2010 Book RNATechnologiesAndTheirApplica

RNA Technologies
For further volumes:

http://www.springer.com/series/8619
.
Volker A. Erdmann l Jan Barciszewski
Editors
RNA Technologies and

Their Applications
Editors
Prof. Dr. Volker A. Erdmann Prof. Dr. Jan Barciszewski
FU Berlin PAN Poznan
Inst. Chemie und Biochemie Inst. Bioorganic Chemistry
Thielallee 63 ul. Z. Noskowskiego 12/14
14195 Berlin 61-704 Poznan
Germany Poland
erdmann@chemie.fu-berlin.de jan.barciszewski@ibch.poznan.pl
ISBN 978-3-642-12167-8 e-ISBN 978-3-642-12168-5

DOI 10.1007/978-3-642-12168-5
Springer Heidelberg Dordrecht London New York
Library of Congress Control Number: 2010934375
# Springer-Verlag Berlin Heidelberg 2010

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication
or parts thereof is permitted only under the provisions of the German Copyright Law of September 9,
1965, in its current version, and permission for use must always be obtained from Springer. Violations
are liable to prosecution under the German Copyright Law.
The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective
laws and regulations and therefore free for general use.
Cover design: deblik, Berlin
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Preface
For almost 50 years, our understanding of molecular processes was very much
influenced by a statement formulated by Francis Crick, which is known as the
“central dogma of molecular biology.” According to this dogma, DNA is converted
to an mRNA and the mRNA to a protein. Another way of stating this dogma is that
one gene is converted to one mRNA and the mRNA to one protein with one
function. The present day molecular biology relies basically on genomic data for
creating new hypotheses, which allow the replacement of the term “descriptive
science” by the much more attractive “discovery science.” The discovery science
has revolutionized biology and gave new tools for hypothesis-driven research,
which concerns primarily, but not exclusively, nucleic acids.
The interest to apply RNA structural and functional characteristics in molecular
biology and medicine began in the late 1980s, when catalytic RNAs and in vitro
selection approaches were an exciting new frontier. Now, thirty years after those
discoveries, we begin to understand the novel aspects of RNA biology. Although
the pathways and molecular components involved in RNA-mediated gene regula-
tion are being elucidated very rapidly, the chemical and mechanistic basis still has
to be worked out. The understanding of molecular mechanisms, and the possibi-
lities for employing these processes for therapeutic purposes, falls surely into the
realm of chemical biology.
The causal relationship between sequence, structure, and function significantly
affects the interaction of RNA molecules with proteins, metabolites, and other
nucleic acids, making RNA a malleable and attractive molecule to drive program-
mable function. RNA molecules, which derive sophisticated behavior from an
ability to adopt complex structures, can be generated from potentially all possible
sequence combinations, leading to diverse secondary structures and functions.
These structures can exist in the form of modular domains, which confer specific
and unique functionality. RNA molecules have evolved to regulate gene expression
in a wide variety of ways in cells and viruses. Despite that, we are only beginning to
appreciate how much of known phenotypic variation can be explained by these
novel classes of RNA regulators.
v
vi Preface
The recognition of the biological roles of small molecular weight RNAs has
been one of the most significant discoveries in molecular biology. These RNA
molecules influence the translation of messenger RNAs (mRNAs) in posttranscrip-
tional manner that makes the regulation of RNAs even more complex.
Recent advances in RNA biology and nucleic acid engineering are inspiring the
use of RNA molecules for the construction of different RNA tools. RNA has
become a focus of investigations into novel therapeutic schemes. Oligonucleotide-
based approaches depend on the Watson–Crick base pairing of oligonucleotides to
their corresponding mRNA target. This leads to posttranscriptional gene silencing
by mRNA cleavage or translational inhibition. Oligonucleotide-based therapies
have great potential for the treatment of RNA virus infections and various
diseases. They include antisense oligonucleotides and their derivatives such as
peptide nucleic acids, locked nucleic acids and morpholinos oligonucleotides
(ONs), RNAi, microRNA ribozymes, aptamers, and Spiegelmers. Beyond sequence
conservation, a very important point is the fact that the RNA target must be
accessible for oligonucleotide interaction. Although the effects of antisense
RNAs on the corresponding sense RNAs have not been clearly established, a
number of examples indicate that they may exert control at various levels of gene
expression, such as transcription, mRNA processing, splicing, stability, transport,
and translation.
Multiple challenges, such as optimization of selectivity, stability, delivery, and
long-term safety, have to be addressed in order for RNA drugs to become a
successful therapeutic tool. Not all RNA classes (e.g., ribozymes or RNA decoys)
were so far successfully developed as drugs. The use of RNA-mediated interference
(RNAi) for gene silencing has provided a powerful tool for loss-of-function studies
in a variety of metazoans. siRNA-mediated gene silencing by degradation of target
messenger RNA has been widely used for the functional characterization of genes.
The secondary structure of mRNA target sites has been reported to strongly
influence RNAi activity. Compared with the laborious, time-consuming, and very
costly gene knockout models, siRNA provides an efficient, specific, and economic
solution for inhibiting the expression of target genes. Efficient siRNA delivery is
essential for the success of specific gene silencing and is therefore understandable
that a number of different laboratories are currently working on the problem of
siRNA delivery in living organisms. Because high doses of siRNAs do provoke an
altered expression of many other genes, selection of an optimal condition could be
very helpful to minimize potential side effects. The advantage of the system lies in
the application of short RNAs, which can be synthesized relatively cheap and can
evolve quickly, to regulate a large and complex protein synthesis.
In this volume, 10 papers out of 19 are dealing with various aspects of RNA
interference. They cover basic issues of the technique and its application in biology
and medicine. There are also three contributions on antisense RNA approaches,
which show a high therapeutic potential.
It is becoming clear that microRNAs are essential regulators of many of the key
pathways implicated in tumor pathogenesis. While adding another layer of com-
plexity, the discovery of the role of miRNAs in tumorigenesis has revealed a new
Preface vii
category of therapeutic targets. As miRNA studies continue to be developed, novel

