mitogenic effects on hepatocytes. Mice with low
levels of serotonin (deficient in tyrosine hydroxylase) also have
deficient LR [61]. On the other hand, mice with germline elimi-
nation of the serotonin transporter resulting in severe decrease
of serotonin in platelets and peripheral blood, do not have defi-
cient LR [62]. Serotonin has multiple beneficial effects on liver
in clinical studies and stimulation of VEGF production may
play a role [63].
Wnt/beta‐catenin
The family of Wnt proteins has emerged as a critical signaling
pathway for tissue repair and regeneration in most tissues. They
operate through the Frizzled family of receptors, with LRP5/6
acting as coreceptors. There are multiple members to the Wnt
family and they are expressed in different proportions from all
hepatic cell types. Beta‐catenin is the signaling vehicle for Wnt
proteins. (For more information on this important group of
hepatocyte signaling molecules, please see Chapter 46.)
Hedgehog (Hh) signaling
Emerging literature has demonstrated multiple roles of the Hh
signaling pathway in liver pathobiology. In relation to LR, all
known hepatic cell types produce Hh and respond to Hh signal-
ing. Inhibition of Hh signaling by cyclopamine delayed liver
regeneration by 48 hours. There was suppression of hepatocyte
and cholangiocyte proliferation and decreased activation of stel-
late cells [64]. Glypican 3 (GPC3), a GPI‐linked protein on
hepatocyte plasma membrane binding to tetraspanin CD81,
binds and retains Hh proteins in normal liver. Following PHx,
GPC3 ceases binding to CD81 and releases Hh proteins, while
Gli1, a transcription factor controlled by Hh, appears promi-
nently in hepatocyte nuclei at day 2 after PHx [65]. Hh signaling
regulates expression of Yap in stellate cells, and Yap itself is
involved in stellate cell activation and production of ECM [66].
Elimination of Smoothened, the cytoplasmic signaling mediator
of Hh, in stellate cells also dramatically decreased expression of
Yap in hepatocytes after PHx [67]. These results demonstrate
that much is yet to be discovered on the regulation of LR by this
very complex system of important signaling regulators.
Insulin
Insulin can best be described as an enabler of all hepatocyte
functions, including proliferation and differentiation. Produced
by the beta cells of pancreatic islets, it is supplied directly to the
liver via portal vein, before it goes to any other organ in the
body. In the absence of insulin, mitogenic effects of HGF and
EGF in hepatocyte cultures are abolished and the viability of
hepatocytes is severely curtailed [6]. Diversion of the portal cir-
culation directly to the vena cava (portacaval shunt) causes liver
atrophy. Administration of insulin in animals with portacaval
shunt restores liver weight, and this process is mediated by pro-
liferation of hepatocytes [68]. Insulin, however, is not by itself
mitogenic in hepatocyte cultures [53] and does not cause hepat-
ocyte proliferation when administered to normal animals.
Growth hormone (GH)
There are isolated studies on effects of GH in liver regeneration
but there is no clear evidence that GH is a major regulator. It
controls production of “somatomedins” insulin‐like growth fac-
tor 1 (IGF1) and IGF2 by hepatocytes, which play important
roles in liver development. GH has been linked to proper timing
of hepatocyte DNA synthesis and to activation of EGFR [69]
and enhances liver regeneration in aged animals [70].
Purinergic signals and NK cells
There is evidence that ATP is rapidly released immediately after
PHx due to increase in portal vein pressure, derived from hepato-
cyte and macrophage lysosomes. The phenomenon is of short
duration. The released ATP is rapidly hydrolyzed by ecto‐ATPases
present on plasma membrane of most cells. Adenosine, the final
derivative of ATP hydrolysis, can interact with purinergic
 45:  Liver Regeneration 571
receptors. Blockade of P2Y2 purinergic receptor by a specific
inhibitor delayed hepatocyte proliferation [71]. Other studies
showed that the stimulation of purinergic receptors primarily
affects hepatic natural killer (NK) cells and interference with NK
cell function also delays liver regeneration [72]. Hepatocytes
express two main types and multiple subtypes of purinergic
receptors, and their stimulation regulates multiple signaling path-
ways associated with hepatocyte proliferation, including phos-
phorylation of ERK1 [71].
Immediate biochemical or physical signals
associated with initiation of liver regeneration
There are two immediate consequences of PHx:
1. Immediate increase of portal flow to the remaining liver
lobes (approximately one‐third of the original liver mass).
2. Acute changes in peripheral blood and tissue constituents
that inform on the required, “hepatostat” driven, liver to body
weight ratio.
The identity of the ultimate early signals that trigger LR must be
sought in these two fundamental and rapid changes. The early
observation by Moolten and Bucher that PHx in one member of
a pair of parabiotic rats stimulated liver regeneration in the
unoperated member of the pair [73], gave rise to all subsequent
studies that resulted in isolation of HGF and documentation of
PHx‐related changes in plasma concentrations of IL6, TNF, NE,
bile acids, and so on.
In addition to the effects of specific molecular/biochemical
signals, the very physical forces generated by the increased por-
tal flow through the remnant liver may also play a role.
Preventing rise in portal vein pressure after PHx results in
reduced early events of LR and deficient activation of HGF [74].
Fluid shear stress due to acute increase in flow over endothelial
monolayers is associated with production of specific signals.
Stability of urokinase plasminogen activator (uPA) mRNA and
increase in the amount of uPA protein occurs by fluid shear
stress in endothelial monolayers [75]. uPA activity increases as
the first, so far, detected signal, within 1 minute after PHx [10].
uPA, however, is produced by both endothelial cells and hepato-
cytes and both cells may be the source of very early rise of uPA
[76]. Wnt proteins are also released by endothelial cells, and this
may relate to the appearance of beta‐catenin in hepatocyte
nuclei within 5 minutes after PHx [77]. There is no study as yet,
however, to assess the immediate impact of PHx on Wnt release
by sinusoidal endothelial cells.
Immediate and early LR‐related signals
following PHx (Table 45.2)
Urokinase plasminogen activator (uPA)
and ECM remodeling
uPA activity in liver increases within 1 minute after PHx [10]. In
addition to its recognized roles in initiating ECM remodeling
and activation of plasminogen to plasmin (leading to a cascade
of activation of metalloproteinases), uPA also activates HGF by
converting the ECM‐embedded, single peptide HGF to its heter-
odimeric form. There are no other agents that can be invoked for
Table 45.2  Activation of signaling pathways and changes occurring
early after partial hepatectomy
• Increase in urokinase activity (first 5 minutes)
• Translocation of N(otch)ICD to the nucleus (15 minutes)
• Translocation of beta‐catenin to the nucleus (5–10 minutes to 6 hours)
• Decrease in HGF biomatrix stores (30 minutes to 3 hours)
• Activation of the HGF receptor (within 30–60 minutes)
• Activation of the EGF receptor (within 30–60 minutes)
• Increase of HGF, norepinephrine, IL6, TNF, TGFb1, serotonin, and
hyaluronic acid in the plasma within 1–2 hours.
• Activation of AP1, NFkB, and STAT3 within 2–3 hours.
• Extensive gene expression reprogramming of hepatocytes within 30
minutes after PHx
• Remodeling of extracellular matrix within the first 2–3 hours
this event, since anti‐uPA antibodies inhibit activation of the
single peptide, pro‐ HGF in whole liver homogenates taken
within 5 minutes after PHx [78]. uPA is also involved with the
remodeling of ECM observed within the first 2 hours after PHx.
The sequential steps (conversion of plasminogen to plasmin,
subsequent degradation of fibrinogen, and activation of metal-
loproteinases MMP2 and MMP9 by plasmin) become noted
shortly after PHx, and are eventually downregulated by a subse-
quent elevation of TIMP1 by 6–18 hours [5, 35, 79]. There are
changes in multiple proteins of ECM (fibronectin prominently
decreasing in periportal area within 5 minutes after PHx) fol-
lowed by an increase in mRNA of several of these proteins,
probably aiming to resynthesize them as a compensatory mech-
anism [80]. Hyaluronic acid, also an ECM component, is rap-
idly released in the peripheral blood and TGFβ1, present in the
extracellular space bound to hepatocyte decorin, is also rapidly
released with the same kinetics [5, 35]. HGF, embedded in the
extracellular matrix as pro‐HGF form, is released to the plasma
as active HGF and rises very rapidly to more than tenfold
increase within 1 hour after PHx [81]. New HGF synthesis and
the appearance of new HGF mRNA starts at 3 hours after PHx
and reaches plateau at 24 hours. Increased expression of HGF
mRNA is also seen in lung and kidneys after PHx. The rise of
HGF in the plasma occurs 3 hours before any rise in HGF
mRNA and is a result of release of HGF stored in hepatic ECM
[4, 5, 35, 82, 83]. It should be noted that HGF is heavily embed-
ded in hepatic ECM; following whole‐body elimination of the
HGF gene, the hepatic ECM concentration of HGF did not
decrease until after two sequential episodes of LR triggered by
CCl4 [84]. The overall changes of ECM remodeling in LR are
quite complex with multiple factors involved, including plasmi-
nogen activator inhibitors (PAI), tissue inhibitors of metallopro-
teinases (TIMP), regulation of HGF activation, and so on. [85].
ECM proteins and associated glycosaminoglycans act as co‐
receptors for many LR associated ligands, and also directly
transmit not fully understood altered signals through integrins
[86–88]. This will be discussed in relation to signals controlling
termination of liver regeneration (see below, Termination of
liver regeneration).
Mobilization of Wnt/beta‐catenin and Notch signaling
There is increase in beta‐catenin in hepatocyte nuclei within 5
minutes after PHx and lasting beyond 6 hours [89]. This is pre-
sumably driven by Wnt derived from sinusoidal endothelial
cells, although that linkage in terms of chronology is not directly
572 THE LIVER:  SIGNALS REGULATING LR
established. Notch is expressed in hepatocytes, cholangiocytes,
and endothelial cells. Its ligand, Jagged‐1, is expressed in hepat-
ocytes and cholangiocytes. The active proteolytic fragment of
Notch, NICD, appears in hepatocyte nuclei within 15 minutes
after PHx, and is associated with subsequent expression of
Notch gene targets [90]. This event may be connected to the
immediate early production by stellate cells of high levels of the
non‐canonical Notch ligand DLK1 within minutes after PHx
[91]. Both Notch and its ligand Jagged remain upregulated until
at least day 4 after PHx. The Notch changes are clearly associ-
ated with hepatocyte proliferation; addition of Jagged‐1 protein
in hepatocyte cultures induced DNA synthesis and silencing
RNA against Notch or Jagged decreased hepatocyte prolifera-
tion until day 4 after PHx [90].
Changes of LR‐related signals in the peripheral blood
Earlier studies by Moolten and Bucher [73] employed parabiotic
rats and demonstrated that PHx in one of the members of the
parabiotic pair led to hepatocyte DNA synthesis not only in the
operated rat but also in the liver of the non‐operated one. Similar
findings were also noted in hepatocytes transplanted to adipose
tissue following PHx of the in situ normal liver [92]. These find-
ings clearly demonstrated the rise of regenerative signals in the
peripheral blood following PHx. Following several decades of
studies, we now know that many signals associated with LR rise
in the peripheral blood after PHx. HGF and NE rise within less
than an hour, with overlapping but subsequent rise of IL6, TNF,
and TGFβ, serotonin, and so on. (see above sections describing
changes related to the specific signaling molecules) [4, 5, 35, 82,
83]. An often‐overlooked consequence of PHx is the acute drop
of glucose concentration in the peripheral blood. The importance
of this for stimulation of LR was clearly demonstrated by admin-
istration of dextrose following PHx; this resulted in temporary
suppression of DNA synthesis at day 2 in mice. Surprisingly,
most of the changes in hepatocyte transcription factors, activa-
tion of HGF, and so on, were not affected; there was, however, a
rise in the mitoinhibitory signaling proteins p21 and p27 and
suppression of FoxM1, a major regulator of cyclin D1 and many
other proteins associated with hepatocyte DNA synthesis [93]. In
addition to specific blood signals, platelets attach to the sinusoi-
dal endothelial cells and become activated at the early stages of
LR [94]. Platelets contain a mixture of stimulatory (HGF, seroto-
nin) and mitoinhibitory (TGFβ) signals.
Cell proliferation kinetics
and intercellular signaling
Hepatocytes
LR is not dependent on stem cells. The vast majority of existing
hepatocytes participate in LR and undergo DNA synthesis. The
percent of hepatocyte participation in DNA synthesis and cell
proliferation varies with age. LR, however, is by no means
impaired in older mice. More than 95% of hepatocytes partici-
pate in LR in rats younger than twenty months of age and the
percent decreases only to 75% for older animals [95]. LR is also
remarkable in that it can be repeated multiple times (up to 12
times reported) in the same animal and the performance of
regeneration is not affected [96]. While there are profound
changes in gene expression of hepatocytes during LR, the gene
expression patterns associated with LR are different than those
operating during liver embryogenesis [97]. Following PHx,
hepatocytes are the first cells to exhibit rapid changes and
responses to mitogenic signals. As mentioned above, there is
rapid migration of beta‐catenin and Notch (NICD) to hepatocyte
nuclei within minutes after PHx [89, 90]. Taub et  al. have
reported rapid changes in hepatocyte gene expression, within 30
minutes [98]. In preparation for mitosis, the structure of canali-
culi becomes temporarily simplified. There are changes in gap
and tight junctions from 24–72 hours, regulated by p38‐MAP
kinase [99]. MET and EGFR are activated within 30 minutes
[100]. STAT3, regulated by IL6 and JAK kinases [101] and
NFκB, regulated by TNF [102] become activated in the first 3–5
hours. IL6 also regulates enhanced expression and action of
C/EBPβ and has reverse effects on C/EBPα [103]. N‐terminal
truncated isoforms of these factors (LIP and LAP, each other’s
antagonists) are generated during LR and regulate many of the
process of the mature C/EBP isoforms [104]. Hepatocyte
FoxM1b plays a key role for all events associated with progres-
sion through the S‐phase and mitosis, and its expression
increases at 24 hours, prior to the G1/S interphase. In mice defi-
cient in FoxM1b, there is delay of DNA synthesis associated
with increase in p21 [105, 106]. Genes associated with genera-
tion of IPS cells, including Oct3/4, Nanog, and Myc, increase
very significantly after PHx [107]. Critical signals affecting
hepatocyte entry into S‐phase are regulated by cyclins D1 and
D2. They both turn on expression of multiple genes associated
with cell proliferation. In addition, cyclin D1 regulates tran-
scription of enzymes in pathways of metabolism of carbohy-
drates, lipids, and amino acids [108]. Cyclin D1 expression is
regulated at the translational level by mir‐21 [109]. Cyclins D1
and D2 regulate activity of cyclin dependent kinases (CDK),
which mediate many of the cyclin D1/2 events. CDK activities
are inhibited by the p53‐regulated protein p21 [110]. The
expression of p21, however, rapidly increases within few hours
after PHx [111], suggesting a complex time‐dependent interplay
of regulatory events, guaranteeing precision of chronology in
the progression of the various steps associated with hepatocytes
in G1. The chronology of these events varies between species.
Peak of hepatocyte DNA synthesis is seen at 24 hours in the rat
but at 36–48 hours in the mouse. However, not all hepatocytes
enter into DNA synthesis at the same time. LR and DNA syn-
thesis proceed from the periportal to the pericentral areas of the
lobules, with time kinetics also varying between species [112].
Klochendler et al. were able to isolate cells from different stages
of the cell cycle and documented decreased expression of genes
characteristic of mature hepatocytes during DNA synthesis
[113]. This suppression is in part explained by the decreased
expression and suppression of action of HNF4α very early in
LR [114]. The waveform transition of hepatocyte replication is
thus guaranteeing that not all hepatocytes in the liver will
undergo decrease in essential gene expression patterns and pre-
serves sufficient homeostatic liver function during LR. A
marked evidence of altered gene expression patterns in hepato-
cytes is the accumulation of triglyceride droplets during days 2
and 3 of LR. This is dependent on signaling changes associated
with blood‐borne signals, as it can be induced in hepatocyte cul-
tures exposed to serum of partially hepatectomized rats [115].
 45:  Liver Regeneration 573
Lipid accumulation in hepatocytes in G1 is associated with
induction of lipogenic enzymes and regulated by leptin [116]
and by EGFR but not by MET [19]. Caveolin is present at the
interface between lipid droplets and cytoplasm and regulates
lipid processing and metabolomics of regeneration [117].
Altered gene expression patterns are also affected by expression
of multiple miRNA at different phases of the cycle, operating at
the translational level of gene expression [118]. Changes in
Hippo pathway kinases and Yap also occur during LR. The
Hippo pathway kinases MST1/2 and LATS1/2 phosphorylate
and deactivate nuclear Yap, whereas nuclear levels of Yap are
associated with regulation of liver size [119]. Yap controls
expression of amphiregulin, an EGFR ligand involved in LR
regulation (see above section on Epidermal growth factor recep-
tor). In normal liver, Yap is expressed in cholangiocytes, with
little or no detectable expression in hepatocytes [120]. Yap
increases in hepatocyte nuclei within 24 hours, associated with
decreased activity of the Hippo kinases. Levels of Yap and
enhanced activity of Hippo pathway return to the lower normal
by day 7 [119]. Through currently unexplained pathways,
nuclear Yap in hepatocytes during LR is regulated by Hedgehog‐
mediated actions on stellate cells [67]. Autophagy regulation is
also an important part of LR. Elimination of Atg5, an important
component of the autophagic pathway, has severe suppressive
effects on DNA synthesis and liver weight is restored mostly by
hepatocyte hypertrophy in ATG5 KO mice [121]. Hepatocytes
in the cell cycle produce and receive mitogenic signals from the
other hepatic cells types. TGFα, FGF1/2, angiopoietins 1 and 2,
VEGF, and GM‐CSF expressed and secreted by hepatocytes are
known to be mitogenic and probably have paracrine effects on
stellate cells, endothelial cells, and Kupffer macrophages, con-
tributing to formation of proper lobular histologic microarchi-
tecture [4, 5, 35, 82, 83]. Hepatocytes also receive newly
synthesized HGF from stellate and endothelial cells throughout
LR [13]. The net result is precise alignment of newly synthe-
sized hepatocytes along the orientation of the closest sinusoid in
a very precise manner [122]. Though the above events related to
hepatocyte proliferation during LR, several studies suggest that
even after 2/3 PHx, and more definitively after hepatectomies
less than 2/3, the pathways involving restoration of liver weight
involve not only hepatocyte proliferation but also hepatocyte
hypertrophy, with hepatocyte enlargement being a solid compo-
nent of the overall LR strategy [123, 124]. The signaling path-
ways controlling regulation of the precise percent of involvement
of hepatocyte proliferation versus mere hypertrophy are not
understood at this point.
Cholangiocytes
Proliferation of cholangiocytes in portal ductules follows the
same time frame as hepatocytes. Cells of the biliary system
respond to the same tyrosine receptor kinases (MET and EGFR)
as hepatocytes. LR proceeds by enlargement of existing lobules;
there is no formation of new portal triads following PHx. The
receptor TGR5, a G‐protein couple receptor, also regulates pro-
liferation of cholangiocytes in culture in response to bile acids,
with some of the bile acids having stimulatory or inhibitory
effects [60]. TGR5 has complex effects during LR, protecting
hepatocytes from increase in biliary flux from the intestine [125].
PHx in mice with germline deletion of TGR5 is associated with
severe jaundice, increased mortality, and hepatocyte necrosis.
Melatonin has overall inhibitory effects on cholangiocyte prolif-
eration [126] whereas histamine released by mast cells has stim-
ulatory effects [127]. While these effects are well documented,
there is no detailed analysis of their impact on cholangiocyte
proliferation during LR. Cholangiocytes also produce platelet
derived growth factor (PDGF), which is mitogenic for stellate
cells [128]. It should be noted that even though then term “chol-
angiocytes” is used to characterize the entire biliary epithelial
compartment, there is evidence from multiple sources that there
are many subtypes, characterized by size and/or expression of
specific receptors and the overall regenerative biology of these
subtypes may differ [129].
Endothelial cells
Sinusoidal endothelial cells enter into proliferation later than
hepatocytes and cholangiocytes, with active DNA synthesis
peaking at days 4–7 after PHx. Their proliferation is associated
with activation of VEGF receptors 1, 2, angiopoietin receptors
Tie 1, 2, PDGFRβ, EGFR, and MET [130]. Since hepatocytes
are the first to enter into proliferation, they form avascular small
clusters and start synthesizing VEGF, which attracts and stimu-
lates proliferation of endothelial cells penetrating the clusters
and establishing formation of sinusoids [131]. This activity is
associated with stimulation of VEGFR2. Simultaneous stimula-
tion of endothelial VEGF receptor 1, however, induces produc-
tion of HGF by the endothelial cells, which contributes toward
stimulation of hepatocyte proliferation [13]. The endothelial
cells in the newly formed capillaries gradually transform into
fenestrated endothelial cells through a complex process [132].
Recent studies have shown that LR is also associated with
migration of endothelial precursors from bone marrow to the
liver, which convert from typical endothelial cells into fenes-
trated endothelial cells and participate in the restructuring of the
hepatic sinusoids in the enlarged lobules that result from LR
[14]. The migrating sinusoidal progenitor cells (“sprocs”)
actively produce HGF. Their migration is controlled by hepatic
production of circulating VEGF and mediated via stromal cell
derived factor 1 (SDF1) [133]. HGF and TGFβ1 are known to
stimulate tubule formation in endothelial cells and are very
likely play a role in the restructuring of the sinusoidal network
[4, 5, 35]. Despite complex pathways of cell migration, endothe-
lial cells regulate LR in many ways by producing cytokines
favoring regenerative response. These “angiocrine” effects were
described recently by Ding et al. [134].
Stellate cells
These cells exist in multiple organs of endodermal origin (lungs,
pancreas, etc.). Hepatic stellate cells store vitamin A in lipid
droplets, and they express many genes in common with glial
cells of the brain, even though stellate cells are derived from
nestin‐positive cells of the cardiac mesenchyme and not the neu-
ral crest. There is strong evidence that stellate cells function as
part of a neuroendocrine control of normal liver and respond to
parasympathetic and sympathetic innervation [135]. They also
function as antigen presenting cells for NK cells [136]. Most of
the studies related to stellate cells focus on pathways resulting
574 THE LIVER:  SIGNALS REGULATING LR
in stellate cell “activation” during chronic liver injury, and their
contribution to liver fibrosis and cirrhosis [137]. Their participa-
tion in LR after PHx has also focused mostly on their contribu-
tion of signaling molecules (including HGF, TGFβ, epimorphin,
pleiotropin, etc.) and the production of most of ECM proteins in
different stages of LR. Despite their vital role in normal and
regenerating liver, there is no detailed study of their prolifera-
tion rates during LR. In normal liver, stellate cells make contact
with both hepatocytes and sinusoidal endothelial cells by very
long processes (Figure  45.1). This relationship suggests that
there are regulatory pathways operating by direct communica-
tion between stellate cells and the cell they contact; these inter-
actions need to be explored. Histologic observations demonstrate
that stellate cell numbers increase in parallel with the restoration
of the sinusoidal network. Stellate cell numbers peak in
Figure 45.1  Intricate connections between stellate cells (SC), sinu-
soidal endothelial cells (SEC) and hepatocytes (Hep). A single stellate
cell (green) is making contacts with multiple other cells. The nature
and purpose of these contacts is not fully understood. They are, how-
ever, likely to be involved in bilateral communications regulating
extracellular matrix deposition and exchange of growth‐regulatory
signaling molecules. Image courtesy of Dr. Donna Stolz, University of
Pittsburgh.
regenerated lobules at day 7 of LR, in parallel with the sinusoi-
dal endothelial cells. Increase in desmin‐negative cells resem-
bling immature stellate cells is seen at day 1 of LR [138].
Kupffer cells (hepatic macrophages)
Several studies have documented that the Kupffer cell mac-
rophages lining the hepatic sinusoids, in addition to standard
macrophage functions (phagocytosis of particles, participation
in eliminating tissue debris after local damage, etc.) also have
some unique properties, related to aspects of immunity in which
liver is an active participator. During LR, there is evidence that
many Kupffer cells proliferate locally [139]. There is evidence,
however, that mononuclear cells deriving from bone marrow
also enter the liver and become typical Kupffer cells during LR
[140]. Kupffer cells are recognized as F4/80+ and there distin-
guished into two categories. The CD11b+ Kupffer cells appear
to be derived from bone marrow and their proportion increases
at day 3, after the peak of hepatocyte proliferation. The CD68+
Kupffer cells do not seem to change in numbers during LR
[141]. Proliferating hepatocytes produce GM‐CSF, a mitogenic
signal for Kupffer cells; stellate cells produce M‐CSF, and
Kupffer cells produce IL6, TNF, TGFβ, and TGFα during LR
[4, 5, 35]. Overall, there are no precise measurements of popu-
lation changes of Kupffer cells during LR but their numbers
vary along with their participation in restoration of sinusoidal
vascular network.
Termination of liver regeneration
Initiation of LR can be precisely timed by the performance of
PHx. Termination of LR, however, is more difficult to define. If
the liver weight to body ratio is to be used as a criterion, normal
ratios (depending on species) are approached by end of the sec-
ond week. In terms of hepatocyte proliferation, a return of nuclear
DNA synthesis to normal levels is approached by days 5–6 [4, 5,
35]. Analysis of liver transcriptomics, however shows a slower
return to normal gene expression values by some time beyond day
14 after PHx [88] (Figure 45.2). Several signaling molecules have
been considered as potentially associated with “termination” of
LR. Hepatocyte‐specific TGFβ1 transgenic mice have normal
liver regeneration, even though TGFβ is mito‐inhibitory for
hepatocytes [142]. Hepatocyte targeted elimination of TGFβ
receptors or activin receptors does not prolong liver regeneration;
extended liver regeneration is seen only when both receptors are
eliminated [143]. An important regulator associated with termina-
tion of liver regeneration and restoration of normal liver function
is likely to be ECM. Addition of ECM preparations (collagen
gels, Matrigel) stabilizes differentiation but inhibits proliferation
of hepatocytes in culture [6]. ECM is partially degraded and
remodeled at the beginning of LR, but it is restored toward the
end of LR [40, 144, 145]. Signaling between ECM and hepato-
cytes is mediated by integrin α3β1; the β1 intracellular domain
makes contact with integrin linked kinase (ILK). The latter trans-
mits growth suppressor and differentiation enhancement signals
through multiple pathways [87]. Hepatocyte‐specific elimination
of ILK prolongs liver regeneration and results in livers of size
158% of the ­original. All protein partners associated with ILK
increase during LR until the end of hepatocyte proliferation, by
 45:  Liver Regeneration 575

Figure 45.2  Expression of the top 250 genes expressed in normal mouse liver at different times after 2/3 partial hepatectomy (PHx). There is rapid
up and down change in expression of most of the 250 genes within 1 day after PHx, with changes in both direction continuing until beyond day 14
of regeneration. This makes it difficult to determine a precise point at which the processes of LR should be considered as “terminated”. A few
specific genes are massively overexpressed during LR, exceeding expression values of any gene seen in normal resting liver. The reason for the
selective massive overexpression of a few genes is not clear. Mup1 is not present in humans. Selenoprotein P1 is the major vehicle by which liver
sends selenium to peripheral tissues, essential for regulation of redox functions in most cells of body. Transthyretin is the major transport vehicle
in plasma for thyroxin (T4) and retinol bound to the retinol binding protein. Cyp2c9 protein is approximately 18% of all proteins of members of the
liver Cyp family and is associated with metabolism of most xenobiotics and many endogenous metabolic compounds including arachidonic acid.
Jak3 gene encodes JAK3 kinase protein, associated with gp130 receptor of IL6 and activating STAT3 transcription factor. Psg28 is a glycoprotein
produced in liver during pregnancy. There is no known function for this protein in liver regeneration.

day 6 [88]. These results suggest that ECM signaling is important
for regulation of termination of liver regeneration, but the path-
ways remain to be further understood. Glypican 3 (GPC3), a GPI‐
linked plasma membrane protein over‐expressed in liver cancer
[146], is nonetheless a growth suppressor for most organs, as
demonstrated by the Simpson–Golabi–Bechmel syndrome in
humans with loss of function of GPC3 [147]. Hepatocyte‐specific
transgenic expression of GPC3 results in substantially suppressed
LR [148]. GPC3, normally interacting with tetraspanin CD81, is
involved during LR with many complex interactions with differ-
ent pathways, including Hedgehog, Wnt, and Hippo signaling.
GPC3 dissociates from CD81 during LR but associates with it
again at days 6–7 [65]. Recent evidence has shown that GPC3
bound to CD81 enhances activity of the Hippo pathway and
results in decreased levels of nuclear Yap. This may be a very
important regulating signal by which GPC3, via Yap, contributes
to the termination of LR [149]. LSP‐1 is another protein, deleted
in about 50% of HCC, and regulating the RAF‐MEK‐ERK path-
way. Deletion of hepatocyte LSP‐1 prolongs LR and transgenic
expression suppresses LR [150]. Beta‐catenin is an important
mitogenic signal contributing to LR; beta‐catenin signaling is
downregulated by Wnt5a whose expression also increases toward
the end of LR and restricts further beta‐catenin signaling [151].
Overcoming deficiencies of extracellular
signals and the critical role of receptor
tyrosine kinases
Elimination of single extracellular signals (IL6, TNF, Wnt/beta‐
catenin, Hedgehog, bile acids, etc.) involved in regulation of
LR, delays but does not abolish LR. This also applies to
elimination of single receptor tyrosine kinases (MET and
EGFR). The flexibility of cybernetics of hepatocyte transcrip-
tomics, which allows overcoming such obstacles, is remarkable.
Absence of many of the extracellular signals normally involved
to initiate hepatocyte DNA synthesis (e.g. IL6 and STAT3; TNF
and NFκB) can be corrected by MET and EGFR [45]. Systemic,
whole‐body, elimination of the HGF gene does not affect LR,
because the heavily embedded HGF in ECM is sufficient in
quantity to support several LR episodes prior to be depleted
[84]. The only extracellular signaling intervention that com-
pletely abolishes LR is the combined elimination of signaling of
both MET and EGFR [19]. When signaling of both of these
receptors is eliminated 5 days prior to PHx, there is a small
amount of hepatocyte proliferation at post‐PHx day 2; prolifera-
tion, however, stops entirely after that. This is followed by
decrease in hepatocyte size to 35% of normal and decrease in
the size of hepatic lobules. There is no net gain in liver size by
the end of day 14. There is significant decrease in hepatocyte‐
specific gene expression, especially for genes involved in lipid
and sterol biosynthesis, glycogen synthesis, and urea cycle.
There is no activation of mTOR and AKT, enhanced activation
of PTEN and AMPKα, and inactivation of enzymes related to
fatty acid synthesis and redox regulation. Despite these pro-
found changes, there is dramatic upregulation of expression of
genes involved in plasma protein synthesis and all members of
CYP family genes (with the exception of Cyp7a1, the first
enzyme involved in bile acid synthesis). Mice die by day 14
after PHx, with low glucose, marked ascites, and high ammonia
levels in peripheral blood. Remarkably, there is no increase
in  hepatocyte death, but there is increased apoptosis of non‐
parenchymal cells, especially stellate cells. Mice die not because
576 THE LIVER:  SIGNALS REGULATING LR

liver dies, but because liver ceases to function in the absence of
activation of EGFR and MET. The same phenomena, with some
variations in activation pathways, were observed and resulted
in death of normal, non‐hepatectomized mice [152]. In both
of those studies, MET and EGFR signaling was eliminated
throughout the body. Yet no abnormalities were detected in
other tissues. Proliferation of stem cells in intestinal crypts was
unaffected. Why is liver so uniquely dependent on activation of
EGFR and MET? (Figure 45.3). Both receptors are activated by
Tyr phosphorylation in normal resting liver [19]. Activation of
MET can be explained by the heavy presence of HGF in hepatic
ECM [84]. The story for EGFR is more complex. As mentioned
above, EGF is synthesized by the Brunner’s glands residing
under the duodenal mucosa [21]. EGF protein is directly
secreted in the intestinal lumen, and a portion is absorbed intact
and provided to hepatocytes continually through the portal cir-
culation [21]. The observations for constant supply of HGF and
EGF to hepatocytes suggest that the combined signaling of HGF
and EGF generates a fundamental cybernetic platform required
for the proper coordination of all other signaling pathways char-
acteristic of normal hepatocytes. Collapse of coordination of the
other signaling pathways occurs when these two receptor tyros-
ine kinases (RTK) are simultaneously abolished. The findings
may have implications for pathogenesis in human liver failure.
Plasma HGF in fulminant hepatitis rises to high levels, but is not
activated [153], in contrast to the active HGF released in plasma
after PHx [154]. There is also closure of the fenestrae of the
sinusoidal endothelial cells in fulminant hepatic failure, poten-
tially interfering with EGF availability [155].
Facultative stem cell relations between
hepatocytes and cholangiocytes
Hepatocytes and cholangiocytes are the two epithelial cell types
in the liver. They typically enter into proliferation to restore the
proper cell numbers in their own compartment, as required for
normal liver function. This is done without participation of stem
cells. There are situations, however, in which either hepatocytes
or cholangiocytes cannot enter into proliferation to repair defi-
ciencies in their own compartment. In such situations, hepato-
cytes and cholangiocytes follow steps of transdifferentiation to
transform themselves into each other [156]. This plasticity of

Figure 45.3  Dependence of hepatocytes on HGF/MET and EGFR signaling. Hepatocytes are continually exposed to HGF and EGF. HGF is
produced by stellate cells and deposited as inactive single peptide form in the extracellular matrix. EGF is produced by Brunner’s glands of the
duodenum, shown by black arrows in the histologic picture (hematoxylin eosin stain; magnification 100×). The histology of Brunner’s glands is
similar to salivary glands, which produce and secrete EGF in saliva. EGF is produced and secreted in the intestinal lumen, and then a portion is
absorbed intact and enters the liver through the portal circulation. HGF receptor MET and EGFR are activated in normal, resting liver. For details
and references, see text.
 45:  Liver Regeneration 577
phenotype confers regenerative advantages to the liver, espe-
cially in situations of liver failure. It is much faster to restore a
hepatocyte or cholangiocyte compartment by a rapid transdif-
ferentiation of one type of cell to another, than to rely on a slow
growth and expansion of a true stem cell compartment. There
are now numerous studies that have documented the capacity of
either cell type to enter into proliferation, transdifferentiate, and
restore the numbers of cells in the other compartment suffering
with defective proliferation [156]. This has been shown in sev-
eral experimental models in rats and mice. Blockade of rat liver
regeneration by acetylaminofluorene (AAF), resulting in
expression of DNA adducts and expression of p21 in hepato-
cytes [157], causes the appearance of a transitory population of
“oval” cells with morphologic properties intermediate between
hepatocytes and cholangiocytes [158]. These cells eventually
evolve into mature hepatocytes [159]. The emergence of the
oval cells from cholangiocytes in rats is supported by the expres-
sion of hepatocyte‐associated transcription factors in cholangio-
cytes immediately after AAF‐PHx [160] and the lack of
emergence of oval cells by pretreatment with the cholangiocyte‐
specific toxin DAPM (4,4’‐diaminodiphenylmethane) prior to
AAF‐PHx [161]. Mice cannot activate AAF, and the evidence
for cholangiocyte conversion to hepatocytes had to exploit more
difficult pathways that cause enhanced expression of p21 in
hepatocytes. This was achieved by Forbes et  al. in mice with
hepatocyte specific deletion of Mdm2, an E3 ubiquitin‐protein
ligase which specifically degrades p53. In the absence of
MDM2, there is enhanced expression of p53 which subse-
quently induces p21. There is increased hepatocyte death and
senescence followed by the appearance of “progenitor” cells
with properties similar to the oval cells seen in rats. The pro-
genitor cells converted into hepatocytes and restored hepatocyte
populations. The biliary origin of these progenitor cells was
established by detecting the newly emergent hepatocytes as car-
rying lineage tagging performed on cholangiocytes [162].
Extended studies have also documented the transformation of
hepatocytes to cholangiocytes in settings where cholangiocyte
proliferation is expected but cannot occur, as, for example, in
rats pretreated with DAPM and subjected to bile duct ligation
(BDL). Up to 50% of the newly produced biliary ductules in the
DAPM–BDL model carry markers specific for hepatocytes
[163]. Studies with hepatic organoids demonstrated that this
conversion is dependent on EGFR and MET [164]. Similar
results have been shown in mice. Conversion of hepatocytes to
cholangiocytes was shown in mice fed diethyldithiocarbamate
(DDC) diet [165]. The most recent evidence for hepatocyte
transdifferentiation to cholangiocytes was recently presented by
Huppert and Willenbring, who demonstrated a complete de
novo generation of a biliary system from hepatocytes by path-
ways controlled by TGFβ [166]. The pathways mediating this
interconversion are complex and, in addition to TGFβ, also
involve Hippo/Yap [120] and Notch [167]. It is not clear whether
all hepatocytes, anywhere in the lobule, can participate in these
interconversions. Demetris et  al. demonstrated cells with
mixed  expression of hepatocyte and cholangiocyte transcrip-
tion factors at the terminal end of the canals of Hering, suggest-
ing a population of pre‐existing cells in normal liver with
progenitor cell phenotypes [168]. Also, surprisingly, in LR sup-
pressed by AAF, there are more progenitor cells appearing after
centrilobular damage induced by CCL4 than after regeneration
following PHx. In AAF‐suppressed centrilobular damage, the
progenitor cells are likely to appear from the cholangioles that
penetrate deep into the hepatic lobule (canals of Hering) [169].
Lemaigre has traced the histology of formation of the ductal
plate during embryogenesis and demonstrated that most newly
formed ductal cells migrate into the space next the portal and
arterial branches to form the ductules of the portal triads. Cells
that fail to migrate, revert to hepatocytes and remain immedi-
ately proximal to the portal triad [170]. Studies in the rat dem-
onstrated that these hepatocytes preferentially contribute to
formation of biliary ductules [163]. In hepatic organoids in
roller bottle cultures maintained in the presence of insulin, HGF,
and EGF, epithelial cells retain a hepatobiliary phenotype with
mixed gene expression patterns. Addition of corticosteroids
induces the separation of the mixed lineage and the emergence
of mature biliary cells immediately under the culture medium
and mature hepatocytes in the tissue underlying the biliary cells.
Conversion of hepatocytes to biliary cells in that system cannot
occur in the absence of HGF and EGF (either one alone can
sustain the conversion) [39, 171]. In the same system, either
HGF or EGF induce TGFβ expression. Given the above recent
findings by Schaub et al. [166] on the role of TGFβ involved in
this transdifferentiation, it is likely that not only HGF and EGF,
but also TGFβ, are involved in the conversion of hepatocytes to
cholangiocytes in the organoid cultures. It should be noted that
whereas cells of mixed hepatobiliary phenotype are rarely seen
in normal human liver, they become the predominant type in
acute fulminant hepatitis, regardless of etiology [172]. It is not
clear whether this is a pathway to healing of fulminant hepatitis
and restoration of normal structure, because the evolution of the
hepatobiliary cells has not been traced in the few cases of fulmi-
nant hepatitis that heal spontaneously.
Regeneration following xenobiotic‐induced
centrilobular liver injury
Hepatocytes in the central portion of hepatic lobules express the
(phase I) CYP family and (phase II) other enzymes involved
with metabolism and processing of xenobiotics. In most
instances, xenobiotics are appropriately processed for elimina-
tion through blood (kidneys) or bile (feces). In rare instances,
however, the processing of some xenobiotics results in genera-
tion of reactive electrophiles, capable of reacting with the nucle-
ophile residues of proteins and nucleic acids, resulting in death
of centrilobular hepatocytes. Experimental models to study this
injury and its repair pathways have most often utilized carbon
tetrachloride (CCl4) or acetaminophen as the damaging agents,
in rats and mice [173, 174]. The extent of centrilobular necrosis
and the lethality depend on the dose of the xenobiotic, and the
lethality of doses vary with the strain of the animals, other die-
tetic components, and so on, as these parameters affect gene
expression patterns in the centrilobular regions [175]. Overall,
this type of injury is more representative of the type of damage
seen in human livers, especially given the fact that inappropri-
ately high doses of acetaminophen are often used for suicide,
resulting in massive liver necrosis, salvaged, whenever possible,
with liver transplantation. The first evidence of this type of
injury is a hepatocyte necrosis in the centrilobular areas. This is
578 THE LIVER:  SIGNALS REGULATING LR
followed by invasion of monocytes, which proliferate very
actively in the site of damage and produce active macrophages.
The latter proceed to remove the area of necrosis. There is rapid
activation of EGFR within minutes after acetaminophen dose
[176]. Hepatocytes enter into the cell cycle (proliferating cell
nuclear antigen [PCNA]‐positive) by the end of day 1, primarily
around the area of necrosis. Most of the cells of the rest of the
lobule, however, become PCNA‐positive by the end of day 2.
Hepatocytes positive for Ki‐67, a biomarker for active DNA
synthesis, appear at day 2 and increase to cover the entire lob-
ule. HGF levels increase within the liver and in the peripheral
blood [81]. mRNA for TGFα and HGF increased in two peaks,
at 12 and 48 hours [177]. In addition to contributions of growth
factors and stimuli seen in LR, macrophages may also contrib-
ute mitogenic stimuli, as they are known to produce HGF,
TGFα, TGFβ, PDGF, IL6, and TNF. The chronology of expres-
sion of these signals from macrophages in the context of repair
of this type of injury has not been assessed. The fact that the first
PCNA‐positive hepatocytes at day 1 first emerge around the
area of necrosis suggests that mitogenic stimuli produced by
macrophages, actively infiltrating and proliferating in that area,
must play a very early, perhaps the first, role. The rate of removal
of the necrotic hepatocytes and the restoration of normal
­histology depends on the dose of the offending agent. Normal
histology, with moderate experimental doses, is usually seen
within one week. The mechanistic details of repair of this type
of injury, though clinically relevant, are less defined compared
to LR. This is due to the comingling of inflammatory events at
the earliest stages, making it difficult to distinguish signals
strictly related to cell proliferation versus inflammation. An
additional issue is the variation of the response based on the
dose of the offending agent [175]. It should be stated, however,
that all mitogenic signals identified in LR have also been
identified in this type of injury, though their sources and time‐
kinetics are not be identical.
Maintenance of steady mass in normal
resting livers
In all previous sections of this chapter we discussed the mito-
genic signals and complex responses triggered by massive loss
of liver parenchyma. The elicited hepatocyte proliferation is a
compensatory response, aimed to restore the full functional
capacity of liver. The loss of hepatocytes is the primary reason
for the emergence of signaling pathways driving the compensa-
tory response. In a normal resting rodent liver, however, at any
given moment, the percent of dead or proliferating hepatocytes
typically does not exceed 0.2% of the total number of cells.
Even though this would involve very small numbers of cells,
there must be pathways that trigger compensatory proliferation
at a micro‐scale, so that the total loss over time does not result
in decreased liver mass. The mechanisms underlying this “hepa-
tostat” are not well understood. It is reasonable to postulate that
the signals regulating this process may be different than the sig-
nals elicited by a large hepatocyte loss. As mentioned above,
based on constant availability of HGF (from ECM) and EGF
(from the Brunner glands of duodenum), MET and EGFR are
always activated in normal liver, though we do not know if this
applies to all or a fraction of hepatocytes. Early studies using
continuous DNA labeling of hepatocytes in normal liver claimed
a slow streaming of proliferating hepatocytes arising in peripor-
tal areas and descending to pericentral areas of the lobule [178].
Subsequent studies with pulse labeling of DNA, however,
showed that replicating hepatocytes could be found in all areas
of the hepatic lobule, not just in periportal, and that “streaming”
of hepatocytes was not the mechanism for maintaining liver size
[179]. There have been several major recent studies, that have
provided new information about the role of different hepatocyte
subpopulations in maintenance of the “hepatostat”. Studies by
Nusse et  al. documented a unique population of hepatocytes
immediately around the central venules. These cells express
Axin2, a protein which is part of the beta‐catenin ubiquitination
complex. These cells express Tbx3, a biomarker for fetal hepa-
toblasts, and are diploid. They are also positive for the beta‐
catenin dependent glutamine synthetase. Wang et  al. tagged
these cells and documented that over time they generate a line-
age of cells which expand around the central area and lead to
progeny covering up to 40% of the cells of the lobule [180]. The
results were impressive, though there is a concern that beta‐
catenin levels may be elevated in these cells. There is only one
of the Axin2 genes active in this model, and the results of poten-
tial Axin2 haplo‐insufficiency may cause elevated levels of
beta‐catenin, enhancing the proliferative rate of the affected
hepatocytes. From the other pole of the lobule, the periportal
area, Furuyama et al. provided evidence that Sox9+ lineage pre-
cursors in liver and pancreas generate progeny of cells that grad-
ually expand toward the other areas of the lobule, differentiating
into Sox− hepatocytes and pancreatic acinar cells [181]. Sox9 is
expressed in cholangiocytes, and in that study, it was thought
that cholangiocytes continually differentiate into hepatocytes
maintaining liver mass homeostasis. Other studies, however,
demonstrated that immediate periportal hepatocytes also express
low levels of Sox9 and they also proliferate to establish expanded
progeny migrating toward the midzonal and centrilobular areas,
especially after chronic liver injury [182]. These cells are the
same hepatocytes earlier identified as being responsible for gen-
erating biliary ductules in inhibited biliary repair [163], previ-
ously identified by Lemaigre et al. as remnant cells of the ductal
plate which failed to enter into the portal triad and returned to
hepatocyte differentiation [170]. Thus, the findings of Furuyama
et al. could be explained in a different way; the lineage described
by Furuyama did not originate from cholangiocytes, but from
Sox9+ immediate periportal hepatocytes. Given, however, the
earlier studies with pulse labeling which demonstrated that
hepatocytes can be tagged in DNA synthesis in all areas of the
lobule, it is logical to consider that both the centrilobular and
periportal hepatocyte lineages operate, proceeding in opposite
directions, and replenishing hepatocytes. In this scenario, DNA
pulse labeling would tend to label hepatocytes in all areas of the
lobule. Supporting evidence from a pan‐lobular hepatocyte par-
ticipation in the process of continual renewal came from another
study. Tchorz et al. demonstrated that liver zonation depends on
a Wnt/beta‐catenin gradient with intensity increasing from peri-
portal to centrilobular areas, and that the formation of this gradi-
ent is dependent on angiocrine signals mediated through
R‐spondin (RSPO) ligands and their receptors (Lgr4/5). Lgr4+
hepatocytes had enhanced proliferative capacities during LR
and they were located throughout the lobule [183]. Finally, in a
 45:  Liver Regeneration 579
very recent study by Artandi et al., lineage tagging of subset of
hepatocytes expressing high levels of telomerase demonstrated
that they exist in all zones of the lobule and they give rise to
tagged progeny that can over time cover the entire lobule [184].
It is very likely that all these pathways (periportal, pericentral,
pan‐lobular) operate at the same time to give progeny that main-
tains a sufficient population of hepatocytes to guarantee ade-
quate mass for liver function. What is not identified however, is
perhaps the most crucial function of the hepatostat. What are the
stimuli that regulate all these processes and guarantee that liver
size does not exceed the required size?
Chronic regeneration in chronic liver disease
Most chronic liver diseases are associated with attrition of
hepatocytes, to a variable extent. This occurs regardless of etiol-
ogy, be that viral, toxic (including alcohol), metabolic, steato-
hepatitis, and so on. Loss of hepatocytes triggers regenerative
responses. These responses can be documented by enhanced
immunohistochemistry for biomarkers associated with hepato-
cyte proliferation (PCNA, Ki67), as well as biomarkers associ-
ated with hepatocyte death (TUNEL assays, expression of
activated caspases). The precise signaling pathways triggering
the regenerative response are likely to be similar with those dis-
covered following acute injury, discussed above. But, there also
likely to be peculiarities related to the injurious agent, probably
affecting regenerative signaling in different ways. The supreme
“prerogative” of loss of hepatocytes is to cause compensatory
hepatocyte proliferation, in order to maintain the required liver
to body weight ratio, best thought of as a “hepatostat”. In con-
trast to any other organ in the body, a required and fixed amount
of liver tissue is needed for body homeostasis, and on that regu-
latory principle, loss of liver “mass” is not tolerated. This phe-
nomenon is unique to the liver; it is not the case with other
organs (with the exception of endocrine glands under control by
pituitary). Loss of half of pancreas, or one of the kidneys, is not
associated with regeneration to increase the remnant tissue to
the original size of the total pancreas, and the residual kidney
becomes slightly larger but not twice the weight [185]. As a
consequence of the expectations of the hepatostat, there is
enhanced proliferation of the surviving hepatocytes at any time
there is continuous hepatocyte loss. This chronic compensatory
proliferation of surviving hepatocytes to make up for the loss of
hepatocytes induced by the chronic disease has several adverse
consequences. Several studies have shown that continuous pro-
liferation in chronic liver disease is associated with decrease in
hepatocyte ploidy. Hepatocytes in both in humans and rodents
are typically polyploid. The average ploidy of human hepato-
cytes is 4n binucleate. In mice, ploidy levels reaching 8n or 16n
are commonly seen. Experimental studies in rodents have shown
that during chronic hepatocyte proliferation, the “ploidy con-
veyor” reverses course and most polyploid hepatocytes revert to
diploid status [186]. Several studies in human liver have docu-
mented increased diploidy levels in most chronic liver diseases,
including cirrhosis [187]. Polyploidy reversal to diploidy, in and
by itself, should not be expected to have negative consequences.
Unfortunately, however, a large percentage of human and rodent
polyploid hepatocytes are also aneuploid, randomly missing
single copies of chromosomes [188]. In a polyploidy status, the
functions of such missing alleles are provided by the presence
of several copies of the single missing chromosome, present in
the polyploid status. As hepatocytes revert to diploidy, however,
some of the diploid cells are likely to randomly have only single
copies of some chromosomes. For example, an 8n polyploid
hepatocyte missing a single copy of chromosome 13 would gen-
erate four diploid hepatocytes of which one would only have a
single copy of chromosome 13. This has implications for devel-
opment of liver cancer. In chronic liver disease, hepatocyte pro-
liferation takes place in an inflammatory environment with
associated genotoxic products, including reactive electrophiles,
lipid peroxides, O2− radicals, and so on. Any inflicted genotoxic
damage on the single copy chromosome at the diploid stage
would not be balanced by the genes of the missing chromosome.
In this setting, otherwise non‐dominant mutations in the exist-
ing single chromosome copy may function as dominant (driver)
mutations. This enhances the risk of developing neoplasia. In
view of this, it should not be a surprise that all types of chronic
liver disease are associated with increased risk for development
of hepatocellular carcinoma [1]. The risk for developing liver
cancer is also enhanced in situations in which proliferation of
hepatocytes is overall impaired. The classic experiment by Solt
et al. demonstrated that “resistant hepatocytes” against any par-
ticular mito‐inhibitory block will rapidly develop into nodules
of monoclonal hepatocyte growth when challenged with acute
liver injury (partial hepatectomy), and the resultant nodules will
allow restoration of the hepatostat (liver to body weight ratio)
[189]. These “resistant hepatocytes” are more exposed to the
risks of developing aneuploid diploidy due to their enhanced
rates of proliferation. The few “resistant cells” are the only ones
that can proliferate to meet hepatostat demands for the whole
liver. The phenotypes of liver cancer in chronic liver disease due
to errors of metabolism associated with chronic inflammation,
are a clear example of this situation. Most cancers developing in
hemochromatosis do not store excessive iron; cancers emerging
in glycogen storage diseases, lipid storage diseases, and so on,
in their majority do not store the abnormal metabolic product;
most cancers seen in patients with alpha‐1 antitrypsin deficiency
do not carry the characteristic PAS‐diastase positive globules
characteristic of the misfolded ATZ protein [1]. In all of these
situations, hepatocytes that for random reasons do not carry the
metabolic abnormality are the “resistant hepatocytes” of Solt
et al. In their effort to maintain the hepatostat, their proliferation
is faster than the vast majority of hepatocytes affected with the
metabolic disease. In a recent study, the same phenomenon
appears to apply to hepatocellular carcinomas associated with
HCV infection. HCV impairs hepatocyte proliferation by inter-
acting with tetraspanin CD81 and activating the Hippo pathway,
thus causing decrease in nuclear Yap [149]. Most hepatocellular
carcinomas do not express CD81 in plasma membrane and are
thus uninfectable by HCV [190]. This makes them resistant to
the mito‐inhibitory effects of Hippo activation and Yap decrease
affecting the normal hepatocytes infected by HCV. The “HCV
resistant” CD81‐negative hepatocytes would proliferate faster
to maintain the hepatostat, and compensate for the loss of hepat-
ocytes caused by HCV infection. In this situation, HCV func-
tions as a “promoter” for enhanced development of hepatocellular
carcinomas, with most of the latter being composed of “HCV
resistant” neoplastic hepatocytes.
580 THE LIVER:  REFERENCES
CONCLUDING REMARKS
Even though LR after partial hepatectomy is not representative
of types of liver injury most commonly seen in liver disease, it
has been a very useful model for understanding the fundamental
signaling mechanisms controlling liver regeneration. It is not
complicated by considerations of inflammatory phenomena
associated with acute hepatocyte necrosis. Much recent knowl-
edge has surfaced for homeostatic events related to liver mass
maintenance under normal conditions. The signaling pathways,
however, regulating the hepatostat are not yet well understood
[1]. We understand that nuclear Yap plays a role in regulating
liver size and that combined signaling of HGF/MET and EGF/
EGFR are essential for maintaining the hepatostat; combined
elimination of these RTK signals causes liver decompensation
and functional collapse. It should not be forgotten at the end
of this chapter, that hepatocytes have very high, almost forever,
proliferative capacity in liver recolonization models. In the
Fah−/− model, after ten sequential recolonization events in which
it was shown that both polyploid and diploid hepatocytes equally
participate, calculations demonstrated that 1 hepatocyte could,
in mathematic principle, generate 50 mouse livers [191]. In that
sense, hepatocytes are unique amongst the epithelial cells in
our  body, and much more needs to be learned not only about
what turns on their proliferation, but also what harnesses that
unlimited proliferative capacity and channels it to meaningful,
hepatostat‐defined, boundaries, so the liver continues to ­function
normally.
REFERENCES
1. Michalopoulos, G.K. Hepatostat: liver regeneration and normal liver tissue
maintenance. Hepatology, 2017;65(4):1384–92.
2. Ware, A.J., Eigenbrodt, E.H., Shorey, J., and Combes, B. Viral hepatitis compli-
cating the Dubin–Johnson syndrome. Gastroenterology, 1972;63(2):331–9.
3. Higgins, G.M. and Anderson, R.M. Experimental pathology of the liver, 1:
Restoration of the liver of the white rat following partial surgical removal.
Arch Pathol, 1931;12:186–202.
4. Michalopoulos, G.K. and DeFrances, M.C. Liver regeneration. Science,
1997;276(5309):60–6.
5. Michalopoulos, G.K. Liver regeneration. J Cell Physiol, 2007;213(2):286–300.
6. Block, G.D., Locker, J., Bowen, W.C. et  al. Population expansion, clonal
growth, and specific differentiation patterns in primary cultures of hepato-
cytes induced by HGF/SF, EGF and TGF alpha in a chemically defined
(HGM) medium. J Cell Biol, 1996;132(6):1133–49.
7. Zarnegar, R. and Michalopoulos, G. Purification and biological characteriza-
tion of human hepatopoietin A., a polypeptide growth factor for hepatocytes.
Cancer Res, 1989;49(12):3314–20.
8. Nakamura, T., Nishizawa, T., Hagiya, M. et  al. Molecular cloning and
expression of human hepatocyte growth factor. Nature, 1989;342(6248):
440–3.
9. Liu, Y., Michalopoulos, G.K., and Zarnegar, R. Molecular cloning and char-
acterization of cDNA encoding mouse hepatocyte growth factor. Biochim
Biophys Acta, 1993;1216(2):299–303.
10. Mars, W.M., Liu, M.L., Kitson, R.P., Goldfarb, R.H., Gabauer, M.K., and
Michalopoulos, G.K. Immediate early detection of urokinase receptor after
partial hepatectomy and its implications for initiation of liver regeneration.
Hepatology, 1995;21(6):1695–701.
11. Schuppan, D., Schmid, M., Somasundaram, R. et  al. Collagens in the
liver  extraular matrix bind hepatocyte growth factor. Gastroenterology,.
1998;114(1):139–52.
12. Passino, M.A., Adams, R.A., Sikorski, S.L., and Akassoglou, K. Regulation
of hepatic stellate cell differentiation by the neurotrophin receptor p75NTR.
Science, 2007;315(5820):1853–6.
13. LeCouter, J., Moritz, D.R., Li, B. et al. Angiogenesis‐independent endothe-
lial protection of liver: role of VEGFR‐1. Science, 2003;299(5608):890–3.
14. Wang, L., Wang, X., Xie, G., Wang, L., Hill, C.K., and DeLeve, L.D. Liver
sinusoidal endothelial cell progenitor cells promote liver regeneration in rats.
J Clin Invest, 2012;122(4):1567–73.
15. Achim, C.L., Katyal, S., Wiley, C.A. et al. Expression of HGF and cMet in
the developing and adult brain. Brain Res Dev Brain Res 1997;102(2):
299–303.
16. Wolf, H.K., Zarnegar, R., Oliver, L., and Michalopoulos, G.K. Hepatocyte
growth factor in human placenta and trophoblastic disease. Am J Pathol,
1991;138(4):1035–43.
17. Trusolino, L., Bertotti, A., and Comoglio, P.M. MET signalling: principles
and functions in development, organ regeneration and cancer. Nat Rev Mol
Cell Cell Biol, 2010;11(12):834–48.
18. Bard‐Chapeau, E.A., Yuan, J., Droin, N. et al. Concerted functions of Gab1
and Shp2 in liver regeneration and hepatoprotection. Mol Cell Cell Biol,
2006;26(12):4664–74.
19. Paranjpe, S., Bowen, W.C., Mars, W.M. et al. Combined systemic elimina-
tion of MET and epidermal growth factor receptor signaling completely
abolishes liver regeneration and leads to liver decompensation. Hepatology,
2016;64(5):1711–24.
20. Carver, R.S., Stevenson, M.C., Scheving, L.A., and Russell, W.E. Diverse
expression of ErbB receptor proteins during rat liver development and regen-
eration. Gastroenterology, 2002;123(6):2017–27.
21. Skov Olsen, P., Boesby, S., Kirkegaard, P. et  al. Influence of epidermal
growth factor on liver regeneration after partial hepatectomy in rats.
Hepatology, 1988;8(5):992–6.
22. Olsen, P.S., Poulsen, S.S., and Kirkegaard, P. Adrenergic effects on secretion
of epidermal growth factor from Brunner’s glands. Gut, 1985;26(9):920–7.
23. Mead, J.E. and Fausto, N. Transforming growth factor alpha may be a physi-
ological regulator of liver regeneration by means of an autocrine mechanism.
Proc Natl Acad Sci USA, 1989;86(5):1558–62.
24. Berasain, C., Garcia‐Trevijano, E.R., Castillo, J. et  al. Amphiregulin: an
early trigger of liver regeneration in mice.[see comment]. Gastroenterology,
2005;128(2):424–32.
25. Khai, N.C., Takahashi, T., Ushikoshi, H. et  al. In vivo hepatic HB‐EGF
gene transduction inhibits Fas‐induced liver injury and induces liver regen-
eration in mice: a comparative study to HGF. J Hepatol, 2006;44(6):
1046–54.
26. Mitchell, C., Nivison, M., Jackson, L.F. et  al. Heparin‐binding epidermal
growth factor‐like growth factor links hepatocyte priming with cell cycle
progression during liver regeneration. J Biol Chem, 2005;280(4):2562–8.
27. Huang, X., Yu, C., Jin, C. et al. Ectopic activity of fibroblast growth factor
receptor 1 in hepatocytes accelerates hepatocarcinogenesis by driving prolif-
eration and vascular endothelial growth factor‐induced angiogenesis. Cancer
Res, 2006;66(3):1481–90.
28. Luo, Y., Yang, C., Lu, W. et al. Metabolic regulator betaKlotho interacts with
fibroblast growth factor receptor 4 (FGFR4) to induce apoptosis and inhibit
tumor Cell, proliferation. J Biol Chem, 2010;285(39):30069–78.
29. Padrissa‐Altes, S., Bachofner, M., Bogorad, R.L. et al. Control of hepatocyte
proliferation and survival by Fgf receptors is essential for liver regeneration
in mice. Gut, 2015;64(9):1444–53.
30. Cicione, C., Degirolamo, C., and Moschetta, A. Emerging role of fibroblast
growth factors 15/19 and 21 as metabolic integrators in the liver. Hepatology,
2012;56(6):2404–11.
31. Kong, B., Huang, J., Zhu, Y. et  al. Fibroblast growth factor 15 deficiency
impairs liver regeneration in mice. Am J Physiol Gastrointest Liver Physiol,
2014;306(10):G893–902.
32. Yang, C., Lu, W., Lin, T. et al. Activation of Liver FGF21 in hepatocarcino-
genesis and during hepatic stress. BMC Gastroenterol, 2013;13:67.
33. Schumacher, J.D. and Guo, G.L. Regulation of hepatic stellate cells and
fibrogenesis by fibroblast growth factors. Biomed Res Int, 2016;2016:
8323747.
34. Jirtle, R.L., Carr, B.I., and Scott, C.D. Modulation of insulin‐like growth
factor‐II/mannose 6‐phosphate receptors and transforming growth factor‐
beta 1 during liver regeneration (published erratum appears in J Biol Chem,
1991;266(36):24860). J Biol Chem, 1991;266(3):22444–50.
35. Michalopoulos, G.K. Principles of liver regeneration and growth homeostasis.
Compr Physiol, 2013;3(1):485–513.
36. Jakowlew, S.B., Mead, J.E., Danielpour, D., Wu, J., Roberts, A.B., and
Fausto, N. Transforming growth factor‐beta (TGF‐beta) isoforms in rat liver
regeneration: messenger RNA expression and activation of latent TGF‐ beta.
Cell Regul, 1991;2(7):535–48.
 45:  Liver Regeneration 581
37. Chari, R.S., Price, D.T., Sue, S.R., Meyers, W.C., and Jirtle, R.L. Down‐
regulation of transforming growth factor beta receptor type I., I.I., and III
during liver regeneration. Am J Surg, 1995;169(1):126–31.
38. Houck, K.A., Cruise, J.L., and Michalopoulos, G. Norepinephrine modulates
the growth‐inhibitory effect of transforming growth factor‐beta in primary
rat hepatocyte cultures. J Cell Physiol, 1988;135(3):551–5.
39. Michalopoulos, G.K., Bowen, W.C., Mule, K., and Stolz, D.B. Histological
organization in hepatocyte organoid cultures. Am J Pathol, 2001;159(5):
1877–87.
40. Rudolph, K.L., Trautwein, C., Kubicka, S. et al. Differential regulation of
extracellular matrix synthesis during liver regeneration after partial hepatec-
tomy in rats. Hepatology, 1999;30(5):1159–66.
41. Thenappan, A., Shukla, V., Abdul Khalek, F.J. et  al. Loss of transforming
growth factor beta adaptor protein beta‐2 spectrin leads to delayed liver
regeneration in mice. Hepatology, 2011;53(5):1641–50.
42. Ichikawa, T., Zhang, Y.Q., Kogure, K. et al. Transforming growth factor beta
and activin tonically inhibit DNA synthesis in the rat liver. Hepatology,
2001;34(5):918–25.
43. Cressman, D.E., Greenbaum, L.E., DeAngelis, R.A. et al. Liver failure and
defective hepatocyte regeneration in interleukin‐6‐deficient mice. Science,
1996;274(5291):1379–83.
44. Sakamoto, T., Liu, Z., Murase, N. et al. Mitosis and apoptosis in the liver
of  interleukin‐6‐deficient mice after partial hepatectomy. Hepatology,
1999;29(2):403–11.
45. Runge, D.M., Runge, D., Foth, H., Strom, S.C., and Michalopoulos, G.K.
STAT 1alpha/1beta, STAT 3 and STAT 5: expression and association with
c‐ MET and EGF‐receptor in long‐term cultures of human hepatocytes.
Biochem Biophys Res Commun, 1999;265(2):376–81.
46. Norris, C.A., He, M., Kang, L.I. et al. Synthesis of IL‐6 by hepatocytes is a
normal response to common hepatic stimuli. PLoS One, 2014;9(4):e96053.
47. Fausto, N. Involvement of the innate immune system in liver regeneration
and injury. J Hepatol, 2006;45(3):347–9.
48. Chaisson, M.L., Brooling, J.T., Ladiges, W., Tsai, S., and Fausto, N.
Hepatocyte‐specific inhibition of NF‐kappaB leads to apoptosis after TNF
treatment, but not after partial hepatectomy. J Clin Invest, 2002;110(2):
193–202.
49. Yamada, Y., Webber, E.M., Kirillova, I., Peschon, J.J., and Fausto, N.
Analysis of liver regeneration in mice lacking type 1 or type 2 tumor necrosis
factor receptor: requirement for type 1 but not type 2 receptor [comment].
Hepatology, 1998;28(4):959–70.
50. Argast, G.M., Campbell, J.S., Brooling, J.T., and Fausto, N. Epidermal
growth factor receptor transactivation mediates tumor necrosis factor‐
induced hepatocyte replication. J Biol Chem, 2004;279(3):34530–6.
51. Desbarats, J. and Newell, M.K. Fas engagement accelerates liver regeneration
after partial hepatectomy. Nat Med, 2000;6(8):920–3.
52. Oben, J.A., Roskams, T., Yang, S. et  al. Hepatic fibrogenesis requires
­sympathetic neurotransmitters. Gut, 2004;53(3):438–45.
53. Cruise, J.L., Houck, K.A., and Michalopoulos, G.K. Induction of DNA
­synthesis in cultured rat hepatocytes through stimulation of alpha 1 adreno-
receptor by norepinephrine. Science, 1985;227(4688):749–51.
54. Han, C., Bowen, W.C., Michalopoulos, G.K., and Wu, T. Alpha‐1 adrenergic
receptor transactivates signal transducer and activator of transcription‐3
(Stat3) through activation of Src and epidermal growth factor receptor
(EGFR) in hepatocytes. J Cell Physiol, 2008;216(2):486–97.
55. Cruise, J.L., Knechtle, S.J., Bollinger, R.R., Kuhn, C., and Michalopoulos,
G. Alpha 1‐adrenergic effects and liver regeneration. Hepatology, 1987;7(6):
1189–94.
56. Broten, J., Michalopoulos, G., Petersen, B., and Cruise, J. Adrenergic
­stimulation of hepatocyte growth factor expression. Biochem Biophys Res
Commun, 1999;262(1):76–9.
57. Yanagita, K., Nagaike, M., Ishibashi, H., Niho, Y., Matsumoto, K., and
Nakamura, T. Lung may have an endocrine function producing hepatocyte
growth factor in response to injury of distal organs. Biochem Biophys Res
Commun, 1992;182(2):802–9.
58. Huang, W., Ma, K. Zhang, J. et al. Nuclear receptor‐dependent bile acid sign-
aling is required for normal liver regeneration. Science, 2006;312:233–6.
59. Borude, P., Edwards, G., Walesky, C. et al. Hepatocyte‐specific deletion of
farnesoid X receptor delays but does not inhibit liver regeneration after par-
tial hepatectomy in mice. Hepatology, 2012;56(6):2344–52.
60. Keitel, V. and Haussinger, D. TGR5 in the biliary tree. Dig Dis,
2011;29(1):45–7.
61. Lesurtel, M., Graf, R., Aleil, B. et  al. Platelet‐derived serotonin mediates
liver regeneration. Science, 2006;312:104–7.
62. Matondo, R.B., Punt, C., Homberg, J. et  al. Deletion of the serotonin
transporter in rats disturbs serotonin homeostasis without impairing liver
regeneration. Am J Physiol Gastrointest Liver Physiol, 2009;296(4):
G963–8.
63. Furrer, K., Rickenbacher, A., Tian, Y. et al. Serotonin reverts age‐related cap-
illarization and failure of regeneration in the liver through a VEGF‐depend-
ent pathway. Proc Natl Acad Sci USA, 2011;108(7):2945–50.
64. Ochoa, B., Syn, W.K., Delgado, I. et al. Hedgehog signaling is critical for
normal liver regeneration after partial hepatectomy in mice. Hepatology,
2010;51(5):1712–23.
65. Bhave, V.S., Mars, W., Donthamsetty, S. et al. Regulation of liver growth by
glypican 3, CD81, hedgehog, and Hhex. Am J Pathol, 2013;183(1):153–9.
66. Sicklick, J.K., Li, Y.X., Choi, S.S. et  al. Role for hedgehog signaling in
hepatic stellate cell activation and viability. Lab Invest, 2005;85(1):1368–80.
67. Swiderska‐Syn, M., Xie, G., Michelotti, G.A. et al. Hedgehog regulates yes‐
associated protein 1 in regenerating mouse liver. Hepatology, 2016;64(1):
232–44.
68. Francavilla, A., Starzl, T.E., Porter, K. et al. Screening for candidate hepatic
growth factors by selective portal infusion after canine Eck’s fistula.
Hepatology, 1991;14(4 Pt 1):665–70.
69. Zerrad‐Saadi, A., Lambert‐Blot, M., Mitchell, C. et al. GH receptor plays a
major role in liver regeneration through the control of EGFR and ERK1/2
activation. Endocrinology, 2011;152(7):2731–41.
70. Jin, J., Wang, G.L., Shi, X., Darlington, G.J.,and Timchenko, N.A. The age‐
associated decline of glycogen synthase kinase 3beta plays a critical role in
the inhibition of liver regeneration. Mol Cell Biol, 2009;29(14):3867–80.
71. Tackett, B.C., Sun, H., Mei, Y. et al. P2Y2 purinergic receptor activation is
essential for efficient hepatocyte proliferation in response to partial hepatec-
tomy. Am J Physiol Gastrointest Liver Physiol, 2014;307(1):G1073–87.
72. Graubardt, N., Fahrner, R., Trochsler, M, et al. Promotion of liver regeneration
by natural killer cells in a murine model is dependent on extracellular adeno-
sine triphosphate phosphohydrolysis. Hepatology, 2013;57(5):1969–79.
73. Moolten, F.L. and Bucher, N.L. Regeneration of rat liver: transfer of humoral
agent by cross circulation. Science, 1967;158(798):272–4.
74. Marubashi, S., Sakon, M., Nagano, H. et al. Effect of portal hemodynamics
on liver regeneration studied in a novel portohepatic shunt rat model.
Surgery, 2004;136(5):1028–37.
75. Sokabe, T., Yamamoto, K., Ohura, N. et al. Differential regulation of uroki-
nase‐type plasminogen activator expression by fluid shear stress in human
coronary artery endothelial cells. Am J Physiol, 2004;287(5):H2027–34.
76. Mars, W.M., Kim, T.H., Stolz, D.B., Liu, M.L., and Michalopoulos, G.K.
Presence of urokinase in serum‐free primary rat hepatocyte cultures and
its  role in activating hepatocyte growth factor. Cancer Res, 1996;56(12):
2837–43.
77. Nejak‐Bowen, K., Moghe, A., Cornuet, P., Preziosi, M., Nagarajan, S., and
Monga, S.P. Role and regulation of p65/beta‐catenin association during liver
injury and regeneration: a "complex" relationship. Gene Expr, 2017;17(3):
219–35.
78. Mars, W.M., Zarnegar, R., and Michalopoulos, G.K. Activation of hepato-
cyte growth factor by the plasminogen activators uPA and tPA. Am J Pathol,
1993;143(3):949–58.
79. Mohammed, F.F., Pennington, C.J., Kassiri, Z. et  al. Metalloproteinase
inhibitor TIMP‐1 affects hepatocyte cell cycle via HGF activation in murine
liver regeneration. Hepatology, 2005;41(4):857–67.
80. Kim, T.H., Bowen, W.C., Stolz, D.B., Runge, D., Mars, W.M., and
Michalopoulos, G.K. Differential expression and distribution of focal adhe-
sion and cell adhesion molecules in rat hepatocyte differentiation. Exp Cell
Res, 1998;244(1):93–104.
81. Lindroos, P.M., Zarnegar, R., and Michalopoulos, G.K. Hepatocyte growth
factor (hepatopoietin A) rapidly increases in plasma before DNA synthesis
and liver regeneration stimulated by partial hepatectomy and carbon tetra-
chloride administration. Hepatology, 1991;13(4):743–50.
82. Michalopoulos, G.K. Liver regeneration after partial hepatectomy: critical
analysis of mechanistic dilemmas. Am J Pathol, 2010;176(1):2–13.
83. Michalopoulos, G.K. Advances in liver regeneration. Exp Rev Gastroenterol
Hepatol, 2014:1–11.
84. Nejak‐Bowen, K., Orr, A., Bowen, W.C., Jr., and Michalopoulos, G.K.
Conditional genetic elimination of hepatocyte growth factor in mice compro-
mises liver regeneration after partial hepatectomy. PLoS One, 2013;8(3):
e59836.
85. Hojilla, C.V., Mohammed, F.F., and Khokha, R. Matrix metalloproteinases
and their tissue inhibitors direct Cell, fate during cancer development. Br J
Cancer, 2003;89(10):1817–21.
582 THE LIVER:  REFERENCES
86. Gkretsi, V., Bowen, W.C., Yang, Y., Wu, C., and Michalopoulos, G.K.
Integrin‐linked kinase is involved in matrix‐induced hepatocyte differentia-
tion. Biochem Biophys Res Commun, 2007;353(3):638–43.
87. Gkretsi, V., Apte, U., Mars, W.M. et al. Liver‐specific ablation of integrin‐
linked kinase in mice results in abnormal histology, enhanced cell prolifera-
tion, and hepatomegaly. Hepatology, 2008;48(6):1932–41.
88. Apte, U., Gkretsi, V., Bowen, W.C. et al. Enhanced liver regeneration fol-
lowing changes induced by hepatocyte‐specific genetic ablation of integrin‐
linked kinase. Hepatology, 2009;50(3):844–51.
89. Monga, S.P., Pediaditakis, P., Mule, K., Stolz, D.B., and Michalopoulos,
G.K. Changes in WNT/beta‐catenin pathway during regulated growth in rat
liver regeneration. Hepatology, 2001;33(5):1098–109.
90. Kohler, C., Bell, A.W., Bowen, W.C., Monga, S.P., Fleig, W., and
Michalopoulos, G.K. Expression of Notch‐1 and its ligand Jagged‐1 in rat
liver during liver regeneration. Hepatology, 2004;39(4):1056–65.
91. Zhu, N.L., Asahina, K., Wang, J. et al. Hepatic stellate cell‐derived delta‐
like  homolog 1 (DLK1) protein in liver regeneration. J Biol Chem,
2012;287(13):10355–67.
92. Jirtle, R.L. and Michalopoulos, G. Effects of partial hepatectomy on
­transplanted hepatocytes. Cancer Res, 1982;42(8):3000–4.
93. Weymann, A., Hartman, E., Gazit, V. et  al. p21 is required for dextrose‐
mediated inhibition of mouse liver regeneration. Hepatology, 2009;50(1):
207–15.
94. Shido, K., Chavez, D., Cao, Z., Ko, J., Rafii, S., and Ding, B.S. Platelets
prime hematopoietic and vascular niche to drive angiocrine‐mediated liver
regeneration. Signal Transduct Target Ther, 2017;2.
95. Stocker, E. and Heine, W.D. Regeneration of liver parenchyma under nor-
mal and pathological conditions. Beitr Pathol, 1971;144(4):400–8.
96. Stocker, E., Wullstein, H.K., and Brau, G. Capacity of regeneration in
liver  epithelia of juvenile, repeated partially hepatectomized rats.
Autoradiographic studies after continous infusion of 3H‐thymidine
(author’s transl). Virchows Arch B Cell Pathol, 1973;14(2):93–103.
97. Kelley‐Loughnane, N., Sabla, G.E., Ley‐Ebert, C., Aronow, B.J., and
Bezerra, J.A. Independent and overlapping transcriptional activation during
liver development and regeneration in mice. Hepatology, 2002;35(3):
525–34.
98. Taub, R., Greenbaum, L.E., and Peng, Y. Transcriptional regulatory signals
define cytokine‐dependent and ‐independent pathways in liver regenera-
tion. Semin Liver Dis, 1999;19(2):117–27.
99. Yamamoto, T., Kojima, T., Murata, M. et  al. p38 MAP‐kinase regulates
function of gap and tight junctions during regeneration of rat hepatocytes. J
Hepatol, 2005;42(5):707–18.
100. Stolz, D.B., Mars, W.M., Petersen, B.E., Kim, T.H., and Michalopoulos,
G.K. Growth factor signal transduction immediately after two‐thirds partial
hepatectomy in the rat. Cancer Res, 1999;59(16):3954–60.
101. Li, W., Liang, X., Kellendonk, C., Poli, V.,and Taub, R. STAT3 contributes
to the mitogenic response of hepatocytes during liver regeneration. J Biol
Chem, 2002;277(32):28411–7.
102. Chaisson, M.L., Brooling, J.T., Ladiges, W., Tsai, S., and Fausto, N.
Hepatocyte‐specific inhibition of NF‐kappaB leads to apoptosis after TNF
treatment, but not after partial hepatectomy. J Clin Invest, 2002;110(2):
193–202.
103. Greenbaum, L.E., Cressman, D.E., Haber, B.A., and Taub, R. Coexistence
of C/EBP alpha, beta, growth‐induced proteins and DNA synthesis in hepat-
ocytes during liver regeneration. Implications for maintenance of the differ-
entiated state during liver growth. J Clin Invest, 1995;96(3):1351–65.
104. Luedde, T., Duderstadt, M., Streetz, K.L. et al. C/EBP beta isoforms LIP
and LAP modulate progression of the cell cycle in the regenerating mouse
liver. Hepatology, 2004;40(2):356–65.
105. Wang, X., Quail, E., Hung, N.J., Tan, Y., Ye, H., and Costa, R.H. Increased
levels of forkhead box M1B transcription factor in transgenic mouse hepat-
ocytes prevent age‐related proliferation defects in regenerating liver. Proc
Natl Acad Sci USA, 2001;98(20):11468–73.
106. Wang, X., Bhattacharyya, D., Dennewitz, M.B. et  al. Rapid hepatocyte
nuclear translocation of the Forkhead Box M1B (FoxM1B) transcription
factor caused a transient increase in size of regenerating transgenic hepato-
cytes. Gene Expr, 2003;11(3–4):149–62.
107. Bhave, V.S., Paranjpe, S., Bowen, W.C. et al. Genes inducing iPS pheno-
type play a role in hepatocyte survival and proliferation in vitro and liver
regeneration in vivo. Hepatology, 2011;54(4):1360–70.
108. Mullany, L.K., White, P., Hanse, E.A. et al. Distinct proliferative and tran-
scriptional effects of the D‐type cyclins in vivo. Cell Cycle, 2008;7(14):
2215–24.
109. Ng, R., Song, G., Roll, G.R., Frandsen, N.M., and Willenbring, H. A
microRNA‐21 surge facilitates rapid cyclin D1 translation and Cell,
cycle progression in mouse liver regeneration. J Clin Invest, 2012;122(3):
1097–108.
110. Albrecht, J.H., Poon, R.Y., Ahonen, C.L., Rieland, B.M., Deng, C., and
Crary, G.S. Involvement of p21 and p27 in the regulation of CDK activity
and Cell, cycle progression in the regenerating liver. Oncogene,
1998;16(16):2141–50.
111. Fausto, N. Protooncogenes and growth factors associated with normal and
abnormal liver growth. Dig Dis Sci, 1991;36(5):653–8.
112. Rabes, H.M. Kinetics of hepatocellular proliferation as a function of the
microvascular structure and functional state of the liver. Ciba Found Symp,
1977(5):31–53.
113. Klochendler, A., Weinberg‐Corem, N., Moran, M. et al. A transgenic mouse
marking live replicating Cell,s reveals in vivo transcriptional program of
proliferation. Dev Cell Cell, 2012;23(4):681–90.
114. Jiao, H., Zhu, Y., Lu, S., Zheng, Y., and Chen, H. An integrated approach
for  the identification of HNF4alpha‐centered transcriptional regulatory
­networks during early liver regeneration. Cell Physiol Biochem, 2015;36(6):
2317–26.
115. Michalopoulos, G., Cianciulli, H.D., Novotny, A.R., Kligerman, A.D.,
Strom, S.C., and Jirtle, R.L. Liver regeneration studies with rat hepatocytes
in primary culture. Cancer Res, 1982;42(1):4673–82.
116. Shteyer, E., Liao, Y., Muglia, L.J., Hruz, P.W., and Rudnick, D.A. Disruption
of hepatic adipogenesis is associated with impaired liver regeneration in
mice (see comment). Hepatology, 2004;40(6):1322–32.
117. Fernandez‐Rojo, M.A., Restall, C., Ferguson, C. et al. Caveolin‐1 orches-
trates the balance between glucose and lipid‐dependent energy metabolism:
implications for liver regeneration. Hepatology, 2012;55(5):1574–84.
118. Shu, J., Kren, B.T., Xia, Z. et al. Genomewide microRNA down‐regulation
as a negative feedback mechanism in the early phases of liver regeneration.
Hepatology, 2011;54(2):609–19.
119. Grijalva, J.L., Huizenga, M., Mueller, K. et  al. Dynamic alterations in
Hippo signaling pathway and YAP activation during liver regeneration. Am
J Physiol Gastrointest Liver Physiol, 2014;307(2):G196–204.
120. Yimlamai, D., Christodoulou, C., Galli, G.G. et al. Hippo pathway activity
influences liver cell fate. Cell, 2014;157(6):1324–38.
121. Toshima, T., Shirabe, K., Fukuhara, T. et al. Suppression of autophagy dur-
ing liver regeneration impairs energy charge and hepatocyte senescence in
mice. Hepatology, 2014;60(1):290–300.
122. Hoehme, S., Brulport, M., Bauer, A. et al. Prediction and validation of cell
alignment along microvessels as order principle to restore tissue architec-
ture in liver regeneration. Proc Nat Acad Sci USA, 2010;107(23):10371–6.
123. Miyaoka, Y. and Miyajima, A. To divide or not to divide: revisiting liver
regeneration. Cell Div, 2013;8(1):8.
124. Meier, M., Knudsen, A.R., Andersen, K.J., Bjerregaard, N.C., Jensen, U.B.,
and Mortensen, F.V. Gene expression in the liver remnant is significantly
affected by the size of partial hepatectomy: an experimental rat study. Gene
Expr, 2017;17(4):289–99.
125. Pean, N., Doignon, I., Garcin, I. et al. The receptor TGR5 protects the liver
from bile acid overload during liver regeneration in mice. Hepatology,
2013;58(4):1451–60.
126. Glaser, S., Han, Y., Francis, H., and Alpini, G. Melatonin regulation of bil-
iary functions. Hepatobiliary Surg Nutr, 2014;3(1):35–43.
127. Kennedy, L., Hargrove, L., Demieville, J. et al. Blocking H1/H2 histamine
receptors inhibits damage/fibrosis in Mdr2(‐/‐) mice and human cholangio-
carcinoma tumorigenesis. Hepatology, 2018.
128. Grappone, C., Pinzani, M., Parola, M. et al. Expression of platelet‐derived
growth factor in newly formed cholangiocytes during experimental biliary
fibrosis in rats. J Hepatol, 1999;31(1):100–9.
129. Li, B., Dorrell, C., Canaday, P.S. et al. Adult mouse liver contains two dis-
tinct populations of cholangiocytes. Stem Cell Reports, 2017;9(2):478–89.
130. Ross, M.A., Sander, C.M., Kleeb, T.B., Watkins, S.C., and Stolz, D.B.
Spatiotemporal expression of angiogenesis growth factor receptors during the
revascularization of regenerating rat liver. Hepatology, 2001;34(6):1135–48.
131. Modis, L. and Martinez‐Hernandez, A. Hepatocytes modulate the hepatic
microvascular phenotype. Lab Invest, 1991;65(6):661–70.
 45:  Liver Regeneration 583
132. Wack, K.E., Ross, M.A., Zegarra, V., Sysko, L.R., Watkins, S.C., and Stolz,
D.B. Sinusoidal ultrastructure evaluated during the revascularization of
regenerating rat liver. Hepatology, 2001;33(2):363–78.
133. DeLeve, L.D., Wang, X., and Wang, L. VEGF‐sdf1 recruitment of CXCR7+
bone marrow progenitors of liver sinusoidal endothelial cells promotes rat
liver regeneration. Am J Physiol Gastrointest Liver Physiol< 2016;310(9):
G739–46.
134. Ding, B.S., Nolan, D.J., Butler, J.M. et  al. Inductive angiocrine signals
from  sinusoidal endothelium are required for liver regeneration. Nature,
2010;468(7321):310–5.
135. Roskams, T., Cassiman, D., De Vos, R., and Libbrecht, L. Neuroregulation
of the neuroendocrine compartment of the liver. Anat Rec A Discov Mol
Cell Cell Evol Biol, 2004;280(1):910–23.
136. Unanue, E.R. Ito cells, stellate cells, and myofibroblasts: new actors in anti-
gen presentation. Immunity, 2007;26(1):9–10.
137. Gandhi, C.R. Hepatic stellate cell activation and pro‐fibrogenic signals.
J Hepatol, 2017;67(5):1104–5.
138. Zimmermann, A. Liver regeneration: the emergence of new pathways. Med
Sci Monit, 2002;8(3):RA53–63.
139. Widmann, J.J. and Fahimi, H.D. Proliferation of mononuclear phagocytes
(Kupffer cells) and endothelial cells in regenerating rat liver. A light and
electron microscopic cytochemical study. Am J Pathol, 1975;80(3):
349–66.
140. Fujii, H., Hirose, T., Oe, S. et al. Contribution of bone marrow cells to liver
regeneration after partial hepatectomy in mice. J Hepatol, 2002;36(5):653–9.
141. Ikarashi, M., Nakashima, H., Kinoshita, M. et  al. Distinct development
and  functions of resident and recruited liver Kupffer cells/macrophages.
J Leukoc Biol, 2013;94(6):1325–36.
142. Sanderson, N., Factor, V., Nagy, P. et  al. Hepatic expression of mature
­transforming growth factor beta 1 in transgenic mice results in multiple
tissue lesions. Proc Nat Acad Sci USA, 1995;92(7):2572–6.
143. Oe, S., Lemmer, E.R., Conner, E.A. et al. Intact signaling by transforming
growth factor beta is not required for termination of liver regeneration in
mice. Hepatology, 2004;40(5):1098–105.
144. Kim, T.H., Mars, W.M., Stolz, D.B., and Michalopoulos, G.K. Expression
and activation of pro‐MMP‐2 and pro‐MMP‐9 during rat liver regeneration.
Hepatology, 2000;31(1):75–82.
145. Kim, T.H., Mars, W.M., Stolz, D.B., Petersen, B.E., and Michalopoulos,
G.K. Extracellular matrix remodeling at the early stages of liver regenera-
tion in the rat. Hepatology, 1997;26(4):896–904.
146. Luo, J.H., Ren, B., Keryanov, S. et al. Transcriptomic and genomic analysis
of human hepatocellular carcinomas and hepatoblastomas. Hepatology,
2006;44(4):1012–24.
147. Pilia, G., Hughes‐Benzie, R.M., MacKenzie, A. et al. Mutations in GPC3, a
glypican gene, cause the Simpson‐Golabi‐Behmel overgrowth syndrome.
Nat Genet, 1996;12(3):241–7.
148. Liu, B., Bell, A.W., Paranjpe, S. et al. Suppression of liver regeneration and
hepatocyte proliferation in hepatocyte‐targeted glypican 3 transgenic mice.
Hepatology, 2010;52(3):1060–7.
149. Xue, Y., Mars, W.M., Bowen, W., Singhi, A.D., Stoops, J., and
Michalopoulos, G.K. Hepatitis C virus mimics effects of Glypican‐3 on
CD81 and promotes development of hepatocellular carcinomas via activa-
tion of hippo pathway in hepatocytes. Am J Pathol, 2018;188(6):1469–77.
150. Koral, K., Paranjpe, S., Bowen, W.C., Mars, W., Luo, J., and Michalopoulos,
G.K. Leukocyte‐specific protein 1: a novel regulator of hepatocellular pro-
liferation and migration deleted in human hepatocellular carcinoma.
Hepatology, 2015;61(2):537–47.
151. Yang, J., Cusimano, A., Monga, J.K. et al. WNT5A inhibits hepatocyte pro-
liferation and concludes beta‐catenin signaling in liver regeneration. Am J
Pathol, 2015;185(8):2194–205.
152. Tsagianni, A., Mars, W.M., Bhushan, B. et al. Combined systemic disrup-
tion of MET and epidermal growth factor receptor signaling causes liver
failure in normal mice. Am J Pathol, 2018;188(10):2223–35.
153. Arakaki, N., Kawakami, S., Nakamura, O. et al. Evidence for the presence
of an inactive precursor of human hepatocyte growth factor in plasma and
sera of patients with liver diseases. Hepatology, 1995;22(6):1728–34.
154. Pediaditakis, P., Lopez‐Talavera, J.C., Petersen, B., Monga, S.P., and
Michalopoulos, G.K. The processing and utilization of hepatocyte growth
factor/scatter factor following partial hepatectomy in the rat. Hepatology,
2001;34(4 Pt 1):688–93.
155. Le Bail, B., Bioulac‐Sage, P., Senuita, R., Quinton, A., Saric, J., and
Balabaud, C. Fine structure of hepatic sinusoids and sinusoidal cells in dis-
ease. J Electron Microsc Tech, 1990;14(3):257–82.
156. Michalopoulos, G.K. and Khan, Z. Liver stem cells: experimental findings and
implications for human liver disease. Gastroenterology, 2015;149(4):876–82.
157. Trautwein, C., Will, M., Kubicka, S., Rakemann, T., Flemming, P., and
Manns, M.P. 2‐acetaminofluorene blocks cell cycle progression after hepa-
tectomy by p21 induction and lack of cyclin E expression. Oncogene,
1999;18(47):6443–53.
158. Tatematsu, M., Ho, R.H., Kaku, T., Ekem, J.K., and Farber, E. Studies on
the proliferation and fate of oval cells in the liver of rats treated with 2‐
acetylaminofluorene and partial hepatectomy. Am J Pathol, 1984;114(3):
418–30.
159. Evarts, R.P., Nagy, P., Nakatsukasa, H., Marsden, E., and Thorgeirsson,
S.S. In vivo differentiation of rat liver oval cells into hepatocytes. Cancer
Res, 1989;49(6):1541–7.
160. Bisgaard, H.C., Nagy, P., Santoni‐Rugiu, E., and Thorgeirsson, S.S.
Proliferation, apoptosis, and induction of hepatic transcription factors are
characteristics of the early response of biliary epithelial (oval) cells to
chemical carcinogens. Hepatology, 1996;23(1):62–70.
161. Petersen, B.E., Zajac, V.F., and Michalopoulos, G.K. Bile ductular damage
induced by methylene dianiline inhibits oval cell activation. J Pathol,
1997;151(4):905–9.
162. Lu, W.Y., Bird, T.G., Boulter, L. et al. Hepatic progenitor cells of biliary
origin with liver repopulation capacity. Nat Cell Biol, 2015.
163. Michalopoulos, G.K., Barua, L., and Bowen, W.C. Transdifferentiation of
rat hepatocytes into biliary cells after bile duct ligation and toxic biliary
injury. Hepatology, 2005;41(3):535–44.
164. Limaye, P.B., Bowen, W.C., Orr, A.V., Luo, J., Tseng, G.C., and
Michalopoulos, G.K. Mechanisms of hepatocyte growth factor‐mediated
and epidermal growth factor‐mediated signaling in transdifferentiation of
rat hepatocytes to biliary epithelium. Hepatology, 2008;47(5):1702–13.
165. Sekiya, S., and Suzuki, A. Hepatocytes, rather than cholangiocytes, can be
the major source of primitive ductules in the chronically injured mouse
liver. Am J Pathol, 2014;184(5):1468–78.
166. Schaub, J.R., Huppert, K.A., Kurial, S.N.T. et al. De novo formation of the
biliary system by TGFbeta‐mediated hepatocyte transdifferentiation.
Nature, 2018;557(7704):247–51.
167. Sparks, E.E., Huppert, K.A., Brown, M.A., Washington, M.K., and Huppert,
S.S. Notch signaling regulates formation of the three‐dimensional architec-
ture of intrahepatic bile ducts in mice. Hepatology, 2010;51(4):1391–400.
168. Isse, K., Lesniak, A., Grama, K. et  al. Preexisting epithelial diversity in
normal human livers: a tissue‐tethered cytometric analysis in portal/peri-
portal epithelial cells. Hepatology, 2013;57(4):1632–43.
169. Petersen, B.E., Zajac, V.F., and Michalopoulos, G.K. Hepatic oval cell acti-
vation in response to injury following chemically induced periportal or
pericentral damage in rats. Hepatology, 1998;27(4):1030–8.
170. Carpentier, R., Suner, R.E., van Hul, N. et al. Embryonic ductal plate cells
give rise to cholangiocytes, periportal hepatocytes, and adult liver progeni-
tor cells. Gastroenterology,. 2011;141(4):1432–8, 8 e1–4.
171. Michalopoulos, G.K., Bowen, W.C., Mule, K., and Luo, J. HGF‐, EGF‐,
and dexamethasone‐induced gene expression patterns during formation of
tissue in hepatic organoid cultures. Gene Expr, 2003;11(2):55–75.
172. Hattoum, A., Rubin, E., Orr, A., and Michalopoulos, G.K. Expression of
hepatocyte epidermal growth factor receptor, FAS and glypican 3 in
EpCAM‐positive regenerative clusters of hepatocytes, cholangiocytes, and
progenitor cells in human liver failure. Hum Pathol, 2013;44(5):743–9.
173. Jaeschke, H. and Bajt, M.L. Intracellular signaling mechanisms of acetami-
nophen‐induced liver cell death. Toxicol Sci, 2006;89(1):31–41.
174. Slater, T.F., Cheeseman, K.H., and Ingold, K.U. Carbon tetrachloride toxic-
ity as a model for studying free‐radical mediated liver injury. Phil Trans R
Soc Lond B Biol Sci, 1985;311(1152):633–45.
175. Bhushan, B., Walesky, C., Manley, M. et al. Pro‐regenerative signaling after
acetaminophen‐induced acute liver injury in mice identified using a novel
incremental dose model. Am J Pathol, 2014;184(1):3013–25.
176. Bhushan, B., Chavan, H., Borude, P. et al. Dual role of epidermal growth
factor receptor in liver injury and regeneration after acetaminophen over-
dose in mice. Toxicol Sci, 2017;155(2):363–78.
177. Webber, E.M., FitzGerald, M.J., Brown, P.I., Bartlett, M.H., and Fausto, N.
Transforming growth factor‐alpha expression during liver regeneration
584 THE LIVER:  REFERENCES
after partial hepatectomy and toxic injury, and potential interactions
between transforming growth factor‐alpha and hepatocyte growth factor.
Hepatology, 1993;18(6):1422–31.
178. Lamers, W.H. The streaming liver: can the age of a hepatocyte be
determined from its position on the portohepatic radius? Hepatology,
­ nephrectomy kidney enlargement in patients with contralateral renal
1990;12(2):372–4.
179. Kennedy, S., Rettinger, S., Flye, M.W., and Ponder, K.P. Experiments in
transgenic mice show that hepatocytes are the source for postnatal liver
growth and do not stream. Hepatology, 1995;22(1):160–8.
180. Wang, B., Zhao, L., Fish, M., Logan, C.Y., and Nusse, R. Self‐renewing
diploid Axin2(+) cells fuel homeostatic renewal of the liver. Nature,
2015;524(7564):180–5.
181. Furuyama, K., Kawaguchi, Y., Akiyama, H. et al. Continuous cell supply
from a Sox9‐expressing progenitor zone in adult liver, exocrine pancreas
and intestine. Nat Genet, 2011;43(1):34–41.
182. Font‐Burgada, J., Shalapour, S., Ramaswamy, S. et  al. Hybrid periportal
hepatocytes regenerate the injured liver without giving rise to cancer. Cell,
2015;162(4):766–79.
183. Planas‐Paz, L., Orsini, V., Boulter, L. et  al. The RSPO‐LGR4/5‐ZNRF3/
RNF43 module controls liver zonation and size. Nat Cell Biol,
2016;18(5):467–79.
184. Lin, S., Nascimento, E.M., Gajera, C.R. et  al. Distributed hepatocytes
expressing telomerase repopulate the liver in homeostasis and injury.
Nature, 2018;556(7700):244–8.
185. Prassopoulos, P., Cavouras, D., and Gourtsoyiannis, N. Pre‐ and post‐
­cancer. Eur Urol, 1993;24(1):58–61.
186. Duncan, A.W., Taylor, M.H., Hickey, R.D. et al. The ploidy conveyor of mature
hepatocytes as a source of genetic variation. Nature, 2010;467(7316):707–10.
187. Anti, M., Marra, G., Rapaccini, G.L. et al. DNA ploidy pattern in human
chronic liver diseases and hepatic nodular lesions. Flow cytometric analysis
on echo‐guided needle liver biopsy. Cancer, 1994;73(2):281–8.
188. Duncan, A.W., Hanlon Newell, A.E., Smith, L. et al. Frequent aneuploidy
among normal human hepatocytes. Gastroenterology, 2012;142(1):25–8.
189. Solt, D.B., Medline, A.,and Farber, E. Rapid emergence of carcinogen‐
induced hyperplastic lesions in a new model for the sequential analysis of
liver carcinogenesis. Am J Pathol, 1977;88(3):595–618.
190. Zhang, Y.Y., Zhang, B.H., Ishii, K., and Liang, T.J. Novel function of CD81
in controlling hepatitis C virus replication. J Virol, 2010;84(7):3396–407.
191. Overturf, K., al‐Dhalimy, M., Ou, C.N., Finegold, M.,and Grompe, M.
Serial transplantation reveals the stem‐cell‐like regenerative potential of
adult mouse hepatocytes. Am J Pathol, 1997;151(5):1273–80.
 β‐Catenin Signaling
46 Satdarshan P.S. Monga
Pittsburgh Liver Research Center, University of Pittsburgh, School of Medicine and UPMC, Pittsburgh,
PA, USA
BACKGROUND
Genetic studies in species such as Xenopus, Drosophila, and
Caenorhabditis have lent themselves well to improve our under-
standing of the molecular basis of human diseases. A classic
example is the identification and characterization of the Wnt/β‐
catenin pathway that is crucial in normal development including
embryogenesis and organogenesis, and at the same time its
deregulation is implicated in disorders such as cancers (reviewed
in [1–3]). This pathway has remained conserved through the
evolutionary process. In Drosophila, the role of Wnt or Wingless
(Wg) was initially identified in normal wing development, how-
ever it was later recognized for multiple functions such as induc-
ing segment polarity and anterior–posterior patterning that were
imperative for a viable embryo [4–6]. As the importance of Wnt
emerged, several key components of this pathway were identi-
fied. The discovery of armadillo (or β‐catenin) added a signifi-
cant player to this orchestra and although circumstantial
evidence suggesting such a relationship existed earlier, it was a
few years later that β‐catenin was positively identified as a cen-
tral component of the canonical Wg pathway [4, 7–9]. These
studies led to the emergence of a model system for cell adhesion
and signal transduction [10]. This was also the beginning of the
understanding of the Wnt/β‐catenin pathway and its role in
complex cellular processes such as cell–cell adhesion, mitogen-
esis, motogenesis and morphogenesis in the vertebrates.
The next several years focused on the discovery of various
components of this pathway that improved our understanding
of  the regulation of this complex signaling cascade in normal
physiology and disease. Several important players such as the
Wnt receptor frizzled (Fzd), zeste‐white 3 kinase or glycogen
synthase kinase 3β (GSK3β), adenomatous polyposis gene
product (APC), axin, dishevelled (Dsh), and T cell factor/
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
lymphoid enhancement factor (TCF/LEF) family of transcrip-
tion factors, and their interactions were identified that were
directly influenced by the Wnt signaling. Many new compo-
nents and interactions as well as expanding list of target genes
are being identified. Also, research is focused on their role in
regulation of this pathway in health and disease. Furthermore,
crosstalk has now been established between the Wnt pathway
and other prominent pathways such as the Hippo, Hedgehog,
Jagged/Notch, HGF, EGF, and TGFβ signaling that could have
additional implications.
Presently, the role of Wnt/β‐catenin pathway is well estab-
lished in vertebrates in development, tissue homeostasis, and
carcinogenesis [11, 12]. β‐catenin knockout yields an embryonic
lethal phenotype in mice due to a defect in gastrulation [13].
Availability of conditional knockouts to overcome embryonic
lethality has been key to understanding a more ubiquitous role of
β‐catenin and other Wnt components in the development of
many organs such as kidneys, lungs, brain, limbs, muscles, and
skin [14–19]. The role of Wnt/β‐catenin signaling in liver patho-
biology has gained significant importance in the last 15–20
years. This chapter will highlight the role of the Wnt/β‐catenin
signaling pathway and of β‐catenin independent of the canonical
Wnt pathway in liver physiology and pathobiology (Table 46.1).
THE WNT SIGNAL TRANSDUCTION
PATHWAY
The binding of an extracellular secreted glycoprotein Wnt to
its  cell surface receptor Fzd or alternate receptors induces
­specific downstream cascades with unique biological functions.
Although the most characterized signaling downstream of the
586 THE LIVER:  ALTERNATE WNT SIGNALING PATHWAYS

Table 46.1  Selected Wnt/β‐catenin pathway targets in liver

Target genes Context Change Reference
Axin‐2 Liver tumors Upregulated [212]
Claudin‐2 Liver baseline/injury Downregulated [109, 112,
(β‐catenin knockout) 132]
Cyclin‐D1 Liver tumors (hepatoblastoma) Upregulated [101, 216,
Liver regeneration 217]
Cyclooxygenase‐2 Liver tumors (cell line) Upregulated [218]
Cyp1a2 Liver tumors Upregulated [219]
Cyp2e1 Liver tumors Upregulated [219]
Cyp7a1 Liver baseline/injury (β‐catenin knockout) Downregulated [112, 132,
133]
Cyp27 Liver baseline/injury (β‐catenin knockout) Downregulated [112, 132,
133]
Epidermal growth factor receptor Transgenic liver (β‐catenin overexpression) Upregulated [119]
Fibronectin Liver tumors (hepatoblastoma) Upregulated [216]
G‐protein‐coupled receptor 49 Hepatocellular cancer Upregulated [220]
(Gpr49 or Lgr5)
Glutamate transporter‐1(GLT‐1) Transgenic liver (mutant β‐catenin overexpression) Upregulated [99]
Glutamine synthetase (GS, Glul) Transgenic liver (mutant β‐catenin overexpression) Upregulated [99]
Lect2 Transgenic liver (mutant β‐catenin overexpression) Upregulated [221]
Ornithine aminotransferase Transgenic liver (mutant β‐catenin overexpression) Upregulated [99]
Regenerating iselet‐derived‐3α Liver tumors (HCC and hepatoblastoma Upregulated [222]
Regucalcin Baseline liver and liver tumors (β‐catenin transgenic and Upregulated (transgenic) [107]
knockout) Downregulated (knockout)
Tbx3 Liver tumors (hepatoblastoma) Upregulated [223]
TGFα Liver regeneration Upregulated [217]
VEGF Liver tumors (cell line) Upregulated [224]

Wnt is via the β‐catenin‐TCF‐dependent gene expression, the
signaling can occur through other effectors. In this, a broad
overview of the various signaling cascades activated down-
stream of Wnt are discussed.
THE WNT/β‐CATENIN SIGNALING
In a normal steady state, in the absence of a Wnt signal, the free
monomeric form of β‐catenin in the cytoplasm is actively targeted
for degradation by ubiquitination. In this “off” state, β‐catenin is
phosphorylated at serine45 (Ser45), Ser33, Ser33, and threo-
nine‐41 (Thr41) by casein kinase Iα (CKIα) and GSK3β [20, 21].
CK and GSK3β are part of a larger multiprotein degradation com-
plex that includes axin and APC. Once phosphorylated this larger
complex enables recognition and ubiquitination of β‐catenin by
β‐transducin repeat‐containing protein (βTrCP) and its ensuing
proteosomal degradation (Figure 46.1a, left panel) [22].
Wnt proteins usually act on a cell in a paracrine or autocrine
manner. However, Wnt proteins, in order to be biologically
active, need to undergo specific post‐translational modifica-
tions. Porcupine, located in the endoplasmic reticulum of a cell,
is essential for glycosylation and acylation of Wnts (Figure 46.1a,
right panel) [23]. Following acylation, Wnts become hydropho-
bic and require a specific cargo receptor, Wntless (Wls) or
Evenness Interrupted (Evi), to transport Wnts from Golgi to
membrane for secretion (Figure 46.1a, right panel) [24].
Once secreted, Wnts bind to cell surface Fzd receptor and
low‐density lipoprotein receptor‐related protein (LRP) 5/6 co‐
receptors [25–27]. This triggers phosphorylation of LRP5/6,
recruitment of Dishevelled (Dvl), and recruitment of Axin
to  the  plasma membrane (Figure  46.1, right panel) [28, 29].
Interestingly, the family of R‐spondin secreted proteins enhance
Wnt signaling through binding to the leucine‐rich repeat‐con-
taining G‐protein coupled receptor‐4 (LGR4) and LGR5 recep-
tors, which in turn increase Wnt‐dependent phosphorylation of
LRP6 [30]. The recruitment of Axin to the plasma membrane
disrupts the β‐catenin destruction complex, promoting its cyto-
plasmic stabilization and nuclear translocation where it binds
and transactivates the TCF/LEF family of transcription factors
to mediate target gene expression [31]. Several target genes of
the pathway have now been identified, which are stage‐ and tis-
sue‐specific, and the liver related targets will be discussed
throughout the text. It is important to note that several Wnt path-
way components such as AXIN, DKK, dFzd7, Fzd2, FRP2,
WISP, βTrCP, and TCF are themselves targets, suggesting exist-
ence of several regulatory loops within this pathway.
ALTERNATE WNT SIGNALING
PATHWAYS
There are 19 Wnts and 10 Frizzled receptors in the mammalian
genome. Understandably, not all Wnt signaling converges on
β‐catenin and various Wnt‐Fzd or Wnt‐non‐Fzd interactions can
lead to activation of alternate signaling pathways that are briefly
discussed here. It should be noted that some Wnts that were clas-
sically dubbed to be non‐canonical have now been shown to acti-
vate or inactivate β‐catenin based on interaction with specific
receptors and the final outcome of a particular signaling is con-
text‐dependent. The Wnt signaling usually activates but rarely
can inhibit β‐catenin‐TCF transactivation. An example is that of
Wnt5a, which can activate or inhibit β‐catenin based on the recep-
tor it binds [32]. It was shown that Wnt5a activated β‐catenin
signaling in the presence of Fz4 while it inhibited Wnt3a‐depend-
ent β‐catenin activation in the presence of Ror2 (Figure 46.1b).
 β-CATENIN SIGNALING
46:  587

(a) WNT
WNT
sFRP
WNT
WIF
LRP5/6
WNT Fzd
LRP5/6
Fzd Wls Dsh
Dkk APC
GSK3 Axin
Axin
WNT CK β-catenin
APC P Porcn
β-catenin
GSK3
CK TCF/LEF β-catenin
P
β-catenin
Target
genes
TCF/LEF

Serine 33 37 45
β-Catenin HGF EGF Fer Src (pp60)

Threonine 41 Y654 Y654 Y142 Y86 β-Catenin

Y670 Y654

(b) (c)
β-Catenin AJ
β-Catenin
WNT5a WNT5a WNT5a Actin cytoskeleton

Fzd4 Fzd2 Ror2 E-Cadherin

Tyrosine-
Occludin Phospho-
JAM-A β-Catenin
Activate Inactivate
β-catenin β-catenin Claudin-2
?
Claudin-1
TJ TCF

Figure 46.1  β‐catenin as a component of the Wnt pathway and cell–cell junctions. (a) On the left, in the absence of Wnt binding to its receptor
frizzled (Fzd) and co‐receptor LDL related protein 5/6 (LRP5/6) due to either sequestration of Wnt by soluble Frizzled related protein (sFRP), or
in the presence of Wnt inhibitory factor (WIF) or inhibition of Fzd‐LRP5/6 interactions by Dikkopf (Dkk), β‐catenin in cytoplasm is bound to its
degradation complex composed of Axin, APC, GSK3, and casein kinase (CK), which leads to its phosphorylation sequentially at Ser45 and Sr33,
Ser37 and Thr41. After phosphorylation, β‐catenin is ubiquitinated by βTrCP and degraded. On the right, Wnt is glycosylated and palmitoylated by
Porcupine (Porcn) in the endoplasmic reticulum and is bound to Wntless (Wls) in the Golgi apparatus to be transported to the membrane for secre-
tion. Wnt binds to receptor Fzd and co‐receptor LRP5/6, which recruits the β‐catenin degradation complex through Dishevelled (Dsh) to phosphor‐
LRP5/6, inactivating the degradation complex, leading to hypophosphorylation of β‐catenin, which is released and eventually translocates to the
nucleus to interact with T cell factor (TCF)/lymphoid enhancer factor (LEF) family of transcription factors to induce target genes. (b) Based on the
receptor binding, Wnt5a can activate or inactivate β‐catenin. If it binds to Fzd4, it can activate β‐catenin while if it binds to Ror2 or Fzd2, it can
inactivate β‐catenin through Wnt/calcium pathway or other mechanisms. (c) β‐Catenin is a component of adherens junctions (AJ) where it links the
cytoplasmic tail of E‐cadherin to α‐catenin and actin cytoskeleton. AJ in the liver are next to tight junctions (TJ) which are known for their barrier
function and are located close to the apical surface of hepatocytes preventing leakage of bile along the lateral surface of hepatocytes into the blood.
TJ are composed of proteins like occludin, junctional adhesion molecule‐A (JAM‐A) and claudins. Claudin‐2 is also a Wnt/β‐catenin target and may
be one basis of junctional crosstalk between the components of the AJ and TJ. Growth factors like HGF, EGF, Fer kinase, and Src kinase can phos-
phorylate β‐catenin at various tyrosine residues located at the C‐terminal of β‐catenin and disrupt β‐catenin‐E‐cadherin interactions and may also
induce β‐catenin nuclear translocation and activation.

The Wnt/Ca2+ pathway

Wnt ligands such as Wnt5a can bind to Fzd2 or Fzd7 or the tyros-
ine kinase‐like orphan receptor 2 (Ror2) to activate this pathway
[33]. Binding of Wnt leads to formation of a complex between
Fzd, Dishevelled, and G proteins, promoting the ­activation of
phospholipase C (PLC), which cleaves phosphatidylinositol
4,5 biphosphate (PIP2) into diacylglycerol (DAG) and inositol
1,4,5‐triphosphate (IP3). DAG activates protein kinase C
(PKC) and IP3 promotes the levels of intracellular calcium,
which in turn activates calcium‐calmodulin dependent kinase II
(CaMKII) and calcineurin (CaN) to regulate cell migration and
proliferation [34].
588 THE LIVER:  β-CATENIN AND PROTEIN KINASE A (PKA)
The planar cell polarity pathway
In the PCP pathway, binding of Wnt ligands to the Ror2/Fzd/
Dishevelled complex triggers the activation of Rho‐family
small GTPases including RhoA and Rac to then activate the
Rho‐associated protein kinase (ROCK) and c‐Jun N‐terminal
kinase (JNK) with an eventual impact on cell polarity and
migration [34, 35].
The Wnt/STOP pathway
This pathway leads to Wnt‐dependent stabilization of proteins
[36]. GSK3β is the central mediator of this pathway as it can
phosphorylate many additional proteins besides β‐catenin to tar-
get them for proteasomal degradation [37]. Wnt binding to its
co‐receptors can trigger the sequestration of GSK3β in multive-
sicular bodies, allowing the cytoplasmic accumulation of
GSK3β‐target proteins [38]. The biological effects of this path-
way can range from effect on cell cycle, cell division, regulation
of cytoskeleton, cell size regulation, and DNA remodeling [39].
Wnt/TOR signaling
In this pathway GSK3β phosphorylates and activates tuberous
sclerosis complex 2 (TSC2), which in turn inhibits the function
of mTOR complex 1 (mTORC1). The presence of Wnt ligands
inhibits GSK3β‐mediated phosphorylation and degradation of
TSC2, which activates the mTORC1 signaling pathway and
stimulates protein translation [40].
WNT INDEPENDENT ROLES
OF β‐CATENIN
Apart from playing a central role in the canonical Wnt pathway,
β‐catenin can interact with a number of proteins at membrane
and in the cytoplasm or nuclei to modulate their signaling and
biological function. A complete discussion on such interactions
is out of the scope of the current chapter but some key interac-
tions relevant in liver pathophysiology and pathobiology are
briefly discussed here.
β‐CATENIN AT ADHERENS JUNCTIONS
A crucial function of β‐catenin is to act as a bridge between the cyto-
plasmic domain of the cadherins and the actin‐containing‐cytoskel-
eton as component of the adherens junctions (AJ) [41]. Cadherins
consist of an extracellular domain, a transmembrane domain, and a
cytoplasmic tail that is the most conserved region among various
subtypes. Structurally, the cytoplasmic tails of cadherins show
dimerization and connect to the actin‐cytoskeleton via p120,
β‐catenin, and α‐catenin. Specific β‐catenin‐binding sites on the
cytoplasmic domain of cadherins have been characterized [42, 43].
The significance of regulation of β‐catenin‐cadherin interactions at
AJ is not only important in modulating cell–cell adhesion but has
been extended to transcriptional activation function of β‐catenin as
well, however, this remains controversial (Figure 46.1c). The inter-
actions between β‐catenin and cadherins are regulated by tyrosine
phosphorylation at the C‐terminal of β‐catenin (reviewed in [44]).
Phosphorylation of β‐catenin destabilizes cadherin‐β‐catenin bond,
α‐catenin‐β‐catenin complex, uncouples cadherin from actin
cytoskeleton, and promotes loss of intracellular adhesion [45, 46].
Conversely, dephosphorylating β‐catenin at tyrosine residues
enhanced E‐cadherin, β‐catenin, and α‐catenin reassembly [47].
Following tyrosine phosphorylation of β‐catenin, its cytosolic pool
is greatly increased, as is its ability to bind to TATA‐box binding
protein (TBP) to eventually increase transcriptional activity of the
β‐catenin/TCF complex [48]. This has also been narrowed down to
tyrosine residue 654. Another important ramification of tyrosine
phosphorylation of β‐catenin and dissociation of β‐catenin‐E‐cad-
herin complex is that it leaves the cytoplasmic domain of E‐cadherin
unstructured and vulnerable to degradation [42].
Several factors can regulate tyrosine phosphorylation of
β‐catenin including: (i) nonreceptor kinases, Src and Fer [49,
50]; (ii) transmembrane kinases, EGF receptor (EGFR) and Met
(HGF receptor) [51–55]; and (iii) various protein tyrosine phos-
phatases (Figure 46.1c).
We reported a novel Met‐β‐catenin complex at the hepatocyte
membrane [56]. HGF induced tyrosine phosphorylation‐
dependent nuclear translocation of β‐catenin. In a follow‐up
study, we identified tyrosine residues 654 and 670 as targets of
HGF‐induced β‐catenin phosphorylation (Figure  46.1c) [57].
Other reports had identified a similar effect of HGF on posi-
tively regulating β‐catenin/TCF transactivation, albeit via other
mechanisms [58–60]. These observations are relevant as high
levels of HGF have been observed in patients with liver patholo-
gies that might be influencing β‐catenin redistribution and alter-
ing the disease course [61, 62].
Similarly, direct interactions of β‐catenin with EGFR have
been reported as well (Figure  46.1c) [52]. In fact, ErbB2 has
also been shown to be associated with β‐catenin [53]. Although
the fate of β‐catenin and nuclear signaling in this context is
unclear its effect on cell–cell adhesion is well defined [63].
β‐CATENIN AND PROTEIN
KINASE A (PKA)
β‐catenin can also be phosphorylated by cAMP/PKA in vitro and
in vivo at Ser552 and Ser675, which leads to increased β‐catenin‐
TCF activation independent of any input from GSK3β [64]. This
has also been shown to occur in liver pathophysiology. In evalu-
ating ways to activate the Wnt/β‐catenin signaling to promote
liver regeneration after hepatectomy, the effect of triiodothyro-
nine (T3) or another thyroid hormone receptor‐β agonist GC‐1
(Sobetirome) was identified [65, 66]. In fact, we showed that
β‐catenin activation after T3/GC‐1 was at least partially depend-
ent on PKA‐mediated phosphorylation on Ser552 and Ser675.
Owing to the increased levels of cAMP observed in various
genetic diseases associated with the loss of fibrocystin function
such as congenital hepatic fibrosis and Caroli disease, there was
increased Ser675‐β‐catenin evident in the cholangiocytes [67].
This contributed to the increased motility of Pkhd1‐deficient
cholangiocytes and thus the cAMP/PKA/β‐catenin/TCF axis
 β-CATENIN SIGNALING
46:  589
may be playing an important result in the disease pathogenesis
and represent a therapeutic target [67].
β‐CATENIN AND NF‐κB
β‐catenin and p65 subunit of NF‐κB were shown to complex in
breast cancer [68]. We also discovered this complex in the liver, in
hepatocytes [69]. This complex seems to be inhibitory for p65
activation in hepatocytes. In fact, hepatocyte‐specific knockouts
of β‐catenin were protected from LPS injury due to increased
NF‐κB activation resulting in pro‐survival gene expression. There
is also an increased baseline immune cell infiltration in the liver
in the absence of β‐catenin and the significance of this is unclear.
Recently, we have also shown that while β‐catenin and NF‐κB are
both activated after partial hepatectomy, these ensue in distinct
cellular compartments with the former occurring in the hepato-
cytes while the latter observed in the non‐parenchymal cells [70].
Overexpression of β‐catenin has also been shown to decrease
NF‐κB reporter activity [71]. Cancer cells harboring β‐catenin
mutations showed reduced NF‐κB activity while enhanced
activity was seen in cells lacking basal β‐catenin activity.
Coimmunoprecipitation studies also confirmed formation of a
complex of β‐catenin with both p65 and p50 subunits of the
NF‐κB protein. Expression of NF‐κB targets like iNOS and Fas
also inversely correlated with β‐catenin levels in hepatocellular
cancer (HCC) samples. These findings are highly relevant as
majority of HCCs occur in livers with chronic injury and inflam-
mation, which may in part be due to the complex interplay of
β‐catenin and NF‐κB at least in a subset of cases.
β‐CATENIN AND FXR
More recently, a complex between farnesoid X receptor (FXR)
and β‐catenin in liver cells has also been shown. Very similar to
its interaction with NF‐κB, β‐catenin also complexes with FXR
and suppresses it. This observation was made after β‐catenin
conditional knockout mice were shown to be less injured after
bile duct ligation as compared to the control littermates [72]. It
appears that β‐catenin functions as both an inhibitor of nuclear
translocation and a nuclear corepressor through formation of a
physical complex with FXR. Loss of β‐catenin expedited FXR
nuclear localization and FXR‐retinoic X receptor (RXR)‐α
association, culminating in small heterodimer protein (SHP)
promoter occupancy and activation in response to bile acids
(BA) or FXR agonist. Conversely, excessive β‐catenin seques-
ters FXR, thus inhibiting its activation. This complex could be
playing a role in pathogenesis of cholestatic injury and may also
have therapeutic implications in various hepatic diseases.
WNT/β‐CATENIN SIGNALING IN LIVER
DEVELOPMENT
In 2003, the first study identified the role of Wnt/β‐catenin
signaling in liver development [73]. Utilizing in vitro organ
cultures and a comprehensive ontogenic analysis, the role of
β‐catenin was demonstrated in early liver development [73,
74]. Several groups have now complemented and furthered the
understanding of how extremely tight regulation of Wnt/β‐
catenin signaling is essential at multiple steps during hepatic
development [75].
Wnt signaling promotes the posterior endodermal fate and
suppresses the anterior endodermal fate during gastrulation and
early somitogenesis [76]. At the same time suppression of Wnt
signaling by secreted frizzled‐related protein 5 (sFRP5) has
been shown to maintain foregut fate in the anterior endoderm
and allows for liver development to initiate [76, 77]. However,
once anteroposterior endoderm patterning is established, Wnt
signaling now positively regulates liver specification [76].
These highly temporal and opposite roles of Wnt/β‐catenin
signaling in liver formation during early development are also
observed in zebrafish [78]. In fact, an unbiased forward‐genetic
screen in zebrafish led to the identification of wnt2bb mutants
that have very small or no liver buds [79]. Wnt2 knockdown in
wnt2bb mutants blocked liver recovery and importantly resulted
in no hhex expression in the liver‐forming region [80], indicat-
ing the essential role of Wnt signaling in liver specification.
Both wnt2 [80] and wnt2bb [79] are expressed in the lateral
plate mesoderm adjacent to the liver‐forming region.
Wnt/β‐catenin signaling is not only necessary but also suffi-
cient for liver specification. Gain‐of‐function studies in zebrafish
showed that overexpression of Wnt2bb [80] or Wnt8a [81] in
entire tissues induced ectopic hepatoblast and hepatocyte for-
mation in the posterior endoderm that normally gives rise to the
intestine. Wnt8a overexpression in non‐hepatic‐destined region
of dorsal endoderm in zebrafish was also shown to induce
ectopic hepatoblasts [82].
A mouse model in which Wnt/β‐catenin signaling is activated
or inactivated in the foregut endoderm after anteroposterior
endoderm patterning but prior to liver specification, will be nec-
essary to define the role of Wnt/β‐catenin signaling in liver
specification. We have performed foxa3‐cre driven β‐catenin
deletion that was evident at E9.5 in mouse hepatoblasts [83].
However, this did not affect the hepatoblast compartment and
HNF4α‐positive hepatoblasts were seen unequivocally in these
conditional knockout embryos. While this may imply that
Wnt/β‐catenin signaling is dispensable for hepatic induction in
mice, it is a technical challenge to abolish β‐catenin expression
at the right time and at the right place.
β‐catenin gene and protein expression peaks at E10–14 in
mouse embryonic liver and during this time β‐catenin is local-
ized in the nucleus, cytoplasm, and membrane in different epi-
thelial cells and coincided well with ongoing cell proliferation
[74]. Mouse embryonic livers from E9.5–10 stages cultured in
the presence of a β‐catenin antisense oligonucleotides, showed
decreased proliferation and a simultaneous increase in apopto-
sis, two processes vital to hepatic morphogenesis that follow
hepatic specification and induction [73]. This correlated well
with a subsequent study that found overexpression of β‐catenin
in developing chicken livers leads to a threefold increase in liver
size, which is due at least in part to an expanded hepatoblast
population [84].
Role of β‐catenin in bile duct morphogenesis during develop-
ment is also intriguing. Antisense against β‐catenin in
590 THE LIVER:  ROLE OF WNT/β-CATENIN AND CELL JUNCTIONS IN BLOOD BILE BARRIER
embryonic liver cultures led to decreased bile duct differentia-
tion and the addition of Wnt3a to the embryonic liver cultures
induced a biliary phenotype [85]. These observations were sup-
ported by in vivo studies that showed suboptimal biliary differ-
entiation in β‐catenin conditional null livers [83]. Conversely,
enhanced biliary differentiation of hepatoblasts was observed in
APC‐null livers that showed increased β‐catenin during prenatal
development, and led to an untimely demise of the embryos
[86]. More recent analysis shows β‐catenin to be dispensable for
differentiation of cholangiocytes to hepatocytes, although
β‐catenin needs to be kept in tight control during development
[87]. A cell nonautonomous role of Wnt/β‐catenin signaling in
biliary morphogenesis has also been reported whereby hepato-
cyte‐specific suppression of Wnt/β‐catenin signaling reduced
Notch activity in biliary cells by reducing expression of jag1ab
and jag2b in hepatocytes [88]. Wnt5a was shown to inhibit bile
duct differentiation of hepatoblasts [89]. Wnt5a deletion from
mesenchymal cells in mid‐gestational liver led to increased
expression of Sox9 and the numbers of hepatocyte nuclear fac-
tor (HNF)‐1β‐positive cholangiocyte precursors. In an in vitro
differentiation assay this effect was CaMKII‐dependent. The
authors did not address the state of β‐catenin signaling in this
study. Wnt5a treatment can activate NF‐AT to induce β‐catenin
degradation and thus may regulate biliary differentiation par-
tially through such a mechanism [90].
The role of β‐catenin in hepatocyte maturation during devel-
opment is important. Antisense‐mediated β‐catenin knockdown
in embryonic liver cultures led to persistent expression of stem
cell markers in hepatocytes [73]. Lack of β‐catenin in hepato-
blasts resulted in dramatic decreases in nuclear enriched tran-
scription factors CEBPα and HNF4α, impairing hepatocyte
maturation and fetal viability [83].
How is β‐catenin spatio‐temporally regulated during devel-
opment? Wnt2bb appears as the upstream effector of β‐catenin
during the earliest phases of hepatic morphogenesis [79]. Wnt9a
expression was reported in endothelial and stellate cells of the
embryonic sinusoidal wall in the developing liver [91]. This
report also showed presence of Frizzled 4, 7, and 9 on hepato-
cytes. Wnt5a was shown to be expressed in mesenchymal cells
during hepatic development although its effect on β‐catenin was
not directly addressed [89]. Based on interaction of Met and β‐
catenin and the role of HGF/Met in liver development, HGF/
Met/β‐catenin signaling may be relevant [56, 92–94]. Expression
of FGF‐10 in the mouse liver correlates with peak β‐catenin
activation; moreover, release of FGF‐10 from stellate cells stim-
ulates β‐catenin expression in hepatoblasts [95]. FGF‐2, FGF‐4,
and FGF‐8 can impact β‐catenin activation in embryonic liver
cultures [96].
ROLE OF WNT/β‐CATENIN
IN METABOLIC ZONATION
Liver occupies a strategic location in the body and receives
nutrient‐ and toxin‐enriched blood through portal circulation.
Histologically, this enters the hepatic lobule through the portal
vein located in the portal triad and mixes with blood from the
hepatic artery in sinusoids to eventually travel to the central
vein. While hepatocytes that are organized radially flanking
sinusoids along this porto‐central axis look grossly similar,
these are destined to perform differing functions. This charac-
teristic within a hepatic lobule is called metabolic zonation,
which has been known for several decades [97]. Wnt/β‐catenin
signaling has now been shown to be a key regulator of the gene
expression in the pericentral or zone three hepatocytes [98]. In
fact, β‐catenin activation controls expression of several genes
involved in glutamine metabolism and xenobiotic metabolism
including glutamine synthetase (GS), ornithine aminotrans-
ferase, and the glutamate transporter GLT‐1 [99–101].
The overall basis of pericentral Wnt/β‐catenin signaling is a
culmination of several factors. APC, the protein responsible for
β‐catenin degradation is expressed at highest levels in zones one
and two and absent in zone three. Wnt2 and Wnt9b have now
been shown to be basally secreted from endothelial cells lining
the central veins [102] to activate β‐catenin in pericentral hepat-
ocytes through as yet unknown Fzd receptor but requiring Wnt
co‐receptors LRP5/6 (Figure 46.2). This has been shown in vari-
ous genetic knockout mice. Mice lacking β‐catenin in hepato-
cytes [100, 101] or Wnt co‐receptors in hepatocytes [103], or
lacking Wntless in endothelial cells lining central veins prevent-
ing all Wnt secretion from these cells [102, 104, 105], all lack
Wnt/β‐catenin targets in zone three hepatocytes. Similarly, the
R‐Spondin/Lgr4/5 axis was shown to optimize Wnt signaling
and also contribute to activation of β‐catenin in zone three
hepatocytes [106].
Nuclear regulation of β‐catenin‐TCF4 complex during zona-
tion is also relevant to discuss. It was shown that in the presence
of β‐catenin, TCF4 is able to bind to Wnt‐responsive elements
on target gene promoters, while in its absence TCF4 binds
HNF4‐α responsive elements. This switch back and forth of
TCF4 may eventually be key to regulation of pericentral versus
periportal gene expression but needs to be further ascertained.
Several known and new targets of β‐catenin/TCF4 signaling in
the liver revealed in the study include Axin2, GS, constitutive
androstane receptor, CYP1A2, aryl hydrocarbon receptor,
Mrp2, glutathione S‐transferases, and Cyp27A1. Other intrigu-
ing targets of β‐catenin reported in the liver include regucalcin
or senescent marker protein‐30 and 4‐gulonolactonase, both of
which are essential for ascorbic acid biosynthesis in murine
liver [107]. Lect2 is also an important target of β‐catenin with
relevance in liver tumors and inflammation.
ROLE OF WNT/β‐CATENIN AND CELL
JUNCTIONS IN BLOOD BILE BARRIER
Liver is a highly vascular organ. Likewise, it also is the largest
gland, and among other factors, one of its main functions is the
secretion of bile. Blood and bile are kept segregated at a micro-
scopic level owing to the function of tight junctions (TJ) in
hepatocytes which allow bile to be secreted apically into biliary
canaliculi moving from the pericentral to periportal zone even-
tually to be carried into bile ductules, while blood traverses the
sinusoids at the basolateral surface of the hepatocytes.
Disruption of this barrier has been reported in a subset of cases
with progressive familial intrahepatic cholestasis (PFIC) due to
 β-CATENIN SIGNALING
46:  591

Figure 46.2  Cell–molecule circuitry of the β‐catenin activation in a baseline liver and after partial hepatectomy. At baseline, endothelial cells
lining the central vein constitutively release Wnt2 and Wnt9b, which act in a paracrine manner on proximal hepatocytes to activate β‐catenin‐TCF4
dependent expression of genes such as glutamine synthetase (GS), cytochrome P450 2e1 (Cyp2e1), and Cyp1a2. After partial hepatectomy, the
increased shear stress caused by sinusoidal blood flow on sinusoidal endothelial cells (red double sided arrows) likely triggers release of Wnt2 and
Wnt9b from these cells (1) to act on Fzd‐LRP5/6 receptors on hepatocytes in proximity (2) to induce β‐catenin stabilization, nuclear translocation,
and binding to TCF4 (3). This leads to increased transcription of Ccnd1 as early at 12 hours (12h) after PH (4) and can last up to 72h. The protein
increase of cyclin‐D1 is observed at 24 h which in turn starts to promote G1 to S phase transition of hepatocytes which peaks at 40h (5). Around
the same time, hepatocytes begin to express increased levels of Wnt5a (6) which is then released and through autocrine signaling is able to inactivate
β‐catenin signaling between 40–72h and thus contribute to termination of this mitogenic signal to eventually help in cessation of liver regeneration
when normal liver size is realized after partial hepatectomy.

loss‐of‐function mutation in the gene encoding for the TJ pro-
tein‐2 or ZO‐2 [108]. The role of β‐catenin as part of the AJ or
of AJ in general in the barrier function within the liver is not
well understood.
Liver‐specific β‐catenin conditional knockout mice also
showed increase in γ‐catenin protein levels, which associated
with E‐cadherin in lieu of β‐catenin to maintain cell–cell junc-
tions [109]. A follow‐up study showed that while not optimal,
γ‐catenin compensation at AJ via its interaction with E‐cadherin
did prevent E‐cadherin degradation from ubiquitin proteasome,
but E‐cadherin levels were still overall decreased in β‐catenin
knockout livers despite γ‐catenin upregulation [110]. Further,
increased serine/threonine phosphorylation of γ‐catenin after
β‐catenin knockdown was identified and at least partially PKA‐
dependent [110]. Despite γ‐catenin increase, all pericentral
Wnt/β‐catenin targets in the liver such as GS, cyp2e1, and
cyp1a2 were absent in the β‐catenin knockouts, showing lack of
compensation of γ‐catenin in the Wnt pathway [109, 110].
To conclusively address the functional relevance of γ‐catenin
at the AJ in the absence of β‐catenin, β‐ and γ‐catenin were
dually deleted in both hepatocytes and cholangiocytes using
Albumin‐Cre. This led to a dramatic phenotype consisting of
intrahepatic cholestasis, cholemia, biliary fibrosis, failure to
thrive, and mortality. While 80% of double knockouts died in
less than 30 days after birth, the survivors progressed to severe
fibrosis and HCC and succumbed by two and a half months of
age. The loss of the two catenins led to two major types of
molecular perturbations. Due to loss of β‐catenin and its role
within the Wnt signaling pathway, there was reduced expression
of TJ proteins such as claudin‐2 which is a known target of the
Wnt pathway [111]. However, such loss was also evident in
β‐catenin single knockout which did not have this phenotype of
the dual knockouts. Loss of γ‐catenin on top of β‐catenin loss,
we posit, caused a second hit deregulating junctional integrity
and decompensation. β‐catenin is known to interact with E‐
cadherin and through this association, it masks the PEST domain
in E‐cadherin, thus preventing its degradation [42]. β‐catenin loss
in the liver was compensated by increased γ‐catenin which
bound to E‐cadherin and hence prevented its degradation
although not as effectively as β‐catenin as some decrease in
E‐cadherin was evident upon β‐catenin loss despite increased
γ‐catenin [109, 110]. However, this prevented a major decom-
pensation and mice were able to survive and showed only mini-
mal increases in hepatic bile acids, serum bilirubin, and
reduction in bile flow due especially with age [72, 112]. Dual
loss of catenins in the liver, led to a notable E‐cadherin loss and
in addition also destabilized occludin, another TJ protein, lead-
ing to its loss as well [113]. Thus, dual loss of catenins, led to
592 THE LIVER:  ROLE IN LIVER REGENERATION AFTER PARTIAL HEPATECTOMY AND TOXICANT‐INDUCED INJURY
loss of claudin‐2 and occludin, disrupting TJ and leading to dis-
ruption of the blood bile barrier and a phenotype reminiscent
of PFIC. While dual loss of catenins has not been reported thus
far in PFIC cases, there remains a subset of these cases whose
pathogenesis remains unknown [114]. Further, alterations in
catenins may be a more global phenomenon contributing to
cholestatic injury and cholemia in various other hepatic patholo-
gies and more in depth study remains to be done.
ROLE IN POSTNATAL LIVER GROWTH
The liver continues to grow during neonatal stages. In fact, in
mice an early postnatal hepatic growth spurt occurs during the
first month after birth. Wnt/β‐catenin signaling was identified to
be active during these stages and correlated well with ongoing
hepatocyte proliferation through regulation of cyclin‐D1 expres-
sion [115]. Several additional models have been employed that
demonstrate a positive role of β‐catenin in liver growth. A trans-
genic mouse overexpressing a stable mutant of β‐catenin gener-
ated under the transcriptional control of calbindin‐D9K
(CaBP9K) promoter and liver‐specific enhancer of the aldolase
B gene displayed three to four times larger livers due to increased
cell proliferation [116]. Interestingly they did not detect any
changes in any of the conventional target genes of the pathway
such as c‐myc and cyclin‐D1. Subsequent analysis of transgenic
livers and subtractive hybridization led to identification genes
involved in glutamine metabolism as targets of the Wnt/β‐
catenin pathway. However, no signs of neoplastic transforma-
tion were reported in these animals, although APC‐conditional
null mice that exhibits β‐catenin stabilization leads to develop-
ment of robust HCCs [117]. Other transgenic mice have shown
similar hepatic growth advantage [118, 119] and also enabled
identification of new targets such as the epidermal growth factor
receptor [119]. Transgenic mice expressing Ser45‐mutant β‐
catenin under an albumin promoter also showed an initial
increase in liver size but then adapted via enhanced association
of β‐catenin to E‐cadherin at the membrane [120]. It was only
after stimulation of liver growth via hepatectomy or chemical
carcinogenesis that a notable growth advantage of active mutant
of β‐catenin was visible as compared to controls.
Two independent groups including ours have reported condi-
tional hepatic loss of β‐catenin affecting liver size [100, 101].
The decreased liver weight : body weight ratio was chiefly due
to reduced postnatal hepatocyte proliferation that led to
decreased hepatic mass which was sustained throughout the ani-
mal’s lifespan. The major mechanism of reduced proliferation
appears to be lower cyclin‐D1 expression, a known target of
β‐catenin in various tissues such as liver and colon [101, 121].
ROLE IN LIVER REGENERATION AFTER
PARTIAL HEPATECTOMY AND
TOXICANT‐INDUCED INJURY
The Wnt/β‐catenin pathway has been comprehensively studied
in liver regeneration (LR) that ensues after partial hepatectomy
(PH). The earliest study done in a rat model of PH showed tran-
sient stabilization of β‐catenin protein due to post‐translational
modification followed by its nuclear translocation within min-
utes after surgery [122]. Interestingly, the overall increase was
transient due to activation of β‐catenin degradation, β‐catenin
persisted in the nuclei of the hepatocytes until around 24 hours.
Knockdown of β‐catenin at the time of PH in rats using phospho
morpholino antisense oligonucleotides led to decreased hepato-
cyte proliferation and recovery of liver mass [123].
In mice, β‐catenin nuclear translocation has been observed as
early as 3–6 hours after PH [70]. Liver‐specific β‐catenin knock-
out mice when subjected to PH showed notable delay in peak
hepatocyte proliferation, which occurred at 72 hours instead of
40 hours [101, 124]. This occurred chiefly due to decreased cyc-
lin‐D1 along with other cyclins like A and E, thereby reducing
the number of hepatocytes in S‐phase in the knockout by at least
50% at 40 hours after PH. While redundancy among signaling
pathways ensuring successful LR after PH has been appreci-
ated, what drives it in the absence of β‐catenin remains unknown
[125]. To address what regulated β‐catenin activation during
LR, we utilized liver‐specific knockouts of Wnt co‐receptors
LRP5/6. Dual loss of these redundant receptors from hepato-
cytes led to a delay in LR after PH very similar to β‐catenin
knockouts [103]. This was associated with lack of β‐catenin
activation and decreased cyclin‐D1 expression in hepatocytes
and hence a decrease in the number of hepatocytes in S‐phase at
40 hours after PH. LR improved by 72–96 hours in these ani-
mals. Thus, β‐catenin is under the control of Wnt signaling dur-
ing LR. To determine which cells in the liver may be the source
of Wnts during LR, we conditionally knocked out Wntless from
hepatocytes and cholangiocytes, macrophages, and endothelial
cells. Only loss of Wnt secretion capability through deletion of
Wntless from endothelial cells phenocopied β‐catenin knock-
outs and LRP5/6‐double knockouts [105]. Deletion from hepat-
ocytes and cholangiocytes had no effect on peak hepatocyte
proliferation whereas deletion from macrophages had a modest
decrease suggesting contribution from these cells as secondary
source of Wnts during LR [103, 105]. Intriguingly, when
endothelial cells were isolated from regenerating livers at 12
hours after PH, the same two Wnts (Wnt2 and Wnt9b) that are
essential for pericentral β‐catenin activation at baseline, were
many‐fold upregulated in the endothelial cells as well [105].
The most upstream effectors that are responsible for basal Wnt2
and Wnt9b expression in central vein endothelial cells versus
induced Wnt2 and Wnt9b expression in sinusoidal endothelial
cells after PH remain unknown. We have recently speculated
that increased shear stress on sinusoidal endothelial cells imme-
diately after PH may be upstream of increased expression and
release of Wnt2 and Wnt9b. Indeed, when primary or immortal-
ized sinusoidal endothelial cells are subjected to orbital shear
stress, there was increased expression of several Wnts including
Wnt2 and Wnt9b [105].
Eventually, prolonged hepatocyte proliferation was observed
after PH when Wntless was deleted from hepatocytes. This led
to discovery of Wnt5a as a termination signal for the Wnt/
β‐catenin pathway which is secreted by hepatocytes when
β‐catenin activation is no longer needed [126].
Thus, the Wnt2/9b‐LRP5/6‐β‐catenin axis is the key regula-
tor of both the metabolic zonation to control expression of
 β-CATENIN SIGNALING
46:  593
genes in zone three as well as in regulating cyclin‐D1 expres-
sion and driving LR after PH (Figure  46.2). Additionally,
Wnt5a appears to terminate β‐catenin activation following
achievement of appropriate hepatocyte proliferation and recov-
ery of hepatic mass.
β‐catenin’s signaling in LR has also been shown in zebrafish.
Transgenic zebrafish lines that express dominant‐negative TCF
in the hepatocytes showed a blunted regenerative response to
hepatectomy due to decreased cell proliferation [78]. Similarly,
zebrafish expressing APC mutant or overexpressing Wnt8a
showed enhanced LR due to increased β‐catenin [78].
In patients, we have shown β‐catenin activation to positively
correlate with hepatocyte proliferation albeit in patients with
acetaminophen overdose [127]. First, β‐catenin was shown to
be relevant in toxicant‐induced liver injury and repair. Sublethal
dose of acetaminophen causes centrilobular necrosis that is fol-
lowed by spontaneous hepatocyte proliferation in adjacent
zone that begins the repair process. Following administration
of a single intra‐peritoneal dose of 500 mg kg−1 of acetami-
nophen to CD1 male mice, ALT levels increased from 3–12
hours and returned to normal by 24 hours [127]. PCNA‐posi-
tive hepatocytes were visible from 3–12 hours as well and
decreased by 24 hours at which time liver appeared almost nor-
mal. β‐catenin stabilization and activation was evident at 1–6
hours after acetaminophen injury and cyclin‐D1 increased from
at 3–12 hours. Liver‐specific β‐catenin knockouts lack cyp2e1
and cyp1a2, and hence are protected from acetaminophen over-
dose since they can’t metabolize acetaminophen to generate the
reactive metabolite [100, 101]. When expression of these two
P450’s was induced in the β‐catenin knockout mice, it enabled
partial metabolism of acetaminophen and in turn led to a mild
injury. When hepatocyte proliferation was compared between
controls and β‐catenin knockouts at equitoxic doses, there was
a deficit in hepatocyte proliferation in the knockouts support-
ing the role of β‐catenin in toxicant‐induced LR as well [127].
In a retrospective study in patients, β‐catenin nuclear/cytoplas-
mic localization correlated with high PCNA and spontaneous
LR, while predominantly membranous β‐catenin without any
nuclear/cytoplasmic redistribution correlated with decreased
proliferation index and need for liver transplantation [127].
Thus β‐catenin is also relevant in regulating hepatocyte prolif-
eration in humans.
To address if β‐catenin activation could promote LR, trans-
genic mice expressing a stable S45‐mutant‐β‐catenin were sub-
jected to PH. These animals showed a more pronounced
regenerative response with a shift to the left in regeneration
kinetics [120]. Likewise, when naked Wnt‐1 DNA was injected
weekly followed by PH, there was increased β‐catenin activa-
tion and premature hepatocyte proliferation [120]. More
recently we have identified an important role of triiodothyro-
nine (T3) or small molecule T3 agonist called GC‐1 (Sobetirome)
in inducing β‐catenin activation and in turn cyclin‐D1 expres-
sion. Both these agents showed a positive effect on hepatocyte
proliferation and promoting LR after PH [65, 66]. Interestingly,
these agents did not promote pro‐tumorigenic activity of onco-
genic β‐catenin in cancer models, but only promoted activation
of hepatocyte proliferation in normal hepatocytes through
β‐catenin activation [128]. Thus use of these agents is a highly
relevant means to stimulate β‐catenin and could have potential
clinical utility in various scenarios [129]. This would be relevant
in transplantation settings such as to induce regeneration in liv-
ing donors, small‐for‐size syndrome, or generally after liver
transplantation. It may also be useful in promoting LR after
toxicant‐induced liver injury.
ROLE OF WNT/β‐CATENIN SIGNALING
IN HEPATIC BILE ACID HOMEOSTASIS
AND INTRAHEPATIC CHOLESTASIS
Hepatocytes are responsible for the conversion of cholesterol
into bile acids. Bile acids, which are detergents and toxic to
hepatocytes, are secreted and eliminated into bile canaliculi for
eventual transport to the gut lumen to assist in the digestion of
dietary lipids and cholesterol. Two of the key enzymes responsi-
ble for the bile acid biosynthesis, Cyp7a1 and Cyp27, are mainly
expressed in the pericentral hepatocytes, suggesting they may
be regulated by Wnt/β‐catenin signaling [130].
Intrahepatic cholestasis encompasses defects in bile flow in
the liver within the intrahepatic bile ducts and biliary canaliculi.
The injury is likely due to the detergent action of the retained bile
acids within the liver that result in local injury to hepatocytes,
ensuing ductular reaction and associated inflammation and fibro-
sis. It has also been reported that HCCs with CTNNB1 mutation
and β‐catenin activation often display intratumoral cholestasis
[131]. Liver‐specific β‐catenin knockout mice showed increased
susceptibility to steatohepatitis with the methionine and choline
deficient diet [132]. These mice also showed increased basal
­levels of hepatic bile acids. This is at least in part thought to be
due to a defect in biliary canaliculi as they showed dilatation,
tortuosity, and loss of canalicular microvilli, likely due to absence
of β‐catenin targets claudin‐2 and regucalcin [112].
Initially it was thought that these increased bile acids may be
responsible for decreased expression of Cyp7a1 and Cyp27,
which are involved in bile acid synthesis from cholesterol, in the
β‐catenin knockouts as feedback mechanism [112, 132].
However, chromatin immunoprecipitation has shown these two
genes to also be direct targets of the Wnt pathway and thus play-
ing a role in bile acid metabolism [133].
An intriguing finding was made when liver‐specific β‐catenin
knockout mice were subjected to bile duct ligation [72]. There
was a notable decrease in hepatic injury observed which was a
result of failure of elevation of bile acid levels in the knockouts
after the surgery although these levels were notably upregulated
in control mice after the same procedure. It was shown that there
was overall decrease in hepatic bile acid biosynthesis due to
decreased expression of Cyp7a1 and Cyp27 which was in part
due to excessive FXR activation. In fact, we showed that
β‐catenin functioned as both an inhibitor of nuclear transloca-
tion and a nuclear corepressor through formation of a physical
complex with FXR. Loss of β‐catenin promoted FXR nuclear
localization and FXR/RXRα association, resulting in SHP pro-
moter activation in response bile acid or FXR agonists.
Thus β‐catenin is an important regulator of bile acid metabo-
lism and in specific cases, therapeutic inhibition of β‐catenin
especially with an FXR agonist may be beneficial specially to
decrease overall intrahepatic bile acid burden and related injury.
594 THE LIVER:  ROLE OF WNT/β-CATENIN SIGNALING IN HEPATOBLASTOMAS
ROLE IN HEPATIC FIBROSIS
Chronic injury to the liver results in hepatic fibrosis, a wound
healing response, which can lead to cirrhosis if the insult contin-
ues. Due to lack of effective treatment, other than alleviating the
injury, hepatic fibrosis and cirrhosis remain a major unmet clini-
cal need. Hepatic fibrosis is mostly a function of hepatic stellate
cells (HSCs) that are the source of extracellular matrix deposi-
tion within injured livers [134].
Wnt signaling is beginning to be assessed in hepatic stellate
cell biology. The first studies that showed increased expression
of the WNT pathway components in hepatic fibrosis were using
genomic analysis from primary biliary cholangitis livers [135,
136]. These studies identified increased expression of Wnt13,
Wnt5a, β‐catenin, and others, although a cause and effect rela-
tionship was not addressed. A study directly investigating gene
expression differences in quiescent versus activated HSCs iden-
tified, among others, an upregulated expression of Wnt5a and
Fz2 [137]. Subsequent studies have further identified aberrant
Wnt5a expression in fibrotic livers, and its suppression led to
reduced HSC activation [138, 139]. Wnt5a can have opposing
roles on β‐catenin signaling based on receptors expressed in a
cell. Whether the role of Wnt5a in promoting fibrosis is due to
inhibition, activation or independent of β‐catenin needs further
clarification. Unfortunately, there is evidence for all three in the
literature at the present time. As mentioned before, one study
showed no change in nuclear translocation of β‐catenin follow-
ing Wnt5a increase that was evident during stellate cell activa-
tion concomitant with upregulation of Fzd2 [137].
Another study showed that activation of Wnt/β‐catenin sign-
aling in primary rat HSC by an inhibitor of GSK3β decreased
synthesis of α‐smooth muscle actin and Wnt5a, and induced the
expression of glial fibrillary acidic protein [140]. This study fur-
ther showed that activation of Wnt signaling lowered DNA syn-
thesis and prevented HSCs from entering the cell cycle to
eventually demonstrate the role of Wnt signaling to β‐catenin in
maintaining their quiescence. However, a growing body of lit-
erature supports the activation of β‐catenin by WNT signaling
during the process of HSC activation and fibrosis [141–143]. In
fact, different molecules have been employed to block Wnt
signaling, directly or indirectly, to demonstrate an overall anti‐
fibrotic effect both in vitro and in vivo.
Activating pregnane X receptor by rifampicin, among other
things, also inhibited Wnt signaling and reduced HSC prolifera-
tion and transdifferentiation to active myofibroblasts [144].
Another study shows that necdin, a melanoma antigen family
protein that promotes neuronal and myogenic differentiation
while inhibiting adipogenesis, which expressed in HSCs.
Necdin is induced during HSC activation and its silencing
reversed them to quiescence through PPARγ and suppression of
Wnt/β‐catenin signaling [145]. Another study showed that
SEPT4, a subunit of the septin cytoskeleton specifically
expressed in quiescent HSCs, is downregulated through trans-
differentiation to activated myofibroblasts. Loss of SEPT4 in
HSCs coincided with decreased expression of Wnt inhibitor
DKK2, increased Wnt signaling via β‐catenin, and increased
fibrosis [146]. Another study supported that concept that the
canonical Wnt pathway promotes fibrogenesis, based on analy-
sis of mesoderm‐specific transcript homologue, a strong
negative regulator of Wnt/β‐catenin signaling [147]. The authors
showed that mesoderm‐specific transcript homologue expres-
sion in HSCs alleviated carbon tetrachloride‐induced collagen
deposition in liver tissue by decreasing the expression of
β‐catenin, α‐smooth muscle actin, and Smad3 both in vivo and
in vitro.
Wnt‐β‐catenin signaling might lead to myofibroblastic acti-
vation of HSCs via negative regulation of adipogenesis [148].
Expression of Wnt11 or β‐catenin with the S33Y mutation in
preadipocytes prevented their differentiation to adipocytes at
least in part through inhibition of expression of CCAAT
enhancer binding protein‐α (CEBPα) and PPARγ. Conversely,
blockade of β‐catenin signaling either through dominant‐nega-
tive TCF4 or axin, facilitated adipogenic differentiation of the
preadipocytes. Quiescent HSCs contain lipid droplets and their
activation to myofibroblasts involves loss of adipogenic genes
including PPARG and CEBPA, so this event is analogous to dif-
ferentiation of adipocytes to preadipocytes. Gain‐of‐function
mutations in adipogenic transcription factors such as PPARG
and sterol regulatory element binding protein‐1c can reverse
culture‐induced myofibroblast into a quiescent HSC phenotype.
Yet another study showed HSC activation to myofibroblasts by
the Wnt pathway to be mediated by miR‐132 and miR‐212
which led to epigenetic repression of methyl‐CpG binding pro-
tein 2 (MeCP2) leading to loss of PPARγ [149]. More recently,
in HSCs expression of stearoyl‐CoA desaturase expression was
shown to be regulated by the Wnt‐β‐catenin signaling. This
enzyme generates monounsaturated fatty acids, which were
shown to provide a feed forward loop to augment Wnt signaling
through the stabilization of Lrp5 and Lrp6 mRNAs, which
encode for Wnt co‐receptors. This was shown to eventually con-
tribute to liver fibrosis through HSC activation [150]. Taken
together these observations indicate that activation of Wnt sign-
aling via β‐catenin could be involved in HSC activation by
inhibiting the adipogenic programs; inhibition of this signaling
pathway could contribute to adipogenic gene profile of a quies-
cent HSC. Inhibiting Wnt signaling to β‐catenin might therefore
block hepatic fibrosis.
ROLE OF WNT/β‐CATENIN SIGNALING
IN HEPATOBLASTOMAS
Hepatoblastoma (HB) is the commonest malignant hepatic
tumor of childhood. These tumors are frequently sporadic; how-
ever, the incidence is highest in patients of familial adenoma-
tous polyposis coli [151]. This led to the identification of APC
mutations in HB in familial cases [152]. Increased frequency of
diverse APC mutations (57%) were reported in sporadic form of
the disease as well [153]. N‐terminal mutations (missense and
deletions) affecting exon 3 of CTNNB1 have also been found in
up to 80–90% of all HBs and are associated with nuclear and
cytoplasmic β‐catenin localization [154]. Mutations in AXIN1
have been identified in less than 10% of these tumors [155]. A
recent comprehensive study in 85 HB patients showed 65 cases
with missense mutations and interstitial deletions in CTNNB1
[156]. They also identified loss‐of‐function mutations in APC
and AXIN1 genes. Thus 82% of HB in this cohort exhibited
 β-CATENIN SIGNALING
46:  595
Wnt/β‐catenin activation. Thus, there is compelling data that β‐
catenin activation is an obligatory event in the etiopathogenesis
of HB, although the exact mechanism of how it leads to HB is
unclear [157].
Activation of Wnt signaling during normal liver development
has already been discussed in this chapter and plays role in both
hepatoblast proliferation and hepatocyte differentiation and
maturation (also reviewed in [75, 158, 159]). While β‐catenin
activation in liver development is spatio‐temporally regulated
and ligand‐dependent, in HB its activation is unrestricted, sus-
tained, and non‐ligand‐dependent due to exon‐3 mutations.
How this contributes to tumors remains to be addressed although
various target genes of Wnt signaling such as c‐Myc, cyclin‐D1,
GS, EGFR/Axin‐2, and others have been reported in various
histological subtypes of HB [160]. Very similar to β‐catenin in
different stages of normal liver development, a more nuclear
β‐catenin is evident in embryonal HB coinciding with lack of
GS, and more membranous, cytoplasmic and nuclear with GS
expression in fetal HB [161]. Intriguingly, when an N‐terminal
deletion mutant of β‐catenin is overexpressed in the liver using
an adenoviral approach or by generation of a transgenic mouse
overexpressing β‐catenin under liver‐specific promoter, the
mice never display HB [116, 118]. Also, transgenic mice
expressing point‐mutant form of β‐catenin do not show HB
[120]. This suggests that oncogenic β‐catenin is insufficient in
inducing HB. How β‐catenin activation leads to HB was unclear
until recently.
A recent study demonstrates β‐catenin mutations frequently
co‐occur with activation of Yes associated protein‐1 (Yap), a
component of the Hippo signaling pathway. In fact, 80% of all
HB cases, irrespective of histological subtype, showed nuclear
β‐catenin due to mutations and nuclear Yap due to unknown
mechanism [162]. Co‐expression of deletion mutant of
β‐catenin and constitutively active Yap by sleeping beauty trans-
poson/transposase and hydrodynamic tail vein injection (SB‐
HTVI), led to development of HB in mice. The role of
Yap‐TEAD and β‐catenin‐TCF appears to be key in regulating
target genes that may be leading to HB development and further
studies will be critical in addressing the mechanism. C‐Myc was
shown to be dispensable for HB development in this model
although it played role in optimum HB development by regulat-
ing tumor metabolism [163]. Likewise, lipocalin‐2, identified as
one of the genes that was many‐fold upregulated in Yap‐β‐
catenin HB in mice and also contained TEAD and TCF binding
sites in its promoter, was also shown to be dispensable for HB
development [164]. However, serum lipocalin‐2 was identified
as a sensitive marker of HB tumor burden in the mouse model
and may be of relevance clinically as well. Future studies would
be invaluable in understanding the role of these pathways in
tumor biology.
ROLE OF WNT/β‐CATENIN SIGNALING
IN HEPATOCELLULAR ADENOMA
Hepatocellular adenomas (HCAs) are characterized by mono-
clonal proliferation of well‐differentiated hepatocytes that are
usually arranged in sheets and cords. There is a classical absence
of a portal triads and interlobular bile ducts within HCAs. While
initially considered as a well‐defined homogeneous entity, it is
now known that several molecular classes of this tumor‐type
exist which dictates tumor origin, behavior, and eventually
determines prognosis and helps determine treatment options
(reviewed in [165]). Around 10–15% display Wnt/β‐catenin
activation secondary to mutations in the CTNNB1 gene [166].
β‐catenin‐mutated HCAs are exclusive from the HNF1A muta-
tion group but may occur in combination with gp130 or GNAS
mutations. Around half of the β‐catenin‐active HCAs are inflam-
matory. GS is upregulated in β‐catenin‐mutated HCA. Since
detection of nuclear β‐catenin by immunostaining is often chal-
lenging due to technical issues as well as heterogeneity in stain-
ing patterns, it may not be conclusive. GS staining can aid in the
diagnosis of β‐catenin‐mutated HCA with greater sensitivity
[167]. β‐catenin‐mutated HCA can occur in males. CTNNB1‐
mutated HCA histologically exhibit cholestasis and cellular
dysplasia. Most importantly, these subsets of HCA have greater
propensity for malignant transformation [166]. Another subset
of HCA has activation of the JAK/STAT pathway and show
polymorphic inflammatory infiltrates. Mutations in IL6ST
(coding for gp130), STAT3, and GNAS, can activate JAK/STAT
signaling. A subset of this tumor‐type also exhibits CTNNB1
mutation, irrespective of the molecular driver, and poses an
increased risk of malignant transformation as well.
ROLE OF WNT/β‐CATENIN SIGNALING
IN HEPATOCELLULAR CANCER
Wnt/β‐catenin activation has been implicated in a major sub-
set of HCC cases. Abnormal localization of cadherins and
catenins in liver cancer was first shown by immunohisto-
chemistry [168]. A more comprehensive study identified
anomalous β‐catenin expression as well as mutations in the
CTNNB1 in around 25% of all HCC cases and up to 50% of
all hepatic tumors in transgenic lines such as c‐myc or H‐ras
[169]. Several subsequent human studies corroborated these
observations and currently around 8–44% of HCCs show
mutations in β‐catenin gene, mostly in exon‐3 although recent
studies have also identified a region in exon 7/8 of CTNNB1
to be a hot spot [170, 171] (Table  46.2). In fact, CTNNB1
mutations have a trunk role in HCC‐like mutations affecting
TERT and p53 [172].
Mutations have also been reported in other components of the
degradation complex of β‐catenin including AXIN1 in around
3–16% [155, 173, 174] and AXIN2 in around 3% of all HCC
cases  [155]. Additional mechanisms have also been described
and include overexpression of FRZ7 [175, 176], Wnt3 upregula-
tion [177], inactivation of GSK3β [178], methylation of soluble
frizzled related protein1 (sFRP1) [179], epigenetic inactivation
of several sFRPs [180], and TGFβ‐dependent activation of
β‐catenin [181]. In HCC patients, exome sequencing has also
revealed cooperating mutations of CTNNB1 with mutations in
ARID2, NFE2L2, TERT, APOB, and MLL2. Likewise, certain
mutations were also always mutually exclusive from CTNNB1
and demonstrated lack of cooperation in development of HCC.
These included TP53 and AXIN1.
596 THE LIVER:  ROLE OF WNT/β-CATENIN SIGNALING IN HEPATOBLASTOMAS

Table 46.2  List of studies showing spectra of mutations in Ctnnb1 into account geographical factors, etiology, co‐morbidities, and
gene kind of mutations may be essential to clarify this discrepancy.
Cases with Decreased fibrosis was reported in a few studies and remains
mutations in exon‐3 an intriguing hallmark of β‐catenin mutated tumors [182, 188].
Study of CTNNB1 (%) Additional information
Are β‐catenin mutations a risk factor for development of HCC
[171] 36/194 (18.5%) Additional 15/194 cases had mutations independent of cirrhosis or are they merely decreasing the
(TCGA) and 90/249 in exon 7/8 of CTNNB1
(36%) (French Additional 3/249 cases had mutations
threshold of neoplastic transformation? It is also intriguing to
cohort) in exon 7/8 of CTNNB1 note that in a recent study, a small but significant subset of HCV
[170] 90/249 (36%) Overall, 52% HCC cases showed patients developed HCC without evidence of advanced fibrosis
activation of Wnt/β‐catenin including [189]. Along the same lines, it is relevant to note that the small
mutations in AXIN1 and exon 7/8 of
CTNNB1 subset of hepatic adenomas that progress to HCC in patients
[225] 41/125 (33%) Additional 15.2% showed mutations in often exhibit β‐catenin gene mutations [166]. This neoplastic
AXIN1 transformation of adenomas occurs in a healthy liver without
[182] 9/32 (28%) Tyrosine‐654 phosphorylated β‐catenin
was observed in fibrolamellar variety
any evidence of fibrosis and further supports the role of
of HCC β‐catenin in fibrosis‐independent HCC. To address the relation-
[174] 20/45 (44%) Additional seven patients had AXIN1 ship of advanced fibrosis, β‐catenin mutations and HCC, we
mutations used an experimental approach in which ser‐45‐mutant
[226] 15/45 (33%) No GSK3β mutations
[220] 16/38 (42%) Multiple mutations in two patients β‐catenin transgenic mice and control mice were fed thioaceta-
[155] 14/73 (19%) One insertion between S33 and G34 mide diet for a prolonged period [190]. No difference in HCC in
[227] 5/62 (8%) Aflatoxin study the two groups of mice suggests that β‐catenin mutations do not
[228] 7/60 (12%) 62% of tumors had cytoplasmic
enhance evolution of cirrhosis to HCC supporting β‐catenin
β‐catenin staining
[229] 57/434 (13%) 34 had mutations at GSK3β sites. 17 gene mutations and cirrhosis to be independent contributors to
showed mutations at codons 32 and 34 tumorigenesis that do not cooperate.
[230] 9/22 (41%) Multiple mutations in one patient The Wnt/β‐catenin pathway in HCC in experimental
[186] 12/35 (34%) Multiple mutations in two patients
[231] 21/119 (18%) Multiple mutations in a patient ­models  deserves mention. The first question that has been
[32] 9/38 (24%) Aberrant accumulation of β‐catenin in answered is if β‐catenin mutation, activation, or overexpression
nucleus, cytoplasm, and membrane by itself is sufficient for initiation of HCC. None of the trans-
was seen in 39% cases genic mice overexpressing either wild‐type or stable‐mutants
[169] 8/31 (26%) Two patients had mutations at D32
of β‐catenin thus far have exhibited spontaneous HCC [116,
Deletions usually involved one of the key sites: S33, S37, S45, or T41. 118–120]. However, several studies now suggest that β‐catenin
­collaborates with other signaling pathways to contribute to
It is presently unclear if the extent of Wnt activation due to hepatocarcinogenesis. β‐catenin was shown to cooperate with
these diverse mechanisms is comparable. At the same time, it is activated Ha‐ras in HCC [190]. Mice heterozygous for Lkb1
unknown if downstream signaling in the form of target gene deletion showed an accelerated progression to HCC when
expression is also varied in these disparate mechanisms of mated with adenovirus‐inducible β‐catenin mutant mice [192].
β‐catenin activation. A study showed significant correlation Similarly, when chemical carcinogen diethylnitrosamine
between CTNNB1 mutations and overexpression of target genes (DEN), which normally induces HCC through Ha‐ras activa-
GS, G‐protein‐coupled receptor (GPR)49, and glutamate trans- tion [193], was injected in serine‐45‐mutant β‐catenin trans-
porter (GLT)‐1 (P = 0.0001), but not for other target genes like genic mouse, these animals developed HCC earlier and
ornithine aminotransferase, LECT2, c‐myc, or cyclin D1 [174]. more  profoundly [120]. These findings strongly suggest that
This study showed GS to be a good immunohistochemical β‐catenin mutation is one of the hits that may be critical to the
marker of β‐catenin activation in HCC, which was also con- development of HCC, however, additional aberrations are nec-
firmed independently [182]. However, no increase in expression essary for tumorigenesis. This is also in agreement with clinical
of GS, GPR49, or GLT‐1 was evident in loss of Axin‐1 function studies where CTNNB1 mutations significantly coexisted with
due to AXIN1 mutations. Thus, it is likely that functional equiv- mutations in ARID2, NFE2L2, TERT, APOB, and MLL2 [170,
alence of various modes of β‐catenin activation is distinct and 194]. Since concomitant Met overexpression/activation and
may have differing consequences in tumor phenotype. Indeed mutations in CTNNB1 were found in around 11% of all HCC
β‐catenin‐active HCC due to mutations in CTNNB1, AXIN1, or cases, these two proto‐oncogenes were co‐expressed using SB‐
additional modes of β‐catenin activation, have all been shown to HTVI. These mice developed HCC with gene expression pro-
have distinct phenotype in the transcriptomic classification of files that displayed high correlation with the gene profiles of a
HCC [174, 181, 183]. Furthermore, recent studies have even subset of human HCC patients with both CTNNB1 mutations
shown disparity in extent of Wnt signaling based on differences and Met activation signatures [171]. To address if Ras activa-
in mutations within exon‐3 of CTNNB1 [184]. tion downstream of Met could be contributing to Met‐β‐catenin
The effect of β‐catenin mutations on prognosis remains HCC, G12D‐KRAS and mutant‐β‐catenin were expressed using
debatable. β‐catenin mutations have been associated with better SB‐HTVI, which also yielded HCC with around 90% molecu-
prognosis and a more differentiated tumor type in some studies lar similarity to Met‐β‐catenin HCC [195]. It will be useful to
[183, 185]. Others have noted high nuclear and cytoplasmic generate relevant models using SB‐HTVI that represent subsets
β‐catenin in more proliferating and poorly differentiated HCC of human HCC which can then be assessed for biological as
[181, 186, 187]. Eventually, more careful meta‐analysis taking well as therapeutic studies.
 β-CATENIN SIGNALING
46:  597
Several proof of principle preclinical studies have demon-
strated an important benefit of therapeutic inhibition of β‐
catenin for treatment of HCC [196]. Various cox2 inhibitors
such as rofecoxib have shown efficacy in decreasing β‐catenin
levels along with shrinkage of tumors [197]. R‐Etodolac, an
enantiomer of a cox2 inhibitor that lacks an inhibitory effect on
cox2 has also shown anti‐β‐catenin effect [198]. Another group
of agents including Exisulind and analogues that are inhibitors
of cyclic GMP phosphodiesterases (PDE) have been shown to
activate protein kinase G (PKG) that in turn decrease β‐catenin
levels via a novel GSK3β‐independent processing mechanism
[199]. ICG‐001, a small molecule known to inhibit β‐catenin’s
interaction with CREB‐binding protein (CBP) was shown to
affect β‐catenin‐TCF‐dependent target gene expression [200].
Its new generation analogue PRI‐724 is in clinical trial for vari-
ous malignancies that exhibit β‐catenin activation [201]. Using
computational chemoinformatics to identify another small mol-
ecule with structural similarity to ICG‐001, we identified a com-
pound we labeled as PMED‐1 [202]. This inhibitor showed in
vitro and in vivo efficacy against β‐catenin in HCC cells and in
zebrafish, respectively. Several of these inhibitors may have
potential therapeutic utility in the correct subset of HCC
patients. While GS as a biomarker of β‐catenin mutations is
helpful in identifying that subset, biopsies may not be feasible in
majority of HCC patients due to underlying liver disease and
cirrhosis. Thus, there is a need for secreted biomarkers to detect
β‐catenin mutations to select eligible patients for anti‐β‐catenin
therapies since β‐catenin is not a global therapeutic target in
HCC [203]. Due to this unmet need, we attempted and identi-
fied Lect2 in cell culture and in mice to be a faithful biomarker
in identifying activating β‐catenin mutations [204]. However,
serum Lect2 did not correspond to β‐catenin mutations in HCC
patients although levels greater than 50 ng ml−1 in patients had a
positive predictive value of 97% in detecting HCC [204].
Using SB‐HTVI to co‐express mutant β‐catenin and mutant
Kras, we showed HCC development in these mice that resem-
bled by gene expression the Met‐β‐catenin HCC, which in turn
showed around 70% resemblance by gene expression studies to
around 11% of all human HCC [171, 195]. We used these mod-
els to examine the effect of Met inhibition on HCC development
and showed a lack of any notable response using two different
modalities [128, 205]. On the other hand, use of a lipid nanopar-
ticle to deliver siRNA to β‐catenin showed a complete response
due to β‐catenin suppression which in turn affected cell prolif-
eration and survival [195]. Previous in vivo studies had also
shown β‐catenin suppression albeit using locked nucleic acid
antisense in a chemical carcinogenesis induced HCC model, to
also lead to complete response [206]. We believe β‐catenin to be
a viable therapeutic target in HCC, once patients are carefully
selected and verified especially for β‐catenin gene mutations.
Despite no FDA approved anti‐β‐catenin agents currently,
a  recent study demonstrates the role of mTOR inhibition in
β‐catenin‐mutated liver tumors [207]. This study showed locali-
zation of p‐mTOR‐S2448, marker of mTORC1 activation, in the
same cells as GS in a normal adult liver. This localization of GS
and p‐mTOR‐S2448 in zone three hepatocytes was absent in
knockouts of hepatocyte β‐catenin or Wnt co‐receptors LRP5/6.
Forced re‐expression of β‐catenin or GS led to re‐expression of
p‐mTOR‐S2448. More importantly, β‐catenin‐mutated HCC
that showed uniform positivity for GS, also showed positivity
for p‐mTOR‐S2448 in mice and in patients. Based on these
observations, Met‐β‐catenin HCC model showed a dramatic
response to mTOR inhibitor rapamycin. Additionally, with the
reports showing an efficacy of immune checkpoint inhibitors in
small subsets of HCC, a recent study showed exclusion of
immune cells from HCC that are mutated for β‐catenin gene,
making them unlikely candidates for these agents [208]. Thus
assessing β‐catenin mutation ­status to stratify cases for receiv-
ing specific agents or excluding them from receiving specific
drugs could form the basis of ­personalized medicine in HCC.
Since β‐catenin is also present at the cell surface in associa-
tion with E‐cadherin, a concern remains if β‐catenin suppres-
sion would lead to destabilization of the AJ and may inadvertently
promote tumor cell migration and metastasis. However, multi-
ple studies have now shown that β‐catenin knockdown is com-
pensated by increased γ‐catenin which maintains AJ through
association with E‐cadherin and actin cytoskeleton [109, 110].
Intriguingly, we did not identify any compensation by γ‐catenin
in the Wnt signaling upon β‐catenin loss. We also failed to
detect nuclear γ‐catenin in β‐catenin knockout mice even after
PH and via TopFlash (β‐catenin‐TCF) reporter assay in vitro.
Thus, the redundancy in catenins at the AJ but not in the Wnt
signaling is reassuring and makes β‐catenin a viable therapeutic
target in HCC. It will be critical to further address the mecha-
nisms whereby γ‐catenin is stabilized at AJ following β‐catenin
suppression [110].
WNT/β‐CATENIN SIGNALING IN BILE
DUCT AND GALL BLADDER TUMORS
The most common tumor that arises in the biliary tree is the
cholangiocarcinoma that can either originate from the intrahe-
patic portion, intrahepatic cholangiocarcinoma (ICC), or the
hilum (hilar cholangiocarcinoma) (reviewed in [206]). In chol-
angiocarcinoma, reduced expression of β‐catenin and E‐cad-
herin at the membrane is observed as compared to the
surrounding non‐cancerous ducts [208]. Nuclear localization of
β‐catenin is seen in a subset of tumors based on histology and
location of the tumor (reviewed in [210]). For most ICCs, aber-
rant nuclear localization is observed in around 15% of cases and
a decrease in membranous localization is related to poorer his-
tological differentiation [211]. No mutations in CTNNB1 were
identified although mutations in any other components of the
Wnt pathway were not assessed in this study. Our personal
experience in 62 ICCs showed only 2/62 tumors with nuclear
β‐catenin [162]. Another study detected exon 3 mutations in
7.5% of biliary tract cancer and in 57% of gallbladder adenomas
[213]. Higher frequency of mutations was seen in ampullary and
gallbladder carcinomas than the bile duct cancers. A higher cor-
relation of CTNBN1 mutations and papillary adenocarcinoma
was also observed. Intraductal papillary neoplasms also showed
anomalous nuclear localization of β‐catenin in around 25% of
patients without any exon‐3 CTNNB1 mutation [214]. Thus,
while the Wnt/β‐catenin pathway may be active in a subset of
biliary tract neoplasms, more studies are needed to comprehend
the mechanism of its observed deregulation.
598 THE LIVER:  REFERENCES