therapeutic targets for different types of tumors will continue to emerge.
The discovery of regulatory RNAs has revolutionized the traditional concept of
RNA function and gene regulation. The fact that the protein coding RNA portion
represents less than 1.5% of the total transcriptional output and the rest represent
noncoding RNA (ncRNA) implies that apart from cis regulatory DNA sequences,
ncRNA could also perform much of the regulatory tasks of complex organisms.
ncRNAs have been shown to control every level of the multilevel-regulated gene
expression pathway, including gene silencing. Small ncRNAs are highly conserved
at the sequence level and regulate transcriptional and posttranscriptional gene
silencing through specific pairing with their target genes, whereas long ncRNAs
are poorly conserved and regulate transcriptional silencing ranging from a single
gene to an entire chromosome through diverse mechanisms not involving any base
pair interactions with the target genes. In the book, we have also included three
chapters on ncRNAs and their functions and therapeutic potential.
We hope that the book will be of interest for biochemists and life scientists and
that it will stimulate their future research.
Berlin, Germany Volker A. Erdmann

Poznan, Poland Jan Barciszewski
March, 2010
.
Contents
The Key Features of RNA Silencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Kuniaki Saito, Keita Miyoshi, Mikiko C. Siomi, and Haruhiko Siomi
Selected Strategies for the Delivery of siRNA In Vitro

and In Vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Sandra D. Laufer, Anke Detzer, Georg Sczakiel, and Tobias Restle
RNAi Suppression and Its Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Xiaoping Yi and Rui Lu
Strategies to Prevent siRNA-Triggered Cellular Toxicity . . . . . . . . . . . . . . . . . 93

Matthias Bauer
RNAi in Malignant Brain Tumors: Relevance to Molecular

and Translational Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Mitsutoshi Nakada, Daisuke Kita, Yutaka Hayashi, Kazuyuki Kawakami,
Jun-ichiro Hamada, and Toshinari Minamoto
Silencing Huntington’s Disease Gene with RNAi . . . . . . . . . . . . . . . . . . . . . . . . . . 131

Yu Zhang and Robert M. Friedlander
Application of Dicer-Substrate siRNA in Pain Research . . . . . . . . . . . . . . . . . 161

Philippe Sarret, Louis Doré-Savard, Pascal Tétreault,
Valérie Bégin-Lavallée, and Nicolas Beaudet
RNAi Treatment of HIV-1 Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

Karin J. von Eije and Ben Berkhout
ix
x Contents
Application of RNA Interference to Treat Conditions Associated

with Dysregulation of Transient Receptor Potential
Vanilloid 1 Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Vickram Ramkumar, Debashree Mukherjea, Sarvesh Jajoo,
Tejbeer Kaur, and Leonard P. Rybak
Harnessing RNAi-Based Functional Genomics to Unravel

the Molecular Complexity Underlying Skin Pigment Variation . . . . . . . . . 227
Hsiang Ho, Jayavani Aruri, Safoora Ahmed, and Anand K. Ganesan
mRNA Structure and its Effects on Posttranscriptional

Gene Silencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Stephen I. Rudnick, Veenu Aishwarya, and Alan M. Gewirtz
Antisense RNA-Mediated Regulation of the p53 Tumor

Suppressor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Marianne Farnebo and Klas G. Wiman
Antisense Oligonucleotides: Insights from Preclinical Studies

and Clinical Trials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
Doreen Kunze, Kai Kraemer, and Susanne Fuessel
What can the New Hammerhead Ribozyme Structures

Teach us About Design? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
William G. Scott
microRNA Biogenesis and its Impact on RNA Interference . . . . . . . . . . . . . . 325

Stefanie Grund and Sven Diederichs
MicroRNAs in Epithelial Antimicrobial Immunity . . . . . . . . . . . . . . . . . . . . . . . 355

Jun Liu, Guoku Hu, Rui Zhou, Kristen M. Drescher,
and Xian-Ming Chen
Emerging Roles of Long Noncoding RNAs in Gene Expression

and Intracellular Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Tetsuro Hirose
Noncoding RNAs as Therapeutic Targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393

Maciej Szymański and Jan Barciszewski
Noncoding RNAs at H19/IGF2 Locus: Role in Imprinting,

Gene Expression, and Associated Pathologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
Nahalie Berteaux, Nathalie Spruyt, and Eric Adriaenssens
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
Contributors
Eric Adriaenssens Institut de Biology de Lille, CNRS UMR 8161, 1 rue Pr

Calmette, BP 447 59021 Lille Cedex, France, eric.adriaenssens@univ-lille1.fr
Safoora Ahmed Department of Dermatology and Biological Chemistry, University

of California, 324 Sprague Hall, Irvine, CA 92697-2400, USA
Veenu Aishwarya Division of Hematology/Oncology, Department of Medicine,

University of Pennsylvania School of Medicine, Philadelphia, PA, USA
Jayavani Aruri Department of Dermatology and Biological Chemistry, University

of California, 324 Sprague Hall, Irvine, CA 92697-2400, USA
Jan Barciszewski Institute of Bioorganic Chemistry of the Polish Academy of

Sciences, Noskowskiego 12, 61-704 Poznan, Poland, Jan.Barciszewski@ibch.
poznan.pl
Matthias Bauer Department of Neurology, Hertie Institute for Clinical Brain

Research, University of Tübingen, Tübingen, Germany, matthias.bauer@helmh
oltz-muenchen.de; Department of Protein Sciences, Helmholtz Center München,
German Research Center for Environmental Health, München-Neuherberg,
Germany; Institute for Human Genetics, Klinikum rechts der Isar, TU München,
München, Germany
Nicolas Beaudet Department of Physiology and Biophysics, Faculty of Medicine

and Health Sciences, Université de Sherbrooke, 3001, 12Ème Avenue Nord,
Sherbrooke, Québec, Canada J1H 5N4; Centre des Neurosciences de Sherbrooke,
Université de Sherbrooke, 3001, 12Ème Avenue Nord, Sherbrooke, Québec, Canada
J1H 5N4
xi
xii Contributors
Valérie Bégin-Lavallée Department of Physiology and Biophysics, Faculty of

Medicine and Health Sciences, Université de Sherbrooke, 3001, 12Ème Avenue
Nord, Sherbrooke, Québec, Canada J1H 5N4; Centre des Neurosciences de
Sherbrooke, Université de Sherbrooke, 3001, 12Ème Avenue Nord, Sherbrooke,
Québec, Canada J1H 5N4
Ben Berkhout Center for Infection and Immunity Amsterdam (CINIMA),

Department of Medical Microbiology, Laboratory of Experimental Virology,
Academic Medical Center of the University of Amsterdam, Amsterdam, The
Netherlands, b.berkhout@amc.uva.nl
Nahalie Berteaux Institut de Biology de Lille, CNRS UMR 8161, 1 rue Pr

Calmette, BP 447 59021, Lille Cedex, France
Xian-Ming Chen Department of Medical Microbiology and Immunology,

Creighton University Medical Center, Omaha, NE 68178, USA, xianmingchen@
creighton.edu
Anke Detzer Institut für Molekulare Medizin, ZMSZ, Universität zu Lübeck &
Universitätsklinikum Schleswig-Holstein, Ratzeburger Allee 160, 23538 Lübeck,
Germany
Sven Diederichs Helmholtz-University-Group “Molecular RNA Biology &