REFERENCES 26. Pinson, K.I., Brennan, J., Monkley, S., Avery, B.J., and Skarnes, W.C. An
LDL‐receptor‐related protein mediates Wnt signalling in mice. Nature,
2000;407(6803):535–8.
  1. Moon, R.T., Bowerman, B., Boutros, M., and Perrimon, N. The promise and
27. Tamai, K., Semenov, M., Kato, Y. et  al. LDL‐receptor‐related proteins in
perils of Wnt signaling through beta‐catenin. Science, 2002;296(5573):1644–6.
Wnt signal transduction. Nature, 2000;407(6803):530–5.
  2. Peifer, M. and Polakis, P. Wnt signaling in oncogenesis and embryogenesis –
28. Tamai, K., Zeng, X., Liu, C. et al. A mechanism for Wnt coreceptor activa-
a look outside the nucleus. Science, 2000;287(5458):1606–9.
tion. Mol Cell, 2004;13(1):149–56.
  3. Arias, A.M. Epithelial mesenchymal interactions in cancer and development.
29. MacDonald, B.T., Tamai, K., and He, X. Wnt/beta‐catenin signaling: compo-
Cell, 2001;105(4):425–31.
nents, mechanisms, and diseases. Dev Cell, 2009;17(1):9–26.
  4. Perrimon, N. and Mahowald, A.P. Multiple functions of segment polarity
30. Carmon, K.S., Gong, X., Lin, Q., Thomas, A., and Liu, Q. R‐spondins func-
genes in Drosophila. Dev Biol, 1987;119(2):587–600.
tion as ligands of the orphan receptors LGR4 and LGR5 to regulate Wnt/
  5. Babu, P. Early developmental subdivisions of the wing disk in Drosophila.
beta‐catenin signaling. Proc Natl Acad Sci USA, 2011;108(28):11452–7.
Mol Gen Genet, 1977;151(3):289–94.
31. Cadigan, K.M. and Nusse, R. Wnt signaling: a common theme in animal
  6. Sharma, R.P. and Chopra, V.L. Effect of the Wingless (wg1) mutation on
development. Genes Dev, 1997;11(24):3286–305.
wing and haltere development in Drosophila melanogaster. Dev Biol,
32. Mikels, A.J. and Nusse, R. Purified Wnt5a protein activates or inhibits beta‐
1976;48(2):461–5.
catenin‐TCF signaling depending on receptor context. PLoS Biol,
 7. Patel, N.H., Schafer, B., Goodman, C.S., and Holmgren, R. The role of
2006;4(4):e115.
­segment polarity genes during Drosophila neurogenesis. Genes Dev, 1989;
33. Oishi, I., Suzuki, H., Onishi, N. et al. The receptor tyrosine kinase Ror2 is
3(6):890–904.
involved in non‐canonical Wnt5a/JNK signalling pathway. Genes Cells,
  8. Peifer, M., Rauskolb, C., Williams, M., Riggleman, B., and Wieschaus, E.
2003;8(7):645–54.
The segment polarity gene armadillo interacts with the wingless signaling
34. Pez, F., Lopez, A., Kim, M., Wands, J.R., Caron de Fromentel, C.,and Merle,
pathway in both embryonic and adult pattern formation. Development,
P. Wnt signaling and hepatocarcinogenesis: molecular targets for the devel-
1991;111(4):1029–43.
opment of innovative anticancer drugs. J Hepatol, 2013;59(5):1107–17.
 9. Riggleman, B., Schedl, P., and Wieschaus, E. Spatial expression of the
35. Nishita, M., Enomoto, M., Yamagata, K., and Minami, Y. Cell/tissue‐tropic
Drosophila segment polarity gene armadillo is posttranscriptionally regu-
functions of Wnt5a signaling in normal and cancer cells. Trends Cell Biol,
lated by wingless. Cell, 1990;63(3):549–60.
2010;20(6):346–54.
10. Peifer, M., Orsulic, S., Pai, L.M., and Loureiro, J. A model system for
36. Acebron, S.P., Karaulanov, E., Berger, B.S., Huang, Y.L., and Niehrs, C.
cell  adhesion and signal transduction in Drosophila. Dev Suppl, 1993:
Mitotic wnt signaling promotes protein stabilization and regulates cell size.
163–76.
Mol Cell, 2014;54(4):663–74.
11. Clevers, H. and Nusse, R. Wnt/beta‐catenin signaling and disease. Cell,
37. Koch, S., Acebron, S.P., Herbst, J., Hatiboglu, G., and Niehrs, C. Post‐tran-
2012;149(6):1192–205.
scriptional Wnt signaling governs epididymal sperm maturation. Cell,
12. Nusse, R. and Clevers, H. Wnt/beta‐catenin signaling, disease, and emerging
2015;163(5):1225–36.
therapeutic modalities. Cell, 2017;169(6):985–99.
38. Vinyoles, M., Del Valle‐Perez, B., Curto, J. et  al. Multivesicular GSK3
13. Haegel, H., Larue, L., Ohsugi, M., Fedorov, L., Herrenknecht, K., and
sequestration upon Wnt signaling is controlled by p120‐catenin/cadherin
Kemler, R. Lack of beta‐catenin affects mouse development at gastrulation.
interaction with LRP5/6. Mol Cell, 2014;53(3):444–57.
Development, 1995;121(11):3529–37.
39. Acebron, S.P. and Niehrs, C. Beta‐catenin‐independent roles of Wnt/LRP6
14. Hari, L., Brault, V., Kleber, M. et al. Lineage‐specific requirements of beta‐
signaling. Trends Cell Biol, 2016;26(12):956–67.
catenin in neural crest development. J Cell Biol, 2002;159(5):867–80.
40. Inoki, K., Ouyang, H., Zhu, T. et al. TSC2 integrates Wnt and energy signals
15. Brault, V., Moore, R., Kutsch, S. et al. Inactivation of the beta‐catenin gene
via a coordinated phosphorylation by AMPK and GSK3 to regulate cell
by Wnt1‐Cre‐mediated deletion results in dramatic brain malformation and
growth. Cell, 2006;126(5):955–68.
failure of craniofacial development. Development, 2001;128(8):1253–64.
41. Knudsen, K.A., Soler, A.P., Johnson, K.R., and Wheelock, M.J. Interaction
16. Lickert, H., Kutsch, S., Kanzler, B., Tamai, Y., Taketo, M.M., and Kemler, R.
of alpha‐actinin with the cadherin/catenin cell‐cell adhesion complex via
Formation of multiple hearts in mice following deletion of beta‐catenin in
alpha‐catenin. J Cell Biol, 1995;130(1):67–77.
the embryonic endoderm. Dev Cell, 2002;3(2):171–81.
42. Huber, A.H. and Weis, W.I. The structure of the beta‐catenin/E‐cadherin
17. Mucenski, M.L., Wert, S.E., Nation, J.M. et al. beta‐Catenin is required for
complex and the molecular basis of diverse ligand recognition by beta‐
specification of proximal/distal cell fate during lung morphogenesis. J Biol
catenin. Cell, 2001;105(3):391–402.
Chem, 2003;278(41):40231–8.
43. Jou, T.S., Stewart, D.B., Stappert, J., Nelson, W.J., and Marrs, J.A. Genetic
18. Huelsken, J., Vogel, R., Erdmann, B., Cotsarelis, G., and Birchmeier, W.
and biochemical dissection of protein linkages in the cadherin‐catenin com-
beta‐Catenin controls hair follicle morphogenesis and stem cell differentia-
plex. Proc Natl Acad Sci USA, 1995;92(11):5067–71.
tion in the skin. Cell, 2001;105(4):533–45.
44. Lilien, J., Balsamo, J., Arregui, C., and Xu, G. Turn‐off, drop‐out: functional
19. Kispert, A., Vainio, S.,and McMahon, A.P. Wnt‐4 is a mesenchymal signal
state switching of cadherins. Dev Dyn, 2002;224(1):18–29.
for epithelial transformation of metanephric mesenchyme in the developing
45. Ozawa, M. and Kemler, R. Altered cell adhesion activity by pervanadate due
kidney. Development, 1998;125(21):4225–34.
to the dissociation of alpha‐catenin from the E‐cadherin.catenin complex. J
20. Behrens, J., Jerchow, B.A., Wurtele, M. et al. Functional interaction of an
Biol Chem, 1998;273(11):6166–70.
axin homolog, conductin, with beta‐catenin, APC, and GSK3beta. Science,
46. Roura, S., Miravet, S., Piedra, J., Garcia de Herreros, A., and Dunach, M.
1998;280(5363):596–9.
Regulation of E‐cadherin/catenin association by tyrosine phosphorylation. J
21. Amit, S., Hatzubai, A., Birman, Y. et al. Axin‐mediated CKI phosphorylation
Biol Chem, 1999;274(51):36734–40.
of beta‐catenin at Ser 45: a molecular switch for the Wnt pathway. Genes
47. Hu, P., O’Keefe, E.J., and Rubenstein, D.S. Tyrosine phosphorylation of human
Dev, 2002;16(9):1066–76.
keratinocyte beta‐catenin and plakoglobin reversibly regulates their binding to
22. Aberle, H., Bauer, A., Stappert, J., Kispert, A., and Kemler, R. Beta‐catenin
E‐cadherin and alpha‐catenin. J Invest Dermatol, 2001;117(5):1059–67.
is a target for the ubiquitin‐proteasome pathway. EMBO, J, 1997;
48. Piedra, J., Martinez, D., Castano, J., Miravet, S., Dunach, M., and de
16(13):3797–804.
Herreros, A.G. Regulation of beta‐catenin structure and activity by tyrosine
23. Barrott, J.J., Cash, G.M., Smith, A.P., Barrow, J.R., and Murtaugh, L.C.
phosphorylation. J Biol Chem, 2001;276(23):20436–43.
Deletion of mouse Porcn blocks Wnt ligand secretion and reveals an ectoder-
49. Rosato, R., Veltmaat, J.M., Groffen, J., and Heisterkamp, N. Involvement of
mal etiology of human focal dermal hypoplasia/Goltz syndrome. Proc Natl
the tyrosine kinase fer in cell adhesion. Mol Cell Biol, 1998;18(10):5762–70.
Acad Sci USA, 2011;108(31):12752–7.
50. Behrens, J., Vakaet, L., Friis, R. et al. Loss of epithelial differentiation and
24. Bartscherer, K., Pelte, N., Ingelfinger, D., and Boutros, M. Secretion of Wnt
gain of invasiveness correlates with tyrosine phosphorylation of the E‐cad-
ligands requires Evi, a conserved transmembrane protein. Cell,
herin/beta‐catenin complex in cells transformed with a temperature‐sensitive
2006;125(3):523–33.
v‐SRC gene. J Cell Biol, 1993;120(3):757–66.
25. Bhanot, P., Brink, M., Samos, C.H. et al. A new member of the frizzled fam-
51. Shibamoto, S., Hayakawa, M., Takeuchi, K. et al. Tyrosine phosphorylation
ily from Drosophila functions as a Wingless receptor. Nature, 1996;
of beta‐catenin and plakoglobin enhanced by hepatocyte growth factor and
382(6588):225–30.
 β-CATENIN SIGNALING
46:  599
epidermal growth factor in human carcinoma cells. Cell Adhes Commun,
1994;1(4):295–305.
52. Hoschuetzky, H., Aberle, H., and Kemler, R. Beta‐catenin mediates the inter-
action of the cadherin‐catenin complex with epidermal growth factor recep-
tor. J Cell Biol, 1994;127(5):1375–80.
53. Kanai, Y., Ochiai, A., Shibata, T. et al. c‐erbB‐2 gene product directly associ-
ates with beta‐catenin and plakoglobin. Biochem Biophys Res Commun,
1995;208(3):1067–72.
54. Takahashi, K., Suzuki, K.,and Tsukatani, Y. Induction of tyrosine phospho-
rylation and association of beta‐catenin with EGF receptor upon tryptic
digestion of quiescent cells at confluence. Oncogene, 1997;15(1):71–8.
55. Birchmeier, C., Birchmeier, W., Gherardi, E., and Vande Woude, G.F. Met,
metastasis, motility and more. Nat Rev Mol Cell Biol, 2003;4(12):915–25.
56. Monga, S.P., Mars, W.M., Pediaditakis, P. et  al. Hepatocyte growth factor
induces Wnt‐independent nuclear translocation of beta‐catenin after Met‐
beta‐catenin dissociation in hepatocytes. Cancer Res, 2002;62(7):2064–71.
57. Zeng, G., Apte, U., Micsenyi, A., Bell, A., and Monga, S.P. Tyrosine residues
654 and 670 in beta‐catenin are crucial in regulation of Met‐beta‐catenin
interactions. Exp Cell Res, 2006;312(18):3620–30.
58. Hiscox, S. and Jiang, W.G. Association of the HGF/SF receptor, c‐met, with
the cell‐surface adhesion molecule, E‐cadherin, and catenins in human
tumor cells. Biochem Biophys Res Commun, 1999;261(2):406–11.
59. Papkoff, J. and Aikawa, M. WNT‐1 and HGF regulate GSK3 beta activity
and beta‐catenin signaling in mammary epithelial cells. Biochem Biophys
Res Commun, 1998;247(3):851–8.
60. Danilkovitch‐Miagkova, A., Miagkov, A., Skeel, A., Nakaigawa, N., Zbar,
B., and Leonard, E.J. Oncogenic mutants of RON and MET receptor tyrosine
kinases cause activation of the beta‐catenin pathway. Mol Cell Biol,
2001;21(17):5857–68.
61. Shiota, G., Umeki, K., Okano, J., and Kawasaki, H. Hepatocyte growth fac-
tor and acute phase proteins in patients with chronic liver diseases. J Med,
1995;26(5–6):295–308.
62. Yamagami, H., Moriyama, M., Tanaka, N., and Arakawa, Y. Detection of
serum and intrahepatic human hepatocyte growth factor in patients with type
C liver diseases. Intervirology, 2001;44(1):36–42.
63. Bonvini, P., An, W.G., Rosolen, A. et  al. Geldanamycin abrogates ErbB2
association with proteasome‐resistant beta‐catenin in melanoma cells,
increases beta‐catenin‐E‐cadherin association, and decreases beta‐catenin‐
sensitive transcription. Cancer Res, 2001;61(4):1671–7.
64. Taurin, S., Sandbo, N., Qin, Y., Browning, D., and Dulin, N.O.
Phosphorylation of beta‐catenin by cyclic AMP‐dependent protein kinase. J
Biol Chem, 2006;281(15):9971–6.
65. Alvarado, T.F., Puliga, E., Preziosi, M. et al. Thyroid hormone receptor beta
agonist induces beta‐catenin‐dependent hepatocyte proliferation in mice:
implications in hepatic regeneration. Gene Expr, 2016;17(1):19–34.
66. Fanti, M., Singh, S., Ledda‐Columbano, G.M., Columbano, A., and Monga,
S.P. Tri‐iodothyronine induces hepatocyte proliferation by protein kinase A‐
dependent beta‐catenin activation in rodents. Hepatology,
2014;59(6):2309–20.
67. Spirli, C., Locatelli, L., Morell, C.M. et  al. Protein kinase A‐dependent
pSer(675) ‐beta‐catenin, a novel signaling defect in a mouse model of con-
genital hepatic fibrosis. Hepatology, 2013;58(5):1713–23.
68. Deng, J., Miller, S.A., Wang, H.Y. et  al. beta‐catenin interacts with and
inhibits NF‐kappa B in human colon and breast cancer. Cancer Cell,
2002;2(4):323–34.
69. Nejak‐Bowen, K., Kikuchi, A., and Monga, S.P. Beta‐catenin‐NF‐kappaB
interactions in murine hepatocytes: a complex to die for. Hepatology,
2013;57(2):763–74.
70. Nejak‐Bowen, K., Moghe, A., Cornuet, P., Preziosi, M., Nagarajan, S., and
Monga, S.P. Role and regulation of p65/beta‐catenin association during liver
injury and regeneration: a “complex” relationship. Gene Expr,
2017;17(3):219–35.
71. Du, Q., Zhang, X., Cardinal, J. et al. Wnt/beta‐catenin signaling regulates
cytokine‐induced human inducible nitric oxide synthase expression by
inhibiting nuclear factor‐kappaB activation in cancer cells. Cancer Res,
2009;69(9):3764–71.
72. Thompson, M.D., Moghe, A., Cornuet, P. et al. beta‐Catenin regulation of
farnesoid X receptor signaling and bile acid metabolism during murine chol-
estasis. Hepatology, 2018;67(3):955–71.
73. Monga, S.P., Monga, H.K., Tan, X., Mule, K., Pediaditakis, P., and
Michalopoulos, G.K. Beta‐catenin antisense studies in embryonic liver cul-
tures: role in proliferation, apoptosis, and lineage specification.
Gastroenterology, 2003;124(1):202–16.
74. Micsenyi, A., Tan, X., Sneddon, T., Luo, J.H., Michalopoulos, G.K., and
Monga, S.P. Beta‐catenin is temporally regulated during normal liver devel-
opment. Gastroenterology, 2004;126(4):1134–46.
75. Nejak‐Bowen, K. and Monga, S.P. Wnt/beta‐catenin signaling in hepatic
organogenesis. Organogenesis, 2008;4(2):92–9.
76. McLin, V.A., Rankin, S.A., and Zorn, A.M. Repression of Wnt/beta‐catenin
signaling in the anterior endoderm is essential for liver and pancreas devel-
opment. Development, 2007;134(12):2207–17.
77. Li, Y., Rankin, S.A., Sinner, D., Kenny, A.P., Krieg, P.A., and Zorn, A.M. Sfrp5
coordinates foregut specification and morphogenesis by antagonizing both canon-
ical and noncanonical Wnt11 signaling. Genes Dev, 2008;22(21):3050–63.
78. Goessling, W., North, T.E., Lord, A.M. et al. APC mutant zebrafish uncover
a changing temporal requirement for wnt signaling in liver development. Dev
Biol, 2008;320(1):161–74.
79. Ober, E.A., Verkade, H., Field, H.A., and Stainier, D.Y. Mesodermal Wnt2b sig-
nalling positively regulates liver specification. Nature, 2006;442(7103):688–91.
80. Poulain, M. and Ober, E.A. Interplay between Wnt2 and Wnt2bb controls
multiple steps of early foregut‐derived organ development. Development,
2011;138(16):3557–68.
81. Shin, D., Lee, Y., Poss, K.D., and Stainier, D.Y.R. Restriction of hepatic
competence by Fgf signaling. Development, 2011;138(7):1339–48.
82. So, J., Martin, B.L., Kimelman, D., and Shin, D. Wnt/beta‐catenin signaling
cell‐autonomously converts non‐hepatic endodermal cells to a liver fate. Biol
Open, 2013;2(1):30–6.
83. Tan, X., Yuan, Y., Zeng, G. et al. Beta‐catenin deletion in hepatoblasts dis-
rupts hepatic morphogenesis and survival during mouse development.
Hepatology, 2008;47(5):1667–79.
84. Suksaweang, S., Lin, C.M., Jiang, T.X., Hughes, M.W., Widelitz, R.B., and
Chuong, C.M. Morphogenesis of chicken liver: identification of localized
growth zones and the role of beta‐catenin/Wnt in size regulation. Dev Biol,
2004;266(1):109–22.
85. Hussain, S.Z., Sneddon, T., Tan, X., Micsenyi, A., Michalopoulos, G.K., and
Monga, S.P. Wnt impacts growth and differentiation in ex vivo liver develop-
ment. Exp Cell Res, 2004;292(1):157–69.
86. Decaens, T., Godard, C., de Reynies, A. et al. Stabilization of beta‐catenin
affects mouse embryonic liver growth and hepatoblast fate. Hepatology,
2008;47(1):247–58.
87. Cordi, S., Godard, C., Saandi, T. et al. Role of beta‐catenin in development
of bile ducts. Differentiation, 2016;91(1–3):42–9.
88. So, J., Khaliq, M., Evason, K. et  al. Wnt/beta‐catenin signaling controls
intrahepatic biliary network formation in zebrafish by regulating notch activ-
ity. Hepatology, 2018;67(6):2352–66.
89. Kiyohashi, K., Kakinuma, S., Kamiya, A. et al. Wnt5a signaling mediates
biliary differentiation of fetal hepatic stem/progenitor cells in mice.
Hepatology, 2013;57(6):2502–13.
90. Saneyoshi, T., Kume, S., Amasaki, Y., and Mikoshiba, K. The Wnt/calcium
pathway activates NF‐AT and promotes ventral cell fate in Xenopus embryos.
Nature, 2002;417(6886):295–9.
91. Matsumoto, K., Miki, R., Nakayama, M., Tatsumi, N., and Yokouchi, Y.
Wnt9a secreted from the walls of hepatic sinusoids is essential for morpho-
genesis, proliferation, and glycogen accumulation of chick hepatic epithe-
lium. Dev Biol, 2008;319(2):234–47.
92. Apte, U., Zeng, G., Muller, P. et al. Activation of Wnt/beta‐catenin pathway
during hepatocyte growth factor‐induced hepatomegaly in mice. Hepatology,
2006;44(4):992–1002.
93. Schmidt, C., Bladt, F., Goedecke, S. et al. Scatter factor/hepatocyte growth
factor is essential for liver development. Nature, 1995;373(6516):699–702.
94. Uehara, Y., Minowa, O., Mori, C. et  al. Placental defect and embryonic
lethality in mice lacking hepatocyte growth factor/scatter factor. Nature,
1995;373(6516):702–5.
95. Berg, T., Rountree, C.B., Lee, L. et al. Fibroblast growth factor 10 is critical
for liver growth during embryogenesis and controls hepatoblast survival via
beta‐catenin activation. Hepatology, 2007;46(4):1187–97.
96. Sekhon, S.S., Tan, X., Micsenyi, A., Bowen, W.C., and Monga, S.P.
Fibroblast growth factor enriches the embryonic liver cultures for hepatic
progenitors. Am J Pathol, 2004;164(6):2229–40.
97. Gebhardt, R. and Mecke, D. Heterogeneous distribution of glutamine syn-
thetase among rat liver parenchymal cells in situ and in primary culture.
EMBO J, 1983;2(4):567–70.
600 THE LIVER:  REFERENCES
  98. Benhamouche, S., Decaens, T., Godard, C. et al. Apc tumor suppressor gene
is the “zonation‐keeper” of mouse liver. Dev Cell, 2006;10(6):759–70.
  99. Cadoret, A., Ovejero, C., Terris, B. et al. New targets of beta‐catenin signal-
ing in the liver are involved in the glutamine metabolism. Oncogene,
2002;21(54):8293–301.
100. Sekine, S., Lan, B.Y., Bedolli, M., Feng, S., and Hebrok, M. Liver‐specific
loss of beta‐catenin blocks glutamine synthesis pathway activity and
cytochrome p450 expression in mice. Hepatology, 2006;43(4):817–25.
101. Tan, X., Behari, J., Cieply, B., Michalopoulos, G.K., and Monga, S.P.
Conditional deletion of beta‐catenin reveals its role in liver growth and
regeneration. Gastroenterology, 2006;131(5):1561–72.
102. Wang, B., Zhao, L., Fish, M., Logan, C.Y., and Nusse, R. Self‐renewing
diploid Axin2(+) cells fuel homeostatic renewal of the liver. Nature,
2015;524(7564):180–5.
103. Yang, J., Mowry, L.E., Nejak‐Bowen, K.N. et al. Beta‐catenin signaling in
murine liver zonation and regeneration: a Wnt‐Wnt situation! Hepatology,
2014;60(3):964–76.
104. Leibing, T., Geraud, C., Augustin, I. et al. Angiocrine Wnt signaling con-
trols liver growth and metabolic maturation in mice. Hepatology,
2018;68(2):707–22.
105. Preziosi, M., Okabe, H., Poddar, M., Singh, S., and Monga, S.P. Endothelial
Wnts regulate β‐catenin signaling in murine liver zonation and regenera-
tion: a sequel to the Wnt–Wnt situation. Hepatol Commun, 2018;2(7).
106. Planas‐Paz, L., Orsini, V., Boulter, L. et  al. The RSPO‐LGR4/5‐ZNRF3/
RNF43 module controls liver zonation and size. Nat Cell Biol,
2016;18(5):467–79.
107. Nejak‐Bowen, K.N., Zeng, G., Tan, X., Cieply, B., and Monga, S.P. Beta‐
catenin regulates vitamin C biosynthesis and cell survival in murine liver. J
Biol Chem, 2009;284(41):28115–27.
108. Sambrotta, M., Strautnieks, S., Papouli, E. et al. Mutations in TJP2 cause
progressive cholestatic liver disease. Nat Genet, 2014;46(4):326–8.
109. Wickline, E.D., Awuah, P.K., Behari, J., Ross, M., Stolz, D.B., and Monga,
S.P. Hepatocyte gamma‐catenin compensates for conditionally deleted
beta‐catenin at adherens junctions. J Hepatol, 2011;55(6):1256–62.
110. Wickline, E.D., Du, Y., Stolz, D.B., Kahn, M., and Monga, S.P. Gamma‐catenin
at adherens junctions: mechanism and biologic implications in hepatocellular
cancer after beta‐catenin knockdown. Neoplasia, 2013;15(4):421–34.
111. Mankertz, J., Hillenbrand, B., Tavalali, S., Huber, O., Fromm, M., and
Schulzke, J.D. Functional crosstalk between Wnt signaling and Cdx‐related
transcriptional activation in the regulation of the claudin‐2 promoter activ-
ity. Biochem Biophys Res Commun, 2004;314(4):1001–7.
112. Yeh, T.H., Krauland, L., Singh, V. et al. Liver‐specific beta‐catenin knock-
out mice have bile canalicular abnormalities, bile secretory defect, and
intrahepatic cholestasis. Hepatology, 2010;52(4):1410–9.
113. Pradhan‐Sundd, T., Zhou, L., Vats, R. et  al. Dual catenin loss in murine
liver causes tight junctional deregulation and progressive intrahepatic chol-
estasis. Hepatology, 2018;67(6):2320–37.
114. Vitale, G., Gitto, S., Raimondi, F. et al. Cryptogenic cholestasis in young
and adults: ATP8B1, ABCB11, ABCB4, and TJP2 gene variants analysis by
high‐throughput sequencing. J Gastroenterol, 2018;53(8):945–58.
115. Apte, U., Zeng, G., Thompson, M.D. et al. Beta‐catenin is critical for early
postnatal liver growth. Am J Physiol Gastrointest Liver Physiol,
2007;292(6):G1578–85.
116. Cadoret, A., Ovejero, C., Saadi‐Kheddouci, S. et al. Hepatomegaly in trans-
genic mice expressing an oncogenic form of beta‐catenin. Cancer Res,
2001;61(8):3245–9.
117. Colnot, S., Decaens, T., Niwa‐Kawakita, M. et al. Liver‐targeted disruption
of Apc in mice activates beta‐catenin signaling and leads to hepatocellular
carcinomas. Proc Natl Acad Sci USA, 2004;101(49):17216–21.
118. Harada, N., Miyoshi, H., Murai, N. et  al. Lack of tumorigenesis in the
mouse liver after adenovirus‐mediated expression of a dominant stable
mutant of beta‐catenin. Cancer Res, 2002;62(7):1971–7.
119. Tan, X., Apte, U., Micsenyi, A. et al. Epidermal growth factor receptor: a
novel target of the Wnt/beta‐catenin pathway in liver. Gastroenterology,
2005;129(1):285–302.
120. Nejak‐Bowen, K.N., Thompson, M.D., Singh, S. et  al. Accelerated liver
regeneration and hepatocarcinogenesis in mice overexpressing serine‐45
mutant beta‐catenin. Hepatology, 2010;51(5):1603–13.
121. Tetsu, O. and McCormick, F. Beta‐catenin regulates expression of cyclin
D1 in colon carcinoma cells. Nature, 1999;398(6726):422–6.
122. Monga, S.P., Pediaditakis, P., Mule, K., Stolz, D.B., and Michalopoulos,
G.K. Changes in WNT/beta‐catenin pathway during regulated growth in rat
liver regeneration. Hepatology, 2001;33(5):1098–109.
123. Sodhi, D., Micsenyi, A., Bowen, W.C., Monga, D.K., Talavera, J.C.,
and  Monga, S.P. Morpholino oligonucleotide‐triggered beta‐catenin
knockdown compromises normal liver regeneration. J Hepatol, 2005;
­
43(1):132–41.
124. Sekine, S., Gutierrez, P.J., Lan, B.Y., Feng, S., and Hebrok, M. Liver‐­
specific loss of beta‐catenin results in delayed hepatocyte proliferation after
partial hepatectomy. Hepatology, 2007;45(2):361–8.
125. Michalopoulos, G.K. Principles of liver regeneration and growth homeosta-
sis. Compr Physiol, 2013;3(1):485–513.
126. Yang, J., Cusimano, A., Monga, J.K. et al. WNT5A inhibits hepatocyte pro-
liferation and concludes beta‐catenin signaling in liver regeneration. Am J
Pathol, 2015;185(8):2194–205.
127. Apte, U., Singh, S., Zeng, G. et al. Beta‐catenin activation promotes liver
regeneration after acetaminophen‐induced injury. Am J Pathol, 2009;
175(3):1056–65.
128. Puliga, E., Min, Q., Tao, J. et al. Thyroid hormone receptor‐beta agonist
GC‐1 inhibits Met‐beta‐catenin‐driven hepatocellular cancer. Am J Pathol,
2017;187(11):2473–85.
129. Gebhardt, R. Speeding up hepatocyte proliferation: how triiodothyronine
and beta‐catenin join forces. Hepatology, 2014;59(6):2074–6.
130. Gebhardt, R. Metabolic zonation of the liver: regulation and implications
for liver function. Pharmacol Ther, 1992;53(3):275–354.
131. Audard, V., Grimber, G., Elie, C. et al. Cholestasis is a marker for hepato-
cellular carcinomas displaying beta‐catenin mutations. J Pathol, 2007;
212(3):345–52.
132. Behari, J., Yeh, T.H., Krauland, L. et al. Liver‐specific beta‐catenin knock-
out mice exhibit defective bile acid and cholesterol homeostasis and
increased susceptibility to diet‐induced steatohepatitis. Am J Pathol,
2010;176(2):744–53.
133. Gougelet, A., Torre, C., Veber, P. et al. T‐cell factor 4 and beta‐catenin chro-
matin occupancies pattern zonal liver metabolism in mice. Hepatology,
2014;59(6):2344–57.
134. Tsuchida, T. and Friedman, S.L. Mechanisms of hepatic stellate cell activa-
tion. Nat Rev Gastroenterol Hepatol, 2017;14(7):397–411.
135. Shackel, N.A., McGuinness, P.H., Abbott, C.A., Gorrell, M.D., and
McCaughan, G.W. Identification of novel molecules and pathogenic path-
ways in primary biliary cirrhosis: cDNA array analysis of intrahepatic
­differential gene expression. Gut, 2001;49(4):565–76.
136. Tanaka, A., Leung, P.S., Kenny, T.P. et al. Genomic analysis of differen-
tially expressed genes in liver and biliary epithelial cells of patients with
primary biliary cirrhosis. J Autoimmun, 2001;17(1):89–98.
137. Jiang, F., Parsons, C.J., and Stefanovic, B. Gene expression profile of qui-
escent and activated rat hepatic stellate cells implicates Wnt signaling path-
way in activation. J Hepatol, 2006;45(3):401–9.
138. Rashid, S.T., Humphries, J.D., Byron, A. et al. Proteomic analysis of extra-
cellular matrix from the hepatic stellate cell line LX‐2 identifies CYR61
and Wnt‐5a as novel constituents of fibrotic liver. J Proteome Res,
2012;11(8):4052–64.
139. Xiong, W.J., Hu, L.J., Jian, Y.C. et al. Wnt5a participates in hepatic stellate
cell activation observed by gene expression profile and functional assays.
World J Gastroenterol, 2012;18(15):1745–52.
140. Kordes, C., Sawitza, I., and Haussinger, D. Canonical Wnt signaling main-
tains the quiescent stage of hepatic stellate cells. Biochem Biophys Res
Commun, 2008;367(1):116–23.
141. Ge, W.S., Wang, Y.J., Wu, J.X., Fan, J.G., Chen, Y.W., and Zhu, L. beta‐
catenin is overexpressed in hepatic fibrosis and blockage of Wnt/­
beta‐catenin signaling inhibits hepatic stellate cell activation. Mol Med Rep,
2014;9(6):2145–51.
142. Cheng, J.H., She, H., Han, Y.P. et al. Wnt antagonism inhibits hepatic stel-
late cell activation and liver fibrosis. Am J Physiol Gastrointest Liver
Physiol, 2008;294(1):G39–49.
143. Myung, S.J., Yoon, J.H., Gwak, G.Y. et al. Wnt signaling enhances the acti-
vation and survival of human hepatic stellate cells. FEBS Lett,
2007;581(16):2954–8.
144. Haughton, E.L., Tucker, S.J., Marek, C.J. et al. Pregnane X receptor activa-
tors inhibit human hepatic stellate cell transdifferentiation in vitro.
Gastroenterology, 2006;131(1):194–209.
 β-CATENIN SIGNALING
46:  601
145. Zhu, N.L., Wang, J., and Tsukamoto, H. The Necdin‐Wnt pathway causes
epigenetic peroxisome proliferator‐activated receptor gamma repression in
hepatic stellate cells. J Biol Chem, 2010;285(40):30463–71.
146. Yanagida, A., Iwaisako, K., Hatano, E. et al. Downregulation of the Wnt
antagonist Dkk2 links the loss of Sept4 and myofibroblastic transformation
of hepatic stellate cells. Biochim Biophys Acta, 2011;1812(11):1403–11.
147. Li, W., Zhu, C., Li, Y., Wu, Q., and Gao, R. Mest attenuates CCl4‐induced
liver fibrosis in rats by inhibiting the Wnt/beta‐catenin signaling pathway.
Gut Liver, 2014;8(3):282–91.
148. Ross, S.E., Hemati, N., Longo, K.A. et al. Inhibition of adipogenesis by
Wnt signaling. Science, 2000;289(5481):950–3.
149. Kweon, S.M., Chi, F., Higashiyama, R., Lai, K., and Tsukamoto, H. Wnt
pathway stabilizes MeCP2 protein to repress PPAR‐gamma in activation of
hepatic stellate cells. PLoS One, 2016;11(5):e0156111.
150. Lai, K.K.Y., Kweon, S.M., Chi, F. et al. Stearoyl‐CoA desaturase promotes
liver fibrosis and tumor development in mice via a Wnt positive‐signaling
loop by stabilization of low‐density lipoprotein‐receptor‐related proteins 5
and 6. Gastroenterology, 2017;152(6):1477–91.
151. Hughes, L.J. and Michels, V.V. Risk of hepatoblastoma in familial adeno-
matous polyposis. Am J Med Genet, 1992;43(6):1023–5.
152. Kurahashi, H., Takami, K., Oue, T. et al. Biallelic inactivation of the APC
gene in hepatoblastoma. Cancer Res, 1995;55(21):5007–11.
153. Oda, H., Imai, Y., Nakatsuru, Y., Hata, J., and Ishikawa, T. Somatic muta-
tions of the APC gene in sporadic hepatoblastomas. Cancer Res,
1996;56(14):3320–3.
154. Koch, A., Denkhaus, D., Albrecht, S., Leuschner, I., von Schweinitz, D., and
Pietsch, T. Childhood hepatoblastomas frequently carry a mutated degrada-
tion targeting box of the beta‐catenin gene. Cancer Res, 1999;59(2):269–73.
155. Taniguchi, K., Roberts, L.R., Aderca, I.N. et  al. Mutational spectrum of
beta‐catenin, AXIN1, and AXIN2 in hepatocellular carcinomas and hepato-
blastomas. Oncogene, 2002;21(31):4863–71.
156. Cairo, S., Armengol, C., De Reynies, A. et al. Hepatic stem‐like phenotype
and interplay of Wnt/beta‐catenin and Myc signaling in aggressive child-
hood liver cancer. Cancer Cell, 2008;14(6):471–84.
157. Bell, D., Ranganathan, S., Tao, J., and Monga, S.P. Novel advances in
understanding of molecular pathogenesis of hepatoblastoma: a Wnt/
beta‐catenin perspective. Gene Expr, 2017;17(2):141–54.
158. Lade, A.G. and Monga, S.P. Beta‐catenin signaling in hepatic development
and progenitors: which way does the WNT blow? Dev Dyn, 2011;240(3):
486–500.
159. Shin, D. and Monga, S.P. Cellular and molecular basis of liver develop-
ment. Compr Physiol, 2013;3(2):799–815.
160. Lopez‐Terrada, D., Gunaratne, P.H., Adesina, A.M. et al. Histologic sub-
types of hepatoblastoma are characterized by differential canonical Wnt
and Notch pathway activation in DLK+ precursors. Hum Pathol,
2009;40(6):783–94.
161. Armengol, C., Cairo, S., Fabre, M., and Buendia, M.A. Wnt signaling and
hepatocarcinogenesis: the hepatoblastoma model. Int J Biochem Cell Biol,
2011;43(2):265–70.
162. Tao, J., Calvisi, D.F., Ranganathan, S. et al. Activation of beta‐catenin and
Yap1 in human hepatoblastoma and induction of hepatocarcinogenesis in
mice. Gastroenterology, 2014;147(3):690–701.
163. Wang, H., Lu, J., Edmunds, L.R. et al. Coordinated activities of multiple
Myc‐dependent and Myc‐independent biosynthetic pathways in hepato-
blastoma. J Biol Chem, 2016;291(51):26241–51.
164. Molina, L., Bell, D., Tao, J. et al. Hepatocyte‐derived lipocalin 2 is a poten-
tial serum biomarker reflecting tumor burden in hepatoblastoma. Am J
Pathol, 2018;188(8):1895–909.
165. Nault, J.C., Bioulac‐Sage, P., and Zucman‐Rossi, J. Hepatocellular benign
tumors‐from molecular classification to personalized clinical care.
Gastroenterology, 2013;144(5):888–902.
166. Zucman‐Rossi, J., Jeannot, E., Nhieu, J.T. et al. Genotype‐phenotype cor-
relation in hepatocellular adenoma: new classification and relationship with
HCC. Hepatology, 2006;43(3):515–24.
167. Bioulac‐Sage, P., Rebouissou, S., Thomas, C. et  al. Hepatocellular ade-
noma subtype classification using molecular markers and immunohisto-
chemistry. Hepatology, 2007;46(3):740–8.
168. Ihara, A., Koizumi, H., Hashizume, R., and Uchikoshi, T. Expression of
epithelial cadherin and alpha‐ and beta‐catenins in nontumoral livers and
hepatocellular carcinomas. Hepatology, 1996;23(6):1441–7.
169. de La Coste, A., Romagnolo, B., Billuart, P. et al. Somatic mutations of the
beta‐catenin gene are frequent in mouse and human hepatocellular carcino-
mas. Proc Natl Acad Sci USA, 1998;95(15):8847–51.
170. Schulze, K., Imbeaud, S., Letouze, E. et al. Exome sequencing of hepato-
cellular carcinomas identifies new mutational signatures and potential
therapeutic targets. Nat Genet, 2015;47(5):505–11.
171. Tao, J., Xu, E., Zhao, Y. et al. Modeling a human hepatocellular carcinoma
subset in mice through coexpression of met and point‐mutant beta‐catenin.
Hepatology, 2016;64(5):1587–605.
172. Torrecilla, S., Sia, D., and Harrington, A.N. Trunk mutational events pre-
sent minimal intra‐ and inter‐tumoral heterogeneity in hepatocellular carci-
noma. J Hepatol, 2017;67(6):1222–31.
173. Satoh, S., Daigo, Y., Furukawa, Y. et al. AXIN1 mutations in hepatocellular
carcinomas, and growth suppression in cancer cells by virus‐mediated
transfer of AXIN1. Nat Genet, 2000;24(3):245–50.
174. Zucman‐Rossi, J., Benhamouche, S., Godard, C. et al. Differential effects
of inactivated Axin1 and activated beta‐catenin mutations in human hepato-
cellular carcinomas. Oncogene, 2007;26(5):774–80.
175. Merle, P., de la Monte, S., Kim, M. et al. Functional consequences of friz-
zled‐7 receptor overexpression in human hepatocellular carcinoma.
Gastroenterology, 2004;127(4):1110–22.
176. Nambotin, S.B., Lefrancois, L., Sainsily, X. et al. Pharmacological inhibi-
tion of Frizzled‐7 displays anti‐tumor properties in hepatocellular carci-
noma. J Hepatol, 2010.
177. Kim, M., Lee, H.C., Tsedensodnom, O. et al. Functional interaction between
Wnt3 and Frizzled‐7 leads to activation of the Wnt/beta‐catenin signaling
pathway in hepatocellular carcinoma cells. J Hepatol, 2008;48(5):780–91.
178. Ban, K.C., Singh, H., Krishnan, R., and Seow, H.F. GSK‐3beta phospho-
rylation and alteration of beta‐catenin in hepatocellular carcinoma. Cancer
Lett, 2003;199(2):201–8.
179. Shih, Y.L., Shyu, R.Y., Hsieh, C.B. et  al. Promoter methylation of the
secreted frizzled‐related protein 1 gene SFRP1 is frequent in hepatocellular
carcinoma. Cancer, 2006;107(3):579–90.
180. Takagi, H., Sasaki, S., Suzuki, H. et al. Frequent epigenetic inactivation of
SFRP genes in hepatocellular carcinoma. J Gastroenterol, 2008;43(5):
378–89.
181. Hoshida, Y., Nijman, S.M., Kobayashi, M. et al. Integrative transcriptome
analysis reveals common molecular subclasses of human hepatocellular
carcinoma. Cancer Res, 2009;69(18):7385–92.
182. Cieply, B., Zeng, G., Proverbs‐Singh, T., Geller, D.A., and Monga, S.P.
Unique phenotype of hepatocellular cancers with exon‐3 mutations in beta‐
catenin gene. Hepatology, 2009;49(3):821–31.
183. Boyault, S., Rickman, D.S., de Reynies, A. et al. Transcriptome classifica-
tion of HCC is related to gene alterations and to new therapeutic targets.
Hepatology, 2007;45(1):42–52.
184. Rebouissou, S., Franconi, A., Calderaro, J. et al. Genotype‐phenotype cor-
relation of CTNNB1 mutations reveals different ss‐catenin activity associ-
ated with liver tumor progression. Hepatology, 2016;64(6):2047–61.
185. Yuan, R.H., Jeng, Y.M., Hu, R.H. et al. Role of p53 and beta‐catenin muta-
tions in conjunction with CK19 expression on early tumor recurrence and
prognosis of hepatocellular carcinoma. J Gastrointest Surg, 2011;15(2):
321–9.
186. Nhieu, J.T., Renard, C.A., Wei, Y., Cherqui, D., Zafrani, E.S., and Buendia,
M.A. Nuclear accumulation of mutated beta‐catenin in hepatocellular car-
cinoma is associated with increased cell proliferation. Am J Pathol,
1999;155(3):703–10.
187. Suzuki, T., Yano, H., Nakashima, Y., Nakashima, O., and Kojiro, M. Beta‐
catenin expression in hepatocellular carcinoma: a possible participation of
beta‐catenin in the dedifferentiation process. J Gastroenterol Hepatol,
2002;17(9):994–1000.
188. Dal Bello, B., Rosa, L., Campanini, N. et al. Glutamine synthetase immu-
nostaining correlates with pathologic features of hepatocellular carcinoma
and better survival after radiofrequency thermal ablation. Clin Cancer
Res,Cancer Res, 2010;16(7):2157–66.
189. Lok, A.S., Seeff, L.B., Morgan, T.R. et al. Incidence of hepatocellular car-
cinoma and associated risk factors in hepatitis C‐related advanced liver
disease. Gastroenterology, 2009;136(1):138–48.
190. Lee, J.M., Yang, J., Newell, P. et al. beta‐Catenin signaling in hepatocellu-
lar cancer: Implications in inflammation, fibrosis, and proliferation. Cancer
Lett, 2014;343(1):90–7.
602 THE LIVER:  REFERENCES
191. Harada, N., Oshima, H., Katoh, M., Tamai, Y., Oshima, M., and Taketo,
M.M. Hepatocarcinogenesis in mice with beta‐catenin and Ha‐ras gene
mutations. Cancer Res, 2004;64(1):48–54.
192. Miyoshi, H., Deguchi, A., Nakau, M. et  al. Hepatocellular carcinoma
development induced by conditional beta‐catenin activation in Lkb1+/‐
mice. Cancer Sci, 2009;100(11):2046–53.
193. Bauer‐Hofmann, R., Klimek, F., Buchmann, A., Muller, O., Bannasch, P.,
and Schwarz, M. Role of mutations at codon 61 of the c‐Ha‐ras gene during
diethylnitrosamine‐induced hepatocarcinogenesis in C3H/He mice. Mol
Carcinog, 1992;6(1):60–7.
194. Nault, J.C., Mallet, M., Pilati, C. et  al. High frequency of telomerase
reverse‐transcriptase promoter somatic mutations in hepatocellular carci-
noma and preneoplastic lesions. Nat Commun, 2013;4:2218.
195. Tao, J., Zhang, R., Singh, S. et al. Targeting beta‐catenin in hepatocellular
cancers induced by coexpression of mutant beta‐catenin and K‐Ras in mice.
Hepatology, 2017;65(5):1581–99.
196. Zeng, G., Apte, U., Cieply, B., Singh, S., and Monga, S.P. siRNA‐mediated
beta‐catenin knockdown in human hepatoma cells results in decreased
growth and survival. Neoplasia, 2007;9(11):951–9.
197. Yao, M., Kargman, S., Lam, E.C. et al. Inhibition of cyclooxygenase‐2 by
rofecoxib attenuates the growth and metastatic potential of colorectal carci-
noma in mice. Cancer Res, 2003;63(3):586–92.
198. Behari, J., Zeng, G., Otruba, W. et al. R‐Etodolac decreases beta‐catenin
levels along with survival and proliferation of hepatoma cells. J Hepatol,
2007;46(5):849–57.
199. Li, H., Pamukcu, R., and Thompson, W.J. Beta‐catenin signaling: therapeu-
tic strategies in oncology. Cancer Biol Ther, 2002;1(6):621–5.
200. Emami, K.H., Nguyen, C., Ma, H. et al. A small molecule inhibitor of beta‐
catenin/CREB‐binding protein transcription [corrected]. Proc Natl Acad
Sci USA, 2004;101(34):12682–7.
201. Lenz, H.J. and Kahn, M. Safely targeting cancer stem cells via selective
catenin coactivator antagonism. Cancer Sci, 2014.
202. Delgado, E.R., Yang, J., So, J. et al. Identification and characterization of a
novel small‐molecule inhibitor of beta‐catenin signaling. Am J Pathol,
2014;184(7):2111–22.
203. Nejak‐Bowen, K.N. and Monga, S.P. Beta‐catenin signaling, liver regener-
ation and hepatocellular cancer: sorting the good from the bad. Semin
Cancer Biol, 2011;21(1):44–58.
204. Okabe, H., Delgado, E., Lee, J.M. et  al. Role of leukocyte cell‐derived
chemotaxin 2 as a biomarker in hepatocellular carcinoma. PLoS One,
2014;9(6):e98817.
205. Zhan, N., Michael, A.A., Wu, K. et al. The effect of selective c‐MET inhibi-
tor on hepatocellular carcinoma in the MET‐active, beta‐catenin‐mutated
mouse model. Gene Expr, 2018;18(2):135–47.
206. Delgado, E., Okabe, H., Preziosi, M. et al. Complete response of Ctnnb1‐
mutated tumours to beta‐catenin suppression by locked nucleic acid anti-
sense in a mouse hepatocarcinogenesis model. J Hepatol, 2014.
207. Adebayo Michael, A.O., Ko, S., and Tao, J. Inhibiting glutamine‐dependent
mTORC1 activation ameliorates liver cancers driven by β‐catenin muta-
tions. Cell Metab, 2019;29(5):1135–50.
208. Ruiz de Galarreta, M., Bresnahan, E., and Molina‐Sanchez, P. β‐catenin
activation promotes immune escape and resistance to anti‐PD‐1 therapy in
hepatocellular carcinoma. Cancer Discov, 2019. doi: 10.1158/2159‐8290.
CD‐19‐0074
209. Nakanuma, Y., Harada, K., Ishikawa, A., Zen, Y., and Sasaki, M. Anatomic
and molecular pathology of intrahepatic cholangiocarcinoma. J Hepatobiliary
Pancreat Surg, 2003;10(4):265–81.
210. Ashida, K., Terada, T., Kitamura, Y., and Kaibara, N. Expression of E‐cad-
herin, alpha‐catenin, beta‐catenin, and CD44 (standard and variant iso-
forms) in human cholangiocarcinoma: an immunohistochemical study.
Hepatology, 1998;27(4):974–82.
211. Rashid, A. Cellular and molecular biology of biliary tract cancers. Surg
Oncol Clin N Am, 2002;11(4):995–1009.
212. Sugimachi, K., Taguchi, K., Aishima, S. et al. Altered expression of beta‐
catenin without genetic mutation in intrahepatic cholangiocarcinoma. Mod
Pathol, 2001;14(9):900–5.
213. Rashid, A., Gao, Y.T., Bhakta, S. et  al. Beta‐catenin mutations in biliary
tract cancers: a population‐based study in China. Cancer Res, 2001;
61(8):3406–9.
214. Abraham, S.C., Lee, J.H., Hruban, R.H., Argani, P., Furth, E.E., and Wu,
T.T. Molecular and immunohistochemical analysis of intraductal papillary
neoplasms of the biliary tract. Hum Pathol, 2003;34(9):902–10.
215. Lustig, B., Jerchow, B., Sachs, M. et  al. Negative feedback loop of Wnt
signaling through upregulation of conductin/axin2 in colorectal and liver
tumors. Mol Cell Biol, 2002;22(4):1184–93.
216. Takayasu, H., Horie, H., Hiyama, E. et al. Frequent deletions and mutations
of the beta‐catenin gene are associated with overexpression of cyclin D1
and fibronectin and poorly differentiated histology in childhood hepato-
blastoma. Clin Cancer Res, 2001;7(4):901–8.
217. Torre, C., Benhamouche, S., Mitchell, C. et al. The transforming growth
factor‐alpha and cyclin D1 genes are direct targets of beta‐catenin signaling
in hepatocyte proliferation. J Hepatol, 2011;55(1):86–95.
218. Araki, Y., Okamura, S., Hussain, S.P. et al. Regulation of cyclooxygenase‐2
expression by the Wnt and ras pathways. Cancer Res, 2003;63(3):728–34.
219. Loeppen, S., Koehle, C., Buchmann, A., and Schwarz, M. A beta‐catenin‐
dependent pathway regulates expression of cytochrome P450 isoforms in
mouse liver tumors. Carcinogenesis, 2005;26(1):239–48.
220. Yamamoto, Y., Sakamoto, M., Fujii, G. et al. Overexpression of orphan G‐
protein‐coupled receptor, Gpr49, in human hepatocellular carcinomas with
beta‐catenin mutations. Hepatology, 2003;37(3):528–33.
221. Ovejero, C., Cavard, C., Perianin, A. et al. Identification of the leukocyte
cell‐derived chemotaxin 2 as a direct target gene of beta‐catenin in the liver.
Hepatology, 2004;40(1):167–76.
222. Cavard, C., Terris, B., Grimber, G. et al. Overexpression of regenerating
islet‐derived 1 alpha and 3 alpha genes in human primary liver tumors with
beta‐catenin mutations. Oncogene, 2006;25(4):599–608.
223. Renard, C.A., Labalette, C., Armengol, C. et al. Tbx3 is a downstream tar-
get of the Wnt/beta‐catenin pathway and a critical mediator of beta‐catenin
survival functions in liver cancer. Cancer Res, 2007;67(3):901–10.
224. Delgado, E., Bahal, R., Yang, J., Lee, J.M., Ly, D.H., and Monga, S.P. Beta‐
catenin knockdown in liver tumor cells by a cell permeable gamma guanidine‐
based peptide nucleic acid. Curr Cancer Drug Targets, 2013;13(8):
867–78.
225. Guichard, C., Amaddeo, G., Imbeaud, S. et  al. Integrated analysis of
somatic mutations and focal copy‐number changes identifies key genes and
pathways in hepatocellular carcinoma. Nat Genet, 2012;44(6):694–8.
226. Cui, J., Zhou, X., Liu, Y., Tang, Z., and Romeih, M. Alterations of
beta‐catenin and Tcf‐4 instead of GSK‐3beta contribute to activation
of  Wnt pathway in hepatocellular carcinoma. Chin Med J (Engl),
2003;116(12):1885–92.
227. Devereux, T.R., Stern, M.C., Flake, G.P. et  al. CTNNB1 mutations and
beta‐catenin protein accumulation in human hepatocellular carcinomas
associated with high exposure to aflatoxin B1. Mol Carcinog, 2001;31(2):
68–73.
228. Wong, C.M., Fan, S.T., and Ng, I.O. Beta‐catenin mutation and overexpres-
sion in hepatocellular carcinoma: clinicopathologic and prognostic signifi-
cance. Cancer, 2001;92(1):136–45.
229. Hsu, H.C., Jeng, Y.M., Mao, T.L., Chu, J.S., Lai, P.L., and Peng, S.Y. Beta‐
catenin mutations are associated with a subset of low‐stage hepatocellular
carcinoma negative for hepatitis B virus and with favorable prognosis. Am
J Pathol, 2000;157(3):763–70.
230. Huang, H., Fujii, H., Sankila, A. et al. Beta‐catenin mutations are frequent
in human hepatocellular carcinomas associated with hepatitis C virus infec-
tion. Am J Pathol, 1999;155(6):1795–801.
231. Legoix, P., Bluteau, O., Bayer, J. et al. Beta‐catenin mutations in hepatocel-
lular carcinoma correlate with a low rate of loss of heterozygosity.
Oncogene, 1999;18(27):4044–6.
232. Kondo, Y., Kanai, Y., Sakamoto, M. et al. Beta‐catenin accumulation and
mutation of exon 3 of the beta‐catenin gene in hepatocellular carcinoma.
Jpn J Cancer Res, 1999;90(12):1301–9.
 Polyploidy in Liver Function,
47 Mitochondrial Metabolism,
and Cancer
Evan R. Delgado, Elizabeth C. Stahl, Nairita Roy, Patrick D. Wilkinson,
and Andrew W. Duncan
Department of Pathology, McGowan Institute for Regenerative Medicine, Pittsburgh Liver Research Center,
University of Pittsburgh, Pittsburgh, PA, USA
HEPATIC CHROMOSOMAL VARIATIONS
AND LIVER FUNCTION
The development of polyploid cells in tissues and organs is a
highly regulated process. In the liver, failure to complete the
final stage of mitosis, known as cytokinesis, drives the retention
of nuclear content to a single cell. The heterogeneity of poly-
ploid cells in the liver is further nuanced by the occurrence of
aneuploidy, where single chromosomes may be gained or lost,
and by reductive mitosis, all driving the “ploidy conveyor.”
Polyploidy in somatic tissues and the liver
The majority of mammalian cells have a diploid nuclear content
and contain two sets of chromosomes. However, there is a phe-
nomenon called polyploidy, which describes cells with addi-
tional sets of chromosomes. Polyploid cells are an innate
component of many mammalian species, including rodents and
humans [1, 2]. The formation and location of these polyploid
cells is tissue‐dependent. For example, myofibrils and osteo-
clasts become multinucleate, polyploid cells via cell fusion
[3, 4]. Others, including megakaryocytes and trophoblast giant
cells, become polyploid through endoreduplication. Here, pro-
liferating cells progress through the cell cycle but fail to com-
plete nuclear division during mitosis, resulting in a polyploid
cell with a single nucleus [5, 6]. On the other hand, failure of
cytokinesis (the final step of cell division where the parent cell’s
cytoplasm divides to produce two daughter cells) primarily
drives the generation of polyploid cardiomyocytes and hepato-
cytes [2, 7, 8].
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
Polyploidy levels change profoundly during liver growth and
maturation. In the early stages of liver development, the liver is
primarily populated with small, uniform diploid (2n) hepato-
cytes [9, 10]. As the hepatocytes proliferate, the diploids fail to
complete cytokinesis, and in turn generate a daughter cell with
two diploid nuclei (2n × 2n), referred to as a binucleate tetra-
ploid (Figure  47.1a). Next, tetraploid cells with two diploid
nuclei replicate their DNA and generate two mononucleate
tetraploid (4n) daughter cells by successfully completing cytoki-
nesis. Mononucleate tetraploid cells can produce mono‐ and
binucleate octaploid (8n and 4n × 4n) hepatocytes through sub-
sequent cell divisions with complete and incomplete cytokine-
sis. The process continues, producing hexadecaploid (16n) cells
and hepatocytes with even greater ploidy states [2]. While fail-
ure to complete cytokinesis is the most common mechanism by
which hepatocytes increase ploidy states, heterotypic cell fusion
between macrophages and hepatocytes has been shown to occur,
albeit infrequently (at approximately 1/100 000 hepatocytes)
[11]. In general, as the liver matures the hepatocyte population
becomes increasingly heterogeneous in size, function, and
nuclear content, including number of nuclei per cell and chro-
mosome sets per nucleus [2, 8, 12].
The rate and percentage of liver polyploidization differ
between species [2]. At birth, the majority of human hepato-
cytes in the liver are diploid, with polyploids representing less
than 10% [13]. Notably, tetraploid hepatocytes appear within
the first few weeks of development, followed by octaploid
hepatocytes at two to three months after birth [14]. By adult-
hood, the percentage of polyploidy is nearly doubled to 20%,
and after the age of 60, approximately 50% of the human liver is
comprised of polyploid hepatocytes [13, 15]. Alternatively, the
604 THE LIVER:  HEPATIC CHROMOSOMAL VARIATIONS AND LIVER FUNCTION

Figure 47.1  Polyploidy and aneuploidy in the liver. (a) Hepatocytes are mononucleate or binucleate, and ploidy is determined by the number of
nuclei per cell and DNA content of each nucleus. Cell division associated with failed cytokinesis drives hepatic polyploidization. The cartoon shows
how a single diploid hepatocyte can give rise to mononucleate and binucleate tetraploid and octaploid hepatocytes. (b) In humans, diploid, tetra-
ploid, and octaploid hepatocytes contain 46, 92, and 184 chromosomes (Chr), respectively. Euploid karyotypes contain balanced sets of chromo-
somes, whereas aneuploid karyotypes contain one or more random chromosome gains and/or losses. (c) Polyploid hepatocytes can undergo
chromosome segregation errors (e.g. lagging chromosomes) and multipolar cell divisions to generate aneuploid daughter cells with reduced ploidy.
The flowchart outlines the various types of daughter cells that can be generated by a dividing tetraploid hepatocyte undergoing bipolar (left) or
multipolar (middle and right) nuclear segregation. (d) The ploidy conveyor model describes hepatic ploidy changes that occur during postnatal
development and proliferation. Cell divisions involving cytokinesis failure occur frequently and lead to increased ploidy. In contrast, cell divisions
associated with multipolar nuclear segregation are infrequent and lead to ploidy reversal and formation of diploid or near‐diploid daughters.
 47:  Polyploidy in Liver Function, Mitochondrial Metabolism, and Cancer 605
livers of mice are already 50% polyploid at weaning, whereas
the livers of rats begin to polyploidize upon weaning, triggered
by changes in insulin AKT‐dependent signaling that result in
cytokinesis failure [16]. By eight to twelve weeks of age both
mouse and rat livers become 80% polyploid, a faster rate than
human polyploidization, making rodents a useful model to
study polyploidization in the span of a few weeks [17]. The sig-
nals regulating the steady accumulation of polyploid hepato-
cytes in developing and aging humans remain unknown but may
also be related to the insulin‐AKT pathway. Interestingly, the
frequency of cells within each ploidy state remains relatively
constant among individuals of the same age, suggesting the
existence of a ploidy sensor to maintain homeostasis in the liver.
Liver aneuploidy and the ploidy conveyor
While hepatocytes are heterogeneous in regard to the number of
chromosome sets, they also exhibit diversity at the individual
chromosome level. G‐banding, fluorescence in situ hybridiza-
tion, and single cell DNA sequencing studies have indicated
4–50% of hepatocytes in healthy livers of mice and humans are
aneuploid, having gained or lost individual chromosomes
(Figure 47.1b) [13, 17, 18]. Recently, it was shown that the liv-
ers of mice deficient in microRNA‐122 had reduced polyploid
levels and were enriched with diploid hepatocytes. Moreover,
the livers had abnormally low aneuploidy levels [19].
Additionally, studies have shown that polyploidy and aneu-
ploidy levels follow similar trends, as both increase with age
and are the highest in adult murine and human livers [2, 20]. The
degree of aneuploidy plateaus in mice from four to fifteen
months of age [17]. Similarly, in humans, hepatic aneuploidy
levels remain stable between the ages of 10 and 60 years [13].
During mitosis, polyploid hepatocytes can form multipolar
spindles and drive “reductive mitosis,” or a reduction in nuclear
content, a process called “ploidy reversal” (Figure  47.1c,d).
One example of ploidy reversal is illustrated with a proliferat-
ing tetraploid hepatocyte that forms a tripolar spindle during
mitosis (Figure 47.1c, middle panel). The chromosomes of the
dividing cell are pulled toward three poles, resulting in two
daughter cells with approximately diploid nuclei and one
daughter cell with a ~tetraploid nucleus. Additionally, poly-
ploid cells can perform ploidy reversal by undergoing double
mitosis (Figure  47.1c, right panel). In this case, proliferating
tetraploids can generate up to four ~diploid daughter cells and
octaploids can generate up to eight ~diploid daughter cells [17,
21, 22]. Chromosome segregation errors often occur during
multipolar cell division [17, 21]. For instance, microtubules
associated with the outer poles occasionally attach to the same
kinetochore during multipolar spindle construction. If left
unchecked, this error prevents chromosomes from migrating
to discrete poles during anaphase. Consequently, these “lag-
ging” chromosomes are frequently excluded from the daughter
cell nuclei and form what is considered a micronucleus.
Cytogenetic analyses of hepatocytes have shown that liver
aneuploidy is widespread, with chromosome gains and losses
occurring in an unbiased fashion [13, 23]. Polyploidization,
ploidy reversal and aneuploidy, which are collectively called
the “ploidy conveyor”, prominently drive hepatocyte heteroge-
neity (Figure 47.1c,d) [2].
THE IMPACT OF POLYPLOIDY
ON GENE EXPRESSION AND LIVER
REGENERATION
The liver is a vitally important organ situated to metabolize hun-
dreds of molecules for energy regulation and drug toxicity, as
well as to produce important blood anticoagulation proteins,
immune complexes, and digestive bile acids. These processes
require the orchestration of complex gene networks and neces-
sitate the maintenance of liver function for survival. As such, the
liver has a robust capacity to adapt to injury and undergo com-
plete regeneration. The impact of polyploidy on gene expression
and liver regeneration is just beginning to be understood.
Polyploidy in liver gene expression
and function
While the mechanisms that produce polyploid hepatocytes have
been well characterized, the relationship between ploidy and
gene expression remains unclear. Experiments have shown that
polyploidy is associated with specific gene expression patterns
in yeast [24]. In regards to mammals, studies have demonstrated
that nuclear content affects expression levels of genes regulating
megakaryocyte differentiation [25] and that giant trophoblast
cells could exhibit biallelic X‐chromosome gene expression due
to incomplete X‐inactivation [26]. Additionally, a large‐scale
genome comparative study illustrated that cardiac stress genes
were expressed at different levels between diploid and polyploid
cardiomyocytes [27].
It has been postulated that hepatic ploidy state could affect
gene expression. One hypothesis is that gene expression levels
increase proportionally with ploidy. In this case, tetraploids and
octaploids would have 2× and 4× greater gene expression levels,
respectively, compared to diploids. Thus, tetraploid and octa-
ploid cells would also potentially contain 2× and 4× the total
RNA and protein content as diploid cells [2]. Another hypothe-
sis is that diploid and polyploid hepatocytes exhibit differential
gene expression. In this scenario, gene expression levels would
vary on a gene to gene basis between ploidy populations, with
polyploids exhibiting greater expression of specific genes ver-
sus diploids and vice versa. A 2007 study by Lu et al. compared
gene expression between quiescent diploid, tetraploid, and octa-
ploid hepatocyte populations isolated by flow cytometry.
Surprisingly, the study found gene expression was mostly
equivalent between ploidy populations, and the magnitude of
difference was small for those genes with different expression
levels [28]. This suggested that few genes are differentially
expressed, at least in quiescent hepatocytes. Further studies are
needed to determine if differential gene expression occurs in
disease conditions or in response to specific stimuli because
variations at the molecular level could ultimately endow certain
ploidy subsets with unique functional capabilities [2, 29].
Liver ploidy supports hepatic adaptation
It has been hypothesized that the diverse population of polyploid
and aneuploid hepatocytes endows the liver with the ability to
adapt to a variety of environmental stresses [23]. Interestingly,
606 THE LIVER:  THE IMPACT OF POLYPLOIDY ON GENE EXPRESSION AND LIVER REGENERATION
Figure 47.2  Proof of concept studies demonstrated that disease‐resistant aneuploid hepatocytes promote adaptation to chronic liver injury.
Healthy mouse livers contain randomly aneuploid hepatocytes, which are illustrated by multicolored cells. In response to chronic injury
(induced by tyrosinemia in Hgd+/− Fah−/− mice), liver function was severely impaired. Hepatocytes lacking chromosome 16 (blue) were resist-
ant to the injury and capable of proliferation. Clonal expansion by these injury‐resistant hepatocytes spontaneously repopulated the liver and
restored liver function.
studies have shown that budding yeast strains with specific chro-
mosome numbers are able to survive toxic insults and deleterious
mutations [30, 31]. In the context of the liver, it has been demon-
strated that aneuploid hepatocytes could protect against chronic
liver disease. Notably, a study illustrated that mice suffering
from tyrosinemia due to loss of fumarylacetoacetate hydrolase
(FAH) an enzyme in the tyrosine catabolic pathway, developed
resistance to the disease partially through hepatic aneuploidy
[23] (Figure  47.2). Mice suffering from tyrosinemia can be
treated and kept healthy by blocking the tyrosine catabolic path-
way upstream of FAH using the drug 2‐(2‐nitro‐4‐trifluoro‐
methylbenzoyl)‐1,3‐cyclo‐hexanedione (NTBC) or deleting the
enzyme homogentisic acid dioxygenase (HGD) [32, 33]. In the
aforementioned study, NTBC was withdrawn from Fah−/− Hgd+/−
mice and, contrary to the expectation that all mice would suc-
cumb to liver failure, some mice developed numerous small,
healthy nodules in their livers. Karyotyping and array compara-
tive genomic hybridization analyses revealed that many of the
healthy nodules were comprised of aneuploid hepatocytes lack-
ing a copy of chromosome 16, which contains the wild‐type
copy of Hgd. Thus, the healthy liver nodules were populated
with Hgd‐deficient hepatocytes. It was hypothesized that chronic
injury was toxic to the majority of hepatocytes except those that
had previously lost the chromosome with the Hgd wild‐type
copy. The disease‐resistant hepatocytes (monosomic for chro-
mosome 16), proliferated and repopulated the liver and restored
normal liver function [23].
Polyploidy and liver regeneration
The regenerative capacity of the liver has been well documented,
but the role of polyploidy in this process remains unclear. It is
hypothesized that diploid and polyploid populations could play
unique roles in the process. Originally, it was believed that poly-
ploid hepatocytes were mature, terminally differentiated cells
with little proliferative capacity. This was suggested from studies
demonstrating that hepatocytes in livers of mice and rats become
increasingly polyploid with age, and that more than 99% of
hepatocytes in adult livers were quiescent [34, 35]. However,
polyploids have been observed to proliferate in response to par-
tial hepatectomy (PH), the surgical removal of up to two‐thirds
of the liver mass [36]. While these experiments confirmed that
polyploid hepatocytes can proliferate, it is unclear if the
proliferative capacities of diploids and polyploids are equivalent.
One study demonstrated that livers from rats that had undergone
PH had increased polyploidy levels and changes associated with
cell senescence and aging [37]. Another study showed that mice
with drug‐induced necrosis and cirrhosis developed regenerating
nodules enriched with diploid hepatocytes [38]. The idea that
diploids are more proliferative than polyploid hepatocytes is also
supported by observations that diploid hepatocytes are enriched
in patients and rodents with hepatocellular carcinomas (HCC)
[38–41]. In contrast, it was shown that transplanted octaploid
and diploid hepatocytes proliferate equivalently in the Fah−/−
liver repopulation model. However, a limitation of this study was
that the ploidy states of the transplanted cells changed as the
octaploid cells underwent ploidy reversal and the diploids poly-
ploidized during the course of repopulation (Figure 47.1c) [17].
Taken together, these experiments illustrate that further studies
are needed to determine the precise proliferative capacities of
diploid and polyploid hepatocyte populations, and the mecha-
nisms regulating proliferation of each ploidy population.
Polyploidy in aging and impaired
liver regeneration
Aging is known to lead to the accumulation of senescent cells in
multiple tissues and organ systems, including the liver [42].
Senescence is the irreversible exit of the cell cycle, driven by the
shortening of telomeres (i.e. the protective caps of chromo-
somes) through exhaustive replication or other stressors. Many
hypothesize that senescence is a protective mechanism against
tumorigenesis that might arise from genomic instability, but
senescence also has the effect of limiting tissue regeneration
and promoting the secretion of inflammatory mediators that can
be damaging to surrounding tissue [43]. Aged octaploid cells
were found to express a higher proportion of senescence mark-
ers compared to tetraploid and diploid cells, including p16, p21,
and p53, which also serve important tumor suppressor roles
[44]. Methods to deplete senescent cells in aged organs and
­tissues, such as the INK‐ATTAC mouse, were not capable of
depleting senescent cells in the liver but reduced the spontane-
ous occurrence of liver tumors by depleting senescent cells at
peripheral sites, thus demonstrating the role of the aged
­microenvironment in the etiology of liver cancers and other age‐
related diseases [42, 43].
 47:  Polyploidy in Liver Function, Mitochondrial Metabolism, and Cancer 607
Unlike senescent diploid cells, which must be cleared by the
immune system, senescent polyploid cells have the unique
capacity to undergo ploidy reversal [44]. Aged hepatocytes that
undergo ploidy reversal also downregulate expression of senes-
cence markers, acting as a type of cell rejuvenation [44]. Ploidy
reversal might partially explain the capacity of the aged liver to
regenerate, although regeneration is substantially reduced in
elderly individuals. After PH in mice, hepatocytes in aged livers
have delayed cell cycle entry and the livers are significantly
deficient in the number of proliferating hepatocytes [45–47].
The age‐related decline in hepatocyte proliferation has been
attributed to transcription factor C/EBPα complexing with Brm,
a chromatin remodeling protein detected only in aged livers
[48], which inhibits E2f‐regulated gene expression [49].
Notably, when the circulatory systems of young and aged mice
are connected through heterochronic parabiosis, the number of
hepatocytes undergoing proliferation increased in the aged liv-
ers. While the ploidy of the rejuvenated hepatocytes was not
documented, a reduction in the C/EBPα–Brm complexes was
observed [48].
When aged diploid or octaploid hepatocytes are implanted
into young Fah−/− mice, the octaploid hepatocytes undergo
ploidy reversal, giving rise to lower hepatic ploidy states that no
longer express markers of senescence [17, 44]. This study not
only implicated the ability of a young systemic environment to
rejuvenate hepatocyte proliferative capacity, but also suggested
that hepatocytes with differential ploidy states have equivalent
proliferation kinetics during long‐term repopulation. Future
studies will need to decipher the role of cell intrinsic factors,
such as ploidy and gene stability, mitochondrial function, and
autophagy, versus extrinsic factors in the local or systemic envi-
ronment, in the process of hepatocyte function and regeneration
after injury and during aging.
THE MITOCHONDRIA‐PLOIDY NEXUS
IN LIVER DISEASE
Liver regeneration and a host of liver diseases are associated
with hepatic ploidy alterations and mitochondrial metabolic
dysfunction [50]. These observations suggest that the regulation
of energy expenditure or the breakdown of mitochondrial func-
tion is linked to the development of differential ploidy states.
This section describes the metabolic states and molecular regu-
lators that connect mitochondrial metabolism and polyploidy in
the liver.
Mitochondrial metabolism and
metabolic disease
Mitochondria serve an important role in managing energy
expenditure of the liver, through metabolism of lipids and other
molecules for ATP generation, which is essential for many cell
functions. Dysregulated mitochondrial metabolism has been
reported in liver diseases such as drug‐induced liver injury, alco-
holic liver disease, non‐alcoholic fatty liver disease (NAFLD),
viral hepatitis, liver cancer, and hemochromatosis [50, 51].
Mitochondrial dysfunction can manifest in several ways,
including disrupted mitochondrial membranes, dysregulation of
oxidative phosphorylation (OXPHOS), production of reactive
oxygen species (superoxides, nitric oxides, and lipid peroxides),
glutathione depletion, release of cytochrome c from the mito-
chondrial matrix, defects in mitochondrial fission and fusion,
abnormal mitophagy, and defective retrograde or anterograde
signaling between the mitochondria and nucleus [50].
Insulin resistance and hyperlipidemia are the major underly-
ing metabolic states of NAFLD [52]. Insulin signaling fortifies
the integrity of the electron transport chain in the mitochondria
by suppressing FOXO1 and maintaining the NAD+ : NADH
ratio [53]. This NAD+ : NADH ratio then determines the activity
of PGC1α, which is the major regulator of mitochondrial bio-
genesis [54]. Desdouets and colleagues recently showed that
advanced fatty liver disease, known as non‐alcoholic steatohep-
atitis, is associated with enhanced polyploidy [55]. Livers from
mice fed with diets that induce NAFLD, such as the methionine‐
choline‐deficient diet or high‐fat diet, were also dramatically
enriched with polyploid hepatocytes. Furthermore, in these fatty
liver models, oxidative stress was found to be the primary driver
of increased polyploidy. Thus, there appears to be an intricate
nexus between mitochondrial dysfunction, correlating with
metabolic liver disease and polyploidization.
The nexus of hepatic polyploidy
and mitochondrial metabolism
Liver regeneration and diseases associated with altered mito-
chondrial metabolism, including NAFLD and hemochromato-
sis, are associated with alterations in ploidy [37, 55–57].
Moreover, many genes that regulate mitochondrial metabolism
are reported to alter hepatic polyploidy such as Mir122, mem-
bers of the E2F family (E2F1, E2F7, and E2F8), Birc5, Ercc1,
Myc, p53, Rb, and Skp2 [2, 19, 58, 59]. Table 47.1 summarizes
the dual role of key molecular regulators involved in mitochon-
drial metabolism and hepatic polyploidy. Together, these obser-
vations strongly suggest that mitochondrial metabolism and
hepatic polyploidy are closely linked and could play a role in
progression of liver diseases.
Mitochondria–ploidy nexus in liver
regeneration
Hepatic polyploidy has been reported to transiently increase
when the liver regenerates following PH, during which concur-
rent changes occur in mitochondrial metabolism [37, 56]. First,
there is a strong positive correlation between the regenerative
capacity of the liver and the efficiency of OXPHOS, as demon-
strated by a subsequent increase in ATP levels and respiratory
control ratio, an indicator of OXPHOS efficiency [84]. Second,
expression of proteins involved in carbohydrate metabolism,
lipid metabolism, and OXPHOS increases 24 hours post PH
[85]. Third, there is heterogeneity among the type of mitochon-
drial populations in hepatocytes. These mitochondrial subtypes
differ in iron uptake capacity, iron metabolism, and complex IV
activity, ranging from 2–42 days post PH [86]. In addition, oxi-
dative stress is associated with liver regeneration, and this could
be caused, at least in part, by regeneration‐associated OXPHOS
[56]. Collectively, these reports highlight the heavy demand for
 Table 47.1  The nexus of hepatic polyploidy and mitochondrial metabolism
Gene Model Hepatic ploidy status Relevance to mitochondrial metabolism
E2f7, E2f8 Deletion via Reduced polyploidy The E2F family is known to regulate cell proliferation genes where E2F7 and E2F8 cooperate to repress E2F1 expression and function [62].
knockout [60, 61] 1. E2F1 regulates metabolism in organs such as liver, muscle, pancreas, and adipose tissue that determine metabolic homeostasis. Target
E2f1 Deletion via Increased polyploidy genes of E2F1 are responsible for lipid synthesis (FAS), oxidative metabolism (TOP1MT, EVOVL2, NANOG), and glycolysis (PFKB,
knockout [61] Sirt6, PDK) [63]
2. E2F7 and E2F8‐mediated regulation of E2F1 also occurs in metabolic scenarios such as oxidative stress induced DNA damage response
[64]
Mir122 Knockout Reduced polyploidy Mir122 deficiency has been reported in diseases with dysregulated mitochondrial metabolism such as HCC [65] and NAFLD.
[11] 1. Mice lacking miR‐122 have a severe defect in lipid metabolism [66]
2. miR‐122 regulates the expression of PGC1α, which is a master regulator of mitochondrial biogenesis [67]
3. Other targets of miR‐122, such as LCMT1, PPP1CC, ATF4, MEKK3, and MAPKAP2, are believed to regulate the expression of PGC1α
[67]
4. Another direct target of miR‐122, pyruvate kinase (PKM2) [68], is a key regulator of glycolysis and, thus, can influence mitochondrial
metabolism
Myc Overexpression Increased polyploidy Myc provides the clearest example of programmed expansion of mitochondrial content linked to the cell cycle [69]
via transgenic [58] 1. Myc regulates the expression of TFAM [70], a major determinant of mitochondrial DNA replication and maintenance, as well as
mouse PGC‐1α and PGC‐1β, which are prime regulators of mitochondrial mass and energy metabolism
Deletion via Reduced polyploidy 2. Myc targets include genes involved in bi‐genomic and mitochondrial transcription, mitochondrial translation, protein import, and ETC
knockout [58] complex assembly [71]
3. Myc also indirectly regulates mitochondrial gene expression via repression of microRNAs controlling nuclear genes encoding
mitochondrial proteins (e.g. miR‐23a/b and glutaminase) [72, 73]
Trp53 Deletion via Increased polyploidy 1. p53 transcriptionally represses glucose transporters GLUT1 and GLUT4 along with the insulin receptor (IR) to inhibit cellular glucose
knockout [58] uptake into the cell [74]
2. Through transcriptional activation of TP53‐induced glycolysis and apoptosis regulator (TIGAR), p53 decreases the rate of glycolysis
and redirects glycolytic intermediates into the pentose phosphate pathway (PPP) [74]
3. Glycolysis is also dampened by negative regulation of phosphor‐glycerate‐mutase (PGM) by p53. Transcriptional activation of
hexokinase II (HK II) by mutant p53 stimulates glycolysis [74]
4. p53 promotes OXPHOS through transcriptional activation of synthesis of cytochrome c oxidase 2 (SCO2), a regulator of complex IV,
and apoptosis‐inducing factor (AIF), which acts directly on complex I. By regulating transcription and stability of ribonucleotide
reductase subunit (p53R2), p53 maintains mitochondrial homeostasis and mitochondrial genome integrity [74]
5. p53 is able to transcriptionally regulate and interact with the nuclear encoded mitochondrial transcription factor A (TFAM) and plays a
role in mitochondrial DNA (mtDNA) transcription and in regulating mtDNA content [74]
6. Genotoxic stress signals trigger cytoplasmic p53 to undergo MDM2‐dependent mono‐ubiquitination that induces translocation of p53 to
the mitochondria, which can trigger mitochondrial dependent apoptosis [74]
7. Mitochondrial p53 has been demonstrated to block the antioxidant function of manganese superoxide dismutase (MnSOD), thus creating
a state of oxidative stress [74]
Rb1 Deletion via Increased polyploidy 1. Proteomic effects of Rb1 ablation reveal a striking decrease in multiple mitochondrial proteins [75]
knockout [58] 2. Rb1‐deficient cells have decreased mitochondrial mass and reduced OXPHOS activity [75]
Skp2 Deletion via Increased polyploidy One of the prime targets of Skp2 ubiquitination‐mediated degradation is mitochondrial sirtuin 3 (SIRT3) [76], a histone NAD+ deactylase
knockout [58] 1. SIRT3 controls the flow of mitochondrial oxidative pathways and, consequently, the rate of production of reactive oxygen species [77]
2. SIRT3 controls energy demand during stress conditions (fasting and exercise) and has the ability to quench reactive oxygen species [77]
3. SIRT3 targets enzymes involved in energy metabolism processes, including the respiratory chain, tricarboxylic acid cycle, fatty acid
β‐oxidation and ketogenesis [77]
Birc5 Deletion via Increased polyploidy Proteomic analysis of Birc5‐deleted livers reveal overexpression of mTOR [78], which is a key player in the mitochondrial metabolism
(Survivin) knockout [58] [79]. This axis of BIRC5 and mTOR can possibly act as a node between ploidy and mitochondrial metabolism.
1. mTOR responds to a number of metabolic stimuli such as insulin, insulin growth factor, nutrients (amino acids), energy status, and
oxygen levels and, thus, regulates proliferation [80]
2. mTOR also inhibits mitochondrial autophagy [81].
Independent of mTOR, BIRC5 regulates mitochondrial fission and complex I activity in neuroblastoma cell lines [82], suggesting that it
might be involved in determining mitochondrial metabolism in the liver as well
Ercc1 Deletion via Increased polyploidy Ercc1 deficiency leads to increased oxidative damage as well as decreased levels of genes that are responsible for oxidative metabolism in
knockout [58] the liver [83].

c47.indd 608 03-11-2019 19:52:11

 47:  Polyploidy in Liver Function, Mitochondrial Metabolism, and Cancer 609

energy during liver regeneration and underscore an essential role
for mitochondrial metabolism in this process [87]. Additional
studies are required to elucidate the mechanisms that connect
liver regeneration with mitochondrial metabolism and ploidy.
PLOIDY AND LIVER CANCER
Polyploidy has been considered a hallmark of cancer for nearly
a century [88–90]. In the liver, however, some have suggested
that polyploidy serves as a protective mechanism by providing
multiple copies of tumor suppressor genes in the event that one
is damaged [91]. Many forms of cancer exhibit polyploidy at
both early and late stages of disease (e.g. renal cell carcinoma,
myeloid leukemia, glioblastoma, pre‐esophageal cancer, and
lung cancer [92]), but in the context of HCC a reduction of
ploidy has been observed. Although HCC is associated with
chromosomal variations [93], the relationship between poly-
ploidy, HCC development and progression, and HCC drug
resistance is poorly understood.
Polyploidy and liver cancer
HCC is the most common form of liver cancer and is associated
with high mortality and morbidity worldwide [94]. Studies con-
ducted in the 1990s and early 2000s of patients with HCC
revealed HCC tumors contained more diploid cells than poly-
ploids [94] (Figure 47.3a). Consistent with these observations,
work conducted on rats exposed to diethylnitrosamine (DEN) or
2‐acetyl‐aminofluorene indicated that the cells within the
diseased, HCC tissue became predominately diploid [41, 95,
96]. It is unclear if this shift from polyploid to diploid cells after
malignant HCC transformation is a cause or effect in pathogen-
esis, but a potential explanation for the enrichment of diploid
hepatocytes in HCCs is that polyploid hepatocytes protect from
oncogenic transformation. This concept may seem counterin-
tuitive, however, because polyploidy is traditionally associated
with increased disease severity in other cancers [92].
Recent studies support the idea that polyploidy plays a protec-
tive role in the liver. Using a tunable system for altering hepatic
ploidy in vivo, Zhu and colleagues showed that Anln deficient
mice with highly polyploid livers were protected from HCC in
the DEN tumor initiation model, whereas E2f8 deficient mice
with predominately diploid livers were predisposed to HCC for-
mation [97]. Most importantly, these studies revealed that the
polyploid state protects against transformation by providing
extra tumor suppressor alleles. With two homologous chromo-
somes per cell, diploid hepatocytes have two alleles of each
tumor suppressor. If one tumor suppressor allele is mutated and
inactivated, the cell is protected by only one functional allele.
However, the cell is highly susceptible to tumorigenesis if the
second allele is subsequently inactivated. In contrast, polyploid
hepatocytes have four or more homologous chromosomes and a
corresponding number of tumor suppressor alleles. When a sin-
gle tumor suppressor is inactivated in a polyploid hepatocyte, the
cell is protected or “buffered” by three or more functional alleles.
Thus, even if a second mutation event inactivates a remaining
allele, the cell still retains a minimum of two functional alleles
and can maintain tumor suppressor activity. Taken together, stud-
ies in patients and rodent models strongly support the notion that
diploid hepatocytes are more susceptible to oncogenic

Figure 47.3  Putative role of hepatic ploidy populations in liver cancer. Diploid hepatocytes are enriched in human HCC and are speculated to
have enhanced proliferative potential. Polyploid hepatocytes are reduced in HCC and may have reduced proliferative potential. Recent studies
indicate that polyploid hepatocytes protect against oncogenesis induced by tumor suppressor loss/inactivation; wild‐type copies of the tumor sup-
pressor found on additional chromosomes serve as “back‐ups” that are capable of restricting oncogenic proliferation.
610 THE LIVER:  CONCLUSION
transformation, while polyploid hepatocytes provide a protective
mechanism against HCC formation (Figure 47.3a).
Polyploidy and drug resistance in cancer
Increased polyploidy in cancers is associated with poor prognosis
primarily driven by an association with spontaneous drug resist-
ance [92]. In colon cancer cell lines, for instance, polyploidy is
attributed to increased resistance to DNA damage caused by
camptothecin, cisplatin, oxaliplatin, gamma irradiation, and ultra-
violet irradiation [98, 99]. Observations supporting a link between
liver ploidy and drug sensitivity have also been made in in vivo
studies. For example, the livers of Mir122 liver‐specific knockout
mice are predominantly diploid, and when challenged with aceta-
minophen, Mir122 knockouts are more susceptible to liver dam-
age compared to wild‐type mice with polyploid hepatocytes [100].
However, the data are difficult to interpret as Mir122 knockout
mice express elevated levels of cytochrome P450 family 2 subfam-
ily E member 1 (CYP2E1) and cytochrome P450 family 1 sub-
family A member 2 (CYP1A2), which metabolize acetaminophen
to its toxic byproduct N‐acetyl‐p‐benzoquinone imine (NAPQI)
[100]. These findings indicate that the relationship between ploidy
and drug resistance in the liver needs to be further investigated.
Chemoresistance is a major concern as it may lead to inappro-
priate treatment regimens, poor disease prognosis, and decreased
quality of life for patients with liver cancer. In the clinic, the
multi‐tyrosine kinase inhibitor sorafenib and the topoisomerase II
inhibitor doxorubicin are used to treat patients with unresectable
HCC [101]. Unfortunately, chemotherapeutic intervention is min-
imally effective and frequently leads to tumors adapting and
becoming refractory to the agent(s) administered [101]. The
mechanisms by which HCC becomes chemoresistant are poorly
understood, but in other cancers, polyploidy has been associated
with drug resistance [92]. The shift in human HCC from poly-
ploid to diploid cell predominance within the diseased tissue indi-
cates that chemoresistance in liver cancer may not be derived
from polyploidy but rather by other mechanisms. Overall, further
studies are necessary to determine how polyploidy affects disease
severity or drug resistance, which may lead to the development of
novel, innovative, and effective therapeutics.
Aneuploidy and liver cancer
Aneuploidy and chromosomal mis‐segregations have been
linked to cancer and neoplastic transformation for almost one
hundred years [88]. Poor disease prognosis and spontaneous
drug resistance are strongly linked to aneuploidy and chromo-
somal instability as these events can contribute to cancer cell
survival (Figure 47.3b) [92].
Although polyploidy and aneuploidy involve chromosomal
alterations, it is critical to understand the difference between the
two when considering disease. Polyploidy is a whole number
increase in complete chromosome sets in a euploid cell. Depending
on the context, aneuploidy can include whole chromosome gains
and losses or structural changes in individual chromosomes. For
example, these changes can include complete chromosomal rear-
rangements or even chromosome segments that are gained or lost.
In the healthy adult liver of humans and mice, whole chromosome
gains and losses are a normal feature [13, 17, 18, 22].
Previous work from Michalopoulos and colleagues found
461 copy number variations (CNVs) in 98 human HCC sam-
ples, demonstrating that chromosomal variations are common in
liver cancer [93]. The CNVs identified in this study contained
discrete genomic amplifications and deletions but not gains or
losses of entire chromosomes. While aneuploidy in patient HCC
tumors is not well studied, hepatoblastoma tumors have been
extensively characterized and shown to contain near‐diploid or
near‐tetraploid cells that have gained or lost entire chromo-
somes [102]. Additionally, underlying pathologies of hepatocel-
lular carcinogenesis, such as chronic hepatitis B virus infection,
aflatoxin exposure, and oxidative damage, are associated with
chromosomal variations [103, 104].
It is largely accepted that increased aneuploidy, or chromo-
some instability (CIN), associates with worse disease prognosis
[105]. Unexpectedly, patients with lung, breast, ovarian, and
gastric cancers with the highest degree of CIN were reported to
have better outcomes than patients with moderate CIN [105].
One potential explanation is that high‐degree CIN leads to
mitotic catastrophe and cell death, which effectively kills tumors
with the most aneuploidy. It is difficult to determine if aneu-
ploidy promotes HCC cancer cell survival because experimen-
tally‐induced aneuploidy and chromosomal rearrangements
have produced ambiguous and conflicting results [105]. To
study aneuploidy’s role in HCC development and progression in
vitro and in vivo, researchers have used genotoxic agents or spe-
cifically tailored genetic models [105]. Recent work suggests
that the relationship between aneuploidy and HCC can be stud-
ied by modulating miR‐122 levels in vitro and in vivo because
Mir122 deficiency results in an increase in mononucleate hepat-
ocytes with a concomitant reduction in aneuploidy [19]. To
effectively investigate if aneuploidy contributes to HCC devel-
opment, the ability to manipulate ploidy in cancer models is
critical. Ultimately, the development of innovative strategies
that can evade chemoresistance and eliminate tumors will
depend upon determining the relationship between aneuploidy,
polyploidy, and cancer cell survival.
CONCLUSION
In adults, the majority of hepatocytes in the liver are polyploid.
This suggests that polyploid cells are critical to the function and
homeostasis of the liver. The accumulation of polyploid hepato-
cytes with age suggests they may serve as a reservoir for prolifera-
tion and adaptation in the case of damage and injury. Recent
studies have shown that polyploid hepatocytes act as a critical
reserve for cells of lower ploidy states, which can be generated by
ploidy reversal. Also, polyploid hepatocytes appear to be more
resistant to toxic liver injury and cancer, and they may play a dis-
crete role in liver metabolism and regeneration. The molecular
mechanisms affecting differential activity of diploid and polyploid
hepatocytes are poorly understood, and a key question that remains
is whether gene expression in hepatocytes is dictated by ploidy
state. A better understanding of hepatic polyploidy will ultimately
benefit patients in the future by driving the improvement and
development of innovative strategies that prevent disease, stimu-
late liver regeneration, and advance artificial organ construction.
 47:  Polyploidy in Liver Function, Mitochondrial Metabolism, and Cancer 611
ACKNOWLEDGMENTS
This work was supported by grants to AWD from the NIH (R01
DK103645), the Commonwealth of Pennsylvania, and the
University of Pittsburgh Physicians Academic Foundation and
to ECS from the NIH (F31 DK112633) and the NIBIB training
grant (T32 EB001026) entitled “Cellular Approaches to Tissue
Engineering and Regeneration.” Thanks to Frances Alencastro
for helpful discussions and critical reading of the manuscript.
The authors of this chapter apologize to the authors whose
works were not cited because of space restrictions.
REFERENCES
1. Orr‐Weaver, T.L. When bigger is better: the role of polyploidy in organogen-
esis. Trends Genet, 2015;31(6):307–15.
2. Duncan, A.W. Aneuploidy, polyploidy and ploidy reversal in the liver. Semin
Cell Dev Biol, 2013;24(4):347–56.
3. Abmayr, S.M. and Pavlath, G.K. Myoblast fusion: lessons from flies and
mice. Development, 2012;139(4):641–56.
4. Helming, L. and Gordon, S. Molecular mediators of macrophage fusion.
Trends Cell Biol, 2009;19(10):514–22.
5. Edgar, B.A. and Orr‐Weaver, T.L. Endoreplication cell cycles: more for less.
Cell, 2001;105(3):297–306.
6. Lee, H.O., Davidson, J.M., and Duronio, R.J. Endoreplication: polyploidy
with purpose. Genes Dev, 2009;23(21):2461–77.
7. Engel, F.B., Schebesta, M., and Keating, M.T. Anillin localization defect in
cardiomyocyte binucleation. J Mol Cell Cardiol, 2006;41(4):601–12.
8. Margall‐Ducos, G., Celton‐Morizur, S., Couton, D., Bregerie, O., and
Desdouets, C. Liver tetraploidization is controlled by a new process of
incomplete cytokinesis. J Cell Sci, 2007;120(20):3633–9.
9. Bohm, N. and Noltemeyer, N. Development of binuclearity and DNA‐poly-
ploidization in the growing mouse liver. Histochemistry, 1981;72(1):55–61.
10. Celton‐Morizur, S., Merlen, G., Couton, D., and Desdouets, C. Polyploidy
and liver proliferation: central role of insulin signaling. Cell Cycle,
2010;9(3):460–6.
11. Wang, X., Montini, E., Al‐Dhalimy, M., Lagasse, E., Finegold, M., and
Grompe, M. Kinetics of liver repopulation after bone marrow transplanta-
tion. Am J Pathol, 2002;161(2):565–74.
12. Wilson, J.W. and Leduc, E.H. The occurrence and formation of binucleate
and multinucleate cells and polyploid nuclei in the mouse liver. Am J Anat,
1948;82(3):353–91.
13. Duncan, A.W., Hanlon Newell, A.E., Smith, L. et al. Frequent aneuploidy
among normal human hepatocytes. Gastroenterology, 2012;142(1):25–8.
14. Wang, M.J., Chen, F., Lau, J.T.Y., and Hu, Y.P. Hepatocyte polyploidization
and its association with pathophysiological processes. Cell Death Dis,
2017;8(5):e2805.
15. Watanabe, T. and Tanaka, Y. Age‐related alterations in the size of human
hepatocytes. A study of mononuclear and binucleate cells. Virchows Arch B
Cell Pathol Incl Mol Pathol, 1982;39(1):9–20.
16. Celton‐Morizur, S., Merlen, G., Couton, D., Margall‐Ducos, G., and
Desdouets, C. The insulin/Akt pathway controls a specific cell division pro-
gram that leads to generation of binucleated tetraploid liver cells in rodents.
J Clin Invest, 2009;119(7):1880–7.
17. Duncan, A.W., Taylor, M.H., Hickey, R.D. et al. The ploidy conveyor of mature
hepatocytes as a source of genetic variation. Nature, 2010;467(7316):707–10.
18. Knouse, K.A., Wu, J., Whittaker, C.A., and Amon, A. Single cell sequencing
reveals low levels of aneuploidy across mammalian tissues. Proc Natl Acad
Sci USA, 2014;111(37):13409–14.
19. Hsu, S.H., Delgado, E.R., Otero, P.A. et al. MicroRNA‐122 regulates poly-
ploidization in the murine liver. Hepatology, 2016;64(2):599–615.
20. Kudryavtsev, B.N., Kudryavtseva, M.V., Sakuta, G.A., and Stein, G.I.
Human hepatocyte polyploidization kinetics in the course of life cycle.
Virchows Arch B Cell Pathol Incl Mol Pathol, 1993;64(6):387–93.
21. Duncan, A.W., Hickey, R.D., Paulk, N.K. et al. Ploidy reductions in murine
fusion‐derived hepatocytes. PLoS Genet, 2009;5(2):e1000385.
22. Faggioli, F., Vezzoni, P., and Montagna, C. Single‐cell analysis of ploidy and
centrosomes underscores the peculiarity of normal hepatocytes. PloS One,
2011;6(10):e26080.
23. Duncan, A.W., Hanlon Newell, A.E., Bi, W. et al. Aneuploidy as a mechanism
for stress‐induced liver adaptation. J Clin Invest, 2012;122(9):3307–15.
24. Galitski, T., Saldanha, A.J., Styles, C.A., Lander, E.S., and Fink, G.R. Ploidy
regulation of gene expression. Science, 1999;285(5425):251–4.
25. Raslova, H., Kauffmann, A., Sekkai, D. et  al. Interrelation between poly-
ploidization and megakaryocyte differentiation: a gene profiling approach.
Blood, 2007;109(8):3225–34.
26. Corbel, C., Diabangouaya, P., Gendrel, A.V., Chow, J.C., and Heard, E.
Unusual chromatin status and organization of the inactive X chromosome in
murine trophoblast giant cells. Development, 2013;140(4):861–72.
27. Anatskaya, O.V. and Vinogradov, A.E. Genome multiplication as adaptation
to tissue survival: evidence from gene expression in mammalian heart and
liver. Genomics, 2007;89(1):70–80.
28. Lu, P., Prost, S., Caldwell, H., Tugwood, J.D., Betton, G.R., and Harrison,
D.J. Microarray analysis of gene expression of mouse hepatocytes of differ-
ent ploidy. Mamm Genom, 2007;18(9):617–26.
29. Rajvanshi, P., Liu, D., Ott, M., Gagandeep, S., Schilsky, M.L., and Gupta, S.
Fractionation of rat hepatocyte subpopulations with varying metabolic
potential, proliferative capacity, and retroviral gene transfer efficiency. Exp
Cell Res, 1998;244(2):405–19.
30. Pavelka, N., Rancati, G., Zhu, J. et al. Aneuploidy confers quantitative pro-
teome changes and phenotypic variation in budding yeast. Nature,
2010;468(7321):321–5.
31. Rancati, G., Pavelka, N., Fleharty, B. et al. Aneuploidy underlies rapid adap-
tive evolution of yeast cells deprived of a conserved cytokinesis motor. Cell,
2008;135(5):879–93.
32. Grompe, M., Lindstedt, S., al‐Dhalimy, M. et al. Pharmacological correction
of neonatal lethal hepatic dysfunction in a murine model of hereditary tyrosi-
naemia type I. Nat Genet, 1995;10(4):453–60.
33. Manning, K., Al‐Dhalimy, M., Finegold, M., and Grompe, M. In vivo sup-
pressor mutations correct a murine model of hereditary tyrosinemia type I.
Proc Natl Acad Sci USA, 1999;96(21):11928–33.
34. Gupta, S. Hepatic polyploidy and liver growth control. Semin Cancer Biol,
2000;10(3):161–71.
35. Fausto, N. and Campbell, J.S. The role of hepatocytes and oval cells in liver
regeneration and repopulation. Mech Dev, 2003;120(1):117–30.
36. Miyaoka, Y., Ebato, K., Kato, H., Arakawa, S., Shimizu, S., and Miyajima,
A. Hypertrophy and unconventional cell division of hepatocytes underlie
liver regeneration. Curr Biol, 2012;22(13):1166–75.
37. Sigal, S.H., Rajvanshi, P., Gorla, G.R. et  al. Partial hepatectomy‐induced
polyploidy attenuates hepatocyte replication and activates cell aging events.
Am J Physiol, 1999;276(5 Pt 1):G1260–72.
38. Gandillet, A., Alexandre, E., Royer, C., Cinqualbre, J., Jaeck, D., and
Richert, L. Hepatocyte ploidy in regenerating livers after partial hepatec-
tomy, drug‐induced necrosis, and cirrhosis. Eur Surg Res, 2003;35(3):
148–60.
39. Anti, M., Marra, G., Rapaccini, G.L. et  al. DNA ploidy pattern in human
chronic liver diseases and hepatic nodular lesions. Flow cytometric analysis
on echo‐guided needle liver biopsy. Cancer, 1994;73(2):281–8.
40. Rua, S., Comino, A., Fruttero, A. et  al. Flow cytometric DNA analysis of
cirrhotic liver cells in patients with hepatocellular carcinoma can provide a
new prognostic factor. Cancer, 1996;78(6):1195–202.
41. Schwarze, P.E., Pettersen, E.O., Shoaib, M.C., and Seglen, P.O. Emergence
of a population of small, diploid hepatocytes during hepatocarcinogenesis.
Carcinogenesis, 1984;5(10):1267–75.
42. Baker, D.J., Childs, B.G., Durik, M. et al. Naturally occurring p16(Ink4a)‐
positive cells shorten healthy lifespan. Nature, 2016;530(7589):184–9.
43. Sturmlechner, I., Durik, M., Sieben, C.J., Baker, D.J., and van Deursen, J.M.
Cellular senescence in renal ageing and disease. Nat Rev Nephrol,
2017;13(2):77–89.
44. Wang, M.J., Chen, F., Li, J.X. et al. Reversal of hepatocyte senescence after
continuous in vivo cell proliferation. Hepatology, 2014;60(1):349–61.
45. Bucher, N.L., Swaffield, M.N., and Ditroia, J.F. The influence of age upon
the incorporation of thymidine‐2‐C14 into the DNA of regenerating rat liver.
Cancer Res, 1964;24:509–12.
46. Fry, M., Silber, J., Loeb, L.A., and Martin, G.M. Delayed and reduced cell
replication and diminishing levels of DNA polymerase‐alpha in regenerating
liver of aging mice. J Cell Physiol, 1984;118(3):225–32.
612 THE LIVER:  REFERENCES
47. Timchenko, N.A., Wilde, M., Kosai, K.I. et al. Regenerating livers of old rats
contain high levels of C/EBPalpha that correlate with altered expression of
cell cycle associated proteins. Nucleic Acids Res, 1998;26(13):3293–9.
48. Conboy, I.M., Conboy, M.J., Wagers, A.J., Girma, E.R., Weissman, I.L., and
Rando, T.A. Rejuvenation of aged progenitor cells by exposure to a young
systemic environment. Nature, 2005;433(7027):760–4.
49. Iakova, P., Awad, S.S., and Timchenko, N.A. Aging reduces proliferative
capacities of liver by switching pathways of C/EBPalpha growth arrest. Cell,
2003;113(4):495–506.
50. Grattagliano, I., Russmann, S., Diogo, C. et al. Mitochondria in chronic liver
disease. Curr Drug Targets, 2011;12(6):879–93.
51. Du, K., Ramachandran, A., and Jaeschke, H. Oxidative stress during aceta-
minophen hepatotoxicity: sources, pathophysiological role and therapeutic
potential. Redox Biol, 2016;10:148–56.
52. Lonardo, A., Ballestri, S., Marchesini, G., Angulo, P., and Loria, P.
Nonalcoholic fatty liver disease: a precursor of the metabolic syndrome. Dig
Liver Dis, 2015;47(3):181–90.
53. Cheng, Z., Tseng, Y., and White, M.F. Insulin signaling meets mitochondria
in metabolism. Trends Endocrinol Metab, 2010;21(10):589–98.
54. Scarpulla, R.C. Metabolic control of mitochondrial biogenesis through the
PGC‐1 family regulatory network. Biochim Biophys Acta, 2011;1813(7):
1269–78.
55. Gentric, G., Maillet, V., Paradis, V. et al. Oxidative stress promotes patho-
logic polyploidization in nonalcoholic fatty liver disease. J Clin Invest,
2015;125(3):981–92.
56. Gorla, G.R., Malhi, H., and Gupta, S. Polyploidy associated with oxidative
injury attenuates proliferative potential of cells. J Cell Sci, 2001;114(16):
2943–51.
57. Troadec, M.B., Courselaud, B., Detivaud, L. et al. Iron overload promotes
Cyclin D1 expression and alters cell cycle in mouse hepatocytes. J Hepatol,
2006;44(2):391–9.
58. Duncan, A.W. Changes in hepatocyte ploidy during liver regeneration, in
Liver Regeneration: Basic Mechanisms, Relevant Models and Clinical
Applications, (ed. U. Apte) Elsevier, 2015.
59. Pandit, S.K., Westendorp, B., Nantasanti, S. et al. E2F8 is essential for poly-
ploidization in mammalian cells. Nat Cell Biol, 2012;14(11):1181–91.
60. Kent, L.N., Rakijas, J.B., Pandit, S.K. et al. E2f8 mediates tumor suppression
in postnatal liver development. J Clin Invest, 2016;126(8):2955–69.
61. Pandit, S.K., Westendorp, B., and de Bruin, A. Physiological significance of
polyploidization in mammalian cells. Trends Cell Biol,
2013;23(11):556–66.
62. Moon, N.S. and Dyson, N. E2F7 and E2F8 keep the E2F family in balance.
Dev Cell, 2008;14(1):1–3.
63. Denechaud, P.D., Fajas, L., and Giralt, A. E2F1, a novel regulator of metabo-
lism. Front Endocrinol, 2017;8:311.
64. Zalmas, L.P., Zhao, X., Graham, A.L. et al. DNA‐damage response control
of E2F7 and E2F8. EMBO Rep, 2008;9(3):252–9.
65. Kutay, H., Bai, S., Datta, J. et al. Downregulation of miR‐122 in the rodent
and human hepatocellular carcinomas. J Cell Biochem, 2006;99(3):671–8.
66. Hsu, S.H., Wang, B., Kota, J. et al. Essential metabolic, anti‐inflammatory,
and anti‐tumorigenic functions of miR‐122 in liver. J Clin Invest,
2012;122(8):2871–83.
67. Burchard, J., Zhang, C., Liu, A.M. et  al. microRNA‐122 as a regulator of
mitochondrial metabolic gene network in hepatocellular carcinoma. Mol Syst
Biol, 2010;6:402.
68. Liu, A.M., Xu, Z., Shek, F.H. et al. miR‐122 targets pyruvate kinase M2 and
affects metabolism of hepatocellular carcinoma. PloS One, 2014;9(1):e86872.
69. Morrish, F. and Hockenbery, D. MYC and mitochondrial biogenesis. Cold
Spring Harb Perspect Med, 2014;4(5).
70. Ahuja, P., Zhao, P., Angelis, E. et al. Myc controls transcriptional regulation
of cardiac metabolism and mitochondrial biogenesis in response to patho-
logical stress in mice. J Clin Invest, 2010;120(5):1494–505.
71. Li, F., Wang, Y., Zeller, K.I. et al. Myc stimulates nuclearly encoded mito-
chondrial genes and mitochondrial biogenesis. Mol Cell Biol, 2005;
25(14):6225–34.
72. Dang, C.V., Le, A., and Gao, P. MYC‐induced cancer cell energy metabolism
and therapeutic opportunities. Clin Cancer Res, 2009;15(21):6479–83.
73. Psathas, J.N. and Thomas‐Tikhonenko, A. MYC and the art of microRNA
maintenance. Cold Spring Harb Perspect Med, 2014;4(8).
74. Itahana, Y. and Itahana, K. Emerging roles of p53 family members in glucose
metabolism. Int J Mol Sci, 2018;19(3).
75. Nicolay, B.N., Danielian, P.S., Kottakis, F. et al. Proteomic analysis of pRb
loss highlights a signature of decreased mitochondrial oxidative phospho-
rylation. Genes Dev, 2015;29(17):1875–89.
76. Iwahara, T., Bonasio, R., Narendra, V., and Reinberg, D. SIRT3 functions in
the nucleus in the control of stress‐related gene expression. Mol Cell Biol,
2012;32(24):5022–34.
77. Hirschey, M.D., Shimazu, T., Huang, J.Y., Schwer, B., and Verdin, E.
SIRT3 regulates mitochondrial protein acetylation and intermediary metab-
olism. Cold Spring Harb Symp Quant Biol, 2011;76:267–77.
78. Bracht, T., Hagemann, S., Loscha, M. et al. Proteome analysis of a hepato-
cyte‐specific BIRC5 (survivin)‐knockout mouse model during liver regen-
eration. J Proteome Res, 2014;13(6):2771–82.
79. Morita, M., Gravel, S.P., Hulea, L. et al. mTOR coordinates protein synthesis,
mitochondrial activity and proliferation. Cell Cycle, 2015;14(4):473–80.
80. Laplante, M. and Sabatini, D.M. mTOR signaling in growth control and
disease. Cell, 2012;149(2):274–93.
81. Kim, J., Kundu, M., Viollet, B., and Guan, K.L. AMPK and mTOR regulate
autophagy through direct phosphorylation of Ulk1. Nat Cell Biol,
2011;13(2):132–41.
82. Hagenbuchner, J., Kuznetsov, A.V., Obexer, P., and Ausserlechner, M.J.
BIRC5/Survivin enhances aerobic glycolysis and drug resistance by altered
regulation of the mitochondrial fusion/fission machinery. Oncogene,
2013;32(40):4748–57.
83. Gregg, S.Q., Gutierrez, V., Robinson, A.R. et al. A mouse model of acceler-
ated liver aging caused by a defect in DNA repair. Hepatology,
2012;55(2):609–21.
84. Guerrieri, F., Muolo, L., Cocco, T. et  al. Correlation between rat liver
regeneration and mitochondrial energy metabolism. Biochim Biophys Acta,
1995;1272(2):95–100.
85. Sun, Q., Miao, M., Jia, X. et al. Subproteomic analysis of the mitochondrial
proteins in rats 24 h after partial hepatectomy. J Cell Biochem,
2008;105(1):176–84.
86. Gear, A.R. Observations on iron uptake, iron metabolism, cytochrome c
content, cytochrome a content and cytochrome c‐oxidase activity in regen-
erating rat liver. Biochem J, 1965;97(2):532–9.
87. Huang, J. and Rudnick, D.A. Elucidating the metabolic regulation of liver
regeneration. Am J Pathol, 2014;184(2):309–21.
88. Boveri, T. Zur Frage der Entstehung maligner Tumoren (The Origin of
Malignant Tumors), Jena, Gustav Fischer, 1914.
89. Wolff, J. The Theory of Cancer, Science History Publishers, Sagamore
Beach, 1907.
90. Duensing, A. and Duensing, S. Centrosomes, polyploidy and cancer, in
Polyploidization and Cancer, (ed. R.Y.C. Poon), Springer, New York, 2010,
pp. 93–103.
91. Nowell, P.C. The clonal evolution of tumor cell populations. Science,
1976;194(4260):23–8.
92. Merlo, L.M., Wang, L.S., Pepper, J.W., Rabinovitch, P.S., and Maley, C.C.
Polyploidy, aneuploidy and the evolution of cancer. Adv Exp Med Biol,
2010;676:1–13.
93. Nalesnik, M.A., Tseng, G., Ding, Y. et al. Gene deletions and amplifications
in human hepatocellular carcinomas: correlation with hepatocyte growth
regulation. Am J Pathol, 2012;180(4):1495–508.
94. Nagasue, N., Kohno, H., Chang, Y.C. et al. DNA ploidy pattern in synchro-
nous and metachronous hepatocellular carcinomas. J Hepatol, 1992;
16(1–2):208–14.
95. Saeter, G., Schwarze, P.E., Nesland, J.M., Juul, N., Pettersen, E.O., and
Seglen, P.O. The polyploidizing growth pattern of normal rat liver is
replaced by divisional, diploid growth in hepatocellular nodules and carci-
nomas. Carcinogenesis, 1988;9(6):939–45.
96. Schwarze, P.E., Saeter, G., Armstrong, D. et al. Diploid growth pattern of
hepatocellular tumours induced by various carcinogenic treatments.
Carcinogenesis, 1991;12(2):325–7.
97. Zhang, S., Zhou, K., Luo, X. et al. The polyploid state plays a tumor‐sup-
pressive role in the liver. Dev Cell, 2018;44(4):447–59 e5.
98. Castedo, M., Coquelle, A., Vitale, I. et al. Selective resistance of tetraploid
cancer cells against DNA damage‐induced apoptosis. Ann N Y Acad Sci,
2006;1090:35–49.
99. Castedo, M., Coquelle, A., Vivet, S. et al. Apoptosis regulation in tetraploid
cancer cells. EMBO J, 2006;25(11):2584–95.
100. Chowdhary, V., Teng, K.Y., Thakral, S. et al. miRNA‐122 protects mice and
human hepatocytes from acetaminophen toxicity by regulating cytochrome
 47:  Polyploidy in Liver Function, Mitochondrial Metabolism, and Cancer 613
P450 family 1 subfamily A member 2 and family 2 subfamily E member 1
expression. Am J Pathol, 2017;187(12):2758–74.
101. Nguyen, K., Jack, K., and Sun, W. Hepatocellular carcinoma: past and
future of molecular target therapy. Diseases, 2016;4(1):1.
102. Weaver, B.A. and Cleveland, D.W. Does aneuploidy cause cancer? Curr
Opin Cell Biol, 2006;18(6):658–67.
103. Turner, P.C., Sylla, A., Diallo, M.S., Castegnaro, J.J., Hall, A.J., and
Wild, C.P. The role of aflatoxins and hepatitis viruses in the etiopatho-
genesis of hepatocellular carcinoma: a basis for primary prevention in
Guinea‐Conakry, West Africa. J Gastroenterol Hepatol, 2002;17
Suppl:S441–8.
104. Slagle, B.L., Zhou, Y.Z., and Butel, J.S. Hepatitis B virus integration event
in human chromosome 17p near the p53 gene identifies the region of the
chromosome commonly deleted in virus‐positive hepatocellular carcino-
mas. Cancer Res, 1991;51(1):49–54.
105. Zasadil, L.M., Britigan, E.M., and Weaver, B.A. 2n or not 2n: aneuploidy,
polyploidy and chromosomal instability in primary and tumor cells. Semin
Cell Dev Biol, 2013;24(4):370–9.
PART FOUR:
PATHOBIOLOGY OF LIVER
DISEASE
 Hepatic Encephalopathy
48 Roger F. Butterworth
Department of Medicine, University of Montreal, Montreal, Canada
INTRODUCTION
Hepatic encephalopathy (HE) is a serious central nervous
­system (CNS) complication of liver diseases that is character-
ized by a range of neurological and neuropsychiatric symptoms
that include perturbations of psychomotor, neuro‐cognitive, and
fine motor function. Deficits in attention, visual perception,
visuo‐spatial construction in addition to motor speed and
­accuracy are among the early symptoms of HE [1]. These symp-
toms are generally considered to be potentially reversible.
The trans‐jugular intrahepatic portosystemic shunt (TIPS)
procedure is effective for the prevention and treatment of com-
plications of cirrhosis such as portal hypertension, gastrointesti-
nal (variceal) bleeding, and refractory ascites. However, the
procedure may result in new or worsening episodes of encepha-
lopathy in up to 50% of patients [2].
The burden of cirrhosis and HE is multidimensional and is
increasing. Direct economic costs related to hospitalizations,
medical procedures, and medications are substantial and to
these must be added the economic cost to the patient resulting
from lost income. Cognitive dysfunction is associated with
worsening financial status, employment prospects, and q­ uality
of life with consequent burden to the patient’s family and car-
egivers [3].
NEW CLASSIFICATION
OF HE SYNDROMES
The current practice guidelines provide a system of classifica-
tion according to the nature of the underlying disease. In this
way, HE is subdivided into three major types:
Type A: HE associated with acute liver failure
Type B: HE associated with portosystemic bypass/shunting
Type C: HE associated with cirrhosis
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
Further subdivisions then occur as a function of the time course
of HE as depicted in Table 48.1.
Episodic HE
Recurrent HE characterized by bouts of HE that occur within
a period of six months or less
Permanent HE in which symptoms are unremitting
West Haven criteria for grading
of HE severity
In patients with cirrhosis, clinically symptomatic (overt) HE
(OHE) heralds the decompensated phase of the disease and grad-
ing of the severity of OHE has traditionally made use of a battery
of neuropsychiatric tests known as the West Haven criteria where
grade I characterized by shortened attention span and sleep dis-
orders is followed by changes in personality, disorientation, and
asterixis (grade II). Grade III is typified by disorientation, confu-
sion, and semi‐stupor followed by the comatose state (grade IV).
If required, assessment of the depth and duration of coma/
impaired consciousness can be made using the Glasgow coma
scale, a points scale based upon motor responsiveness, verbal
performance, and eye opening in response to stimuli.
Minimal HE (MHE), a subtype of type C HE in which clini-
cally‐manifest symptoms are absent, occurs in up to 80% of
patients with cirrhosis and has a significant impact on health‐
related quality of life (HRQOL) of the patient [3]. MHE is diag-
nosed by one or more of a series of well‐established psychometric
tests. MHE is associated with increased numbers of falls and
road traffic accidents [3, 4].
In recent times, largely due to the inherent subjective nature
of the neuropsychiatric symptoms defining grade I HE by
West Haven, efforts were made to modify the nomenclature
and classification of HE. In the resulting AASLD/EASL
guidelines, a new term pertinent to HE classification, namely
covert HE (CHE), has appeared defined as MHE together
with what was formerly grade I HE [4]. This change was
618 THE LIVER:  NEUROPATHOLOGY

accompanied by a streamlining of the diagnosis of overt HE
into grades II, III, and IV as shown in Table 48.1.
CLINICAL PRESENTATION
AND DIAGNOSIS
HE is a diagnosis of exclusion of other metabolic disorders,
infectious diseases, intracranial vascular events, and intracranial
space‐occupying lesions all of which may potentially result in
neuropsychiatric symptomatology [5]. Covert HE is accompa-
nied by alterations of performance in psychometric tests directed
toward attention, working memory, psychomotor speed, and
visuospatial ability. Overt HE is considered to start with asterixis
or flapping tremor [6] and staging of severity in OHE by West
Haven. Identification of well‐established precipitating factors
for HE such as infection, gastrointestinal bleeding, and consti-
pation may support the diagnosis of HE in cirrhosis. For patients
with an evidently altered state of consciousness, the Glasgow
coma scale is widely employed.
Despite strong evidence for a role of ammonia in the pathogen-
esis of HE, measurements of venous ammonia add little informa-
tion by way of diagnostic or prognostic value in patients with
cirrhosis. On the other hand, arterial ammonia concentrations
Table 48.1  Updated nomenclature of hepatic encephalopathy (HE)
syndromes is dependent upon the type of HE (A, B, or C), the grade
of HE (minimal, I, II, III, IV) that is now subdivided into covert
(a combination of MHE and grade I overt) and overt (II, III, and IV).
Depending upon the time course, HE is then rated as episodic,
recurrent, or permanent and may fall into one of two categories
namely spontaneous or precipitated
Type HE grade Time course
Minimal
A I Covert Episodic
B II Recurrent
III Overt
C Persistent
IV
may have prognostic value for prediction of the severe neurologi-
cal complications of acute liver failure (ALF) [7].
Grading of HE in patients with cirrhosis using West Haven
criteria has been criticized due to its inherent subjective nature
with the likelihood to lead to high inter‐examiner variability.
One alternative technique that is gaining in popularity makes
use of critical flicker frequency (CFF), a quantitative measure
based upon the correlation between cerebral processing of oscil-
latory visual stimuli impairment and increasing severity of HE.
The CFF procedure is used extensively for the grading of neuro-
cognitive changes in a variety of neurological disorders that
includes low‐grade HE in cirrhosis where a CFF cut‐off of 39
Hz differentiates low‐grade overt HE from non‐HE patients [8].
NEUROPATHOLOGY
Glial pathology
Glial cells manifest the most conspicuous and reproducible
morphological and functional changes related to HE in both
ALF and in cirrhosis. However, the nature and extent of these
changes are determined by the type of liver injury (acute versus
chronic), the extent of portal‐systemic shunting, and the number
and duration of the episodes of HE. Glial pathology in HE
involves one of two major types of cells, namely astrocytes and
microglia.
Astrocytes
Pathological evaluation of brain sections from patients with
end‐stage ALF reveals cytotoxic brain edema as shown in
Figure 48.1a where the astrocyte manifests massive swelling of
an astrocyte end foot (A) [9]. The astrocyte swelling has the
potential to result in the development of intracranial hyperten-
sion leading ultimately to brain herniation which remains one of
the major causes of mortality in ALF.
In contrast to ALF, the cardinal neuropathological feature
of decompensated cirrhosis is a characteristic morphological

(a) (b)