Cancer”, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 280
(B150), 69120 Heidelberg, Germany, s.diederichs@dkfz.de; Institute of Pathology,
University of Heidelberg, Im Neuenheimer Feld 280 (B150), 69120 Heidelberg,
Germany
Louis Doré-Savard Department of Physiology and Biophysics, Faculty of

Medicine and Health Sciences, Université de Sherbrooke, 3001, 12Ème Avenue
Nord, Sherbrooke, Québec, Canada J1H 5N4; Centre des Neurosciences de
Sherbrooke, Université de Sherbrooke, 3001, 12Ème Avenue Nord, Sherbrooke,
Québec, Canada J1H 5N4
Kristen M. Drescher Department of Medical Microbiology and Immunology,

Creighton University Medical Center, Omaha, NE 68178, USA
Karin von Eije Center for Infection and Immunity Amsterdam (CINIMA),
Department of Medical Microbiology, Laboratory of Experimental Virology,
Academic Medical Center of the University of Amsterdam, Amsterdam, The
Netherlands
Marianne Farnebo Department of Oncology-Pathology, Cancer Center Karolinska

(CCK), Karolinska Institutet, SE-171 76 Stockholm, Sweden, Marianne.Farnebo@ki.se
Contributors xiii
Robert M. Friedlander Neuroapoptosis Laboratory, Department of Neurosurgery,

Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, 02115,
USA, rfriedlander@rics.bwh.harvard.edu
Susanne Fuessel Department of Urology, Medical Faculty, Dresden University of

Technology, Fetscherstraße 74, 01307 Dresden, Germany
Anand K. Ganesan Department of Dermatology and Biological Chemistry,

University of California, 324 Sprague Hall, Irvine, CA 92697-2400, USA
Alan M. Gewirtz Division of Hematology/Oncology, Department of Medicine,

University of Pennsylvania School of Medicine, Philadelphia, PA, USA, gewirtz@
mail.med.upenn.edu
Stefanie Grund Helmholtz-University-Group “Molecular RNA Biology &

Cancer”, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 280
(B150), 69120 Heidelberg, Germany, s.grund@dkfz.de; Institute of Pathology,
University of Heidelberg, Im Neuenheimer Feld 280 (B150), 69120 Heidelberg,
Germany
Jun-ichiro Hamada Department of Neurosurgery, Graduate School of Medical

Science, Kanazawa University, 13-1 Takara-machi, Kanazawa 920-8640, Japan
Yutaka Hayashi Department of Neurosurgery, Graduate School of Medical

Science, Kanazawa University, 13-1 Takara-machi, Kanazawa 920-8640, Japan
Tetsuro Hirose Biomedicinal Information Research Center, National Institute

of Advanced Industrial Science and Technology (AIST), 2-42 Aomi, Koutou
135-0064, Tokyo, Japan, tets-hirose@aist.go.jp
Hsiang Ho Department of Dermatology and Biological Chemistry, University of

California, 324 Sprague Hall, Irvine, CA 92697-2400, USA
Guoku Hu Department of Medical Microbiology and Immunology, Creighton

University Medical Center, Omaha, NE 68178, USA
Sarvesh Jajoo Department of Pharmacology, Southern Illinois University School

of Medicine, PO Box 19629, Springfield, IL 62794-9629, USA
Tejbeer Kaur Department of Pharmacology, Southern Illinois University School

of Medicine, PO Box 19629, Springfield, IL 62794-9629, USA
Kazuyuki Kawakami Divisions of Translational and Clinical Oncology, Kana-

zawa University, 13-1 Takara-machi, Kanazawa 920-8640, Japan
xiv Contributors
Daisuke Kita Department of Neurosurgery, Graduate School of Medical Science,

Kanazawa University, 13-1 Takara-machi, Kanazawa 920-8640, Japan
Kai Kraemer Department of Urology, Medical Faculty, Dresden University of

Technology, Fetscherstraße 74, 01307 Dresden, Germany
Doreen Kunze Department of Urology, Medical Faculty, Dresden University

of Technology, Fetscherstraße 74, 01307 Dresden, Germany, doreen.kunze@
uniklinikum-dresden.de
Sandra D. Laufer Institut für Molekulare Medizin, ZMSZ, Universität zu Lübeck

& Universitätsklinikum Schleswig-Holstein, Ratzeburger Allee 160, 23538
Lübeck, Germany, laufer@imm.uni-luebeck.de
Jun Liu Department of Medical Microbiology and Immunology, Creighton

Rui Lu Department of Biological Sciences, Louisiana State University, Baton

Rouge, LA 70803, USA, ruilu@lsu.edu
Toshinari Minamoto Divisions of Translational and Clinical Oncology, Kanazawa

University, 13-1 Takara-machi, Kanazawa 920-8640, Japan, minamoto@staff.
kanazawa-u.ac.jp; Surgical Oncology, Cancer Research Institute, Kanazawa Uni-
versity, 13-1 Takara-machi, Kanazawa, 920-8640, Japan
Keita Miyoshi Department of Molecular Biology, Keio University School of

Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
Debashree Mukherjea Department of Surgery, Southern Illinois University

School of Medicine, PO Box 19629, Springfield, IL 62794-9629, USA
Mitsutoshi Nakada Department of Neurosurgery, Graduate School of Medical

Science, Kanazawa University, 13-1 Takara-machi, Kanazawa 920-8640, Japan,
nakada@ns.m.kanazawa-u.ac.jp
Vickram Ramkumar Department of Pharmacology, Southern Illinois University

School of Medicine, PO Box 19629, Springfield, IL 62794-9629, USA, vramkumar@
siumed.edu
Tobias Restle Institut für Molekulare Medizin, ZMSZ, Universität zu Lübeck &
Germany
Contributors xv
Stephen I. Rudnick Division of Hematology/Oncology, Department of Medicine,

University of Pennsylvania School of Medicine, Philadelphia, PA, USA; Fox
Chase Cancer Center, Philadelphia, PA, USA
Leonard P. Rybak Department of Pharmacology, Southern Illinois University

School of Medicine, PO Box 19629, Springfield, IL, 62794-9629, USA; Department
of Surgery, Southern Illinois University School of Medicine, PO Box 19629,
Springfield, IL 62794-9629, USA
Kuniaki Saito Department of Molecular Biology, Keio University School of

Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
Philippe Sarret Department of Physiology and Biophysics, Faculty of Medicine

Université de Sherbrooke, 3001, 12Ème Avenue Nord, Sherbrooke, Québec,
Canada J1H 5N4
William G. Scott Department of Chemistry and Biochemistry and The Center

for the Molecular Biology of RNA, University of California at Santa Cruz, 228
Sinsheimer Laboratories, Santa Cruz, CA 95064, USA, wgscott@chemistry.ucsc.edu
Georg Sczakiel Institut für Molekulare Medizin, ZMSZ, Universität zu Lübeck &
Germany
Mikiko C. Siomi Department of Molecular Biology, Keio University School of

Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo, 160-8582, Japan, siomim@sc.
itc.keio.ac.jp
Nathalie Spruyt Institut de Biology de Lille, CNRS UMR 8161, 1 rue Pr Calmette,
BP 447 59021, Lille Cedex, France
Maciej Szymański Institute of Bioorganic Chemistry of the Polish Academy of

Sciences, Noskowskiego 12, 61-704 Poznan, Poland, mszyman@ibch.poznan.pl
Pascal Tétreault Department of Physiology and Biophysics, Faculty of Medicine

Université de Sherbrooke, 3001, 12Ème Avenue Nord, Sherbrooke, Québec,
Canada J1H 5N4
Klas G. Wiman Department of Oncology-Pathology, Cancer Center Karolinska

(CCK), Karolinska Institutet, SE-171 76 Stockholm, Sweden
xvi Contributors
Xiaoping Yi Department of Biological Sciences, Louisiana State University,

Baton Rouge, LA 70803, USA
Yu Zhang Neuroapoptosis Laboratory, Department of Neurosurgery, Brigham and

Women’s Hospital, Harvard Medical School, Boston, MA 02115, USA
Rui Zhou Department of Medical Microbiology and Immunology, Creighton

The Key Features of RNA Silencing
Kuniaki Saito, Keita Miyoshi, Mikiko C. Siomi, and Haruhiko Siomi
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 RNA Silencing Effector as a Two-Component System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3 Small RNA Biogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.1 miRNAs and siRNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.2 Dicer-Independent Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.3 Endo-siRNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4 RISC Loading, Sorting, and Target-Sensing of Small RNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.1 Sorting by Precursor Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2 Sorting by the 50 Ends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3 Sorting by Dicer Processing Polarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.4 Target-Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5 Safeguards for RNA Silencing Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
6 Effector Modes of RNA Silencing Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
7 Regulations of RNA Silencing Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
7.1 Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
7.2 Modification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
7.3 RISC Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
8 Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Abstract The discovery by Fire and Mello in 1998 of sequence-specific gene

silencing as a response to double-stranded RNAs (dsRNAs), termed RNA inter-
ference (RNAi), has had an enormous impact on biology. In RNAi and related
pathways, small noncoding RNAs of 20–30 nucleotides (nt) guide regulatory
complexes to RNA targets via base-pairing and promote the inactivation of
homologous sequences by a variety of mechanisms, thereby adding a great level
K. Saito, K. Miyoshi, M.C. Siomi (*), and H. Siomi