Figure 48.1  Characteristic ultrastructural changes in the brain in acute and in chronic liver failure. (a) Electron micrograph from a patient with
acute liver failure due to acetaminophen overdose showing cytotoxic brain edema with marked swelling of perivascular astrocyte (A), and dilatation
of endoplasmic reticulum (arrows) and mitochondria (M). (b) Electron micrograph from a 51‐year‐old cirrhotic patient who died in hepatic coma
presenting Alzheimer type II astrocytes with a large, pale nucleus and margination of chromatin (Alz) compared with normal astrocytes with normal
chromatin pattern (N).
 48:  Hepatic Encephalopathy 619
alteration of the astrocyte nucleus known as Alzheimer type 2
astrocytosis [10]. The Alzheimer type 2 phenotype consists of
nuclear pallor and swelling of both the cell itself and its
nucleus. Margination of the chromatin pattern and deposition
of glycogen is also characteristic of the Alzheimer type 2
change as depicted in Figure  48.1b. Such changes are une-
venly distributed in brain with particularly high densities
observed in cerebral cortex, basal ganglia, and cerebellum
[11]. It is important to note that there is also evidence of brain
edema in cirrhosis although, in contrast to the situation in
ALF, the edema in cirrhosis is low grade in nature, rarely
leading to alterations of intracranial pressure. Low‐grade
edema in cirrhosis has been demonstrated in basal ganglia and
corticospinal tract of patients with cirrhosis using magnetic
resonance imaging (MRI) where decreases of magnetization
transfer ratios and increases of water apparent diffusion coef-
ficients were reported [12].
Microglia
Microglia are the immunomodulatory cells of the brain and
results of studies in animal models of acute or chronic liver fail-
ure as well as in material from patients with these conditions
confirm that activation of microglia is a common occurrence in
HE [13]. Activation of microglia is indicative of neuroinflamma-
tion and is observed in a wide range of neurodegenerative CNS
disorders including HE. Neuroinflammation associated with
microglial activation is currently considered to be a feature of
HE in both acute and chronic liver diseases [13,14]. This issue is
discussed further in a subsequent section on Neuroinflammation
in this chapter.
Neuronal pathology
Although the principal neuropathological change typical of HE
in cirrhosis is primarily glial in nature, it is important to bear in
mind that neuronal cell death frequently also occurs. Changes of
neuronal morphological and functional characteristics are not
uncommon [15]. These changes occur in distinct brain struc-
tures and have been attributed to a spectrum of diverse patho-
physiologic mechanisms. The neuropathological and clinical
characteristics of these neurodegenerative disorders are summa-
rized in the following sections.
Acquired non‐Wilsonian hepatocerebral
degeneration (ANHD)
ANHD occurs in patients with cirrhosis following a protracted
clinical course often with multiple episodes of coma and evi-
dence of portal‐systemic shunting. The neuropathology consists
of spongiform degeneration in deep cerebrocortical layers as
well as in basal ganglia structures, cerebellum, and subcortical
white matter.
Wernicke’s encephalopathy
Unsuspected Wernicke‐type hemorrhagic lesions in thalamus
occur in up to 30% of patients with end‐stage cirrhosis of alco-
holic etiology [11] where lesions may be acute or chronic (long-
standing). The higher incidence of Wernicke‐type lesions in
these patients compared to noncirrhotic alcoholics likely relates
to the fact that liver is a key site for thiamine synthesis and stor-
age in humans and Wernicke’s encephalopathy results solely
from thiamine deficiency.
Cerebellar degeneration
Mild to severe cerebellar degeneration characterized by lesions
in cerebellar vermis and severe loss of Purkinje cells occurs in
both alcoholic and non‐alcoholic patients with end‐stage cirrho-
sis although the severity of neuropathologic changes is more
severe in the alcoholics [11]. There are no clear correlations
between the incidence and extent of cerebellar degeneration,
Wernicke‐type thalamic lesions or Alzheimer type 2 astrocyto-
sis suggesting that these neuropathologic phenotypes are under-
pinned by distinct mechanisms.
Post‐shunt myelopathy
Although rare, myelopathy may occur in patients with cirrhosis
following multiple episodes of coma or following the TIPS pro-
cedure. Symptoms of spastic paresis or paralysis of the lower
limbs are apparent and result from demyelination of both direct
and crossed corticospinal tracts in this condition.
Parkinsonism in cirrhosis
Characterized by extrapyramidal symptoms (hypokinesia,
tremor, rigidity) the condition is rapidly progressive and may
occur independent of the extent of cognitive dysfunction. With a
prevalence as high as 21% in one report of patients with cirrhosis
listed for transplantation [16], the disorder has been attributed to
dopaminergic neuronal deficits resulting from manganese depo-
sition in basal ganglia [17]. Extrapyramidal symptoms in these
patients may respond to liver transplantation and, in some cases
to L‐DOPA therapy. One useful procedure that is available for
confirmation of diagnosis of Parkinsonism in cirrhosis ­consists
of T1‐weighted MRI where bilateral signal hyperintensities are
observed in globus pallidus and substantia nigra. Further discus-
sion relating to brain manganese deposition in cirrhosis appears
in the Manganese section of the current chapter.
Global brain reserve
A novel concept composed of a structural component (brain
reserve [BR]) together with a patient’s ability to tolerate changes
(cognitive reserve), BR is dependent upon disease etiology,
severity, and progression. Decreased BR examined by magnetic
resonance techniques is related to three factors namely ­structural
changes in white and gray matter of multiple brain structures,
brain edema, and altered brain metabolism.
Results of a multi‐modal MRI study reveal that patients with
alcoholic cirrhosis have, despite abstinence, a poor BR whereas
patients with non‐alcoholic cirrhosis have a greater potential for
deterioration of BR following HE or TIPS. Enhancement of the
direct effects of chronic alcohol on brain (Wernicke’s encepha-
lopathy, cerebellar degeneration) or indirectly via the liver
resulting in ANHD could form part of the basis for the structural
changes characteristic of BR. Worsening of BR as a c­ onsequence
of alcoholic etiology of cirrhosis (despite abstinence) has the
potential to modulate the impact of HE on disease p­ rogression
and suitability for transplantation [18].
620 THE LIVER:  PATHOPHYSIOLOGY
PATHOPHYSIOLOGY
Ammonia
Liver is the organ primarily tasked with the body’s removal of
excess ammonia in the form of urea via the urea cycle in peri-
portal hepatocytes and glutamine via the enzyme glutamine syn-
thetase (GS) localized in perivenous hepatocytes as shown in a
simplified manner in Figure 48.2. Patients with cirrhosis com-
monly develop intra‐ and extra‐hepatic portal‐systemic shunts
and this shunting combined with the loss of metabolic capacity
of hepatocytes results in impaired removal of ammonia and con-
sequent hyperammonemia.
One key adaptive consequence of impairment of hepatic
ammonia removal in cirrhosis is activation of the gene coding
for the enzyme protein GS in skeletal muscle [19]. This meta-
bolic adaptation provides an alternate pathway for ammonia
removal in the form of muscle glutamine production as shown
schematically in Figure 48.2.
Arterial blood and brain ammonia levels are increased sev-
eral‐fold in patients with cirrhosis and HE and dynamic 13NH3‐
positron emission tomography (PET) studies demonstrate
significant increases of brain ammonia and of the cerebral meta-
bolic rate for ammonia (CMRA, defined as the rate at which
ammonia is captured by brain and incorporated into metabo-
lites) in these patients [20]. These studies went on to demon-
strate that, contrary to earlier reports, this increase of brain
ammonia in patients with cirrhosis was primarily a consequence
of increased arterial blood concentrations rather than to altered
kinetics of blood–brain transfer of ammonia.
Venous ammonia measurements are of little predictive value
in terms of HE occurrence or severity in patients with cirrhosis.
In contrast, arterial ammonia concentrations are useful predic-
tors of high risk for the development of severe complications of
brain edema in patients with ALF such as intracranial hyperten-
sion [21] and brain herniation [7] that remains a major cause of
mortality in this condition.
The brain is devoid of an effective urea cycle. Consequently,
removal of excess ammonia by the brain depends almost
BRAIN
urea LIVER
Periportal
hepatocytes
Perivenous
hepatocytes
NH3 GUT
MUSCLE
GLUTAMINE KIDNEY
Normal
Figure 48.2  Schematic representation of inter‐organ trafficking of ammonia and glutamine under normal physiological conditions and in liver
failure. Gut‐derived ammonia is normally removed as urea (periportal hepatocytes) or glutamine (perivenous hepatocytes). In liver failure, ammonia
removal by both types of hepatocytes is decreased. Brain ammonia uptake increases but there is limited capacity for further increases in brain glu-
tamine synthesis. In contrast, skeletal muscle becomes the principal route for ammonia removal due to post‐translational increase in the enzyme
glutamine synthetase.
exclusively on the synthesis of glutamine and the enzyme
responsible, GS, is preferentially expressed by the astrocyte.
1
H‐magnetic resonance spectroscopic studies of patients with
cirrhosis reveal increased concentrations of brain glutamine as a
function of the severity of HE in these patients [22].
Consequently, the accumulation of glutamine in the astrocyte
may be causally‐related to the pathogenesis of brain edema in
ALF. However, studies in experimental animal models with
ALF resulting from toxic liver injury failed to demonstrate a
significant correlation between brain glutamine content, its syn-
thesis or turnover, and severity of HE or brain edema [23] sug-
gesting the presence of alternative or additional mechanisms.
Established mechanisms responsible for the deleterious
effects of ammonia on brain function include the following:
Direct effects of the ammonium ion (NH4 +)
on the neuronal membrane
Concentrations of ammonia equivalent to those reported in
brain in acute and/or chronic liver failure are known to exert
direct deleterious effects on both inhibitory and excitatory neu-
rotransmission [24] by virtue of the fact that, under normal
physiological conditions, ammonia exists primarily in its proto-
nated form (NH4+) an entity with comparable ionic radius to that
of potassium (K+) which is an ion with potent depolarizing
properties on the neuronal membrane.
Effects of ammonia on brain energy metabolism
Ammonia, in low millimolar concentrations equivalent to those
observed in HE is also a potent inhibitor of the rate‐limiting
tricarboxylic acid (TCA) cycle enzyme α‐ketoglutarate dehy-
drogenase [25] with the potential to result in impairment of
glucose oxidation, brain lactate production, and impending
­
brain energy failure.
Effects of ammonia on the brain GABA system
Recent studies show that ammonia is an activator of translocator
protein (TLP), a protein located on the mitochondrial membrane
BRAIN
urea LIVER
Periportal
hepatocytes
Perivenous
hepatocytes
NH3 GUT
MUSCLE
GLUTAMINE KIDNEY
Liver Failure
 48:  Hepatic Encephalopathy 621
of glial cells (astrocytes and microglia) where it serves in the
transport of cholesterol. Activation by ammonia results in
increased uptake and conversion to a novel series of compounds
known as neurosteroids (NS) one of which, allopregnanolone, is
a potent agonist of the modulatory site on the gamma‐amino
butyric acid (GABA) receptor. In this way, ammonia serves as
an activator of GABAergic transmission with the potential to
increase chloride flux and neuroinhibition. This mechanism is
covered in greater detail in the section on Inhibitory neurotrans-
mission of the current chapter.
Manganese
T1‐weighted MRI of patients with cirrhosis consistently reveals
bilateral symmetric signal hyperintensities in substructures of
the basal ganglia particularly the globus pallidus and substantia
nigra. In a landmark study of 51 patients with cirrhosis who
were listed for liver transplantation and were assessed prospec-
tively over a one‐year period, 11 patients (21.6%) exhibited
definite Parkinsonism together with typical MRI signal hyperin-
tensities. The degree of signal hyperintensity was not correlated
with the etiology of cirrhosis, Child–Pugh scores, or fasting
blood ammonia and no patients had OHE at the time of imaging.
Neurological symptoms included a symmetric akinetic‐rigid
syndrome, tremor, stooped posture, and gait impairment with
rapid progression over months [16]. Blood manganese concen-
trations were elevated up to sevenfold in all of nine patients in
whom measurements were made and cerebrospinal fluid (CSF)
manganese concentrations were increased in all of three cases in
which CSF was available. Liver transplantation resulted in nor-
malization of MRI signal hyperintensities and circulating man-
ganese levels as well as improved neurological symptoms.
Treatment of two of these patients with L‐DOPA resulted in
substantial improvements of motor function [16]. Motor impair-
ments including the akinetic‐rigid syndrome that are character-
istic of Parkinsonism in cirrhosis could contribute to the poor
performance in psychometric testing using paper and pencil
tests such as NCT‐A and NCT‐B in which there is a significant
motor component thus casting some doubt on the sole use of
these tests for the assessment of cognitive function and diagno-
sis of MHE or covert HE in patients with cirrhosis. Alternative
tests with less dependency on unimpaired motor performance
are available.
In a separate study, basal ganglia tissue obtained at autopsy
from patients with cirrhosis who died in hepatic coma were
found to contain several‐fold increased manganese concentra-
tions [26] as well as alterations of marker proteins and metabo-
lite patterns related to the dopamine (DA) neurotransmitter
system that are characteristic of idiopathic (non‐cirrhosis‐
related) Parkinson’s disease [17].
Systemic inflammation
Systemic inflammation resulting from infection and/or hepato-
cellular damage is a common occurrence in acute or chronic
liver failure and the acquisition of a systemic inflammatory
response (SIRS) is a major predictor of HE in patients with cir-
rhosis or ALF [27]. Cellular damage to liver tissue resulting
from toxins, particularly ethanol, has the potential to exacerbate
SIRS the extent of which increases as a function of the nature
and severity of liver injury. SIRS results from the release into
the circulation of proinflammatory cytokines such as tumor
necrosis factor alpha (TNFα), interleukin‐1beta (IL‐1β), and
interleukin‐6 (IL‐6). Circulating levels of TNFα are invariably
increased in patients with cirrhosis and the magnitude of the
increase correlates well with the grade of OHE [28].
Systemic inflammation resulting from lipopolysaccharide
(LPS, endotoxin) has been shown to precipitate HE and increase
blood–brain barrier (BBB) permeability in mice with ALF
resulting from toxic liver injury [29]. These findings suggest
that both cytotoxic (cell swelling) and vasogenic (BBB break-
down) mechanisms could contribute to the pathogenesis of
brain edema and its CNS complications in ALF.
The concept of synergism
Interest in synergistic mechanisms as an integral part of the
pathogenesis of HE in cirrhosis began with the pioneering work
of Les Zieve in the 1970s where synergistic actions of estab-
lished liver‐derived toxins including ammonia, methanethiol,
and octanoic acid when administered to laboratory animals
were shown to act in concert leading to HE [30]. Since that time,
evidence of synergism between three of the major entities impli-
cated in the pathogenesis of HE in cirrhosis namely ammonia,
manganese, and proinflammatory cytokines has been reported.
In patients with cirrhosis and documented infection, induc-
tion of hyperammonemia is associated with worsening of neu-
ropsychiatric status [31] and there is evidence indicating that
synergism between ammonia, manganese, and proinflammatory
cytokines represents a cascade of mechanisms resulting in the
worsening of HE [32]. The trail of evidence consists of the
­following: glutamate is the principal excitatory neurotransmitter
of mammalian brain and the termination of its action relies on
transport into surrounding astrocytes via specific transporters.
Both ammonia and manganese have been shown to inhibit these
transporters resulting in an excess of glutamate in the synaptic
cleft leading to hyperexcitability and activation of post‐synaptic
glutamate receptors. This cascade results in nitration and conse-
quent inactivation of key proteins via a process known as pro-
tein tyrosine nitration (PTN) [33]. One such tyrosine‐nitrated
protein is the enzyme GS that is solely responsible for ammonia
removal by the brain. Under conditions of PTN, brain ammonia
levels become increased and the synergistic cycle is amplified.
Neuroinflammation
Recent studies provide convincing evidence in support of a
novel concept namely the role of neuroinflammation (inflam-
mation of the brain per se) in the pathogenesis of HE in patients
with ALF [34]. Neuroinflammation in ALF is characterized by
activation of microglial cells together with increased production
of proinflammatory cytokines such as TNFα and the interleu-
kins IL‐1β and IL‐6 in the brain. The grade of HE severity cor-
relates well with that of proinflammatory cytokine production
and the use of proinflammatory gene deletion strategies results
in slowing of HE progression in experimental ALF [35].
Microglial activation has also been reported in autopsied brain
tissue samples from ALF patients with grade IV HE [13].
622 THE LIVER:  PATHOPHYSIOLOGY
Microglial activation also occurs in patients with cirrhosis who
died in hepatic coma [14].
Activated microglia are known to express transcripts for the
mitochondrial protein translocator protein (TLP) and using
PET and the selective TLP ligand 11C‐PK11195, activation
of  microglia indicative of a neuroinflammatory response is
observed in patients with cirrhosis and MHE. Microglial acti-
vation appears to be a relatively early occurrence in patients
with decompensated cirrhosis. Particularly intense signals were
observed in the anterior cingulate cortex, a brain structure
known to be associated with the control of attention in MHE
patients [36].
Liver‐brain proinflammatory signaling occurs in HE and a
growing list of possible mechanisms have been proposed [34].
At the cellular level, human cerebrovascular endothelial cells
exposed to TNFα manifest increased transport of ammonia [37]
and cultured neural cells exposed to combinations of ammonia
and recombinant proinflammatory cytokines display increased
expression of genes coding for proteins implicated in cellular
dysfunction in HE suggestive of synergism. Factors other than
ammonia and manganese that are increased in brain in liver fail-
ure also have the potential to elicit a neuroinflammatory
response. For example, millimolar concentrations of lactate are
known to substantially increase the release of TNFα and IL‐6
from cultured microglial cells [38].
Brain energy metabolism
There is no convincing evidence from either biochemical or
spectroscopic investigations in support of the hypothesis that
HE is primarily caused by a sufficiently severe loss in concen-
tration of high energy phosphates indicative of cerebral energy
failure. However, alterations of uptake and metabolism of the
brain’s primary energy substrate, glucose, have consistently
been described in both acute and chronic liver diseases. Studies
using PET and the non‐metabolized glucose transport ligand
18
fluorodeoxyglucose in patients with cirrhosis and mild HE
show significant decreases in uptake in anterior cingulate cor-
tex, a brain structure implicated in the monitoring of responses
to visual stimuli [39]. Not surprisingly, decreased brain glu-
cose uptake in these patients was correlated with impaired per-
formance on psychometric testing. This decrease of brain
glucose uptake is currently considered to be the conse-
quence of decreased brain energy requirements resulting from
decreased neural activation in (rather than the cause) of HE
[39]. Pathophysiologically‐relevant concentrations of ammo-
nia inhibit the TCA cycle, an essential metabolic pathway
involved in the maintenance of brain energy requirements
[25]. Slowing of the TCA cycle results in the shuttling of pyru-
vate to lactate and cerebrospinal fluid lactate concentrations
correlate well with severity of HE in patients with cirrhosis
[40] as well as in brain microdialysates from patients with
ALF where surges of increased intracranial pressure were fre-
quently found to accompany increased lactate concentrations
[41]. Brain lactate concentrations in excess of 12 mM have
been reported at coma stages of encephalopathy in animal
models of ALF [25]. Such concentrations are beyond
the brain’s capacity to buffer lactate with the potential to result
in acidosis.
Cerebral blood flow (CBF)
Although global CBF in patients with cirrhosis is generally
somewhat reduced, PET studies reveal that CBF changes in
these patients with mild HE occurred in a region‐selective
manner whereby flow to cerebral cortical regions was
decreased while flow to basal ganglia, cerebellar, and thalamic
structures was significantly increased. Since this pattern of
changes in CBF parallels the regional changes of brain glucose
utilization, CBF autoregulation (defined as the capacity of
CBF to match brain activity independent of changes of sys-
temic arterial pressure) appears to be preserved in patients
with cirrhosis [42].
In contrast to the situation in cirrhosis, loss of CBF autoreg-
ulation is common in patients with ALF where the variability of
CBF rates are influenced by a variety of factors including the
complex interplay between local energy demands, ammonia
neurotoxicity, systemic arterial pressure, and intracranial
hypertension. A five‐phase classification system based upon a
retrospective analysis of sequential intracranial pressure (ICP)
and CBF measurements in ALF patients has been proposed
[43] as follows:
Phase 1: An initial phase of low CBF resulting from lower neu-
ronal activity
Phase 2: A progressive increase of CBF
Phase 3: A subsequent increase of ICP
Phase 4: Final stage where increased ICP results in decreased CBF
Phase 5: Brain death
The classification integrates well with previously‐published
data; it has a mechanistic correlation and is expected to facilitate
prognostic evaluation in patients with ALF.
Cell volume regulation and brain edema
Brain edema has been described in patients with cirrhosis and in
patients with ALF. However, the nature and pathological char-
acteristics of the brain edema in the two conditions are quite
distinct. In ALF brain edema has been characterized consisting
primarily of cytotoxic brain edema that, if uncontrolled, may
progress to intracranial hypertension (ICH) and brain hernia-
tion. Brain edema in ALF appears to have multiple causes
including the accumulation of brain lactate [25] as described
above as well as proinflammatory mechanisms resulting from
microglial activation and release of cytokines [34].
Brain edema leading to ICH is uncommon in cirrhosis since
the swelling is low‐grade. However, it has the potential to result
in impairments of cell–cell signalling and to cause malfunction
of astrocytic proteins. NMR spectroscopic studies in patients
with cirrhosis have consistently shown disturbances of brain
organic osmolytes including increases of glutamate/glutamine
with compensatory reductions of myo‐inositol [44].
Inhibitory neurotransmission
GABA is the principal inhibitory neurotransmitter system of
mammalian brain and alterations of GABAergic transmission
are implicated in a wide range of neurodegenerative and meta-
bolic brain diseases including HE.
 48:  Hepatic Encephalopathy 623
GABAergic neurotransmission is mediated by the amino acid
GABA via activation of a post‐synaptic GABA‐receptor com-
plex (GRC), a protein complex and ligand‐gated ion channel
that is selective for chloride (Cl−) ion. Activation of the GRC by
GABA results in the entry of chloride through its pore resulting
in hyperpolarization of the post‐synaptic neuron. This results in
inhibition of central neurotransmission by diminishing the
chance that an action potential occurs.
Activation of GABAergic transmission (also frequently
referred to as “increased GABAergic tone”) was originally pro-
posed in the 1980s based on visual evoked response patterns in
experimental animals with toxic liver injury and HE that were
identical to patterns observed in normal animals treated with
well‐established activators of the GRC [45]. These findings led
to intense interest in activation of the GRC as a major factor
implicated in the pathogenesis of HE, an interest that continues
to this day. Initially, studies were focused on the measurement
of integral components of the GABA system including GABA
concentrations and metabolites, activities of enzymes responsi-
ble for GABA synthesis as well as post‐synaptic GABA recep-
tors. In all cases, these parameters were found to be present in
normal amounts in animal models of acute and chronic liver
failure and, more importantly, in autopsied brain tissue from
patients with decompensated cirrhosis who died in hepatic coma
[46]. These findings led to a change in focus of research in this
area and a renewal of interest in the multiple modulatory sites
and associated subunit proteins of the GRC itself.
The GRC is composed of an active binding site for GABA,
often referred to as the GABA recognition site that, together with
allosteric sites, modulate the activity of the GRC. These sites are
targets for a range of substances including benzodiazepines and
neuroactive steroids also known as neurosteroids (NS).
The benzodiazepine modulatory site on the GRC has been
extensively studied in brain tissue obtained at autopsy from
patients with cirrhosis who died in stage IV HE as well as in
living patients with mild HE by use of the selective benzodi-
azepine modulatory site antagonist PET ligand Ro15‐1788
(flumazenil). No significant alterations of binding site affini-
ties or densities of these sites were observed when compared
to data from age‐matched control material [46]. Furthermore,
subsequent studies failed to demonstrate any significant alter-
ations of the allosteric coupling between the benzodiazepine
modulatory site and the GABA recognition in the HE patient
material [47]. Together, these findings clearly demonstrate
that the GRC with respect to the benzodiazepine modulatory
site is functioning normally in patients with HE related to
cirrhosis.
The reports of no significant alterations in expression or cou-
pling of the GABA receptor machinery of the brain in HE led to
a renewed interest in the search for so‐called “endogenous
ligands” for these receptors. Several such substances were
detected. Quantitative mass spectrometric analysis followed to
unequivocally establish the identity of these substances. They
were identified as well‐established pharmaceutical benzodiaz-
epines or their metabolites [48] that were traced to their admin-
istration as muscle relaxants during prior endoscopic work‐up
or as sedatives.
In contrast to the disappointing findings in relation to the
benzodiazepine modulatory site and its potential role in the
pathogenesis of HE in chronic liver diseases, investigations of
the other modulatory site on the GRC, the NS modulatory site in
HE patient material yielded more positive results.
The molecular steps involved in NS synthesis in the brain and
the relationship between ammonia exposure and NS synthesis is
shown in a simplified schematic form in Figure 48.3.
The process starts with the activation of a mitochondrial pro-
tein located in glia (astrocytes and microglia) known as translo-
cator protein (TLP) resulting in the entry of cholesterol into the
mitochondrion. A series of well‐established metabolic steps
then transform cholesterol into NS that are then available for
release into the synaptic cleft.
Two of these NS namely allopregnanolone (ALLO) and tet-
rahydro‐deoxy corticosterone (THDOC), are potent positive
allosteric modulators of the GRC and, hence, are activators of
GABAergic neurotransmission leading to enhanced central
neuroinhibition.
The first direct evidence for a role of NS in the pathogenesis
of HE in cirrhosis was provided by the report of significant
increases of ALLO in brain tissue obtained at autopsy from
patients with decompensated cirrhosis who died in stage IV HE
(coma) [49]. ALLO concentrations in control material from
patients with no liver disorders, patients with decompensated
cirrhosis but no encephalopathy, as well as one patient who died
in uremic coma, were all within normal limits. Furthermore,
ALLO concentrations in brain extracts from HE patients were
Ammonia Cholesterol
TLP
Perineuronal GABA
Pregnenolone
astrocyte MM Nerve
or
SER
Terminal
microglia
Astrocyte
Allopregnanolone GABA
GABA
NS
site
site
Post-Synaptic
neuron
Cl
Neuroinhibition
HE
Figure 48.3  Pathophysiologic link between ammonia toxicity and
“increased GABAergic tone” in HE. A schematic representation of the
sequence of steps whereby activation of the translocator protein (TLP)
situated on the mitochondrial membrane (MM) of the astrocyte or
microglial cell occurs. This results in increased transport of cholesterol
into the cell and its conversion to pregnenolone and the NS allopregna-
nolone. Stimulation of the NS modulatory site on the GABA‐A receptor
complex by allopregnanolone then amplifies the signal produced by the
neurotransmitter (GABA) acting on the adjacent GABA recognition
site. The net result is increased entry of chloride (Cl−) and enhanced
neuroinhibition that contributes to the pathogenesis of HE in cirrhosis.
624 THE LIVER:  MANAGEMENT AND THERAPEUTIC STRATEGIES FOR HE
present in sufficiently high concentrations to result in a 53%
increase in binding of the GABA agonist ligand 3H‐muscimol to
brain membrane preparations thus satisfying the requirement
that concentrations of ALLO in HE patient material were suffi-
cient to result in “increased GABAergic tone” [50]. Novel com-
pounds with antagonist action at the NS modulatory site on the
GRC have recently been synthesized. Further information and
results of preliminary clinical trials are summarized in the sec-
tion, GABA receptor modulators in this chapter.
MANAGEMENT AND THERAPEUTIC
STRATEGIES FOR HE
The ultimate test of the pertinence of the diverse theories of the
pathogenesis of HE rests in the clinic where the impact of
appropriate treatments based upon the nutritional, metabolic
and toxic factors identified in this chapter may be undertaken.
Nutritional management
The pathogenesis of malnutrition in liver disease is complex and
multifactorial involving reduction of dietary intake due to ano-
rexia and dietary restriction, alterations of nutrient biosynthesis,
impaired intestinal absorption, disturbances of substrate utiliza-
tion, increased protein loss, and abnormal metabolism resulting
from, for example, the proinflammatory state.
The functional integrity of the liver is essential for the provi-
sion, inter‐organ transfer, and metabolism of essential nutrients.
As outlined in the present review, nitrogen metabolism plays a
key role in the development of HE in cirrhosis and in conditions
of liver failure, skeletal muscle adapts metabolically to serve as
the major back‐up organ for ammonia removal. Sarcopenia or
loss of muscle mass is commonly encountered in cirrhosis where
it has adverse effects on survival, health‐related quality of life, as
well as outcome following liver transplantation [51]. Sarcopenia
may also lead to worsening of the complications of cirrhosis
including portal hypertension, ascites, and HE. Indeed, results of
a prospective study reveal that muscle depletion in patients with
cirrhosis increases the risk of both MHE and OHE [52].
Obesity and its associated alterations of nutrient intake are
frequent occurrences in patients with non‐alcoholic fatty liver
disease (NAFLD) and play an important role in determining
rates of progression to non‐alcoholic steatohepatitis (NASH)
and cirrhosis. A recent consensus document from the
International Society for Hepatic Encephalopathy and Nitrogen
Metabolism (ISHEN) on guidelines for the nutritional manage-
ment of HE in cirrhosis has appeared [53]. The stated objectives
of nutritional management in patients with cirrhosis include the
correction of specific nutritional deficiencies, prevention and
treatment of the complications of cirrhosis as well as the com-
plications of liver transplantation, and the support of liver
regeneration.
Energy requirements of 35–45 kcal g−1 daily are recom-
mended for patients with cirrhosis with small meals evenly dis-
tributed throughout the day and a late‐night snack of complex
carbohydrates to minimize protein utilization. The practice of
dietary protein restriction that was prevalent until the late 1990s
[54] has now been widely discontinued following the report [55]
that patients with cirrhosis benefit from normal protein (1.2–1.5 g
kg−1 daily). Diets rich in vegetable and dairy protein are
encouraged and patients intolerant to dietary protein may con-
sider branched chain amino acid supplements as an alternative
to protein [53].
Vitamin deficiencies in cirrhosis result from impaired hepatic
function and diminished hepatic reserves as well as inadequate
dietary intake and malabsorption. A neuropathologic study
examining brain tissue from patients with cirrhosis who died in
stage IV HE revealed lesions in thalamus and mamillary bodies
that are characteristic of Wernicke’s encephalopathy together
with cerebellar degeneration consistent with vitamin B1 defi-
ciency [11]. The B1 deficiency in these patients was attributed to
a loss of hepatic stores of the vitamin and the prevalence of
lesions was threefold higher than that reported in material from
non‐liver disease controls. It was recommended that B1 supple-
ments be administered to all patients with decompensated cir-
rhosis particularly those with alcoholic etiology [51].
Deficiencies of other vitamins including B2, B6, and B12 have
also been reported in cirrhosis associated with alcoholic liver
damage and/or diminished hepatic storage but causal relation-
ships with complications of cirrhosis were not established. Zinc
deficiency is common in cirrhosis but evidence for beneficial
effects of zinc supplements on HE is equivocal [56].
It has been suggested that malnutrition is a factor that
increases both costs and post‐transplant complications [57].
Neurological complications post transplantation are numerous
and may include diffuse encephalopathy, seizures, intracranial
hemorrhage, stroke, progressive neurological deterioration, cen-
tral pontine myelinolysis, ataxia, psychosis, confabulation, and
peripheral neuropathy. While some of these complications
likely result from specific nutritional and metabolic deficits
alluded to above, further studies are required to determine the
precise nutritional factors implicated and facilitate specific
nutritional recommendations.
Ammonia‐lowering strategies
Given the wealth of evidence derived from biochemical, neuro-
imaging, and spectroscopic studies in support of a central role of
ammonia in the pathogenesis of HE, it is not surprising that the
therapeutic strategy that has so far stood the test of time relates
to the reduction of hyperammonemia. Several treatments aimed
at reducing the production of ammonia and/or its removal have
been found to be effective in this regard.
Non‐absorbable disaccharides
Non‐absorbable disaccharides such as lactulose and lactitol
remain first‐line agents for the lowering of the production and
absorption of ammonia. They are metabolized by bacteria in the
colon with production of acetic and lactic acids and the conse-
quent acidification of the colon creates a hostile environment for
urease‐rich intestinal bacteria that are responsible for the pro-
duction of ammonia. In addition, these agents facilitate the for-
mation of NH4+ leading to reduced ammonia production and
their cathartic properties may result in up to fourfold increases
of fecal nitrogen excretion [58].
 48:  Hepatic Encephalopathy 625
Although non‐absorbable disaccharides are widely employed
as first‐line therapy for HE in cirrhosis, systematic reviews and
meta‐analyses of trials have yielded mixed results. In an analy-
sis of ten studies, it was concluded that there was insufficient
evidence to recommend them for standard therapy until they
were shown to confer benefit over placebo [59]. The analysis
was subsequently challenged based upon shortcomings in
experimental design and methodology [60] and further studies
including a second meta‐analysis of lactulose revealed signifi-
cant improvements in cognitive function and health‐related
quality of life in MHE [61] as well for the primary prophylaxis
of overt HE [62] in patients with cirrhosis. A subsequent meta‐
analysis showed that, compared with placebo/no intervention,
the non‐absorbable disaccharides were associated with signifi-
cant overall benefits on HE [63].
Antibiotics
The aminoglycoside antibiotic neomycin is still used for the
treatment of patients with cirrhosis and HE despite its poten-
tially serious side‐effects that include ototoxicity and nephro-
toxicity since the drug is partially absorbed. Rifaximin is a
poorly absorbed, broad spectrum antibiotic that is effective for
lowering of blood ammonia and improvement of mental state in
patients with MHE as well as in those with overt HE where its
efficacy has been found to exceed that of neomycin. Results of
a large, randomized, double‐blind, placebo‐controlled trial
demonstrated that rifaximin reduced the risk of repeat episodes
of overt HE and reduced the time to first hospitalization with no
serious adverse events [64]. A subsequent systematic review and
meta‐analysis of 19 trials with 1390 patients confirmed that
rifaximin is useful for the management of HE with a beneficial
effect on HE severity and for treatment of acute episodes as well
as on patient mortality [65]. Rifaximin also improves psycho-
metric test scores and HRQOL in patients with MHE [66].
Probiotics
Urease‐producing bacteria are plentiful in the gut where they
convert urea to carbamate and ammonia. Modifications of the
gut microflora balance by the introduction of bacteria that com-
pete with those expressing urease (probiotics) are effective
ammonia‐lowering agents. Treatment of patients with non‐alco-
holic cirrhosis and MHE with probiotic yoghurt results in slow-
ing of progression to overt HE [67]. Furthermore, randomized
controlled trials comparing probiotics with placebo reveal
improvements in arterial ammonia concentrations, psychomet-
ric test scores, and HRQOL in MHE patients [68, 69]. The
degree of improvement was equivalent to that of lactulose [68]
or rifaximin [69]. A subsequent meta‐analysis of 14 studies
totaling 1152 patients confirmed the comparable beneficial
effects of probiotics and lactulose and went on to demonstrate
efficacy of probiotics in reducing hospitalization rates as well as
rates of progression to overt HE.
L‐ornithine L‐aspartate (LOLA)
LOLA is a mixture of two endogenous amino acids with estab-
lished ammonia‐lowering properties. Multiple mechanisms
responsible for the ammonia‐lowering actions of LOLA include
an ability to increase urea production by residual hepatocytes
(L‐ornithine is a urea cycle substrate) and increased glutamine
synthesis in skeletal muscle [70]. There is evidence that LOLA
has direct hepatoprotective actions in patients with cirrhosis
and HE [71].
Beneficial effects of LOLA for the management and treat-
ment of HE in cirrhosis have been reported in over 20 rand-
omized clinical trials over the last 20 years. The intravenous
formulation of LOLA is particularly effective in the treatment of
low‐grade [72] and bouts of overt [73] HE in cirrhosis. A recent
meta‐analysis demonstrated that the oral formulation of LOLA
is particularly effective for the treatment of MHE [74].
LOLA is reportedly equivalent or superior in efficacy com-
pared to rifaximin [69], probiotics [68, 69], and lactulose [68]
for the treatment of MHE and a network meta‐analysis con-
firmed these findings [75]. However, in the only trial conducted
to date, LOLA did not appear to be effective for the treatment of
HE in ALF [76].
Branched chain amino acids (BCAAs)
Results of a Cochrane review on the efficacy of BCAAs for
improvement of mental state concluded that patients treated
with BCAAs are more likely to recover from HE compared to
patients receiving control regimens that included lactulose and
neomycin [77]. Two large studies revealed beneficial effects of
BCAAs on nutrition and survival, effects that may outweigh any
effects on HE per se [77, 78]. Mechanisms responsible for the
beneficial effects of BCAAs have not been fully elucidated
despite their continuing use for over 35 years. Suggested mech-
anisms include beneficial effects on their use as substrates for
hepatic protein synthesis as well as stimulation of liver regen-
eration and increases of cerebral perfusion.
Benzoate, phenylacetate
Benzoate and phenylacetate are agents that have been used
extensively for the lowering of blood ammonia in children
with congenital urea cycle disorders. Phenylacetate is gener-
ally given orally as its precursor phenylbutyrate. Phenylacetate
condenses with glutamine to form phenylacetyl glutamine
whereas benzoate condenses with glycine to form hippurate.
Both condensation products are excreted in the urine. The
ammonia‐lowering action results from removal of these
ammoniagenic substrates as well as the diversion of available
ammonia for the replenishment of the amino acid pools.
Sodium benzoate was reported to manifest comparable ­efficacy
to lactulose in a prospective study in patients with cirrhosis
and acute HE [79].
Mixtures of phenylacetate with other agents such as glycerol
or L‐ornithine have recently been introduced as possible ammo-
nia‐lowering strategies for possible use in the treatment of HE in
cirrhosis. In the case of glycerol phenylbutyrate (GBP), a multi-
center phase 2 trial, patients in the GBP arm manifested lower-
ing of blood ammonia and experienced less HE events as well as
less HE hospitalizations compared to placebo [80].
Ornithine phenylacetate (OP), a mixture of L‐ornithine and
phenylacetate has been shown to reduce blood ammonia in
experimental animal models of HE, a result that is not surprising
given that both the L‐ornithine and phenylacetate moieties are
626 THE LIVER:  MANAGEMENT AND THERAPEUTIC STRATEGIES FOR HE
independently well‐established ammonia‐lowering compounds.
A study on the safety, tolerability, and pharmacokinetics of OP
in patients with acute liver injury or failure showed significant
urinary ammonia excretion, good tolerability, and no safety sig-
nals [81]. Randomized, controlled studies are now required to
determine its use as an ammonia‐scavenging agent in patients
with HE.
L‐carnitine
L‐carnitine has been shown to manifest a significant protective
effect in patients with cirrhosis and ammonia‐precipitated
encephalopathy [82]. The protective effect of L‐carnitine was
evident for patients with overt HE grades I and II as well as for
patients with MHE as assessed by improvements in scores for
NCT‐A. Studies in both experimental congenital hyperam-
monemia [83] and portal‐systemic encephalopathy [84] showed
a significant protective effect of L‐carnitine on brain energy
metabolism characterized by improved brain glucose oxidative
capacity and prevention of cerebrospinal fluid lactate accumula-
tions [84]. As alluded to earlier in this chapter, decreased glu-
cose oxidation in brain in liver failure is the consequence of
ammonia‐induced inhibition of the TCA cycle by ammonia
leading to lactate accumulation [25].
These novel beneficial effects of L‐carnitine are the conse-
quence of inhibition of one of the major metabolic consequences
of exposure to ammonia namely that of decreased brain glucose
oxidation rather than a direct effect on ammonia production or
removal.
Neuropharmacological approaches
GABA receptor modulators
A clinical trial of the efficacy of flumazenil, a potent antagonist
of the benzodiazepine modulatory site on the GABA receptor
complex was undertaken in patients with cirrhosis and moderate
to severe HE. The trial demonstrated clinical improvement of
HE in a significant subset of patients [85] and this was subse-
quently confirmed in a systematic review and meta‐analysis
[86] although, as expected, the benefit was relatively short‐act-
ing due to the short half‐life of flumazenil.
The findings of increased brain concentrations of allopreg-
nanolone (described in the section, Inhibitory neurotransmis-
sion in the current chapter) the potent agonist of the NS site on
the GABA‐A receptor in material from HE patients resulted in
a renewed interest in the discovery of compounds with the nec-
essary configuration to block the actions of NS on the NS
­modulatory site on the GRC [47]. So‐called “GABA‐receptor‐
modulating steroid antagonists” (GAMSA) were synthesized
and tested in animal models. One example, 3β‐20β‐
dihydroxy‐5α‐pregnane (UC1011) was shown previously to
inhibit the ALLO‐induced increases of GABA‐induced
increases of chloride uptake into cerebral cortical and hip-
pocampal preparations making it an ideal candidate for further
study [87]. More recently, a second member of the GAMSA
family, GR 3027 has been shown to improve spatial learning,
motor coordination, and circadian rhythms in an experimental
model of HE in chronic liver disease [88]. Controlled clinical
trials are now required.
Anti‐inflammatory agents
Recognition of the key roles of systemic inflammation and neuro-
inflammation in the pathogenesis of HE in both ALF and in cir-
rhosis has resulted in renewed interest in the search for agents with
anti‐inflammatory properties as potential therapies. Ibuprofen
attenuates the neurologic deficits in learning ability in an experi-
mental animal model of type B HE [89]. A central role of TNFα has
been proposed to explain HE in cirrhosis based upon the observa-
tions that many precipitating factors including infection, gastroin-
testinal bleeding, constipation, and renal failure are associated with
increased circulating levels of the cytokine. Moreover, many thera-
pies known to lower ammonia such as lactulose, antibiotics, and
probiotics also effectively reduce TNFα production [28].
Following up on the report that deletion of the gene coding
for the TNFα receptor resulted in delayed onset of HE and pre-
vention of brain edema in ALF due to toxic liver injury [35], a
more recent study showed that the systemic sequestration of
TNFα by administration of the TNFα‐neutralizing compound
etanercept resulted in reduction of microglial activation and
delayed progression of HE in mice with acute liver injury [90].
Minocycline is a semi‐synthetic tetracycline antibiotic that
limits microglial activation by a mechanism that is independent of
its antimicrobial properties and studies in the liver ischemic rat
model of ALF showed that minocycline prevents microglial acti-
vation and attenuates the increased synthesis of proinflammatory
cytokines including TNFα in the brain while leading to a slowing
of progression of HE and prevention of brain edema [91].
The beneficial effect of mild hypothermia in ALF involves
anti‐inflammatory mechanisms at both hepatic and cerebral
levels [92].
Other centrally‐acting agents
Evidence of a dopaminergic deficit together with neurological
symptoms characteristic of Parkinson’s disease are well established
findings in patients with HE associated with cirrhosis. It is there-
fore not surprising that attempts have been made to treat these
patients with agents with the potential to reactivate the dopaminer-
gic system. Dopamine‐replacement therapy by the precursor amino
acid L‐DOPA was found to be ineffective for improvement of cog-
nitive symptoms [93] but, subsequently, using a more appropriate
Parkinson disease rating scale for assessment of clinical efficacy,
L‐DOPA was found to result in improvements in motor perfor-
mance [16]. Clinical trials with the dopamine receptor agonist bro-
mocriptine have yielded conflicting results [94, 95] but, again,
clinical endpoints such as the portal‐systemic encephalopathy
(PSE) index rather than tests aimed at assessment of dopaminergic
function and motor coordination were used making interpretation
of the findings problematic. Therapeutic options for the treatment
of Parkinsonism in cirrhosis have been reviewed [96].
Liver transplantation and HE
Liver transplantation (LT) is the ultimate treatment for decom-
pensated cirrhosis. Although LT generally confers a level of
reversibility of some of the symptoms of HE, there is a growing
body of evidence suggesting that this reversibility is far from
complete [97] and patients with a history of overt HE have an
increased probability of persistent neurocognitive dysfunction
 48:  Hepatic Encephalopathy 627
post‐LT compared to those without prior overt HE. This finding
begs the question whether it is the episode(s) of overt HE or the
LT that is responsible for the persistent deficits. This issue was
subsequently addressed in a prospective study in which it was
demonstrated that even a single episode of overt HE was accom-
panied by a learning deficit and that multiple episodes led to
deficits in reaction times, attention, psychomotor speed, and
working memory [98]. Overt HE patients may suffer a loss of
“cognitive reserve” that this could have important implications
for the assignment of priority for LT. Other possible causes of
post‐LT cognitive dysfunction include anoxic intraoperative
complications, immunosuppressant medication, comorbidities,
manganese neurotoxicity [17], and thiamine deficiency [11].
In a study comparing the incidence and severity of neurologi-
cal complications after cadaveric versus living donor LT, the most
common complications (encephalopathy and seizures) were sig-
nificantly lower following living donor LT. These findings could
be accounted for, at least in part, to the shorter cold ischemia time.
The incidence of neurological complications post‐LT was not
dependent on the nature of the immunosuppressant used [99].
Future directions for the field of HE
When the author of this chapter began his research in HE some
four decades ago, there were two major alternative agents
employed for the management and treatment of HE in cirrhosis;
a non‐absorbable disaccharide and an aminoglycoside antibi-
otic. Both acted at the level of the gut, both resulted in lowering
of blood ammonia; both had serious side‐effects. Many of us
wonder if anything has changed. Well, the disaccharide is still
with us but at least we appear to have found a better antibiotic.
It appears that the search for agents, some new, some recycled,
targeting the lowering of blood ammonia is here to stay.
Progress in the search for more rational therapies for HE is
slow. Despite evidence to the contrary, it is widely believed that
HE results principally, if not entirely, from the neurotoxic
effects of ammonia. Consequently, search for new therapies
remains focused on the gut.
HE is, after all, a brain disorder and advances in our under-
standing of intercellular signalling involving defined neuro-
transmitter and neuromodulators in brain areas implicated in the
maintenance of consciousness, cognitive, and motor function in
liver disorders continue to be elucidated as do novel concepts
involving central anti‐inflammatory molecules. Yet translation
of these discoveries, and sometimes the use of new molecules
themselves, remains sluggish. One exception presented in this
chapter relates to the discovery of the effectiveness of an antag-
onist of the neurosteroid modulatory site on the GABA‐A recep-
tor complex of brain. Clinical trials are ongoing.
The future directions of HE research and patient care will
undoubtedly be based to an increasing extent on targeting of the
brain per se.
ACKNOWLEDGMENTS
Studies and related publications from the author’s research unit
were funded by operating grants from the Canadian Institutes of
Health Research (CIHR) and the Canadian Association for the
study of the Liver (CASL). The author is grateful to Mr Jonas
Eric Pilling for the design of Figure 48.3.
REFERENCES
1. Mullen, K.D. and Prakash, R.K. (Eds.) Hepatic Encephalopathy, Clinical
Gastroenterology, Humana Press/Springer New York, pp. 97–102
2. Rossle, M. and Euringer, W. Hepatic encephalopathy in patients with tran-
sjugular intrahepatic portosystemic shunt (TIPS), in Hepatic encephalopa-
thy: Clinical Gastroenterology, (eds. K.D. Mullen and R.K. Prakash),
Humana Press/Springer, New York, 2012, pp. 211–20.
3. Bajaj, J.S., Wade, J.B., Gibson, D.P. et al. The multi‐dimensional burden of
cirrhosis and hepatic encephalopathy on patients and caregivers. Am J
Gastroenterol, 2011;10:6646–53.
4. Bajaj, J.S. Introduction and setting the scene: new nomenclature of hepatic
encephalopathy and American Association for the study of liver diseases/
European Association for the study of the liver guidelines. Clin Liv Dis,
2017;9:48–51.
5. Blei, A.T. and Cordoba, J. Hepatic encephalopathy practise guidelines. Am J
Gastroenterol, 2001;96:1968–75.
6. Bajaj, J.S., Cordoba, J., Mullen, K.D. et  al. Review article: the design of
clinical trials in hepatic encephalopathy‐an ISHEN consensus statement.
Aliment Pharmacol Ther, 2011;3:339–47.
7. Clemmesen, J.O., Larsen, F.S., Kondrup, J. et  al. Cerebral herniation in
patients with acute liver failure is correlated with arterial ammonia concen-
tration. Hepatology, 1999;2:948–53.
8. Kircheis, G., Wettstein, M., Timmermann, L. et al. Critical flicker frequency
for quantification of low‐grade hepatic encephalopathy. Hepatology,
2002;35:357–66.
9. Kato, M., Hughes, R.D., and Keays, R.T. Electron microscopic study of
brain capillaries in cerebral edema from fulminant hepatic failure.
Hepatology, 1992;15:1060–6.
10. Butterworth, R.F., Giguere, J.F., Michaud, J. et al. Ammonia: key factor
in the pathogenesis of hepatic encephalopathy. Neurochem Pathol,
1987:6–12.
11. Kril, J.J. and Butterworth, R.F. Diencephalic and cerebellar pathology in
alcoholic and non‐alcoholic patients with end‐stage liver disease.
Hepatology, 1997;2:637–41.
12. Cordoba, J., Alonso, J, Rovira, A. et  al. The development of low‐grade
cerebral edema in cirrhosis is supported by the evolution of 1H‐magnetic
resonance abnormalities after liver transplantation. J Hepatol,
2001;3:598–604.
13. Butterworth, R.F. Hepatic encephalopathy: a central neuroinflammatory dis-
order? Hepatology, 2011;53:372–6
14. Zemtsova, I., Gorg, B., Keitel, V. et  al. Microglial activation in hepatic
encephalopathy in rats and humans. Hepatology, 2011;5:404–15.
15. Butterworth, R.F., Neuronal cell death in hepatic encephalopathy. Metab
Brain Dis, 2007;2:209–20.
16. Burkhard, P.R., Delavelle, J., Du Pasquier, R. et al. Chronic Parkinsonism
associated with cirrhosis. Arch Neurol, 2003;60:521–8.
17. Butterworth, R.F., Spahr, L., Fontaine, S. et al. Manganese toxicity, dopa-
minergic dysfunction and hepatic encephalopathy. Metab Brain Dis,
1995;10:559–67.
18. Ahluwalia, V., Wade, J.B., Moeller, F.G. et al. The etiology of cirrhosis is a
strong determinant of brain reserve: a multi‐modal MR imaging study. Liver
Transpl, 2015;2:1123–32.
19. Desjardins, P., Rao, K.V., Michalak, A. et al. Effect of portacaval anastomo-
sis on glutamine synthetase protein and gene expression in brain, liver and
skeletal muscle. Metab Brain Dis, 1999;14:273–80.
20. Sorensen, M. and Ott, P. Cerebral ammonia metabolism in cirrhosis. Funct
Mol Imaging Hepatol, 2012:153–9.
21. Bernal, W., Hall, C., Karvellas, C.J. et al. Arterial ammonia and clinical risk
factors for encephalopathy and intracranial hypertension in acute liver fail-
ure. Hepatology, 2007;46:1844–52.
22. Laubenberger, J., Haussinger, D., Boyer, S. et al. Proton magnetic resonance
spectroscopy of brain in symptomatic and asymptomatic patients with liver
cirrhosis. Gastroenterology, 1997;112:1610–6.
628 THE LIVER:  REFERENCES
23. Zwingmann, C., Chatauret, N., Liebfritz, D. et al. Selective increase of brain
lactate synthesis in experimental acute liver failure: results of a 13C/1H
nuclear magnetic resonance study. Hepatology, 2003;3:420–8.
24. Felipo, V. and Butterworth, R.F. Neurobiology of ammonia. Prog Neurobiol,
2002;67:259–79.
25. Lai, J.C.K. and Cooper, A.J.L. Brain α‐ketoglutarate dehydrogenase: kinetic
properties, regional distribution and effects of inhibitors. J Neurochem,
1986;47:376–86.
26. Pomier Layrargues, G., Spahr, L. and Butterworth, R.F. Increased manganese
concentrations in pallidum of cirrhotic patients, Lancet, 1995;34:735–41.
27. Rolando, N., Wade, J., Davalos, M. et  al. The systemic inflammatory
response syndrome in acute liver failure. Hepatology, 2000;3:734–9.
28. Odeh, M. Pathogenesis of hepatic encephalopathy: the tumour necrosis fac-
tor‐alpha theory. Eur J Clin Invest, 2007;3:291–304.
29. Chastre, A., Belanger, M., Nguyen, B.N. et al. Lipopolysaccharide precipi-
tates hepatic encephalopathy and increases blood‐brain barrier permeability
in mice with acute liver failure. Liver Int, 2013;3:353–61
30. Zieve, L., Doizaki, W.M., and Zieve, F.J. Synergism between mercaptans and
ammonia or fatty acids in the production of coma: a possible role of mercap-
tans in the pathogenesis of hepatic coma. J Lab Clin Med, 1974;8:26–8.
31. Shawcross, D.L., Wright, G., Olde Daminck, S.W.M. et al. Role of ammonia
and inflammation in minimal hepatic encephalopathy. Metab Brain Dis,
2007;2:125–38.
32. Butterworth, R.F. Pathogenesis of hepatic encephalopathy in cirrhosis: the
concept of synergism revisited. Metab Brain Dis, 2016;31(6):1211–15.
33. Schliess, F., Gorg, B., Fischer, R. et al. Ammonia induces MK801‐sensitive
nitration and phosphorylation of protein tyrosine residues in rat astrocytes.
FASEB J, 2002;1:739–41.
34. Butterworth, R.F. The liver‐brain axis in liver failure: neuroinflammation
and encephalopathy. Nat Rev Gastroenterol Hepatol, 2013;1:522–8.
35. Bemeur, C., Qu, H., Desjardins, P. et al. IL‐1 or TNF receptor gene deletion
delays onset of encephalopathy and attenuates brain edema in experimental
acute liver failure. Neurochem Int, 2010;5:213–5.
36. Cagnin, A., Taylor‐Robinson, S.D., Forton, D.M. et al. In vivo imaging of
cerebral “peripheral benzodiazepine binding sites” in patients with hepatic
encephalopathy. Gut, 2006;5:547–53.
37. Duchini, A., Govindarajan, S., Santucci, M. et al. Effects of tumor necrosis
factor‐a and interleukin‐6 on fluid‐phase permeability in CNS‐derived
endothelial cells. J Invest Med, 1996;4:474–82.
38. Andersson, A.K., Adermark, L., Persson, M. et  al. Lactate contributes to
ammonia‐mediated astroglial dysfunction during hyperammonemia,
Neurochem Res, 2009;3:555–65.
39. Lockwood, A.H., Weissenborn, K., and Butterworth, R.F. An image of the
brain in patients with liver disease. Curr Opin Neurol, 1997;1:525–33.
40. Yao, H., Sadoshima, S., Fijii, K. et al. Cerebrospinal fluid lactate in patients
with hepatic encephalopathy. Eur Neurol, 1987;2:182–7.
41. Tofteng, F., Jorgensen, L., Hansen, B.A. et  al. Cerebral microdialysis in
patients with fulminant hepatic failure, Hepatology, 2002;3(6):1333–40.
42. Larsen, F.S., Olsen, K.S., Ejlersen, E. et al. Cerebral blood flow autoregula-
tion and transcranial Doppler sonography in patients with cirrhosis.
Hepatology, 1995;22(3):730–6.
43. Aggarwal, S., Obrist, W., Yonas, H. et al. Cerebral hemodynamic and meta-
bolic profiles in fulminant hepatic failure: relationship to outcome. Liver
Transpl, 2005;11(11):1353–60.
44. Haussinger, D. Low grade cerebral edema and the pathogenesis of hepatic
encephalopathy in cirrhosis. Hepatology, 2006;43:1187–90.
45. Jones, E.A. Ammonia, the GABA neurotransmitter system and hepatic
encephalopathy. Metab Brain Dis, 2002;17:275–81.
46. Butterworth, R.F., Lavoie, J., Giguere, J.F. et al. Affinities and densities of
high‐affinity [3H] muscimol (GABA‐A) binding sites and of central benzo-
diazepine receptors are unchanged in autopsied brain tissue from cirrhotic
patients with hepatic encephalopathy. Hepatology, 1988;8(5):1084–8.
47. Ahboucha, S. and Butterworth, R.F. Pathogenesis of hepatic encephalopa-
thy: a new look at GABA from the molecular standpoint. Metab Brain Dis,
2004;19;331–43.
48. Butterworth, R.F. and Wells, J. Detection of benzodiazepines in hepatic
encephalopathy: reply. Hepatology, 1995;21:604–5.
49. Ahboucha, S., Pomier Layrargues, G., Mamer, O. et al. Increased brain con-
centrations of a neuroinhibitory steroid in human hepatic encephalopathy.
Ann Neurol, 2005;58:169–70.
50. Butterworth, R.F. Neurosteroids in hepatic encephalopathy: novel insights
and new therapeutic opportunities. J Steroid Biochem Mol Biol,
2015;160:94–7.
51. Bemeur, C. and Butterworth, R.F. Nutrition in the management of cirrhosis
and its neurological complications. J Clin Exp Hepatol, 2014;4(2):141–50.
52. Merli, M., Giusto, M., Lucidi, C. et al. Muscle depletion increases the risk of
overt and minimal hepatic encephalopathy: results of a prospective study.
Metab Brain Dis, 2013;28:281–84.
53. Amodio, P., Bemeur, C., Butterworth, R.F. et al. The nutritional management
of hepatic encephalopathy in patients with cirrhosis: International Society
for Hepatic Encephalopathy and Nitrogen Metabolism consensus.
Hepatology, 2013;58:325–36.
54. Soulsby, C.T. and Morgan, M.Y. Dietary management of hepatic encepha-
lopathy in cirrhotic patients; survey of current practice in United Kingdom.
BMJ, 1999;31:8391.
55. Cordoba, J., Lopez‐Hellin, J., Planas, M. et al. Normal protein diet for epi-
sodic hepatic encephalopathy: results of a randomized study. J Hepatol,
2004;41(1):38–43.
56. Bresci, G., Parisi, G., and Bant, S. Management of hepatic encephalopathy
with oral zinc supplementation: a long‐term treatment. Eur J Med,
1993;2(7):414–6.
57. Merli, M., Giusto, M., Gentili, F. et al. Nutritional status: its influence on the
outcome of patients undergoing liver transplantation. Liver Int,
2009;30:208–14.
58. Sharma, P. and Sharma, B.J. Management of overt hepatic encephalopathy. J
Clin Exp Hepatol, 2015;5:S82–7.
59. Als‐Nielsen, B., Gluud, L.L., and Gluud, C. Non‐absorbable disaccharides
for hepatic encephalopathy: systematic review of randomized trials. BMJ,
2004. doi:10.1136/bmj.38048.506134.EE
60. Morgan, M.Y., Blei, A., Grungreiff, K. et  al. The treatment of hepatic
encephalopathy. Metab Brain Dis, 2007;22:389–405.
61. Prasad, S., Dhiman, R.K., Duseja, A. et  al. Lactulose improves cognitive
functions and health‐related quality of life in patients who have minimal
hepatic encephalopathy. Hepatology, 2007;4:549–59.
62. Sharma, P., Sharma, B.C., Agrawal, A. et al. Primary prophylaxis of overt
hepatic encephalopathy in patients with cirrhosis: an open‐labelled rand-
omized controlled trial of lactulose versus no lactulose. J Gastroenterol
Hepatol, 2012;27:1329–35.
63. Gluud, L.L., Vilstrup, H., and Morgan, M.Y. Non‐absorbable disaccharides
versus placebo/no intervention and lactulose versus lactitol for the preven-
tion and treatment of hepatic encephalopathy in people with cirrhosis.
Cochrane Database Syst Rev, 2016;4, CD003044.
64. Bass, N.M., Mullen, K.D., Sanyal, A. et al. Rifaximin treatment in hepatic
encephalopathy. N Engl J Med, 2010;362:1071–81.
65. Kimer, N., Krag, A., Moller, S. et al. Systematic review with meta‐analysis:
the effects of rifaximin in hepatic encephalopathy. Aliment Pharmacol Ther,
2014;40:123–32.
66. Sidhu, S.S., Goyal, O., Mishra, B.P. et al. Rifaximin improves psychometric
performance and health‐related quality of life in patients with minimal
hepatic encephalopathy (the RIME trial). Am J Gastroenterol,
2011;106:307–16.
67. Bajaj, J.S., Saoian, K., Christensen, K.M. et al. Probiotic yogurt for the treat-
ment of minimal hepatic encephalopathy. Am J Gastroenterol,
2008;103:1707–15.
68. Mittal, V.V., Sharma, B.C., Sharma, P. et al. A randomized controlled trial
comparing lactulose, probiotics and L‐ornithine L‐aspartate in treatment of
minimal hepatic encephalopathy. Eur J Gastroenterol Hepatol,
2011;23:725–32.
69. Sharma, K., Pant, S., Misra, S. et al. Effect of rifaximin, probiotics and L‐
ornithine L‐aspartate on minimal hepatic encephalopathy: a randomized
controlled trial. Saudi J Gastroenterol, 2014;20:225–32.
70. Rose, C., Michalak, A., Pannunzio, P. et al. L‐ornithine L‐aspartate in experi-
mental portal‐systemic encephalopathy: therapeutic efficacy and mechanism
of action. Metab Brain Dis, 1998;13:147–57.
71. Butterworth, R.F. and Grüngreiff, K. L‐ornithine L‐aspartate (LOLA) for the
treatment of hepatic encephalopathy in cirrhosis: evidence for novel hepato-
protective mechanisms. JSM Liver Clin Res, 2019;3:5.
72. Kircheis, G., Nilius, R., Held, C. et al. Therapeutic efficacy of L‐ornithine
L‐aspartate infusions in patients with cirrhosis and hepatic encephalopathy.
Hepatology, 1997;25:1351–60.
 48:  Hepatic Encephalopathy 629
73. Sidhu, S.S., Sharma, B.C., Goyal, O. et al. L‐ornithine L‐aspartate in bouts
of overt hepatic encephalopathy. Hepatology, 2018;67:700–10.
74. Butterworth, R.F., Kircheis, G., Hilger, N. et al. Efficacy of L‐ornithine L‐
aspartate for the treatment of hepatic encephalopathy and hyperammonemia
in cirrhosis: systematic review and meta‐analysis of randomized controlled
trials. J Clin Exp Hepatol. 2018;8(3):301–13.
75. Zhu, G.Q., Shi, K.Q., Huang, S. et al. Systematic review with network meta‐
analysis: the comparative effectiveness and safety of interventions in patients
with overt hepatic encephalopathy. Aliment Pharmacol Ther, 2015;41:624–35.
76. Acharya, S.K., Bhatia, V., Sreenivas, V. et  al. Efficacy of L‐ornithine L‐
aspartate in acute liver failure: a double‐blind, randomized, placebo‐con-
trolled study. Gastroenterology, 2009;136:2159–68.
77. Marchesini, G., Bianchi, G., Merli, M. et  al. Nutritional supplementation
with branched chain amino acids in advanced cirrhosis: a double‐blind rand-
omized trial. Gastroenterology, 2003;124:1792–801.
78. Muto, Y., Sato, S., Watanabe, A. et al. Effects of oral branched‐chain amino
acid granules on event‐free survival in patients with liver cirrhosis. Clin
Gastroenterol Hepatol, 2005;3(7):705–13.
79. Sushma, S., Dasarathy, S., Tandon, R.K. et al. Sodium benzoate in the treat-
ment of acute hepatic encephalopathy: a double‐blind randomized trial,
Hepatology, 1992;16:138–44.
80. Rockey, D.C., Vierling, J.M., Mantry, P. et  al. Randomized double‐blind,
controlled study of glycerol phenylbutyrate in hepatic encephalopathy,
Hepatology, 2014;59:1073–83.
81. Stravitz, R.T., Gottfried, M., Durkalski, V. et al. Safety, tolerability and phar-
macokinetics of L‐ornithine phenylacetate in patients with acute liver injury/
failure and hyperammonemia. Hepatology, 2018;67:1003–13.
82. Malaguarnera, M., Pisone, G., Astutu, M. et al. L‐carnitine in the treatment
of mild or moderate encephalopathy. Digest Dis, 2003;21:271–75.
83. Ratnakumari, L., Qureshi, I.A., and Butterworth, R.F. Effect of L‐carnitine
on cerebral and hepatic energy metabolites in congenitally hyperammone-
mic sparse‐fur mice and its role during benzoate therapy. Metabolism,
1993;42:1039–46.
84. Therrien, G., Butterworth, J., Rose, C. et al. Protective effect of L‐Carnitine
in ammonia‐precipitated encephalopathy in portacaval‐shunted rats: evi-
dence for a central mechanism of action. Hepatology, 1997;25:551–6.
85. Pomier Layrargues, G., Giguere, J.F., Lavoie, J. et al. Flumazenil in cirrhotic
patients in hepatic coma: a randomized double‐blind placebo‐controlled
crossover trial. Hepatology, 1994;19:32–7.
86. Als‐Nielsen, B., Gluud, L.L., and Gluud, C. Benzodiazepine receptor antag-
onists for hepatic encephalopathy. Cochrane Database Syst Rev,
2004;2:CD002798.
87. Turkmen, S., Lundgren, P., Birzniece, V. et  al. 3Beta‐20beta‐dihydroxy‐5‐
alpha‐pregnane (UC1011) antagonism of the GABA potentiation and the
learning impairment induced in rats by allopregnanolone. Eur J Neurosci,
2004;20:1604–12.
88. Johansson, M., Agusti, A., Llansola, M. et  al. GR3027 antagonizes
GABA‐A receptor‐potentiating neurosteroids and restores spatial learn-
ing and motor coordination in rats with chronic hyperammonemia and
hepatic encephalopathy. Am J Physiol Gastrointest Liver Physiol,
2015;309:G400–9.
89. Cauli, O., Rodrigo, R., Piedrafita, B. et al. Inflammation and hepatic enceph-
alopathy: ibuprophen restores learning ability in rats with portacaval shunts.
Hepatology, 2007;46:514–9.
90. Chastre, A., Belanger, M., Beauchesne, E. et  al. Inflammatory cascades
driven by tumor necrosis factor alpha play a major role in the progression of
acute liver failure and its neurological complications. PloS One,
2012;7:e49670.
91. Jiang, W., Desjardins, P., and Butterworth, R.F. Cerebral inflammation con-
tributes to encephalopathy and brain edema in acute liver failure: protective
effect of minocycline. J Neurochem, 2009;109:485–93.
92. Vaquero, J. and Butterworth, R.F. Mild hypothermia for the treatment of
acute liver failure: what are we waiting for? Nat Clin Pract Gastroenterol
Hepatol, 2007;10:528–9.
93. Michel, H., Solere, M., Granier, P. et  al. Treatment of cirrhotic hepatic
encephalopathy with L‐DOPA. A controlled trial. Gastroenterology,
1980;29:555–61.
94. Uribe, M., Farca, A., Marquez, M.A. et al. Treatment of chronic portal‐sys-
temic encephalopathy with bromocriptine: a double‐blind controlled trial.
Gastroenterology, 1979;76:1347–51.
95. Morgan, M.Y., Jacobovits, A.W., James, I.M. et al. Successful use of bro-
mocriptine in the treatment of chronic hepatic encephalopathy.
Gastroenterology, 1980;78:663–70.
96. Butterworth, R.F. Parkinsonism in cirrhosis: pathogenesis and current thera-
peutic options. Metab Brain Dis, 2013;28(2):261–7.
97. Sotil, E.U., Gottstein, J., Ayala, E. et al. Impact of preoperative overt hepatic
encephalopathy on neurocognitive function after liver transplantation. Liver
Transpl, 2009;15:184–92.
98. Bajaj, J.S., Schubert, C.M., Heuman, D.M. et  al. Persistence of cognitive
impairment after resolution of overt hepatic encephalopathy.
Gastroenterology, 2010;138:2332–40.
99. Saner, F.H., Gu, Y., Minouchehr, S. et al. Neurological complications after
cadaveric and living donor transplantation. J Neurol, 2006;253:612–7.
 The Kidney in Liver Disease
49 Moshe Levi1, Shogo Takahashi1, Xiaoxin X. Wang1, and Marilyn E. Levi2
1
Department of Biochemistry and Molecular and Cellular Biology, Georgetown University,
Washington, DC, USA
2
Department of Medicine, Division of Infectious Diseases, University of Colorado, Aurora, CO, USA

IMPAIRED RENAL FUNCTION kidney’s ability to counterbalance the effects of vasoconstrictors

IN LIVER DISEASE
Impaired renal function is common in liver diseases, either as
part of multi‐organ involvement in acute illness or secondary to
advanced liver disease [1–4].
on the renal circulation, (v) increased gut permeability resulting
in bacterial translocation, and (vi) systemic inflammation with
release of cytokines, damage‐associated molecular patterns
(DAMPS), and reactive oxygen species (ROS).
These hemodynamic factors can accelerate kidney injury in
the presence of intrinsic kidney disease caused by various
disorders.

ACUTE KIDNEY INJURY

IN LIVER DISEASE
Acute kidney injury (AKI) previously known as acute renal
failure (ARF) or acute tubular necrosis (ATN) occurs in
approximately 20% of hospitalized patients with cirrhosis. In
contrast, the incidence of AKI in acute liver failure varies from
40 to 80%, especially in the setting of infections such as viral
hemorrhagic fever, leptospirosis, bacterial peritonitis, and
­ In cirrhotic patients with ascites, most of the cases of acute
toxin‐induced injuries such as acetaminophen poisoning,
administration of nephrotoxic antibiotics (such as aminogly-
coside antibiotics), or high doses of non‐steroidal anti‐inflam-
matory drugs.
The mechanisms leading to acute renal injury are multiple and
additive and include: (i) changes in the systemic arterial circula-
tion, including systemic arterial vasodilatation and sometimes in
certain alcoholic patients the concomitant presence of cardio-
myopathy, (ii) portal hypertension, (iii) activation of renal vaso-
constrictive hormones, including the renin–angiotensin system,
the sympathetic nervous system (SNS), arginine ­vasopressin,
endothelin (ET), thromboxane A2, and leukotrienes, which all
additively impair renal blood flow, (iv) suppression of renal vas-
odilatory factors including prostacyclins which impair the
HEPATORENAL SYNDROME
Hepatorenal syndrome (HRS) is a unique form of functional
renal failure that often complicates advanced liver disease,
hepatic failure, or portal hypertension. The International Club of
Ascites has recently updated the diagnostic criteria for AKI and
HRS‐AKI [5, 6], which are outlined in Table 49.1.
renal failure are mediated by: (i) pre‐renal failure and (ii) acute
kidney injury (acute tubular necrosis).
HRS accounts for approximately 20% of renal failure. HRS
is characterized by: (i) an extreme expression of the profound
circulatory dysfunction in cirrhosis, with marked splanchnic
arterial and systemic vasodilatation, insufficient cardiac output,
severe reduction of effective blood volume, homeostatic activa-
tion of vasoactive systems, and intense renal vasoconstriction,
resulting in a critical decrease in renal blood flow and glomeru-
lar filtration rate, (ii) absence of pathological changes in renal
tissue (vasculitis, glomerulonephritis, or tubulointerstitial
­fibrosis), and (iii) preserved renal tubular function [7].
Precipitating events that cause the development of HRS
include: (i) bacterial peritonitis and/or sepsis of other origin,

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
 49:  The Kidney in Liver Disease 631
Table 49.1  Current diagnostic and treatment criteria for HRS‐AKI
(A) Diagnostic criteria of HRS‐AKI
Cirrhosis and ascites
Diagnosis of AKI according to ICA‐AKI criteria: increase in SCr
greater than or equal to 0.3 mg dL−1 within 48 hours
Absence of shock
No response after two consecutive days of diuretic withdrawal and
plasma volume expansion with albumin (1 g kg−1 of body weight)
No current or recent use of nephrotoxic drugs (NSAIDs,
aminoglycosides, iodinated contrast media, etc.)
No macroscopic signs of structural kidney injury, defined as:
• absence of proteinuria (greater than 500 mg d−1)
• absence of microhematuria (greater than 50 RBCs per high
power field),
• normal findings on renal ultrasonography
(B) Treatment criteria of HRS‐AKI
Meeting all the diagnostic criteria of HRS‐AKI
AKI stage greater than or equal to 1B (after plasma albumin expansion)
No contraindication to vasoconstrictor therapy
Criteria for treatment individualized
AKI, acute kidney injury; HRS‐AKI, hepatorenal syndrome‐AKI; ICA, International
Club of Ascites; NSAIDs, non‐steroidal anti‐inflammatory drugs; RBCs, red blood cells.
Reproduced with permission of John Wiley & Sons [5].
(ii) increased doses of diuretics, (iii) large‐volume paracentesis
without volume expansion, and (iv) gastrointestinal hemor-
rhage, which all result in further compromise of the systemic
and renal circulation.
Clinically, two different forms of HRS can be distinguished:
type 1 HRS (HRS‐1) is characterized by rapid progression of
renal failure, whereas type 2 HRS (HRS‐2) is characterized by a
less severe and more stable form of renal impairment.
In recent years, there have been several potentially life‐saving
and/or survival‐prolonging therapies aimed at patients with
HRS type 1 (HRS‐1) [8–10].
Liver transplantation
Liver transplantation, when possible, is the ideal form of treat-
ment for HRS and end‐stage liver disease. Although subjects
with HRS do have more frequent complications both pre‐ and
post‐operation, including AKI requiring dialysis, nevertheless
the overall survival rate averages over 50% at 3 years. However,
due to ongoing organ shortage, most patients with end‐stage
liver disease and HRS die before a liver becomes available.
Therefore, alternative temporizing measures become very
important [10].
Use of vasoconstrictors and albumin
There is increasing evidence that shows that the use of vasocon-
strictors especially terlipressin in combination with plasma
expanders such as intravenous albumin may reverse HRS by
decreasing splanchnic vasodilatation and increasing effective
arterial blood volume, resulting in decreases in endogenous
renal vasoconstrictors and increased renal perfusion and
­glomerular filtration rate.
Recently two randomized (phase 3) studies OT‐0401 and
REVERSE compared the efficacy of terlipressin plus albumin
versus placebo plus albumin in patients with HRS‐1 [8].
Pooled patient‐level data were analyzed for HRS reversal
(serum creatinine [SCr] value less than or equal to 133 μmol
L−1], 90‐day survival, need for renal replacement therapy, and
predictors of HRS reversal. Patients received intravenous terli-
pressin 1–2 mg every 6 hours plus albumin or placebo plus albu-
min up to 14 days. The pooled analysis comprised 308 patients
(terlipressin: n = 153; placebo: n = 155). HRS reversal was
significantly more frequent with terlipressin versus placebo
­
(27% versus 14%; P = 0.004). Terlipressin was associated with
a more significant improvement in renal function from baseline
until end of treatment, with a mean between‐group difference in
SCr concentration of −53.0 μmol L−1 (P < 0.0001). Lower SCr,
lower mean arterial pressure, and lower total bilirubin and
absence of known precipitating factors for HRS were independ-
ent predictors of HRS reversal and longer survival in terlipres-
sin‐treated patients.
Terlipressin plus albumin therefore resulted in a significantly
higher rate of HRS reversal versus albumin alone in patients with
HRS‐1 (ClinicalTrials.gov identifier: OT‐0401, NCT00089570;
REVERSE, NCT01143246).
While terlipressin is widely used in Europe and Asia, it is
not yet available in the United States and instead intravenous
albumin infusion plus noradrenaline has been used. The com-
bination vasoconstrictive therapy of midodrine plus octreotide
has also been used. However, the combination of midodrine
plus octreotide is not as effective as terlipressin or noradrena-
line [9, 11].
VIRAL INFECTIONS
There are several viral diseases that affect both the liver and the
kidney, including hepatitis B and hepatitis C. Viral infections
can cause renal injury by a number of mechanisms: (i) circulat-
ing immune complexes that consist of viral antigens and the
host antiviral antibodies; (ii) in situ immune‐mediated mecha-
nisms involving viral antigens bound to glomerular structures;
(iii) expression of viral proteins or pathogenic proinflammatory
factors including cytokines, chemokines, and growth factors in
renal tissue; and (iv) direct cytopathogenic effects on glomeru-
lar and tubulointerstitial cells.
Hepatitis B
Aside from hepatitis, hepatitis B causes several extrahepatic
manifestations, including reactive arthritis, skin rashes, vasculi-
tis, and glomerulonephritis.
Three forms of glomerulonephritis are associated with ­hepatitis
B: (i) membranous glomerulonephritis is most frequently reported
in the Asian population and in children, (ii) membranoprolifera-
tive glomerulonephritis, and (iii) IgA nephropathy, most com-
monly seen in adults. The causal role of hepatitis B virus (HBV)
infection in glomerulonephritis has been suggested by the finding
of viral antigens in immune complex deposits in the glomeruli
and virus‐specific cytotoxic T lymphocytes. This is particularly
relevant in kidney transplant recipients, where graft function may
be impacted.
Hepatitis B glomerulonephritis has a more favorable course
in children than adults. Adults with nephrotic syndrome and
abnormal liver function tests have the worst prognosis and are
associated with progression to end‐stage renal disease (ESRD).
632 THE LIVER: VIRAL INFECTIONS
Lamivudine, emtricitabine, tenofovir, and entecavir inhibit hep-
atitis B DNA polymerase and are used for treatment, with the
caveat that resistance may develop, particularly with lamivu-
dine. Therefore, treatment of hepatitis B with these agents will
require monitoring for breakthrough viremia. Lamivudine and
emtricitabine, nucleoside reverse transcriptase inhibitors used
for the treatment of hepatitis B and HIV requires a lower dose
for the treatment of hepatitis B monoinfection. Tenofovir diso-
proxil fumarate (TDF) which is also used for the treatment of
both hepatitis B and HIV is potentially nephrotoxic. The prod-
rug TDF is converted rapidly to tenofovir in the gut and circu-
lates exclusively in plasma and filtered at the glomerulus. TNF
is also actively transported into proximal renal tubular cells
from the interstitial fluid and leads to mitochondrial toxicity and
proximal tubular injury, resulting in Fanconi’s syndrome.
Manifestation of Fanconi’s syndrome includes proteinuria,
hypophosphatemia and phosphaturia, normoglycemic glycosu-
ria, uricosuria, hyperuricemia, and aminoaciduria. By contrast,
a newer prodrug of tenofovir, tenofovir alafenamide (TAF) has
more potent anti‐HIV activity and is not considered nephrotoxic
as it does not accumulate in proximal tubular cells [12].
Entecavir has similar mechanisms of nephrotoxicity although to
a lesser degree. All of these agents are cleared by the kidney and
may require dose adjustment.
Hepatitis C
More than 80% of subjects acutely infected with hepatitis C
develop chronic infection and chronic liver disease. Hepatitis C
virus (HCV) infection is also associated with high prevalence of
renal disease. HCV‐positive subjects have 40% higher odds for
development of chronic kidney disease as compared with HCV‐
negative individuals.
Hepatitis C is commonly associated with membranoprolifer-
ative glomerulonephritis, with and without cryoglobulinemia,
membranous glomerulonephritis, focal segmental glomerulo-
sclerosis (FSGS), and polyarteritis nodosa. Hepatitis C infection
has also been associated with diabetes‐related nephropathy.
Common laboratory findings in subjects with hepatitis
C‐­associated glomerulonephritis include cryoglobulinemia
(mixed type II), increased monoclonal rheumatoid factor (IgM
kappa), decreased early complements C4, C1q, and CH50, and
normal or slightly decreased C3.
Previous treatment of hepatitis C included the use of pegylated
interferon (PEG‐IFN) and ribavirin with associated adverse
reactions including fevers, anemia, and limited efficacy. More
recently, PEG‐IFN has been replaced by directly acting antivi-
rals (DAAs) that have revolutionized the treatment of hepatitis
C, achieving cure in 95–98% of cases, defined as a sustained
virologic response at twelve weeks, or SVR12. Successful treat-
ment of HCV often leads to remission of cryoglobulinemic glo-
merulonephritis or mixed cryoglobulinemia. In the presence of
severe vasculitis, rapidly progressive glomerulonephritis or
nephrotic syndrome, immunosuppression should be initiated
early to prevent progression of renal disease [13], including cor-
ticosteroids, rituximab, and plasma exchange [13]. There are
three classes of DAAs: (i) NS3/4 protease inhibitors (PIs),
(ii) NS5B polymerase inhibitors, and (iii) NS5A inhibitor class.
Commonly, combination therapy using 1 or more DAAs from
different classes is used to decrease the development of resist-
ance and enhance potency of the regimen. Regimens are deter-
mined by the HCV genotype, specifically genotypes 1, 4, 5, or
6. Some DAAs are potentially nephrotoxic such as sofosbuvir
which is not recommended for patients with chronic kidney dis-
ease stage 4 or 5 with eGFR less than 30 or ESRD.
Hepatitis C infection prevalence in patients with ESRD ranges
between 3–70% depending on the country. Therefore, all dialysis
patients require initial HCV antibody screening and if negative,
repeated every six months. If the HCV antibody is positive, hep-
atitis C RNA should be obtained to look for evidence of viremia
as well as a genotype. Treatment should be initiated in the pres-
ence of HCV viremia with regimens based on the genotype.
Patients who are infected with hepatitis C will suppress hepa-
titis B viremia if coinfected. When these patients undergo treat-
ment of hepatitis C with either PEG‐IFN or directly acting
antivirals, the underlying hepatitis B may reactivate and pro-
gress [14]. Liver failure from hepatitis B resulting in death and
liver transplantation has been reported. Consequently, the Food
and Drug Administration placed a black box warning on DAAs,
with recommendations to screen all patients with hepatitis C for
any positive serologies for hepatitis B. The most common posi-
tive serology to be consistent with active hepatitis B is hepatitis
B surface antigen (HBsAg), and if detected, indicates a risk for
progression of hepatitis B when HCV is successfully treated.
Therefore, if hepatitis B treatment is indicated, this should be
initiated prior to DAA therapy for hepatitis C to avoid active
hepatitis B ­infection. In the presence of a positive hepatitis B
core antibody and negative HBsAg, the risk of hepatitis B
reactivation is less likely. In this scenario, HBV DNA moni-
toring is performed monthly and continued for three months
after DAA is completed [15].
Human immunodeficiency virus
While HIV‐1 causes a wide spectrum of renal diseases, the role of
HIV‐1 in causing primary hepatitis is less certain. Liver function
abnormalities in HIV‐1‐infected individuals have been associated
with drug adverse reactions and interactions, alcohol abuse, HCV
or HBV coinfections, opportunistic infections such as HHV‐8
(associated with Kaposi’s sarcoma), and Hodgkin’s or non‐
Hodgkin’s lymphoma. With initiation of antiretroviral therapy
(ART), autoimmune hepatitis and liver failure may develop as a
complication of immune reconstitution syndrome (IRIS), a par-
adoxical worsening of the individual patient’s HIV‐associated
condition at the time of initiation of antiretroviral therapy and may
be responsive to corticosteroids. IRIS may develop in response
to autoantigens associated with autoimmune conditions that may
manifest after antiretroviral therapy is initiated such as systemic
lupus erythematosus, Graves’ disease, sarcoidosis, and rheuma-
toid arthritis. IRIS may also develop in reaction to underlying
pathogens such as Mycobacterium tuberculosis or opportunistic
infections such as Pneumocystis jiroveci (formerly P. carinii),
Mycobacterium avium complex, Toxoplasmosis gondii,
Cryptococcus spp., Histoplasmosis capsulatum, Cytomegalovirus
(CMV), and other Herpesviridae and JC virus [16]. IRIS has also
been reported to unmask underlying Hodgkin’s and non‐Hodgkin’s
lymphoma with potential liver infiltration within six months of
initiation of ART. Management of IRIS includes treatment of
 49:  The Kidney in Liver Disease 633
opportunistic infections, hepatitis B or C or lymphoma, and con-
tinuation of antiretroviral therapy with or without steroids.
Antiretroviral drug hepatotoxicity has been well described. A
previous study conducted in subjects with unexplained transam-
inase (ALT and AST) elevations in HIV‐1 monoinfected patients
on antiviral therapy found a high rate of liver lesions, mostly
consistent with non‐alcoholic steatohepatitis (NASH). It is now
recognized that nucleoside reverse transcriptase inhibitors
(NRTIs) including AZT, lamivudine, tenofovir, and emtricit-
abine may induce hepatitis steatosis. The NRTI abacavir induces
a unique hypersensitivity reaction in patients who are positive
for HLA B5701 and are at risk for liver failure and death.
Therefore, a screening blood test to identify the presence of
HLA B5701 should be performed prior to initiation of abacavir.
Non‐nucleoside reverse transcriptase inhibitors (NNRTI) such
as efavirenz may be associated with severe liver injury. Despite
these observations, a more recent study following a cohort of
10 083 HIV infected patients in the United States between
2004–2010 showed infrequent elevations in aminotransferases.
However, this study also noted that the risk of hepatotoxicity
was greater with HIV protease inhibitor use in hepatitis B or C
coinfected patients [17].
In HIV positive patients who are coinfected with hepatitis B
and/or C, more rapid progression of hepatic decompensation
including drug‐induced liver injury is seen compared to HIV
monoinfected individuals. Treatment with antiretroviral therapy
has been shown to control the advancement of hepatic failure
and is an integral part of HBV and HCV management. Initial
antiretroviral regimens in coinfected patients are the same as in
HIV monoinfected patients with the caveat that drug selection
may require adjustment due to potential drug interactions with
directly acting agents. In addition, when treating HCV with
DAA regimens that include PIs, HIV coinfected patients who
are maintained on HIV protease inhibitors should be switched to
an HIV non‐PI containing regimen.
Kidney disease is a common complication of HIV infection.
The spectrum of HIV‐associated renal disease includes direct
association with infection, diseases that are linked with the sys-
temic response to infection, diseases that develop because of
superinfections, and diseases that are associated with treatment of
HIV infection [12]. Specifically, AKI may result from acute tubu-
lar necrosis, acute tubulointerstitial nephritis, rhabdomyolysis,
antiretroviral therapy, HIV‐associated nephropathy, HIV immune
complex disease, thrombotic microangiopathies, and urinary tract
obstruction. Chronic kidney disease (CKD) may also result from
these causes. In addition, there are several potential opportunistic
infections of the kidney parenchyma, including viruses such as
cytomegalovirus (CMV) and BK, fungal, both typical and atypical
mycobacterial infections, as well as infiltrative lesions of the kid-
ney that may result from lymphoma and Kaposi’s sarcoma [12].
Current antiretroviral therapy regimens may result in chronic
inflammation, premature aging, and metabolic syndrome and its
complications including chronic kidney disease [18].
HIV patients of African descent have increased predisposi-
tion to HIV‐associated nephropathy due to variants in ApoL1
gene encoding apolipoprotein L1 [19, 20]. In addition to HIV‐
associated nephropathy, APOL1 risk variants are also associated
with focal segmental glomerulosclerosis and hypertension‐
associated arterionephrosclerosis [19, 21].
Treatment of HIV‐associated nephropathy centers around use
of highly active antiretroviral therapy and renin angiotensin
aldosterone antagonists.
Viral infections in kidney transplant recipients
Prevention of kidney graft rejection requires immunosuppres-
sive agents that primarily inhibit T‐lymphocyte function.
Consequently, these patients are at risk for viral complications
so that vaccinations, including live vaccines should be adminis-
tered prior to transplantation such as Zostavax for herpes zoster.
Following transplantation, live vaccines are contraindicated
due  to risk for vaccine‐induced infection in the presence of
immunosuppression.
Specific issues related to transplantation
1. Reactivation of previous infections that have remained in the
latent phase such as cytomegalovirus, herpes simplex, herpes
zoster, and Epstein‐Barr viruses, all potentially manifest as
acute hepatitis. Epstein‐Barr virus is also associated with post‐
transplant lymphoproliferative disorder (PTLD) that may
involve the liver and responds to lowering of immunosuppres-
sion. Specific cases with CD20+ markers may respond to
rituximab +/− chemotherapy including cyclophosphamide,
doxorubicin, vincristine, and prednisone (CHOP). Reactivation
may also occur with the use of antilymphocyte antibodies
such as antithymocyte globulin (ATG) and thymoglobulin and
requires the use of valganciclovir or ganciclovir for CMV
prophylaxis. Reactivation may be predicted by the net state of
immunosuppression, whereby excessive lowering of T and B
cell function may impair the ability to control latent infections,
resulting in acute reactivation [22]. Vaccination for herpes
zoster reactivation
Reactivation of BK virus, a polyomavirus is seen primar-
ily in post‐kidney transplant recipients. This virus is believed
to be acquired during childhood and is maintained in a latent
phase in the genitourinary tract. During periods of immuno-
suppression, BK virus may reactivate and cause multiple
complications including ureteral stenosis, hematuria, kidney
graft dysfunction, and organ loss. Monitoring of BK in
plasma and urine is used to predict graft dysfunction and is
treated preemptively with lowering of immunosuppression.
2. Donor derived transmission of viral infections: this is com-
monly related to the presence of viral antibodies (e.g. CMV
IgG) in the donor and absence of protective antibody in the
recipient termed mismatches resulting in primary infections.
Use of prophylaxis to prevent primary infections with the
appropriate antivirals may prevent donor transmission such
as valganciclovir or ganciclovir for cytomegalovirus for six
months and acyclovir for herpes simplex and zoster.
Prophylaxis for EBV is controversial, and EBV DNA moni-
toring is recommended.
3. Community acquired infections: The most common viral
pathogens observed include Herpesviridae such as cytomeg-
alovirus, herpes simplex, herpes zoster, and Epstein‐Barr, all
of which may present as primary infections and involve the
liver and kidneys. Adenovirus, a cause of hemorrhagic
­cystitis, nephritis, and graft failure in addition to hepatitis
634 THE LIVER:  NAFLD, NASH, AND CKD
responds to lowering of immunosuppression and use of cido-
fovir, although nephrotoxicity is a significant issue. A prod-
rug of cidofovir, brincidofovir is currently available on a
compassionate use basis for adenoviral infections with no
significant nephrotoxicity reported, although gastrointestinal
intolerance is frequently seen. Influenza infections are a sig-
nificant cause of morbidity and mortality in transplant recipi-
ents, and vaccination is imperative.
METABOLIC SYNDROME AND OBESITY
It is now increasingly recognized that metabolic syndrome and
obesity affect both the kidney and the liver. The proportion of
obese adults worldwide has increased to 37% in men and simi-
larly to 38% in women. In the United States the prevalence of
obesity has similarly increased to 39.8% in adults [23].
OBESITY AND KIDNEY DISEASE
The obesity epidemic has resulted in an increased incidence of
obesity‐related glomerulopathy (ORG), a distinct entity pre-
senting as proteinuria, enlarged glomeruli, progressive glomeru-
losclerosis, and decline in renal function [24–26]. Pathologically
glomerular hypertrophy and adaptive focal segmental glomeru-
losclerosis are characteristics of ORG.
Obesity‐induced increases in glomerular filtration rate, renal
plasma flow, filtration fraction, and renal tubular sodium reab-
sorption results in glomerular hypertrophy.
In addition to the hemodynamic changes, insulin resistance,
altered fatty acid and cholesterol metabolism, mitochondrial dys-
function, oxidative stress, ER stress, inflammation, microbiome,
altered bile acid metabolism, and increased activity of profibro-
genic factors also play a role in the pathogenesis of ORG.
Approximately one third of the patients develop progressive
increase in proteinuria and decline in renal function, resulting in
ESRD. Renin angiotensin aldosterone system blockade, and
especially weight loss, and bariatric surgery result in reductions
of proteinuria and progression of kidney disease.
Studies in experimental models of diet‐induced obesity and
kidney disease indicate that activation of the bile acid regulated
nuclear receptor farnesoid X receptor (FXR) and/or G protein‐
coupled receptor TGR5 protects against kidney injury. The
mechanisms of action include prevention of lipid accumulation,
oxidative stress, inflammation, ER stress, mitochondrial
­dysfunction, and fibrosis [27–30].
OBESITY AND LIVER DISEASE
Obesity and metabolic syndrome, especially in the presence of
diabetes mellitus, has become the leading cause of non‐alco-
holic fatty liver disease (NAFLD), in its whole spectrum of dis-
ease ranging from simple steatosis to NASH, cirrhosis, and
hepatocellular carcinoma (HCC). The prevalence of NAFLD is
projected to increase from the current estimate of 25% of the
population. The rising disease prevalence is expected to result in
increasing number of patients with cirrhosis and end stage liver
disease requiring liver transplantation and in addition an
increase in HCC [31].
NAFLD, NASH, AND CKD
There is now increasing evidence that NAFLD, is associated
with a nearly 40% increase in the long‐term risk of incident
CKD [1–4]. The mechanisms by which NAFLD increases CKD
risk are not fully known. However, several factors including
insulin resistance, altered fatty acid and cholesterol metabolism,
mitochondrial dysfunction, oxidative stress, ER stress, inflam-
mation, microbiome, altered bile acid metabolism, and increased
activity of profibrogenic factors can result in progression of
NAFLD and play a role on increased incidence of CKD
(Figures 49.1 and 49.2).
Insulin resistance
Insulin resistance is a common feature of NAFLD that also
contributes to its pathogenesis. For example, insulin resist-
ance in adipose tissue leads to release of fatty acids through
dysregulated lipolysis of triglycerides. Insulin resistant adi-
pocytes also secrete cytokines like IL6 that has proinflamma-
tory effects in the liver [32, 33]. This results in further
augmentation of insulin resistance. In addition, the fatty acids
released by adipocytes are taken up by the liver, which con-
tributes to increased lipogenesis and lipid droplet formation
in the liver [32, 34].
Insulin resistance also plays an important role in pathogen-
esis of CKD by modulating renal hemodynamics, glomerular
mesangial cells and podocytes, and renal tubular function [33,
35]. In insulin resistance despite increased renovascular resist-
ance due to impaired nitric oxide (NO) generation, reduced
tubuloglomerular feedback, and dilation of afferent arterioles
due to increased reabsorption of glucose and sodium result in
glomerular hyperfiltration. Reduced insulin signaling in
mesangial cells results in mesangial cell hypertrophy, hyper-
trophy, and extracellular matrix deposition. Insulin resistance
in podocytes results through lysosomal and proteasomal deg-
radation of the insulin receptor by a nephrin dependent mecha-
nism [36]. This induces podocyte apoptosis, effacement of its
foot processes, albuminuria, thickening of the glomerular
basement membrane, and increased glomerulosclerosis. The
renal tubular consequences of insulin resistance are more com-
plex but include gluconeogenesis in the proximal tubule and
increased sodium absorption in the distal tubule [33, 35].
While the effects of insulin resistance per se in regulation of
proximal tubule glucose reabsorption via SGLT‐2 is debatable,
nevertheless SGLT‐2 expression and activity are increased in
humans and animal models of obesity and diabetes [37]. SGLT‐2
inhibitors including empagliflozin and canagliflozin have renal
and cardiovascular disease beneficial effects in human subject.
Furthermore SGLT‐2 inhibitors decrease renal disease and liver
disease in an animal model of diet‐induced obesity and insulin
resistance [38].
 49:  The Kidney in Liver Disease 635

Figure 49.1  Organ crosstalk in the pathophysiology of non‐alcoholic fatty liver disease (NAFLD) and chronic kidney disease (CKD). Many
­factors, such as caloric intake, dietary factors (such as high fructose consumption and low vitamin D levels), genetic factors, and inflammation of
visceral adipose tissue can increase an individual’s risk of NAFLD. Increased amounts of non‐esterified fatty acids (NEFAs) resulting from the
expansion of intra‐abdominal visceral adipose tissue are associated with an increase in NF‐κB and inflammatory pathways, dysregulation of
­adipokine production leading to decreased adiponectin levels, and impaired insulin signaling. Progression of liver dysfunction triggers pathways
that might influence the development of CKD. For example, insulin resistance and atherogenic dyslipidemia as well as proinflammatory factors,
prothrombotic factors, and profibrogenic molecules can promote vascular and renal damage. Reduced activation of the energy sensor, 5′‐AMP
activated protein kinase (AMPK), in response to reduced adiponectin levels further stimulates proinflammatory and profibrogenic mechanisms.
Activation of the renin‐angiotensin system (RAS) and endothelial cells might also contribute to liver and kidney dysfunction by increasing oxidative
stress, inflammation, and coagulation pathways. Increased production of uremic toxins by intestinal microbiota in the setting of CKD might induce
further renal, liver, and cardiovascular damage through inflammatory, oxidative, and fibrotic pathways. Dysbiosis of the intestinal microbiota,
which often occurs with obesity, potentially influences NAFLD, CKD, and type 2 diabetes mellitus (T2DM) through complex mechanisms. Finally,
cardiovascular disease (CVD), the risk of which is increased in the setting of NAFLD, T2DM, or intestinal dysbiosis, can influence the development
of renal dysfunction (and vice versa) through cardiorenal interactions. AGE, advanced glycation end‐product; NASH, non‐alcoholic steatohepatitis;
ROS, reactive oxygen species; SCFA, short‐chain fatty acid; TMAO, trimethylamine oxide. Reproduced with permission of Springer Nature, Figure
2 in [1].

Lipid metabolism
Excessive lipid accumulation can occur ectopically in nonfat
tissue, contributing to their damage through toxic processes
named lipotoxicity. Liver lipid accumulation results from an
imbalance between lipid acquisition (uptake of circulating fatty
acids and de novo lipogenesis [DNL]) and lipid removal (fatty
acid oxidation [FAO], export as a component of very low‐­
density lipoproteins [VLDL] particles, lipid droplet formation
and lipolysis) (Figure 49.3).
The hepatic lipid uptake is largely dependent on fatty acid
transporters [39], predominately mediated by fatty acid transport
proteins (FATPs), cluster of differentiation 36 (CD36), and cave-
olins located in the hepatocyte plasma membrane. Of the six
mammalian FATP isoforms, FATP2 and FATP5 are found pri-
marily in the liver and their knockdown in mice decreases uptake
of fatty acids and ameliorates hepatic steatosis. CD36 facilitates
the transport of long‐chain fatty acids and is regulated by peroxi-
some proliferator‐activated receptor (PPAR) γ, ­ pregnane X
receptor, and liver X receptor (LXR). CD36 expression in the
liver is increased in NAFLD and is thought to ­stimulate increased
uptake of fatty acids. The caveolins comprise a family of three
membrane proteins for lipid trafficking and lipid droplets forma-
tion. Caveolin 1 is increased in the liver of mice with NAFLD,
and mainly localized in the centrilobular zone three, where the
steatosis is most severe. Following uptake, fatty acids in the cyto-
sol need specific intracellular fatty acid binding proteins (FABP)
636 THE LIVER:  NAFLD, NASH, AND CKD

Figure 49.2  Potential mechanisms by which intestinal dysbiosis might promote the development of non‐alcoholic fatty liver disease (NAFLD)
and chronic kidney disease (CKD). An imbalance in intestinal microbiota (dysbiosis) resulting in an increase in Gram‐negative bacteria leads to an
increase in lipopolysaccharide (LPS) production, which can damage the intestinal epithelium resulting in increased gut permeability. This increased
gut permeability enables further egress of bacterial content, LPS, and small molecules into the portal and systemic circulations causing inflamma-
tion. Dysbiosis also promotes the increased production of secondary bile acids such as deoxycholic acid. Following their return to the liver as part
of the enterohepatic circulation, secondary bile acids have been linked to chronic inflammation, cholestasis and carcinogenesis through their actions
on farnesoid X Receptor (FXR; with downstream effects on cholesterol metabolism) and by inducing cellular senescence and causing damage to
mitochondrial and cell membranes. The intestinal microbiota also generates molecules such as trimethylamine (TMA), p‐cresol, and indole from
dietary choline, phenylalanine/tyrosine, and tryptophan, respectively. TMA is oxidized in the liver to trimethylamine oxide (TMAO), which pro-
motes atherosclerotic vascular disease. Indole and p‐cresol are metabolized in the liver to indole sulfate and p‐cresol sulfate, which are cleared by
the proximal tubules and are potentially nephrotoxic. Other molecules produced by microbiota metabolism, such as phenylacetic acid and hippuric
acid, are also potentially toxic to the kidney. Reduced levels of short‐chain fatty acids (SCFAs) as a consequence of dysbiosis can also induce
decreased lipogenesis and increased gluconeogenesis, leading to insulin resistance, further dysfunction of the liver and kidney, and to the develop-
ment of type 2 diabetes mellitus (T2DM). Reproduced with permission of Springer Nature, Figure 3 in [1].

to facilitate their transportation, s­ torage, and utilization. FABP1,
also known as liver‐type FABP, is the predominant isoform in the
liver. Its expression is increased in NAFLD but declines as the
disease progresses to NASH.
DNL results from the synthesis of new fatty acids from acetyl‐
CoA subunits produced primarily through glycolysis and the
metabolism of carbohydrates. In addition to glucose which most
commonly supplies carbon units for DNL, fructose also produces
acetyl‐CoA that can enter the lipogenic pathway, important when
considering the increasing use of fructose in corn syrup as a sweet-
ener. DNL starts with acetyl‐CoA that is converted to malonyl‐
CoA by acetyl‐CoA carboxylase (ACC) and further converted to
palmitate by fatty acid synthase (FASN). New fatty acid may then
undergo a range of desaturation, elongation, and esterification
steps before ultimately being stored as triglycerides in lipid drop-
lets or exported as VLDL particles. Dysregulated DNL is a central
feature of liver lipid accumulation in NAFLD patients [31].
Lipogenic genes are coordinately regulated at the transcription
level. Transcription factors, such as sterol regulatory element‐
binding protein 1c (SREBP1c), which is activated by insulin and
LXRα, and carbohydrate regulatory element‐binding protein
(ChREBP), which is activated by carbohydrates, play critical roles
in this process. SREBP1c expression is enhanced in patients with
NAFLD, while SREBP1c knockout mice display decreased
expression of lipogenic enzymes. ChREBP is the bona fide tran-
scription ­factor that is primarily responsive to glucose. ChREBP
­regulates not only enzymes in glucose metabolism, but also lipo-
genic enzymes. Therefore, ChREBP causes complex metabolic
changes in loss‐ and gain‐of function studies. ChREBP expression
in liver biopsies from patients with NASH is increased when stea-
tosis is found to be greater than 50% but decreased in the presence
of severe insulin resistance [40], indicating ChREBP might disso-
ciate hepatosteatosis from insulin resistance. Recent studies fur-
ther illustrate how different post‐­translational modifications, such
as phosphorylation, acetylation, O‐GlcNAcylation, or ubiquitina-
tion, operate together in regulating lipogenic gene transcription.
For example, deacetylation of SREBP‐1c by SIRT1 inhibited
binding to its target lipogenic promoters [41]. Insulin signaling
triggers activation of series of kinases downstream of PI3K, such
as Akt and mTORC1/2, to regulate SREBP‐1c ­activity [42].
 49:  The Kidney in Liver Disease 637

Figure 49.3  The substrate‐overload liver injury model of NASH pathogenesis. Free fatty acids are central to the pathogenesis of NASH. Free fatty
acids that originate from lipolysis of triglyceride in adipose tissue are delivered through blood to the liver. The other major contributor to the free
fatty acid flux through the liver is DNL, the process by which hepatocytes convert excess carbohydrates, especially fructose, to fatty acids. The two
major fates of fatty acids in hepatocytes are mitochondrial beta‐oxidation and re‐esterification to form triglyceride. Triglyceride can be exported
into the blood as VLDL or stored in lipid droplets. Lipid droplet triglyceride undergoes regulated lipolysis to release fatty acids back into the hepato-
cyte free fatty acid pool. PNPLA3 participates in this lipolytic process, and a single‐nucleotide variant of PNPLA3 is strongly associated with
NASH progression, underscoring the importance of the regulation of this lipolysis. When the disposal of fatty acids through beta‐oxidation or for-
mation of triglyceride is overwhelmed, fatty acids can contribute to the formation of lipotoxic species that lead to ER stress, oxidant stress, and
inflammasome activation. These processes are responsible for the phenotype of NASH with hepatocellular injury, inflammation, stellate cell activa-
tion, and progressive accumulation of excess extracellular matrix. Lifestyle modifications that include healthy eating habits and regular exercise
reduce the substrate overload through decreased intake and diversion of metabolic substrates to metabolically active tissues and can thereby prevent
or reverse NASH. SCD, stearoyl CoA‐desaturase; FAS, fatty acid synthase; NKT, natural killer T cell; Tregs, regulatory T cells; PMNs, polymor-
phonuclear leukocytes. Reproduced with permission of Springer Nature, Figure 1 in [31].