Department of Molecular Biology, Keio University School of Medicine, 35 Shinanomachi,
Shinjuku-ku, Tokyo 160-8582, Japan
e-mail: siomim@sc.itc.keio.ac.jp
V.A. Erdmann and J. Barciszewski (eds.), RNA Technologies and Their Applications, 1
RNA Technologies, DOI 10.1007/978-3-642-12168-5_1,
2 K. Saito et al.
of complexity to the way cells regulate protein levels. These pathways, which are
collectively referred to as RNA silencing, mediate biological activities that fall
into two broad categories; genomic surveillance and gene regulation. RNA silenc-
ing occurs in a variety of organisms and is evolutionarily conserved. Central to
these processes is small RNA generation by Dicer and inactivation of cognate
RNA targets by small RNA–Argonaute complexes acting in combination with a
multitude of interacting and collaborating proteins. In some systems, silencing
signals are amplified and small RNAs are produced by a Dicer-independent
pathway, which challenges our perception and definition of RNAi. There has
been remarkable progress in our understanding of the mechanisms underlying
RNAi and related silencing processes, which hails the prospect of fully decipher-
ing the RNAi machinery.
Keywords RNA Silencing RNAi Argonaute Dicer small RNA
1 Introduction
The mechanism underlying the RNA-induced defects appears to be distinct from that of
conventional antisense RNA because both the sense and antisense RNA strands cause
similar defects. Because the mechanism is not known, we will refer to this technique as
RNAi, for RNA-mediated interference (Rocheleau et al. 1997).
Craig Mello chose the term “RNAi” to refer to a mysterious gene silencing process
first observed in 1997. Then, in 1998, the key molecule in the process was found to
be dsRNA (Fire et al. 1998). The discovery of RNAi heralded a new RNA revolu-
tion and led to the discovery of “hidden layers” of gene expression regulation, in
which many previously unidentified families of small RNAs, 20–30 nt in length,
mediate gene silencing. These findings led to the unification of a number of
different RNA-based silencing pathways, including microRNA (miRNA)- and
small interfering RNA (siRNA)-mediated silencing in plants and animals, cosup-
pression and paramutation in plants, quelling in fungi, heterochromatin formation
in fission yeast, and RNA-directed DNA methylation (Stefani and Slack 2008; Ding
and Voinnet 2007; Girard and Hannon 2008; B€ uhler and Moazed 2007). In addi-
tion, RNAi has rapidly become one of the most powerful and indispensable
functional genomics tools and is also considered as a novel and invaluable clinical
therapeutic approach to specifically target genes associated with a variety of dis-
eases (Novina and Sharp 2004). Key steps in the RNAi pathway are shared by a
diverse set of gene regulatory mechanisms, including mechanisms that silence
endogenous genes, particularly genes involved in development and stem cell
maintenance, and mechanisms that restrain the expression of transposons or viruses
and that direct transcriptional gene silencing (Stefani and Slack 2008; Ding and
Voinnet 2007; Girard and Hannon 2008; B€ uhler and Moazed 2007; Siomi and
Siomi 2008).
The Key Features of RNA Silencing 3
Long dsRNA is the trigger molecule in RNAi and can be derived from various
sources, such as simultaneous sense and antisense transcription of specific genomic
loci, foldback-structured transcripts from repetitive sequences, and viral replication
intermediates. The basic biochemical requirements of an RNAi response can be
broken down into three steps (Meister and Tuschl 2004; Tomari and Zamore 2005;
Siomi and Siomi 2009): (a) long dsRNA is processed by a ribonuclease (RNase) III
enzyme called Dicer into small interfering RNA (siRNA) duplexes; (b) these are
subsequently unwound, and one strand, the so-called guide strand (as compared
with the complementary passenger strand), is preferentially loaded onto the RNA-
induced silencing complex (RISC); (c) siRNAs serve as the sequence determinants
of the RNAi pathway by scanning the resident population of mRNAs (possibly any
single-stranded RNA) and directing multiple rounds of cleavage to homologous
mRNAs, via Slicer, a RISC component endonuclease (Fig. 1). Despite using
divergent proteins and mechanisms, organisms employ strikingly convergent stra-
tegies, which comprise a rather simple two-component system; small RNAs act as
specificity factors and Argonaute proteins act as effectors for repression. Depending
on both the nature of the Argonaute protein and the degree of complementarity
between the small RNA and the target sequence, association of RISC with target
RNAs can result in a range of functions. These findings together with structural and
biochemical studies culminated with the revelation that Argonaute is the RNA-
guided Mg2+-dependent RNA endonuclease of RISC that cleaves a single phos-
phodiester bond in the target, otherwise known as the Slicer (Meister et al. 2004;
Liu et al. 2004; Song et al. 2004). Once loaded with a small RNA, Argonaute
proteins exhibit a variety of functions for controlling protein synthesis and RNA
stability, maintaining genome integrity and mediating the production of a specific
set of small RNAs (Peters and Meister 2007; Hutvagner and Simard 2008).
Although the establishment of functional specificity for the different Argonaute
proteins remains to be understood, some of these specialized functions are likely to
be caused by structures and associated proteins specific to each Argonaute.
Emerging data also support Dicer as a RISC component, thus mechanistically
connecting the initiation phase with the effector phase of gene silencing. Refine-
ments of RNAi-related systems include built-in molecular rulers that define the size
of small RNAs, apparatuses that determine small RNA strand selection or polarity,
mechanisms that direct further amplification rounds using RNA-dependent RNA
polymerase (RdRP) activities on additional templates or by forming a cleavage-
mediated cycle, and safeguards for off-target silencing. The analysis of the biogen-
esis and targeting of small RNAs has also benefited from a classic genetic analysis
combined with a more recent approach that uses novel high-throughput cDNA
sequencing technologies with sophisticated bioinformatics tools (Mardis 2008).
The picture emerging from these analyses is of multiple variations on the core
RNA silencing mechanisms. Importantly, it is becoming clear that the activity of
RNA silencing pathways is subject to intense regulation at different levels, from
biogenesis of small RNAs to silencing modes of RISC.
RNAi is of great biomedical interest as it has potential for the study of gene
function, the validation of candidate drug targets, and even the treatment of disease
4 K. Saito et al.
a b
transposon locus flamenco
? Primary ?
Aub Piwi
Drosha/Pasha U U
Dicer2/R2D2
Dicer2/Loqs Aub Aub

AGO2 U U
Dicer1/Loqs C3PO A
transposon
mRNA
secondary
antisense (Ping-pong)
strand of
transposon
mRNA U
A A
AGO1 AGO2 AGO2 AGO3 AGO3
c d
21U-RNA primary siRNA/
Core sequence Argonaute complex
CTGTTTCA YRNT
A/T rich A/T rich Chr. IV
Spacer 3' 5'
Large motif Small motif
(~20nt ) 5' 3'
~34nt 4nt
5'
RdRP
P
PP
3'
U
5' 3'
SAGO1 SAGO2 CSR1

U PRG1
PPP PPP PPP
Fig. 1 Small RNA production and loading onto Argonautes. (a) In Drosophila, dsRNAs from an
endogenous or exogenous source are processed by RNase III domain nucleases, such as Drosha
and Dicer, into small RNA duplexes of about 21 nucleotides. Accessory proteins for RNase III
domain nucleases, such as Loqs and Pasha, help to distinguish the dsRNA precursors depending on
secondary structures or the degree of complementarity. Upon this association, one strand of the
small RNA duplex is selected and loaded onto one of the Argonautes. (b) In Drosophila gonad,
PIWI-subfamily proteins (Piwi, Aub and AGO3) associate with piRNAs and act in transposon
silencing. In the primary processing pathway, piRNAs are produced from single-stranded
precursors transcribed from the genome. In the secondary pathway (also known as the amplifica-
tion loop), piRNAs are produced by a Slicer-dependent mechanism. Aub and AGO3 generate the
50 end of piRNAs associating with AGO3 and Aub, respectively. The amplification loop is
independent of de-novo synthesis of piRNAs. (c) In C. elegans, the 21U-RNAs are predominantly
generated from chromosome IV and characterized by a specific upstream motif. Similar to
Drosophila piRNAs, 21U-RNAs are associated with a PIWI-subfamily protein, PRG-1, and are
produced in a Dicer-independent manner. (d) In C. elegans, Primary siRNAs associate with RDE-
1 and are then guided to the target RNA. Target RNAs are used as a template for producing
secondary siRNAs by RNA-directed RNA polymerase (RdRP). The secondary siRNAs can then
engage in another round of silencing
(Hannon 2002). Here, we review the biogenesis of small guide RNAs and discuss
our understanding of the molecular mechanisms of RNA silencing based on recent
insights into the regulatory circuitry of the silencing state. Although much of the
specific evidence discussed here refers to Drosophila and mammalian systems, in