FAO is controlled by nuclear receptor PPARα and occurs
mainly in the mitochondria, providing a source of energy to gen-
erate ATP especially when circulating glucose concentrations are
low [31]. Activation of PPARα induces the transcription of a
range of genes related to FAO thereby reducing hepatic lipid lev-
els. PPARα expression modulates not only lipid ­homeostasis, but
inflammation as well. Expression of genes related to FAO was
higher in patients with more severe steatosis compared to patients
with less severe steatosis or non‐steatotic controls. FAO, meas-
ured indirectly as plasma β‐hydroxybutyrate levels, was higher
in patients with NASH compared to steatosis or normal controls.
Increased FAO may be an adaptive response in patients with
NAFLD attempting to reduce the lipid overload and lipotoxicity,
but it also produces ROS and excessive FAO may exhaust the
capacity of the antioxidant defense system and induce oxidative
stress. Accordingly, hepatic oxidative stress and changes in
mitochondrial ultrastructure were increased alongside FAO in
patients with NASH. Antioxidant markers, glutathione, glu-
tathione peroxidase, and superoxide dismutase were decreased
in liver biopsies from NAFLD patients and in mitochondria from
animal models of NAFLD [43].
Hepatic lipid can be exported from the liver into the blood after
being packed into water‐soluble very low‐density lipoprotein
(VLDL) particles alongside cholesterol, phospholipids, and apoli-
poproteins [44]. The assembly of VLDL particles occurs in the ER
where apolipoprotein B100 (apoB100) is lipidated by the enzyme
microsomal triglyceride transfer protein (MTTP). apoB100 and
MTTP are key components in hepatic VLDL secretion. In patients
with genetic defects in the apoB or MTTP gene, hepatic steatosis
is common due to compromised triglyceride export.
Excessive lipid not secreted into the blood can form lipid
droplets in hepatocytes, a defining feature of NAFLD. The
638 THE LIVER:  NAFLD, NASH, AND CKD
triglycerides stored in lipid droplets have to be mobilized via
hydrolysis to release fatty acids for utilization [45]. Triglyceride
hydrolysis consists of sequential reactions of adipose triglycer-
ide lipase (ATGL, annotated as patatin‐like phospholipase
domain containing protein 2, PNPLA2 or desnutrin) hydrolyz-
ing triglycerides to diacylglycerol (DG) and fatty acid,
­hormone‐sensitive lipase (HSL) breaking down DG to monoa-
cylglycerol (MG) and fatty acid and monoacylglycerol lipase
(MAGL) completing the lipolytic reaction of MG to glycerol
and fatty acid [45]. Various studies suggest a critical role of
intracellular lipid droplet homeostasis in regulating hepatic fat
content. A tight regulation between lipid synthesis, hydrolysis,
secretion, and FAO is required to prevent hepatic lipid overload
and subsequent intracellular lipid accumulation, leading to
steatosis, lipotoxicity, and liver damage, and promoting disease
progression and fibrosis.
The specific lipotoxic lipids that promote NASH phenotype
include DG, ceramides, and lysophosphatidylcholine (LPC)
species [31]. Hepatic free cholesterol is also considered a criti-
cal lipotoxic molecule in NASH [46]. The main pathways by
which hepatocytes acquire cholesterol, include endogenous
synthesis of cholesterol via master regulator SREBP‐2 and its
target 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR),
the rate‐limiting enzyme; uptake of LDL and chylomicron rem-
nants via LDL receptor‐mediated endocytosis and subsequent
processing through the endosomal/lysosomal compartment; and
uptake of HDL cholesterol directly via scavenger receptor class
B type I (SR-BI). The main pathways by which cholesterol can
be removed from hepatocytes, include conversion to bile acids
by the rate‐limiting enzymes CYP7A1 and CYP27A1 and
excretion of bile acids into bile by bile salt export pump (BSEP);
excretion of cholesterol into bile by ATP‐binding cassette sub‐
family G member 5 and 8 (ABCG5/G8); incorporation into
VLDL and secretion into the circulation; and efflux of choles-
terol onto circulating apolipoprotein AI and nascent HDL parti-
cles. Extensive dysregulation of cholesterol homeostasis has
been documented in NAFLD, causing both increased synthesis
and uptake of cholesterol as well as decreased removal of cho-
lesterol, and leading to increased hepatic cholesterol levels.
Among the agents that target lipotoxicity [31, 47], aramchol,
is an antagonist for stearoyl-CoA desaturase-1 (SCD-1), which is
an enzyme that catalyzes a rate‐limiting step in the synthesis of
monounsaturated fatty acids such as oleic acid. This agent also
activates cholesterol efflux by stimulating the ABCA1 trans-
porter, a widely expressed cholesterol export pump. A multi-
center phase 2b trial (NCT02279524) in NASH patients is
ongoing. For another enzyme in DNL, ACC, two inhibitors
PF‐05221304 and GS‐0976 are in clinical trials in patients with
NAFLD and NASH.
In the kidney, abnormal lipid metabolism promotes increased
triglyceride and cholesterol accumulation [24]. Renal triglyceride
accumulation can occur because of increased fatty acid synthesis
mediated by SREBP‐1c and its target enzymes including ACC,
FASN, and SCD‐1; and/or mediated by ChREBP and its target
enzymes including liver‐pyruvate kinase (L‐PK). In the murine
high‐fat diet‐induced obesity model SREBP‐1 expression and
activity results in increased renal lipid accumulation and renal
disease [48–50]. The effects of high‐fat diet on kidney disease are
prevented in SREBP‐1c knockout mice. Increased SREBP‐1
expression is also found in the glomeruli of patients with obesity
related glomerulopathy [51]. In a search for in vivo inhibitors of
SREBP‐1, the FXR agonists have been demonstrated to inhibit
SREBP‐1 expression, lipid accumulation, inflammation, and
fibrosis in animal models of diet‐induced obesity and insulin
resistance [28, 52, 53]. Renal triglyceride accumulation can also
occur by increased uptake via CD36 or FATPs. Renal CD36
expression is upregulated in patients with CKD, particularly those
with diabetic nephropathy [54]. Blockade or knockout of CD36
can prevent kidney injury in experimental animals. Decreased
FAO mediated by PPARα and its target enzymes is also responsi-
ble for renal triglyceride accumulation and is reported in kidney
biopsy samples from patients with ORG and diabetes [55]. In this
regard, PPARα agonist fenofibrate has been shown to prevent
development of kidney disease [24].
Studies in mice with high fat‐induced obesity have shown
increased cholesterol synthesis and accumulation as well as
the  development of renal disease [49, 56] associated with
increased expression and activity of SREBP‐2, a master regulator
of cholesterol synthesis and cholesterol metabolism. Statin treat-
ment has also been shown to reduce lipid accumulation in the
proximal tubules in high‐fat diet‐fed mice. Increased cholesterol
accumulation can also occur because of decreased cholesterol
efflux. The nuclear receptor LXR plays a major role in cholesterol
efflux and its expression and target enzymes that regulate choles-
terol efflux, including ABCA1 and ABCG1, are decreased in
human kidney biopsy samples from diabetic patients with ORG
[55] and in ­animal models of diabetes [57]. Treatment with LXR
agonists or induction of cholesterol efflux with cyclodextrin can
reduce cholesterol accumulation and improves renal disease in
diabetic mice.
Mitochondrial dysfunction
Lipotoxicity is mediated by different cellular mechanisms
including direct damage to mitochondria [58]. Mitochondria,
the powerhouse of the cells, play a pivotal role in the final oxi-
dation of metabolites such as fatty acids and glucose. The mito-
chondrial oxidative phosphorylation (OXPHOS) is responsible
for the production of most cellular total energy. In addition, it is
also involved in the regulation of intrinsic signaling pathways of
apoptosis. Since the first studies on NAFLD, much evidence
pointed out that it was primarily characterized by the presence
of mitochondrial dysfunction [59]. The alteration of mitochon-
drial functions is evident with electron microscopy analysis by
some ultrastructural changes such as megamitochondria, loss of
cristae, and paracrystalline inclusion bodies in the matrix.
When the disposal of fatty acids through FAO or formation of
triglyceride is overwhelmed in NAFLD, fatty acids can contribute
to the formation of lipotoxic species that lead to oxidative stress
and inflammation [31]. Oxidative stress can lead to the peroxida-
tion of phospholipids, such as cardiolipin (diphosphatidyl glyc-
erol), one of the major phospholipids in the inner mitochondrial
membrane. Since cardiolipin is assumed to enhance the respira-
tory chain activity, particularly of the complex I, the oxidation of
cardiolipin leads to an imbalance of OXPHOS. Oxidative stress‐
induced damage to mitochondrial DNA can also result in the
impairment of OXPHOS and further increases the generation
of  ROS. The mitochondrial damage may eventually lead to
 49:  The Kidney in Liver Disease 639
apoptotic death of hepatocytes. An increased expression of apop-
totic proteins and enzymes, such as Bcl‐2, was found in the
murine models of fatty liver disease. The apoptosis along with the
generated cytokines from the liver stellate and Kupffer cells fur-
ther augment the fibrotic changes to advance the disease.
Sirtuins are a group of nicotinamide adenine dinucleotide
(NAD)‐dependent deacetylases, share multiple cellular functions
related to mitochondrial energy homeostasis and antioxidant
activity in liver and kidney diseases [60, 61]. SIRT1, the most
studied of these enzymes, has an indirect regulatory effect on oxi-
dative stress, activating forkhead proteins and PGC‐1α, transcrip-
tion factors involved in transcription of mitochondrial biogenesis
genes and antioxidant enzyme genes and in ROS‐detoxifying
capacity. SIRT1 level is decreased in a rat model of NAFLD.
SIRT3 localizes in the mitochondrial matrix and can increase
FAO by activation of long‐chain acyl‐CoA dehydrogenases, and
its activity is found decreased in animal models with fatty liver.
NAD+, as the co‐substrate for SIRT1 and SIRT3, controls various
metabolic improvements. In this way, NAD+ depletion may con-
tribute to mitochondrial dysfunction, obstructing the adaptive
response mediated by sirtuins to high hepatic lipid levels.
Hence, alleviation of the mitochondrial impairment, particu-
larly in the early stages of NAFLD, may prevent the progression of
the disease. Among the various experimentally studied mitochon-
drial‐targeted agents, triphenylphosphonium cation ligated ubiqui-
none Q10 and vitamin E, Szeto‐Scheller peptides, and superoxide
dismutase mimetic‐salen manganese complexes (EUK‐8 and
EUK‐134) have been found to be most promising [47].
Mitochondrial dysfunction is also the main cause of renal
pathology induced by high fat diet. Given that the kidney is an
organ that demands continuous high‐energy provision, mostly
from FAO, lipid overload and impaired FAO lead to a distur-
bance in fatty acid uptake and utilization, further aggravating
lipid accumulation in kidney cells and tissue [62]. This will cre-
ate a vicious cycle whereby lipids can damage mitochondria and
generate ROS, further limiting mitochondrial FAO and causing
more cellular lipid accumulation, resulting in more mitochon-
drial ROS levels. As an approach to target mitochondria dys-
function, activation of membrane bile acid receptor TGR5 has
been shown to decrease mitochondrial ROS generation and
increase mitochondrial biogenesis, mitochondrial antioxidant
generation, and mitochondrial FAO in kidney disease in obesity
and diabetes [29].
Oxidative stress
An overload of fatty acids into mitochondria, after an increased
intake or an insulin‐resistance condition, may lead to an
increase in the permeability of the inner mitochondrial mem-
brane. This occurrence leads to the dissipation of the mem-
brane potential and the loss of ATP synthesis capacity, resulting
in mitochondrial function impairment and an enhanced ROS
generation. The increase of FAO, inducing an increased elec-
tron flux in the electron transport chain (ETC), may generate
an “electron leakage” due to reduction of the activity of ETC
complexes, thus ensuring a direct reaction between electrons
and oxygen, leading to the formation of ROS, rather than the
normal reaction mediated by cytochrome c oxidase that com-
bines oxygen and protons in order to form water [63]. The
incomplete or suboptimal FAO leads to accumulation of long
chain acylcarnitines, ceramides, and diacylglycerols, lipotoxic
intermediates that may act as indirect sources of ROS. In addi-
tion to pro‐oxidant mechanisms, in an experimental model of
NASH, a decreased activity of several detoxifying enzymes
was observed. Glutathione peroxidase (GPx) activity is
reduced probably in consequence of GSH depletion and
impaired transport of cytosolic GSH into the mitochondrial
matrix [64]. The polymorphism C47T of the SOD2 gene,
encoding for manganese superoxide dismutase, is associated
with a reduction of activity of this enzyme resulting in an
increased ROS production and a high susceptibility to devel-
oping NASH and advanced fibrosis in NAFLD [65].
The oxidative stress‐induced impairment of mitochondrial
permeability can generate an influx of calcium (Ca2+) and
iron into the mitochondria. Iron in the presence of H2O2
favors Fenton’s reaction to generate more hydroxyl radicals.
Further, Ca2+ can stimulate the inducible NO synthase‐medi-
ated nitric oxide radical (NO●) production and apoptotic cell
death. A significant increase in the NO● level during severe
hepatocyte injury may induce the progression to necrotic
cell  death [66]. The peroxynitrite produced from NO● and
superoxide radical is one of the important mediators of free
radical toxicity.
Endoplasmic reticulum stress
The ER stress and the unfolded protein response (UPR) play
important roles in NAFLD/NASH and the associated kidney
disease. The activation of ER stress may initially be protective,
while chronic ER stress may result in apoptosis and fibrosis.
The UPR activation in the ER is mediated by three major sign-
aling pathways, including activating transcription factor 6
(ATF6), inositol‐requiring enzyme 1α (IRE1α), and PRKR‐
like ER kinase (PERK). Upon accumulation of misfolded pro-
teins in the ER, BiP (GRP78) disassociates from the sensors
and binds to unfolded proteins, which then activates the
sensors.
In the liver while lipid accumulation in hepatocytes can ini-
tiate ER stress, activation of ER stress can also induce lipogen-
esis and thus creating a vicious cycle for the progression of
steatosis. ER stress induces insulin induced gene (INSIG) deg-
radation and activation of the SREBP pathway [32, 67]. In
addition, ER stress also induces increased expression of the
VLDL receptor which results in increased lipoprotein delivery
to the liver [32]. Increased lipid accumulation and chronic ER
stress in the hepatocyte results increased inflammation and
causes hepatocyte apoptosis. In addition, ER stress also stimu-
lates hepatic stellate cell (HSC) collagen I secretion, which
mediates fibrosis [68]. CHOP, the C/EBP homologous protein,
plays an important role in ER stress induced apoptosis and
fibrosis [68].
In the kidney chronic ER stress and UPR activation [69], as it
occurs in the setting of obesity and the metabolic syndrome
[30], contributes to podocyte injury, apoptosis, proteinuria, and
CKD. Furthermore, albumin has been shown to induce ER
stress in renal tubular cells by increasing intracellular calcium
levels, inducing UPR‐dependent upregulation of lipocalin 2,
resulting in apoptosis.
640 THE LIVER:  NAFLD, NASH, AND CKD
Inflammation
Inflammation is a physiological response to tissue injury or
infection that leads to secretion of various inflammatory media-
tors, such as cytokines, chemokines, and eicosanoids, which
coordinate cellular defense mechanisms and tissue repair. The
persistence of inflammatory activity over time results in chronic
inflammatory changes that exacerbate tissue injury and may
result in an abnormal wound healing response. In the case of
NAFLD, inflammation contributes to the development of NASH
and liver fibrosis. NAFLD refers to a spectrum of histological
abnormalities in the liver, ranging from isolated steatosis with
no or minimal inflammatory activity and no evidence of cell
damage (NAFL) to NASH, which is characterized by steatosis,
inflammation, and hepatocellular injury, hallmarked by the
presence of hepatocyte ballooning, with different degrees of
fibrosis [70]. Many factors are associated with chronic systemic
immune response, a phenotype that is common in many indi-
viduals with NAFLD, CKD, T2DM, or CVD. A central issue in
this field is identification of those factors that trigger inflamma-
tion, thus fueling the transition from non‐alcoholic fatty liver to
NASH [71]. These triggers of liver inflammation may have their
origins inside as well as outside of the liver.
Lipotoxicity is one of the major mechanisms underlying
hepatocyte dysfunction leading to disease progression in
NASH [72]. Increased hepatocyte free fatty acid (FFA) influx
results in the formation of lipotoxic intermediates in the liver,
such as ceramides, diacylglycerols, LPC, and oxidized fatty
acid and cholesterol metabolites, which act as ROS [73].
Recent studies have revealed that hepatocyte inflammasome
activation may be an important link between the initial meta-
bolic stress and subsequent hepatocyte death and stimulation
of fibrosis in NASH [74]. The accumulation of free cholesterol
in Kupffer cells and hepatic stellate cells can also activate the
NLRP3 inflammasome [75].
While liver and kidney are important organs for lipid metabo-
lism, the increased proinflammatory cytokine production by
lipotoxicity, which causes ectopic lipid deposition that occurs in
more advanced forms of NAFLD is also likely to have a patho-
genetic role in the development of extra‐hepatic complications,
such as CVD and CKD.
Deterioration to NASH results in the production or activa-
tion of multiple inflammatory mediators such as NF‐κB path-
way, cytokines [76], lipopolysaccharides (LPS), and ROS,
which can lead to insulin resistance, endothelial dysfunction,
and a tissue inflammatory infiltrate that can amplify systemic
chronic inflammation. A study in patients with T2DM, with or
without persistent hepatic inflammation (owing to chronic
HBV infection) suggested that the presence of liver inflamma-
tion is a key mediator in the increased risk of CKD. Although
the presence of T2DM undoubtedly increases the risk of CVD
in patients with NAFLD, several studies have shown that
potential mediators of vascular and renal damage occur more
frequently in patients with NAFLD regardless of whether or
not they also have T2DM [77]. Administration of sera from
patients with CKD or uninephrectomy to rodents or cultured
adipocytes induces lipodystrophy, ectopic fat redistribution
from adipose tissue to liver and muscle, insulin resistance, and
glucose intolerance [78, 79].
Although obesity and NAFLD injure the kidney, growing
e­vidence suggests CKD may also contribute to NAFLD and
insulin resistance. Taken together, increased levels of lipids
and  the increased inflammatory response are both thought to
be important pathogenetic factors that are not only involved in
the development of NASH but are also significant factors in the
development and progression of CKD.
Microbiome
The liver is the key metabolic organ exposed to high levels of
intestinal products through the tight bidirectional linkages in the
biliary tract, portal vein, and systemic circulation. An altered
microbiome (or “dysbiosis”) may lead to an increase in intesti-
nal permeability, which can amplify many of these effects
derived from intestine [80].
Studies of the microbiome are relatively nascent; however, it
is anticipated that there will be considerable progress in linking
its role to NASH. The changes in specific microbial products,
secondary to altered gut microbial composition, and the changes
in intestinal permeability and function can affect hepatic struc-
ture and function to further increase the risk of NAFLD. Patients
with NAFLD have a higher prevalence of microbial dysbiosis.
Human studies have documented that the gut microbiome
among patients with NASH is less complex than that of healthy
subjects [81, 82] and indicate that weight loss also alters the
microbiome. Adults with NAFLD showed increased serum tri-
methylamine N‐oxide (TMAO) [83] and decreased production
of phosphatidylcholine [84]. Plausible mechanistic links
between an altered microbiome and fatty liver are emerging and
include the potential for bacterial proteins to function as ligands
for G protein‐coupled receptors [85, 86] (Figure 49.4).
Dysbiosis has been described in patients with obesity [87] or
other features of metabolic syndrome [88] and in those with
established NAFLD [89], T2DM [90, 91], CVD [92], or CKD
[93]. Several potential pathways, factors, and processes might
link dysbiosis, or mediators of the gut microbiota, and NAFLD
to CKD risk factors and vascular and renal diseases (Figure 49.2)
[1]. Dysbiosis can potentially influence NAFLD, CKD, and
obesity through multiple and complex mechanisms. It has been
reported that changes in the gut microbiome of patients with
CKD relate to lower levels of Bifidobacteriaceae and
Lactobacillaceae and to higher levels of Enterobacteriaceae
[94]. Microbial fermentation of dietary fiber in the intestine by
anaerobic bacteria such as Bifidobacteria and Lactobacilli
results in the formation of short‐chain fatty acids (SFCAs)
including acetate, propionate, and butyrate, which have the
potential to influence hepatic lipogenesis and gluconeogenesis.
Thus, decreased levels of SCFAs in the setting of NAFLD might
contribute to the development of liver adiposity and hepatic
insulin resistance [95]. The intestinal microbiota also produces
trimethylamine, p‐cresol, and indole from dietary nutrients such
as choline, phenylalanine/tyrosine, and tryptophan, respectively.
Further metabolism in the liver by oxidation or sulfation pro-
duces ionically charged water‐soluble molecules, such as
TMAO, p‐cresol sulfate, and indole sulfate, which can be
excreted in the urine. Plasma TMAO levels are elevated in
patients with CKD and portend poorer long‐term survival [96].
Indole sulfate, which is cleared by the proximal tubules, is
Figure 49.4  Interplay between the liver and gut microbiota in alcoholic liver disease and NAFLD. Intestinal dysbiosis and bacterial overgrowth
are observed in both alcoholic liver disease (ALD) (part a) and non‐alcoholic fatty liver disease (NAFLD) (part b). Bacterial overgrowth causes an
increase in secondary bile acids (BAs), which modulates farnesoid X receptor (FXR)‐mediated hepatic synthesis of BA, leading to an overall
increase in hepatic BA synthesis. A reduction in hepatic phosphatidylcholine is also seen in both ALD and NAFLD, which causes triglyceride
accumulation in the liver (fatty liver). While ALD‐associated dysbiosis is characterized by reduction in Lactobacillus and Candida overgrowth,
patients with NAFLD have a higher abundance of Lactobacillus (effects on fungal population remain to be investigated). In ALD and NAFLD,
increased ethanol and its metabolite acetaldehyde in the intestinal lumen mediate weakening of intestinal tight junctions. Consequently, increased
translocation of microbial‐associated molecular patterns (MAMPs) (seen in ALD and NAFLD) and gut metabolites, such as acetaldehyde, acetate
(seen in ALD) and trimethylamine (TMA, seen in NAFLD), elicits intestinal and hepatic inflammatory responses, leading to progressive liver dam-
age. AMP, antimicrobial peptides; EtOH, ethanol; HFD, high‐fat diet; LCFAs, long‐chain fatty acids, TMAO, trimethylamine N‐oxide. Reproduced
with permission of Springer Nature, Figure 3 in [86].
642 THE LIVER:  REFERENCES
proinflammatory and potentially toxic to the kidneys as it
increases the risk of tubulointerstitial fibrosis [97].
It is now widely accepted that liver and kidney damage can
result from extensive interaction with the gut microbiota through
specialized molecules, such as TMA. As the role of the microbiota
in liver disease progression, prognosis, and treatment is increas-
ingly recognized, it is necessary to make focused and conscious
efforts to study the effect of the microbiome to efficiently address
the socioeconomic burden of this spectrum of liver diseases.
Increasing evidence suggests that the liver and kidneys share path-
ways including microbiome and dysbiosis that are intrinsically
linked to each other. These indicate that the prevalence of CKD is
markedly increased among patients with NAFLD, and that the
presence and severity of NAFLD is associated with an increased
incidence of CKD, and that this association could be independent
of multiple cardiorenal risk factors. Taken together, more active
and systematic research for the link between the microbiome and
liver and kidney disease are needed to understand the mechanisms
and causal association between these diseases.
Bile acid signaling
Bile acids are produced in the liver from cholesterol and are
metabolized by enzymes derived from the gut microbiome.
Evidence shows that bile acids are critically important for main-
taining a healthy gut microbiota [98]. The synthesis of bile acids
is tightly regulated by negative feedback inhibition through the
nuclear receptor FXR [99], which is a transcription factor that
binds to the promoter region and initiates the expression of a
wide range of target genes [100]. FXR is expressed in several
tissues including the liver, intestine, and kidney [101]. In the
liver, bile acid‐activated FXR induces the expression of small
heterodimer partner (SHP), which binds to liver receptor
homolog‐1 (LRH‐1) and thereby inhibits expression of the
Cyp7a1 gene [102]. The primary bile acids produced in humans
are chenodeoxycholic acid (CDCA) and cholic acid (CA), while
rodents produce CA and muricholic acids (MCAs), predomi-
nantly beta‐MCA (βMCA) [103].
The most potent endogenous ligand for FXR is CDCA, fol-
lowed by CA, DCA, and LCA [104]. UDCA, TαMCA, TβMCA,
and GβMCA inhibit FXR activation [98].
A positive phase 2b clinical trial of obeticholic acid (OCA)
(25 mg per day) was terminated early because of demonstration
of efficacy during an interim analysis planned a priori [105].
However, OCA, in 25 mg per day doses, causes both pruritus
and moderate increases in low‐density lipoprotein (LDL) cho-
lesterol in some individuals. Several small‐molecule agonists of
FXR that do not have a bile acid structural backbone have been
developed on the premise that they will not increase LDL cho-
lesterol or cause pruritus, but this has not yet been proven.
Additionally, FXR induces the release of the growth factor
FGF19 from the intestine upon bile acid binding to the recep-
tor [106], which has beneficial effects on NASH in animal
models, although some results are conflicting [107, 108]. In
a phase 2a study, FGF19 analog yielded a dramatic reduction
in hepatic fat and liver enzymes in patients with biopsy‐
confirmed NASH [107].
TGR5 is another bile acid‐responsive receptor involved in
host metabolism. TGR5 is a plasma membrane‐bound G
protein‐coupled receptor ubiquitously expressed with high
expression in gallbladder, placenta, lung, spleen, intestine, liver,
brown and white adipose tissue, skeletal muscle, and bone mar-
row [109]. TGR5 is activated mainly by the secondary bile acids
LCA and DCA [110]. TGR5 signaling controls glucose homeo-
stasis by increased energy expenditure in brown adipose tissue
and muscle and by increased GLP‐1 release in intestinal L cells
[111]. At present it is unclear whether FXR and TGR5 signal
directly or indirectly via alterations in the microbiota and bile
acid metabolism that produce agonistic or antagonistic signaling
through each other.
Glucagon‐like peptide (GLP)‐1 is an intestinal hormone gen-
erated through the proteolytic processing of proglucagon that
stimulates insulin secretion and inhibits secretion of glucagon.
GLP‐1 is also an insulin sensitizer with additional metabolic
effects that contribute to its anti‐NASH activity [112]. Liraglutide,
a GLP‐1 agonist requiring daily injection, improved NASH his-
tology in a small pilot study [113]. Interestingly, L cells express
FXR and FXR also regulates GLP‐1 synthesis [114].
REFERENCES
1. Targher, G. and Byrne, C.D. Non‐alcoholic fatty liver disease: an emerging
driving force in chronic kidney disease. Nat Rev Nephrol, 2017;13(5):
297–310.
2. Yeung, M.W., Wong, G.L., Choi, K.C. et al. Advanced liver fibrosis but not
steatosis is independently associated with albuminuria in Chinese patients
with type 2 diabetes. J Hepatol, 2017.
3. Sinn, D.H., Kang, D., Jang, H.R. et al. Development of chronic kidney dis-
ease in patients with non‐alcoholic fatty liver disease: a cohort study. J
Hepatol, 2017;67(6):1274–80.
4. Mantovani, A., Zaza, G., Byrne, C.D. et al. Nonalcoholic fatty liver disease
increases risk of incident chronic kidney disease: a systematic review and
meta‐analysis. Metabolism, 2018;79:64–76.
5. Sole, C., Pose, E., Sola, E., and Gines, P. Hepatorenal syndrome in the era of
acute kidney injury. Liver Int, 2018;38(11):1891–1901.
6. Angeli, P., Gines, P., Wong, F. et  al. Diagnosis and management of acute
kidney injury in patients with cirrhosis: revised consensus recommendations
of the International Club of Ascites. J Hepatol, 2015;62(4):968–74.
7. Durand, F., Graupera, I., Gines, P., Olson, J.C., and Nadim, M.K.
Pathogenesis of hepatorenal syndrome: implications for therapy. Am J
Kidney Dis, 2016;67(2):318–28.
8. Sanyal, A.J., Boyer, T.D., Frederick, R.T. et al. Reversal of hepatorenal syn-
drome type 1 with terlipressin plus albumin vs. placebo plus albumin in a
pooled analysis of the OT‐0401 and REVERSE randomised clinical studies.
Aliment Pharmacol Ther, 2017;45(11):1390–402.
9. Glass, L. and Sharma, P. Evidence‐based therapeutic options for hepatorenal
syndrome. Gastroenterology, 2016;150(4):1031–3.
10. Wong, F., Leung, W., Al Beshir, M., Marquez, M., and Renner, E.L.
Outcomes of patients with cirrhosis and hepatorenal syndrome type 1 treated
with liver transplantation. Liver Transpl, 2015;21(3):300–7.
11. Cavallin, M., Kamath, P.S., Merli, M. et al. Terlipressin plus albumin versus
midodrine and octreotide plus albumin in the treatment of hepatorenal syn-
drome: a randomized trial. Hepatology, 2015;62(2):567–74.
12. Cohen, S.D., Kopp, J.B., and Kimmel, P.L. Kidney diseases associated with
human immunodeficiency virus infection. N Engl J Med, 2017;377(24):
2363–74.
13. Rutledge, S.M., Chung, R.T., and Sise, M.E. Treatment of hepatitis C virus
infection in patients with mixed cryoglobulinemic syndrome and cryoglobu-
linemic glomerulonephritis. Hemodial Int, 2018;22(1):S81–96.
14. Collins, J.M., Raphael, K.L., Terry, C. et al. Hepatitis B virus reactivation
during successful treatment of hepatitis C virus with sofosbuvir and sime-
previr. Clin Infect Dis, 2015;61(8):1304–6.
15. Sise, M.E. Hepatitis C virus infection and the kidney. Nephrol Dial
Transplant, 2018;34(3):415–8.
 49:  The Kidney in Liver Disease 643
16. Rodrigues, E.M.F., Fernandes, R., Susin, R., and Fior, B. Immune reconstitu-
tion inflammatory syndrome as a cause of autoimmune hepatitis and acute
liver failure. Rev Bras Ter Intensiva, 2017;29(3):382–5.
17. Gowda, C., Newcomb, C.W., Liu, Q. et al. Risk of acute liver injury with
antiretroviral therapy by viral hepatitis status. Open Forum Infect Dis,
2017;4(2):ofx012.
18. Nadkarni, G.N., Konstantinidis, I., and Wyatt, C.M. HIV and the aging kid-
ney. Curr Opin HIV AIDS, 2014;9(4):340–5.
19. Genovese, G., Friedman, D.J., Ross, M.D. et al. Association of trypanolytic
ApoL1 variants with kidney disease in African Americans. Science,
2010;329(5993):841–5.
20. Tzur, S., Rosset, S., Shemer, R. et al. Missense mutations in the APOL1 gene
are highly associated with end stage kidney disease risk previously attributed
to the MYH9 gene. Hum Genet, 2010;128(3):345–50.
21. Kopp, J.B., Nelson, G.W., Sampath, K. et  al. APOL1 genetic variants in
focal segmental glomerulosclerosis and HIV‐associated nephropathy. J Am
Soc Nephrol, 2011;22(11):2129–37.
22. Fishman, J.A. Infection in organ transplantation. Am J Transplant,
2017;17(4):856–79.
23. Hales, C.M., Fryar, C.D., Carroll, M.D., Freedman, D.S., and Ogden, C.L.
Trends in obesity and severe obesity prevalence in US youth and adults by
sex and age, 2007–2008 to 2015–2016. JAMA, 2018;319(16):1723–5.
24. D’Agati, V.D., Chagnac, A., de Vries, A.P. et al. Obesity‐related glomeru-
lopathy: clinical and pathologic characteristics and pathogenesis. Nat Rev
Nephrol, 2016;12(8):453–71.
25. Lakkis, J.I. and Weir, M.R. Obesity and kidney disease. Prog Cardiovasc
Dis, 2018;61(2):157–67.
26. Whaley‐Connell, A. and Sowers, J.R. Obesity and kidney disease: from
population to basic science and the search for new therapeutic targets.
Kidney Int, 2017;92(2):313–23.
27. Wang, X.X., Wang, D., Luo, Y. et al. FXR/TGR5 dual agonist prevents pro-
gression of nephropathy in diabetes and obesity. J Am Soc Nephrol,
2018;29(1):118–37.
28. Wang, X.X., Jiang, T., Shen, Y., Adorini, L. et al. The farnesoid X receptor
modulates renal lipid metabolism and diet‐induced renal inflammation, fibro-
sis, and proteinuria. Am J Physiol Renal Physiol, 2009;297(6):F1587–96.
29. Wang, X.X., Edelstein, M.H., Gafter, U. et al. G protein‐coupled bile acid
receptor TGR5 activation inhibits kidney disease in obesity and diabetes. J
Am Soc Nephrol, 2016;27(5):1362–78.
30. Gai, Z., Gui, T., Hiller, C., and Kullak‐Ublick, G.A. Farnesoid X receptor
protects against kidney injury in uninephrectomized obese mice. J Biol
Chem, 2016;291(5):2397–411.
31. Friedman, S.L., Neuschwander‐Tetri, B.A., Rinella, M., and Sanyal, A.J.
Mechanisms of NAFLD development and therapeutic strategies. Nat Med,
2018;24(7):908–22.
32. Baiceanu, A., Mesdom, P., Lagouge, M., and Foufelle, F. Endoplasmic retic-
ulum proteostasis in hepatic steatosis. Nat Rev Endocrinol, 2016;12(12):
710–22.
33. Artunc, F., Schleicher, E., Weigert, C., Fritsche, A., Stefan, N., and Haring,
H.U. The impact of insulin resistance on the kidney and vasculature. Nat Rev
Nephrol, 2016;12(12):721–37.
34. Samuel, V.T. and Shulman, G.I. Nonalcoholic fatty liver disease as a nexus
of metabolic and hepatic diseases. Cell Metab, 2018;27(1):22–41.
35. Gnudi, L., Coward, R.J.M., and Long, D.A. Diabetic nephropathy: perspec-
tive on novel molecular mechanisms. Trends Endocrinol Metab,
2016;27(11):820–30.
36. Lay, A.C., Hurcombe, J.A., Betin, V.M.S. et  al. Prolonged exposure of
mouse and human podocytes to insulin induces insulin resistance through
lysosomal and proteasomal degradation of the insulin receptor. Diabetologia,
2017;60(11):2299–311.
37. Wang, X.X., Levi, J., Luo, Y. et al. SGLT2 protein expression is increased in
human diabetic nephropathy: SGLT2 protein inhibition decreases renal lipid
accumulation, inflammation, and the development of nephropathy in dia-
betic mice. J Biol Chem, 2017;292(13):5335–48.
38. Wang, D., Luo, Y., Wang, X. et al. The sodium‐glucose cotransporter 2 inhib-
itor dapagliflozin prevents renal and liver disease in western diet induced
obesity mice. Int J Mol Sci, 2018;19(1).
39. Koo, S.H. Nonalcoholic fatty liver disease: molecular mechanisms for the
hepatic steatosis. Clin Mol Hepatol, 2013;19(3):210–5.
40. Benhamed, F., Denechaud, P.D., Lemoine, M. et al. The lipogenic transcrip-
tion factor ChREBP dissociates hepatic steatosis from insulin resistance in
mice and humans. J Clin Invest, 2012;122(6):2176–94.
41. Ponugoti, B., Kim, D.H., Xiao, Z. et  al. SIRT1 deacetylates and inhibits
SREBP‐1C activity in regulation of hepatic lipid metabolism. J Biol Chem,
2010;285(44):33959–70.
42. Porstmann, T., Santos, C.R., Griffiths, B. et al. SREBP activity is regulated
by mTORC1 and contributes to Akt‐dependent cell growth. Cell Metab,
2008;8(3):224–36.
43. Begriche, K., Massart, J., Robin, M.A., Bonnet, F., and Fromenty, B.
Mitochondrial adaptations and dysfunctions in nonalcoholic fatty liver dis-
ease. Hepatology, 2013;58(4):1497–507.
44. Fabbrini, E., Mohammed, B.S., Magkos, F., Korenblat, K.M., Patterson,
B.W., and Klein, S. Alterations in adipose tissue and hepatic lipid kinetics in
obese men and women with nonalcoholic fatty liver disease. Gastroenterology,
2008;134(2):424–31.
45. Quiroga, A.D. and Lehner, R. Pharmacological intervention of liver triacylg-
lycerol lipolysis: the good, the bad and the ugly. Biochem Pharmacol,
2018;155:233–41.
46. Ioannou, G.N. The role of cholesterol in the pathogenesis of NASH. Trends
Endocrinol Metab, 2016;27(2):84–95.
47. Fiorucci, S., Biagioli, M., and Distrutti, E. Future trends in the treatment of
non‐alcoholic steatohepatitis. Pharmacol Res, 2018;134:289–98.
48. Sun, L., Halaihel, N., Zhang, W., Rogers, T., and Levi, M. Role of sterol
regulatory element‐binding protein 1 in regulation of renal lipid metabolism
and glomerulosclerosis in diabetes mellitus. J Biol Chem, 2002;277(21):
18919–27.
49. Jiang, T., Wang, Z., Proctor, G. et al. Diet‐induced obesity in C57BL/6J mice
causes increased renal lipid accumulation and glomerulosclerosis via a sterol
regulatory element‐binding protein‐1c‐dependent pathway. J Biol Chem,
2005;280(37):32317–25.
50. Kume, S., Uzu, T., Araki, S. et al. Role of altered renal lipid metabolism in
the development of renal injury induced by a high‐fat diet. J Am Soc Nephrol,
2007;18(10):2715–23.
51. Wu, Y., Liu, Z., Xiang, Z. et  al. Obesity‐related glomerulopathy: insights
from gene expression profiles of the glomeruli derived from renal biopsy
samples. Endocrinology, 2006;147(1):44–50.
52. Jiang, T., Wang, X.X., Scherzer, P. et  al. Farnesoid X receptor modulates
renal lipid metabolism, fibrosis, and diabetic nephropathy. Diabetes,
2007;56(10):2485–93.
53. Wang, X.X., Jiang, T., Shen, Y. et al. Diabetic nephropathy is accelerated by
farnesoid X receptor deficiency and inhibited by farnesoid X receptor activa-
tion in a type 1 diabetes model. Diabetes, 2010;59(11):2916–27.
54. Yang, X., Okamura, D.M., Lu, X. et  al. CD36 in chronic kidney disease:
novel insights and therapeutic opportunities. Nat Rev Nephrol, 2017;13(12):
769–81.
55. Herman‐Edelstein, M., Scherzer, P., Tobar, A., Levi, M., and Gafter, U.
Altered renal lipid metabolism and renal lipid accumulation in human dia-
betic nephropathy. J Lipid Res, 2014;55(3):561–72.
56. Wang, X.X., Jiang, T., Shen, Y. et al. Vitamin D receptor agonist doxercalcif-
erol modulates dietary fat‐induced renal disease and renal lipid metabolism.
Am J Physiol Renal Physiol, 2011;300(3):F801–10.
57. Proctor, G., Jiang, T., Iwahashi, M., Wang, Z., Li, J., and Levi, M. Regulation
of renal fatty acid and cholesterol metabolism, inflammation, and fibrosis in
Akita and OVE26 mice with type 1 diabetes. Diabetes, 2006;55(9):2502–9.
58. Marra, F. and Svegliati‐Baroni, G. Lipotoxicity and the gut‐liver axis in
NASH pathogenesis. J Hepatol, 2018;68(2):280–95.
59. Caldwell, S.H., Swerdlow, R.H., Khan, E.M. et al. Mitochondrial abnormali-
ties in non‐alcoholic steatohepatitis. J Hepatol, 1999;31(3):430–4.
60. Ding, R.B., Bao, J., and Deng, C.X. Emerging roles of SIRT1 in fatty liver
diseases. Int J Biol Sci, 2017;13(7):852–67.
61. Morigi, M., Perico, L., and Benigni, A. Sirtuins in renal health and disease. J
Am Soc Nephrol, 2018;29(7):1799–809.
62. Tang, C., Cai, J., and Dong, Z. Mitochondrial dysfunction in obesity‐related
kidney disease: a novel therapeutic target. Kidney Int, 2016;90(5):930–3.
63. Grattagliano, I., de Bari, O., Bernardo, T.C., Oliveira, P.J., Wang, D.Q., and
Portincasa, P. Role of mitochondria in nonalcoholic fatty liver disease‐‐from
origin to propagation. Clin Biochem, 2012;45(9):610–8.
64. Caballero, F., Fernandez, A., Matias, N. et  al. Specific contribution of
methionine and choline in nutritional nonalcoholic steatohepatitis: impact on
mitochondrial S‐adenosyl‐L‐methionine and glutathione. J Biol Chem,
2010;285(24):18528–36.
65. Al‐Serri, A., Anstee, Q.M., Valenti, L. et al. The SOD2 C47T polymorphism
influences NAFLD fibrosis severity: evidence from case‐control and intra‐
familial allele association studies. J Hepatol, 2012;56(2):448–54.
644 THE LIVER:  REFERENCES
66. Ajith, T.A. Role of mitochondria and mitochondria‐targeted agents in non‐
alcoholic fatty liver disease. Clin Exp Pharmacol Physiol, 2018;45(5):
413–21.
67. Lebeaupin, C., Vallee, D., Hazari, Y., Hetz, C., Chevet, E., and Bailly‐Maitre,
B. Endoplasmic reticulum stress signaling and the pathogenesis of non‐alco-
holic fatty liver disease. J Hepatol, 2018;69(4):927–47.
68. Kropski, J.A. and Blackwell, T.S. Endoplasmic reticulum stress in the patho-
genesis of fibrotic disease. J Clin Invest, 2018;128(1):64–73.
69. Cybulsky, A.V. Endoplasmic reticulum stress, the unfolded protein response
and autophagy in kidney diseases. Nat Rev Nephrol, 2017;13(11):681–96.
70. Brunt, E.M., Wong, V.W., Nobili, V. et al. Nonalcoholic fatty liver disease.
Nat Rev Dis Primers, 2015;1:15080.
71. Schuster, S., Cabrera, D., Arrese, M., and Feldstein, A.E. Triggering and
resolution of inflammation in NASH. Nat Rev Gastroenterol Hepatol,
2018;15(6):349–64.
72. Wree, A., Broderick, L., Canbay, A., Hoffman, H.M., and Feldstein, A.E.
From NAFLD to NASH to cirrhosis‐new insights into disease mechanisms.
Nat Rev Gastroenterol Hepatol, 2013;10(11):627–36.
73. Han, M.S., Park, S.Y., Shinzawa, K. et  al. Lysophosphatidylcholine as a
death effector in the lipoapoptosis of hepatocytes. J Lipid Res, 2008;
49(1):84–97.
74. Csak, T., Ganz, M., Pespisa, J., Kodys, K., Dolganiuc, A., and Szabo, G. Fatty
acid and endotoxin activate inflammasomes in mouse hepatocytes that release
danger signals to stimulate immune cells. Hepatology, 2011;54(1):133–44.
75. Wree, A., McGeough, M.D., Inzaugarat, M.E. et al. NLRP3 inflammasome
driven liver injury and fibrosis: roles of IL‐17 and TNF in mice. Hepatology,
2017.
76. Sharma, M., Mitnala, S., Vishnubhotla, R.K., Mukherjee, R., Reddy, D.N.,
and Rao, P.N. The riddle of nonalcoholic fatty liver disease: progression
from nonalcoholic fatty liver to nonalcoholic steatohepatitis. J Clin Exp
Hepatol, 2015;5(2):147–58.
77. Targher, G., Day, C.P., and Bonora, E. Risk of cardiovascular disease in
patients with nonalcoholic fatty liver disease. N Engl J Med, 2010;
363(14):1341–50.
78. Pelletier, C.C., Koppe, L., Croze, M.L. et al. White adipose tissue overpro-
duces the lipid‐mobilizing factor zinc alpha2‐glycoprotein in chronic kidney
disease. Kidney Int, 2013;83(5):878–86.
79. Axelsson, J., Åström, G., Sjölin, E. et al. Uraemic sera stimulate lipolysis in
human adipocytes: role of perilipin. Nephrol Dial Transplant, 2011;
26(8):2485–91.
80. Leung, C., Rivera, L., Furness, J.B., and Angus, P.W. The role of the gut
microbiota in NAFLD. Nat Rev Gastroenterol Hepatol, 2016;13(7):
412–25.
81. Betrapally, N.S., Gillevet, P.M., and Bajaj, J.S. Changes in the intestinal
microbiome and alcoholic and nonalcoholic liver diseases: causes or effects?
Gastroenterology, 2016;150(8):1745–55.e3.
82. Loomba, R., Seguritan, V., Li, W. et al. Gut microbiome based metagenomic
signature for non‐invasive detection of advanced fibrosis in human nonalco-
holic fatty liver disease. Cell Metab, 2017;25(5):1054–62.e5.
83. Chen, Y.M., Liu, Y., Zhou, R.F. et  al. Associations of gut‐flora‐dependent
metabolite trimethylamine‐N‐oxide, betaine and choline with non‐alcoholic
fatty liver disease in adults. Sci Rep, 2016;6:19076.
84. Arendt, B.M., Ma, D.W., Simons, B. et al. Nonalcoholic fatty liver disease is
associated with lower hepatic and erythrocyte ratios of phosphatidylcholine
to phosphatidylethanolamine. Appl Physiol Nutr Metab, 2013;38(3):
334–40.
85. Cohen, L.J., Esterhazy, D., Kim, S.H. et  al. Commensal bacteria make
GPCR ligands that mimic human signalling molecules. Nature,
2017;549(7670):48–53.
86. Tripathi, A., Debelius, J., Brenner, D.A. et  al. The gut‐liver axis and the
intersection with the microbiome. Nat Rev Gastroenterol Hepatol,
2018;15(7):397–411.
87. Liu, R., Hong, J., Xu, X. et  al. Gut microbiome and serum metabolome
alterations in obesity and after weight‐loss intervention. Nat Med, 2017;
23:859.
88. Mehal, W.Z. The Gordian knot of dysbiosis, obesity and NAFLD. Nat Rev
Gastroenterol Hepatol, 2013;10(11):637–44.
89. Tilg, H., Cani, P.D., and Mayer, E.A. Gut microbiome and liver diseases.
Gut, 2016;65(12):2035–44.
90. Qin, J., Li, Y., Cai, Z. et  al. A metagenome‐wide association study of gut
microbiota in type 2 diabetes. Nature, 2012;490(7418):55–60.
  91. Utzschneider, K.M., Kratz, M., Damman, C.J., and Hullar, M. Mechanisms
linking the gut microbiome and glucose metabolism. J Clin Endocrinol
Metab, 2016;101(4):1445–54.
  92. Koopen, A.M., Groen, A.K., and Nieuwdorp, M. Human microbiome as
therapeutic intervention target to reduce cardiovascular disease risk. Curr
Opin Lipidol, 2016;27(6):615–22.
  93. Briskey, D., Tucker, P.S., Johnson, D.W., and Coombes, J.S. Microbiota
and the nitrogen cycle: Implications in the development and progression of
CVD and CKD. Nitric Oxide, 2016;57:64–70.
  94. Sampaio‐Maia, B., Simoes‐Silva, L., Pestana, M., Araujo, R., and Soares‐
Silva, I.J. The role of the gut microbiome on chronic kidney disease. Adv
Appl Microbiol, 2016;96:65–94.
  95. Scorletti, E. and Byrne, C.D. Extrahepatic diseases and NAFLD: the trian-
gular relationship between NAFLD, type 2‐diabetes and dysbiosis. Dig Dis,
2016;34(1):11–8.
  96. Tang, W.H., Wang, Z., Kennedy, D.J. et al. Gut microbiota‐dependent tri-
methylamine N‐oxide (TMAO) pathway contributes to both development
of renal insufficiency and mortality risk in chronic kidney disease. Circ
Res, 2015;116(3):448–55.
  97. Nallu, A., Sharma, S., Ramezani, A., Muralidharan, J., and Raj, D. Gut
microbiome in chronic kidney disease: challenges and opportunities. Transl
Res, 2017;179:24–37.
  98. Wahlström, A., Sayin Sama I., Marschall, H‐U., and Bäckhed, F. Intestinal
crosstalk between bile acids and microbiota and its impact on host metabo-
lism. Cell Metab, 2016;24(1):41–50.
  99. Sinal, C.J., Tohkin, M., Miyata, M., Ward, J.M., Lambert, G., and Gonzalez,
F.J. Targeted disruption of the nuclear receptor FXR/BAR impairs bile acid
and lipid homeostasis. Cell, 2000;102(6):731–44.
100. Teodoro, J.S., Rolo, A.P., and Palmeira, C.M. Hepatic FXR: key regulator
of whole‐body energy metabolism. Trends Endocrinol Metab, 2011;
22(11):458–66.
101. Lefebvre, P., Cariou, B., Lien, F., Kuipers, F., and Staels, B. Role of bile
acids and bile acid receptors in metabolic regulation. Physiol Rev,
2009;89(1):147–91.
102. Goodwin, B., Jones, S.A., Price, R.R. et  al. A regulatory cascade of the
nuclear receptors FXR, SHP‐1, and LRH‐1 represses bile acid biosynthesis.
Mol Cell. 2000;6(3):517–26.
103. Takahashi, S., Fukami, T., Masuo, Y. et al. Cyp2c70 is responsible for the
species difference in bile acid metabolism between mice and humans. J
Lipid Res, 2016;57(12):2130–7.
104. Makishima, M., Okamoto, A.Y., Repa, J.J. et al. Identification of a nuclear
receptor for bile acids. Science, 1999;284(5418):1362–5.
105. Neuschwander‐Tetri, B.A., Loomba, R., Sanyal, A.J. et  al. Farnesoid X
nuclear receptor ligand obeticholic acid for non‐cirrhotic, non‐alcoholic
steatohepatitis (FLINT): a multicentre, randomised, placebo‐controlled
trial. Lancet, 2015;385(9972):956–65.
106. Nies, V.J., Sancar, G., Liu, W. et al. Fibroblast growth factor signaling in
metabolic regulation. Front Endocrinol,2015;6:193.
107. Harrison, S.A., Rinella, M.E., Abdelmalek, M.F. et al. NGM282 for treat-
ment of non‐alcoholic steatohepatitis: a multicentre, randomised, double‐
blind, placebo‐controlled, phase 2 trial. Lancet, 2018;391(10126):1174–85.
108. Jiang, C., Xie, C., Li, F. et al. Intestinal farnesoid X receptor signaling pro-
motes nonalcoholic fatty liver disease. J Clin Invest, 2015;125(1):386–402.
109. Kawamata, Y., Fujii, R., Hosoya, M. et  al. A G protein‐coupled receptor
responsive to bile acids. J Biol Chem, 2003;278(11):9435–40.
110. Chen, X., Lou, G., Meng, Z., and Huang, W. TGR5: a novel target for
weight maintenance and glucose metabolism. Exp Diabetes Res, 2011;
2011:853501.
111. Thomas, C., Gioiello, A., Noriega, L. et al. TGR5‐mediated bile acid sens-
ing controls glucose homeostasis. Cell Metab, 2009;10(3):167–77.
112. Jinnouchi, H., Sugiyama, S., Yoshida, A. et al. Liraglutide, a glucagon‐like
peptide‐1 analog, increased insulin sensitivity assessed by hyperinsuline-
mic‐euglycemic clamp examination in patients with uncontrolled type 2
diabetes mellitus. J Diabetes Res, 2015;2015:8.
113. Armstrong, M.J., Gaunt, P., Aithal, G.P. et al. Liraglutide safety and effi-
cacy in patients with non‐alcoholic steatohepatitis (LEAN): a multicentre,
double‐blind, randomised, placebo‐controlled phase 2 study. Lancet,
2016;387(10019):679–90.
114. Trabelsi, M‐S., Daoudi, M., Prawitt, J. et al. Farnesoid X receptor inhibits
glucagon‐like peptide‐1 production by enteroendocrine L cells. Nat
Commun, 2015;6:7629.
 α1‐Antitrypsin Deficiency
50 David A. Rudnick and David H. Perlmutter
Department of Pediatrics, Division of Gastroenterology, Hepatology, and Nutrition and Department of
Developmental Biology, Washington University School of Medicine in St. Louis, St. Louis Children’s
Hospital, St. Louis, MO, USA
INTRODUCTION
The classical form of α1‐antitrypsin (α1‐AT) deficiency,
homozygous for the mutant α1‐ATZ allele, is a relatively com-
mon disease. It affects approximately 1 in 1600 to 1 in 2000 live
births in most populations of Northern European ancestry [1, 2].
Although only a subgroup of deficient individuals develop liver
disease, it represents the most common metabolic cause of liver
disease in children [3] and can be associated with chronic liver
disease and hepatocellular carcinoma in adults [4]. This defi-
ciency also causes premature development of pulmonary
emphysema in adults.
The α1‐AT molecule is a single‐chain secretory glycoprotein
that inhibits destructive neutrophil proteases including elastase,
cathepsin G, and proteinase 3. It is often referred to as a hepatic
acute phase reactant in that plasma α1‐AT is predominantly
derived from the liver and plasma levels increase three‐ to five-
fold during the host response to tissue injury/inflammation. It is
the archetype of a family of structurally related circulating ser-
ine protease inhibitors termed SERPINs. In the deficient state,
there is an approximately 85% to 90% reduction in serum α1‐
AT concentrations. A single amino acid substitution results in a
mutant protein unable to traverse the secretory pathway. This
α1‐ATZ protein is retained in the endoplasmic reticulum (ER)
rather than secreted into the blood and body fluids.
The classical deficient state is unique as a genetic disease in
that it causes injury to one target organ, lung, by loss‐of‐func-
tion and injury to another target organ, liver, by what appears to
be a gain‐of‐function mechanism. Most data in the literature
indicate that emphysema results from a decreased number of
α1‐AT molecules within the lower respiratory tract, allowing
unregulated elastolytic attack on the lung’s connective tissue
matrix [5]. Oxidative inactivation of residual α1‐AT as a result
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
of cigarette smoking accelerates lung injury [5]. Moreover, the
elastase–antielastase theory for the pathogenesis of emphysema
is based on the concept that oxidative inactivation of α1‐AT as a
result of cigarette smoking plays a key role in the emphysema of
α1‐AT‐sufficient individuals, the vast majority of cases of
emphysema [5]. It is more difficult to explain the pathogenesis
of liver injury in this deficiency. Results of transgenic animal
experiments provide further evidence that the liver disease does
not result from a deficiency in antielastase activity [6, 7]. Most
data in the literature corroborate the concept that liver injury in
α1‐AT deficiency results from the hepatotoxicity of mutant­
α1‐ ATZ accumulation in hepatocytes.
Although it is a single gene defect, there is extraordinary
variation in the phenotypic expression of disease in the classical
form of α1‐AT deficiency. For instance, nationwide prospective
screening studies done by Sveger in Sweden showed that only
8% of an unbiased cohort developed clinically significant liver
disease through the fourth decade of life [1, 8]. These data indi-
cate that other genetic traits and/or environmental factors pre-
dispose a subgroup of PIZZ individuals, that is, those
homozygous for the α1‐ATZ mutant allele, to liver injury. There
is also variation in incidence and severity of lung injury among
α1‐AT‐deficient individuals. Environmental factors, such as
cigarette smoking, obviously play an important role [2, 5].
However, there are well‐documented examples of siblings and
other relatives of deficient individuals with severe emphysema
who have the same genotype, a history of heavy cigarette smok-
ing and only mild, subclinical pulmonary function abnormali-
ties even at advanced ages [9].
Historically, the diagnosis of α1‐AT deficiency was based on
the altered migration of the abnormal α1‐ATZ molecule in
serum specimens subjected to isoelectric focusing gel analysis.
Contemporary diagnostic strategies may now employ genomic
646 THE LIVER:  CLINICAL MANIFESTATIONS OF LIVER DISEASE
analyses for specific gene mutations. Treatment of α1‐AT defi-
ciency‐associated liver disease remains mostly supportive. Liver
replacement therapy has been used successfully for severe liver
injury. α1‐AT‐deficient patients with severe emphysema have
undergone lung transplantation. Intravenous administration of
purified plasma α1‐AT augmentation therapy was approved for
α1‐AT deficiency lung disease on the basis of biochemical effi-
cacy studies [10]. Recently, a randomized, placebo‐controlled
trial provided some initial, albeit limited, evidence that augmen-
tation therapy slows the progression of lung destruction in α1‐
AT deficient subjects [11].
CLINICAL MANIFESTATIONS OF
LIVER DISEASE
Liver involvement is often noticed in early infancy because of
persistent jaundice, elevated serum transaminases, and increased
conjugated bilirubin levels [12, 13]. Some infants are recog-
nized because of cholestasis characterized by pruritus and
hypercholesterolemia. The clinical picture resembles extrahe-
patic biliary atresia but liver histology shows paucity of intrahe-
patic bile ducts [14]. Alternatively, liver disease may first be
discovered in later childhood or early adolescence when the
affected individual presents with hepatosplenomegaly, ascites,
or esophageal variceal hemorrhage. In some but not all cases, a
history of unexplained prolonged obstructive jaundice during
the neonatal period is identified.
The incidence and natural history of liver disease in α1‐AT
deficiency was determined in a nationwide screening study of
newborns in Sweden by Sveger [1]. From 200 000 infants, 127
PIZZ newborns were identified and prospectively followed,
now through age 37–40 years [8]. Initial analysis of this cohort
identified 14 PIZZ infants with prolonged obstructive jaundice,
another 8 with other reasons for clinical suspicion of liver dis-
ease, and the others with no clinical evidence for liver disease
[1]. Five PIZZ subjects subsequently died in early childhood,
two with known cirrhosis and two others with incidentally‐
detected liver disease at autopsy [15]. At reassessment of this
cohort at age 37–40 years, no participating subjects showed evi-
dence of active liver disease [8]. However, abnormalities in
liver‐related laboratory and/or imaging studies were reported.
One issue not addressed by the Sveger study is whether any
PIZZ subjects have subclinical histologic abnormalities that
lead to progressive disease in later adulthood.
It is now evident that this deficiency causes liver disease with
initial onset in adults. In our retrospective analysis of US liver
transplantation databases, we found that 77% of α1‐AT‐defi-
cient subjects undergoing liver transplantation over the last 20
years were adults, with peak age of transplantation in those sub-
jects of 50–64 years [16]. Many are probably heterozygotes
with other potential causes of liver disease, and some may be
poorly substantiated diagnoses. Nevertheless, our analysis sug-
gests that severe liver disease requiring transplantation in sub-
jects with ZZ or SZ phenotypes is more frequent in adults than
children. This observation is consistent with the discovery that
autophagy plays a critical role in the pathobiology of this liver
disease and the understanding that a physiological decline in
autophagy function in middle age may play an important role in
age‐dependent degenerative diseases in general.
Earlier studies also reported incidentally identified signifi-
cant liver pathology at autopsy in some elderly patients with
α1‐AT deficiency without known liver disease. A Swedish study
showed a higher risk of cirrhosis and primary liver cancer in
such subjects than previously suspected [4]. Based on these con-
siderations, α1‐AT deficiency should be considered in the dif-
ferential diagnosis of any adult who presents with chronic
hepatitis, cirrhosis, portal hypertension, or hepatocellular carci-
noma of unknown origin. Taken together, the overwhelming
clinical experience with this disease indicates wide variation in
liver disease phenotype with many “protected” from or having
slowly progressing liver disease.
It still remains unclear whether liver injury results from the
heterozygous α1‐AT PIMZ (PI genotype heterozygous for the
wild‐type M and mutant Z alleles) state by itself. Liver biopsy
and transplant database studies have identified heterozygous
patients with severe liver disease without other apparent expla-
nation (reviewed in [3]). However, these studies are generally
unable to exclude environmental causes of liver disease. Indeed,
one cross‐sectional study of re‐examined α1‐AT deficiency
patients in a referral‐based Austrian university hospital sug-
gested that liver disease in heterozygotes was largely accounted
for by hepatitis B or C virus, autoimmune disease, or alcohol
abuse (reviewed in [3]). Other studies have subsequently
reported significant correlations between the MZ state and
development of, presentation with, or more rapid progression to
advanced liver disease and liver transplantation in subjects with
other chronic liver diseases [16, 17].
Liver disease has also been associated with other allelic vari-
ants of α1‐AT. For example, children with compound PISZ het-
erozygosity are affected by liver injury in a manner similar to
PIZZ children [1], and liver disease was reported in subjects
with α1‐AT deficiency type PIMmalton (PI genotype for the
Mmalton allele, Table 50.1) [18]. These observations are inter-
esting because the α1‐ATS molecule forms heteropolymers with
α1‐ATZ [19], and, like α1‐ATZ, the PIMmalton molecule
undergoes polymerization and ER retention [20]. Liver disease
has also been detected in single patients with other α1‐AT allelic
variants [3], but in those cases other causes of liver injury have
generally not been completely excluded.
Historically, the diagnosis was established by a serum α1‐AT
phenotype determination by isoelectric focusing or by agarose
electrophoresis at acid pH. Modern genomic methodologies
permit specific analyses for known liver disease‐associated α1‐
AT gene mutations, and analyses for gene duplications, dele-
tions, or previously unidentified mutations in the AT gene
coding region. Liver histology is characterized by periodic
acid–Schiff‐positive, diastase‐resistant globules in hepatocyte
ER. These globules are most prominent in periportal hepato-
cytes, and may also be seen in Kupffer cells and cells of biliary
ductular lineage (reviewed in [3]). There may be variable hepa-
tocellular necrosis, inflammatory cell infiltration, periportal
fibrosis, and/or cirrhosis. There is often evidence of bile duct
epithelial cell destruction, and, occasionally, paucity of intrahe-
patic bile ducts. There can also be an intense autophagic reac-
tion with nascent and degradative‐type autophagic vacuoles
detected by electron microscopy on liver biopsies [21].
 α1-ANTITRYPSIN DEFICIENCY
50:  647

Table 50.1  Deficiency variants of α1‐antitrypsin

Clinical disease
Variant Defect Site Liver Lung Cellular defect
Z Single base substitution M1 [Ala213] Glu342‐Lys + + IC accumulation
S Single base substitution Glu264‐Val − − IC accumulation
MHeerlen Single base substitution Pro369‐Leu − + IC accumulation
MProcida Single base substitution Leu41‐Pro − + IC accumulation
MMalton Single base deletion Phe52 + + IC accumulation
MDuarte Unknown Unknown +? + Unknown
MMineral Springs Single base substitution Gly57‐Glu − + No function; EC degradation?
Siiyama Single base substitution Ser53‐Phe − + IC accumulation
PDuarte Two base substitution Arg101‐His +? + Unknown
Asp256‐Val
PLowell Single base substitution Asp256‐Val − + IC degradation?
WBethesda Single base substitution Ala336‐Thre − + Accelerated catabolism?
ZWrexham Single base substitution Ser19‐Leu ? ? Unknown
F Single base substitution Arg223‐Cys − − Unknown
T Single base substitution Glu264‐Val − − Unknown
I Single base substitution Arg39‐Cys − − IC degradation
MPalermo Single base deletion Phe51 − − Unknown
MNichinan Single base deletion and Phe52 − − Unknown
single base substitution Gly148‐Arg
ZAusburg Single base substitution Glu342‐Lys − − Unknown
King’s Single base substitution His334‐Asp + IC accumulation
Mpisa Single base substitution Lys259‐Ile − +? IC accumulation
Etaurisano Single base substitution Lys368‐Glu − + IC accumulation
Yorzinuovi Single base substitution Pro391‐His +? − IC accumulation
PiSDonosti S variant defect + single base S variant site + Ser14‐Phe − +? IC polymerization, degradation; reduced secretion
substitution
PiTijarafe Single base substitution Ile50‐Asn − − IC polymerization, stabilization; reduced secretion
PiSevilla Single base substitution Ala58‐Asp − +? IC polymerization, stabilization; reduced secretion
PiCadiz Single base substitution Glu151‐Lys − +? None
PiTarragona Single base substitution Phe227‐Cys − +? IC polymerization, degradation; reduced secretion
 PiPuerto Real Single base substitution Thr249‐Ala − +? IC polymerization, stabilization; altered glycosylation
PiValencia Single base substitution Lys328‐Glu − − IC stabilization (reduced enzyme activity)

STRUCTURE, FUNCTION, most important determinant of functional specificity for each

AND PHYSIOLOGY
α1‐AT is encoded by a 12.2 kb gene containing seven exons and
six introns on human chromosome 14q31‐32.3 (reviewed in
[3]). The first three exons and a short 5’ segment of the fourth
exon encode variable, tissue‐specific 5’ untranslated regions
(UTRs) of the α1‐AT mRNA, which are generated by alternative
post‐transcriptional exon splicing [22, 23]. Alternate splicing of
the α1‐AT 5’ UTR results in variable inclusion or exclusion of
long upstream open reading frames (ORFs) whose presence or
absence alters the efficiency of α1‐AT translation [24] and could
potentially affect hepatic phenotypes in PIZZ subjects. Most of
the fourth exon and the remaining three exons encode the α1‐AT
protein sequence. The α1‐AT protein is a single‐chain, 55 kDa
polypeptide with 394 amino acids and 3 asparagine‐linked com-
plex carbohydrate side‐chains. Two major serum isoforms differ
in the configuration of the carbohydrate side‐chains.
α1‐AT is the archetype of the SERPIN family of protease
inhibitors [3]. Most SERPINs function as suicide inhibitors
of specific target proteases. Others are not inhibitory, but
rather form complexes with but do not inactivate their hor-
mone ligands. The reactive site “P1” residue of α1‐AT is the
SERPIN molecule. This concept was dramatically confirmed
by the discovery of α1‐AT Pittsburgh, a variant in which the
P1 methionine 358 residue of α1‐AT is replaced by arginine.
In this variant, α1‐AT functions as a thrombin inhibitor, and
severe bleeding diathesis results [25]. α1‐AT is an inhibitor
of serine proteases in general, but under physiologic condi-
tions its targets are probably only neutrophil elastase, cathep-
sin G, and proteinase 3, proteases released by activated
neutrophils because the kinetics of association with these
enzymes are dramatically more favorable than with other ser-
ine protease [26].
α1‐AT acts competitively by allowing target enzymes to bind
directly to its P1 Met residue‐containing reactive loop. The
resulting complex contains one molecule of each reactant. The
complex of α1‐AT and serine protease is a covalently stabilized
structure, resistant to dissociation by denaturing compounds.
The interaction between α1‐AT and serine protease is “suicidal”
in that the inhibitor is irreversibly modified and no longer able
to bind or inactivate enzyme. Studies also show that the irrevers-
ible trapping of the target enzyme is mediated by a profound
conformational change in α1‐AT, which Carrell and Lomas lik-
ened to a “mousetrap, with the active inhibitor circulating in the
648 THE LIVER:  MECHANISM OF DEFICIENCY
metastable stressed‐form and then springing into the stable,
relaxed form to lock the complex with its target protease” [27].
The functional activity of α1‐AT in vivo may be regulated by
several factors. For one, it may be rendered inactive as an elastase
inhibitor by active oxygen products, intermediates of activated
neutrophils and macrophages that can oxidize the reactive‐site
methionine of α1‐AT [5]. This effect is thought to constitute the
basis for increased susceptibility to emphysema in smokers,
whether deficient in α1‐AT or not. The α1‐AT molecule may also
be inactivated in vivo by proteases including collagenase and
pseudomonas elastase [26, 28]. Several studies indicate that α1‐
AT protects experimental animals from the lethal effects of tumor
necrosis factor (TNF) [29]. This protective effect is thought to be
due to inhibition of the synthesis and release of platelet‐activating
factor from neutrophils [30], presumably through the inhibition
of neutrophil‐derived proteases. α1‐AT also appears to have func-
tional activities that do not involve inhibition of neutrophil pro-
teases. The C‐terminal fragment of α1‐AT, which can be generated
during the formation of a complex with serine protease or during
proteolytic inactivation by thiol‐ or metalloproteases, is a potent
neutrophil chemoattractant [31].
The predominant site of synthesis of plasma α1‐AT is the
liver [32]. Tissue‐specific expression of α1‐AT in human
hepatoma cells is directed by structural elements recognized by
nuclear transcription factors including HNF‐1α and HNF‐1β,
C‐EBP, HNF‐4, and HNF‐3 (reviewed in [3]). Plasma concen-
trations of α1‐AT increase three‐ to fivefold during the host
response to inflammation and/or tissue injury [3]. The source of
this additional α1‐AT has always been considered the liver;
thus, α1‐AT is classified as a hepatic acute phase reactant.
Synthesis of α1‐AT in human hepatoma cells (HepG2, Hep3B)
is upregulated by interleukin 6 (IL‐6) but not interleukin 1
(IL‐1) or TNF [33]. Plasma concentrations also increase during
oral contraceptive therapy and pregnancy [34].
α1‐AT is expressed in monocytes and macrophages [35] and
mRNA has been isolated from multiple tissues in transgenic
mice [36, 37], but only in some cases have studies distinguished
whether such mRNA is in ubiquitous tissue macrophages or
other cell types. For instance, α1‐AT is synthesized in entero-
cytes and intestinal paneth cells, as determined by studies in
intestinal epithelial cell lines, ribonuclease protection assays of
human intestinal RNA, and in situ hybridization analysis in cry-
ostat sections of human intestinal mucosa [23, 38]. α1‐AT is
also synthesized by pulmonary epithelial cells, with such syn-
thesis less responsive to IL‐6 than to a related cytokine, oncos-
tatin M [39].
The half‐life of α1‐antitrypsin in plasma is approximately 5
days [40]. It is estimated that the daily production rate of α1‐AT
is 34 mg kg−1 body weight, with 33% of the intravascular pool
degraded daily. Several physiologic factors may affect the rate of
α1‐AT catabolism. First, desialylated α1‐AT is cleared from the
circulation in minutes [3], probably via hepatic asialoglycopro-
tein receptor‐mediated endocytosis. Second, α1‐AT in complex
with elastase or proteolytically modified is cleared more rapidly
than native α1‐AT [41]. Because its ligand specificity is similar to
that required for in vivo clearance of serpin‐enzyme complexes
(SECs), the SEC receptor may also be involved in the clearance
and catabolism of α1‐AT‐elastase and other serpin‐enzyme com-
plexes [42]. The low‐density protein receptor‐related protein
(LRP) can also mediate clearance and catabolism of α1‐AT‐
elastase complexes [43]. α1‐AT diffuses into most tissues and is
found in most body fluids [5].
DEFICIENCY VARIANTS
OF α1‐ANTITRYPSIN
Human α1‐AT variants are classified according to the protease
inhibitor (PI1,2) phenotype system as defined by agarose electro-
phoresis or isoelectric focusing of plasma in polyacrylamide at
acid pH [44]. The PI classification assigns a letter to variants,
according to migration of the major isoform, using alphabetic
order from anode to cathode, or from low to high isoelectric
point. The most common normal variant migrates to an interme-
diate isoelectric point, designated M, and according to PI geno-
type nomenclature individuals homozygous for this variant are
designated PIMM. The most common severe deficiency allelic
variant migrates to a high isoelectric point, designated Z, and
individuals homozygous for this variant are designated PIZZ.
Several null and dysfunctional variants, with absent or reduced
serum levels or activity, have been reported, some associated
with premature development of emphysema.
In addition to Z, other deficiency variants with reduced serum
α1‐AT concentrations have been reported (Table  50.1). Some
are not associated with clinical disease such as the S variant [45,
46]. Several are associated with premature lung disease
(Table 50.1). With respect to other AT variants associated with
liver disease, hepatocyte α1‐AT inclusions and liver disease
were reported in two persons with MMalton and one with
MDuarte [3, 18]. One person with the deficiency variant
Siiyama, with emphysema and hepatocyte inclusions but with-
out liver disease, was also reported [47]. Several novel defi-
ciency variants with a tendency to accumulate and polymerize
in cell culture models were recently discovered by gene sequenc-
ing of subjects with discrepancies between serum α1‐AT con-
centration and targeted variant genotyping for Z and S variants
[48]. Another recent study reported that the Z variant, which is
the most common deficiency variant and the variant most often
associated with α1‐AT deficiency liver disease, is able to heter-
opolymerize with S, Mmalton, and Mwurzburg, in cell culture
models [49].
MECHANISM OF DEFICIENCY
A point mutation results in the substitution of lysine for gluta-
mate 342 [27] and renders the mutant α1‐ATZ molecule
impaired in its capacity to transverse the secretory pathway.
Newly synthesized α1‐ATZ accumulates predominantly in the
ER, and perhaps other as yet undefined compartments that do
not stain with conventional markers of the ER [50], with rela-
tively limited proportions reaching the extracellular fluid [51].
This impairment is observed in liver cells, macrophages, trans-
fected cell lines, and iPS (induced pluripotent stem cell)‐derived
hepatocyte‐like cells (iHeps) [3, 50]. Site‐directed mutagenesis
studies have shown that the single amino acid substitution
 α1-ANTITRYPSIN DEFICIENCY
50:  649
(E342K) is sufficient to produce the defect in secretion [50, 52].
The mutant α1‐ATZ molecule is partially functionally active,
having about 50% to 80% of the elastase inhibitory capacity of
wild‐type α1‐ATM [53]. There is a modest increase in the rate
of in vivo clearance/catabolism of radiolabeled α1‐ATZ com-
pared with wild‐type α1‐ATM when infused into normal indi-
viduals, but this difference does not account for the decrease in
blood levels of α1‐AT in deficient individuals [41].
The point mutation also renders the α1‐ATZ molecule prone
to polymerization and aggregation. Carrell and Lomas have
demonstrated the presence of polymers and aggregates in liver
biopsy specimens and plasma of patients with homozygous
PIZZ α1‐antitrypsin deficiency [27]. Their studies suggested
that a “loop‐sheet insertion” mechanism was responsible for the
polymerization [54]. Similar polymers have been found in
the plasma of patients with the PISiiyama α1‐AT variant and the
PIMMalton α1‐AT variant [20, 55]. The mutation in α1‐AT
PISiiyama (Ser 53 to Phe) [47], and in α1‐ATPIMMalton (Phe
52 deletion) [18] affect residues that provide a ridge for the slid-
ing movement that opens the A sheet. Thus, these mutations
would be expected to interfere with the insertion of the reactive
center loop into the gap in the A sheet, and therefore leave the
gap in the A sheet available for spontaneous loop‐sheet polym-
erization. It is indeed interesting that hepatocellular α1‐AT
globules have been observed in a few patients with these two
variants. Recent observations suggest that the α1‐ATS variant
also undergoes loop‐sheet polymerization [19] and that this may
account for its retention in the ER, albeit a milder degree of
retention than that for α1‐ATZ [45]. Moreover, α1‐ATS can
apparently form heteropolymer with α1‐ATZ [19], providing a
potential explanation for liver disease in patients with the SZ
phenotype [56].
More recent studies by Huntington and colleagues have sug-
gested a different mechanism for polymerogenic and aggrega-
tion‐prone properties of ATZ that involves at least two different
domain‐swapping phenomena [57–59]. This mechanism seems
most consistent with recent characterization of a putative ATZ
monomer crystal structure [60].
By themselves, however, these data do not prove that the
polymerization of α1‐ATZ results in retention in the ER. In fact,
many polypeptides must assemble into oligomeric or polymeric
complexes to traverse the ER and reach their destination within
the cell, at the surface of the plasma membrane, or into the
extracellular fluid [61]. Evidence that polymerization results in
the retention of α1‐ATZ in the ER was originally attributed to
studies in which the fate of α1‐ATZ was examined after the
introduction of additional mutations into the molecule. For
instance, Kim et al. introduced a mutation into the α1‐AT mol-
ecule at amino acid 51, F51L [62]. This mutation is remote from
the Z mutation, E342K, but apparently impedes polymerization
and prevents insertion of a synthetic peptide into the gap in the
A sheet, implying that the mutation leads to closing of this gap.
The double‐mutated F51L α1‐ATZ molecule was less prone
to  polymerization and folded more efficiently in vitro than
α1‐ATZ. Moreover, the introduction of the F51L mutation par-
tially corrected the intracellular retention properties of α1‐ATZ
in microinjected Xenopus oocytes [63] and in yeast [64].
Nevertheless, several lines of evidence suggest that misfold-
ing is the primary defect responsible for impaired secretion/
intracellular accumulation of ATZ and, in this conceptualiza-
tion, polymerization/aggregation is a result of, rather than, the
primary defect itself. First, only 18% of the intracellular pool of
ATZ is in polymers at steady state in a mammalian cell line
model that faithfully recapitulates the intracellular accumula-
tion/fate of ATZ [65]. Second, a naturally occurring variant of
ATZ, which has the same E342K mutation as ATZ and also a
C‐terminal truncation, accumulates in the ER even though it
does not form polymers, suggesting that misfolding is sufficient
to lead to intracellular accumulation of ATZ [66]. Third, the
results of Sidhar et al. [63] and Kang et al. [64] do not exclude
the possibility that diminished secretion of ATZ is partially
­corrected by the second engineered mutation because this sec-
ond mutation also prevents the primary misfolding defect.
Furthermore, in a very interesting study a small molecule that
prevents polymerization in vitro does not correct the secretory
defect of ATZ in vivo but rather leads to enhanced degradation
[67]. Taken together, the data suggests that misfolding is the
primary defect and that polymerization is a time‐dependent
effect of misfolding and accumulation. This conceptualization
is also consistent with the domain‐swapping mechanism of
polymerization described by Huntington [57–59] and older
studies of folding kinetics by Yu et al. [68] in which polymeriza-
tion is viewed as a “kinetic” result of delayed folding of mono-
meric ATZ, and explains how some ATZ gets secreted. The
distinction between misfolding or polymerization as the pri-
mary inciting event for intracellular accumulation/impaired
secretion is important when considering potential therapeutic
approaches. If misfolding is the primary event a therapy that
prevents polymerization but not misfolding might fail to alter
the accumulation and/or the impaired secretion.
In either case, however, there is powerful evidence that the
tendency for α1‐ATZ to misfold, polymerize and aberrantly
accumulate within the secretory pathway is integral to its pro-
teotoxicity and disease‐causing potential. In studies of two fam-
ilies with autosomal dominantly inherited dementia, unique
neuronal inclusion bodies were shown to be associated with
polymers of a neuron‐specific member of the SERPIN family,
neuroserpin [69]. Moreover, the mutation in neuroserpin in one
family is homologous to the mutation in the α1‐AT Siiyama
allele that is associated with polymerization and inclusions in
the ER of liver cells.
In one of the most interesting studies on the mechanism for
ATZ accumulation in the ER, Nyfeler et al. provided evidence
for the lectin ER Golgi intermediate compartment 53 kDa pro-
tein (ERGIC‐53) as an export receptor for α1‐antitrypsin [70].
Most importantly, ERGIC‐53 failed to recognize mutant
α1‐ATZ. This study did not address whether polymerization
prevents α1‐ATZ from being presented to ERGIC‐53 or the
altered folding pathway of α1‐ATZ prevents an essential ligand
domain from being available to bind to ERGIC‐53.
PATHOGENESIS OF LIVER DISEASE
Although there are still many gaps in our understanding of how
the deficiency causes liver disease in some of the homozygotes,
we know that aberrant intracellular accumulation of α1‐ATZ is
650 THE LIVER:  PATHWAYS FOR INTRACELLULAR DEGRADATION OF MUTANT α1-ATZ
the unifying characteristic of what would be called a gain‐of‐
toxic‐function mechanism. Experimental results in transgenic
mice provide the most powerful evidence. Transgenic mice car-
rying the mutant human α1‐ATZ allele develop periodic acid–
Schiff‐positive, diastase‐resistant intrahepatic globules and liver
injury [6, 7]. Because there are normal levels of antielastases in
these animals, as directed by endogenous genes, the liver injury
cannot be attributed to a loss‐of‐function mechanism. The most
extensively characterized transgenic mouse model, the PiZ
mouse, generated with a human genomic fragment encompass-
ing all of the exons and introns of the human ATZ gene [6],
recapitulates the hepatic pathology of the human disease with
intrahepatocytic globules reflecting accumulation of misfolded
ATZ in the ER, slowly progressing fibrosis, low‐grade inflam-
mation, dysplasia, and increased prevalence of hepatocellular
carcinoma [71, 72]. In work with this mouse model over many
years we have found that marked progressive nodular regenera-
tion is the dominant hepatic pathological characteristic, strongly
resembling liver pathology in the human disease.
There is still relatively limited information about how pro-
teinopathy/intracellular α1‐ATZ accumulation leads to liver
injury. Although structural and functional alterations in mito-
chondria and caspase‐3 activation have been observed in liver
tissue from the PiZ mouse model and from α1‐antitrypsin defi-
ciency patients [73, 74], mitochondrial dysfunction probably
has mostly a cytostatic effect because apoptosis, necrosis, and
inflammation are not major pathological characteristics of the
liver in α1‐antitrypsin deficiency.
The mechanisms responsible for the hepatic fibrotic response
to proteinopathy are also not well understood. Genomic analysis
of a mouse model with hepatocyte‐specific inducible expression
of mutant α1‐ATZ, ideal for elucidating signal transduction
pathways that are activated by the proteinopathy, has identified
a relatively rich network of downstream targets of the TGFβ
pathway, including upregulation of connective tissue growth
factor [75], and this pathway is known to play a central role in
fibrosis [76]. We have also demonstrated that activation of the
NFκB signaling pathway is an important and specific effect of
cellular ATZ accumulation [77] and is designed to prevent fibro-
genesis by activating collagen‐degrading metalloproteases [78].
Many recent studies have also suggested that fibrosis results
from proteinopathies in several other tissues. Lung fibrosis has
been described in several rare diseases characterized by pro-
teinopathy in respiratory epithelial cells, including surfactant
protein C deficiency and Hermansky–Pudlak syndrome [79,
80]. Similarly myocardial fibrosis has been described for desmi-
nopathy that affects cardiomyocytes [81] and skeletal muscle
fibrosis in inclusion‐body myositis [82]. Interestingly, by
enhancing the degradation of misfolded proteins, autophagy has
been shown to mitigate cardiac fibrosis from desminopathy [81]
and skeletal muscle fibrosis from inclusion‐body myositis [83]
just as it does for hepatic fibrosis in the PiZ model of α1‐antit-
rypsin deficiency.
The liver of the PiZ mouse also shows glycogen depletion [84]
and defective ureagenesis [85]. The latter has been attributed to
downregulation of hepatocyte nuclear factor‐4α [85]. Because
these functional abnormalities are seen clinically in severe forms
of liver disease, this report provides additional evidence for the
validity of the PiZ mouse as a model of the human disease.
For many years it was conceptually difficult to reconcile the
gain‐of‐toxic function mechanism with the observations of
Sveger, showing that only a subset of α1‐AT‐deficient homozy-
gotes develop significant liver damage. In 1994, we published
a series of experiments investigating the hypothesis that a sub-
set of the PIZZ population is more susceptible to liver injury
because of one or more additional inherited traits or environ-
mental factors that exaggerate either the intracellular accumula-
tion of the mutant α1‐ATZ or the cellular pathophysiological
consequence of mutant α1‐ATZ accumulation [86]. Furthermore,
we hypothesized that these putative genetic/environmental
modifiers would affect intracellular degradation of α1‐ATZ or
adaptive signaling pathways that are activated by the intracel-
lular accumulation of α1‐ATZ. This hypothetical model was
validated by showing a lag in intracellular degradation of α1‐
ATZ in skin fibroblasts from human PIZZ homozygotes
already known to have severe liver disease (“susceptible
hosts”) as compared to fibroblasts from homozygotes who had
no liver disease (“protected hosts”). The fibroblasts had been
engineered to express α1‐ATZ using gene transfer by ampho-
tropic recombinant retroviral particles. More recently we
found similar results in iPS‐derived hepatocyte cells (iHeps)
[50]. The rate of degradation of α1‐ATZ was slower in iHeps
from susceptible hosts than in iHeps from protected hosts.
Over the years since the 1994 skin fibroblast experiments we
have characterized the putative pathways for intracellular
­degradation of α1‐ATZ and the putative adaptive signaling
pathways activated by intracellular accumulation of α1‐ATZ
(Figure 50.1).
PATHWAYS FOR INTRACELLULAR
DEGRADATION OF MUTANT α1‐ATZ
Early studies using yeast and mammalian cell lines showed that
the proteasomal pathway participates in intracellular disposal of
α1‐ATZ by a process that is now known as ER‐associated deg-
radation (ERAD) in which the substrate is extracted retrograde
from the ER to the cytoplasm [87, 88]. In fact, α1‐ATZ was one
of the first identified substrates of the ERAD pathway [88].
Autophagy was subsequently identified as a second major
pathway for disposal of misfolded α1‐ATZ [21]. Autophagy is
an intracellular catabolic pathway by which cells digest subcel-
lular structures and cytoplasm to generate amino acids as a sur-
vival mechanism (Figure  50.2). It is characterized by double
membrane vacuoles called autophagosomes, which fuse with
lysosomes for degradation of the internal constituents. Increased
numbers of autophagosomes were observed in human fibroblast
cell lines engineered for expression of mutant α1‐ATZ, in the
liver of PiZ transgenic mice and in liver biopsy specimens from
patients with ATD. Definitive evidence for the role of autophagy
in ATZ disposal was provided by genetic studies in which α1‐
ATZ disposal was delayed in autophagy‐deficient [Atg5‐null]
murine embryonic fibroblast cell lines [89] and Atg6‐null yeast
strains [90, 91].
The importance of the autophagic pathway in intracellular ATZ
degradation has been further validated by recent studies demon-
strating that autophagy enhancer drugs promote intracellular ATZ
 α1-ANTITRYPSIN DEFICIENCY
50:  651

Figure 50.1  Conceptual model for mechanisms of proteotoxicity in α1‐AT deficiency liver disease. Cellular factors that determine whether an
AT‐deficient individual is protected or susceptible to liver disease. In the susceptible host, there is greater accumulation of misfolded ATZ in the ER
because of subtle alternations in putative proteostasis network regulatory mechanisms. Here these proteostasis regulatory mechanisms are envi-
sioned as either ER degradation or signaling pathways that represent cellular protective responses. Reproduced with permission of Cold Spring
Harbor Laboratory Press from [122].

Figure 50.2  (a) Intracellular molecules and organelles destined for autophagic degradation (termed “cargo”) are surrounded by a double‐mem-
brane structure termed the “isolation membrane” (IM). The mechanisms regulating formation of the IM and its targeting to cargo remain incom-
pletely defined. The IM matures into the autophagosome vacuole, separating the cargo from other cellular compartments, and then fuses with the
lysosome to expose the cargo to degradative enzymes. Image adapted from [123] and available under the terms of the Creative Commons Attribution
License, CCBY 4.0.

disposal and attenuate hepatic fibrosis in the PiZ mouse model of
ATD in vivo [71]. One of the concepts originating from studies of
ATZ disposal in autophagy‐deficient yeast strains is that
autophagy becomes particularly important at higher levels of ATZ
expression (reviewed in [92]). These results taken together with
the structural constraints of the proteasome have led to the sup-
position that the proteasomal pathway degrades soluble mono-
meric forms of ATZ whereas autophagy is needed for soluble and
insoluble polymers. However, it is also possible that autophagy
plays a role in the disposal of soluble monomeric ATZ that accu-
mulates at levels of expression that exceed the capacity of the
proteasome. Another important result from the studies by Kruse
et al. in autophagy‐deficient yeast showed that a misfolded fibrin-
ogen variant associated with liver disease in a rare inherited form
of hypofibrinogenemia is degraded by autophagy in a manner
almost identical to that of misfolded ATZ [91].
652 THE LIVER:  HEPATIC CARCINOMA IN α1-ANTITRYPSIN DEFICIENCY
Pathways for intracellular disposal of ATZ other than the pro-
teasomal system and the canonical macroautophagy system are
highly likely. For example, a sortilin‐mediated pathway from
Golgi to lysosome has been described to participate in degrada-
tion of ATZ in yeast and mammalian cell line models [90, 93].
Another pathway for ATZ disposal which diverges from the
canonical autophagy system was recently identified in a power-
ful Caenorhabditis elegans model of ATD and found to be pre-
sent in a mammalian cell line model as well [94]. This pathway
is particularly interesting because it is suppressed by insulin
signaling and when upregulated by knocking down components
of the insulin signaling pathway it can completely mitigate ATZ
proteotoxicity.
SIGNALING PATHWAYS ACTIVATED BY
ACCUMULATION OF α1‐ATZ
IN THE ENDOPLASMIC RETICULUM
To determine which signaling pathways are activated when
α1‐ATZ accumulated in the ER, we developed cell line and
mouse model systems with inducible rather than constitutive
expression of α1‐ATZ because the latter would potentially per-
mit adaptations that could obscure the primary signaling
effects. A series of studies using these kinds of systems have
shown that the autophagic response and the nuclear factor κB
(NFκB) signaling pathway, but not the unfolded protein
response, are a­ ctivated when α1‐ATZ accumulates in the ER
[75, 77, 89].
Activation of NFκB appears to be a hallmark of α1‐ATZ
­accumulation [75, 77]. In recent studies in which the PiZ mouse
is mated to two different mouse models engineered for absence
of NFκB signaling we found more severe inflammation, fibrosis,
steatosis, dysplasia, and more hepatocytes with globules [78]
indicating that NFκB signaling is intended to protect the liver
from the effects of α1‐ATZ accumulation. Interestingly the effect
of NFκB signaling pathway under these circumstances is through
a small number of downstream target genes: Egr‐1, a transcrip-
tion factor that is essential for liver regeneration [95]; RGS16, an
inhibitor of G‐protein signaling that has been implicated in acti-
vation of autophagy [75]; and matrix metalloproteases MMP7
and MMP12. Each of these targets appear to be mediating parts of
the hepatic response to α1‐ATZ accumulation: decreased prolif-
eration of hepatocytes with massive ATZ accumulation, known as
globule‐containing hepatocytes, due to Egr1‐downregulation;
increased autophagy due to RGS16 upregulation; and counteract-
ing of fibrogenesis by MMP7 and MMP12 upregulation.
The hepatic transcriptomic analysis of the Z mouse also
shows changes in gene expression indicative of TGFβ signaling
and this is consistent with the fibrotic response that represents
the dominant pathological characteristics of the liver in α1‐­
antitrypsin deficiency [75]. Furthermore we have recently found
that accumulation of the α1‐ATZ variant in respiratory epithelial
cells elicits fibrosis in the lungs with evidence for fibrosis in the
lungs of α1‐antitrypsin deficiency patients with very severe
COPD (chronic obstructive pulmonary disease) [96]. The mech-
anism by which accumulation of misfolded proteins in the ER
elicits TGFβ signaling has not been studied.
c‐Jun N‐terminal kinase activation was recently demonstrated
in the liver of the PiZ mouse model, as well as in liver speci-
mens from patients with α1‐antitrypsin deficiency [97].
Interestingly this signaling effect leads to increased expression
of α1‐ATZ and therefore likely plays a role in amplifying the
pathobiology of misfolded protein accumulation.
HEPATIC CARCINOMA
IN α1‐ANTITRYPSIN DEFICIENCY
Although increased susceptibility to hepatic cancer was shown
years ago [4], there have been only a few studies of potential
mechanisms for carcinogenesis. Theorizing that hepatocellular
hyperproliferation was likely to be involved, Rudnick et  al.
investigated the proliferation of liver cells by BrdU labeling in
the PiZ mouse model [98]. In these studies hepatocellular pro-
liferation was increased around sevenfold in the PiZ mouse
compared to a wild‐type control and this degree of proliferation
appeared to reflect the slowly progressing chronic nature of α1‐
antitrypsin deficiency liver disease. Most importantly, by using
double immunohistochemical staining it was shown that divid-
ing hepatocytes were almost exclusively the ones that lacked
intracellular α1‐ATZ inclusions, called the globule‐devoid
hepatocytes. Furthermore it was shown that hyperproliferation
of globule‐devoid hepatocytes was driven by the number of
adjacent globule‐containing hepatocytes. This last conclusion
was based on the observation that the number of globule‐­
containing hepatocytes was markedly increased in male PiZ
mice or in testosterone‐treated female PiZ mice and this corre-
lated directly with the degree of hyperproliferation of globule‐
devoid hepatocytes.
Taken together these observations led to a theory that hepat-
ocyte hyperproliferation was elicited by crosstalk between
globule‐containing and globule‐devoid hepatocytes ([99] and
Figure 50.3). The globule‐devoid hepatocytes were viewed as
younger cells capable of responding to trans‐acting regenera-
tive signals derived from the globule‐containing hepatocytes.
The globule‐containing hepatocytes were viewed as having
greater proteotoxic α1‐ATZ accumulation and unable to
respond to the existing regenerative signals because of the pro-
teotoxic effect on cell proliferation. Thus, the globule‐contain-
ing hepatocytes are sick but not dead and stimulate the
regeneration of the globule‐devoid hepatocytes which have a
selective proliferative advantage. Interestingly, the replicative
defect in the globule‐containing hepatocytes was shown to be
relative because these cells could proliferate as well as globule‐
devoid hepatocytes when the regenerative stimulus was par-
ticularly powerful, such as in PiZ mice that survived
experimental partial hepatectomy [98]. The nature of the differ-
ences between the globule‐containing and globule‐devoid cells
is not well elucidated. A study by Linblad et al. has suggested
that the globule‐devoid hepatocytes have lesser accumulation
of ATZ [74] and this would be consistent with younger cells
that have had less time to accumulate α1‐ATZ. It is also possi-
ble that globule‐devoid hepatocytes are derived from globule‐
containing hepatocytes as they increase capacity for degradation
of α1‐ATZ. Several observations militate against this latter
 α1-ANTITRYPSIN DEFICIENCY
50:  653

Figure 50.3  Hypothetical model for hepatocarcinogenesis in ATD. Globule‐containing hepatocytes (pale pink) tend to be periportal. They are
“sick but not dead” and generate chronic regenerative signals which can only be received effectively in “trans” by globule‐devoid hepatocytes (deep
pink). The globule‐devoid hepatocytes tend to be in the centrilobular regions. When regenerative signals are received by globule‐devoid hepatocytes
by this crosstalk, it drives mitosis and ultimately carcinogenesis (dark red) in the globule‐devoid regions. Reproduced with permission of John
Wiley & Sons from [99].

possibility. The number of globule‐containing hepatocytes
decrease with age [98], and Ding et al. [100] showed that trans-
planted hepatocytes have a selective proliferative advantage
that also depends on the number of adjacent globule‐containing
hepatocytes in that it was much more evident in male PiZ mice
that had significantly more globule‐containing hepatocytes
than female PiZ mice. This is associated with enhanced apop-
tosis of the host hepatocytes, hepatic repopulation with donor
hepatocytes, and resolution of the liver fibrosis that occurs in
untreated PiZ mice [100].
It is interesting to note that hepatocellular carcinoma develops
with aging in male PiZ mice [72] and males with α1‐antitrypsin
deficiency were also disproportionately affected by hepatic
­cancer in the autopsy studies of Eriksson et al. [4]. Moreover,
in cases of hepatocellular carcinoma associated with α1‐anti-
trypsin deficiency one observes a staining pattern in which the
carcinoma is negative for inclusions but surrounded by adjacent
liver cells that are positive for inclusions which is entirely con-
sistent with the carcinogenesis theory proposed by Rudnick and
Perlmutter [99].
654 THE LIVER:  TREATMENT
MODIFIERS OF THE HEPATIC
PHENOTYPYE OF α1‐ANTITRYPSIN
DEFICIENCY
Studies designed to identify genetic and environmental modifiers
of the liver disease phenotype in human populations have begun
to appear in the literature in recent years. In one interesting study,
a single nucleotide polymorphism (SNP) in the MAN1B1 gene
was found to be statistically over‐represented in a series of infants
with end‐stage liver disease [101]. The variant was shown to
reduce intracellular levels of the mannosidase [102]. Recent
experiments have shown that Man1B1 is actually localized to the
Golgi but it plays a role in regulation of protein secretion as a part
of the protein quality control network which is recently recog-
nized to be localized in the Golgi [103]. Furthermore those
­experiments have provided a basis for how reduced levels of
Man1B1 could theoretically lead to greater intracellular
α1‐ATZ ­accumulation [103]. These results for the Man1B1 vari-
ant would appear to validate our hypothesis that intracellular deg-
radation pathways are targets of liver disease modifiers but further
population studies of this variant would be reassuring [104, 105].
A SNP in the upstream flanking region of the α1‐antitrypsin
gene has also been implicated in susceptibility to liver disease
[106]. However, the nature of that variant could not be recon-
ciled with how it might affect liver disease susceptibility and
its  statistical association with variation in the liver disease
­phenotype was dependent on a questionable classification of
population subgroups.
Our hypothesis for variation in liver disease susceptibility
also identifies signaling pathways that could increase or
decrease α1‐ATZ proteotoxicity as potential targets for disease
modifiers. As of yet we have not encountered an example of this
potential scenario, but we predict that further studies of iHeps
from patients with different forms of α1‐antitrypsin deficiency
liver disease, in terms of age of onset and type of hepatic pathol-
ogy, will identify such a mechanism in the near future.
TREATMENT
The most important principle in the treatment of α1‐AT
­deficiency is avoidance of cigarette smoking. Cigarette smoking
markedly accelerates the destructive lung disease that is associ-
ated with α1‐AT deficiency, reduces the quality of life, and
­significantly shortens the longevity of these individuals [107].
There is no specific therapy for α1‐AT deficiency‐associated
liver disease. Therefore, clinical care largely involves support-
ive management of symptoms due to liver dysfunction and for
the prevention of complications. Progressive liver dysfunction
in α1‐AT‐deficient patients has been treated by orthotopic liver
transplantation, with 5‐year survival rates approaching 90% in
children and 80% in adults at 1 year and 80% at 5 years [108].
Several novel strategies for treatment of α1‐antitrypsin defi-
ciency liver disease that would obviate the need for organ trans-
plantation and chronic immunosuppression are currently under
investigation and at various stages of development. One of the
relatively newer strategies targets intracellular degradation
­pathways using autophagy enhancer drugs.
Autophagy was considered an excellent target because it is
specifically activated when α1‐ATZ accumulates in cells and it
also plays a critical role in intracellular disposal of ATZ. At the
time when this approach was first investigated several drugs
which could enhance autophagic degradation of other misfolded
proteins, such as mutant polyQ proteins that cause Huntington
disease, were being described [109]. It had also become apparent
that α1‐antitrypsin deficiency liver disease could more frequently
have its onset at 50–65 years of age [16] coincident with the
decline in autophagy function that is believed to trigger other age‐
dependent degenerative diseases associated with misfolded
­proteins. Hidvegi et al. first investigated the drug carbamazepine
(CBZ), which is known for its widespread use in humans as an
anticonvulsant and mood stabilizer and found that it enhanced
autophagic degradation of α1‐ATZ in mammalian cell line mod-
els [71]. Moreover, administration of this drug by oral gavage to
the PiZ mouse model over a three‐week period significantly
reduced hepatic α1‐ATZ load and hepatic fibrosis in vivo. Because
CBZ is already FDA‐approved it could be moved immediately
into a phase 2/3 clinical trial for treatment of severe liver disease
due to ATD. Several other drugs with autophagy enhancer proper-
ties have been identified by high‐throughput screening of drug
libraries and these are currently under investigation [92].
A number of relatively new gene therapy strategies are being
investigated. One of these involves new methods for silencing
gene expression using vectors that are also capable of encoding
wild‐type α1‐antitrypsin to address both gain‐ and loss‐of
function sequelae of α1‐antitrypsin deficiency, respectively
­
[110, 111]. In one approach, Li et al. utilized adeno‐associated
virus harboring short‐hairpin RNA to knockdown endogenous
α1‐ATZ expression together with a codon‐optimized wild‐type
α1‐antitrypsin transgene cassette [110]. In another approach,
Mueller et  al. utilized an adeno‐associated virus harboring
microRNA to silence endogenous α1‐ATZ gene expression
together with a microRNA‐resistant wild‐type α1‐antitrypsin
gene [111]. In each case hepatic α1‐ATZ load was reduced and
levels of human α1‐antitrypsin increased in the serum of a trans-
genic mouse model. However, the effect on liver fibrosis by this
strategy was not as compelling and hence further studies are
required to test whether more potent and widespread silencing
would be more effective. Another study using antisense oligo-
nucleotides by systemic administration to silence α1‐ATZ gene
expression had a more impressive effect in reducing hepatic
fibrosis in the PiZ mouse model system [112].
Another potential gene therapy approach being investigated
is the transfer of genes that activate autophagy and therein
reduce α1‐ATZ accumulation and proteotoxicity. Pastore et al
have championed this approach using TFEB, a master transcrip-
tional activator of the autophagolysosomal system [113]. Using
helper‐dependent adenovirus for systemic delivery of TFEB and
targeting of its expression to liver, this approach significantly
reduced hepatic α1‐ATZ load and liver fibrosis in the PiZ mouse
model. In vitro studies also validated that TFEB reduces cellular
α1‐ATZ levels in an autophagy‐dependent manner [113].
Although it will not address the loss‐of‐function mechanisms
associated with α1‐antitrypsin deficiency lung disease, gene
therapy with TFEB or drugs that target TFEB activation
­constitute exciting potential therapeutic strategies for liver dis-
ease associated with α1‐antitrypsin deficiency.
 α1-ANTITRYPSIN DEFICIENCY
50:  655
Ultimately, genomic editing will be considered to definitively
correct the genetic defect that causes ATD. The most recent
development in this area, CRISPR/Cas‐9‐mediated genome
editing, has been used in the PiZ mouse model in vivo, showing
reduced α1‐ATZ in the liver and low levels of wild‐type α1AT in
the liver [114] and in serum [115].
Several research groups are exploring a “structure‐based”
screening strategy that aims to generate peptides to prevent
polymerization of the mutant α1‐ATZ with the hypothesis that
this would facilitate secretion. A small molecule compound
designed against a lateral hydrophobic cavity in α1‐ATZ pre-
vented its polymerization, however, further experiments in a cell
line model revealed that this compound enhanced intracellular
degradation only, with minimal effect on secretion [67]. These
results provide further evidence that it is the misfolding of
α1‐ATZ, independent of its tendency to polymerize, which is
primarily responsible for impaired secretion. Another small
molecule based on a peptide that targets the reactive center loop
of α1‐antitrypsin has been designed and introduced in cell line
model systems with evidence for improved secretion of α1‐ATZ
[116]. However, the efficacy of this peptide in an animal model
system for either increasing secretion or reducing liver damage
in vivo remains to be tested. It also remains possible that this
type of peptide binding changes the conformation of the mutant
protein in such a way that both misfolding and polymerization
are reduced independently.
Chemical chaperones that can non‐selectively facilitate fold-
ing of diverse misfolded proteins have also been investigated as
a potential therapeutic option for α1‐antitrypsin deficiency liver
disease. Glycerol and 4‐phenylbutyric acid (PBA) were found to
mediate a robust enhancement in the secretion of α1‐ATZ in a
mammalian cell line model and its oral administration in PiZ
mice increased blood levels of human α1‐antitrypsin reaching
20–50% of the levels present in PiM mice and normal humans
[117]. However, a pilot clinical trial involving 10 patients with
α1‐antitrypsin deficiency‐associated liver disease, failed to
reveal any significant increase in serum levels of α1‐antitrypsin
after 14 days of treatment with PBA [118]. It is not clear why the
drug lacked effect but the large doses required are known to be
quite challenging to tolerate and so it may be worthwhile to test
in the future if newer, more tolerable formulations are developed.
Recently, suberoylanilide hydroxamic acid (SAHA), another
drug which has many pharmacological similarities to PBA, has
been found to enhance secretion of α1‐ATZ in cell line models of
α1‐antitrypsin deficiency [119]. However, SAHA has not yet
been tested in animal models. Moreover, detailed studies are
needed to delineate whether this effect is due to increased syn-
thesis of α1‐ATZ or due to its ability to reduce α1‐ATZ accumu-
lation in cells or both. If increased secretion of α1‐ATZ is because
of, or even associated with, increased synthesis, the treatment
could produce more rather than less cellular proteotoxicity.
Hepatocyte transplantation therapy has also been investigated
as a potential treatment for α1‐antitrypsin deficiency. It has been
tested in the past as a treatment for several metabolic liver dis-
eases [120]. Compared to orthotopic liver transplantation it has
the advantage of being a minimally invasive procedure with lit-
tle known morbidity, and is considerably less expensive than
protein replacement therapy or liver transplantation. Importantly,
recent studies have revealed that wild‐type donor hepatocytes
can repopulate almost the entire liver of the PiZ mouse model
[100]. In the PIZ mouse model, the donor cells replaced both
globule‐containing and globule‐devoid cells, indicating that
both types of affected hepatocytes have impaired proliferative
capacity compared to wild‐type hepatocytes. Because the trans-
planted hepatocytes have a selective proliferative advantage
over α1‐ATZ‐containing endogenous hepatocytes and can sub-
stitute for the latter in a diseased liver, this option of therapy
may be considered for α1‐antitrypsin deficiency lung and liver
disease.
Another exciting therapeutic strategy in which genomic edit-
ing is combined with hepatocyte transplantation has been tested
in a transgenic mouse model of α1‐antitrypsin deficiency.
Studies by Yusa et  al. have shown that the mutation in the
α1‐antitrypsin gene could be corrected in human induced pluri-
potent stem (iPS) cells derived from a α1‐antitrypsin deficiency
patient using a combination of zinc‐finger nucleases and trans-
poson techniques [121]. Importantly, the corrected iPS cell lines
could then be engrafted into the liver of the transgenic mouse
model system and, based on the observations of Ding et  al.
[100], the corrected cells should expand significantly because
they will have a selective proliferative advantage. This strategy,
if it proves successful in further preclinical models, has the
potential to address both the loss‐ and gain‐of‐function mecha-
nisms of organ damage and the advantage of personalized treat-
ment options without any need for immunosuppression.
REFERENCES
  1. Sveger, T. Liver disease in alpha1‐antitrypsin deficiency detected by screen-
ing of 200,000 infants. N Engl J Med, 1976;294(24):1316–21.
  2. Silverman, E.K. and Sandhaus, R.A. Clinical practice. Alpha1‐antitrypsin
deficiency. N Engl J Med, 2009;360(26):2749–57.
  3. Teckman, J.H., Qu, D., and Perlmutter, D.H. Molecular pathogenesis of liver
disease in alpha1‐antitrypsin deficiency. Hepatology, 1996;24(6):1504–16.
  4. Eriksson, S., Carlson, J., and Velez, R. Risk of cirrhosis and primary liver
cancer in alpha 1‐antitrypsin deficiency. N Engl J Med, 1986;314(12):736–9.
  5. Crystal, R.G. Alpha 1‐antitrypsin deficiency, emphysema, and liver disease.
Genetic basis and strategies for therapy. J Clin Invest, 1990;85(5):1343–52.
  6. Carlson, J.A., Rogers, B.B., Sifers, R.N. et al. Accumulation of PiZ alpha
1‐antitrypsin causes liver damage in transgenic mice. J Clin Invest, 1989;
83(4):1183–90.
  7. Dycaico, M.J., Grant, S.G., Felts, K. et al. Neonatal hepatitis induced by alpha
1‐antitrypsin: a transgenic mouse model. Science, 1988;242(4884):1409–12.
  8. Mostafavi, B., Diaz, S., Tanash, H.A., and Piitulainen, E. Liver function in
alpha‐1‐antitrypsin deficient individuals at 37 to 40 years of age. Medicine
(Baltimore), 2017;96(12):e6180.
  9. Silverman, E.K., Province, M.A., Rao, D.C., Pierce, J.A., and Campbell, E.J.
A family study of the variability of pulmonary function in alpha 1‐antitrypsin
deficiency. Quantitative phenotypes. Am Rev Respir Dis, 1990;142(5):
1015–21.
10. Crystal, R.G. Augmentation treatment for alpha1 antitrypsin deficiency.
Lancet, 2015;386(9991):318–20.
11. McElvaney, N.G., Burdon, J., Holmes, M. et  al. Long‐term efficacy and
safety of alpha1 proteinase inhibitor treatment for emphysema caused by
severe alpha1 antitrypsin deficiency: an open‐label extension trial (RAPID‐
OLE). Lancet Respir Med, 2017;5(1):51–60.
12. Ibarguen, E., Gross, C.R., Savik, S.K., and Sharp, H.L. Liver disease in
alpha‐1‐antitrypsin deficiency: prognostic indicators. J Pediatr, 1990;
117(6):864–70.
13. Sharp, H.L., Bridges, R.A., Krivit, W., and Freier, E.F. Cirrhosis associated
with alpha‐1‐antitrypsin deficiency: a previously unrecognized inherited dis-
order. J Lab Clin Med, 1969;73(6):934–9.
656 THE LIVER:  REFERENCES
14. Hadchouel, M. and Gautier, M. Histopathologic study of the liver in the
early  cholestatic phase of alpha‐1‐antitrypsin deficiency. J Pediatr, 1976;
89(2):211–5.
15. Sveger, T. The natural history of liver disease in alpha 1‐antitrypsin deficient
children. Acta Paediatr Scand, 1988;77(6):847–51.
16. Chu, A.S., Chopra, K.B., and Perlmutter, D.H. Is severe progressive liver
disease caused by alpha‐1‐antitrypsin deficiency more common in children
or adults? Liver Transpl, 2016;22(7):886–94.
17. Schaefer, B., Mandorfer, M., Viveiros, A. et  al. Heterozygosity for the
alpha‐1‐antitrypsin Z allele in cirrhosis is associated with more advanced
disease. Liver Transpl, 2018;24(6):744–51.
18. Curiel, D.T., Holmes, M.D., Okayama, H. et al. Molecular basis of the liver
and lung disease associated with the alpha 1‐antitrypsin deficiency allele
Mmalton. J Biol Chem, 1989;264(23):13938–45.
19. Mahadeva, R., Chang, W.S., Dafforn, T.R. et al. Heteropolymerization of S,
I, and Z alpha1‐antitrypsin and liver cirrhosis. J Clin Invest, 1999;103(7):
999–1006.
20. Lomas, D.A., Elliott, P.R., Sidhar, S.K. et al. alpha 1‐Antitrypsin Mmalton
(Phe52‐deleted) forms loop‐sheet polymers in vivo. Evidence for the C sheet
mechanism of polymerization. J Biol Chem, 1995;270(28):16864–70.
21. Teckman, J.H. and Perlmutter, D.H. Retention of mutant alpha(1)‐antitrypsin
Z in endoplasmic reticulum is associated with an autophagic response. Am J
Physiol Gastrointest Liver Physiol, 2000;279(5):G961–74.
22. Perlino, E., Cortese, R., and Ciliberto, G. The human alpha 1‐antitrypsin
gene is transcribed from two different promoters in macrophages and hepat-
ocytes. EMBO J, 1987;6(9):2767–71.
23. Hafeez, W., Ciliberto, G., and Perlmutter, D.H. Constitutive and modulated
expression of the human alpha 1 antitrypsin gene. Different transcriptional
initiation sites used in three different cell types. J Clin Invest, 1992;89(4):
1214–22.
24. Corley, M., Solem, A., Phillips, G. et al. An RNA structure‐mediated, post-
transcriptional model of human alpha‐1‐antitrypsin expression. Proc Natl
Acad Sci USA, 2017;114(47):E10244–53.
25. Owen, M.C., Brennan, S.O., Lewis, J.H., and Carrell, R.W. Mutation of anti-
trypsin to antithrombin. alpha 1‐antitrypsin Pittsburgh (358 Met leads to
Arg), a fatal bleeding disorder. N Engl J Med, 1983;309(12):694–8.
26. Mast, A.E., Enghild, J.J., Nagase, H., Suzuki, K., Pizzo, S.V., ND Salvesen,
G. Kinetics and physiologic relevance of the inactivation of alpha 1‐protein-
ase inhibitor, alpha 1‐antichymotrypsin, and antithrombin III by matrix met-
alloproteinases‐1 (tissue collagenase), ‐2 (72‐kDa gelatinase/type IV
collagenase), and ‐3 (stromelysin). J Biol Chem, 1991;266(24):15810–6.
27. Carrell, R.W. AND Lomas, D.A. Conformational disease. Lancet, 1997;
350(9071):134–8.
28. Vissers, M.C., George, P.M., Bathurst, I.C., Brennan, S.O., and Winterbourn,
C.C. Cleavage and inactivation of alpha 1‐antitrypsin by metalloproteinases
released from neutrophils. J Clin Invest, 1988;82(2):706–11.
29. Van Molle, W., Libert, C., Fiers, W., and Brouckaert, P. Alpha 1‐acid glyco-
protein and alpha 1‐antitrypsin inhibit TNF‐induced but not anti‐Fas‐induced
apoptosis of hepatocytes in mice. J Immunol, 1997;159(7):3555–64.
30. Camussi, G., Tetta, C., Bussolino, F., and Baglioni, C. Synthesis and release
of platelet‐activating factor is inhibited by plasma alpha 1‐proteinase inhibi-
tor or alpha 1‐antichymotrypsin and is stimulated by proteinases. J Exp Med,
1988;168(4):1293–306.
31. Joslin, G., Griffin, G.L., August, A.M. et  al. The serpin‐enzyme complex
(SEC) receptor mediates the neutrophil chemotactic effect of alpha‐1 antit-
rypsin‐elastase complexes and amyloid‐beta peptide. J Clin Invest,
1992;90(3):1150–4.
32. Hood, J.M., Koep, L.J., Peters, R.L. et al. Liver transplantation for advanced
liver disease with alpha‐1‐antitrypsin deficiency. N Engl J Med, 1980;
302(5):272–5.
33. Perlmutter, D.H., May, L.T., and Sehgal, P.B. Interferon beta 2/interleukin 6
modulates synthesis of alpha 1‐antitrypsin in human mononuclear phago-
cytes and in human hepatoma cells. J Clin Invest, 1989;84(1):138–44.
34. Laurell, C.B. and Rannevik, G. A comparison of plasma protein changes
induced by danazol, pregnancy, and estrogens. J Clin Endocrinol Metab,
1979;49(5):719–25.
35. Perlmutter, D.H., Cole, F.S., Kilbridge, P., Rossing, T.H., and Colten, H.R.
Expression of the alpha 1‐proteinase inhibitor gene in human monocytes and
macrophages. Proc Natl Acad Sci USA, 1985;82(3):795–9.
36. Kelsey, G.D., Povey, S., Bygrave, A.E., and Lovell‐Badge, R.H. Species‐ and
tissue‐specific expression of human alpha 1‐antitrypsin in transgenic mice.
Genes Dev, 1987;1(2):161–71.
37. Carlson, J.A., Rogers, B.B., Sifers, R.N., Hawkins, H.K., Finegold, M.J., and
Woo, S.L. Multiple tissues express alpha 1‐antitrypsin in transgenic mice
and man. J Clin Invest, 1988;82(1):26–36.
38. Molmenti, E.P., Perlmutter, D.H., and Rubin, D.C. Cell‐specific expression
of alpha 1‐antitrypsin in human intestinal epithelium. J Clin Invest,
1993;92(4):2022–34.
39. Cichy, J., Potempa, J., and Travis, J. Biosynthesis of alpha1‐proteinase inhib-
itor by human lung‐derived epithelial cells. J Biol Chem, 1997;272(13):
8250–5.
40. Laurell, C.B., Nosslin, B., and Jeppsson, J.O. Catabolic rate of alpha1‐antit-
rypsin of Pi type M and Z in man. Clin Sci Mol Med, 1977;52(5):457–61.
41. Mast, A.E., Enghild, J.J., Pizzo, S.V., and Salvesen, G. Analysis of the plasma
elimination kinetics and conformational stabilities of native, proteinase‐com-
plexed, and reactive site cleaved serpins: comparison of alpha 1‐proteinase
inhibitor, alpha 1‐antichymotrypsin, antithrombin II.I., alpha 2‐antiplasmin,
angiotensinogen, and ovalbumin. Biochemistry, 1991;30(6):1723–30.
42. Joslin, G., Wittwer, A., Adams, S., Tollefsen, D.M., August, A., and
Perlmutter, D.H. Cross‐competition for binding of alpha 1‐antitrypsin (alpha
1 AT)‐elastase complexes to the serpin‐enzyme complex receptor by other
serpin‐enzyme complexes and by proteolytically modified alpha 1 A.T. J
Biol Chem, 1993;268(3):1886–93.
43. Kounnas, M.Z., Church, F.C., Argraves, W.S., and Strickland, D.K. Cellular
internalization and degradation of antithrombin III‐thrombin, heparin cofac-
tor II‐thrombin, and alpha 1‐antitrypsin‐trypsin complexes is mediated by
the low density lipoprotein receptor‐related protein. J Biol Chem, 1996;
271(11):6523–9.
44. Pierce, J.A. and Eradio, B.G. Improved identification of antitrypsin pheno-
types through isoelectric focusing with dithioerythritol. J Lab Clin Med,
1979;94(6):826–31.
45. Teckman, J.H. and Perlmutter, D.H. The endoplasmic reticulum degradation
pathway for mutant secretory proteins alpha1‐antitrypsin Z and S is distinct
from that for an unassembled membrane protein. J Biol Chem, 1996;
271(22):13215–20.
46. Long, G.L., Chandra, T., Woo, S.L., Davie, E.W., and Kurachi, K. Complete
sequence of the cDNA for human alpha 1‐antitrypsin and the gene for the S
variant. Biochemistry, 1984;23(21):4828–37.
47. Seyama, K., Nukiwa, T., Takabe, K., Takahashi, H., Miyake, K., and Kira, S.
Siiyama (serine 53 (TCC) to phenylalanine 53 (TTC)). A new alpha 1‐antit-
rypsin‐deficient variant with mutation on a predicted conserved residue of
the serpin backbone. J Biol Chem, 1991;266(19):12627–32.
48. Matamala, N., Lara, B., Gomez‐Mariano, G. et al. Characterization of novel
missense variants of SERPINA1 gene causing alpha‐1 antitrypsin deficiency.
Am J Respir Cell Mol Biol, 2018;58(6):706–16.
49. Laffranchi, M., Berardelli, R., Ronzoni, R., Lomas, D.A., and Fra, A.
Heteropolymerization of alpha‐1‐antitrypsin mutants in cell models mimick-
ing heterozygosity. Hum Mol Genet, 2018;27(10):1785–93.
50. Tafaleng, E.N., Chakraborty, S., Han, B. et al. Induced pluripotent stem cells
model personalized variations in liver disease resulting from alpha1‐antit-
rypsin deficiency. Hepatology, 2015;62(1):147–57.
51. Perlmutter, D.H., Kay, R.M., Cole, F.S., Rossing, T.H., Van Thiel, D., and
Colten, H.R. The cellular defect in alpha 1‐proteinase inhibitor (alpha 1‐PI)
deficiency is expressed in human monocytes and in Xenopus oocytes
injected with human liver mRNA. Proc Natl Acad Sci USA, 1985;82(20):
6918–21.
52. McCracken, A.A., Kruse, K.B., and Brown, J.L. Molecular basis for defec-
tive secretion of the Z variant of human alpha‐1‐proteinase inhibitor: secre-
tion of variants having altered potential for salt bridge formation between
amino acids 290 and 342. Mol Cell Biol, 1989;9(4):1406–14.
53. Lomas, D.A., Evans, D.L., Stone, S.R., Chang, W.S., and Carrell, R.W.
Effect of the Z mutation on the physical and inhibitory properties of alpha
1‐antitrypsin. Biochemistry, 1993;32(2):500–8.
54. Lomas, D.A., Evans, D.L., Finch, J.T., and Carrell, R.W. The mechanism of Z
alpha 1‐antitrypsin accumulation in the liver. Nature, 1992;357(6379):605–7.
55. Lomas, D.A., Finch, J.T., Seyama, K., Nukiwa, T., and Carrell, R.W. Alpha
1‐antitrypsin Siiyama (Ser53‐‐>Phe). Further evidence for intracellular loop‐
sheet polymerization. J Biol Chem, 1993;268(21):15333–5.
56. Dafforn, T.R., Mahadeva, R., Elliott, P.R., Sivasothy, P., and Lomas, D.A. A
kinetic mechanism for the polymerization of alpha1‐antitrypsin. J Biol
Chem, 1999;274(14):9548–55.
57. Yamasaki, M., Li, W., Johnson, D.J., and Huntington, J.A. Crystal structure
of a stable dimer reveals the molecular basis of serpin polymerization.
Nature, 2008;455(7217):1255–8.
 α1-ANTITRYPSIN DEFICIENCY
50:  657
58. Whisstock, J.C., Silverman, G.A., Bird, P.I. et al. Serpins flex their muscle:
I.I. Structural insights into target peptidase recognition, polymerization, and
transport functions. J Biol Chem, 2010;285(32):24307–12.
59. Yamasaki, M., Sendall, T.J., Pearce, M.C., Whisstock, J.C., and Huntington,
J.A. Molecular basis of alpha1‐antitrypsin deficiency revealed by the struc-
ture of a domain‐swapped trimer. EMBO Rep, 2011;12(10):1011–7.
60. Huang, X., Zheng, Y., Zhang, F. et al. Molecular mechanism of Z alpha1‐
antitrypsin deficiency. J Biol Chem, 2016;291(30):15674–86.
61. Hurtley, S.M. and Helenius, A. Protein oligomerization in the endoplasmic
reticulum. Annu Rev Cell Biol, 1989;5:277–307.
62. Kim, J., Lee, K.N., Yi, G.S., and Yu, M.H. A thermostable mutation located
at the hydrophobic core of alpha 1‐antitrypsin suppresses the folding defect
of the Z‐type variant. J Biol Chem, 1995;270(15):8597–601.
63. Sidhar, S.K., Lomas, D.A., Carrell, R.W., and Foreman, R.C. Mutations
which impede loop/sheet polymerization enhance the secretion of human
alpha 1‐antitrypsin deficiency variants. J Biol Chem, 1995;270(15):8393–6.
64. Kang, H.A., Lee, K.N., and Yu, M.H. Folding and stability of the Z and
S(iiyama) genetic variants of human alpha1‐antitrypsin. J Biol Chem,
1997;272(1):510–6.
65. Schmidt, B.Z. and Perlmutter, D.H. Grp78, Grp94, and Grp170 interact with
alpha1‐antitrypsin mutants that are retained in the endoplasmic reticulum.
Am J Physiol Gastrointest Liver Physiol, 2005;289(3):G444–55.
66. Lin, L., Schmidt, B., Teckman, J., and Perlmutter, D.H. A naturally occurring
nonpolymerogenic mutant of alpha 1‐antitrypsin characterized by prolonged
retention in the endoplasmic reticulum. J Biol Chem, 2001;276(36):33893–8.
67. Mallya, M., Phillips, R.L., Saldanha, S.A. et al. Small molecules block the
polymerization of Z alpha1‐antitrypsin and increase the clearance of intra-
cellular aggregates. J Med Chem, 2007;50(22):5357–63.
68. Yu, M.H., Lee, K.N., and Kim, J. The Z type variation of human alpha 1‐anti-
trypsin causes a protein folding defect. Nat Struct Biol, 1995;2(5):363–7.
69. Davis, R.L., Shrimpton, A.E., Holohan, P.D., et al. Familial dementia caused
by polymerization of mutant neuroserpin. Nature, 1999;401(6751):376–9.
70. Nyfeler, B., Reiterer, V., Wendeler, M.W. et al. Identification of ERGIC‐53
as an intracellular transport receptor of alpha1‐antitrypsin. J Cell Biol,
2008;180(4):705–12.
71. Hidvegi, T., Ewing, M., Hale, P. et  al. An autophagy‐enhancing drug pro-
motes degradation of mutant alpha1‐antitrypsin Z and reduces hepatic fibro-
sis. Science, 2010;329(5988):229–32.
72. Marcus, N.Y., Brunt, E.M., Blomenkamp, K. et al. Characteristics of hepato-
cellular carcinoma in a murine model of alpha‐1‐antitrypsin deficiency.
Hepatol Res, 2010;40(6):641–53.
73. Teckman, J.H., An, J.K., Blomenkamp, K., Schmidt, B., and Perlmutter, D.
Mitochondrial autophagy and injury in the liver in alpha 1‐antitrypsin defi-
ciency. Am J Physiol Gastrointest Liver Physiol, 2004;286(5):G851–62.
74. Lindblad, D., Blomenkamp, K., and Teckman, J. Alpha‐1‐antitrypsin mutant
Z protein content in individual hepatocytes correlates with cell death in a
mouse model. Hepatology, 2007;46(4):1228–35.
75. Hidvegi, T., Mirnics, K., Hale, P., Ewing, M., Beckett, C., and Perlmutter,
D.H. Regulator of G signaling 16 is a marker for the distinct endoplasmic
reticulum stress state associated with aggregated mutant alpha1‐antitrypsin
Z in the classical form of alpha1‐antitrypsin deficiency. J Biol Chem,
2007;282(38):27769–80.
76. Dooley, S., Hamzavi, J., Ciuclan, L. et  al. Hepatocyte‐specific Smad7
expression attenuates TGF‐beta‐mediated fibrogenesis and protects against
liver damage. Gastroenterology, 2008;135(2):642–59.
77. Hidvegi, T., Schmidt, B.Z., Hale, P., and Perlmutter, D.H. Accumulation of
mutant alpha1‐antitrypsin Z in the endoplasmic reticulum activates cas-
pases‐4 and ‐12, NFkappaB, and BAP31 but not the unfolded protein
response. J Biol Chem, 2005;280(47):39002–15.
78. Mukherjee, A., Hidvegi, T., Araya, P., Ewing, M., Stolz, D.B., and Perlmutter,
D.H. NFkappaB mitigates the pathological effects of misfolded alpha1‐anti-
trypsin by activating autophagy and an integrated program of proteostasis
mechanisms. Cell Death Differ, 2019;26(3)455–69.
79. Bridges, J.P., Wert, S.E., Nogee, L.M., and Weaver, T.E. Expression of a
human surfactant protein C mutation associated with interstitial lung disease
disrupts lung development in transgenic mice. J Biol Chem, 2003;278(52):
52739–46.
80. Young, L.R., Gulleman, P.M., Bridges, J.P. et  al. The alveolar epithelium
determines susceptibility to lung fibrosis in Hermansky–Pudlak syndrome.
Am J Respir Crit Care Med, 2012;186(10):1014–24.
81. Bhuiyan, M.S., Pattison, J.S., Osinska, H. et al. Enhanced autophagy amelio-
rates cardiac proteinopathy. J Clin Invest, 2013;123(12):5284–97.
  82. Doppler, K., Mittelbronn, M., Lindner, A., and Bornemann, A. Basement
membrane remodelling and segmental fibrosis in sporadic inclusion body
myositis. Neuromuscul Disord, 2009;19(6):406–11.
  83. Nogalska, A., D’Agostino, C., Terracciano, C., Engel, W.K., and Askanas,
V. Impaired autophagy in sporadic inclusion‐body myositis and in endo-
plasmic reticulum stress‐provoked cultured human muscle fibers. Am J
Pathol, 2010;177(3):1377–87.
  84. Hubner, R.H., Leopold, P.L., Kiuru, M., De, B.P., Krause, A., and Crystal,
R.G. Dysfunctional glycogen storage in a mouse model of alpha1‐antit-
rypsin deficiency. Am J Respir Cell Mol Biol, 2009;40(2):239–47.
  85. Piccolo, P., Annunziata, P., Soria, L.R. et al. Down‐regulation of hepatocyte
nuclear factor‐4alpha and defective zonation in livers expressing mutant Z
alpha1‐antitrypsin. Hepatology, 2017;66(1):124–35.
 86. Wu, Y., Whitman, I., Molmenti, E., Moore, K., Hippenmeyer, P., and
Perlmutter, D.H. A lag in intracellular degradation of mutant alpha 1‐antit-
rypsin correlates with the liver disease phenotype in homozygous PiZZ alpha
1‐antitrypsin deficiency. Proc Natl Acad Sci USA, 1994;91(19):9014–8.
  87. Qu, D., Teckman, J.H., Omura, S., and Perlmutter, D.H. Degradation of a
mutant secretory protein, alpha1‐antitrypsin Z., in the endoplasmic reticu-
lum requires proteasome activity. J Biol Chem, 1996;271(37):22791–5.
  88. Werner, E.D., Brodsky, J.L., and McCracken, A.A. Proteasome‐dependent
endoplasmic reticulum‐associated protein degradation: an unconventional
route to a familiar fate. Proc Natl Acad Sci USA, 1996;93(24):13797–801.
  89. Kamimoto, T., Shoji, S., Hidvegi, T. et al. Intracellular inclusions contain-
ing mutant alpha1‐antitrypsin Z are propagated in the absence of autophagic
activity. J Biol Chem, 2006;281(7):4467–76.
  90. Kruse, K.B., Brodsky, J.L., and McCracken, A.A. Characterization of an
ERAD gene as VPS30/ATG6 reveals two alternative and functionally dis-
tinct protein quality control pathways: one for soluble Z variant of human
alpha‐1 proteinase inhibitor (A1PiZ) and another for aggregates of A1PiZ.
Mol Biol Cell, 2006;17(1):203–12.
  91. Kruse, K.B., Dear, A., Kaltenbrun, E.R. et  al. Mutant fibrinogen cleared
from the endoplasmic reticulum via endoplasmic reticulum‐associated pro-
tein degradation and autophagy: an explanation for liver disease. Am J
Pathol, 2006;168(4):1299–308; quiz 404–5.
  92. Wang, Y. and Perlmutter, D.H. Targeting intracellular degradation pathways
for treatment of liver disease caused by alpha1‐antitrypsin deficiency.
Pediatr Res, 2014;75(1–2):133–9.
  93. Gelling, C.L., Dawes, I.W., Perlmutter, D.H., Fisher, E.A., and Brodsky,
J.L. The endosomal protein‐sorting receptor sortilin has a role in trafficking
alpha‐1 antitrypsin. Genetics, 2012;192(3):889–903.
  94. Long, O.S., Benson, J.A., Kwak, J.H. et al. A C. elegans model of human
alpha1‐antitrypsin deficiency links components of the RNAi pathway to
misfolded protein turnover. Hum Mol Genet, 2014;23(19):5109–22.
  95. Liao, Y., Shikapwashya, O.N., Shteyer, E., Dieckgraefe, B.K., Hruz, P.W.,
and Rudnick, D.A. Delayed hepatocellular mitotic progression and
impaired liver regeneration in early growth response‐1‐deficient mice. J
Biol Chem, 2004;279(41):43107–16.
  96. Hidvegi, T., Stolz, D.B., Alcorn, J.F. et al. Enhancing autophagy with drugs
or lung‐directed gene therapy reverses the pathological effects of respira-
tory epithelial cell proteinopathy. J Biol Chem, 2015;290(50):
29742–57.
  97. Pastore, N., Attanasio, S., Granese, B. et al. Activation of the c‐Jun N‐ter-
minal kinase pathway aggravates proteotoxicity of hepatic mutant Z
alpha1‐antitrypsin. Hepatology, 2017;65(6):1865–74.
 98. Rudnick, D.A., Liao, Y., An, J.K., Muglia, L.J., Perlmutter, D.H., and
Teckman, J.H. Analyses of hepatocellular proliferation in a mouse model of
alpha‐1‐antitrypsin deficiency. Hepatology, 2004;39(4):1048–55.
  99. Rudnick, D.A. and Perlmutter, D.H. Alpha‐1‐antitrypsin deficiency: a new
paradigm for hepatocellular carcinoma in genetic liver disease. Hepatology,
2005;42(3):514–21.
100. Ding, J., Yannam, G.R., Roy‐Chowdhury, N. et  al. Spontaneous hepatic
repopulation in transgenic mice expressing mutant human alpha1‐antit-
rypsin by wild‐type donor hepatocytes. J Clin Invest, 2011;121(5):
1930–4.
101. Pan, S., Huang, L., McPherson, J. et al. Single nucleotide polymorphism‐
mediated translational suppression of endoplasmic reticulum mannosidase
I modifies the onset of end‐stage liver disease in alpha1‐antitrypsin defi-
ciency. Hepatology, 2009;50(1):275–81.
102. Pan, S., Wang, S., Utama, B. et al. Golgi localization of ERManI defines
spatial separation of the mammalian glycoprotein quality control system.
Mol Biol Cell, 2011;22(16):2810–22.
658 THE LIVER:  REFERENCES
103. Iannotti, M.J., Figard, L., Sokac, A.M., and Sifers, R.N. A Golgi‐localized
mannosidase (MAN1B1) plays a non‐enzymatic gatekeeper role in protein
biosynthetic quality control. J Biol Chem, 2014;289(17):11844–58.
104. Chappell, S., Guetta‐Baranes, T., Hadzic, N., Stockley, R., and Kalsheker,
N. Polymorphism in the endoplasmic reticulum mannosidase I (MAN1B1)
gene is not associated with liver disease in individuals homozygous for the
Z variant of the alpha1‐antitrypsin protease inhibitor (PiZZ individuals).
Hepatology, 2009;50(4):1315;6.
105. Joly, P., Lachaux, A., Ruiz, M. et al. SERPINA1 and MAN1B1 polymor-
phisms are not linked to severe liver disease in a French cohort of alpha‐1
antitrypsin deficiency children. Liver Int, 2017;37(11):1608–11.
106. Chappell, S., Hadzic, N., Stockley, R., Guetta‐Baranes, T., Morgan, K., and
Kalsheker, N. A polymorphism of the alpha1‐antitrypsin gene represents a
risk factor for liver disease. Hepatology, 2008;47(1):127–32.
107. Janus, E.D., Phillips, N.T., and Carrell, R.W. Smoking, lung function, and
alpha 1‐antitrypsin deficiency. Lancet, 1985;1(8421):152–4.
108. Kemmer, N., Kaiser, T., Zacharias, V., and Neff, G.W. Alpha‐1‐antitrypsin
deficiency: outcomes after liver transplantation. Transplant Proc,
2008;40(5):1492–4.
109. Sarkar, S., Perlstein, E.O., Imarisio, S. et  al. Small molecules enhance
autophagy and reduce toxicity in Huntington’s disease models. Nat Chem
Biol, 2007;3(6):331–8.
110. Li, C., Xiao, P., Gray, S.J., Weinberg, M.S., and Samulski, R.J. Combination
therapy utilizing shRNA knockdown and an optimized resistant transgene
for rescue of diseases caused by misfolded proteins. Proc Natl Acad Sci
USA, 2011;108(34):14258–63.
111. Mueller, C., Tang, Q., Gruntman, A. et  al. Sustained miRNA‐mediated
knockdown of mutant AAT with simultaneous augmentation of wild‐type
AAT has minimal effect on global liver miRNA profiles. Mol Ther,
2012;20(3):590–600.
112. Guo, S., Booten, S.L., Aghajan, M. et al. Antisense oligonucleotide treat-
ment ameliorates alpha‐1 antitrypsin‐related liver disease in mice. J Clin
Invest, 2014;124(1):251–61.
113. Pastore, N., Blomenkamp, K., Annunziata, F. et al. Gene transfer of master
autophagy regulator TFEB results in clearance of toxic protein and correc-
tion of hepatic disease in alpha‐1‐anti‐trypsin deficiency. EMBO Mol Med,
2013;5(3):397–412.
114. Shen, S., Sanchez, M.E., Blomenkamp, K. et al. Amelioration of Alpha‐1
antitrypsin deficiency diseases with genome editing in transgenic mice.
Hum Gene Ther, 2018;29(8):861–73.
115. Song, C.Q., Wang, D., Jiang, T. et  al. In vivo genome editing partially
restores alpha1‐antitrypsin in a murine model of AAT deficiency. Hum
Gene Ther, 2018;29(8):853–60.
116. Alam, S., Wang, J., Janciauskiene, S., and Mahadeva, R. Preventing and
reversing the cellular consequences of Z alpha‐1 antitrypsin accumulation
by targeting s4A. J Hepatol, 2012;57(1):116–24.
117. Burrows, J.A., Willis, L.K., and Perlmutter, D.H. Chemical chaperones
mediate increased secretion of mutant alpha 1‐antitrypsin (alpha 1‐AT) Z:
A potential pharmacological strategy for prevention of liver injury and
emphysema in alpha 1‐AT deficiency. Proc Natl Acad Sci USA,
2000;97(4):1796–801.
118. Teckman, J.H. Lack of effect of oral 4‐phenylbutyrate on serum alpha‐1‐
antitrypsin in patients with alpha‐1‐antitrypsin deficiency: a preliminary
study. J Pediatr Gastroenterol Nutr, 2004;39(1):34–7.
119. Bouchecareilh, M., Hutt, D.M., Szajner, P., Flotte, T.R., and Balch, W.E.
Histone deacetylase inhibitor (HDACi) suberoylanilide hydroxamic acid
(SAHA)‐mediated correction of alpha1‐antitrypsin deficiency. J Biol
Chem, 2012;287(45):38265–78.
120. Fox, I.J., Chowdhury, J.R., Kaufman, S.S. et al. Treatment of the Crigler‐
Najjar syndrome type I with hepatocyte transplantation. N Engl J Med,
1998;338(20):1422–6.
121. Yusa, K., Rashid, S.T., Strick‐Marchand, H. et al. Targeted gene correction
of alpha1‐antitrypsin deficiency in induced pluripotent stem cells. Nature,
2011;478(7369):391–4.
122. Perlmutter, D.H. and Silverman, G.A. Hepatic fibrosis and carcinogenesis
in alpha1‐antitrypsin deficiency: a prototype for chronic tissue damage in
gain‐of‐function disorders. Cold Spring Harb Perspect Biol, 2011;3(3).
123. Juhasz, G. and Neufeld, T.P. Autophagy: a forty‐year search for a missing
membrane source. PLoS Biol, 2006;4(2):e36.
 Pathophysiology of Portal
51 Hypertension
Yasuko Iwakiri and Roberto J. Groszmann
Section of Digestive Diseases, Yale School of Medicine, New Haven, CT, USA
INTRODUCTION
Portal hypertension refers to a pathological increase in blood
pressure within the portal system due to obstruction of portal
blood flow. The most frequent cause of portal hypertension is
liver cirrhosis, and many of lethal complications of cirrhosis
such as ascites and gastroesophageal variceal hemorrhage are
related to portal hypertension. With less frequency, portal hyper-
tension also occurs in noncirrhotic conditions, such as portal
vein thrombosis, chronic schistosomiasis, heart failure, Budd–
Chiari syndrome, and idiopathic portal hypertension. As a disor-
der of portal venous pressure, portal hypertension can be
conceptualized using the hydraulic derivation of Ohm’s Law
(pressure = flow × resistance) [1], composed of variables
grounded in basic vascular biology.
Portal hypertension develops in liver cirrhosis as a result of
increased intrahepatic vascular resistance caused by multiple
pathological events in the sinusoidal circulation, such as struc-
tural distortion by fibrosis, microvascular thrombosis, dysfunc-
tion of liver sinusoidal endothelial cells (LSECs), and hepatic
stellate cell activation [2–5]. LSECs and pericyte‐like hepatic
stellate cells are closely associated with one another in these
pathological events in paracrine and autocrine manners.
Portal hypertension then leads to splanchnic and systemic
arterial vasodilation, which in turn contributes to increases in
splanchnic blood flow to the portal system and thus further
increases in portal pressure despite the formation of portosys-
temic collaterals [3–7]. This condition aggravates portal
­hypertension and facilitates the hyperdynamic circulation char-
acterized by decreased mean arterial pressure, decreased sys-
temic vascular resistance, increased cardiac index, and decreased
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
peripheral resistance. The hyperdynamic circulation along with
increased blood flow to portosystemic collaterals results in
­clinically devastating complications such as gastroesophageal
varices and variceal hemorrhage, hepatic encephalopathy,
ascites, and renal failure due to the hepato‐renal syndrome
[8–10] (Figure 51.1).
This chapter provides current knowledge of the biology of
portal hypertension with a focus on cellular and molecular
events in different segments of vasculatures (intrahepatic versus
extrahepatic circulations) that contribute to the development
and perpetuation of portal hypertension and the subsequent
development of hyperdynamic circulatory syndrome.
INTRAHEPATIC CIRCULATION
An overview
Hepatic sinusoids are specialized vascular beds that constitute
the hepatic microcirculation. Blockage of sinusoids and the
resulting increased hepatic vascular resistance to portal venous
flow is the primary cause of portal hypertension. In cirrhosis,
blockage of sinusoids is largely a result of massive structural
changes in the liver associated with fibrosis/cirrhosis and intra-
hepatic vasoconstriction [5, 11, 12]. Phenotypic changes in
hepatic cells, such as hepatic stellate cells and LSECs, are
known to play pivotal roles in increased intrahepatic vascular
resistance and have been studied extensively. This section
addresses the cellular and molecular mechanisms underlying
increased intrahepatic vascular resistance with a focus on
hepatic stellate cells, LSECs, and thrombosis.
660 THE LIVER:  INTRAHEPATIC CIRCULATION
Functional Structural
Increased resistance
Varices
Porto-systemic
collaterals
PORTAL
HYPERTENSION
Vasodilatation Increased flow
Hypotension Increase in CO
Central hypovolemia
Activation NE, VP, A-II Na and water retention
increase venous return
to the heart
Figure 51.1  Summary of the pathophysiology of portal hyperten-
sion. The increase in hepatic resistance leads to an increase in portal
pressure. This leads to a cascade of disturbances in the splanchnic and
systemic circulation characterized by vasodilation, sodium and water
retention, and plasma volume expansion, which are major players in the
pathogenesis of ascites and hepatorenal syndrome. Additionally, these
alterations lead to an increase in portal blood inflow that contributes to
maintaining and aggravating portal hypertension. Another characteris-
tic feature is the development of portosystemic collaterals, which are
responsible for complications such as variceal bleeding and hepatic
encephalopathy. CO, cardiac output; NE, norepinephrine; VP, vasopres-
sin; A‐II, angiotensin II; Na, sodium.
Hepatic stellate cell biology
Fibrosis and architectural alterations
Hepatic fibrosis is thought to be the primary factor that contrib-
utes to increased intrahepatic vascular resistance in cirrhotic
liver. In a classic study, Bhathal and Grossman demonstrated
through vasodilator challenges in the isolated perfused cirrhotic
rat liver that 80% of the increased intrahepatic resistance to por-
tal venous flow is attributable to structural changes, while the
remaining 20% is due to a reversible, hypercontractile pheno-
type of the hepatic microcirculation [13]. The key pathway of
fibrosis in the liver is proinflammatory signaling that causes
activation of hepatic stellate cells and thereby leads to extracel-
lular matrix deposition. Understanding this pathway has led in
turn to an understanding of its reversal and the regression of
fibrosis, which holds a great therapeutic potential for chronic
liver disease and portal hypertension [14, 15].
Hepatic stellate cells as pericytes
Hepatic stellate cells are thought to play an additional role in
portal hypertension beyond fibrosis. Hepatic stellate cells are
located in the space of Disse, directly underneath LSECs, and
work as hepatic pericytes, perivascular non‐endothelial cells
with a variety of functions including regulation of vascular tone
through smooth muscle‐like contractility and regulation of
endothelial proliferation [16–19]. These notions arise from
studies demonstrating that activated hepatic stellate cells acquire
a myofibroblast‐like phenotype in response to liver injury [20]
and exhibit a contractile phenotype [21]. Thus hepatic stellate
cells, through perivascular contraction in the sinusoidal micro-
circulation, are thought to be key contributors to the dynamic
and reversible component of portal hypertension in cirrhosis.
Regulation of hepatic stellate cell contraction
Endothelin signaling is a key pathway that regulates a contractile
phenotype of hepatic stellate cells. Endothelin binds to G‐protein
coupled receptors, endothelin A (ETA) and endothelin B (ETB),
which are generally found on vascular smooth muscle cells and
endothelial cells, respectively. Endothelin‐1 (ET‐1) is the major
subtype found in liver disease and is preferentially bound to ETA
more than the other two subtypes (ET‐2 and ET‐3) [22]. ET‐1
protein levels are elevated in liver injury along with ET‐1 mRNA
[23]. While endothelial cells produce the majority of ET‐1 in
normal liver, liver injury shifts this production primarily to
hepatic stellate cells [24], which also profoundly upregulate ETA
and ETB receptors [25, 26], suggesting increased sensitivity to
this signal. ET‐1 has been shown to induce contraction of the
sinusoidal vasculature in an experimental model [27], and antag-
onism of ETA has been shown to reduce portal pressure in an
animal model of cirrhosis [28]. Smooth muscle cell or hepatic
stellate cell contraction is a result of myosin light chain (MLC).
A study on the mechanism of ET‐1‐mediated hepatic stellate cell
contraction has revealed that ET‐1 leads to MLC phosphoryla-
tion through multiple pathways [29]. One is a Ca2+‐dependent
pathway in which ET‐1 causes a transient increase in intracellu-
lar Ca2+, leading to MLC kinase activation and MLC phospho-
rylation. Others include activation of the protein kinase C and
Rho‐kinase pathways, both leading to MLC phosphorylation.
Liver sinusoidal endothelial cell biology
Fenestration and capillarization
LSECs have a distinct phenotype from endothelial cells else-
where in the liver, as well as elsewhere in the body. Their most
distinguishing feature is fenestration with fenestrae being
approximately 0.1 microns in size and organized into groups of
sieve plates. It is thought that fenestrae facilitate the transport of
macromolecules from hepatic sinusoids to the space of Disse,
where they can interact with hepatic stellate cells and hepato-
cytes. Another feature of LSECs distinct from endothelial cells
in other organs is the lack of a basement membrane, which
allows maximum permeability between the lumen of the sinu-
soid and the space of Disse [30].
LSECs lose their fenestrae, develop a basement membrane,
and become “capillarized” as a consequence of liver fibrosis [31,
32]. Vascular endothelial growth factor (VEGF) is known as a key
factor for the maintenance of endothelial fenestrae [33]. Inhibition
of VEGF signaling in a transgenic animal model, in which liver‐
specific secretion of a soluble VEGF decoy receptor sequesters
endogenous VEGF, caused loss of LSEC fenestration and resulted
in portal hypertension and hepatic stellate cell activation inde-
pendent of hepatic parenchymal damage, with reversal of portal
hypertension with restoration of VEGF [34]. This VEGF action
on LSEC phenotype is nitric oxide (NO)‐dependent. Accordingly,
 51:  Pathophysiology of Portal Hypertension 661

an NO synthase inhibitor, Nω‐Nitro‐L‐arginine methyl ester
hydrochloride (L‐NAME), can lead to the loss of the LSEC phe-
notype [35].
Besides VEGF, various factors have been shown to alter
LSEC fenestration. Collagens in the space of Disse may play a
role in maintenance or loss of LSEC fenestration [36]. A role of
lipid rafts in regulation of fenestration has also been ­investigated.
Using super‐resolution fluorescence microscopy, researchers
showed an inverse relationship between areas of lipid rafts and
areas of membranes with fenestration in LSECs. They also dem-
onstrated that inhibiting lipid raft formation via 7‐ketocholesterol
or actin disruption increased fenestration, and conversely that
increasing lipid raft formation with a low concentration of
Triton X‐100 decreased fenestration [37].
Crosstalk between LSECs and hepatic stellate cells
While phenotypic changes of LSEC is known as a consequence
of liver fibrosis/cirrhosis as mentioned above [31, 32], it is also
thought that loss of the LSEC phenotype can be permissive for
hepatic stellate cell activation [38] and that the communication
between LSECs and hepatic stellate cells plays a key role in the
pathogenesis of portal hypertension (Figure 51.2) [39]. Research
into LSEC and hepatic stellate cell communication demon-
strated that an isoform of fibronectin produced by LSECs in a
bile duct ligation model of liver damage was able to activate
hepatic stellate cells [40], though a subsequent study revealed it
to be a factor important for hepatic stellate cell motility but not
differentiation to a myofibroblast phenotype [41]. Another study
has shown that LSECs communicate with hepatic stellate cells
via exosomes containing sphingosine kinase‐1 (SK1) and its
product sphingosine‐1 phosphate, providing a signal for hepatic
stellate cell migration [42], which is closely connected to their
activated phenotype.
Nitric oxide produced by LSECs not only regulates vascular
tone in the liver, but also plays a key role in the crosstalk between
LSECs and hepatic stellate cells. LSECs are able to induce
hepatic stellate cell reversion from activation to quiescence via
an NO‐dependent mechanism [43], possibly paracrine signaling
via the Kruppel‐like factor 2 (KLF2)‐NO‐guanylate cyclase
pathway in LSECs [44]. Another study showed that NO donors
were able to inhibit proliferation and chemotaxis of activated
hepatic stellate cells in response to platelet‐derived growth fac-
tor by disrupting its intracellular signaling pathway in prosta-
glandin E2‐mediated manners [45]. Further, NO inhibited
hepatic stellate cell migration via cGMP‐dependent protein
kinase (PKG)‐mediated inhibition of the Rac1 pathway [46,
47]. NO signaling has also been shown to induce hepatic stellate
cell apoptosis [48] via a caspase‐independent mechanism pos-
sibly related to increased mitochondrial oxidative stress and
increased mitochondrial membrane permeability, along with a
possible lysosomal stress component. These observations may
have relevance in diseases such as alcoholic liver injury and
non‐alcoholic fatty liver disease, in which loss of the normal
LSEC phenotype has been shown to occur before fibrosis [38].
Endothelial dysfunction
NO is a critical regulator of normal hepatic vascular tone and
portal pressure [49]. The sources of NO in the hepatic vascula-
ture are LSECs and endothelial cells of blood vessels. These
cells constitutively express endothelial nitric oxide synthase
(eNOS) and produce a low level of NO. Production of NO
increases in response to an increase in blood flow via shear

Figure 51.2  LSEC/HSC crosstalk. (a) Injured liver sinusoidal endothelial cells (LSECs) lead to activation of hepatic stellate cells (HSCs). Liver
injury leads LSECs to produce the EIIIA isoform of fibronectin, which signals to HSCs through integrin α9β1 to promote motility, which is impor-
tant for their activated phenotype. LSECs also signal to promote HSC motility through sphingosine kinase 1 – sphingosine‐1‐phosphate (SK1‐S1P)‐
containing exosomes, which adhere to HSCs via fibronectin binding to an integrin receptor. Nitric oxide (NO) production by sinusoidal endothelial
cells is important in maintaining HSC quiescence, with a Kruppel‐like factor (KLF) 2 pathway enhancing NO and guanylate cyclase production.
(b) NO production from LSECs also inhibiting the Rac/Rho pathway in HSCs. NO also causes HSC apoptosis, and may thereby limit the number
of activated HSCs in the liver. Vit A: Vitamin A. Modified from McConnell and Iwakiri [39] and reproduced with permission of Springer Nature.
HSC are in green; healthy LSEC are in blue; injured LSEC are in red.
662 THE LIVER:  EXTRAHEPATIC CIRCULATION
stress [50] and can also be increased by VEGF [51]. LSECs in
cirrhotic liver express eNOS similarly to LSECs in normal liver.
However, eNOS activity is lower under pathologic conditions
and NO release is diminished in disease [52]. In addition,
LSECs in cirrhosis have a reduced ability to respond to increases
in blood flow compared to the healthy state [53]. These adverse
changes in LSECs result in decreased production of NO and
impaired vasodilation in the hepatic microcirculation in cirrho-
sis and are important contributors to increased intrahepatic vas-
cular resistance observed in portal hypertension. The regulation
of eNOS function involves multiple regulators acting in concert
with stimulatory signaling, including its phosphorylation by the
protein kinase Akt [54], which is facilitated by G‐protein cou-
pled‐receptor kinase interactor‐1 (GIT1) [55, 56]. The function
of eNOS may also be inhibited by binding to caveolin‐1, which
can be disrupted by calmodulin [57].
Because endothelial dysfunction leads to increased vascular
resistance in the sinusoidal microcirculation and promotes acti-
vation of hepatic stellate cells, a pharmacological approach that
reverses the dysfunctional LSEC phenotype could be an effec-
tive therapeutic strategy. An example is the use of statins.
Studies have shown that statins ameliorate portal hypertension
in cirrhotic patients [58] as well as experimental models of por-
tal hypertension [59, 60]. These studies demonstrated that
statins improve endothelial dysfunction and increase NO bioa-
vailability in the sinusoidal microcirculation. Several potential
mechanisms for this increased NO bioavailability have been
proposed. One is the ability of statins to inhibit synthesis of iso-
prenoids, which are critical for membrane anchoring and activa-
tion of small GTPases, such as RhoA. Given that RhoA/
Rho‐kinase signaling could decrease eNOS activity [61] and
expression [62], statins, by decreasing RhoA activity, could
improve eNOS function, thereby enhance NO bioavailability
and decrease intrahepatic vascular resistance [60]. Another
potential mechanism is that statins increase activity of Akt/pro-
tein kinase B, which phosphorylates and activates eNOS,
thereby increasing NO bioavailability [59]. In addition to
improving endothelial cell dysfunction, statins could target the
RhoA/Rho‐kinase pathway in pericytes (e.g. activated hepatic
stellate cells) and decrease their contractility, thereby lowering
intrahepatic vascular resistance and ameliorating portal hyper-
tension [60].
Pathological angiogenesis
Angiogenesis, or the process of new blood vessel formation from
pre‐existing vascular beds, has also been implicated in portal
hypertension. Hepatic angiogenesis is thought to generate irregu-
lar intrahepatic circulatory routes and thus could increase intrahe-
patic vascular resistance. The Notch1 signaling pathway is known
to be important for embryonic vascular development and postna-
tal vascular remodeling. In the postnatal liver, inhibition of the
Notch1 pathway leads to nodular regenerative hyperplasia, a
common etiology of noncirrhotic portal hypertension. Notch1
deletion in mice resulted in an abnormal sinusoidal microcircula-
tion, as indicated by de‐differentiation of LSECs, pathological
remodeling of the hepatic sinusoidal microvasculature, intussus-
ceptive angiogenesis (also known as splitting angiogenesis),
and  dysregulation of ephrinB2/EphB4 and endothelial tyrosine
kinase (i.e. impairment of arterial versus venous specification).
Interestingly, animals with lack of Notch1 gene developed portal
hypertension even before the onset of nodular regenerative hyper-
plasia, and this was thought to result from intrahepatic vascular
disorders, possibly due to intussusceptive angiogenesis in the
liver microcirculation [63]. Angiogenesis has also been shown to
be associated with fibrosis progression in the liver [64, 65].
However, this relationship is complex and could be causative or
correlative due to involvement of hypoxia, a strong inducer of
angiogenesis. Modulation of angiogenesis does not generally pro-
vide a predictable effect on liver fibrosis, either [66].
Microvascular thrombosis/platelet activation
The study of intrahepatic portal hypertension is evolving to
include platelet activation and thrombosis as crucial factors for
its pathophysiology. Ian Wanless and others were instrumental
contributors in this area, observing what they termed “parenchy-
mal extinction” accounting for fibrosis progression due to intra-
hepatic vascular thrombosis [67]. Further, while cirrhosis had
previously been thought to have a bleeding trend, more advanced
physiologic tests to assess coagulation status [68] and system-
atic studies of bleeding complications [69] have led to an impor-
tant consensus that hemostatic status in cirrhosis is rebalanced
or perhaps even prothrombotic.
Platelets are a key player of the typical physiology of vascu-
lar thrombosis. Initially, cirrhotic platelets were thought to be
dysfunctional and predispose patients to a bleeding trend [70,
71]. More recently, however, it has been found that activity of
cirrhotic platelets in hemostasis and thrombosis is potentially
preserved [72] or even increased [73], although conflicting data
also exist [74]. Generally, the function of platelets in various
types of liver injury is quite complex with multiple stage‐spe-
cific factors influencing the role of platelets as profibrotic or
antifibrotic [75]. Further experiments and clinical trials will
continue to expand our knowledge regarding the role of throm-
bosis and platelets in sinusoidal portal hypertension.
EXTRAHEPATIC CIRCULATION
An overview
In addition to the liver vasculature, the mesenteric vasculature
plays a key role in portal hypertension. Foundational knowledge
of the pathophysiology of this vascular bed comes from seminal
studies by Groszmann and others [76–96]. These studies dem-
onstrated that in portal hypertension, even with increased
hepatic vascular resistance, the splanchnic circulation is hyper-
dynamic. Splanchnic arterial vasodilation is a key feature of the
hyperdynamic circulation, as it can perpetuate increased blood
inflow to the portal system and thus exacerbate portal hyperten-
sion [4, 5]. Arterial vasodilation is attributable to abnormal cell
function in different layers of the vasculature, such as endothe-
lial cells, smooth muscle cells, and the adventitial layer that
contains vascular progenitor cells and neuronal termini. This
section discusses the mechanisms of arterial vasodilation and
collateral vessel formation in the splanchnic and systemic circu-
lations in cirrhosis with portal hypertension.
 51:  Pathophysiology of Portal Hypertension 663

Arterial vasodilation in the splanchnic
and systemic circulations
Arterial vasodilation
NO is probably the most important vasodilator molecule that
contributes to excessive vasodilation observed in arteries of the
splanchnic and systemic circulations in portal hypertension.
Experimental models of portal hypertension with or without cir-
rhosis have also shown that other vasodilator molecules, such as
carbon monoxide (CO), prostacyclin (PGI2), endocannabinoids,
and endothelium‐derived hyperpolarizing factor (EDHF), are
also induced [5, 7, 97]. The induction of arterial vasodilation
despite inhibition of NO, CO, and PGI2 has indicated the pres-
ence of other endothelium‐derived vasodilator molecule(s),
known as EDHF [98]. As its name implies, the action of EDHF
is to hyperpolarize vascular smooth muscle cells, causing
vascular relaxation. The candidates of EDHF include ­arachidonic
acid metabolites, epoxyeicosatrienoic acid (EET), potassium
ions (K+), components of gap junctions, and hydrogen peroxide
[5]. Vasoactive molecules known to be involved in the regula-
tion of vascular tone in cirrhosis are summarized in Figure 51.3.
An increase in portal pressure triggers eNOS activation and
subsequent NO overproduction in the extrahepatic circulation.
Changes in portal pressure are detected at different vascular
beds depending on the severity of portal hypertension [51]. An
increment increase in portal pressure is sensed first by the intes-
tinal microcirculation and increases VEGF production with a
subsequent increase in eNOS levels in the intestinal microcircu-
lation. When portal pressure further increases and reaches a
­certain level, arterial vasodilation develops in the splanchnic
circulation (i.e. the mesenteric arteries). In contrast to arterial
vasodilation in the intestinal microcirculation where VEGF

(a) (b) (c)

Normal Cirrhosis Cirrhosis

Intrahepatic circulation Splanchnic and Systemic Circulation

Sinusoidal Endothelial Cells Endothelial Cells

HbCO
Adrenomedullin Hb
VEGF
TNFα Agonists CO
P1 P1
-1 3 -1 3 Shear Stress
ET Ak ET Ak GTP
t GRK2 t 1 Cyclooxygenase
P v-
BR Akt BR Ca IP3
ET ET 1 AK1
P BH4 BH4 v- eNOS BH4
eNOS CaM eNOS CaM Ca PO4
Ca2+
xxx Hsp90 Hsp90
Hsp90
eNOSCa2+ HO-1 COX
COX-1 COX-1

NO CO PGI2
NO
NO NO TXA2
sGC
ETAR sGC sGC AC
cGMP
cGMP cGMP cAMP
Stellate Cells
Smoooth Muscle Cells

Vasoconstriction Vasodilatation
NO
TXA2
CO
NO PGI2
Endothelial hypoactivation Endothelial hyperactivation

Figure 51.3  Vasoactive molecules known to be involved in the regulation of vascular tone in cirrhosis. In the arterial splanchnic and systemic
circulation (right panel), agonists such as adrenomedullin, vascular endothelial growth factor (VEGF), and tumor necrosis factor alpha (TNFα) or
physical stimuli such as shear stress stimulate Akt, which directly phosphorylates and activates endothelial nitric oxide synthase (eNOS). eNOS
requires cofactors such as tetrahydrobiopterin (BH4) for its activity. Heat shock protein 90 (Hsp90) is one of the positive regulators of eNOS. Like
NO, carbon monoxide (CO) produced by hemeoxygenase‐1 (HO‐1) causes vasodilation by activating soluble guanylate cyclase (sGC) to generate
cyclic guanosine monophosphate (cGMP) in vascular smooth muscle cells. Prostacyclin (PGI2) is synthesized by cyclooxygenase (COX) and
elicits smooth muscle relaxation by stimulating adenylate cyclase (AC) and generation of cyclic adenosine monophosphate (cAMP). In the intrahe-
patic circulation in cirrhosis (left panel), decreased NO and increased thromboxane A2 (TXA2) production in SECs results in a net reduction of
vasorelaxation in the intrahepatic circulation. Endothelin‐1 (ET‐1) has dual vasoactive effects, mediating vasoconstriction through binding to
endothelin A (ETA) receptors located on HSCs and causing HSC contraction. Binding of ET‐1 to ETB receptor (ETBR) mediates vasodilation
through Akt phosphorylation and eNOS phosphorylation in normal liver. In cirrhosis, G‐protein‐coupled receptor kinase‐2 (GRK2), an inhibitor of
G protein‐coupled receptor signaling, is upregulated in SECs, leading to the impairment of Akt phosphorylation and a reduction in NO production.
An increased production of COX‐1‐derived vasoconstrictor prostanoid TXA2 is also an example of endothelial dysfunction in cirrhosis. Modified
from Iwakiri and Groszmann [7] and reproduced with permission of Elsevier.
664 THE LIVER:  HYPERDYNAMIC CIRCULATORY SYNDROME
mediates eNOS upregulation, it is thought that mechanical
forces including cyclic strains and shear stress induce eNOS
activation and lead to NO overproduction in the splanchnic cir-
culation [51, 92, 93, 99, 100]. Arterial vasodilation in the
splanchnic circulation facilitates the systemic circulation to be
also hyperdynamic.
Hypocontractility
A decrease in contractility to vasoconstrictors is also typical of
arteries of the splanchnic and systemic circulations in portal
hypertension. This hypocontractility occurs largely due to the
presence of excessive vasodilator molecules (e.g. NO) in the
endothelium, but is to some degree attributable to a decrease in
several vasoconstrictive molecules produced in smooth muscle
cells and neurons. Those molecules include neuropeptide Y
[101], urotensin II [102, 103], angiotensin [104], and brady-
kinin [105, 106]. In arteries of the splanchnic circulation, it has
been shown that vasodilators increase and vasoconstrictors
decrease.
Neural factors
Neural factors are also thought to be involved in the dysfunction
of vascular tone in portal hypertension, especially through the
sympathetic nervous system [101, 107, 108]. It is reported that
sympathetic nerve atrophy/regression observed in mesenteric
arterial beds of portal hypertensive rats leads to vasodilation
and/or hypocontractility of those arterial beds [109, 110]. The
role of neural factors in decreased contractile responses remains
to be fully elucidated.
Structural changes of arteries
The thinning of arterial walls is observed in the splanchnic and
systemic circulations of rats with cirrhotic livers [111, 112].
While this arterial wall thinning can be a consequence of the
hyperdynamic circulation, it may also sustain arterial vasodila-
tion and exacerbate portal hypertension [5, 6]. While NO plays
a role at least in part, the molecular mechanisms responsible for
arterial wall thinning remain to be understood.
Collateral vessel formation
Portosystemic collateral vessels develop in response to an
increase in portal pressure. These collateral vessels develop in
an attempt to decompress the hypertensive portal system and are
formed through the opening of pre‐existing vessels or angiogen-
esis [113, 114]. However, these collateral vessels are also known
to cause serious complications, including variceal bleeding and
hepatic encephalopathy [5, 115, 116]. A change in portal pres-
sure is thought to be detected first by the intestinal microcircula-
tory bed, followed by arteries of the splanchnic circulation [51].
These vascular beds subsequently generate various angiogenic
factors, such as VEGF [96, 117, 118] and placental growth fac-
tor (PlGF) [119], which promote the formation of portosystemic
collaterals.
Experimental studies of portal hypertension and cirrhosis
have shown that portosystemic collaterals are reduced by 18 to
78% with treatment by anti‐VEGFR2 [100], a combination of
anti‐VEGF (rapamycin)/anti‐PDGF (Gleevec) [120], anti‐PlGF
[119], apelin antagonist [121], sorafenib [122, 123], and a can-
nabinoid receptor 2 agonist [124]. However, the reduction of
these collaterals does not necessarily decrease portal pressure
because it does not substantially change the blood flow to the
portal vein. Therefore, the concomitant mitigation of arterial
vasodilation is needed to reduce portal pressure.
HYPERDYNAMIC CIRCULATORY
SYNDROME
An overview
Excessive arterial vasodilation in the splanchnic and systemic
circulations in portal hypertension results in the development of
hyperdynamic circulatory syndrome. This syndrome is charac-
terized by increased cardiac index, decreased systemic vascular
resistance, and decreased mean arterial pressure and eventually
leads to multiple organ failure frequently observed in chronic
liver disease. This section discusses multiple organ failure
observed in the hyperdynamic circulatory state (Figure 51.4).
The systemic circulation and the heart
Although arterial vasodilation in the splanchnic circulation is an
essential initiating factor, hyperdynamic circulation does not
occur without expansion of plasma volume and the development
of portosystemic collaterals [125, 126]. In cirrhotic patients,
blood and plasma volumes are elevated, but are not evenly dis-
tributed among vascular areas [127]. For example, arterial blood
volumes in the heart, lungs, and central arterial tree are decreased
compared to those in the splanchnic circulation, resulting in
central hypovolemia. Decreased blood volumes together with
arterial hypotension lead to baroreceptor activation of potent
vasoconstriction systems such as the sympathetic nervous sys-
tem and the renin angiotensin aldosterone system [128], result-
ing in water retention and plasma volume expansion. The
combination of an expanded plasma volume and a reduction in
peripheral vascular resistance leads to an increase in cardiac
output.
When portal hypertension persists, the heart results in a high
cardiac output syndrome: initial compensation occurs according
to the degree of individual cardiac output, followed by some
degree of cardiac insufficiency. The cardiac index is usually
higher than normal (greater than 4 L min−1 m2) but insufficient
to maintain arterial pressure in the face of progressive arterial
vasodilation [5]. Importantly, high cardiac output failure is
reversible once the initial cause leading to the high cardiac out-
put is treated. This reversal has also been observed in patients
with cirrhosis after liver transplantation [129, 130].
The hyperdynamic splanchnic circulation
As mentioned previously, the hyperdynamic splanchnic circula-
tion is central to the development of the syndrome. Although
it  is commonly recognized as a complication of cirrhosis, it
should be better conceptualized as a complication of portal
 51:  Pathophysiology of Portal Hypertension 665

Splanchnic Renal Systemic Pulmonary Brain

Circulation Circulation Circulation Circulation Circulation

* Acute liver failure

Vasodilation Vasoconstriction Vasodilation Vasodilation Vasodilation

Initiated by local Initiated by local Initiated by local
factors factors factors

• ↑Blood flow • Na/H2O retention • ↑Cardiac output • Arterial hypoxemia • Brain edema
• ↑Portal pressure • Hepatorenal • High output • Hepato-pulmonary
syndrome • Heart failure syndrome
• ↓Peripheral O2
utilization

Figure 51.4  Vasodilation: the source of all evils. *Chronic encephalopathy is associated with a reduced brain blood flow. The mechanism is prob-
ably similar to what is observed in the renal circulation. Modified from Iwakiri and Groszmann [5] and reproduced with permission of John
Wiley & Sons.

hypertension. The term “portal venous inflow” is often used for
splanchnic blood flow entering into the portal system to distin-
guish it from portal blood flow entering into the liver [77].
Portal hypertension is the only known pathophysiological situa-
tion in which portal blood flow entering into the portal system
(portal venous inflow) is different from portal blood flow enter-
ing into the liver. Portal venous inflow entering into the portal
system significantly increases, while portal blood flow entering
into the liver decreases, because portal blood escapes into porto-
systemic collaterals formed as a result of portal hypertension
[5]. Arterial vasodilation in the splanchnic vascular bed is
thought to contribute to this increased portal venous inflow
entering into the portal system and also helps to compensate for
the blood that should escape into portosystemic collaterals.
Thus, the most accepted concept is that arterial vasodilation
starts from the splanchnic vascular bed (i.e. intestinal microcir-
culation) and then proceeds to the whole splanchnic circulation)
and is the key factor that leads to the development of the hyper-
dynamic circulatory syndrome. The increase in portal pressure
itself triggers arterial vasodilation in the splanchnic vascular
bed [51, 93].
The hyperdynamic pulmonary circulation
The hyperdynamic circulation also affects the lungs. The pul-
monary vasodilation is associated with the hepatopulmonary
syndrome, one of the most severe complications of chronic liver
disease (Figure  51.4). Although the intrinsic mechanism that
triggers this syndrome into its full expression is not fully known,
local vasodilation mediated by several endothelial vasodilators,
including NO and CO, plays an important role [131]. Local fac-
tors in the pulmonary circulation may determine why only some
patients develop hepatopulmonary syndrome. High cardiac out-
put may also contribute to the severity of hepatopulmonary syn-
drome by increasing shear stress in the pulmonary vascular
endothelium as well as by shortening a pulmonary and tissue
transit time of red blood cells [132, 133].
The renal circulation
The hyperdynamic circulatory state indirectly influences the
renal circulatory bed (Figure 51.4). It is thought that splanchnic
arterial vasodilation in patients with portal hypertension results
in redistribution of the blood volume, leading to a reduction in
central blood volumes (i.e. a relative hypovolemic state). The
kidney responds to this hypovolemic state by renal arterial vaso-
constriction, a reduction in glomerular filtration, and retention
of sodium and water. The central hypovolemic state activates
signals that induce vasoconstrictive and volume retaining neu-
rohumoral conditions, thereby keeping the sodium and water
retentive state [134–136]. In patients with compensated cirrho-
sis, progressive systemic arterial vasodilation leads to an eleva-
tion of intravascular volume and cardiac output, so that arterial
perfusion pressure can be maintained. As disease conditions
progress, cardiac output continues to increase in response to
progressive arterial vasodilation. Eventually, cardiac response
666 THE LIVER:  REFERENCES
starts to fail to maintain perfusion pressure, renal blood flow
drops, and renal failure develops [137, 138]. This phenomenon
is known as hepatorenal syndrome. Treatment of arterial vasodi-
lation improves renal function [139, 140].
The cerebral circulation
The effects of hyperdynamic circulatory syndrome on the cere-
bral circulation are probably the most difficult to define
(Figure 51.4). Both an increase and decrease in cerebral blood
flow have been described in acute and chronic liver diseases,
respectively [141]. An increase in cerebral blood flow has been
mainly associated with acute liver failure, which could poten-
tially lead to the development of brain edema [142]. In contrast,
cerebral blood flow is decreased in chronic liver disease, and
this decrease runs in parallel with the above‐mentioned reduc-
tion in renal blood flow, suggesting that the mechanisms of
blood flow reductions in chronic liver disease may be similar in
these two organs [143].
CONCLUSION
The pathogenesis of portal hypertension is complex, because
portal hypertension involves not only the hepatic circulation, but
also the splanchnic and systemic circulations of the hyperdy-
namic circulatory state. Because of the disparate conditions of
vascular tone in the intrahepatic and extrahepatic circulations
(i.e. vasoconstriction in the intrahepatic circulation versus vaso-
dilation in the extrahepatic circulation), the organ/tissue or cell‐
specific modulation of vasodilator or vasoconstrictor molecules
is of paramount importance for therapeutic purposes.
REFERENCES
1. Iwakiri, Y., Shah, V., and Rockey, D.C. Vascular pathobiology in chronic
liver disease and cirrhosis – current status and future directions. J Hepatol,
2014;61(4):912–24.
2. Bosch, J., Groszmann, R.J., and Shah, V.H. Evolution in the understanding
of the pathophysiological basis of portal hypertension: How changes in para-
digm are leading to successful new treatments. J Hepatol, 2015;62(1
Suppl):S121–30.
3. Iwakiri, Y. The molecules: mechanisms of arterial vasodilatation observed in
the splanchnic and systemic circulation in portal hypertension. J Clin
Gastroenterol, 2007;41(3):S288–94.
4. Iwakiri, Y. Pathophysiology of portal hypertension. Clin Liver Dis,
2014;18(2):281–91.
5. Iwakiri, Y. and Groszmann, R.J. The hyperdynamic circulation of chronic
liver diseases: from the patient to the molecule. Hepatology, 2006;43(2
Suppl 1):S121–31.
6. Iwakiri, Y. Endothelial dysfunction in the regulation of cirrhosis and portal
hypertension. Liver Int, 2012;32(2):199–213.
7. Iwakiri, Y. and Groszmann, R.J. Vascular endothelial dysfunction in cirrho-
sis. J Hepatol, 2007;46(5):927–34.
8. Tetangco, E.P., Silva, R., and Lerma, E. Portal hypertension: etiology, evalu-
ation, and management. Dis Mon, 2016;62(12):411–26.
9. Cavallin, M., Fasolato, S., Marenco, S. et al. The Treatment of Hepatorenal
Syndrome. Dig Dis, 2015;33(4):548–54.
10. Garcia‐Tsao, G., Abraldes, J.G., Berzigotti, A. et  al. Portal hypertensive
bleeding in cirrhosis: Risk stratification, diagnosis, and management: 2016
practice guidance by the American Association for the study of liver dis-
eases. Hepatology, 2017;65(1):310–35.
11. Rockey, D.C. Cell and molecular mechanisms of increased intrahepatic
resistance and hemodynamic correlates. Sanyal AJ SV, editor. Totowa, NJ:
Humana Press Inc; 2005.
12. Pinzani, M., Vizzutti, F. Anatomy and vascular biology of the cells in the
portal circulation. Sanyal AJ SV, editor. Totowa, NJ: Humana Press Inc;
2005.
13. Bhathal, P.S. and Grossman, H.J. Reduction of the increased portal vascular
resistance of the isolated perfused cirrhotic rat liver by vasodilators. J
Hepatol, 1985;1(4):325–37.
14. Lee, Y.A., Wallace, M.C., and Friedman, S.L. Pathobiology of liver fibrosis:
a translational success story. Gut, 2015;64(5):830–41.
15. Trautwein, C., Friedman, S.L., Schuppan, D. et al. Hepatic fibrosis: Concept
to treatment. J Hepatol, 2015;62(1 Suppl):S15–24.
16. Shepro, D. and Morel, N.M. Pericyte physiology. FASEB, 1993;7(11):
1031–8.
17. Bergers, G. and Song, S. The role of pericytes in blood‐vessel formation and
maintenance. Neuro Oncol, 2005;7(4):452–64.
18. Franco, M., Roswall, P., Cortez, E. et al. Pericytes promote endothelial cell
survival through induction of autocrine VEGF‐A signaling and Bcl‐w
expression. Blood, 2011;118(10):2906–17.
19. LaBarbera, K.E., Hyldahl, R.D., O’Fallon, K.S. et al. Pericyte NF‐kappaB
activation enhances endothelial cell proliferation and proangiogenic cytokine
secretion in vitro. Physiol Rep, 2015;3(4).
20. Rockey, D.C., Boyles, J.K., Gabbiani, G. et al. Rat hepatic lipocytes express
smooth muscle actin upon activation in vivo and in culture. J Submicrosc
Cytol Pathol, 1992;24(2):193–203.
21. Rockey, D.C., Housset, C.N., and Friedman, S.L. Activation‐dependent con-
tractility of rat hepatic lipocytes in culture and in vivo. J Clin Invest,
1993;92(4):1795–804.
22. Rockey, D.C. Vascular mediators in the injured liver. Hepatology,
2003;37(1):4–12.
23. Rockey, D.C., Fouassier, L., Chung, J.J. et  al. Cellular localization of
endothelin‐1 and increased production in liver injury in the rat: potential for
autocrine and paracrine effects on stellate cells. Hepatology, 1998;27(2):
472–80.
24. Shao, R., Yan, W., and Rockey, D.C. Regulation of endothelin‐1 synthesis by
endothelin‐converting enzyme‐1 during wound healing. J Biol Chem,
1999;274(5):3228–34.
25. Rothermund, L., Leggewie, S., Schwarz, A. et al. Regulation of the hepatic
endothelin system in advanced biliary fibrosis in rats. Clin Chem Lab Med,
2000;38(6):507–12.
26. Yokomori, H., Oda, M., Ogi, M. et al. Enhanced expression of endothelin
receptor subtypes in cirrhotic rat liver. Liver, 2001;21(2):114–22.
27. Zhang, J.X., Pegoli, W., Jr., and Clemens, M.G. Endothelin‐1 induces
direct  constriction of hepatic sinusoids. Am J Physiol, 1994;266(4 Pt 1):
G624–32.
28. Feng, H.Q., Weymouth, N.D., and Rockey, D.C. Endothelin antagonism in
portal hypertensive mice: implications for endothelin receptor‐specific sign-
aling in liver disease. Am J Physiol Gastrointest Liver Physiol,
2009;297(1):G27–33.
29. Iizuka, M., Murata, T., Hori, M. et al. Increased contractility of hepatic stel-
late cells in cirrhosis is mediated by enhanced Ca2+‐dependent and Ca2+‐
sensitization pathways. Am J Physiol Gastrointest Liver Physiol, 2011;
300(6):G1010–21.
30. Wisse, E. An electron microscopic study of the fenestrated endothelial lining
of rat liver sinusoids. J Ultrastruct Res, 1970;31(1):125–50.
31. Bhunchet, E. and Fujieda, K. Capillarization and venularization of hepatic
sinusoids in porcine serum‐induced rat liver fibrosis: a mechanism to main-
tain liver blood flow. Hepatology, 1993;18(6):1450–8.
32. Horn, T., Christoffersen, P., and Henriksen, J.H. Alcoholic liver injury:
defenestration in noncirrhotic livers – a scanning electron microscopic study.
Hepatology, 1987;7(1):77–82.
33. Funyu, J., Mochida, S., Inao, M. et al. VEGF can act as vascular permeability
factor in the hepatic sinusoids through upregulation of porosity of endothe-
lial cells. Biochem Biophys Res Commun, 2001;280(2):481–5.
34. May, D., Djonov, V., Zamir, G. et  al. A transgenic model for conditional
induction and rescue of portal hypertension reveals a role of VEGF‐mediated
regulation of sinusoidal fenestrations. PloS One, 2011;6(7):e21478.
 51:  Pathophysiology of Portal Hypertension 667
35. DeLeve, L.D., Wang, X., Hu, L. et al. Rat liver sinusoidal endothelial cell
phenotype is maintained by paracrine and autocrine regulation. Am J Physiol
Gastrointest Liver Physiol, 2004;287(4):G757–63.
36. McGuire, R.F., Bissell, D.M., Boyles, J. et al. Role of extracellular matrix in
regulating fenestrations of sinusoidal endothelial cells isolated from normal
rat liver. Hepatology, 1992;15(6):989–97.
37. Svistounov, D., Warren, A., McNerney, G.P. et al. The relationship between
fenestrations, sieve plates and rafts in liver sinusoidal endothelial cells. PloS
One, 2012;7(9):e46134.
38. DeLeve, L.D. Liver sinusoidal endothelial cells in hepatic fibrosis.
Hepatology, 2015;61(5):1740–6.
39. McConnell, M. and Iwakiri, Y. Biology of portal hypertension. Hepatology
Int, 2018;12(Suppl 1):11–23.
40. Jarnagin, W.R., Rockey, D.C., Koteliansky, V.E. et al. Expression of variant
fibronectins in wound healing: cellular source and biological activity of the
EIIIA segment in rat hepatic fibrogenesis. J Cell Biol, 1994;127(6 Pt
2):2037–48.
41. Olsen, A.L., Sackey, B.K., Marcinkiewicz, C. et  al. Fibronectin extra
domain‐A promotes hepatic stellate cell motility but not differentiation into
myofibroblasts. Gastroenterology, 2012;142(4):928–37 e3.
42. Wang, R., Ding, Q., Yaqoob, U. et al. Exosome adherence and internalization
by hepatic stellate cells triggers sphingosine 1‐phosphate‐dependent migra-
tion. J Biol Chem, 2015;290(52):30684–96.
43. Deleve, L.D., Wang, X., and Guo, Y. Sinusoidal endothelial cells prevent rat
stellate cell activation and promote reversion to quiescence. Hepatology,
2008;48(3):920–30.
44. Marrone, G., Russo, L., Rosado, E. et  al. The transcription factor KLF2
mediates hepatic endothelial protection and paracrine endothelial‐stellate
cell deactivation induced by statins. J Hepatol, 2013;58(1):98–103.
45. Failli, P., De, F.R., Caligiuri, A. et  al. Nitrovasodilators inhibit platelet‐
derived growth factor‐induced proliferation and migration of activated
human hepatic stellate cells. Gastroenterology, 2000;119(2):479–92.
46. Routray, C., Liu, C., Yaqoob, U. et al. Protein kinase G signaling disrupts
Rac1‐dependent focal adhesion assembly in liver specific pericytes. Am J
Physiol Cell Physiol, 2011;301(1):C66–74.
47. Lee, J.S., Kang Decker, N., Chatterjee, S. et  al. Mechanisms of nitric oxide
interplay with Rho GTPase family members in modulation of actin membrane
dynamics in pericytes and fibroblasts. Am J Pathol, 2005;166(6):1861–70.
48. Langer, D.A., Das, A., Semela, D. et al. Nitric oxide promotes caspase‐inde-
pendent hepatic stellate cell apoptosis through the generation of reactive
oxygen species. Hepatology, 2008;47(6):1983–93.
49. Mittal, M.K., Gupta, T.K., Lee, F.Y. et  al. Nitric oxide modulates hepatic
vascular tone in normal rat liver. Am J Physiol, 1994;267(3 Pt 1):G416–22.
50. Shah, V., Haddad, F.G., Garcia‐Cardena, G. et al. Liver sinusoidal endothe-
lial cells are responsible for nitric oxide modulation of resistance in the
hepatic sinusoids. J Clin Invest, 1997;100(11):2923–30.
51. Abraldes, J.G., Iwakiri, Y., Loureiro‐Silva, M. et al. Mild increases in portal
pressure upregulate vascular endothelial growth factor and endothelial nitric
oxide synthase in the intestinal microcirculatory bed, leading to a hyperdy-
namic state. Am J Physiol Gastrointest Liver Physiol, 2006;290(5):G980–7.
52. Rockey, D.C. and Chung, J.J. Reduced nitric oxide production by endothelial
cells in cirrhotic rat liver: endothelial dysfunction in portal hypertension.
Gastroenterology, 1998;114(2):344–51.
53. Shah, V., Toruner, M., Haddad, F. et  al. Impaired endothelial nitric oxide
synthase activity associated with enhanced caveolin binding in experimental
cirrhosis in the rat. Gastroenterology, 1999;117(5):1222–8.
54. Fulton, D., Gratton, J.P., McCabe, T.J. et  al. Regulation of endothelium‐
derived nitric oxide production by the protein kinase Akt. Nature,
1999;399(6736):597–601.
55. Liu, S., Premont, R.T., and Rockey, D.C. G‐protein‐coupled receptor kinase
interactor‐1 (GIT1) is a new endothelial nitric‐oxide synthase (eNOS) inter-
actor with functional effects on vascular homeostasis. J Biol Chem,
2012;287(15):12309–20.
56. Liu, S., Premont, R.T., and Rockey, D.C. Endothelial nitric‐oxide synthase
(eNOS) is activated through G‐protein‐coupled receptor kinase‐interacting
protein 1 (GIT1) tyrosine phosphorylation and Src protein. J Biol Chem,
2014;289(26):18163–74.
57. Michel, J.B., Feron, O., Sacks, D. et al. Reciprocal regulation of endothelial
nitric‐oxide synthase by Ca2+‐calmodulin and caveolin. J Biol Chem,
1997;272(25):15583–6.
58. Abraldes, J.G., Albillos, A., Banares, R. et al. Simvastatin lowers portal pres-
sure in patients with cirrhosis and portal hypertension: a randomized con-
trolled trial. Gastroenterology, 2009;136(5):1651–8.
59. Abraldes, J.G., Rodriguez‐Vilarrupla, A., Graupera, M. et  al. Simvastatin
treatment improves liver sinusoidal endothelial dysfunction in CCl4 cirrhotic
rats. J Hepatol, 2007;46(6):1040–6.
60. Trebicka, J., Hennenberg, M., Laleman, W. et al. Atorvastatin lowers portal
pressure in cirrhotic rats by inhibition of RhoA/Rho‐kinase and activation of
endothelial nitric oxide synthase. Hepatology, 2007;46(1):242–53.
61. Ming, X.F., Viswambharan, H., Barandier, C. et al. Rho GTPase/Rho kinase
negatively regulates endothelial nitric oxide synthase phosphorylation
through the inhibition of protein kinase B/Akt in human endothelial cells.
Mol Cell Biol, 2002;22(24):8467–77.
62. Laufs, U., and Liao, J.K. Post‐transcriptional regulation of endothelial nitric
oxide synthase mRNA stability by Rho GTPase. J Biol Chem, 1998;
273(37):24266–71.
63. Dill, M.T., Rothweiler, S., Djonov, V. et  al. Disruption of Notch1 induces
vascular remodeling, intussusceptive angiogenesis, and angiosarcomas in
livers of mice. Gastroenterology, 2012;142(4):967–77 e2.
64. Corpechot, C., Barbu, V., Wendum, D. et  al. Hypoxia‐induced VEGF and
collagen I expressions are associated with angiogenesis and fibrogenesis in
experimental cirrhosis. Hepatology, 2002;35(5):1010–21.
65. Ehling, J., Bartneck, M., Wei, X. et  al. CCL2‐dependent infiltrating mac-
rophages promote angiogenesis in progressive liver fibrosis. Gut,
2014;63(12):1960–71.
66. Thabut, D. and Shah, V. Intrahepatic angiogenesis and sinusoidal remodeling
in chronic liver disease: new targets for the treatment of portal hypertension?
J Hepatol, 2010;53(5):976–80.
67. Wanless, I.R., Wong, F., Blendis, L.M. et al. Hepatic and portal vein throm-
bosis in cirrhosis: possible role in development of parenchymal extinction
and portal hypertension. Hepatology, 1995;21(5):1238–47.
68. Tripodi, A. Hemostasis abnormalities in cirrhosis. Curr Opin Hematol,
2015;22(5):406–12.
69. De Pietri, L., Bianchini, M., Montalti, R. et al. Thrombelastography‐guided
blood product use before invasive procedures in cirrhosis with severe coagu-
lopathy: a randomized, controlled trial. Hepatology, 2016;63(2):566–73.
70. Ordinas, A., Escolar, G., Cirera, I. et  al. Existence of a platelet‐adhesion
defect in patients with cirrhosis independent of hematocrit: studies under
flow conditions. Hepatology, 1996;24(5):1137–42.
71. Rubin, M.H., Weston, M.J., Langley, P.G. et al. Platelet function in chronic liver
disease: relationship to disease severity. Dig Dis Sci, 1979;24(3):197–202.
72. Lisman, T., Bongers, T.N., Adelmeijer, J. et  al. Elevated levels of von
Willebrand Factor in cirrhosis support platelet adhesion despite reduced
functional capacity. Hepatology, 2006;44(1):53–61.
73. Raparelli, V., Basili, S., Carnevale, R. et  al. Low‐grade endotoxemia and
platelet activation in cirrhosis. Hepatology, 2017;65(2):571–81.
74. Potze, W., Siddiqui, M.S., Boyett, S.L. et al. Preserved hemostatic status in
patients with non‐alcoholic fatty liver disease. J Hepatol, 2016;65(5):
980–7.
75. Chauhan, A., Adams, D.H., Watson, S.P. et al. Platelets: no longer bystanders
in liver disease. Hepatology, 2016;64(5):1774–84.
76. Chojkier, M. and Groszmann, R.J. Measurement of portal‐systemic shunting
in the rat by using gamma‐labeled microspheres. Am J Physiol, 1981;
240(5):G371–5.
77. Groszmann, R.J., Vorobioff, J., and Riley, E. Splanchnic hemodynamics in
portal‐hypertensive rats: measurement with gamma‐labeled microspheres.
Am J Physiol, 1982;242(2):G156–60.
78. Vorobioff, J., Bredfeldt, J.E., and Groszmann, R.J. Hyperdynamic circula-
tion in portal‐hypertensive rat model: a primary factor for maintenance of
chronic portal hypertension. Am J Physiol, 1983;244(1):G52–7.
79. Sikuler, E., Kravetz, D., and Groszmann, R.J. Evolution of portal hyperten-
sion and mechanisms involved in its maintenance in a rat model. Am J
Physiol, 1985;248(6 Pt 1):G618–25.
80. Sarin, S.K., Mosca, P., Sabba, C. et  al. Hyperdynamic circulation in a
chronic murine schistosomiasis model of portal hypertension. Hepatology,
1991;13(3):581–4.
81. Colombato, L.A., Albillos, A., and Groszmann, R.J. Temporal relationship
of peripheral vasodilatation, plasma volume expansion and the hyperdy-
namic circulatory state in portal‐hypertensive rats. Hepatology, 1992;
15(2):323–8.
668 THE LIVER:  REFERENCES
82. Lee, F.Y., Albillos, A., Colombato, L.A. et al. The role of nitric oxide in the
vascular hyporesponsiveness to methoxamine in portal hypertensive rats.
Hepatology, 1992;16(4):1043–8.
83. Sieber, C.C. and Groszmann, R.J. Nitric oxide mediates hyporeactivity to
vasopressors in mesenteric vessels of portal hypertensive rats.
Gastroenterology, 1992;103(1):235–9.
84. Lee, F.Y., Colombato, L.A., Albillos, A. et  al. N omega‐nitro‐L‐arginine
administration corrects peripheral vasodilation and systemic capillary
hypotension and ameliorates plasma volume expansion and sodium reten-
tion in portal hypertensive rats. Hepatology, 1993;17(1):84–90.
85. Sieber, C.C., Lopez‐Talavera, J.C., and Groszmann, R.J. Role of nitric
oxide in the in vitro splanchnic vascular hyporeactivity in ascitic cirrhotic
rats. Gastroenterology, 1993;104(6):1750–4.
86. Lopez‐Talavera, J.C., Merrill, W.W., and Groszmann, R.J. Tumor necrosis
factor alpha: a major contributor to the hyperdynamic circulation in prehe-
patic portal‐hypertensive rats. Gastroenterology, 1995;108(3):761–7.
87. Sieber, C.C., Lee, F.Y., and Groszmann, R.J. Long‐term octreotide treat-
ment prevents vascular hyporeactivity in portal‐hypertensive rats.
Hepatology, 1996;23(5):1218–23.
88. Shah, V., Wiest, R., Garcia‐Cardena, G. et al. Hsp90 regulation of endothe-
lial nitric oxide synthase contributes to vascular control in portal hyperten-
sion. Am J Physiol, 1999;277(2 Pt 1):G463–8.
89. Wiest, R., Das, S., Cadelina, G. et al. Bacterial translocation in cirrhotic
rats stimulates eNOS‐derived NO production and impairs mesenteric vas-
cular contractility. J Clin Invest, 1999;104(9):1223–33.
90. Wiest, R., Shah, V., Sessa, W.C. et al. NO overproduction by eNOS pre-
cedes hyperdynamic splanchnic circulation in portal hypertensive rats. Am
J Physiol, 1999;276(4 Pt 1):G1043–51.
91. Iwakiri, Y., Cadelina, G., Sessa, W.C. et al. Mice with targeted deletion of
eNOS develop hyperdynamic circulation associated with portal hyperten-
sion. Am J Physiol Gastrointest Liver Physiol, 2002;283(5):G1074–81.
92. Iwakiri, Y., Tsai, M.H., McCabe, T.J. et al. Phosphorylation of eNOS initi-
ates excessive NO production in early phases of portal hypertension. Am J
Physiol Heart Circ Physiol, 2002;282(6):H2084–90.
93. Tsai, M.H., Iwakiri, Y., Cadelina, G. et al. Mesenteric vasoconstriction trig-
gers nitric oxide overproduction in the superior mesenteric artery of portal
hypertensive rats. Gastroenterology, 2003;125(5):1452–61.
94. Wiest, R., Cadelina, G., Milstien, S. et al. Bacterial translocation up‐regu-
lates GTP‐cyclohydrolase I in mesenteric vasculature of cirrhotic rats.
Hepatology, 2003;38(6):1508–15.
95. Groszmann, R.J., Garcia‐Tsao, G., Bosch, J. et al. Beta‐blockers to prevent
gastroesophageal varices in patients with cirrhosis. N Engl J Med,
2005;353(21):2254–61.
96. Huang, H.C., Haq, O., Utsumi, T. et al. Intestinal and plasma VEGF levels
in cirrhosis: the role of portal pressure. J Cell Mol Med, 2012;16(5):
1125–33.
97. Iwakiri, Y. The molecules: mechanisms of arterial vasodilatation observed
in the splanchnic and systemic circulation in portal hypertension. J Clin
Gastroenterol, 2007;41(10 Suppl 3):S288–94.
98. Feletou, M. and Vanhoutte, P.M. Endothelium‐dependent hyperpolariza-
tions: past beliefs and present facts. Ann Med, 2007;39(7):495–516.
99. Iwakiri, Y. The systemic and splanchnic circulation, in Chronic Liver
Failure, Mechanisms and Management, (eds. Gines, P. et  al.), Humana
Press, New York, 2011, pp. 305–21.
100. Fernandez, M., Mejias, M., Angermayr, B. et al. Inhibition of VEGF recep-
tor‐2 decreases the development of hyperdynamic splanchnic circulation
and portal‐systemic collateral vessels in portal hypertensive rats. J Hepatol,
2005;43(1):98–103.
101. Moleda, L., Trebicka, J., Dietrich, P. et al. Amelioration of portal hyperten-
sion and the hyperdynamic circulatory syndrome in cirrhotic rats by
­neuropeptide Y via pronounced splanchnic vasoaction. Gut, 2011;60(8):
1122–32.
102. Trebicka, J., Leifeld, L., Hennenberg, M. et  al. Hemodynamic effects of
urotensin II and its specific receptor antagonist palosuran in cirrhotic rats.
Hepatology, 2008;47(4):1264–76.
103. Kemp, W., Krum, H., Colman, J. et  al. Urotensin II: a novel vasoactive
mediator linked to chronic liver disease and portal hypertension. Liver Int,
2007;27(9):1232–9.
104. Hennenberg, M., Trebicka, J., Kohistani, A.Z. et al. Vascular hyporespon-
siveness to angiotensin II in rats with CCl(4)‐induced liver cirrhosis. Eur J
Clin Invest, 2009;39(10):906–13.
105. Chu, C.J., Wu, S.L., Lee, F.Y. et al. Splanchnic hyposensitivity to glypres-
sin in a haemorrhage/transfused rat model of portal hypertension: role of
nitric oxide and bradykinin. Clin Sci (Lond), 2000;99(6):475–82.
106. Chen, C.T., Chu, C.J., Lee, F.Y. et al. Splanchnic hyposensitivity to glypres-
sin in a hemorrhage‐transfused common bile duct‐ligated rat model of portal
hypertension: role of nitric oxide and bradykinin. Hepatogastroenterology,
2009;56(94–95):1261–7.
107. Heinemann, A., Wachter, C.H., Fickert, P. et al. Vasopressin reverses mes-
enteric hyperemia and vasoconstrictor hyporesponsiveness in anesthetized
portal hypertensive rats. Hepatology, 1998;28(3):646–54.
108. Song, D., Liu, H., Sharkey, K.A. et al. Hyperdynamic circulation in portal‐
hypertensive rats is dependent on central c‐fos gene expression. Hepatology,
2002;35(1):159–66.
109. Coll, M., Martell, M., Raurell, I. et al. Atrophy of mesenteric sympathetic
innervation may contribute to splanchnic vasodilation in rat portal hyper-
tension. Liver Int, 2010;30(4):593–602.
110. Ezkurdia, N., Coll, M., Raurell, I. et al. Blockage of the afferent sensitive
pathway prevents sympathetic atrophy and hemodynamic alterations in rat
portal hypertension. Liver Int, 2012;32(8):1295–305.
111. Fernandez‐Varo, G., Ros, J., Morales‐Ruiz, M. et al. Nitric oxide synthase
3‐dependent vascular remodeling and circulatory dysfunction in cirrhosis.
Am J Pathol, 2003;162(6):1985–93.
112. Fernandez‐Varo, G., Morales‐Ruiz, M., Ros, J. et al. Impaired extracellular
matrix degradation in aortic vessels of cirrhotic rats. J Hepatol,
2007;46(3):440–6.
113. Sumanovski, L.T., Battegay, E., Stumm, M. et al. Increased angiogenesis in
portal hypertensive rats: role of nitric oxide. Hepatology, 1999;29(4):
1044–9.
114. Sieber, C.C., Sumanovski, L.T., Stumm, M. et al. In vivo angiogenesis in
normal and portal hypertensive rats: role of basic fibroblast growth factor
and nitric oxide. J Hepatol, 2001;34(5):644–50.
115. Groszmann, R.J., Kotelanski, B., and Cohn, J.N. Different patterns of
porta‐systemic shunting in cirrhosis of the liver studied by an indicator
dilution technique. Acta Gastroenterol Latinoam, 1971;3(3):111–6.
116. Bosch, J., Pizcueta, P., Feu, F. et al. Pathophysiology of portal hyperten-
sion. Gastroenterol Clin North Am, 1992;21(1):1–14.
117. Fernandez, M., Vizzutti, F., Garcia‐Pagan, J.C. et al. Anti‐VEGF receptor‐2
monoclonal antibody prevents portal‐systemic collateral vessel formation
in portal hypertensive mice. Gastroenterology, 2004;126(3):886–94.
118. Geerts, A.M., De Vriese, A.S., Vanheule, E. et al. Increased angiogenesis
and permeability in the mesenteric microvasculature of rats with cirrhosis
and portal hypertension: an in vivo study. Liver Int, 2006;26(7):889–98.
119. Van Steenkiste, C., Geerts, A., Vanheule, E. et al. Role of placental growth
factor in mesenteric neoangiogenesis in a mouse model of portal hyperten-
sion. Gastroenterology, 2009;137(6):2112–24 e1–6.
120. Fernandez, M., Mejias, M., Garcia‐Pras, E. et al. Reversal of portal hyper-
tension and hyperdynamic splanchnic circulation by combined vascular
endothelial growth factor and platelet‐derived growth factor blockade in
rats. Hepatology, 2007;46(4):1208–17.
121. Tiani, C., Garcia‐Pras, E., Mejias, M. et  al. Apelin signaling modulates
splanchnic angiogenesis and portosystemic collateral vessel formation in
rats with portal hypertension. J Hepatol, 2009;50(2):296–305.
122. Mejias, M., Garcia‐Pras, E., Tiani, C. et al. Beneficial effects of sorafenib
on splanchnic, intrahepatic, and portocollateral circulations in portal hyper-
tensive and cirrhotic rats. Hepatology, 2009;49(4):1245–56.
123. Reiberger, T., Angermayr, B., Schwabl, P. et  al. Sorafenib attenuates the
portal hypertensive syndrome in partial portal vein ligated rats. J Hepatol,
2009;51(5):865–73.
124. Huang, H.C., Wang, S.S., Hsin, I.F. et al. Cannabinoid receptor 2 agonist
ameliorates mesenteric angiogenesis and portosystemic collaterals in cir-
rhotic rats. Hepatology, 2012;56(1):248–58.
125. Genecin, P., Polio, J., and Groszmann, R.J. Na restriction blunts expansion
of plasma volume and ameliorates hyperdynamic circulation in portal hyper-
tension. Am J Physiol Gastrointest Liver Physiol, 1990;259(22):G498–503.
126. Morgan, J.S., Groszmann, R.J., Rojkind, M., and Enriquez, R.
Hemodynamic mechanisms of emerging portal hypertension caused by
schistosomiasis in the hamster. Hepatology, 1990;11(1):98–104.
127. Moller, S. and Henriksen, J.H. Cardiovascular complications of cirrhosis.
Gut, 2008;57(2):268–78.
128. Schrier, R.W. Water and sodium retention in edematous disorders: role of
vasopressin and aldosterone. Am J Med, 2006;119(7 Suppl 1):S47–53.
 51:  Pathophysiology of Portal Hypertension 669
129. Aller, R., de Luis, D.A., Moreira, V. et al. The effect of liver transplantation
on circulating levels of estradiol and progesterone in male patients:
­parallelism with hepatopulmonary syndrome and systemic hyperdynamic
circulation improvement. J Endocrinol Invest, 2001;24(7):503–9.
130. Angus, P.W., Vaughan, R.B., and Chin‐Dusting, J.P. Responses to endothe-
lin‐1 in patients with advanced cirrhosis before and after liver transplanta-
tion. Gut, 2004;53(5):773.
131. Fallon, M.B. Mechanisms of pulmonary vascular complications of liver
disease: hepatopulmonary syndrome. J Clin Gastroenterol, 2005;39(4
Suppl 2):S138–42.
132. Katsuta, Y., Honma, H., Zhang, X.J. et al. Pulmonary blood transit time and
impaired arterial oxygenation in patients with chronic liver disease. J
Gastroenterol, 2005;40(1):57–63.
133. Agusti, A.G., Roca, J., Bosch, J. et  al. Effects of propranolol on arterial
oxygenation and oxygen transport to tissues in patients with cirrhosis. Am
Rev Respir Dis, 1990;142(2):306–10.
134. Moller, S., Henriksen, J.H., and Bendtsen, F. Central and noncentral blood
volumes in cirrhosis: relationship to anthropometrics and gender. Am J
Physiol Gastrointest Liver Physiol, 2003;284(6):G970–9.
135. Shapiro, M.D., Nicholls, K.M., Groves, B.M. et  al. Interrelationship
between cardiac output and vascular resistance as determinants of
effective arterial blood volume in cirrhotic patients. Kidney Int,
­
1985;28(2):206–11.
136. Schrier, R.W., Arroyo, V., Bernardi, M. et al. Peripheral arterial vasodila-
tion hypothesis: a proposal for the initiation of renal sodium and water
retention in cirrhosis. Hepatology, 1988;8(5):1151–7.
137. Cohn, J.N. Renal hemodynamic alterations in liver disease. Perspect
Nephrol Hypertens, 1976;3:255–34.
138. Ruiz‐del‐Arbol, L., Monescillo, A., Arocena, C. et al. Circulatory function
and hepatorenal syndrome in cirrhosis. Hepatology, 2005;42(2):439–47.
139. Cohn, J.N., Tristani, F.E., and Khatri, I.M. Systemic vasoconstrictor and
renal vasodilator effects of PLV‐2 (octapressin) in man. Circulation,
1968;38(1):151–7.
140. Lenz, K., Hortnagl, H., Druml, W. et  al. Beneficial effect of 8‐ornithin
vasopressin on renal dysfunction in decompensated cirrhosis. Gut,
1989;30(1):90–6.
141. Blei, A.T. Monitoring cerebral blood flow. A useful clinical tool in acute
liver failure? Liver Transplantation, 2005;11(11):1320–22.
142. Vaquero, J., Chung, C., and Blei, A.T. Cerebral blood flow in acute liver
failure: a finding in search of a mechanism. Metab Brain Dis,
2004;19(3–4):177–94.
143. Guevara, M., Bru, C., Gines, P. et al. Increased cerebrovascular resistance
in cirrhotic patients with ascites. Hepatology, 1998;28(1):39–44.
 Non‐alcoholic Fatty Liver
52 Disease: Mechanisms
and Treatment
Yaron Rotman and Devika Kapuria
Liver Energy and Metabolism Section, Liver Diseases Branch, National Institute of Diabetes and Digestive
and Kidney Diseases, National Institutes of Health, Bethesda, MD, USA
SUMMARY
Non‐alcoholic fatty liver disease (NAFLD), the excess hepatic
accumulation of fat in the form of triglycerides (TG), has
become a global epidemic. The disease spans across a spectrum
ranging from simple steatosis to non‐alcoholic steatohepatitis
(NASH), where the fat accumulation is accompanied by inflam-
mation and injury, and to cirrhosis and hepatocellular carci-
noma. NAFLD is associated with other metabolic diseases
including type 2 diabetes and coronary artery disease and there
is increasing evidence of a link between NAFLD and several
cancers. In this chapter, we review the epidemiology of NAFLD,
discuss mechanisms of injury and progression, the natural his-
tory of the disease, and management options for patients.
DEFINITIONS
NAFLD is defined as the presence of excess hepatic TG by
imaging or histology in the absence of secondary causes of
hepatic fat accumulation [1]. A liver TG content greater than
5.5% is considered abnormal, a threshold that was obtained
from population imaging studies [2] and is consistent with data
from biochemical assays of autopsy samples [3]. There are two
forms of NAFLD, which can only be distinguished histologi-
cally. Non‐alcoholic fatty liver (NAFL) is defined as the pres-
ence of steatosis without evidence for hepatocellular damage
(ballooning), whereas the presence of hepatocyte injury (or bal-
looning) is required to diagnose NASH. Both NAFL and NASH
may present with or without fibrosis.
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
EPIDEMIOLOGY
Prevalence of NAFLD
NAFLD is the most common liver disorder in the Western world
and its prevalence appears to be increasing [4], concordant with the
rise in metabolic syndrome, obesity, and type 2 diabetes. Estimates
of disease prevalence vary depending on the method used to detect
it. In the NHANES III survey (1988–1994) the prevalence of unex-
plained aminotransferase elevation, thought to represent NAFLD,
was 5.4% [5] while in the 1999–2002 NHANES, the rate of ele-
vated transaminases was 9.8% [6]. More recently, Takyar et al. [7]
in an analysis of 3160 “healthy” volunteers reported a 27.9% preva-
lence of presumed NAFLD, based on elevated transaminases and
body mass index (BMI). However, it is well known that normal
aminotransferases do not preclude the presence of NAFLD [8].
This was demonstrated when the same 1988–1994 NHANES III
cohort was assessed by ultrasonography and the prevalence of stea-
tosis was found to be 21.4% [9]. In a later study using ultrasonog-
raphy between 2007–2010 [10], 46% of 328 subjects had evidence
of steatosis. Finally, using 1H‐magnetic resonance spectroscopy
(MRS) in the multiethnic Dallas heart study, steatosis was found in
one third of the study cohort [11].
There is a wide variation in the global prevalence and report-
ing of NAFLD. A recent meta‐analysis [4] that included
57 worldwide studies reported a pooled global NAFLD preva-
lence of 25.24%. The highest prevalence was reported from
South America (30.45%) and the Middle East (31.79%). The
lowest prevalence was reported in Africa (13.48%). Asia,
Europe, and North America respectively have a pooled NAFLD
prevalence of 27.37%, 23.71% and 24.13% [12].
 52:  Non‐alcoholic Fatty Liver Disease: Mechanisms and Treatment 671
A similar trend of increasing NAFLD prevalence is seen in
pediatric populations, where prevalence rose from 3.9% in
1988–1994 to 10.7% in 2007–2010 [13]. An autopsy series span-
ning 1993–2003 estimates the prevalence of pediatric NAFLD as
9.6%, with a prevalence of 38% in obese children [14].
Prevalence of NASH
In contrast to NAFLD, that can be identified reliably through
imaging modalities, the diagnosis of NASH remains a histologi-
cal one. Although several non‐invasive tests and formulas were
suggested as a method to differentiate NAFL from NASH
(­discussed below), none is sufficiently reliable to allow accurate
diagnosis. As a liver biopsy is often not feasible or clinically‐
required in most patient cohorts, estimating the true incidence
of NASH can be complex. In a recent meta‐analysis [4], the
pooled prevalence of NASH among NAFLD patients that under-
went a liver biopsy was 59.1%. However, this is likely overesti-
mating true prevalence due to selection bias. In a study that
performed prospective research liver biopsies in all subjects
with NAFLD irrespective of the clinical indication 29.9% of
subjects with NAFLD were found to have NASH [10].
Extrapolating from these data, the overall prevalence of NASH
is estimated at 1.5–6.5% [4]. Consistent with that, in a large
cohort of obese subjects who underwent bariatric surgery
(expected to be enriched for NAFLD), 7.5% were found to have
biopsy‐proven NASH [15].
NAFLD and the metabolic syndrome
NAFLD often occurs together with obesity, diabetes, and the
metabolic syndrome. There is a clear correlation between BMI
and hepatic TG content [11]. The prevalence of obesity in
NAFLD subjects is reported at 37–71% [4, 16] and over 90% of
severely obese patients undergoing bariatric surgery were found
to have NAFLD. Similarly, up to 33% of subjects with NAFLD
have been found to have type 2 diabetes, whereas 70% of
patients with diabetes have NAFLD; patients with both diabetes
and NAFLD have an increased risk of NASH, fibrosis [17], and
hepatocellular carcinoma [18]. Almost half of NAFLD and up to
71% of all NASH patients have been found to have the meta-
bolic syndrome [4, 19].
Given the co‐occurrence of NAFLD and metabolic abnormali-
ties, it is difficult to ascertain from cross‐sectional studies whether
NAFLD is a consequence of obesity, diabetes, and metabolic syn-
drome, or vice versa. In a longitudinal study [20], subjects with
no NAFLD at baseline were more likely to develop NAFLD after
7 years of follow‐up if they had higher BMI, insulin resistance, or
presence of metabolic syndrome at baseline. Similarly, NAFLD
at baseline was associated with incident diabetes after 6 years of
follow‐up [21]. Thus, a bidirectional association exists between
NAFLD and diabetes and the metabolic syndrome.
Additional risk factors
Longitudinal studies show a distinctly increased incidence of
NAFLD in males compared to females. In women, the incidence
of NAFLD is higher in menopausal and post‐menopausal
women compared to pre‐menopausal women [20, 22].
Race and ethnicity are major risk factors for NAFLD with a
disproportionately higher prevalence in Hispanics and relative
protection in African‐Americans compared to non‐Hispanic
Caucasians [4, 11]. In Asians, NAFLD is being increasingly
­recognized even in subjects with normal body mass indices.
GENETIC EPIDEMIOLOGY
The prevalence and severity of NAFLD have a strong heritable
component, as seen in twin [23] and family [24] studies, and
reflected by the strong racial impact [11]. Genome‐wide asso-
ciation studies (GWAS) and in recent years, sequencing
approaches, identified several single nucleotide polymorphism
(SNPs) that are associated with the disease. Of most interest are
variants that associate with NAFLD independently of estab-
lished risk factors such as obesity or diabetes, reflecting a true
hepatic risk.
PNPLA3
The non‐synonymous SNP rs738409 (c.444C>G, p.I148M) in
the patatin‐like phospholipase domain containing‐3 (PNPLA3)
gene encoding the adiponutrin protein has consistently shown
a strong association with NAFLD. Initially identified to be
associated in the general population with hepatic TG [25], and
liver enzymes [26], PNPLA3‐I148M has been subsequently
associated with histological severity of NAFLD and NASH
and was shown to promote inflammation, injury and fibrosis in
addition to its impact on liver fat content [27, 28]. The high‐
risk allele is also associated with earlier presentation of dis-
ease in pediatric patients [27], risk of hepatocellular carcinoma
[29], alcoholic liver disease [30], and has also been shown to
impact other liver diseases, suggesting it affects a common
injury pathway.
The mechanism by which the I148M mutation affects hepatic
lipid metabolism is not clear. Adiponutrin is present in the endo-
plasmic reticulum and lipid droplet membranes of hepatocytes
and adipose cells and demonstrates both TG hydrolase and acyl-
transferase activities, although this may not be its physiological
function in vivo. The I148M mutation impairs ubiquination,
leading to adiponutrin accumulation on hepatic lipid droplets
[31] and modulation of the droplet TG and phospholipid com-
position [32]. Beyond its function in hepatocytes adiponutrin
has retinyl‐palmitate lipase activity in stellate cells which may
explain its link to hepatic fibrosis [33].
TM6SF2
A SNP in the transmembrane 6 superfamily member 2 (TM6SF2)
gene, rs58542926 (c.499A>G, p.E167K), was found to be asso-
ciated with NAFLD [34]. The minor allele, associated with
increased risk of NAFLD, is also associated with decreased
serum LDL‐cholesterol and TG and with decreased risk for
myocardial infarction [35]. Inhibition of TM6SF2 impairs the
lipidation of nascent very low‐density lipoprotein (VLDL) in
the liver, resulting in decreased TG secretion and increased cel-
lular TG accumulation [36, 37].
672 THE LIVER:  MECHANISMS OF HEPATIC INJURY AND DEVELOPMENT OF STEATOHEPATITIS
HSD17B13
Several SNPs in 17‐beta hydroxysteroid dehydrogenase 13
(HSD17B13) show a strong association with NAFLD‐­associated
liver injury. A splice‐site SNP, rs72613567, generates alterna-
tive loss‐of‐function isoforms that associate with protection
from hepatocyte injury, inflammation and fibrosis but con-
versely with increased steatosis in NAFLD subjects [38, 39].
Other loss‐of‐function variants in the HSD17B13 gene have
been identified, with a similar protective effect on injury
[39, 40]. The gene product is a hepatic lipid‐droplet protein with
retinol dehydrogenase activity in vitro and in vivo [39].
Several other genes were associated with histological f­ eatures
of NAFLD, including MBOAT7 and GCKR [41]. In addition,
epigenetic changes also seem to play a role in disease pathogen-
esis, with data suggesting that methylation of mitochondrial
DNA is associated with histological severity [42].
PATHOGENESIS OF NAFLD
MECHANISMS OF FAT ACCUMULATION
The hallmark of NAFLD is the accumulation of TG in hepato-
cytes. The building blocks of TG are fatty acids (FA), which can
be delivered to the hepatocyte as non‐esterified fatty acids
(NEFA) from adipose tissue lipolysis or as part of chylomicron
TG from a dietary source, and can also be formed within the
hepatocyte through the de novo lipogenesis pathway. Hepatic
TG and FA can be secreted from the liver in VLDL, be oxidized
through β‐oxidation or metabolized to other derivatives. Fatty
liver reflects an imbalance between these input and output path-
ways (Figure 52.1).
Adipose tissue NEFA
In NAFLD, the majority of FA in hepatic TG are derived from
circulating NEFA (59%), while newly‐formed FA contribute
26% and dietary fat 15% [43]. Most circulating NEFA derive
from adipocyte lipolysis (82% in the fasting state and 62% in
the fed state), suggesting the adipocyte is the main source of
hepatic FA. Adipose tissue insulin resistance (IR) is broadly
found in NAFLD, manifesting as failure to suppress postpran-
dial lipolysis and leads to increased delivery of NEFA to the
liver [44]. Interestingly, the concentration of intrahepatic TG in
NAFLD [45] is strongly associated with adipose tissue IR but
not with hepatic IR. Furthermore, the degree of adipose IR has
been shown to be worse in NASH compared to NAFLD [46],
and in NASH patients to be associated with the fibrosis score.
Overall, this demonstrates the important role of adipocyte dys-
function and excess delivery of adipocyte‐derived NEFA to the
liver in the pathogenesis of hepatic TG accumulation.
De novo lipogenesis
De novo lipogenesis (DNL), the formation of FA from acetyl‐
CoA, is highly regulated and key enzymes in the pathway are
upregulated in NAFLD. In healthy lean individuals, DNL
contributes less than 5% of the total TG synthesis [47]. However,
in NAFLD patients, DNL is markedly increased and accounts
for nearly a quarter of the FA in hepatic TG [48]. The presence
of increased hepatic DNL, an insulin‐driven process, despite
hepatic IR suggests the presence of selective IR, either through
impairment of specific signaling pathways [49] or as a response
to altered nutrient fluxes [50]. Thus, the liver in NAFLD is
exposed to increased levels of NEFA derived from dysregulated
adipose tissue lipolysis and hepatic lipogenesis.
VLDL excretion
The excess influx of NEFA to the liver, and resultant excess
formation of hepatic TG are accompanied by increase in VLDL
and VLDL‐TG secretion. However, this increase is limited by
the ability to secrete ApoB100 and is not sufficient to balance
the rate of intrahepatic TG production [51].
β‐oxidation
There are conflicting data regarding the role of hepatic mito-
chondrial β‐oxidation in accumulation of liver TG, where some
studies suggest an increase in hepatic FA oxidation in individu-
als with NAFLD [52] while other studies show a decrease [53].
This discrepance may be due, in part, to difference in study
methodologies.
MECHANISMS OF HEPATIC INJURY
AND DEVELOPMENT
OF STEATOHEPATITIS
Two‐hit versus multi‐hit hypotheses
Steatosis is common but NASH is only seen in a subset of
patients. Similarly, steatosis is relatively easy to generate in
murine models but recapitulating NASH in them has been far
more difficult. Thus, the mechanisms that underly the transition
from steatosis to steatohepatitis need to be elucidated. Initially a
two‐hit hypothesis was proposed [54]. The accumulation of
hepatic TG was suggested as a first hit, sensitizing the liver to
further insults; a second hit would be required to drive the injury
process, mainly through induction of lipid peroxidation. Potential
suggested second hits were medications, gut‐derived endotoxin,
iron overload, and various inducers of oxidative stress. In recent
years it has become clear that a two‐hit model is unable to explain
the complexity of NAFLD and a more complex, “multi‐hit” or
“continuous‐hit” hypothesis has been formed. The major driving
process of hepatocyte injury is thought to be lipotoxicity from
free FA or their derivatives [55], with subsequent activation of
inflammatory responses and fibrosis. Genetic and nutritional fac-
tors are modulators of this process.
Lipotoxicity
As previously discussed, there is an increase in net hepatic FA
flux in NAFLD. Shunting of excess FA to form TG is an impor-
tant defense mechanism against lipotoxicity, which probably
 52:  Non‐alcoholic Fatty Liver Disease: Mechanisms and Treatment 673

(a)
VLDL

Acetyl-CoA
Glucose

TG
β-oxidation
FA
Glucose

CM

NEFA

(b)
VLDL

Acetyl-CoA
Glucose

TG
β-oxidation
FA
Glucose

CM

NEFA

Figure 52.1  Mechanisms of hepatic fat accumulation. (a) Fatty acids in the liver reflect a balance between input and disposal. Input pathways
include delivery of non‐esterified fatty acids from adipose tissue and triglycerides from chylomicrons and generation of fatty acids from acetyl‐CoA
through de novo lipogenesis (double arrow). Disposal includes oxidation or generation of triglycerides. Triglycerides can be secreted in VLDL or
hydrolyzed to fatty acids. (b) Adipose tissue insulin resistance in NAFLD causes an increase in delivery of non‐esterified fatty acids to the liver, while
hepatic and muscle insulin resistance cause hyperglycemia and increased availability of acetyl‐CoA. De novo lipogenesis is driven by hyperinsuline-
mia generating a net increase in fatty acids and increased shunting to form triglycerides. Secretion of VLDL particles is increased but not sufficient
to compensate. CM, chylomicrons; FA, fatty acids; NEFA, non‐esterified fatty acids; TG, triglycerides; VLDL, very low‐density lipoprotein.

explains the relatively benign phenotype of NAFL. In fact, TG
are likely inert and not causing injury by themselves as high-
lighted by animal studies in which inhibition of TG formation
has led to increased liver damage despite the inability to gener-
ate steatosis [56]. When this mechanism is overwhelmed, excess
FA, especially saturated ones, and downstream metabolites such
as ceramides, diacylglycerols, and lysophosphatidylcholine
(LPC) accumulate and initiate a cascade of cell injury and death,
inflammation, and resultant fibrosis [55].
Oxidative and endoplasmic reticulum stress
Oxidative stress has been strongly implicated in the pathogenesis
of NASH [57]. The increased hepatic metabolic load drives an
adaptive process with increased capacity of hepatic mitochon-
dria for oxidative phosphorylation [58]; however, in NASH this
adaptation is impaired, possibly as a result of lipotoxic injury to
the mitochondria, resulting in increased formation of reactive
oxygen species (ROS) and generation of oxidative stress [59].
Increased microsomal and peroxisomal oxidation of FA also
contributes to the formation of excess ROS [60]. Uncontrolled
ROS generate injury to the membranes and DNA and lead to the
formation of highly toxic lipid peroxidation products, especially
malondialdehyde and 4‐hydroxynonenal [61].
Endoplasmic reticulum (ER) stress has been demonstrated in
NAFLD [62], likely due to lipotoxic injury from saturated FA, and
leads to activation of the unfolded protein response. As with mito-
chondria, NASH appears to involve an overwhelming of this adap-
tive response, and chronic ER stress by itself induces increased
DNL and formation of ROS, forming a vicious cycle [63]. When
overwhelmed, ER stress activates IkB and c‐Jun N‐terminal kinases,
initiating a cascade of proinflammatory and pro‐apoptotic responses.
674 THE LIVER:  NATURAL HISTORY OF NAFLD
Extension of damage beyond the hepatocyte
The lipotoxic damage to hepatocytes activates regulated death
pathways in NASH, predominantly apoptosis [64], with activa-
tion of both intrinsic and extrinsic pathways. In addition, there
appears to be a population of stressed cells which undergo suble-
thal injury [65]; the histological correlate of these cells are the
ballooned hepatocytes seen in NASH livers. The apoptotic and
ballooned hepatocytes are releasing a myriad of signals includ-
ing sonic hedgehog [66] and, in addition, these cells are releasing
extracellular vesicles [65] containing a variety of signaling mol-
ecules including ceramides, CXCL10, TRAIL, miRNA, and
mitochondrial DNA [67]. These act as chemokines and damage‐
associated molecular patterns (DAMPs), which, together with
pathogen‐associated molecular patterns (PAMPs) derived from
gut bacteria, are recognized by pattern recognition receptors
(PRRs) including toll‐like receptors (TLR) and NOD‐like recep-
tors (NLR). Of the various TLRs and NLRs, the ones mostly
implicated in NASH pathogenesis are TLR4 and the NLRP3
inflammasome, respectively. The DAMP/PAMP‐PRR cascade
leads to activation of cells of the innate immune system [68].
Immune cell activation
Kupffer cells (KC), the liver‐resident macrophages, are impli-
cated in the initiation of the inflammatory process in NASH.
Recognition of PAMP/DAMP by PRRs, as well as a direct effect
of free fatty acids leads to KC activation and secretion of
cytokines such as tumor necrosis factor (TNF) and interleukin‐1β
and chemokines such as C‐C motif ligand 2 (CCL2) and CCL5.
These, in turn, lead to recruitment and activation of circulating
monocyte‐derived macrophages to the liver, and amplification
of the inflammatory process [69]. Furthermore, KC themselves
are major producers of ROS, further propagating the damage
cascade. Neutrophils are also recruited to the liver by cytokines
and DAMPs/PAMPs and further contribute to the inflammatory
process through their myeloperoxidase activity [70]. Targeted
depletion of KC in mouse models or genetic disruption of TLR4,
the CCL2‐CCR2 axis or neutrophil enzymes, all lead to amelio-
ration of experimental NASH, confirming the important role of
these innate immune cells [69]. There is also emerging data
regarding the role of additional immune cells, including natural
killer (NK) and NKT cells, regulatory T cells, and mucosal
associated invariant T (MAIT) cells [71].
Microbiome
In recent years, multiple studies have demonstrated a difference
in the composition of gut microbial populations between sub-
jects with NAFLD and controls, and between subjects with
NASH and those with NAFL. However, findings have not been
consistent across studies, limiting the ability to generalize those
findings and utilize them to establish a plausible mechanism
[72]. Stronger evidence supporting the role of dysbiosis in
NAFLD comes from animal studies. In a pivotal experiment, the
development of fatty liver on a steatogenic diet in germ‐free
mice was modulated by the microbiome implanted into these
mice [73] and steatosis could be transmitted by co‐housing of
mice with different microbiota [74].
The gut microbiota can affect NAFLD through several mech-
anisms. Disruption or leakage across the gut barrier can lead to
increased hepatic exposure to lipopolysaccharide (LPS) and
other microbial products that act as PAMPs and activate innate
immune responses in the liver, as previously described [75].
Metabolites derived from the interaction of the gut microbiota
with gut content such as trimethylamine (TMA), a metabolite of
choline, short‐chain fatty acids, amino acid metabolites, and
endogenous ethanol, can activate signaling pathways in the liver
and modulate NAFLD and NASH. Finally, gut bacteria metabo-
lize and modify the components of the bile acid pool; this in turn
can affect the activation of intestinal and hepatic farnesoid X
receptor (FXR) and modulate the development of steatosis and
steatohepatitis [76]. Furthermore, microbial metabolism of bile
acids can modify activation of hepatic NKT cells and their anti-
tumor activity, suggesting a link between the gut microbiome
and risk for liver cancer [77]. However, these mechanistic asso-
ciations have to be confirmed in human studies.
NATURAL HISTORY OF NAFLD
NAFLD is a slowly progressing disease, and will affect patients
over a span of decades, if not through their entire lifetime. Although
often asymptomatic, NAFLD is associated with increased overall
and liver‐associated mortality.
Hepatic outcomes of NAFLD
Symptoms of NAFLD, if any, are subtle and nonspecific and
may include hepatic discomfort and fatigue. More serious
symptoms and complications, such as hepatic decompensation,
need for liver transplant, and liver‐related mortality, are typically
associated with progression to cirrhosis [78]. As such, the
­presence of fibrosis is the main determinant of eventual hepatic
outcomes of NAFLD [79–81]. Fibrosis progresses at a variable
rate and can even regress spontaneously [82, 83], likely reflect-
ing the highly dynamic and fluctuating nature of the underlying
disease. In paired‐biopsy series, the most important predictor of
progression of fibrosis is the presence of NASH or inflamma-
tion [83, 84]. Thus, the pathophysiological paradigm of injury
and inflammation as drivers of fibrosis is consistent with the
clinical data.
Beyond progression to hepatic decompensation, patients with
NAFLD can also develop hepatocellular carcinoma (HCC) [85].
As with other liver diseases, most cases of HCC arise in the
context of cirrhosis; however, there is now clear evidence that
NAFLD can lead to HCC in the absence of underlying cirrhosis
and that the risk of noncirrhotic HCC appears to be higher in
NAFLD compared to viral hepatitis [86].
Extrahepatic manifestations
Although the presence of NAFLD (and especially NASH) is
accompanied by an increase in liver‐related mortality, cardio-
vascular disease and cancer are still the first and second leading
causes of death in these patients. Thus, there is a clear associa-
tion of NAFLD with extrahepatic manifestations.
 52:  Non‐alcoholic Fatty Liver Disease: Mechanisms and Treatment 675
Cardiovascular disease
The presence of fatty liver disease is associated with increased
cardiovascular risk. This increased risk spans the spectrum of
cardiovascular disease (CVD) from asymptomatic atherosclero-
sis [87] to increased cardiovascular events [88] and mortality
[81, 89]. Given that NAFLD and CVD share risk factors such as
obesity and insulin resistance, it is still not clear whether the
presence of NAFLD confers an increased risk beyond those risk
factors, and whether there is a direct mechanistic effect of
NAFLD on atherosclerosis development.
Extrahepatic cancers
Beyond the risk of HCC, patients with NAFLD also have higher
rates of extrahepatic malignancies. Several studies have shown
an increased risk of colorectal adenomas and cancer [90], with
more advanced liver disease likely portending a higher risk [91].
An increase in breast cancer [92] as well as in other cancers [93]
was also reported to be associated with NAFLD. As with cardio-
vascular disease risk, it is unclear whether the increased cancer
risk associated with NAFLD reflects an association with obesity
and diabetes, both known to increase cancer risk, or whether it
is independent of them.
Chronic kidney disease
Chronic kidney disease is more common in NAFLD and the risk
is higher in patients with NASH [94], even after controlling for
metabolic risk factors.
EVALUATION OF THE PATIENT
WITH NAFLD
Clinical features and diagnosis
NAFLD is diagnosed when there is evidence of hepatic steatosis
on imaging or histology and there are no secondary causes
of  steatosis present [1]. Patients with NAFLD are typically
asymptomatic, and diagnosis is often incidental. Patients may
occasionally present with fatigue or a vague, nondescript pain or
discomfort in the right upper quadrant and may have hepato-
megaly due to fatty infiltration of the liver. Signs or symptoms
of chronic liver disease are generally absent unless patients
­present with advanced liver disease or cirrhosis.
Role of liver biopsy
The liver biopsy is an important tool in the diagnosis and man-
agement of NAFLD. It is the gold standard and the only reliable
method to differentiate NASH from NAFL and also allows for
assessment of the multiple components of disease and for accu-
rately staging the degree of fibrosis. There are several scoring
schemes that are utilized for semiquantitative assessment of
NAFLD histology, most commonly the NASH–CRN scoring
system [95]. These scores are appropriate for use in clinical tri-
als but should not be used in lieu of an appropriate histological
diagnosis [96]. Despite its usefulness, the use of a liver biopsy is
limited by its inherent risk [97], cost and sampling error.
Currently, guidelines recommend considering a liver biopsy in
patients who are at an increased risk of having advanced fibro-
sis, as well as in patients who are suspected to have competing
etiologies for hepatic steatosis, or to identify the chief cause of
liver damage in case of more than one liver disease [1]. Thus,
there is an unmet need for non‐invasive tools to identify and
stratify NAFLD.
Non‐invasive assessments
Assessment of liver steatosis
Imaging modalities such as ultrasound or computed tomography
(CT) are reasonably sensitive and specific for the detecting the
presence of moderate–severe steatosis [98]. For example, for
detection of greater than 20% steatosis, ultrasound has 91% sensi-
tivity and 99% specificity [99] but sensitivity markedly decreases
for mild steatosis. The degree of hyperechogenicity on ultrasound
or the relative or absolute hepatic attenuation on CT can be used
for semiquantitative assessment of the degree of steatosis but with
limited accuracy. Ultrasound remains the primary modality for
diagnosis of hepatic steatosis, due to its wide availability, lack of
radiation exposure, and low cost. Recently, the controlled attenua-
tion parameter (CAP), quantifying the degree of ultrasound atten-
uation in the liver was found to be a sensitive measure for detection
and quantification of steatosis [100]
In contrast to ultrasound and CT, magnetic resonance imag-
ing (MRI)‐based imaging modalities offer superior sensitivity
and specificity for detecting steatosis. An additional advantage
of MRI is the ability to accurately quantify liver fat using either
MR spectroscopy or MRI proton density fat fraction (MRI‐
PDFF) [101]. While MRI‐based detection methods have a high
sensitivity and specificity, cost and limited availability currently
do not make them a viable option outside of clinical research.
There are several composite scores to calculate the degree of
steatosis that are based on serum biomarkers, with or without
anthropometric parameters, such as the fatty liver index, the
hepatic steatosis index, and the proprietary SteatoTest; these
have reasonable accuracy but it is unclear whether they can
replace the need for an imaging study in clinical care [102].
Identification of NASH
A crucial element in managing patients with NAFLD and in
clinical trial recruitment is the differentiation of NAFL from
NASH. As detailed above, a liver biopsy is the gold standard for
diagnosis, but poses significant limitations in terms of accepta-
bility, risk, and cost. Several blood‐based markers for the diag-
nosis of NASH have been studied, based on disease pathogenesis.
Examples are cytokeratin‐18 fragments as markers of hepato-
cyte apoptosis [103, 104], oxidized lipid products [104], and
several proprietary panels (reviewed in [102]). However, these
all suffer from limited accuracy, incomplete validation, and/or
cost and are not yet ready for routine clinical application.
Assessment of fibrosis
Since fibrosis is the main determinant of mortality and liver‐related
complications in NAFLD, a non‐invasive marker of fibrosis would
provide risk stratification and allow for selection of patients for
676 THE LIVER:  TREATMENT OPTIONS IN NAFLD

treatment. As with assessment of steatosis, investigated biomark-
ers include imaging modalities and blood‐based markers [105].
Liver fibrosis is associated with increased hepatic stiffness.
Several imaging modalities measure liver elasticity, which is
directly related to stiffness, by generating a shear wave in the
liver and measuring its velocity. Vibration controlled transient
elastography (Fibroscan), shear‐wave elastography, and acoustic
radial force imaging (ARFI) are based on ultrasound, while MRI
is used for magnetic resonance elastography (MRE). In general,
elastography‐based tests provide high accuracy in detecting cir-
rhosis and advanced fibrosis but are not sufficiently accurate to
diagnose the presence of earlier fibrosis (stage 2) [102, 106].
Several blood‐based biomarkers or compound scores were
developed to detect fibrosis. Some of these are nonspecific but
employ readily available data such as liver enzymes, platelet
count, and clinical and anthropometric features. These include
scores initially applied to viral hepatitis, such as the AST/ALT
ratio, AST/platelet ratio index (APRI) or FIB‐4, as well as
scores developed specifically for NAFLD such as the NAFLD
fibrosis score (NFS) or BARD score. The accuracy of all of
these scores is limited and despite easy availability, they are not
sufficient for routine clinical use, though they are useful in epi-
demiological studies [105].
Other blood‐based biomarkers are more specific and are
based on measuring markers of the hepatic fibrotic process and
deposition of extracellular matrix. Blood levels of hyaluronic
acid, TIMP1, and pro‐collagen III amino terminal peptide
(PIIINP and Pro‐C3) are among the promising ones. Several
compound and proprietary scores are based on these markers,
including the enhanced liver fibrosis (ELF), the FibroSure/
FibroTest, and FibroMeter NAFLD. In general, these specific
scores provide greater accuracy than the nonspecific ones, but
are currently limited by availability and cost.
TREATMENT OPTIONS IN NAFLD
Despite the rise in prevalence and impact of NAFLD and NASH,
there are currently no approved therapies for the disease.
However, several medications approved for other indications
have shown benefit in treating NASH and a plethora of new
pharmacological agents are being studied. With the increasing
understanding of pathogenic processes, there is a plethora of
agents developed to target these processes (Figure 52.2) [107].
Reviewed here are currently available treatments with proven

Figure 52.2  Mechanistic targets for NAFLD/NASH treatment. Increased availability of carbohydrates and non‐esterified fatty acids, together
with insulin resistance, drive hepatic accumulation of triglycerides and fatty acids, leading to metabolic and oxidative stress, hepatocyte injury,
activation of an inflammatory response and eventually fibrosis. Lifestyle modification and bariatric surgery decrease the excessive caloric load to
the liver. GLP‐1 receptor agonists (such as liraglutide) decrease caloric load (through reduction in appetite and weight loss) and improve insulin
sensitivity with possible direct effect on the liver. PPAR agonists (like pioglitazone or elafibranor) improve insulin resistance in adipose tissue and/
or muscle, reducing the excess load of glucose and fatty acids to the liver, while also decreasing hepatic lipogenesis. Agonists of FXR (such as
obeticholic acid) or FGF‐19 analogs (NGM‐282) decrease bile acid synthesis and hepatic lipogenesis; NGM‐282 may also act as an insulin sensi-
tizer. Inhibitors of acetyl‐coA carboxylase, the first step in hepatic de novo lipogenesis, block hepatic fatty acid synthesis and accumulation.
Downstream of hepatic lipid metabolism, vitamin E, a fat‐soluble antioxidant, ameliorates the oxidative stress that results from lipotoxicity while
ASK1 inhibitors (such selonsertib) block the ASK1‐JNK pathway through which oxidative stress leads to hepatocyte injury and apoptosis.
Cenicriviroc, a C‐C chemokine receptor type 2 (CCR2) and 5 (CCR5) antagonist, blocks the recruitment of inflammatory cells and decreases
inflammatory injury. Finally, anti‐fibrotic agents target the fibrotic process itself. ACCi, acetyl‐coA carboxylase inhibitors; ASK1i, apoptosis sig-
nal‐regulating kinase 1 inhibitors; CCR, C‐C chemokine receptor; CHO, carbohydrates; FA, fatty acids; FGF, fibroblast growth factor; FXR,
farnesoid X receptor; GLP‐1RA, glucagon‐like peptide 1 receptor agonists, NEFA, non‐esterified fatty acids; PPAR, peroxisome proliferator‐acti-
vator receptor; TG, triglycerides.
 52:  Non‐alcoholic Fatty Liver Disease: Mechanisms and Treatment 677
Table 52.1  Select response rates in clinical trials of NASH
Histological
Treatment Duration responsea (%)
Lifestyle (all) 12 months 47
Lifestyle (Subjects 12 months 82
with ≥ 5% weight loss)
Pioglitazone 18 months –  34b–58
96 weeks
Vitamin E 96 weeks 43
Obeticholic acid 72 weeks 45
Elafibranor (120 mg) 52 weeks 48
Liraglutide 48 weeks 43
a
 Histological response criteria varied between trials but typically include at least 2‐point reduction in the NAFLD activity score (NAS). Response percentages calculated by
intention‐to‐treat.
b
 Non‐significant compared to placebo.
benefit, and those that are in advanced clinical trial stages at the
time of writing (Table 52.1).
Reduction of caloric intake
Treatment trials of lifestyle intervention, including diet and
physical activity, have shown not only reduction in liver TG
content but also histological improvement, directly related to
the degree of weight loss [108]. Recently Vilar‐Gomez and oth-
ers [109] reported on the results of a 52‐week lifestyle interven-
tion program in 293 subjects with NASH. Only 30% of
participants achieved greater than 5% weight loss, highlighting
the difficulty of achieving a sustained effect, the main limitation
of lifestyle interventions. However, among those who lost
greater than 5% of their weight, resolution of NASH was seen in
58%. Furthermore, in the few subjects who lost greater than
10% of their weight, resolution of NASH was seen in 90% and
a reduction in fibrosis stage was seen in 81%. Another evidence,
albeit indirect, for the benefit of weight reduction is the marked
rate of histological response in the control arms of therapeutic
trials, which is typically associated with losing weight [110].
Given the proven effects of lifestyle modification for NASH and
its known benefits for associated comorbidities such as diabetes
and metabolic syndrome, it should be advocated in any patient
with NAFLD, irrespective of pharmacological therapy options.
Due to the difficulty of sustaining weight loss with lifestyle
interventions alone, bariatric surgery is useful for patients with
obesity and has shown benefit in treating diabetes. In the largest
series to date [15], histological resolution of NASH was seen in
85% of subjects one year after surgery, and 33% showed
improvement in steatosis. Recently, 5‐year follow‐up data was
reported from the same cohort [111]. The 85% NASH resolu-
tion rate persisted and fibrosis continued to improve between
years 1 and 5. Thus, bariatric surgery may be an attractive ther-
apy for a subset of patients with NASH,
Pharmacological treatment
PPAR activation
The peroxisome proliferator‐activator receptors (PPARs) are a
group of nuclear receptors that transcriptionally regulate a wide
variety of metabolic and inflammatory processes. Their activation
is typically driven by binding various FA and FA derivatives as
Resolution of
NASH (%) Improvement in fibrosis (%) References
25 19b [109]
58 26b [109]
(Subjects with ≥ 10% weight
loss had 45% fibrosis regression)
47–51 44b [112, 113]
36 41b [112]
20b 35 [115]
19 Not reported [114]
35 26b [124]
ligands. PPARγ is expressed predominantly in the adipose tissue
and controls adipogenesis, lipogenesis, and glucose metabolism.
Thiazolidinediones (TZDs) are synthetic agonists of PPARγ and
are approved for the treatment of diabetes. Pioglitazone, the TZD
most commonly used, was studied for the treatment of NASH in
several trials. In the PIVENS study, 163 non‐diabetics with
NASH were treated with pioglitazone or placebo for 96 weeks.
Histological improvement was seen in 34% of subjects treated
with pioglitazone, compared to 19% in placebo, but the difference
did not reach the pre‐specified statistical significance level [112].
In contrast, in a study including 101 diabetics or pre‐diabetics
with NASH, 18 months of pioglitazone treatment were associated
with resolution of NASH in 51% of subjects, which was signifi-
cantly greater than the placebo‐treated arm [113]. Overall, piogl-
itazone treatment appears to be effective for NASH, especially in
subjects with diabetes or pre‐diabetes, but its acceptability may
be limited by concerns about cardiovascular risks and by weight
gain, which is a common side‐effect.
Elafibrinor is a dual PPAR agonist, targeting both PPARα and
PPARδ. The phase IIb GOLDEN‐505 trial compared two doses
of elafibrinor (80 mg and 120 mg) for 52 weeks to placebo.
Treatment with elafibranor was associated with histological
improvement, but only in a post hoc analysis in subjects with
significant disease activity at baseline [114]. A phase III trial
is underway.
Bile acid signaling
FXR acts as an intracellular bile acid sensor and when acti-
vated, inhibits bile acid synthesis and secretion as well as
decreasing lipogenesis. Obeticholic acid (OCA) is a synthetic
bile acid that acts as an FXR agonist. The phase IIb FLINT trial
compared OCA with placebo in patients with NASH; OCA was
superior to placebo in achieving histological improvement
(46%) and NASH resolution (22%) [115] but was associated
with significant pruritus and a mild increase in LDL‐­cholesterol.
A phase III trial to determine efficacy and safety is ongoing and
other FXR agonists are in earlier stages of development.
Fibroblast growth factor 19 (FGF19) is a peptide hormone that
is released in response to FXR activation in the terminal ileum
and controls hepatocyte bile acid synthesis through its receptor
FGFR4. FGF19 also has insulin‐like effects on hepatic gluco-
neogenesis and glycogen synthesis. Phase II studies in patients
with NAFLD using NGM282, an FGF19 analog, demonstrate
678 THE LIVER:  REFERENCES
efficacy in reducing hepatic fat content [116] and improving
NASH histology [117].
Vitamin E
Given the evidence for the role of oxidative stress in the patho-
genesis of NASH, antioxidants could have therapeutic benefit.
Vitamin E (α‐tocopherol), a lipid‐soluble antioxidant, has been
studied in several clinical trials. In the previously mentioned
PIVENS trial, 43% of non‐diabetic patients with NASH treated
with 800 IU/day of natural vitamin E had histological improve-
ment and 36% had resolution of NASH [112] and similar results
were seen in the TONIC pediatric trial [118]. Furthermore, of all
pharmacological therapies, vitamin E is the only one shown to
decrease the rate of clinically meaningful outcomes such as
decompensation or need for liver transplant, albeit in a retro-
spective analysis [119]. However, there have been reports of
increased risk of hemorrhagic stroke [120] and prostate cancer
[121] in subjects receiving high‐dose vitamin E, and a recent
report from a small trial suggests an increase in gastrointestinal
bleeding in NAFLD patients, especially in diabetics [122].
Thus, using vitamin E to treat NASH can be considered but
requires monitoring and assessment of potential risks.
GLP 1 receptor agonists
GLP‐1 (glucagon‐like peptide‐1) is an incretin hormone secreted
by enteroendocrine L cells in response to ingested nutrients. It
acts predominantly in the fed‐state, modulating and augmenting
the insulin response. GLP‐1 potentiates glucose‐stimulated
insulin secretion from beta cells, mediated by activation of
its  G  protein‐coupled receptor. GLP‐1 receptor agonists
(GLP‐1RA) are in clinical use for the treatment of diabetes and
obesity and are thought to mediate their effects through a com-
bination of the insulinotropic effect, weight loss, increased
energy expenditure, and delayed gastric emptying.
In diabetes‐associated NAFLD, GLP‐1RAs lead to reduction
in liver enzymes and liver fat content in human and animal stud-
ies [123]. In the phase II randomized controlled LEAN trial
[124], 52 non‐diabetics with NASH were treated with liraglu-
tide, a GLP‐1RA, or placebo for 48 weeks. Liraglutide provided
histological benefit including resolution of NASH in 39%.
Larger trials with other GLP‐1RA formulations are ongoing.
Given the combined benefit on insulin resistance, weight, liver
histology, and beneficial cardiovascular outcome in diabetics,
GLP‐1RAs are an attractive option to explore further in the
management of NAFLD.
Inhibition of de novo lipogenesis
As detailed above, DNL is increased in the liver of subjects
with NAFLD and DNL‐derived fatty acids may be drivers of
hepatic lipotoxicity. GS‐0976 is a liver‐targeted inhibitor of
acetyl‐CoA carboxylase (ACC), the rate‐limiting enzyme in
DNL. In a prospective randomized phase 2 study, 100 subjects
with likely NASH received GS‐0976 (5 or 20 mg/d) for
12  weeks and were compared with 26 subjects who received
­placebo [125]. GS‐0976 treatment was significantly associated
with a dose‐dependent and marked reduction in liver fat
­content, and was generally well‐tolerated, although marked
hypertriglyceridemia was seen in 16% of treated patients. It is
still unknown whether GS‐0976 treatment can also lead to his-
tological improvement in NASH.
Other treatments
There is a large number of investigative agents currently studied
as potential treatments for NASH, targeting various aspects of
the disease. Among those are agonists directed at the thyroid
hormone receptor β, inhibitors of apoptosis signal‐regulating
kinase 1 (ASK‐1), a key mediator of hepatocyte stress response,
modulators of gut microbiome, combined agonists of receptors
to gut‐derived hormones, an analogue of FGF21, and antifi-
brotic agents. In addition, there are agents in early stages of
development targeting genes that are associated with the dis-
ease. Finally, combination treatment with agents targeting dif-
ferent pathways may be needed to achieve meaningful responses
in a large proportion of subjects.
CONCLUSION
NAFLD has risen to epidemic proportions and directly or indi-
rectly affects the health of many. Increased understanding of
genetic and pathogenic mechanisms will hopefully lead to
effective and safe therapeutic options.
REFERENCES
  1. Chalasani, N., Younossi, Z., Lavine, J.E. et al. The diagnosis and manage-
ment of nonalcoholic fatty liver disease: practice guidance from the American
Association for the Study of Liver Diseases. Hepatology, 2018; 67:328–57.
 2. Szczepaniak, L.S., Nurenberg, P., Leonard, D. et  al. Magnetic resonance
spectroscopy to measure hepatic triglyceride content: prevalence of hepatic
steatosis in the general population. Am J Physiol Endocrinol Metab, 2005;
288:E462–8.
 3. Kwiterovich, P.O., Jr., Sloan, H.R., and Fredrickson, D.S. Glycolipids and
other lipid constituents of normal human liver. J Lipid Res, 1970; 11:322–30.
  4. Younossi, Z., Anstee, Q.M., Marietti, M. et al. Global burden of NAFLD and
NASH: trends, predictions, risk factors and prevention. Nat Rev Gastroenterol
Hepatol, 2018; 15:11–20.
  5. Clark, J.M., Brancati, F.L., and Diehl, A.M. The prevalence and etiology of
elevated aminotransferase levels in the United States. Am J Gastroenterol,
2003; 98:960–7.
  6. Ioannou, G.N., Boyko, E.J., and Lee, S.P. The prevalence and predictors of
elevated serum aminotransferase activity in the United States in 1999–2002.
Am J Gastroenterol, 2006; 101:76–82.
  7. Takyar, V., Nath, A., Beri, A. et  al. How healthy are the “Healthy volun-
teers”? Penetrance of NAFLD in the biomedical research volunteer pool.
Hepatology, 2017; 66:825–33.
  8. Fracanzani, A.L., Valenti, L., Bugianesi, E. et al. Risk of severe liver disease
in nonalcoholic fatty liver disease with normal aminotransferase levels: a
role for insulin resistance and diabetes. Hepatology, 2008; 48:792–8.
  9. Lazo, M., Hernaez, R., Eberhardt, M.S. et al. Prevalence of nonalcoholic fatty
liver disease in the United States: the Third National Health and Nutrition
Examination Survey, 1988–1994. Am J Epidemiol, 2013; 178:38–45.
10. Williams, C.D., Stengel, J., Asike, M.I. et  al. Prevalence of nonalcoholic
fatty liver disease and nonalcoholic steatohepatitis among a largely middle‐
aged population utilizing ultrasound and liver biopsy: a prospective study.
Gastroenterology, 2011; 140:124–31.
11. Browning, J.D., Szczepaniak, L.S., Dobbins, R. et al. Prevalence of hepatic
steatosis in an urban population in the United States: Impact of ethnicity.
Hepatology, 2004; 40:1387–95.
 52:  Non‐alcoholic Fatty Liver Disease: Mechanisms and Treatment 679
12. Younossi, Z.M., Koenig, A.B., Abdelatif, D. et al. Global epidemiology of
nonalcoholic fatty liver disease  –  meta‐analytic assessment of prevalence,
incidence, and outcomes. Hepatology, 2016; 64:73–84.
13. Welsh, J.A., Karpen, S., and Vos, M.B. Increasing prevalence of nonalco-
holic fatty liver disease among United States adolescents, 1988–1994 to
2007–2010. J Pediatr, 2013; 162:496–500.
14. Schwimmer, J.B., Deutsch, R., Kahen, T. et al. Prevalence of fatty liver in
children and adolescents. Pediatrics, 2006; 118:1388–93.
15. Lassailly, G., Caiazzo, R., Buob, D. et al. Bariatric surgery reduces features
of nonalcoholic steatohepatitis in morbidly obese patients. Gastroenterology,
2015; 149:379–88.
16. Sasaki, A., Nitta, H., Otsuka, K. et al. Bariatric surgery and non‐alcoholic
fatty liver disease: current and potential future treatments. Front Endocrinol,
2014; 5:164.
17. Doycheva, I., Patel, N., Peterson, M., and Loomba, R. Prognostic implica-
tion of liver histology in patients with nonalcoholic fatty liver disease in
diabetes. J Diabetes Complications, 2013; 27:293–300.
18. Anstee, Q.M., Targher, G., and Day, C.P. Progression of NAFLD to diabetes
mellitus, cardiovascular disease or cirrhosis. Nat Rev Gastroenterol Hepatol,
2013; 10:330–44.
19. Marchesini, G., Bugianesi, E., Forlani, G. et  al. Nonalcoholic fatty liver,
steatohepatitis, and the metabolic syndrome. Hepatology, 2003; 37:917–23.
20. Zelber‐Sagi, S., Lotan, R., Shlomai, A. et  al. Predictors for incidence and
remission of NAFLD in the general population during a seven‐year prospec-
tive follow‐up. J Hepatol, 2012; 56:1145–51.
21. Ma, J., Hwang, S.J., Pedley, A. et al. Bi‐directional analysis between fatty
liver and cardiovascular disease risk factors. J Hepatol, 2017; 66:390–7.
22. Kojima, S.‐I., Watanabe, N., Numata, M. et al. Increase in the prevalence of
fatty liver in Japan over the past 12 years: analysis of clinical background. J
Gastroenterol, 2003; 38:954–61.
23. Loomba, R., Schork, N., Chen, C.H. et al. Heritability of Hepatic Fibrosis
and Steatosis Based on a Prospective Twin Study. Gastroenterology, 2015;
149:1784–93.
24. Schwimmer, J.B., Celedon, M.A., Lavine, J.E. et al. Heritability of nonalco-
holic fatty liver disease. Gastroenterology, 2009; 136:1585–92.
25. Romeo, S., Kozlitina, J., Xing, C. et al. Genetic variation in PNPLA3 confers
susceptibility to nonalcoholic fatty liver disease. Nature Genet, 2008;
40:1461–5.
26. Yuan, X., Waterworth, D., Perry, J.R.B. et  al. Population‐based genome‐
wide association studies reveal six loci influencing plasma levels of liver
enzymes. Am J Hum Genet, 2008; 83:520–8.
27. Rotman, Y., Koh, C., Zmuda, J.M. et al. The association of genetic variability
in patatin‐like phospholipase domain‐containing protein 3 (PNPLA3) with
histological severity of nonalcoholic fatty liver disease. Hepatology, 2010;
52:894–903.
28. Speliotes, E.K., Butler, J.L., Palmer, C.D. et al. PNPLA3 variants specifi-
cally confer increased risk for histologic nonalcoholic fatty liver disease but
not metabolic disease. Hepatology, 2010; 52:904–12.
29. Hassan, M.M., Kaseb, A., Etzel, C.J. et al. Genetic variation in the PNPLA3
gene and hepatocellular carcinoma in USA: risk and prognosis prediction.
Mol Carcinog, 2013; 52: Suppl 1:E139–47.
30. Tian, C., Stokowski, R.P., Kershenobich, D. et  al. Variant in PNPLA3 is
associated with alcoholic liver disease. Nature Genet, 2010; 42:21–3.
31. Basuray, S., Smagris, E., Cohen, J.C., and Hobbs, H.H. The PNPLA3 variant
associated with fatty liver disease (I148M) accumulates on lipid droplets by
evading ubiquitylation. Hepatology, 2017; 66:1111–24.
32. Mitsche, M.A., Hobbs, H.H., and Cohen, J.C. Patatin‐like phospholipase
domain‐containing protein 3 promotes transfers of essential fatty acids from
triglycerides to phospholipids in hepatic lipid droplets. J Biol Chem, 2018;
293:6958–68.
33. Pirazzi, C., Valenti, L., Motta, B.M. et al. PNPLA3 has retinyl‐palmitate lipase
activity in human hepatic stellate cells. Hum Mol Genet, 2014; 23:4077–85.
34. Kozlitina, J., Smagris, E., Stender, S. et al. Exome‐wide association study
identifies a TM6SF2 variant that confers susceptibility to nonalcoholic fatty
liver disease. Nat Genet, 2014; 46:352–6.
35. Holmen, O.L., Zhang, H., Fan, Y. et  al. Systematic evaluation of coding
­variation identifies a candidate causal variant in TM6SF2 influencing total
cholesterol and myocardial infarction risk. Nat Genet, 2014; 46:345–51.
36. Mahdessian, H., Taxiarchis, A., Popov, S. et al. TM6SF2 is a regulator of
liver fat metabolism influencing triglyceride secretion and hepatic lipid
droplet content. Proc Nat Acad Sci USA, 2014; 111:8913–8.
37. Smagris, E., Gilyard, S., Basuray, S. et al. Inactivation of Tm6sf2, a gene
defective in fatty liver disease, impairs lipidation but not secretion of very
low density lipoproteins. J Biol Chem, 2016; 291:10659–76.
38. Abul‐Husn, N.S., Cheng, X., Li, A.H. et al. A protein‐truncating HSD17B13
variant and protection from chronic liver disease. New Eng J Med, 2018;
378:1096–106.
39. Ma, Y., Belyaeva, O.V., Brown, P.M. et al. HSD17B13 is a hepatic retinol
dehydrogenase associated with histological features of non‐alcoholic fatty
liver disease. Hepatology, 2019;69(4):1504–19.
40. Kozlitina, J., Stender, S., Hobbs, H.H., and Cohen, J.C. HSD17B13 and chronic
liver disease in Blacks and Hispanics. N Engl J Med, 2018; 379:1876–7.
41. Eslam, M., Valenti, L., and Romeo, S. Genetics and epigenetics of NAFLD
and NASH: clinical impact. J Hepatol, 2018; 68:268–79.
42. Pirola, C.J., Gianotti, T.F., Burgueno, A.L. et al. Epigenetic modification of
liver mitochondrial DNA is associated with histological severity of nonalco-
holic fatty liver disease. Gut, 2013; 62:1356–63.
43. Donnelly, K.L., Smith, C.I., Schwarzenberg, S.J. et al. Sources of fatty acids
stored in liver and secreted via lipoproteins in patients with nonalcoholic
fatty liver disease. J Clin Invest, 2005; 115:1343–51.
44. Sanyal, A.J., Campbell‐Sargent, C., Mirshahi, F. et al. Nonalcoholic steato-
hepatitis: association of insulin resistance and mitochondrial abnormalities.
Gastroenterology, 2001; 120:1183–92.
45. Bril, F., Barb, D., Portillo‐Sanchez, P. et al. Metabolic and histological impli-
cations of intrahepatic triglyceride content in nonalcoholic fatty liver dis-
ease. Hepatology, 2017; 65:1132–44.
46. Musso, G., Cassader, M., De Michieli, F. et al. Nonalcoholic steatohepatitis
versus steatosis: Adipose tissue insulin resistance and dysfunctional response
to fat ingestion predict liver injury and altered glucose and lipoprotein
metabolism. Hepatology, 2012; 56:933–42.
47. Schwarz, J.M., Neese, R.A., Turner, S. et al. Short‐term alterations in carbo-
hydrate energy intake in humans. Striking effects on hepatic glucose produc-
tion, de novo lipogenesis, lipolysis, and whole‐body fuel selection. J Clin
Invest, 1995; 96:2735–43.
48. Lambert, J.E., Ramos‐Roman, M.A., Browning, J.D., and Parks, E.J.
Increased de novo lipogenesis is a distinct characteristic of individuals with
nonalcoholic fatty liver disease. Gastroenterology, 2014; 146:726–35.
49. Brown, M.S. and Goldstein, J.L. Selective versus total insulin resistance: a
pathogenic paradox. Cell Metab, 2008; 7:95–6.
50. Otero, Y.F., Stafford, J.M., and Mcguinness, O.P. Pathway‐selective insulin
resistance and metabolic disease: the importance of nutrient flux. J Biol
Chem, 2014; 289:20462–9.
51. Fabbrini, E., Mohammed, B.S., Magkos, F. et al. Alterations in adipose tis-
sue and hepatic lipid kinetics in obese men and women with nonalcoholic
fatty liver disease. Gastroenterology, 2008; 134:424–31.
52. Iozzo, P., Bucci, M., Roivainen, A. et al. Fatty acid metabolism in the liver,
measured by positron emission tomography, is increased in obese individu-
als. Gastroenterology, 2010; 139:846–56, 56 e1–6.
53. Naguib, G., Morris, N., Haynes‐Williams, V. et  al. Beta‐oxidation is
decreased in non‐alcoholic fatty liver disease: a non‐invasive assessment uti-
lizing palmitate and acetate breath tests. Hepatology, 2017; 66:1043.
54. Day, C.P. and James, O.F.W. Steatohepatitis: a tale of two “hits”?
Gastroenterology, 1998; 114:842–5.
55. Neuschwander‐Tetri, B.A. Hepatic lipotoxicity and the pathogenesis of non-
alcoholic steatohepatitis: the central role of nontriglyceride fatty acid metab-
olites. Hepatology, 2010; 52:774–88.
56. Yamaguchi, K., Yang, L., Mccall, S. et  al. Inhibiting triglyceride synthe-
sis  improves hepatic steatosis but exacerbates liver damage and fibrosis
in  obese mice with nonalcoholic steatohepatitis. Hepatology, 2007;
45:1366–74.
57. Videla, L.A., Rodrigo, R., Orellana, M. et al. Oxidative stress‐related param-
eters in the liver of non‐alcoholic fatty liver disease patients. Clin Sci (Lond),
2004; 106:261–8.
58. Sunny, N.E., Parks, E.J., Browning, J.D., and Burgess, S.C. Excessive
hepatic mitochondrial TCA cycle and gluconeogenesis in humans with non-
alcoholic fatty liver disease. Cell Metab, 2011; 14:804–10.
59. Koliaki, C., Szendroedi, J., Kaul, K. et al. Adaptation of hepatic mitochon-
drial function in humans with non‐alcoholic fatty liver is lost in steatohepa-
titis. Cell Metab, 2015; 21:739–46.
60. Bellanti, F., Villani, R., Facciorusso, A. et al. Lipid oxidation products in the
pathogenesis of non‐alcoholic steatohepatitis. Free Radic Biol Med, 2017;
111:173–85.
680 THE LIVER:  REFERENCES
61. Seki, S., Kitada, T., Yamada, T. et al. In situ detection of lipid peroxidation
and oxidative DNA damage in non‐alcoholic fatty liver diseases. J Hepatol,
2002; 37:56–62.
62. Puri, P., Mirshahi, F., Cheung, O. et al. Activation and dysregulation of the
unfolded protein response in nonalcoholic fatty liver disease.
Gastroenterology, 2008; 134:568–76.
63. Lebeaupin, C., Vallee, D., Hazari, Y. et al. Endoplasmic reticulum stress sig-
nalling and the pathogenesis of non‐alcoholic fatty liver disease. J Hepatol,
2018; 69:927–47.
64. Feldstein, A.E., Canbay, A., Angulo, P. et al. Hepatocyte apoptosis and fas
expression are prominent features of human nonalcoholic steatohepatitis.
Gastroenterology, 2003; 125:437–43.
65. Ibrahim, S.H., Hirsova, P., and Gores, G.J. Non‐alcoholic steatohepatitis
pathogenesis: sublethal hepatocyte injury as a driver of liver inflammation.
Gut, 2018; 67:963–72.
66. Rangwala, F., Guy, C.D., Lu, J. et al. Increased production of sonic hedge-
hog by ballooned hepatocytes. J Pathol, 2011; 224:401–10.
67. Garcia‐Martinez, I., Santoro, N., Chen, Y. et al. Hepatocyte mitochondrial
DNA drives nonalcoholic steatohepatitis by activation of TLR9. J Clin
Invest, 2016; 126:859–64.
68. Arrese, M., Cabrera, D., Kalergis, A.M., and Feldstein, A.E. Innate immu-
nity and inflammation in NAFLD/NASH. Dig Dis Sci, 2016; 61:1294–303.
69. Schuster, S., Cabrera, D., Arrese, M., and Feldstein, A.E. Triggering and
resolution of inflammation in NASH. Nat Rev Gastroenterol Hepatol, 2018;
15:349–64.
70. Xu, R., Huang, H., Zhang, Z., and Wang, F.S. The role of neutrophils in the
development of liver diseases. Cell Mol Immunol, 2014; 11:224–31.
71. Byun, J.S. and Yi, H.S. Hepatic immune microenvironment in alcoholic and
nonalcoholic liver disease. Biomed Res Int, 2017; 2017:6862439.
72. Sharpton, S.R., Ajmera, V., and Loomba, R. Emerging role of the gut micro-
biome in nonalcoholic fatty liver disease: from composition to function. Clin
Gastroenterol Hepatol.2019;17(2):296–306.
73. Le Roy, T., Llopis, M., Lepage, P. et  al. Intestinal microbiota determines
development of non‐alcoholic fatty liver disease in mice. Gut, 2013;
62:1787–94.
74. Henao‐Mejia, J., Elinav, E., Jin, C. et al. Inflammasome‐mediated dysbiosis
regulates progression of NAFLD and obesity. Nature, 2012; 482:179–85.
75. Chu, H., Duan, Y., Yang, L., and Schnabl, B. Small metabolites, possible big
changes: a microbiota‐centered view of non‐alcoholic fatty liver disease.
Gut, 2019;68(2):359–70.
76. Jiang, C., Xie, C., Li, F. et al. Intestinal farnesoid X receptor signaling pro-
motes nonalcoholic fatty liver disease. J Clin Invest, 2015; 125:386–402.
77. Ma, C., Han, M., Heinrich, B. et  al. Gut microbiome‐mediated bile acid
metabolism regulates liver cancer via NKT cells. Science, 2018; 360:(6391).
78. Vilar‐Gomez, E., Calzadilla‐Bertot, L., Wai‐Sun Wong, V. et al. Fibrosis sever-
ity as a determinant of cause‐specific mortality in patients with advanced non-
alcoholic fatty liver disease: a multi‐national cohort study. Gastroenterology,
2018; 155:443–57 e17.
79. Angulo, P., Kleiner, D.E., Dam‐Larsen, S. et al. Liver fibrosis, but no other
histologic features, is associated with long‐term outcomes of patients with
nonalcoholic fatty liver disease. Gastroenterology, 2015; 149:389–97.
80. Dulai, P.S., Singh, S., Patel, J. et al. Increased risk of mortality by fibrosis
stage in nonalcoholic fatty liver disease: systematic review and meta‐analy-
sis. Hepatology, 2017; 65:1557–65.
81. Ekstedt, M., Hagström, H., Nasr, P. et al. Fibrosis stage is the strongest pre-
dictor for disease‐specific mortality in NAFLD after up to 33 years of fol-
low‐up. Hepatology, 2015; 61:1547–54.
82. Ekstedt, M., Franzén, L.E., Mathiesen, U.L. et al. Long‐term follow‐up of
patients with NAFLD and elevated liver enzymes. Hepatology, 2006;
44:865–73.
83. Singh, S., Allen, A.M., Wang, Z. et al. Fibrosis progression in nonalcoholic
fatty liver vs nonalcoholic steatohepatitis: a systematic review and meta‐­
analysis of paired‐biopsy studies. Clin Gastroenterol Hepatol, 2015; 13:
643–54.e1–9; quiz e39–40.
84. Argo, C.K., Northup, P.G., Al‐Osaimi, A.M.S., and Caldwell, S.H.
Systematic review of risk factors for fibrosis progression in non‐alcoholic
steatohepatitis. J Hepatol, 2009; 51:371–9.
85. Younossi, Z.M., Otgonsuren, M., Henry, L. et  al. Association of nonalco-
holic fatty liver disease (NAFLD) with hepatocellular carcinoma (HCC) in
the United States from 2004 to 2009. Hepatology, 2015; 62:1723–30.
  86. Mittal, S., El‐Serag, H.B., Sada, Y.H. et al. Hepatocellular carcinoma in the
absence of cirrhosis in United States veterans is associated with nonalco-
holic fatty liver disease. Clin Gastroenterol Hepatol, 2016; 14:124–31 e1.
  87. Pais, R., Giral, P., Khan, J.F. et al. Fatty liver is an independent predictor of
early carotid atherosclerosis. J Hepatol, 2016; 65:95–102.
  88. Targher, G., Byrne, C.D., Lonardo, A. et al. Non‐alcoholic fatty liver dis-
ease and risk of incident cardiovascular disease: a meta‐analysis. J Hepatol,
2016; 65:589–600.
  89. Söderberg, C., Stål, P., Askling, J. et al. Decreased survival of subjects with
elevated liver function tests during a 28‐year follow‐up. Hepatology, 2010;
51:595–602.
  90. Adams, L.A., Anstee, Q.M., Tilg, H., and Targher, G. Non‐alcoholic fatty
liver disease and its relationship with cardiovascular disease and other
extrahepatic diseases. Gut, 2017; 66:1138–53.
  91. Ahn, J.S., Sinn, D.H., Min, Y.W. et al. Non‐alcoholic fatty liver diseases and
risk of colorectal neoplasia. Aliment Pharmacol Ther, 2017; 45:345–53.
  92. Kim, G.A., Lee, H.C., Choe, J. et al. Association between non‐alcoholic
fatty liver disease and cancer incidence rate. J Hepatol.2017;
 93. Hicks, S.B., Mara, K., Larson, J.J. et  al. The incidence of extrahepatic
malignancies in nonalcoholic fatty liver disease (NAFLD). Hepatology,
2018; 68:20A–A.
  94. Musso, G., Gambino, R., Tabibian, J.H. et al. Association of non‐alcoholic
fatty liver disease with chronic kidney disease: a systematic review and
meta‐analysis. PLoS Med, 2014; 11:e1001680.
  95. Kleiner, D.E., Brunt, E.M., Van Natta, M. et al. Design and validation of a
histological scoring system for nonalcoholic fatty liver disease. Hepatology,
2005; 41:1313–21.
  96. Brunt, E.M., Kleiner, D.E., Wilson, L.A. et al. Nonalcoholic fatty liver dis-
ease (NAFLD) activity score and the histopathologic diagnosis in NAFLD:
distinct clinicopathologic meanings. Hepatology, 2011; 53:810–20.
  97. Takyar, V., Etzion, O., Heller, T. et al. Complications of percutaneous liver
biopsy with Klatskin needles: a 36‐year single‐centre experience. Aliment
Pharmacol Ther, 2017; 45:744–53.
 98. Zhang, Y.N., Fowler, K.J., Hamilton, G. et  al. Liver fat imaging‐a clinical
overview of ultrasound, CT, and MR imaging. Br J Radiol, 2018; 91:20170959.
  99. Hernaez, R., Lazo, M., Bonekamp, S. et al. Diagnostic accuracy and relia-
bility of ultrasonography for the detection of fatty liver: a meta‐analysis.
Hepatology, 2011; 54:1082–90.
100. Karlas, T., Petroff, D., Sasso, M. et al. Individual patient data meta‐analysis
of controlled attenuation parameter (CAP) technology for assessing steato-
sis. J Hepatol, 2017; 66:1022–30.
101. Noureddin, M., Lam, J., Peterson, M.R. et al. Utility of magnetic resonance
imaging versus histology for quantifying changes in liver fat in nonalcoholic
fatty liver disease trials. Hepatology (Baltimore, Md), 2013; 58:1930–40.
102. Wong, V.W., Adams, L.A., De Ledinghen, V. et al. Noninvasive biomarkers
in NAFLD and NASH ‐ current progress and future promise. Nat Rev
Gastroenterol Hepatol, 2018; 15:461–78.
103. Kwok, R., Tse, Y.K., Wong, G.L. et al. Systematic review with meta‐analy-
sis: non‐invasive assessment of non‐alcoholic fatty liver disease‐‐the role of
transient elastography and plasma cytokeratin‐18 fragments. Aliment
Pharmacol Ther, 2014; 39:254–69.
104. Feldstein, A.E., Lopez, R., Tamimi, T.A. et al. Mass spectrometric profiling
of oxidized lipid products in human nonalcoholic fatty liver disease and
nonalcoholic steatohepatitis. J Lipid Res, 2010; 51:3046–54.
105. Younossi, Z.M., Loomba, R., Anstee, Q.M. et al. Diagnostic modalities for
nonalcoholic fatty liver disease, nonalcoholic steatohepatitis, and associ-
ated fibrosis. Hepatology, 2018; 68:349–60.
106. Hannah, W.N., Jr. and Harrison, S.A. Noninvasive imaging methods to
determine severity of nonalcoholic fatty liver disease and nonalcoholic
steatohepatitis. Hepatology, 2016; 64:2234–43.
107. Rotman, Y. and Sanyal, A.J. Current and upcoming pharmacotherapy for
non‐alcoholic fatty liver disease. Gut, 2017; 66:180–90.
108. Promrat, K., Kleiner, D.E., Niemeier, H.M. et al. Randomized controlled
trial testing the effects of weight loss on nonalcoholic steatohepatitis.
Hepatology, 2010; 51:121–9.
109. Vilar‐Gomez, E., Martinez‐Perez, Y., Calzadilla‐Bertot, L. et  al. Weight
loss through lifestyle modification significantly reduces features of nonal-
coholic steatohepatitis. Gastroenterology, 2015; 149:367–78.e5.
110. Han, M.a.T., Altayar, O., Hamdeh, S. et al. Rates of and factors associated
with placebo response in trials of pharmacotherapies for nonalcoholic
 52:  Non‐alcoholic Fatty Liver Disease: Mechanisms and Treatment 681
s­teatohepatitis: systematic review and meta‐analysis. Clin Gastroenterol
Hepatol, 2010;17(4):616–29.
111. Lassailly, G., Caiazzo, R., Gnemmi, V. et al. Regression of fibrosis after
disappearance of nash in morbidly obese patients: a prospective bariatric
surgery cohort with sequential liver biopsies. Hepatology, 2018; 68:44a‐5a.
112. Sanyal, A.J., Chalasani, N., Kowdley, K.V. et al. Pioglitazone, vitamin E, or
placebo for nonalcoholic steatohepatitis. New Eng J Med, 2010; 362:1675–85.
113. Cusi, K., Orsak, B., Bril, F. et  al. Long‐term pioglitazone treatment for
patients with nonalcoholic steatohepatitis and prediabetes or type 2 diabe-
tes mellitus. Ann Int Med, 2016; 165:305.
114. Ratziu, V., Harrison, S.A., Francque, S. et al. Elafibranor, an agonist of the
peroxisome proliferator—activated receptor—α and −δ, induces resolution
of nonalcoholic steatohepatitis without fibrosis worsening. Gastroenterology,
2016; 150:1147–59.s
115. Neuschwander‐Tetri, B.A., Loomba, R., Sanyal, A.J. et  al. Farnesoid X
nuclear receptor ligand obeticholic acid for non‐cirrhotic, non‐alcoholic
steatohepatitis (FLINT): a multicentre, randomised, placebo‐controlled
trial. Lancet, 2015; 385:956–65.
116. Harrison, S.A., Rinella, M.E., Abdelmalek, M.F. et al. NGM282 for treat-
ment of non‐alcoholic steatohepatitis: a multicentre, randomised, double‐
blind, placebo‐controlled, phase 2 trial. Lancet, 2018; 391:1174–85.
117. Harrison, S.A., Trotter, J.F., Paredes, A.H. et al. NGM282 rapidly improves
NAFLD activity score (NAS) and fibrosis in 12 weeks in patients with
biopsy‐confirmed nonalcoholic steatohepatitis (NASH): results of a phase
2 multi‐center dose finding study. Hepatology, 2018; 68:66a‐a.
118. Lavine, J.E., Schwimmer, J.B., Van Natta, M.L. et al. Effect of vitamin E or
metformin for treatment of nonalcoholic fatty liver disease in children and
adolescents. JAMA, 2011; 305:1659.
119. Vilar‐Gomez, E., Vuppalanchi, R., Gawrieh, S. et al. Vitamin E improves
transplant‐free survival and hepatic decompensation among patients with
NASH and advanced fibrosis. Hepatology.2018.
120. Schürks, M., Glynn, R.J., Rist, P.M. et al. Effects of vitamin E on stroke
subtypes: meta‐analysis of randomised controlled trials. BMJ, 2010;
341:c5702.
121. Klein, E.A., Thompson, I.M., Jr., Tangen, C.M. et al. Vitamin E and the risk
of prostate cancer: the selenium and vitamin E cancer prevention trial
(SELECT). JAMA, 2011; 306:1549–56.
122. Rotman, Y., Lingala, S., and Morris, N. Subtle iron deficiency is common
during vitamin E treatment for NAFLD. Hepatology, 2016; 64:571a‐2a.
123. Shao, N., Kuang, H.Y., Hao, M. et al. Benefits of exenatide on obesity and
non‐alcoholic fatty liver disease with elevated liver enzymes in patients
with type 2 diabetes. Diabetes Metab Res Rev, 2014; 30:521–9.
124. Armstrong, M.J., Gaunt, P., Aithal, G.P. et al. Liraglutide safety and effi-
cacy in patients with non‐alcoholic steatohepatitis (LEAN): a multicentre,
double‐blind, randomised, placebo‐controlled phase 2 study. Lancet, 2016;
387:679–90.
125. Loomba, R., Kayali, Z., Noureddin, M. et al. GS‐0976 reduces hepatic stea-
tosis and fibrosis markers in patients with nonalcoholic fatty liver disease.
Gastroenterology, 2018; 155:1463–73 e6.
 Alcoholic Liver Disease
53 Bin Gao1, Xiaogang Xiang1,2, Lorenzo Leggio3, and George F. Koob4
1
Laboratory of Liver Diseases, National Institute on Alcohol Abuse and Alcoholism, National Institutes of
Health, Bethesda, MD, USA
2
Department of Infectious Diseases, Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University,
Shanghai, China
3
Section on Clinical Psychoneuroendocrinology and Neuropsychopharmacology, National Institute on
Alcohol Abuse and Alcoholism and National Institute on Drug Abuse, National Institutes of Health,
Bethesda, MD, USA
4
National Institute on Alcohol Abuse and Alcoholism and National Institute on Drug Abuse, National
Institutes of Health, Bethesda, MD, USA

Excessive alcohol consumption leads a spectrum of liver disor-
ders from simple steatosis (fatty liver) to severe forms of liver
injury including steatohepatitis, cirrhosis, and hepatocellular
carcinoma [1, 2]. Many proinflammatory mediators, metabolic
pathways, transcriptional factors, and epigenetic factors have
been identified to contribute to the development and progression
of alcoholic liver disease (ALD), which were discussed in detail
in the previous edition of this book. In this chapter, we briefly
discuss ethanol metabolism, ALD pathogenesis, clinical diag-
nosis and treatment of ALD, and in more detail recent advances
in the pathogenesis and treatment of ALD. We also discuss the
neurobiology of alcohol use disorder (AUD) and treatment of
this disorder in patients with ALD.
ETHANOL METABOLISM
The liver is the major organ for converting ethanol into acetalde-
hyde via the three enzymatic pathways [3]. The liver expresses
high levels of alcohol dehydrogenase (ADH), which is the most
important enzyme for converting ethanol to acetaldehyde.
Human ADH genes encode at least seven isozymes including
ADH1–7 with the major forms of ADH1, ADH2, and ADH3 in
human liver. ADH exists in cytosol and oxidizes ethanol into
acetaldehyde and reduces NAD+ to NADH, which generates a
highly reduced cytosolic environment in cells. Normal liver also
expresses high levels of the cytochrome P450 isozyme CYP2E1,
which is located in the endoplasmic reticulum and is  highly
induced after chronic ethanol consumption. CYP2E1 can oxi-
dize ethanol into acetaldehyde with conversion of
NADPH+H++O2 to NADP++2H2O, thereby generating reactive
oxygen species (ROS) and free radicals that induce lipid peroxi-
dation to promote adduct formation with several macromole-
cules. Finally, another enzyme that can oxidize ethanol is
catalase, which is located in the peroxisome. However, this
pathway is considered a minor pathway of ethanol oxidation
in the liver.
The second step of ethanol metabolism is to convert acetal-
dehyde into acetate by aldehyde dehydrogenase (ALDH) [3].
ALDH also exists in multiple forms with ALDH2 being a key
isoform to metabolize acetaldehyde. ALDH2 is expressed at
the highest levels in the liver and is located in the mitochon-
dria. Acetate generated from acetaldehyde metabolism in
hepatocytes is rapidly secreted into the circulation and is fur-
ther converted into acetyl coenzyme A (acetyl‐CoA) by acetyl‐
CoA synthetases. Acetyl‐CoA then enters into the citric acid
cycle (Krebs cycle) and is finally converted to carbon dioxide
and water.
Human ADH and ALDH2 genes exist as polymorphisms with
the most relevant ALDH2 variants including the ALDH2*1 and
ALDH2*2 alleles. The ALDH2*1 allele encodes active acetalde-
hyde metabolizing enzyme with G at nucleotide position 42421
with corresponding glutamate at amino acid position 487;
whereas the ALDH2*2 allele encodes inactive enzyme with A at
position 42421 and lysine at position 487. People with homozy-
gous ALDH2*1/1 have high levels of acetaldehyde metaboliz-
ing activity in the liver; whereas individuals with homozygous
ALDH2*2/2 have very low or undetectable enzyme activities.
Individuals with heterozygous ALDH2*2/1 have approximately
80–90% lower enzyme activity because ALDH2 is the isote-
tramer enzyme and all four subunits of ALDH2 are needed to be
normal for keeping its full activity. Individuals with the inactive
variants of ALDH2 exhibit acetaldehyde accumulation after

The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
 53:  Alcoholic Liver Disease 683
alcohol consumption, and present an acetaldehyde‐mediated
“flushing syndrome,” which includes facial flushing, palpita-
tions, drowsiness, and other unpleasant symptoms.
SPECTRUM OF ALD
ALD has a broad spectrum of disorders from simple fatty liver
(steatosis) to severe forms of steatohepatitis, cirrhosis, and hepa-
tocellular carcinoma (HCC) [1, 2]. Fatty liver occurs in almost
all chronic heavy drinkers with early fat accumulation in zone 3
(perivenular) hepatocytes, which can extend to zone 2 and even
zone 1 (periportal) hepatocytes when liver damage becomes
severe. Steatohepatitis (steatosis and inflammation) and cirrho-
sis are detected in approximately 20–40% and 8–20% of chronic
heavy drinkers, respectively. Chronic excessive drinking can
also cause HCC, which occurs mostly after cirrhosis in approxi-
mately 3–10% of excessive drinkers. Many patients with chronic
ALD are asymptomatic with fully compensated and preserved
liver function, but some of them may develop episodes of super-
imposed alcoholic hepatitis (AH) with obvious jaundice, fever,
abdominal pain, anorexia, and weight loss [4]. AH has high
short‐term ­mortality especially in those with cirrhosis [4].
GENETIC FACTORS, COMORBIDITY,
AND ALD
Although almost all heavy drinkers develop fatty liver, only
about 35% of them develop advanced ALD, which is probably
because other risk factors are involved in the pathogenesis of
ALD. Indeed, gender, obesity, dietary factors, drinking patterns,
non‐sex‐linked genetic factors, and cigarette smoking have been
suggested to affect susceptibility to ALD.
Gender
Female sex is a well‐documented risk factor for susceptibility to
ALD in rodent models. It is believed that the increased risk
among women is probably due to lower levels of gastric ADH,
a higher proportion of body fat, and the presence of estrogens.
Obesity and dietary factors
Obesity is another important risk factor that accelerates ALD
development and progression in alcohol use disorder (AUD).
Experimental studies indicate that acute or chronic ethanol feed-
ing and obesity synergistically induced liver injury and inflam-
mation via the activation of endoplasmic reticulum stress, hepatic
macrophage and neutrophilic infiltration, and lipotoxicity. For
example, acute ethanol gavage induces acute steatohepatitis in
obese mice induced by high‐fat diet feeding but not in chow‐fed
mice [5]. Acute ethanol gavage upregulates the expression of
chemokine (C‐X‐C motif) ligand 1 (CXCL1) mRNA in the liver
but not in other organs by 30‐fold in obese mice, which attracts
neutrophil accumulation in the liver and subsequently causes
liver injury [5]. In addition, intragastric overfeeding of mice with
a high-fat diet and coadministration of ­ethanol synergistically
induced an elevation of serum alanine aminotransferase (ALT)
levels, hepatic steatosis, liver inflammation, and fibrosis via the
induction of nitrosative, endoplasmic reticulum, and mitochon-
drial stress and upregulation of hepatic Toll‐like receptor 4 [6].
The type of fat (saturated versus unsaturated) may also affect
the outcome of ALD. Saturated fat has been shown to be protec-
tive against ALD, while unsaturated fat is deleterious for ALD
in animal models [7]. However, how dietary factors affect
ALD  in humans is probably complex and remains poorly
characterized.
Drinking patterns
During the last five years, one of the major advances in the field
of ALD research was the discovery that acute ethanol gavage
caused significant neutrophilia, liver injury, and inflammation
in chronically ethanol‐fed mice [8–11] and in high‐fat diet‐fed
mice [5]. In agreement with these findings from animal models,
clinical studies show that individuals with AUD with recent
excessive alcohol consumption have higher levels of circulating
neutrophils, serum alanine transaminase (ALT), and aspartate
transaminase (AST) compared with individuals with AUD with-
out recent drinking [12]. The levels of circulating neutrophils
correlate positively with serum ALT and AST in individuals
with AUD with recent excessive drinking, suggesting that recent
excessive drinking elevates circulating neutrophils and subse-
quently induces liver injury [12]. In addition, early studies sug-
gest that frequent heavy drinking that begins at an early age is a
risk factor for the development of severe forms of ALD com-
pared with episodic or binge drinking [13]. However, a recent
prospective cohort study suggests that recent alcohol drinking
rather than earlier in life is associated with risk of alcoholic
­cirrhosis [14].
Non‐sex‐linked genetic factors
Many non‐sex‐linked genetic factors have been implicated in
affecting the susceptibility to advanced ALD. These include
variations in genes that encode antioxidant enzymes, cytokines
and other inflammatory mediators, and alcohol‐metabolizing
enzymes. However, the evidence that links most of these genetic
factors to ALD is weak, only a few of these factors seem to be
strongly associated with ALD. For example, a single‐nucleotide
variant (I148M) in human patatin‐like phospholipase domain‐
containing 3 gene (PNPLA3), known as I148M variant, is one of
the best characterized variants and has strong correlations with
increased risk of steatosis, fibrosis, and cirrhosis in various
types of liver diseases including ALD [15].
Viral hepatitis
Viral hepatitis infection is a leading cause of chronic liver dis-
ease worldwide. The synergistic effects of excessive alcohol
drinking and viral hepatitis B or C infection on liver disease
progression have been well documented. It has been reported
that excessive alcohol drinking markedly accelerates liver fibro-
sis and HCC development and progression in patients with viral
hepatitis C infection.
684 THE LIVER:  ALCOHOLIC STEATOHEPATITIS (ASH) AND ALCOHOLIC HEPATITIS (AH)
Non‐alcoholic fatty liver disease (NAFLD)
Excessive alcohol consumption likely worsens NAFLD pro-
gression; while the reports regarding the effects of moderate
alcohol drinking on NAFLD have been controversial [16].
Current clinical data are not sufficient to support a recommen-
dation for or against moderate alcohol consumption in
NAFLD [16].
Other comorbidities
Long‐term alcohol consumption may also accelerate liver dis-
ease progression in patients with human immunodeficiency
virus infection, hemochromatosis, and so on. A greater under-
standing of the interaction between alcohol and these comorbid
factors could help us design better therapies for the treatment of
chronic liver disease.
ALCOHOLIC FATTY LIVER
Fat accumulation in the liver (fatty liver) is the earliest response
of the liver to acute and chronic alcohol drinking [17]. Fatty
liver is diagnosed when the liver fat weight is greater than 5 to
10% of the liver’s weight and is characterized by the accumula-
tion of triglycerides, phospholipids, cholesterol esters, and other
types of fat in hepatocytes. Although alcohol drinking causes fat
accumulation in the liver, there is no evidence that ethanol and
its metabolite acetaldehyde directly participate in biosynthesis
of fatty acids (FAs). Acetate, the metabolite of acetaldehyde,
can be converted to acetyl‐CoA by acetyl‐CoA synthetases with
acetyl‐CoA synthetase 2. Once it is formed, acetyl‐CoA enters
the citric acid cycle (Krebs cycle) and contributes to fatty acid
biosynthesis. However, how much this acetate–acetyl‐CoA
pathway directly contributes to the formation of alcoholic fatty
liver is not clear because acetate derived from ethanol metabo-
lism in hepatocytes is rapidly secreted into the circulation and
does not maintain high concentrations in hepatocytes, and the
liver does not express the mitochondrial acetyl‐CoA synthetase
2 that is a major form for converting acetate into acetyl‐CoA.
Emerging evidence suggests that alcohol consumption pro-
motes fat mobilization to the liver and de novo FA biosynthesis
but inhibits fat clearance and FA β‐oxidation in the liver, result-
ing in fat accumulation in the liver [1, 17]. Early studies demon-
strated that acute alcohol administration can increase the supply
of dietary lipids to the liver from the small intestine by increasing
the intestinal lymph flow and output of both dietary and nondie-
tary lipids, but this effect was less evident after chronic alcohol
treatment. Recent studies focused on how alcohol consumption
increases the supply of lipids from adipose tissues to the liver
[18]. It has been shown that alcohol consumption induces lipoly-
sis [19] and adipocyte death [20] and subsequently elevates cir-
culating FAs and their accumulation in the liver. In addition to
augmenting FA supply, alcohol can also increase fat accumula-
tion by promoting lipid formation and stimulating de novo FA
biosynthesis in hepatocytes via the upregulation of fat‐specific
protein 27 (FSP27) and sterol  ­regulatory element‐binding pro-
tein 1c (SREBP‐1c), respectively [1, 9, 17]. FSP27 is expressed
at high levels in adipose tissues, playing a critical role in promot-
ing lipid droplet formation. In normal liver, FSP27 is expressed
at very low levels but highly upregulated post alcohol consump-
tion, especially acute alcohol consumption [9]. Upregulation of
hepatic FSP27 by acute ethanol consumption is attributable to
activation of hepatic endoplasmic reticulum stress, which acti-
vates the cyclic AMP‐responsive element‐binding protein H
(CREBH) and subsequent nuclear translocation of nCREBH [9].
The important role of FSP27 in inducing alcoholic fatty liver is
supported by the fact that disruption of the Fsp27 gene prevents
the development of alcoholic fatty liver [9]. SREBP‐1c is a mas-
ter transcription factor for controlling hepatic expression of
many lipogenic genes that encode proteins and enzymes to
induce FA biosynthesis. Disruption of the Srebp‐1c gene mark-
edly prevents alcoholic fatty liver in mice, suggesting a critical
role of SREBP‐1c in inducing alcoholic fatty liver [21]. Alcohol
consumption can directly upregulate hepatic SREBP‐1c expres-
sion via its metabolite acetaldehyde [22], or indirectly stimu-
late its expression by modulating multiple factors and pathways
that control hepatic expression of SREBP‐1 (see references
therein [1, 17]). In contrast to promotion of FA biosynthesis,
alcohol attenuates FA β‐oxidation, which is another critical
mechanism responsible for the formation of alcoholic fatty
liver. Early studies revealed that alcohol treatment increases the
ratio of reduced nicotinamide adenine dinucleotide/oxidized
nicotinamide adenine dinucleotide in the liver and conse-
quently disrupts the mitochondrial β‐oxidation of FAs and
results in fat accumulation in hepatocytes [23]. Recent studies
unveiled an additional mechanism by which ethanol inhibits
FA β‐oxidation, which is inactivation of peroxisome prolifera-
tor‐activated receptor (PPAR)‐α, a nuclear hormone receptor
that upregulates expression of a range of genes involved in FA
transport and oxidation [24].
Alcohol consumption can also promote hepatic fat accumula-
tion by inhibiting 5’ AMP‐activated protein kinase (AMPK) and
deregulating autophagy. AMPK plays a critical role in control-
ling the activity of several enzymes and transcriptional factors
that regulate fat metabolism [25]. For example, AMPK inhibits
acetyl‐CoA carboxylase (ACC) and SREBP activity and conse-
quently attenuates FA biosynthesis but enhances its oxidation
[25]. Alcohol exposure can suppress AMPK activity and subse-
quently enhance FA synthesis and reduces FA oxidation, result-
ing in alcoholic fatty liver [26]. Autophagy plays an important
role in clearing lipid droplets in hepatocytes, and chronic alco-
hol consumption inhibits autophagy, thereby reducing lipid
clearance and causing hepatic steatosis [27]. However, acute
alcohol intake seems to activate autophagy, which may play a
compensatory role in preventing the development of fatty liver
during the early stages of alcoholic liver injury [28].
ALCOHOLIC STEATOHEPATITIS (ASH)
AND ALCOHOLIC HEPATITIS (AH)
Accumulation of fat in hepatocytes is the first response of the
liver to alcohol consumption [17]; however, alcohol consump-
tion may also cause hepatocyte damage and inflammation,
which is defined as ASH [4]. ASH is a histological concept that
 53:  Alcoholic Liver Disease 685

is characterized by steatosis, hepatocyte death, ballooning
degeneration, neutrophilic infiltration, Mallory–Denk bodies,
and chicken‐wire fibrosis [4]. Most patients with ASH are
asymptomatic for prolonged periods of time and are defined as
walking ASH or subclinical ASH, but some patients with ASH
develop a significant clinical syndrome, which is defined as AH
(see New therapeutic targets for AH) [4]. Over the last four dec-
ades, multiple mechanisms have been identified to contribute to
ASH caused by alcohol drinking (Figure  53.1). First, alcohol
and its metabolite acetaldehyde can directly cause hepatotoxic-
ity by generating ROS and inducing endoplasmic reticulum
stress and mitochondrial dysfunction [29]. Second, damaged
hepatocytes caused by alcohol release damage‐associated
molecular patterns (DAMPs) induce liver inflammation, such as
neutrophil infiltration [30]. Third, chronic alcohol consumption
induces gut bacterial overgrowth and dysbiosis and increases
gut permeability, resulting in elevation of bacteria and their
related products in the liver and circulation and subsequent
inflammation [31].
AH is a severe form of ALD that presents a form of acute‐on‐
chronic liver failure in individuals with AUD with underlying
ALD including ASH and cirrhosis [4, 32, 33]. AH is character-
ized by an abrupt rise in jaundice (serum bilirubin levels) and
liver‐related complications and is often accompanied by fever,
abdominal pain, anorexia, and weight loss [33]. In addition, por-
tal hypertension, ascites, encephalopathy, and variceal bleeding
are associated with severe cases of AH. Most patients with AH
have pre‐existing alcoholic cirrhosis, while some patients with
mild underlying liver disease or silent ALD can also develop
AH. Diagnosis of AH is based on clinical syndromes with an
episode of jaundice with serum bilirubin levels greater than 3
mg dL−1 and AST levels greater than 50 IU mL−1 in chronic
AUD with heavy drinking greater than 5 years [33]. Jaundice is
an essential syndrome for the diagnosis of AH, but not all epi-
sodes of jaundice in AUD with chronic ALD are attributed to
AH [4]. For example, severe sepsis, biliary obstruction, diffuse
liver cancer, drug‐induced liver injury, and ischemic hepatitis
(i.e. due to massive bleeding or cocaine use) can also cause an
episode of jaundice and liver decompensation in AUD, which
should not be diagnosed as AH [4]. Patients with severe AH
especially those with cirrhosis usually have high short‐term
mortality with a mortality rate of 20–50% at three months [32].
Liver histology analysis of AH revealed infiltration of a large
number of inflammatory cells in the liver with predominant
neutrophils and macrophages [30]. Microarray analysis has
identified that a large number of inflammatory mediators are
upregulated in the liver from AH patients [34]. In addition to
massive liver inflammation, a systemic inflammatory response
is also a key feature of AH and correlates with disease severity
[35]. The inflammation in most AH patients is likely triggered
by recent excessive binge drinking, bacterial infection, or both
[4]. Recent excessive binge drinking was reported in many

Figure 53.1  Pathogenesis and molecular mechanisms that trigger liver failure, systemic inflammatory response, and multiorgan failure in alco-
holic hepatitis. Excessive binge drinking in AUD with underlying ALD induces hepatocellular damage, induces systemic inflammation, and impairs
liver regeneration, resulting in liver failure and other complications in alcoholic hepatitis. Modified from [4] and reproduced with permission
from Elsevier.
686 THE LIVER:  EMERGING MECHANISMS

patients with AH, and can cause elevation of serum ALT and promoting HSC activation and liver fibrosis, but ­activation of
AST levels as well as circulating neutrophils in AUD [12]. It is some components of innate immunity negatively regulates HSC
possible that excessive binge drinking causes massive hepato- activation and suppresses liver fibrosis. For example, activation
cyte damage, which releases a large amount of DAMPs that trig- of natural killer (NK) cells can directly kill activated HSCs and
ger systemic inflammation in patients with AUD. For example, produce IFN‐γ that blocks HSC proliferation and induces HSC
damaged hepatocytes can release mitochondrial DNA, which death. Such an antifibrotic effect of NK cells is markedly sup-
can activate various types of inflammatory cells via Toll‐like pressed after chronic alcohol consumption, thereby accelerating
receptors [11]. Bacterial infection is detected in up to 50% of liver fibrosis [38].
AH patients and is associated with high mortality in these
patients [36]. Pathogen-associated molecular patterns (PAMPs)
produced during bacterial infection including endotoxin, can
activate macrophages/Kupffer cells to produce a wide variety of
ALCOHOLIC LIVER CANCER
inflammatory mediators in AH [35].
Liver cirrhosis derived from all etiologies of liver pathology
Normal liver has the great ability to regenerate after injury or
including alcohol consumption is a major risk factor for liver
loss of tissue; however, hepatocytes in the injured liver in AH
cancer, mainly HCC. The mechanisms by which cirrhosis pro-
patients seem unable to efficiently regenerate. Instead, liver
motes HCC are not fully understood. Several hypotheses have
­progenitor cells are markedly expanded and accumulated in AH
been proposed. First, cirrhosis is associated with telomere
patients, and the number of these cells positively correlates with
shortening, which can induce chromosomal instability and
disease severity. It is likely that these liver progenitor cells are
mutations of tumor suppressors and oncogenes. Second, cirrho-
unable to fully differentiate into functional hepatocytes and do
sis is associated with chronic inflammation and elevation of
not contribute to liver regeneration. Thus, impaired liver regen-
many growth factors and cytokines that stimulate HCC growth.
eration is probably an important mechanism contributing to
In addition to these common mechanisms, there are some
liver failure in patients with severe AH (Figure 53.1).
exceptional mechanisms that are believed to specifically con-
tribute to alcoholic liver cancer development. Acetaldehyde, an
ethanol metabolite, is considered a carcinogen with mutagenic
properties. It is well documented that ALDH2‐deficient indi-
ALCOHOLIC LIVER FIBROSIS
viduals have high levels of acetaldehyde after alcohol drinking
and are associated with an increased risk of esophageal cancers;
Fibrosis is a scarring response of the liver to chronic liver dam-
however, the ­association between HCC and individuals with
age caused by virtually all etiologies of liver pathology includ-
AUD with ALDH2 deficiency is less clear. It is known that acet-
ing excessive alcohol drinking. Although all types of chronic
aldehyde is electrophilic and forms an adduct with DNA and
liver injury cause liver fibrosis, alcoholic liver fibrosis has a
interstrand crosslinks, causing DNA damage. In addition, acet-
characteristic pattern of pericellular and sinusoidal chicken‐
aldehyde suppresses DNA repair enzyme activity and subse-
wire fibrosis, which could be extended to panlobular fibrosis
quently blocks DNA repair. Thus, acetaldehyde not only causes
when this characteristic pattern becomes diffused [37]. Liver
DNA damage but also prevents DNA repair, contributing to
fibrosis is characterized by excessive accumulation of collagen
hepatocarcinogenesis. Moreover, heavy drinking can induce
and other extracellular matrix (ECM) proteins. Most of these
aberrant DNA methylation including genome‐wide hypometh-
ECM proteins are produced by activated hepatic stellate cells
ylation, leading to chromosomal instability and subsequently
(HSCs). Several other types of cells can also produce ECM
promoting the development of liver cancer.
­protein, but to a much lesser extent. These cells include portal
Excessive alcohol drinking is known to cause broad immune
fibroblasts and bone marrow‐derived myofibroblasts. Activation
suppression including attenuation of antitumor immunity, which
of HSCs, a key step in developing liver fibrosis, is controlled by
is probably another important mechanism that can accelerate
many inflammatory mediators and growth factors that are com-
HCC development and progression.
monly elevated in ALD and other types of liver diseases [37].
For example, transforming growth factor (TGF)‐β is one of the
most important mediators that induces HSC transformation;
while platelet‐derived growth factor (PDGF) is a critical growth EMERGING MECHANISMS
factor that stimulates HSC proliferation. Both TGF‐β and PDGF
are elevated in ALD in patients and animal models of alcoholic
Gut microbiome
liver fibrosis [37]. In addition, there are some unique mecha-
nisms that promote alcoholic liver fibrosis [37]. First, acetalde- An important role of gut bacteria‐derived LPS in the develop-
hyde, the first ethanol metabolite, can directly stimulate and ment of ALD in animal models was reported more than two
maintain HSC activation by stimulating multiple signaling path- ­decades ago. Recent studies suggest that the gut microbiome
ways and transcription factors. Second, ALD patients exhibit plays a complex role in the pathogenesis of ALD [31]. Chronic
elevated lipopolysaccharides (LPSs) due to increased gut bacte- alcohol consumption can cause bacterial overgrowth in the large
rial growth and gut permeability. LPS is a well‐known mediator and small intestines and dysbiosis in humans and animals [31].
that can directly induce HSC activation, thereby playing an In general, changes in gut microbiome post chronic alcohol
important role in promoting alcoholic liver fibrosis. Third, acti- consumption are complex but are often associated with a
­
vation of innate immunity is known to play a critical role in decrease in “good” bacteria (such as Lactobacillus spp.) but an
 53:  Alcoholic Liver Disease 687
increase in “bad” bacteria (such as Enterobacteriaceae). These
changes could cause pathological bacterial translocation, altera-
tions of intestinal metabolites and bile acid metabolism, result-
ing in elevation of systemic levels of gut‐derived microbial
production and subsequent hepatocellular damage and liver
inflammation [31]. In addition, a recent study suggests that gut
fungi may also play a role in promoting ALD [39]. Chronic
alcohol consumption increases microbiota populations in gut
and treatment with antifungal drugs attenuated intestinal fungal
overgrowth and ameliorated ALD in mice. Patients with ALD
have reduced intestinal fungal diversity and Candida overgrowth
and increased systemic exposure and immune response to myc-
obiota. The levels of these responses correlate with m ­ ortality,
suggesting that altered gut mycobiomes contribute to ALD
pathogenesis [39]. Moreover, transplantation of intestinal
microbiota from a patient with severe AH induced greater liver
inflammation and injury in mice than transplantation of these
from an individual with AUD without AH [40]. Thus, manipu-
lating the intestinal microbiome (both bacteria and fungi) might
be an effective strategy for preventing and treating ALD, and
fecal microbiota transplant might be a potential therapeutic
option for ALD.
Epigenetics
Epigenetics is the investigation of heritable alterations of phe-
notype (appearance) or gene expression that are caused by
mechanisms other than alterations in DNA coding sequences.
Epigenetic modifications include DNA methylation, histone
modifications, and RNA‐based mechanisms. Alcohol consump-
tion is known to affect metabolism of methionine and thereby
DNA methylation [41]. In the liver, homocysteine is methylated
to methionine and then S‐adenosylmethionine (SAMe) in a
transmethylation reaction that is catalyzed by methionine aden-
osyltransferase. SAMe is a principal methyl donor in methyla-
tion and plays a critical role in inducing DNA and histone
methylation. Excessive alcohol consumption reduces hepatic
levels of SAMe and DNA and histone methylation, which sub-
sequently upregulates the expression of genes that control the
endoplasmic reticulum stress response and ALD [41].
miRNAs
miRNAs (small non‐coding RNA molecules with 19–25 nucle-
otides) play a critical role in controlling gene expression by
inducing RNA silencing and by altering post‐transcriptional
regulation of gene expression. Many miRNAs have been shown
to play important roles in the development of ALD. For exam-
ple, hepatocyte‐specific miR122 seems to play a role in protect-
ing against ALD by reducing hypoxia‐inducible factor 1 alpha
(HIF1α) mRNA. Hepatic miR122 is downregulated in liver
samples from patients with ALD and ethanol‐fed mice, thereby
exacerbating liver injury and inflammation [42]. In contrast,
neutrophil‐specific miR‐223 is upregulated in the liver and
serum from ethanol‐fed mice and individuals with AUD com-
pared to their controls [12]. Such elevations likely play a com-
pensatory role in limiting neutrophil activation via inhibition of
the IL‐6–p47phox pathway [12].
Extracellular vesicles (EVs)
EVs are nanometer‐sized, membrane‐bound, extracellular milieu
vesicles derived from cells, playing an important role in connect-
ing cell to cell communication and in promoting inflammation in
a variety of diseases, including liver diseases. EVs can be
grouped as three types based on their cellular biogenesis and
sizes: exosomes (40–150 nm diameter in size), microvesicles/
microparticles (MPs; 50–1000 nm diameter), and apoptotic bod-
ies (greater than 500 nm in diameter). EVs exert their functions
by delivering their contents from one cell to another. For instance,
stressed hepatocytes can activate macrophages and neutrophils
via the release of EVs that contain lipids, proteins, chemokines,
and nucleic acids (e.g. mitochondrial DNA [mtDNA]), playing a
role in the pathogenesis of ALD [43]. Patients with AUD or ALD
had an elevated number of EVs in the circulation, which are
enriched in mtDNA and correlate with elevated circulating neu-
trophils and serum levels of ALT [11]. Compared with pair‐fed
control mice, mice post chronic or chronic‐plus‐binge ethanol
feeding also have elevated circulating EVs, which are enriched
in mRNAs, proteins, mtDNAs, and likely promote liver inflam-
mation and injury in ALD [11, 44]. In addition to being an
important player in ALD, EVs may be used as potential biomark-
ers for diagnosis because EVs from ALD contain specific signa-
tures of proteins, RNAs, and DNAs [43].
Adaptive immunity
The role of innate immunity in the pathogenesis of ALD has
been well documented; however, the involvement of adaptive
immunity in ALD remains obscure. It is generally believed that
excessive alcohol consumption induces oxidative stress, which
subsequently generates lipid peroxidation products, such as
malondialdehyde and 4‐hydroxynonenal. These products can
modify many proteins and induce formation of protein adducts
that can act as neoantigens to induce the activation of adaptive
immunity [45]. Activation of adaptive immunity has not been
characterized or detected in animal models of ALD, but it was
reported in patients with ALD [45]. For example, patients with
ALD are associated with increased levels of circulating antibod-
ies against lipid peroxidation adducts and increased numbers of
intrahepatic T cells [45]. The infiltration of intrahepatic T cells
in ALD is not just bystander activation, and these T cells present
a pronounced oligoclonal nature along with ALD‐associated
clonotypes as demonstrated in a recent high‐throughput T‐cell
receptor sequencing study [46]. If activation of adaptive immu-
nity is a major driver of disease progression in some ALD
patients, immunosuppressive therapy may be required to
­effectively treat these patients.
Innate‐like T cells
The roles of Kupffer cells/macrophages in the pathogenesis of
ALD have been extensively studied over the last two decades.
Recent studies suggest that nature killer T (NKT) cells and
mucosa‐associated invariant T cells (MAIT) also play a role in
the pathogenesis of ALD. Mouse liver lymphocytes contain
approximately 30–40% natural killer T (NKT) cells, which are
a heterogeneous group of T cells that share properties of both
688 THE LIVER:  CLINICAL DIAGNOSIS AND TREATMENT OF ALD
T  cells and NK cells and rapidly produce a large number of
cytokines such as IFN‐γ, IL‐4, and so on. Several recent studies
reported that hepatic NKT cells are activated and increased
post chronic‐plus‐binge ethanol feeding. These activated NKT
cells contribute to Kupffer cell activation and alcoholic liver
injury [47, 48]. Of note, human liver lymphocytes contain a
very low number of NKT cells but are enriched with MAIT
cells, representing 20–50% of intrahepatic T cells in normal
human livers. MAIT cells are detected at low levels in com-
monly used laboratory mouse strains, accounting for less than
1% of intrahepatic T cells in mice. MAIT cells play a key role
in host defense against bacterial infection through invariant
T‐cell receptors (TCRs) that recognize microbial riboflavin/
vitamin B2 metabolites presented by the major histocompatibil-
ity complex class I‐related protein 1. Patients with severe ALD
are associated with a marked reduction of MAIT cells [49],
which may contribute to an increased risk of bacterial infection
in these patients.
Chronic‐plus‐binge ethanol
Over the last 5 years, one important advance in the field of ALD
study is the introduction of binge ethanol in chronically ethanol‐
fed mice [50] or in high‐fat diet‐fed mice [5]. This chronic‐plus‐
binge ethanol feeding model was initially called the NAAA
model [50] and was later also called the Gao‐binge model [51].
Such an ethanol challenge induces more severe alcoholic steato-
hepatitis with marked elevation of serum ALT and AST and sig-
nificant hepatic neutrophil infiltration compared with binge
ethanol alone, chronic ethanol feeding alone, and high‐fat diet
feeding alone [5, 8, 52]. Acute ethanol gavage was also intro-
duced in the chronic intragastric ethanol infusion model, which
causes marked hepatic neutrophil infiltration [53]. By using
these new animal models, researchers have identified many
novel mechanisms that contribute to acute‐on‐chronic liver
injury in ALD. These include activation of endoplasmic reticu-
lum stress, activation of inflammatory cells (e.g. neutrophils,
Kupffer cells, and NKT cells), upregulation of several factors
that promote steatosis (e.g. FSP27), and dysregulation of cell
survival/death signaling pathways (e.g. pyroptosis, apoptosis,
etc.) [10, 53]. For example, several studies suggest that neutro-
phils contribute to hepatocellular damage and hepatic injury in
this early stage of ALD that is induced by chronic‐plus‐binge
ethanol feeding or high‐fat diet‐plus‐binge ethanol feeding
[5, 8]. However, neutrophils can also promote liver repair and
are vital in the control of bacterial infection. Therefore, neutro-
phils likely play both beneficial and detrimental roles in the
­pathogenesis of ALD.
CLINICAL DIAGNOSIS AND TREATMENT
OF ALD
The diagnosis of ALD is based on the combination of clinical
and laboratory findings, including a medical history of signifi-
cant alcohol intake, physical signs of liver disease, supporting
laboratory tests of liver disease, non‐invasive liver imaging, and
invasive liver biopsy [2, 54]. Documentation of excessive
alcohol drinking and evidence of liver disease are the key diag-
nostic factors of ALD.
Medical history
Denial of excessive alcohol use or underreported alcohol intake
in the medical history is a frequent occurrence in AUD patients,
which may make diagnosis of ALD challenging in these
patients. Therefore, indirect evidence of excessive alcohol use
is always considered, such as questionnaires, information from
family members, and laboratory tests to reinforce or confirm
the suspicion of AUD. Various questionnaires for screening
­excessive alcohol use are used by clinicians, such as the CAGE
(cut‐annoyed‐guilty‐eye), MAST (Michigan Alcoholism
Screening Test), and AUDIT (Alcohol use disorders identifica-
tion test) [2]. The CAGE and MAST questionnaires are the
most commonly used.
Clinical symptoms and signs
Anorexia, nausea and vomiting, upper abdominal pain, malaise,
fatigue, dark urine, fever, confusion, and weight loss are the most
commonly presenting nonspecific symptoms of ALD. The most
common sign of AH is rapid development of jaundice; others
include fever, ascites, proximal muscle wasting, tender hepato-
megaly, and hepatic bruit. Severe cases often develop hepatic
encephalopathy and gastrointestinal bleeding. Patients with mild
to moderate ALD are often asymptomatic. In some patients, bilat-
eral parotid gland hypertrophy, muscle wasting, malnutrition,
Dupuytren’s contracture sign, and signs of symmetric peripheral
neuropathy are suggestive of harmful alcohol consumption.
Splenomegaly, gynecomastia, spider angiomas, asterixis, palmar
erythema, and digital clubbing are frequently observed in ALD
patients with cirrhosis [2]. In decompensated cirrhosis, jaundice,
ascites, peripheral edema, and hepatic encephalopathy are always
seen in addition to the cirrhotic physical ­findings [2].
Laboratory tests
There are no specific laboratory biomarkers for ALD. Elevations
of serum enzymes such as AST, ALT, alkaline phosphatase
(ALP), and GGT can only provide clues for the diagnosis of
ALD. The NIAAA Alcoholic Hepatitis Consortia proposed a
working definition of acute AH to include an AST to ALT ratio
above 1.5 with AST and ALT less than 400 U L−1 [33]. Clinically,
GGT is the most frequently used biomarker to detect previous
alcohol consumption, though it can also arise from other etiolo-
gies [55]. Routine blood tests and biochemical detection, such
as elevations of red blood cell mean corpuscular volume (RBC
MCV), white blood cell (WBC) count, GGT, AST, and immu-
noglobulin A (IgA) may indicate early ALD, while a decrease in
albumin, prolonged prothrombin time (PT), elevated total bili-
rubin (TBIL) level, and low platelet (PLT) count are indications
of advanced ALD [56]. Recent studies suggest that a new bio-
marker of apoptotic and necrotic cell death, the M30/M65 frag-
ment of cytokeratin 18 (CK18), is sensitive to liver injury in
ALD patients [57]. In some ALD patients, metabolic abnormali-
ties such as elevated levels of triglycerides and uric acid,
hypokalemia, and hyponatremia are frequently observed.
 53:  Alcoholic Liver Disease 689
Liver imaging
Ultrasound, computerized tomography, and magnetic resonance
imaging can be used to detect the presence of underlying liver
disease, but they cannot provide specific information for diag-
nosing ALD. Ultrasound is routinely used to check hepatocel-
lular carcinoma, biliary obstruction, ascites, splenomegaly,
portal hypertension, and portal vein thrombosis. Non‐contrast
computerized tomography more easily detects macroscopic fat
in the liver in fatty liver disease [58]. Magnetic resonance imag-
ing is the most sensitive and specific imaging modality for
detecting steatosis (95% sensitivity, 98% specificity) except for
patients with iron overload [59]. Recently, the clinical applica-
tions of newer imaging techniques such as transient elastogra-
phy (TE, Fibroscan®) and magnetic resonance elastography
(MRE) improve diagnostic accuracy in quantifying steatosis
and fibrosis [60]. Among all imaging modalities, ultrasound is
the most widely used due to its low cost. By detecting values of
controlled attenuation parameter (CAP value) and liver stiff-
ness, transient elastography (TE, Fibroscan) has demonstrated
good accuracy in quantifying steatosis and fibrosis, and it is a
significantly cheaper alternative [54].
Liver biopsy
Liver biopsy is not routinely recommended for the diagnosis of
ALD in most patients in the United States. However, it is
believed to be the gold standard for diagnosing and assessing
the severity of steatosis and staging of fibrosis, and it is the only
modality that is available to distinguish between simple steato-
sis and steatohepatitis [2]. Liver biopsy is invasive and mostly
­performed percutaneously. A transjugular route of biopsy is
required in patients with coagulopathy or ascites. The histologic
features of ALD on liver biopsy vary based on the extent and
stage of hepatic injury. The typical histological lesions of ALD
are large fat droplets and ballooning of hepatocytes including
Mallory–Denk bodies and mega‐mitochondria as well as neu-
trophil infiltration and intrasinusoidal chicken‐like fibrosis [60].
Macrovesicular steatosis is the earliest and most frequently seen
pattern of alcohol‐induced liver injury. Alcoholic steatohepatitis
is defined by the coexistence of steatosis, hepatocyte balloon-
ing, and neutrophil infiltration [54].
Complications in ALD
Cholestasis, fat embolism, and portal hypertension are occa-
sionally observed in patients with severe fatty liver caused by
alcohol. Alcoholic ketoacidosis is often observed in patients
with chronic alcohol consumption and concurrent malnutrition.
Patients with alcoholic cirrhosis or severe AH always have
­complications such as ascites, spontaneous bacterial peritonitis
(SBP), variceal bleeding, electrolyte disturbance, hepatorenal
syndrome (HRS), hepatic encephalopathy, and HCC [2].
Assessment of ALD
There are several prognostic scoring models for assessing
­disease severity and short‐term mortality of ALD, including the
Maddrey discriminant function (MDF), the Mayo model for
end‐stage liver disease (MELD), the Glasgow alcoholic hepati-
tis score (GAHS), the Age, bilirubin, INR, creatinine (ABIC)
score, the Lille model, and the Child–Turcotte–Pugh (CTP)
score [2]. MDF (4.6 × [patient prothrombin time – control pro-
thrombin time] + serum total bilirubin) is the most widely used
[2]. ALD patients with a MDF value greater than or equal to 32
is indicative of a high risk of short‐term mortality (30–50% at
one month) with improved short‐term clinical outcomes after
receiving corticosteroid [60].
TREATMENT OF ALD
Alcohol abstinence
The backbone therapeutic intervention for patients with ALD is
alcohol abstinence. Following abstinence, relapse is a major risk
in these patients. Continued alcohol use after diagnosis of ALD
has been shown to promote disease progression [2, 54].
Nutritional support
Malnutrition is very common in patients with severe ALD, such
as deficiencies of vitamin A, vitamin D, vitamin E, thiamine,
folate, niacin, pyridoxine, zinc, magnesium, and selenium. A
dietary intake of 1.2–1.5 g of protein kg−1 and 35–40 kcal kg−1
body weight is recommended in these patients [60].
Corticosteroids
Corticosteroids are recommended for patients with severe AH in
the absence of sepsis and infection, which can improve short‐
term survival rates in these patients [2]. However, long‐term fol-
low‐up does not reveal a significant survival difference between
those treated with corticosteroids and controls [2].
Pentoxifylline (PTX)
PTX is a competitive, nonselective, phosphodiesterase inhibitor
that can inhibit tumor necrosis factor and leukotriene synthesis,
inflammation, and innate immunity. PTX can also reduce the
development of hepatorenal syndrome (HRS) in ASH patients,
resulting in improvements in short‐term survival [61].
N‐acetyl cysteine (NAC)
NAC has antioxidant activity and is ineffective when used alone
in ALD patients. However, a randomized trial showed that the
combination of NAC with prednisolone reduced one‐month
mortality and incidence of HRS/infection [62].
Artificial liver support system
Artificial liver support systems are a therapy option for liver
failure associated with severe ALD, which has been controver-
sial for decades. Recently, there was a study that showed that the
application of an artificial liver support system in patients with
severe liver dysfunction secondary to ALD showed some bene-
fits in a subset of the patients [63].
690 THE LIVER:  NEW THERAPEUTIC TARGETS FOR ALCHOLIC HEPATITIS
Liver transplantation
Patients with end‐stage liver disease secondary to alcoholic
­cirrhosis should be considered for liver transplantation. Early
liver transplantation, which has been performed for acute severe
AH, is discussed in the following section.
NEW THERAPEUTIC TARGETS
FOR ALCHOLIC HEPATITIS
Although no new drugs for ALD have successfully been devel-
oped, many therapeutic targets were recently identified from
studies of human ALD samples and animal models, and some of
them are currently in clinical trials for the treatment of AH,
which are discussed below.
Inflammation is considered a critical factor for causing liver
damage in AH and has been actively investigated as a therapeu-
tic target for the treatment of AH [64]. For instance, immuno-
suppressive drug steroids, such as prednisolone, have been used
for the treatment of AH for more than five decades, but accumu-
lating data suggest that prednisolone treatment is beneficial for
short‐term survival but has no benefits for long‐term survival in
AH patients [65]. Steroid drugs have broad immunosuppressive
functions but are ineffective for a variety of neutrophil‐mediated
disorders (e.g. asthma, septic shock), which is, at least in part,
mediated by promoting neutrophil survival. AH is associated
with hepatocyte necrosis and infiltration of neutrophils, which
may contribute to the ineffectiveness of prednisolone treatment
in some AH patients. In addition, steroid treatment is associated
with an increased risk of bacterial infection in patients with AH
due to its broad immunosuppressive function. More specific tar-
gets for inflammation are urgently needed for the treatment of
AH. Indeed, a large number of therapeutic targets for inflamma-
tion have been recently identified and include inflammatory
cytokines and mediators, chemokines and their receptors, gut
microbiota and bacterial products, and so on [66]. Some of them
are currently being tested in clinical trials for the treatment of
AH, including IL‐1 inhibitors, apoptosis signal‐regulating
kinase 1 (ASK1) inhibitors, LPS blockers, and probiotics [66].
Preclinical studies demonstrated that IL‐1 plays an important
role in promoting ALD in animal models. Thus, IL‐1 inhibitors,
which have been approved for the treatment of several types of
inflammatory disease due to their good safety profiles and low
tendency for adverse side‐effects, are currently being tested in
clinical trials for the treatment of AH. ASK1, a serine/threonine
signaling kinase, is activated by oxidative stress, and its activa-
tion leads to activation of p38 mitogen‐activated kinase (p38
MAPK) and c‐Jun N‐terminal kinase (JNK), playing a role in
promoting hepatic inflammation, apoptosis, and fibrosis. ASK1
inhibitors are currently being tested in clinical trials for several
types of liver diseases including non‐alcoholic steatohepatitis
and AH. In addition, a large number of studies have demon-
strated that excessive alcohol consumption induces gut bacterial
overgrowth, dysbiosis, and elevation of circulating LPS, which
contribute to liver inflammation and injury in AH. Several ongo-
ing clinical trials are targeting these factors to treat AH using
probiotics, antibiotics, fecal microbiota transplantation,
hyperimmune bovine colostrum enriched with anti‐LPS anti-
bodies, and so on. Collectively, many inflammatory mediators
have been implicated in liver injury and inflammation in patients
with severe AH. Many of these ­mediators probably synergisti-
cally or additively promote liver inflammation in AH. Clinical
trials are needed to identify the inflammatory mediators that
play critical roles in the pathogenesis of ALD in patients and
can be used as targets for the ­treatment of ALD, which will take
years to complete. The immunosuppressive drug prednisolone
or other steroids will likely continue to be used for the treatment
of severe AH until more specific immunosuppressive drugs are
identified.
AH is associated not only with hepatocellular injury but
also with impairments of liver regeneration. The application of
hepatoprotective agents may provide some benefits in ALD
therapy to protect against hepatocellular damage and promote
liver regeneration. Indeed, the hepatoprotective cytokine IL‐22
is currently being tested in clinical trials for the treatment of
AH. IL‐22, which induces STAT3 activation in hepatocytes, has
many ­beneficial effects in the liver [67] but likely has minimal
side‐effects due to the restricted expression of IL‐22 receptors,
which are mainly expressed in epithelial cells but not in immune
cells. By targeting hepatocytes, IL‐22 plays an important role in
ameliorating hepatocellular damage, promoting liver regenera-
tion, and alleviating liver fibrosis [67]. In addition, IL‐22 treat-
ment may effectively impede bacterial infection and ameliorate
kidney injury, two deleterious conditions that often contribute to
death in AH patients. IL‐22 therapy is currently being tested in
clinical trials for the treatment of patients with severe AH [68].
Because severe ALD and AH are associated with multiple
problems, including inflammation, hepatocellular damage, poor
liver regeneration, and many complications, combination thera-
pies are likely to be needed to treat these severe disorders [68].
These combination therapies include the inhibition of inflamma-
tion, hepatoprotection and stimulation of liver regeneration, and
organ‐specific support in multiple organ failure for critically ill
patients. Liver transplantation is probably the only potentially
life‐saving treatment for many patients with end‐stage ALD and
multiple organ failure who do not respond to medical treatment.
Indeed, ALD now replaces hepatitis C (HCV) infection as the
leading indication for liver transplantation in the United States.
Current treatment guidelines require a six‐month period of alco-
hol abstinence; however, most patients with severe AH do not
survive six months. Thus, several centers in Europe and the
United States have started early liver transplantation with less
than six‐month abstinence, achieving excellent clinical out-
comes with great survival rates and low rates of alcohol relapse
in highly selected patients [69]. Based on these data, the
American Gastroenterological Association Clinical Practices
Update Committee gave best practice advice: “Patients with
severe AH, particularly those with a MELD score > 26 with good
insight into their AUD and good social support should be referred
for evaluation for liver transplantation, as the 90‐day mortality
rate is very high” [70]. However, this recommendation has not
been broadly adopted yet by the majority of transplantation cent-
ers and is still challenged by several issues, such as a shortage of
donor livers, competition for donor livers by other types of liver
diseases, alcohol relapse post transplantation, and potentially
spontaneous recovery in these patients after abstinence.
 53:  Alcoholic Liver Disease 691

NEUROBIOLOGY OF ALCOHOL USE
DISORDER (AUD)
Definitions and conceptual framework
for the neurocircuitry of AUD
AUD can be defined as a chronically relapsing disorder that is
characterized by a compulsion to seek and take the drug (alcohol),
loss of control in limiting drug (alcohol) intake, and the emergence
of a negative emotional state (e.g. dysphoria, anxiety, irritability
[hyperkatifeia]), reflecting a motivational withdrawal syndrome,
when access to the drug (alcohol) is prevented [71]. These key ele-
ments embrace most of the symptoms of AUD as expressed in
moderate to severe AUD in the Diagnostic and Statistical Manual
of Mental Disorders, 5th edition (DSM‐5; American Psychiatric
Association, 2013). AUD and addiction in general have been
­heuristically framed as a three‐stage cycle: binge/intoxication,
withdrawal/negative affect, and preoccupation/anticipation
(“craving”). These three stages r­epresent dysregulation in three
functional domains (incentive salience/pathological habits, nega-
tive emotional states, and executive function, respectively) and are
mediated by three major neurocircuitry elements (basal ganglia,
extended amygdala, and prefrontal cortex, respectively;
Figure  53.2). The three stages are conceptualized as interacting
with each other, becoming more intense, and ultimately leading to
the pathological state known as addiction [71].
In AUD, a pattern of oral drug taking evolves that is often
characterized by binges of alcohol intake that can be daily

Figure 53.2  Conceptual framework for the neurobiological basis of addiction. In the binge/intoxication stage, reinforcing effects of drugs may
engage reward neurotransmitters and associative mechanisms in the nucleus accumbens shell and core and then engage stimulus‐response habits
that depend on the dorsal striatum. In the withdrawal/negative affect stage, the negative emotional state of withdrawal may engage activation of the
extended amygdala. The extended amygdala is composed of several basal forebrain structures, including the bed nucleus of the stria terminalis,
central nucleus of the amygdala, and possibly a transition zone in the medial portion (or shell) of the nucleus accumbens. There are major projec-
tions from the extended amygdala to the hypothalamus and brainstem. The preoccupation/anticipation (craving) stage involves the processing of
conditioned reinforcement in the basolateral amygdala and the processing of contextual information by the hippocampus. Executive control depends
on the prefrontal cortex and includes the representation of contingencies, the representation of outcomes, and their value and subjective states (i.e.
craving and, presumably, feelings) associated with drugs. The subjective effects, termed “drug craving” in humans, involve activation of the orbito-
frontal and anterior cingulate cortices and temporal lobe, including the amygdala. ACC, anterior cingulate cortex; BNST, bed nucleus of the stria
terminalis; CeA, central nucleus of the amygdala; DS, dorsal striatum; dlPFC, dorsolateral prefrontal cortex; GP, globus pallidus; HPC, hippocam-
pus; NAC, nucleus accumbens; OFC, orbitofrontal cortex; Thal, thalamus; vlPFC, ventrolateral prefrontal cortex; vmPFC, ventromedial prefrontal
cortex. Modified from [125] and reproduced with permission of Springer Nature.
692 THE LIVER:  NEUROBIOLOGY OF ALCOHOL USE DISORDER (AUD)

episodes or prolonged days of heavy drinking and is character-
ized by a severe emotional and somatic withdrawal syndrome.
Many individuals with AUD continue with such a binge/with-
drawal pattern for extended periods of time, but some individu-
als can evolve into an opioid addiction‐like situation in which
they must have alcohol available at all times to avoid the nega-
tive consequences of abstinence. Here, intense preoccupation
with obtaining alcohol (craving) develops that is linked not only
to stimuli that are associated with obtaining the drug but also to
stimuli that are associated with the aversive motivational state of
withdrawal. A pattern ultimately develops in moderate to severe
AUD where the drug often is taken to avoid the severe dysphoria
and discomfort of abstinence.
The thesis argued here, derived largely from fundamental work
in neurobiology, is that AUD is a brain neurocircuitry disorder
and that neuroadaptations within specific motivational circuits
play an important role in defining and perpetuating the disorder.
The argument herein is that the excessive engagement of reward
circuitry engages high incentive salience for cues and contexts that
are conditioned to drug seeking (binge/intoxication stage) and sets
up a core deficit of lower reward function and greater activation of
brain stress systems (withdrawal/negative affect stage) and sig-
nificant impairment in executive function, all of which contribute
to the compulsive drinking that is associated with AUD.
Drug addiction has generally been conceptualized as a ­disorder
that involves elements of both impulsivity and ­ compulsivity.
Collapsing the cycles of impulsivity and compulsivity yields a
composite addiction cycle that consists of three stages – preoccu­
pation/anticipation, binge/intoxication, and withdrawal/negative
affect – in which impulsivity often dominates at the early stages and
compulsivity dominates at the terminal stages (Figure 53.2). As an
individual moves from impulsivity to compulsivity, a shift occurs
from positive reinforcement that drives the motivated behavior to
negative reinforcement that drives the motivated behavior [72].
Negative reinforcement can be defined as the process by which
the removal of an aversive stimulus (e.g. negative emotional state
of  drug withdrawal) increases the probability of a response.
Importantly, negative reinforcement is not punishment, although
both involve an aversive stimulus. In punishment, the aversive stim-
ulus suppresses behavior, including drug taking (e.g. disulfiram
[Antabuse]). Negative reinforcement can perhaps be described in
lay terms as reward via relief (i.e. relief reward), such as the removal
of pain or in the case of AUD removal of the negative emotional
state of acute withdrawal or protracted abstinence. Driving negative
reinforcement is a negative emotional state that is a common
­
­presentation in most individuals with AUD during withdrawal and
protracted abstinence. The neurobiological substrates that ­underlie
the motivation to seek alcohol will be reviewed herein using the
heuristic framework that is outlined above.
Neural substrates for the binge/intoxication
stage associated with AUD
Alcohol at intoxicating doses has a wide but selective action on
neurotransmitter systems in the brain reward systems, based on
animal models of the acute reinforcing effects of alcohol, and
animal studies that use selective receptor antagonists for spe-
cific neurochemical systems, and human positron emission
tomography imaging studies. Multiple neurochemical systems
are implicated in the acute reinforcing effects of alcohol, includ-
ing γ‐aminobutyric acid (GABA), opioid peptides, dopamine,
serotonin, and glutamate (Figure 53.3).

Figure 53.3  Schematic diagram describing the etiology of addiction based on an early drawing by Dr Loren Parsons. Notice that as positive
reinforcement fades as a motivating factor, negative reinforcement is conceptualized to gain importance as driving compulsive drug seeking.
Relapse is hypothesized to return one to the addictive process at any point in the cycle, and repeated binges–withdrawals–relapses often speed the
trajectory back to compulsive use, presumably through residual changes in neurocircuitry.
 53:  Alcoholic Liver Disease 693
There has been significant work that shows, at least at the
pharmacological level, a role for GABA in the intoxicating
effects of alcohol [73]. Systemic injections of GABAA receptor
antagonists or inverse agonists reverse the motor‐impairing
effects of alcohol, the anxiolytic‐like effects of alcohol, and
alcohol drinking [73]. Endogenous opioid peptide systems have
long been hypothesized to play a role in the reinforcing effects
of alcohol. Naltrexone decreases alcohol drinking and self‐
administration in a variety of animal models [73], and these
results led to the clinical use of naltrexone in reducing alcohol
consumption and preventing relapse.
Significant evidence also supports a role for the mesolimbic
dopamine system in alcohol reinforcement. Alcohol self‐admin-
istration increases extracellular levels of dopamine in the
nucleus accumbens in nondependent rats [74]. Such increases
occur not only during the actual self‐administration session but
also precede the self‐administration session, possibly reflecting
the incentive motivational properties of cues that are associated
with alcohol [74]. Incentive motivation (i.e. incentive‐salience)
is anchored within the construct of conditioned reinforcement
and is argued to be a phenomenon by which a previously neutral
stimulus acquires incentive value through pairings with a drug
of abuse [75]. Systemic injections of dopamine receptor antago-
nists also decrease responses to alcohol. However, mesolimbic
dopamine does not appear to be essential for the acute reinforc-
ing effects of alcohol, since lesions of the mesolimbic dopamine
system failed to block operant self‐administration of alcohol
[73], suggesting multiple redundant sources of alcohol’s actions
that converge on the nucleus accumbens and amygdala.
A prominent hypothesis is that when drug addiction progresses
from occasional recreational use to compulsive use, drug‐seeking
behavior shifts from reward‐driven to habit‐driven behavior to
compulsive‐like responding. Three cortico‐basal ganglia circuits
form what have been described as cortico‐striatal‐pallidal‐­thalamic
loops that process associative, sensorimotor, and emotional infor-
mation [76]. During this progression to habit‐driven behavior, the
control over drug‐seeking behavior also appears to shift from
the nucleus accumbens to the dorsal striatum. Furthermore, within
the dorsal striatum, there is a shift in function [77].
Neural substrates for the withdrawal/negative
affect stage associated with AUD
The neurocircuitry and neuropharmacology of the withdrawal/
negative affect stage of the addiction cycle support a conceptual
framework that builds on opponent process theory [78] and
extends to an allostatic model of brain motivational systems that
has been proposed to explain persistent changes in motivation
that are associated with dependence in addiction [79, 80]. In this
formulation, addiction is conceptualized as a cycle of increasing
dysregulation of brain reward/anti‐reward mechanisms that
results in a negative emotional state that contributes to the
­compulsive use of drugs. Counteradaptive processes that are
part of the normal homeostatic limitation of reward function
fail to return within the normal homeostatic range. These coun-
teradaptive processes are hypothesized to be mediated by
two mechanisms: within‐system neuroadaptations and between‐­
system neuroadaptations [80].
Within‐system neuroadaptations that
contribute to the compulsivity associated
with the dark side of AUD
One prominent hypothesis is that dopamine and opioid peptide
reward/incentive motivational systems are compromised in cru-
cial phases of the addiction cycle, such as withdrawal and pro-
tracted abstinence. The argument is that a decrease in dopamine
and opioid peptide function is hypothesized to lead to lower
motivation for non‐drug‐related stimuli and greater sensitivity
to cues that are associated with the abused drug (i.e. increase in
incentive salience) [81]. Supporting this hypothesis, decreases
in activity of the mesolimbic dopamine system and decreases in
serotonergic neurotransmission in the nucleus accumbens have
been observed during alcohol withdrawal in animal studies [73].
Parallel to these studies is strong evidence that the firing of ven-
tral tegmental area dopamine neurons dramatically decreases
during acute withdrawal from alcohol [82].
Imaging studies of humans with addiction have consistently
shown long‐lasting decreases in the numbers of dopamine D2
receptors in alcohol‐dependent subjects compared with controls
[83]. Additionally, alcohol‐dependent subjects had dramatically
lower dopamine release in the striatum in response to a pharma-
cological challenge with the stimulant drug methylphenidate
[83]. These findings suggest an overall reduction of the sensitiv-
ity of the dopamine component of reward circuitry to natural
reinforcers and other drugs in individuals with addiction.
Other within‐system neuroadaptations under this conceptual
framework could include greater sensitivity of receptor trans-
duction mechanisms in the nucleus accumbens. Drugs of abuse
have acute receptor actions that are linked to intracellular sign-
aling pathways that may undergo adaptations with chronic treat-
ment. In the context of chronic alcohol administration, multiple
molecular mechanisms have been hypothesized to counteract
the acute effects of alcohol that could be considered within‐­
system neuroadaptations. For example, chronic alcohol
decreases GABA receptor function, possibly through downreg-
ulation of the α1 subunit, and also decreases the acute inhibition
of adenosine reuptake [73]. Whereas acute alcohol activates
adenylate cyclase, withdrawal from chronic alcohol decreases
cyclic adenosine monophosphate response element binding pro-
tein phosphorylation in the amygdala and is linked to decreases
in the function of neuropeptide Y (NPY) and increases in anxi-
ety‐like responses during acute alcohol withdrawal [84].
Between‐system neuroadaptations that
contribute to the compulsivity associated
with the dark side of AUD
Another major neuroadaptation that can contribute to the nega-
tive emotional state that drives negative reinforcement in the
withdrawal/negative affect stage is brain neurocircuits and
­neurochemical systems that are involved in arousal–stress mod-
ulation that are engaged within the neurocircuitry of the brain
stress systems in an attempt to overcome the chronic presence of
the perturbing drug (alcohol) and to restore normal function
despite the presence of drug. The neuroanatomical entity that is
termed the extended amygdala [85] may represent a common
694 THE LIVER:  NEUROBIOLOGY OF ALCOHOL USE DISORDER (AUD)
anatomical substrate that integrates brain arousal–stress systems
with hedonic processing systems to produce the between‐­
system opponent process that is elaborated above. The extended
amygdala forms a separate entity within the basal forebrain [86]
and has been conceptualized to be composed of several basal
forebrain structures, including the bed nucleus of the stria termi-
nalis, central nucleus of the amygdala, sublenticular substantia
innominata, and a transition zone in the medial part of the
nucleus accumbens (e.g. shell) [85]. The extended amygdala
receives numerous afferents from limbic structures, such as the
basolateral amygdala and hippocampus, and sends efferents to
the medial part of the ventral pallidum and a large projection to
the lateral hypothalamus, thus further defining the specific brain
areas that interface classical limbic (emotional) structures with
the extrapyramidal motor system. The extended amygdala
has long been hypothesized to play a key role not only in fear
conditioning [87] but also in the emotional component of pain
processing [88].
The brain stress system that is mediated by corticotropin‐
releasing factor (CRF) systems in both the extended amygdala
and hypothalamic–pituitary–adrenal axis (HPA) are dysregu-
lated by the chronic administration of all major drugs with
dependence or abuse potential, with a common response of
­elevated adrenocorticotropic hormone, corticosterone, and
amygdala CRF during acute withdrawal from chronic drug
administration [80] (Figure 53.3). In animal models of alcohol
dependence, extrahypothalamic CRF systems become hyperac-
tive during alcohol withdrawal, with an increase in extracellular
CRF in the central nucleus of the amygdala and bed nucleus of
the stria terminalis in dependent rats [80]. For example, alcohol
withdrawal reliably produces anxiety‐like responses in animal
models that can be reversed by CRF receptor antagonists
[80].  Indeed, using models of repeated alcohol exposure and
­withdrawal, intermittent alcohol exposure facilitates withdrawal‐
associated anxiety‐like responses, and a CRF1 receptor antago-
nist prevented the sensitization of withdrawal‐induced anxiety
[89]. These results are consistent with a prolonged history of
alcohol exposure that produces persistent upregulation of both
CRF and CRF1 receptors in the brain [89].
Perhaps even more compelling, a peptide CRF1/CRF2 antago-
nist, when administered directly in the central nucleus of the
amygdala, blocked alcohol self‐administration in alcohol‐
dependent rats. Cellular studies have identified the actions of
CRF on GABAergic interneurons in the central nucleus of the
amygdala [90]. Systemic injections of small‐molecule CRF1
receptor antagonists also blocked the increase in alcohol intake
that was associated with acute withdrawal and protracted absti-
nence. A CRF receptor antagonist that was administered chroni-
cally during the development of dependence blocked the
development of compulsive‐like responding for alcohol (for
review, see reference [89]).
Alcohol addiction has long been associated with dysregula-
tion of the HPA axis [91]. Clinical studies have reported impair-
ments in HPA axis responsivity in AUD [91], even to the point
of pseudo‐Cushing’s syndrome, manifested by high levels of
cortisol, in individuals with alcohol addiction [92]. However, a
more common observation is a blunted cortisol response in indi-
viduals with alcohol dependence, again possibly reflecting
adaptations to an initial hyperactive cortisol response [91].
Similar effects have been observed in animal models, with a
blunted corticosterone response in rats that are made dependent
using the chronic intermittent alcohol vapor model [93]. One
hypothesis is that activation of the HPA axis can drive neuroad-
aptive changes in extrahypothalamic CRF systems in the
extended amygdala. High corticosterone increases CRF mRNA
in the central nucleus of the amygdala and lateral bed nucleus of
the stria terminalis and decreases CRF mRNA in the paraven-
tricular nucleus of the hypothalamus. Thus, an initial exposure
to high corticosterone, stimulated by moderate to heavy drink-
ing, may stimulate CRF expression in the central nucleus of the
amygdala and lateral bed nucleus of the stria terminalis, eventu-
ally leading to neuroadaptive changes, including the further sen-
sitization of CRF activation in the extended amygdala and lower
HPA function [93]. Consistent with this hypothesis, chronic
­glucocorticoid receptor blockade with mifepristone during the
course of alcohol vapor exposure prevented the escalation of
alcohol intake and blocked the increase in progressive‐ratio
responding for alcohol in dependent animals [94]. Chronic glu-
cocorticoid receptor antagonism also blocked escalated and
compulsive alcohol drinking during protracted abstinence in
rats with a history of alcohol dependence. These results suggest
a critical role for glucocorticoid receptors in the development
and maintenance of alcohol dependence.
Other brain neurotransmitter or neuromodulatory systems
that have pro‐stress actions also converge on the extended
amygdala, all of which may contribute to negative emotional
states that are associated with drug withdrawal or protracted
abstinence [80]. Chronic administration of psychostimulants
and opioids has long been hypothesized to activate cyclic aden-
osine monophosphate response element binding protein, which
in turn activates dynorphin in medium spiny neurons that in turn
feedback and decrease the activity of ventral tegmental dopa-
mine neurons [95]. Although κ‐opioid receptor agonists sup-
press nondependent drinking, presumably via aversive stimulus
effects, κ‐opioid receptor antagonists block the excessive drink-
ing associated with alcohol withdrawal and dependence, and
this effect may be mediated by the shell of the nucleus accum-
bens [73].  Other neurotransmitter/neuromodulatory systems
that comprise the brain stress system in the extended amyg-
dala include ­vasopressin, hypocretin (orexin), substance P, and
­neuroimmune factors. In addition to CRF and dynorphin, there
is evidence that norepinephrine, vasopressin, substance P, and
hypocretin (orexin) may all contribute to negative emotional
states of drug withdrawal, particularly alcohol withdrawal
(Figure  53.3). Thus, activation of this pro‐stress, pro‐negative
emotional state system is multi‐determined and comprises the
neurochemical bases for hedonic opponent processes [96].
Neurotransmitter systems that are implicated in anti‐stress
actions include NPY, nociceptin, and endocannabinoids. NPY
has powerful orexigenic and anxiolytic effects and has been
hypothesized to act in opposition to the actions of CRF in
addiction [84]. Nociceptin and synthetic ­nociception receptor
agonists have effects on GABA synaptic activity in the central
nucleus of the amygdala that are similar to NPY and can block
high alcohol consumption in a genetically selected line of rats
that is known to be hypersensitive to s­ tressors [97]. Evidence
also implicates endocannabinoids in the regulation of affective
states, in which reductions of cannabinoid CB1 receptor
 53:  Alcoholic Liver Disease 695

signaling produce anxiogenic‐like behavioral effects [98]. hypothesized to reflect the neural representations of reward and
Thus, vulnerability to AUD may involve not only a sensitized incentive salience constructs [104]. One parsimonious view of
stress system but also a hypoactive stress buffer system, and the human imaging data that is consistent with animal model
behavioral and pharmacological interventions that block pro‐ data is that there is a “go” system in the dorsal prefrontal/­
stress and stimulate anti‐stress (described in Neural substrates cingulate cortex that drives impulsivity and craving and a “stop”
for the preoccupation/anticipation stage associated with AUD) system in ventromedial prefrontal cortex that inhibits impulsiv-
systems may be particularly interesting targets for the treatment ity and craving [105]. In humans, stress and stressors have long
of AUD. been associated with relapse and the vulnerability to relapse
In summary, two processes are hypothesized to form the neu- [106]. Individuals with addiction are hypersensitive to pain dur-
robiological basis for the withdrawal/negative affect stage: loss ing withdrawal, particularly in the face of negative affect [107].
of function in the reward systems (within‐system neuroadapta- Indeed, the leading precipitant of relapse is negative emotion/
tion) and recruitment of the brain stress systems (between‐­ affect including elements of anger, frustration, sadness, anxiety,
system neuroadaptation) [71]. As dependence and withdrawal and guilt [108].
develop, brain stress systems, such as CRF, norepinephrine, and Studies of “craving” in animal models can be divided into
dynorphin, are recruited, producing aversive or stress‐like states three domains: alcohol seeking that is induced by the drug itself,
[99]. The combination of decreases in reward neurotransmitter alcohol priming‐induced reinstatement, alcohol seeking that is
function and recruitment of brain stress systems provides a induced by stimuli that are paired with drug taking, cue‐ and
powerful motivation for reengaging in drug taking and drug context‐induced reinstatement, and alcohol seeking that is
seeking. induced by an acute stressor or a state of stress (i.e. stress‐
induced reinstatement) [109]. Most evidence from animal stud-
Neural substrates for the preoccupation/ ies suggests that drug‐induced reinstatement is localized to the
medial prefrontal cortex/ventral striatum circuit and mediated
anticipation stage associated with AUD
by the neurotransmitter glutamate [110]. Neurotransmitter sys-
The preoccupation/anticipation or “craving” stage of the addic- tems that are involved in drug‐induced reinstatement involve a
tion cycle has long been hypothesized to be a key part of the glutamatergic projection from the frontal cortex to the nucleus
neurocircuitry that mediates relapse in humans. Dysregulation accumbens that is modulated by dopamine activity in the frontal
of the frontal cortex mediates not only elements of impulsivity cortex (Figure  53.3). In contrast, neuropharmacological and
and compulsivity but also protracted abstinence and craving. neurobiological studies that use animal models of cue‐induced
Human imaging studies reveal neurocircuitry dysregulation reinstatement involve the basolateral amygdala as a critical sub-
during the preoccupation/anticipation stage in AUD that strate, with a possible feed‐forward mechanism possibly through
includes not only compromises in frontal cortical executive the same prefrontal cortex system that is involved in drug‐
function but also dysregulated substrates that mediate craving. induced reinstatement [109, 111]. Cue‐induced reinstatement
Lower frontal cortex activity parallels deficits in executive func- involves dopamine modulation in the basolateral amygdala and
tion in neuropsychologically challenging tasks in individuals a glutamatergic projection to the nucleus accumbens from both
with AUD with and without Wernicke–Korsakoff’s syndrome the basolateral amygdala and ventral subiculum [111]. Such
(for review, see references [100, 101]). In individuals with AUD, reinstatement in alcohol drinking can be blocked by the sys-
there are impairments in the maintenance of spatial information, temic administration of naltrexone and a selective μ and δ
the disruption of decision making, and impairments in behavio- ­opioid receptor antagonist [109]. These results are consistent
ral inhibition. Such frontal cortex‐derived executive function with human studies that showed that opioid receptor antagonists
disorders in AUD have been linked to deficits in the ability of may blunt the urge to drink that is elicited by the presentation of
behavioral treatments to effect recovery from AUD post‐­ alcohol‐related cues in individuals with AUD [112]. In contrast,
detoxification [100]. Thus, deficits in the prefrontal cortical the stress‐induced reinstatement of drug‐related responding in
control of incentive salience may represent a key mechanism to animal models appears to depend on the activation of both CRF
explain individual differences in the vulnerability to AUD, and and norepinephrine in elements of the extended amygdala
the excessive attribution of incentive salience to drug‐related (­central nucleus of the amygdala and bed nucleus of the stria
cues and residual hypersensitivity of the brain stress systems terminalis) [113].
may perpetuate excessive drug intake, compulsive behavior,
and relapse. Compulsivity in AUD: a negative
Craving responses to cues in human imaging studies activate
an overarching cognitive control network in the brain, involving
reinforcement perspective
dorsolateral prefrontal, anterior cingulate, and parietal cortices, Compulsivity in AUD can derive from neurocircuitry changes at
all of which support a broad range of executive functions [102]. all three stages of the addiction cycle (Figure 53.3). In the binge/
Such studies have led to the hypothesis that a frontal–cingulate– intoxication stage, such changes may involve neurocircuits that
parietal–subcortical cognitive control network is consistently include enhanced incentive salience [75] and the engagement of
recruited across a range of traditional executive function tasks, pathological habit function [114]. In the withdrawal/negative
many of which show deficits in humans with AUDs. Indeed, affect stage, such changes involve the recruitment of neurocir-
the most prominent activation by alcohol‐related cues involves cuits that are involved in negative emotional states. In the preoc­
the dorsolateral prefrontal cortex, cingulate cortex, and orbito- cupation/anticipation stage, such changes involve impairments
frontal cortex [103]. Such drug cue‐evoked responses are in executive function and the dysregulation of inhibitory control
696 THE LIVER:  TREATMENT OF AUD PATIENTS WITH ALD
that regulates impulsivity [115]. Many of the cue‐ and stress‐
related neuroadaptations persist into protracted abstinence,
largely described as dysregulation that persists beyond acute
withdrawal [73].
However, negative emotional states of acute and protracted
abstinence that are induced by chronic high‐dose alcohol have
been largely neglected in the clinical domain. Negative emo-
tional states may strongly impact compulsivity not only by
directly driving negative reinforcement but also by enhancing
the value of incentive salience, enhancing the value of habit
expression, or exacerbating impairments in executive function.
Thus, the overall conceptual theme that is argued herein is that
moderate to severe AUD represents a break with homeostatic
brain regulatory mechanisms that regulate the emotional state of
the individual. One hypothesis is that the dysregulation of
­emotion begins with the binge and subsequent acute withdrawal
but leaves a residual neuroadaptive trace that allows rapid
“re‐addiction” even months and years after detoxification and
abstinence [79]. Thus, the emotional dysregulation of alcohol
addiction represents more than the simple homeostatic dysregu-
lation of hedonic function; it also represents a dynamic break
with the homeostasis of this system that has been termed
allostasis [79]. A further argument is that this hypernegative
emotional state, termed hyperkatifeia, sensitizes over time,
extends into protracted abstinence, and provides the driving
force for another source of motivation for prolonging and main-
taining addiction, that of negative reinforcement (Figure 53.3).
TREATMENT OF AUD PATIENTS
WITH ALD
While reductions below harmful drinking levels is beneficial,
the cornerstone in the treatment of patients with ALD remains
total alcohol abstinence, as abstinence can improve outcome at
all stages of liver disease [116]. However, there are only a few
­randomized studies that have investigated behavioral and/or
pharmacological treatments in patients with AUD and ALD.
Behavioral treatments
Brief interventions are aimed at educating the patient about
harmful drinking, increasing motivation to change behavior, and
reinforcing skills to address problematic drinking. In primary
care settings, studies support the use of brief interventions to
reduce drinking [117]. Furthermore, these interventions may
synergize with other areas of treatment, such as, compliance to
medications and referral to treatment programs, an approach
referred to as screening, brief intervention and referral to treat­
ment (SBIRT). Specific psychosocial and behavioral treatments
for AUD include 12‐step facilitation, cognitive behavioral ther-
apy (CBT), and motivational enhancement therapy (MET)
[116]. Twelve‐step facilitation focuses on abstinence and
involves Alcoholics Anonymous meetings. The goal of CBT is
to develop coping mechanisms where alcohol is replaced by
alcohol‐free circumstances. MET helps patients work through
the dilemma by “rolling with the resistance” to change [116].
Specific studies investigating behavioral treatments for AUD in
patients with ALD are very limited. Lieber et  al. [118] con-
ducted a clinical pharmacology study of liver fibrosis and
showed that the addition of a brief intervention resulted in a
significant reduction in alcohol consumption. Other studies
have investigated behavioral platforms, like CBT and MET, as
ways to facilitate alcohol abstinence in patients with ALD. A
systematic review of these studies was conducted by Khan et al.
[119]. A total of 13 papers were selected for a whole sample of
1945 patients. This systematic review indicated that combining
medical care and behavioral approaches, like CBT and MET,
facilitated alcohol abstinence, a conclusion that supports the
importance of multidisciplinary approaches to treat patients
with AUD and ALD [119].
Medications
Pharmacological approaches to treat AUD patients include med-
ications to treat alcohol withdrawal syndrome (AWS) and those
used to help patients reduce their craving, cut their drinking,
facilitate abstinence, and prevent relapse. As for the treatment of
AWS, it is important to keep in mind that while up to 50% of
individuals with AUD present with alcohol withdrawal symp-
toms after they stop drinking, only a small percentage requires
pharmacological treatment. For these patients, benzodiazepines
are the gold standard because they represent the only class of
medications that reduces the risk of withdrawal seizures and/or
delirium tremens [120]. In those individuals with AUD and
ALD who develop AWS, a symptom‐triggered schedule with
lorazepam or oxazepam is preferred. In fact, these benzodiaz-
epines do not undergo phase I biotransformation; rather, they
only undergo glucuronidation, which is preserved even if liver
function is compromised [116].
A better understanding of the neurobiology of addiction,
reviewed above in this chapter, has played an important role in
the development of medications for AUD, as summarized in
Table 53.1. In the United States, acamprosate, disulfiram, and
naltrexone (oral and intramuscular) are approved by the Food
and Drug Administration (FDA) for treatment of AUD. A recent
meta‐analysis supports the efficacy of naltrexone and acampro-
sate, but not disulfiram, for AUD [121]. Furthermore, nalmefene
was recently approved in Europe for the treatment of AUD, but
it is not approved in the United States. Not only is the number of
approved medications for AUD very limited, but also their use
in patients with ALD is even more narrow. In fact, disulfiram
may cause hepatotoxicity and is not recommended in patients
with ALD. Naltrexone’s potential for causing hepatotoxicity
also exists, even if it seems to be rare. Acamprosate has not for-
mally been tested in AUD patients with liver disease, but it does
not undergo hepatic metabolism and there are no reports of
hepatotoxicity [116].
In the past decades, preclinical and clinical studies have
­provided support for some medications as potential novel treat-
ments for AUD. While none of these medications are FDA‐
approved, some of them have shown efficacy for AUD in phase
2/3 trials. Among them, the most promising are baclofen,
gabapentin, ondansetron, topiramate, and varenicline, as sum-
marized in Table  53.1. However, formal clinical trials testing
these medications in AUD patients with ALD are lacking, except
for baclofen. In fact, while general clinical trials testing baclofen
 53:  Alcoholic Liver Disease 697

Table 53.1  FDA‐approved medications and others tested in alcohol use disorder patientsa
Dosage Pharmacological target Possible use in alcohol use disorder patients with alcoholic liver disease?

FDA‐approved medications for alcohol use disorder

Acamprosate 666 mg TID Possibly NMDA receptor agonist Yes (no hepatic metabolism)
Disulfiram 250–500 mg QD Inhibition of acetaldehyde No (hepatic metabolism; cases of liver toxicity have been reported)
dehydrogenase
Naltrexoneb PO: 50 mg QD μ opioid receptor antagonist With caution (perceptions of liver toxicity limit use in advanced
PO or IM IM: 380 mg monthly alcoholic liver disease)
Not FDA‐approved medications tested for alcohol use disorder
Baclofen 10 mg TID; GABAB receptor agonist Yes (minimal hepatic metabolism)
80 mg QD max Baclofen has been formally tested in clinical studies with alcohol use
disorder patients with liver cirrhosis
Gabapentin 900–1800 mg QD Unclear – the most likely Yes (no hepatic metabolism)
mechanism is blockade of
voltage‐dependent Ca2+ channels.
Ondansetron 1–16 mcg /kg−1 BID 5HT3 antagonist Yes, but with caution because liver toxicity has been reported, albeit
relationship to ondansetron administration is not determined
Topiramate 300 mg QD Anticonvulsant multiple targets: Yes (partial hepatic metabolism mostly by glucuronidation)
−Glutamate/+GABA In patients with hepatic encephalopathy, use with caution: topiramate‐
related cognitive side‐effects may confound the clinical course and
treatment of hepatic encephalopathy
Varenicline 2 mg QD Nicotinic acetylcholine receptor Yes (minimal hepatic metabolism)
partial agonist
a
 Reproduced from [116] with permission of Elsevier.
b
 Nalmefene is not‐FDA approved but was recently approved in Europe for alcohol use disorder. Compared to naltrexone, nalmefene has a longer half‐life and no evidence of
hepatotoxicity.
FDA, Food and Drug Administration; TID, three times a day; NMDA, N‐methyl‐D‐aspartate; QD, once a day; PO, per os (oral); IM, intramuscular; GABA, gamma‐­
aminobutyric acid; BID, twice a day; HT, serotonin.

have generated conflicting results [122], two independent clini-
cal trials testing baclofen in AUD patients with ALD both
­support its efficacy in treating this specific subpopulation [123,
124]. However, the mechanisms by which baclofen may be effi-
cacious, specifically in AUD patients with ALD, are unknown.
Furthermore, although not formally tested in patients with ALD,
the other medications mentioned above (especially gabapentin
and varenicline), might also be useful in AUD individuals with
ALD because they have no evidence of liver toxicity (for addi-
tional details, please see Table 53.1).
In conclusion, the crucial goal in treating patients with AUD
and ALD is to help them to achieve and maintain alcohol absti-
nence. As briefly summarized above, both behavioral and phar-
macological treatments are available and may be effective in
helping AUD patients to quit drinking. These treatments remain
underutilized, and it is therefore very important that their use is
expanded in clinical care, including in hepatology settings
where many patients are referred for their ALD. Therefore, it is
critical that addiction medicine is integrated into treatment and
prescribed by the multidisciplinary teams that care for patients
with AUD and ALD.
REFERENCES
1. Gao, B. and Bataller, R. Alcoholic liver disease: pathogenesis and new thera-
peutic targets. Gastroenterology, 2011;141:1572–85.
2. O’Shea, R.S., Dasarathy, S., McCullough, A.J. et al. Alcoholic liver disease.
Hepatology, 2010;51:307–28.
3. Zakhari, S. and Li, T.K. Determinants of alcohol use and abuse: impact
of  quantity and frequency patterns on liver disease. Hepatology, 2007;
46;2032–9.
4. Mandrekar, P., Bataller, R., Tsukamoto, H. et al. Alcoholic hepatitis: transla-
tional approaches to develop targeted therapies. Hepatology, 2016;64:
1343–55.
5. Chang, B., Xu, M.J., Zhou, Z. et al. Short‐ or long‐term high‐fat diet feeding
plus acute ethanol binge synergistically induce acute liver injury in mice: an
important role for CXCL1. Hepatology, 2015;62:1070–85.
6. Xu, J., Lai, K.K., Verlinsky, A. et al. Synergistic steatohepatitis by moderate
obesity and alcohol in mice despite increased adiponectin and p‐AMPK.
J Hepatol, 2011;55:673–82.
7. You, M., Considine, R.V., Leone, T.C. et  al. Role of adiponectin in the
­protective action of dietary saturated fat against alcoholic fatty liver in mice.
Hepatology, 2005;42:568–77.
8. Bertola, A., Park, O., and Gao, B. Chronic plus binge ethanol feeding syner-
gistically induces neutrophil infiltration and liver injury in mice: a critical
role for E‐selectin. Hepatology, 2013;58:1814–23.
9. Xu, M.J., Cai, Y., Wang, H. et al. Fat‐specific protein 27/CIDEC promotes
development of alcoholic steatohepatitis in mice and humans.
Gastroenterology, 2015;149:1030–41.
10. Gao, B., Xu, M.J., Bertola, A. et al. Animal models of alcoholic liver disease:
pathogenesis and clinical relevance. Gene Expr, 2017;17:173–86.
11. Cai, Y., Xu, M.J., Koritzinsky, E.H. et  al. Mitochondrial DNA‐enriched
microparticles promote acute‐on‐chronic alcoholic neutrophilia and hepato-
toxicity. JCI Insight, 2017;2.
12. Li, M., He, Y., Zhou, Z. et  al. MicroRNA‐223 ameliorates alcoholic liver
injury by inhibiting the IL‐6‐p47phox‐oxidative stress pathway in neutro-
phils. Gut, 2017;66:705–15.
13. Hatton, J., Burton, A., Nash, H. et  al. Drinking patterns, dependency and
life‐time drinking history in alcohol‐related liver disease. Addiction,
2009;104:587–92.
14. Askgaard, G., Gronbaek, M., Kjaer, M.S. et al. Alcohol drinking pattern and
risk of  alcoholic liver cirrhosis: a prospective cohort study. J Hepatol,
2015;62:1061–7.
15. Trepo, E., Romeo, S., Zucman‐Rossi, J. et al. PNPLA3 gene in liver d­ iseases.
J Hepatol, 2016;65:399–412.
16. Ajmera, V.H., Terrault, N.A., and Harrison, S.A. Is moderate alcohol use in
nonalcoholic fatty liver disease good or bad? A critical review. Hepatology,
2017;65:2090–9.
17. Purohit, V., Gao, B., Song, B.J. Molecular mechanisms of alcoholic fatty
liver. Alcohol Clin Exp Res, 2009;33:191–205.
698 THE LIVER:  REFERENCES
18. Parker, R., Kim, S.J., and Gao, B. Alcohol, adipose tissue and liver disease:
mechanistic links and clinical considerations. Nat Rev Gastroenterol
Hepatol, 2018;15:50–9.
19. Zhong, W., Zhao, Y., Tang, Y. et  al. Chronic alcohol exposure stimulates
­adipose tissue lipolysis in mice: role of reverse triglyceride transport in the
pathogenesis of alcoholic steatosis. Am J Pathol, 2012;180:998–1007.
20. Sebastian, B.M., Roychowdhury, S., Tang, H. et  al. Identification of a
cytochrome P4502E1/Bid/C1q‐dependent axis mediating inflammation in
adipose tissue after chronic ethanol feeding to mice. J Biol Chem,
2011;286:35989–97.
21. Ji, C., Chan, C., and Kaplowitz, N. Predominant role of sterol response ele-
ment binding proteins (SREBP) lipogenic pathways in hepatic steatosis in
the murine intragastric ethanol feeding model. J Hepatol, 2006;45:717–24.
22. You, M., Fischer, M., Deeg, M.A. et al. Ethanol induces fatty acid synthesis
pathways by activation of sterol regulatory element‐binding protein
(SREBP). J Biol Chem, 2002;277:29342–7.
23. Baraona, E. and Lieber, C.S. Effects of ethanol on lipid metabolism. J Lipid
Res, 1979;20:289–315.
24. Galli, A., Pinaire, J., Fischer, M. et al. The transcriptional and DNA binding
activity of peroxisome proliferator‐activated receptor alpha is inhibited by
ethanol metabolism. A novel mechanism for the development of ethanol‐
induced fatty liver. J Biol Chem, 2001;276:68–75.
25. O’Neill, H.M., Holloway, G.P., and Steinberg, G.R. AMPK regulation of
fatty acid metabolism and mitochondrial biogenesis: implications for
­obesity. Mol Cell Endocrinol, 2013;366:135–51.
26. You, M., Matsumoto, M., Pacold, C.M. et  al. The role of AMP‐activated
protein kinase in the action of ethanol in the liver. Gastroenterology,
2004;127:1798–1808.
27. Dolganiuc, A., Thomes, P.G., Ding, W.X. et  al. Autophagy in alcohol‐
induced liver diseases. Alcohol Clin Exp Res, 2012;36:1301–8.
28. Ding, W.X., Li, M., Chen, X. et al. Autophagy reduces acute ethanol‐induced
hepatotoxicity and steatosis in mice. Gastroenterology, 2010;139:1740–52.
29. Nagy, L.E., Ding, W.X., Cresci, G. et al. Linking pathogenic mechanisms
of  alcoholic liver disease with clinical phenotypes. Gastroenterology,
2016;150:1756–68.
30. Gao, B. and Tsukamoto, H. Inflammation in alcoholic and nonalcoholic fatty
liver disease: friend or foe? Gastroenterology, 2016;150:1704–9.
31. Hartmann, P., Seebauer, C.T., and Schnabl, B. Alcoholic liver disease: the
gut microbiome and liver cross talk. Alcohol Clin Exp Res, 2015;39:
763–75.
32. Lucey, M.R., Mathurin, P., and Morgan, T.R. Alcoholic hepatitis. N Engl J
Med, 2009;360:2758–69.
33. Crabb, D.W., Bataller, R., Chalasani, N.P. et  al. Standard definitions and
common data elements for clinical trials in patients with alcoholic hepatitis:
recommendation from the NIAAA alcoholic hepatitis consortia.
Gastroenterology, 2016;150:785–90.
34. Affo, S., Dominguez, M., Lozano, J.J. et al. Transcriptome analysis identi-
fies TNF superfamily receptors as potential therapeutic targets in alcoholic
hepatitis. Gut, 2013;62:452–60.
35. Michelena, J., Altamirano, J., Abraldes, J.G. et al. Systemic inflammatory
response and serum lipopolysaccharide levels predict multiple organ failure
and death in alcoholic hepatitis. Hepatology, 2015;62:762–72.
36. Louvet, A., Wartel, F., Castel, H. et al. Infection in patients with severe alco-
holic hepatitis treated with steroids: early response to therapy is the key
­factor. Gastroenterology, 2009;137:541–8.
37. Bataller, R. and Gao, B. Liver fibrosis in alcoholic liver disease. Semin Liver
Dis, 2015;35:146–56.
38. Jeong, W.I., Park, O., and Gao, B. Abrogation of the antifibrotic effects of
natural killer cells/interferon‐gamma contributes to alcohol acceleration
of liver fibrosis. Gastroenterology, 2008;134:248–58.
39. Yang, A.M., Inamine, T., Hochrath, K. et  al. Intestinal fungi contribute to
development of alcoholic liver disease. J Clin Invest, 2017;127:2829–41.
40. Llopis, M., Cassard, A.M., Wrzosek, L. et  al. Intestinal microbiota
contributes to individual susceptibility to alcoholic liver disease. Gut,
­ of alcoholic liver disease. Pharmacol Ther, 2017;180:77–89.
2016;65:830–9.
41. Esfandiari, F., Medici, V., Wong, D.H. et al. Epigenetic regulation of hepatic
endoplasmic reticulum stress pathways in the ethanol‐fed cystathionine beta
synthase‐deficient mouse. Hepatology, 2010;51:932–41.
42. Satishchandran, A., Ambade, A., Rao, S. et al. MicroRNA 122, Regulated by
GRLH2, protects livers of mice and patients from ethanol‐induced liver
­disease. Gastroenterology, 2018;154:238–52.
43. Szabo, G. and Momen‐Heravi, F. Extracellular vesicles in liver disease and
potential as biomarkers and therapeutic targets. Nat Rev Gastroenterol
Hepatol, 2017;14:455–66.
44. Saha, B., Momen‐Heravi, F., Furi, I. et al. Extracellular vesicles from mice
with alcoholic liver disease carry a distinct protein cargo and induce
­macrophage activation via heat shock protein 90. Hepatology, 2018;67(5):
1986–2000.
45. Sutti, S., Bruzzi, S., and Albano, E. The role of immune mechanisms in
alcoholic and nonalcoholic steatohepatitis: a 2015 update. Expert Rev
­
Gastroenterol Hepatol, 2016;10:243–53.
46. Liaskou, E., Klemsdal Henriksen, E.K., Holm, K. et  al. High‐throughput
T‐cell receptor sequencing across chronic liver diseases reveals distinct
­disease‐associated repertoires. Hepatology, 2016;63:1608–19.
47. Mathews, S., Feng, D., Maricic, I. et al. Invariant natural killer T cells con-
tribute to chronic‐plus‐binge ethanol‐mediated liver injury by promoting
hepatic neutrophil infiltration. Cell Mol Immunol, 2016;13:206–16.
48. Cui, K., Yan, G., Xu, C. et al. Invariant NKT cells promote alcohol‐induced
steatohepatitis through interleukin‐1beta in mice. J Hepatol, 2015;
62:1311–8.
49. Riva, A., Patel, V., Kurioka, A. et al. Mucosa‐associated invariant T cells link
intestinal immunity with antibacterial immune defects in alcoholic liver
­disease. Gut, 2017;67(5):918–30.
50. Bertola, A., Mathews, S., Ki, S.H. et al. Mouse model of chronic and binge
ethanol feeding (the NIAAA model). Nat Protoc, 2013;8:627–37.
51. Chao, X., Wang, S., Zhao, K. et  al. Impaired TFEB‐mediated lysosome
­biogenesis and autophagy promote chronic ethanol‐induced liver injury and
steatosis in mice. Gastroenterology, 2018;155(3):865–79.
52. Wang, W., Xu, M.J., Cai, Y. et al. Inflammation is independent of steatosis in
a murine model of steatohepatitis. Hepatology, 2017;66:108–23.
53. Khanova, E., Wu, R., Wang, W. et al. Pyroptosis by caspase11/4‐gasdermin‐
D pathway in alcoholic hepatitis. Hepatology, 2017;67(5):1737–53.
54. European Association for the Study of Liver. EASL clinical practical guide-
lines: management of alcoholic liver disease. J Hepatol, 2012;57:399–420.
55. Alatalo, P., Koivisto, H., Puukka, K. et  al. Biomarkers of liver status in
heavy  drinkers, moderate drinkers and abstainers. Alcohol Alcohol,
2009;44:199–203.
56. Aday, A.W., Mitchell, M.C., Casey, L.C. Alcoholic hepatitis: current trends
in management. Curr Opin Gastroenterol, 2017;33:142–8.
57. Ku, N.O., Strnad, P., Bantel, H. et al. Keratins: Biomarkers and modulators
of apoptotic and necrotic cell death in the liver. Hepatology, 2016;
64:966–76.
58. Mortele, K.J. and Ros, P.R. Imaging of diffuse liver disease. Semin Liver Dis,
2001;21:195–212.
59. Borra, R.J., Salo, S., Dean, K. et al. Nonalcoholic fatty liver disease: rapid
evaluation of liver fat content with in‐phase and out‐of‐phase MR imaging.
Radiology, 2009;250:130–6.
60. Stickel, F., Datz, C., Hampe, J. et al. Pathophysiology and management of
alcoholic liver disease: update 2016. Gut Liver, 2017;11:173–88.
61. Parker, R., Armstrong, M.J., Corbett, C. et al. Systematic review: pentoxifyl-
line for the treatment of severe alcoholic hepatitis. Aliment Pharmacol Ther,
2013;37:845–54.
62. Nguyen‐Khac, E., Thevenot, T., Piquet, M.A. et  al. Glucocorticoids plus
N‐acetylcysteine in severe alcoholic hepatitis. N Engl J Med, 2011
;365:1781–9.
63. Piechota, M. and Piechota, A. An evaluation of the usefulness of extracorpor-
eal liver support techniques in patients hospitalized in the ICU for severe
liver dysfunction secondary to alcoholic liver disease. Hepat Mon,
2016;16:e34127.
64. Wang, H.J., Gao, B., Zakhari, S. et  al. Inflammation in alcoholic liver
­disease. Annu Rev Nutr, 2012;32:343–68.
65. Lieber, S.R., Rice, J.P., Lucey, M.R. et al. Controversies in clinical trials for
alcoholic hepatitis. J Hepatol, 2018;68(3):586–92.
66. Xu, M.J., Zhou, Z., Parker, R. et al. Targeting inflammation for the treatment
67. Kong, X., Feng, D., Mathews, S. et  al. Hepatoprotective and anti‐fibrotic
functions of interleukin‐22:therapeutic potential for the treatment of
­alcoholic liver disease. J Gastroenterol Hepatol, 2013;28(1):56–60.
68. Gao, B. and Shah, V.H. Combination therapy: new hope for alcoholic hepa-
titis? Clin Res Hepatol Gastroenterol, 2015;39(1):S7–11.
69. Kubiliun, M., Patel, S.J., Hur, C. et al. Early liver transplantation for alco-
holic hepatitis: ready for primetime? J Hepatol, 2018;68(3):380–2.
 53:  Alcoholic Liver Disease 699
70. Mitchell, M.C., Friedman, L.S., and McClain, C.J. Medical management
of  severe alcoholic hepatitis: expert review from the clinical practice
updates  committee of the AGA Institute. Clin Gastroenterol Hepatol,
2017;15:5–12.
71. Koob, G.F. and Le Moal, M. Drug abuse: hedonic homeostatic dysregula-
tion. Science, 1997;278:52–8.
72. Koob, G.F. Allostatic view of motivation: implications for psychopathology,
in Motivational Factors in the Etiology of Drug Abuse, (eds. R.A. Bevins and
M.T. Bardo), (series title: Nebraska Symposium on Motivation, Vol 50).
University of Nebraska Press, Lincoln NE, 2004, pp. 1–18.
73. Koob, G.F. Neurocircuitry of alcohol addiction: synthesis from animal
­models, in Alcohol and the Nervous System (eds. E.V. Sullivan and
A.  Pfefferbaum), (series title: Handbook of Clinical Neurology, vol 125).
Elsevier, Amsterdam, 2014, pp. 33–54.
74. Weiss, F., Lorang, M.T., Bloom, F.E. et al. Oral alcohol self‐administration
stimulates dopamine release in the rat nucleus accumbens: genetic and moti-
vational determinants. J Pharmacol Exp Ther, 1993;267:250–8.
75. Robinson, T.E. and Berridge, K.C. The neural basis of drug craving: an
incentive‐sensitization theory of addiction. Brain Res Brain Res Rev,
1993;18:247–91.
76. Haber, S.N., Fudge, J.L., and McFarland, N.R. Striatonigrostriatal pathways
in primates form an ascending spiral from the shell to the dorsolateral stria-
tum. J Neurosci, 2000;20:2369–82.
77. Lovinger, D.M. and Kash, T.L. Mechanisms of neuroplasticity and ethanol’s
effects on plasticity in the striatum and bed nucleus of the stria terminalis.
Alcohol Res, 2015;37:109–24.
78. Solomon, R.L. and Corbit, J.D. An opponent‐process theory of motivation.
I. Temporal dynamics of affect. Psychol Rev, 1974;81:119–45.
79. Koob, G.F. and Le Moal, M. Drug addiction, dysregulation of reward, and
allostasis. Neuropsychopharmacology, 2001;24:97–129.
80. Koob, G.F. and Le Moal, M. Addiction and the brain antireward system.
Annu Rev Psychol, 2008;59:29–53.
81. Melis, M., Spiga, S., and Diana, M. The dopamine hypothesis of drug addic-
tion: hypodopaminergic state. Int Rev Neurobiol, 2005;63:101–54.
82. Diana, M., Pistis, M., Carboni, S. et al. Profound decrement of mesolimbic
dopaminergic neuronal activity during ethanol withdrawal syndrome in rats:
electrophysiological and biochemical evidence. Proc Natl Acad Sci USA,
1993;90:7966–9.
83. Volkow, N.D., Fowler, J.S., Wang, G.J. et  al. Imaging dopamine’s role in
drug abuse and addiction. Neuropharmacology, 2009;56(1):3–8.
84. Kyzar, E.J. and Pandey, S.C. Molecular mechanisms of synaptic remodeling
in alcoholism. Neurosci Lett, 2015;601:11–9.
85. Heimer, L. and Alheid, G.F. Piecing together the puzzle of basal forebrain
anatomy. Adv Exp Med Biol, 1991;295:1–42.
86. Alheid, G.F. and Heimer, L. New perspectives in basal forebrain organiza-
tion of special relevance for neuropsychiatric disorders: the striatopallidal,
amygdaloid, and corticopetal components of substantia innominata.
Neuroscience, 1988;27:1–39.
87. LeDoux, J.E. Emotion circuits in the brain. Annu Rev Neurosci,
2000;23:155–84.
88. Neugebauer, V., Li, W., Bird, G.C. et al. The amygdala and persistent pain.
Neuroscientist, 2004;10:221–34.
89. Zorrilla, E.P., Logrip, M.L., and Koob, G.F. Corticotropin releasing factor: a
key role in the neurobiology of addiction. Front Neuroendocrinol,
2014;35:234–44.
90. Roberto, M., Cruz, M.T., Gilpin, N.W. et al. Corticotropin releasing factor‐
induced amygdala gamma‐aminobutyric acid release plays a key role in
­alcohol dependence. Biol Psychiatry, 2010;67:831–9.
91. Vendruscolo, L.F. and Koob, G.F. Alcohol dependence conceptualized as a
stress disorder, in: Oxford Handbook of Stress and Mental Health, (eds.
K.  Harkness and E.P. Hayden), Oxford University Press, New York, 2018;
in press.
92. Kirkman, S. and Nelson, D.H. Alcohol‐induced pseudo‐Cushing’s disease:
a  study of prevalence with review of the literature. Metabolism,
1988;37:390–4.
93. Richardson, H.N., Lee, S.Y., O’Dell, L.E. et al. Alcohol self‐administration
acutely stimulates the hypothalamic‐pituitary‐adrenal axis, but alcohol
dependence leads to a dampened neuroendocrine state. Eur J Neurosci,
2008;28:1641–53.
94. Vendruscolo, L.F., Barbier, E., Schlosburg, J.E. et al. Corticosteroid‐depend-
ent plasticity mediates compulsive alcohol drinking in rats. J  Neurosci,
2012;32:7563–71.
95. Carlezon, W.A., Jr., Nestler, E.J., and Neve, R.L. Herpes simplex virus‐
mediated gene transfer as a tool for neuropsychiatric research. Crit Rev
Neurobiol, 2000;14:47–67.
96. Koob, G.F. Neurobiology of addiction, in Textbook of Substance Abuse
Treatment, 5th edn, (eds. M. Galanter et  al.), American Psychiatric
Publishing, Washington, DC, 2015, pp. 3–24.
97. Economidou, D., Hansson, A.C., Weiss, F. et al. Dysregulation of nocicep-
tin/orphanin FQ activity in the amygdala is linked to excessive alcohol
drinking in the rat. Biol Psychiatry, 2008;64:211–8.
98. Serrano, A. and Parsons, L.H. Endocannabinoid influence in drug
­reinforcement, dependence and addiction‐related behaviors. Pharmacol
Ther, 2011;132:215–41.
99. Koob, G.F. Alcoholism: allostasis and beyond. Alcohol Clin Exp Res,
2003;27:232–43.
100. Sullivan, E.V. and Pfefferbaum, A. Neurocircuitry in alcoholism: a sub-
strate of disruption and repair. Psychopharmacology, 2005;180:583–94.
101. Oscar‐Berman, M. Function and dysfunction of prefrontal brain circuitry in
alcoholic Korsakoff’s syndrome. Neuropsychol Rev, 2012;22:154–69.
102. Niendam, T.A., Laird, A.R., Ray, K.L. et al. Meta‐analytic evidence for a
superordinate cognitive control network subserving diverse executive func-
tions. Cogn Affect Behav Neurosci, 2012;12:241–68.
103. Olbrich, H.M., Valerius, G., Paris, C. et al. Brain activation during craving
for alcohol measured by positron emission tomography. Aust N Z J
Psychiatry, 2006;40:171–8.
104. Jasinska, A.J., Stein, E.A., Kaiser, J. et al. Factors modulating neural reac-
tivity to drug cues in addiction: a survey of human neuroimaging studies.
Neurosci Biobehav Rev, 2014;38:1–16.
105. Koob, G.F. and Volkow, N.D. Neurobiology of addiction: a neurocircuitry
analysis. Lancet Psychiatry, 2016;3:760–73.
106. Marlatt, G.J. Determinants of relapse: implications for the maintenance of
behavioral change, in Behavioral Medicine: Changing Health Lifestyles,
(eds. P. Davidson and S. Davidson), Brynner/Mazel, New York, 1980,
pp. 410–52.
107. Jochum, T., Boettger, M.K., Burkhardt, C. et al. Increased pain sensitivity
in alcohol withdrawal syndrome. Eur J Pain, 2010;14:713–8.
108. Zywiak, W.H., Connors, G.J., Maisto, S.A. et al. Relapse research and the
reasons for drinking questionnaire: a factor analysis of Marlatt’s relapse
taxonomy. Addiction, 1996;91: S121–30.
109. Martin‐Fardon, R. and Weiss, F. Modeling relapse in animals. Curr Top
Behav Neurosci, 2013;13:403–32.
110. McFarland, K. and Kalivas, P.W. The circuitry mediating cocaine‐induced
reinstatement of drug‐seeking behavior. J Neurosci, 2001;21:8655–63.
111. Everitt, B.J. and Wolf, M.E. Psychomotor stimulant addiction: a neural sys-
tems perspective. J Neurosci, 2002;22:3312–20.
112. Monti, P.M., Rohsenow, D.J., Hutchison, K.E. et al. Naltrexone’s effect on
cue‐elicited craving among alcoholics in treatment. Alcohol Clin Exp Res,
1999;23:1386–94.
113. Shaham, Y., Shalev, U., Lu, L. et  al. The reinstatement model of drug
relapse: history, methodology and major findings. Psychopharmacology,
2003;168:3–20.
114. Everitt, B.J., Belin, D., Economidou, D. et al. Review. Neural mechanisms
underlying the vulnerability to develop compulsive drug‐seeking habits and
addiction. Philos Trans R Soc Lond B Biol Sci, 2008;363:3125–35.
115. Jentsch, J.D. and Taylor, J.R. Impulsivity resulting from frontostriatal dys-
function in drug abuse: implications for the control of behavior by reward‐
related stimuli. Psychopharmacology, 1999;146:373–90.
116. Leggio, L. and Lee, M.R. Treatment of alcohol use disorder in patients with
alcoholic liver disease. Am J Med, 2017;130:124–34.
117. Bertholet, N., Daeppen, J.B., Wietlisbach, V. et  al. Reduction of alcohol
consumption by brief alcohol intervention in primary care: systematic
review and meta‐analysis. Arch Intern Med, 2005;165:986–95.
118. Lieber, C.S., Weiss, D.G., Groszmann, R. et al. II. Veterans affairs coopera-
tive study of polyenylphosphatidylcholine in alcoholic liver disease.
Alcohol Clin Exp Res, 2003;27:1765–72.
119. Khan, A., Tansel, A., White, D.L. et al. Efficacy of psychosocial interven-
tions in inducing and maintaining alcohol abstinence in patients with
chronic liver disease: a systematic review. Clin Gastroenterol Hepatol,
2016;14:191–202.
120. Leggio, L., Kenna, G.A., and Swift, R.M. New developments for the
­pharmacological treatment of alcohol withdrawal syndrome. A focus on
non‐benzodiazepine GABAergic medications. Prog Neuropsychopharmacol
Biol Psychiatry, 2008;32:1106–17.
700 THE LIVER:  REFERENCES

121. Jonas, D.E., Amick, H.R., Feltner, C. et  al. Pharmacotherapy for adults
with alcohol use disorders in outpatient settings: a systematic review and
meta‐analysis. JAMA, 2014;311:1889–900.
122. Litten, R.Z., Falk, D.E., Ryan, M.L. et al. Advances in pharmacotherapy
development: human clinical studies. Handb Exp Pharmacol, 2018;
248:579–613.
123. Addolorato, G., Leggio, L., Ferrulli, A. et al. Effectiveness and safety of
baclofen for maintenance of alcohol abstinence in alcohol‐dependent
patients with liver cirrhosis: randomised, double‐blind controlled study.
Lancet, 2007;370:1915–22.
124. Morley, K.C., Baillie, A., Fraser, I. et  al. Baclofen in the treatment of
­alcohol dependence with or without liver disease: multisite, randomised,
double‐blind, placebo‐controlled trial. Br J Psychiatry, 2018;
212:362–9.
125. Koob, G.F. and Volkow, N.D. Neurocircuitry of addiction. Neuropsycho­
pharmacology, 2010;35:217–38.
 Drug‐Induced Liver Injury
54 Lily Dara and Neil Kaplowitz
Research Center for Liver Disease, Department of Medicine, Division of Gastrointestinal and Liver
Diseases, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
INTRODUCTION
Drug‐induced liver injury (DILI) is an increasingly recognized
clinical problem encompassing over 50% of cases of acute liver
failure (ALF) in the United States [1]. First described as a dis-
tinct clinical entity by Hyman Zimmerman in the 1950s, DILI is
defined as liver injury with alanine aminotransferase (ALT)
more than 3–5 times the upper limit of normal (ULN) or an
alkaline phosphatase (ALP) more than twice the ULN with or
without elevated bilirubin. DILI remains a diagnostic challenge
due to a lack of pathognomonic findings and objective biomark-
ers and a wide range of presentation. DILI has a wide spectrum
of clinical manifestations from acute hepatitis with jaundice to
cholestatic DILI to rarer forms of DILI such as nodular regen-
erative hyperplasia (e.g. azathioprine), sinusoidal obstruction
syndrome (e.g. cyclophosphamide), steatohepatitis (e.g. tamox-
ifen), among others (Table 54.1). DILI can have a rapid or suba-
cute presentation with elevated liver enzymes or occur more
gradually and in rare instances present as a more chronic pro-
gressive disease [2]. As such, DILI is a diagnosis of exclusion
and is suspected when there is temporal association between
taking a drug and the onset of liver injury. Despite the various
potential presentations of drug hepatotoxicity, most often DILI
presents as drug‐induced hepatitis or cholestatic injury or a
combination of the two. Most drugs have a unique signature
which can be determined by calculating the R‐value. The R‐
value is defined as the ratio of ALT/ULN divided by the ALP/
ULN. It is commonly used as an index to determine the pheno-
type of liver injury. By convention, an R‐value ≥ 5 suggests a
drug‐induced hepatitis, an R‐value ≤ 2 suggests cholestatic
DILI, and a value in between (2 < R‐value < 5) suggests a mixed
picture  [3]. The phenotype of liver injury also has prognostic
The Liver: Biology and Pathobiology, Sixth Edition. Edited by Irwin M. Arias, Harvey J. Alter, James L. Boyer, David E. Cohen, David A. Shafritz,
Snorri S. Thorgeirsson, and Allan W. Wolkoff.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
value. Drug‐induced hepatitis, when accompanied by jaundice,
results in 10% mortality without transplant. This phenomenon,
first observed and reported by Hyman Zimmerman, is often
referred to as “Hy’s law”. Drug‐induced cholestasis is a more
indolent disease, associated with elevated ALP often due to
cholangiocyte damage and may be accompanied by pruritus.
Cholestatic DILI is a signature of certain drugs, is more com-
mon in the elderly and often resolves more slowly [4].
In the past few decades, hundreds of different drugs and
herbal compounds have been reported to cause DILI, with vari-
able latency, patterns of injury, and phenotypic presentation.
The underlying mechanism of liver injury and cell death has
also been extensively explored in vivo and in vitro using differ-
ent agents, most notably with the most widely used hepatotoxin,
APAP. In this chapter we will focus on the molecular mecha-
nisms and signaling pathways activated in DILI. We will not
cover herbal and dietary supplement‐induced liver injury and
although these compounds are not studied as systematically, it is
likely that the same principles and mechanisms apply.
DIRECT ACTING VERSUS
IDIOSYNCRATIC TOXICITY
The liver is an important target for drugs because lipophilic
drugs are metabolized in the liver by conversion to more aque-
ous soluble forms which can be excreted. Drug metabolism
involves the participation of phase I cytochrome P (Cyp)‐450
system, phase II conjugation, and phase III transporter proteins,
which modulate transformation and excretion. Although parent
drug accumulation may participate, usually toxic intermediates
702 THE LIVER:  ACETAMINOPHEN
Table 54.1  Phenotype and histopathological presentations of DILI
Clinical phenotype
Hepatocellular injury with non‐specific acute hepatitis on biopsy, zonal
necrosis, or spotty necrosis (mild to severe injury)
Hepatocellular injury with features of autoimmune hepatitis on biopsy
Hepatocellular injury with, steatohepatitis +/− fibrosis on biopsy
Hepatocellular injury with macrovesicular steatosis
Hepatocellular injury with microvesicular steatosis
Cholestasis and inflammation
Bland cholestasis
Chronic cholestasis with ductopenia
Vascular lesions
Portal hypertension and peliosis hepatitis, NRH, or SOS
Vascular lesions
Portal hypertension and sinusoidal obstruction syndrome on biopsy
Adenoma
Angiosarcoma
Hepatocellular carcinoma
participate in mediating toxicity. Certain drugs such as APAP,
aspirin, niacin, and chemotherapeutic agents can cause pre-
dictable, dose‐dependent injury directly to the liver. This direct
and metabolic toxicity is unique to certain drugs as most drugs
cause an idiosyncratic, unpredictable, and dose‐independent
form of hepatotoxicity often abbreviated IDILI (idiosyncratic
DILI). Among the direct hepatotoxins, APAP is the mostly
widely studied drug. First introduced by von Mering in 1893,
APAP was not widely used until the 1960s [5]. The first hepa-
totoxicity cases were reported in the same decade by Davidson
and Eastham who described two patients with fulminant liver
failure and centrilobular hepatic necrosis who subsequently
succumbed to APAP toxicity within three days [6]. The toxic-
ity from APAP was shown to be dose dependent and the drug
was initially considered safe up to 4 g per 24 hours. More
recent studies on healthy volunteers receiving a daily dose of
4 g of APAP have called the concept of a universal safe thresh-
old into question [7]. In addition to dose, fasting, acute illness,
concomitant drugs, and metabolic stress can result in liver
injury from lower doses of APAP, due to depletion of glu-
tathione (GSH) stores or modulation of biotransformation to
toxic intermediary.
In contrast to direct toxins, drugs that cause IDILI do so by
activating the immune system. The metabolism of drugs can
lead to the formation of intermediates which generate hap-
tenic‐peptide antigens. A number of studies have now shown
that patients who develop IDILI harbor specific single nucle-
otide polymorphisms (SNPs) in genes which encode for
human leukocyte antigen (HLA) regions. This association
strongly suggests a genetically determined immune predispo-
sition in certain individuals leading to an adaptive immune
response directed at drug hapten or modified autoantigens.
However, most of the identified HLA polymorphisms are rel-
atively common in the general population and thus carry a
small hazard ratio for the development of IDILI. Therefore,
other contributing factors must be influencing susceptibility
to IDILI [8].
Drugs
Acetaminophen, bromfenac, bupropion, carbon tetrachloride, cocaine,
diclofenac, halothane, heparin, INH, ipilimumab, ketoconazole,
methotrexate, methyldopa, pembrolizumab, phenytoin, propylthiouracil,
nivolumab, rifampin, statins, troglitazone
Methyldopa, minocycline, nitrofurantoin, statins
Amiodarone, methotrexate, tamoxifen
Anti‐psychotics, MTP (microsomal triglyceride transfer protein) inhibitors,
protease inhibitors
Aspirin (Reye’s syndrome), NRTI, tetracycline, valproic acid
Allopurinol, amoxicillin/clavulanic acid, carbamazepine, chlorpromazine,
hydralazine, methyldopa, penicillamine, phenylbutazone, phenytoin,
procainamide, quinidine, sulfasalazine, sulfonamides, sulindac
OCP, cyclosporin A, anabolic steroids
Chlorpromazine, chlorpropamide, chlorothiazide, rarely others
Anabolic steroids, azathioprine, OCPs (Budd–Chiari syndrome), HIV
drugs, busulfan, cyclophosphamide, pyrrolizidine alkaloids,
6‐mercaptopurine
Busulfan, cyclophosphamide, pyrrolizidine alkaloids
Anabolic steroids, oral contraceptives
Anabolic steroids, arsenic, copper, polyvinyl chloride, thorotrast
Aflatoxin, anabolic steroids, danazol
ACETAMINOPHEN
APAP is the single most common cause of ALF in the United
States. [9]. Hepatotoxicity from APAP occurs in a dose‐dependent,
predictable manner, much like the human injury, in rodents
treated with large doses of the drug. Rats are more resistant to
APAP; however mice are susceptible and exhibit characteristic
histologic findings, congestion around the central vein, fol-
lowed by vacuolization, and necrosis by 6 hours [5]. Mouse
models of APAP toxicity especially C57BL6 mice are com-
monly used to study APAP DILI and hepatocyte death. However,
marked mouse strain differences in susceptibility to APAP have
been reported [10]. APAP is the precursor to the toxic metabo-
lite, NAPQI (N‐acetyl‐p‐benzoquinone imine), formed by the
direct two electron oxidation of the parent drug by the Cyps
[11]. Multiple Cyps can generate NAPQI but the most common
enzyme is Cyp2E1. However, Cyp1A2, Cyp3A4, and Cyp2D6
may also participate. NAPQI is highly reactive and covalently
binds to intracellular proteins resulting in organelle stress. At
nontoxic doses, NAPQI is efficiently detoxified by GSH form-
ing an APAP‐GSH conjugate [12]. GSH synthesis is limited by
cysteine availability as the other two amino acids of GSH, glu-
tamate and glycine, are abundant. When GSH is depleted and
cysteine supply is limited, the NAPQI is free to attack protein‐
SH groups throughout the cell, with the mitochondria being the
key target [13]. This led to the development of the antidote,
N‐acetyl‐cysteine (Mucomyst®), which provides cysteine to the
liver and is highly effective if administered within 10 hours of
overdose in humans and 1.5–2 hours in mice [14]. In the absence
of GSH, NAPQI covalently binds intracellular proteins, and
APAP–protein adducts can be detected in the liver and serum.
The appearance of APAP–protein adducts in serum correlates
with hepatotoxicity and serum AST, ALT levels. These adducts
have been shown by immunoblot analysis to originate from the
liver [15]. This has led to the development of a highly sensitive
and specific HPLC‐EC assay for detection of acetaminophen
protein adducts (3‐cysteine‐acetaminophen in proteins) in the
 54:  Drug‐Induced Liver Injury 703
serum as specific biomarkers of lysis of hepatocytes in APAP
DILI [16, 17].
Despite the excellent correlation between NAPQI‐protein
adduct formation and toxicity, no direct causality between
adduct formation and hepatocyte necrosis has been demon-
strated. The removal of APAP–protein adducts though selective
autophagy occurs within 24 hours [18]. NAPQI–protein adducts
in the presence of depleted GSH stores result in mitochondrial
damage and increase of intracellular peroxides and reactive
oxygen species (ROS) and via an iron‐mediated mechanism
generate hydroxyl radical (Fenton reaction), as well as, perox-
ynitrite through the interaction of superoxide with mitochon-
drial nitric oxide (NO) [19]. Hinson and colleagues have
demonstrated the appearance of 3‐nitro‐tyrosine (a marker of
nitrogen stress) in mouse models of APAP in the centrilobular
zone, mainly in the mitochondria [20]. Furthermore, the devel-
opment of the nitrated proteins in hepatocytes correlated with
the necrotic cell death [20]. The importance of mitochondrial
damage in the pathogenesis of APAP‐induced necrosis has been
known since the 1980s when electron microscopic examination
of livers from acetaminophen treated mice indicated mitochon-
drial damage [21]. Functionally, APAP has been shown to alter
mitochondrial respiration and effect the electron transport chain
(ETC) both in vitro and in isolated hepatocytes from in vivo
treated animals [22, 23]. Generation of ROS subsequently exac-
erbates mitochondrial stress and may induce endoplasmic retic-
ulum (ER) stress leading to the activation of intracellular
signaling pathways, most importantly the mitogen activated
protein kinase (MAPK) pathway [24]. Interference with various
MAPK proteins, including mixed lineage kinase protein 3
(MLK3), apoptosis signal inducing kinase (ASK1), mitogen
activated protein kinase 4 (MKK4), and C Jun‐N‐terminal
kinase (JNK) as well as the JNK binding partner, SH3BP (Sab)
markedly protects against APAP induced hepatocyte death [23,
25–29]. Xenobiotic stress, nutrient deprivation, or organelle
stress leads to the activation of upstream higher order MAPKs
which through cascading phosphorylation events ultimately
phosphorylate JNK. Under normal conditions, JNK activation
by default is transient, as sustained p‐JNK leads to cell death.
When stress signals exceed a certain threshold (such as toxic
doses of APAP), p‐JNK interacts with mitochondria by binding
to the kinase interacting motif of Sab, a mitochondrial outer
membrane protein [30]. The interaction of JNK with Sab leads
to the generation of ROS and sustained JNK activation, result-
ing in a feedforward mechanism. JNK does not enter the mito-
chondria. However, an intra‐mitochondrial signaling pathway
involving the tyrosine phosphatase SH2 phosphatase 1 (SHP1),
resulting in dephosphorylation of activated‐Src (proto‐onco-
gene c‐Src; Src short for sarcoma), which occurs on and requires
the platform, DOK4 (an inner membrane protein) has been
demonstrated [31]. The deactivation of Src is thought to increase
mitochondrial ROS generation by dampening electron transport
leading to the buildup of reducing equivalents. The JNK‐
dependent increased mitochondrial ROS then amplifies mito-
chondrial stress and leads to ROS sensitive mitochondrial
permeability transition (MPT) (Figure  54.1). MPT inhibitors
such as cyclosporine A inhibit APAP toxicity in vitro and in vivo
in mice [32, 33]. Cyclophilin D deficient mice are also been
reported to be protected from APAP toxicity and cell death,
further consolidating the importance of mitochondria in this
form of DILI [34]. Following MPT a combination of collapse of
mitochondrial ATP production and release of DNA damaging
proteins from the mitochondria result in cell death. In addition
to the MAPKs, other kinases and signaling molecules have been
implicated in hepatocyte death from APAP. Knockout, knock-
down or inhibition of glycogen synthase kinase 3 beta (GSK3β),
receptor interacting protein kinase‐1 (RIPK1), and dynamin
related protein‐1 (DRP1) have all been shown to abrogate or
prevent hepatocyte death from APAP [35–37]. While receptor
interacting protein kinase 1 (RIPK1) is thought to participate in
APAP toxicity upstream of JNK, the role of receptor interacting
protein kinase‐3 (RIPK3) is more controversial [38]. Although
APAP‐DILI results in necrotic cell death which is clearly regu-
lated through signaling mechanisms, it is not a form of necrop-
tosis, as mixed lineage kinase domain like (MLKL) knockout
does not protect against APAP toxicity [36].
IDIOSYNCRATIC DILI
Idiosyncratic DILI (IDILI) remains a major diagnostic chal-
lenge. It is a diagnosis of exclusion and common liver disorders
such as viral hepatitis, autoimmune hepatitis (AIH), biliary tract
disease, ischemic hepatitis, heart failure, and circulatory dys-
function as well as sepsis and alcoholic liver disease must be
ruled out. If a high index of suspicion is present for drug toxic-
ity, the more common causes of liver disease are ruled out, and
the drug is taken within the correct time frame (latency), and the
injury phenotypically resembles the drug’s signature (if known),
the probability of an IDILI event is high. A liver biopsy can be
helpful if eosinophilic infiltrates, granulomas, or centrilobular
necrosis is noted. Since there is no specific test to determine
causality and even the liver biopsy is usually nonspecific, the
diagnosis of IDILI remains challenging [8].
The exact pathophysiology of IDILI and why it occurs is
likely multifactorial involving both drug and pharmacologic
factors such as dose and lipophilicity as well as host and indi-
vidual factors such as genetics, the immune response, and
defective adaptive responses. Evidence for the involvement of
the immune system in IDILI is strong. For example, certain
drugs such as halothane, dihydralazine, and anticonvulsants are
associated with an allergic type hypersensitivity and present
with rash and eosinophilia. Other notable drugs that have been
reported to present with rash and eosinophilia include: trimetho-
prim/sulfamethoxazole, cefazolin, and ciprofloxacin [39].
Severe allergic skin reactions such as toxic epidermal necrosis
(TEN) and Stevens–Johnson syndrome (SJS) have been seen
with carbamazepine and phenytoin IDILI and carry a poor prog-
nosis [39]. Some drugs, such as nitrofurantoin and minocycline
can cause DILI which mimics autoimmune hepatitis with a
positive anti‐nuclear antibody (ANA) and plasma cell infiltrates
which are often indistinguishable from AIH on liver biopsy
[39]. This has led to the hypothesis that these drugs may be acti-
vating AIH in patients with a genetic predisposition [40].
However, the frequencies of AIH‐associated HLA alleles,
DRB1*0301 and DRB1*0401, were not increased in patients
with AIH type DILI presentation compared to controls [41].
704 THE LIVER:  IDIOSYNCRATIC DILI

Figure 54.1  Signaling pathway in acetaminophen (APAP) induced liver injury. APAP is metabolized by cytochrome P450 2E1 (CYP2E1) to a
reactive metabolite called NAPQI N‐acetyl‐p‐benzoquinone imine which depletes glutathione (GSH) followed by covalent binding to intracellular
proteins which induces cellular and organelle stress. NAPQI targets mitochondria resulting in the generation of reactive oxygen species (ROS).
ROS subsequently activate the mitogen activated protein kinase (MAPK) cascade. Mixed‐lineage protein kinase 3 (MLK3) is activated in the early
phase and apoptosis signal‐inducing kinase‐1 (ASK1) in the later phase of injury. MLK3 and ASK1 phosphorylate mitogen activated protein kinase
4 (MKK4) which goes on to phospho‐activate c‐Jun N‐terminal kinase (JNK). Other kinases such as receptor interacting protein kinase 1 (RIPK1),
glycogen synthase kinase‐3β (GSK3β) and protein kinase C‐α (PKCα) have also been reported to activate JNK. Activated JNK (p‐JNK) binds to
SH3BP (Sab) on the outer mitochondrial membrane. This results in the activation of an intra‐mitochondrial pathway involving a SH2 phosphatase
(SHP1) and the docking protein DOK4 (located on the inner membrane of mitochondria) which ultimately results in the dephosphorylation of
intra‐mitochondrial Src. The deactivation of Src impairs mitochondrial electron transport, increasing mitochondrial ROS generation. The JNK‐
dependent increased mitochondrial ROS then amplifies mitochondrial stress and sustains p‐JNK in an active form. This eventually leads to mito-
chondrial permeability transition (MPT) and release of inter‐mitochondrial proteins which cleave nuclear DNA.

Large databases of IDILI patients have looked for evidence
of a genetic predisposition to IDILI and interestingly many
large genome wide association studies (GWAS) have identified
polymorphism in HLA regions in patients with IDILI due to
various agents. Since the HLA complex codes for the major his-
tocompatibility complex (MHC) proteins in humans, variants in
HLA molecules can result in aberrant antigen presentation and
inappropriate activation of an immune response. The HLA‐A,
HLA‐B, and HLA‐C genes govern the structure of MHC I mol-
ecules, while HLA‐DP, HLA‐DQ, and HLA‐DR control the
structure of MHC II. Most identified SNPs are in the region of
MHC II molecules. This can affect antigen presentation to
T helper cells (CD4+) and thus partly explain why certain indi-
viduals develop IDILI.
Further supporting the contribution of immunity to DILI are
lymphocyte stimulation tests demonstrating drugs or drug–­
protein adducts can activate peripheral blood lymphocytes from
affected individuals [42, 43]. Early studies using the lympho-
cyte transformation test, a simple in vitro assay based on assess-
ment of lymphocyte proliferative responses in drug‐treated and
vehicle control cultures, detected drug‐specific lymphocyte
responses in approximately 50% of patients with DILI [44].
Drug‐specific T cells from IDILI patients have been shown to
be activated in an HLA restricted manner, also suggesting an
adaptive immune pathogenesis [42]. Using healthy donors with
known HLA risk alleles for certain drugs (HLA‐B*15:02,
HLA‐B*57:01, and HLA‐B*58:01) and donors without the risk
mutations, priming of naïve T cells was demonstrated to skew
 54:  Drug‐Induced Liver Injury 705

toward donors expressing the previously identified specific read‐outs have any bearing to the clinical scenario in idiosyn-
HLA alleles in GWAS studies [45]. T cells w