which cell-free systems for analyzing RNAi are available, it also includes a broader
perspective from other systems.
2 RNA Silencing Effector as a Two-Component System
Since the discovery of the RNAi pathway, many endogenous small RNA pathways
that share many common features with the siRNA pathway have been identified.
These small RNAs function by degrading mRNAs and effect translational repression
of mRNAs or regulate gene silencing at the transcriptional level by heterochromatin
modification. A common feature of small RNAs is that they function as specificity
determinants for the repressive activities of Argonaute-containing effector com-
plexes. In principal, the RNA silencing system is an RNA-guided enzyme system
that requires only one (nonsequence-specific) protein for its enzymatic activity.
Sequence specificity is achieved by the small RNA component of the RNA-protein
(RNP) complex (H€ uttenhofer and Schattner 2006). In RNA silencing, one protein,
Argonaute, binds to many small guide RNAs that recognize their target by Watson–
Crick base pairing and thereby guide the Argonaute complex to different substrates.
It should be noted that although small guide RNAs might home in on homologous
DNA sequences, to date they have only been shown to target homologous RNAs.
The core protein component of all RISCs is a member of the Argonaute family of
small RNA-binding proteins (Tabara et al. 1999; Faehnle and Joshua-Tor 2007).
Members of this family are defined by the presence of PAZ and PIWI domains and
they normally consist of one variable N-terminal domain and conserved C-terminal
PAZ, MID and PIWI domains (Faehnle and Joshua-Tor 2007). The PAZ domain
recognizes the 30 end of small RNAs in a sequence-independent manner. The PIWI
domain adopts a folded structure similar to that of RNase H enzymes and exhibits
endonuclease or Slicer activity. Three residues (DDH) within the PIWI domain
form a catalytic triad. 50 monophosphate on the guide RNA strand binds to a
conserved binding pocket contained within a cleft bridging the MID and PIWI
domains. Argonaute proteins are divided into three phylogenetic groups: AGO,
from its founding member Arabidopsis Argonaute 1 (Ago1); PIWI, from Drosophila
Piwi (P-element induced wimpy testis); and WAGO/group III, Caenorhabditis
elegans-specific proteins. Many metazoan organisms have multiple Argonautes,
while single celled organisms like the fission yeast Schizosaccharomyces pombe
and the protozoan parasite Trypanosoma Brucei encode only one Argonaute
(Cerutti and Casas-Mollano 2006). This remarkable diversity of Argonautes raised
the possibility that different members of the family have become specialized in
each organism to perform distinct functions (see below).
A hallmark of RNA silencing is the production of short dsRNA molecules
(21–28 nt) by RNase III enzymes. Expansion of Dicer enzymes has also occurred:
Arabidopsis thaliana has four Dicer-like proteins and Drosophila melanogaster has
two Dicers, whereas mammals and yeast encoded a single Dicer. Despite high levels
of functional conservation, the complexity of the RNA silencing machinery varies
6 K. Saito et al.
greatly between different organisms (Cerutti and Casas-Mollano 2006; Kim et al.
2009; Siomi and Siomi 2009). It has been argued that the last common ancestor of
eukaryotes likely had at least one Argonaute, one Dicer, and one RNA-dependent
RNA polymerase (RdRP) (Cerutti and Casas-Mollano 2006), although interestingly,
among animals, only nematodes are reported to contain genes encoding RdRP. Two
main categories of small RNAs, siRNAs, and miRNAs have been defined, differing
in the nature of their precursors (Fig. 1a). Other small RNA species have been
recently identified, including piRNAs and endogenous siRNAs in flies and mam-
mals (Fig. 1a, b). The biogenesis of these small RNAs challenges the definition of
RNAi, since some do not appear to be produced “in response to dsRNA.”
3 Small RNA Biogenesis
3.1 miRNAs and siRNAs
miRNAs are small endogenous noncoding RNAs involved in posttranscriptional

gene regulation. In retrospect, the identification of lin-4 RNA in C. elegans in 1993
by Ambros was the harbinger of a diverse class of regulatory small RNAs of broad
importance (Lee et al. 1993). In 1999, Hamilton and Baulcombe published work
that links small RNAs of 21 nt to viral gene silencing in plants (Hamilton and
Baulcombe 1999) and then Ruvkun identified let-7, a second lin-4-like worm gene
(Reinhart et al. 2000). Although lin-4 was a worm-specific gene, an ingeniously
simple zoo blot by Ruvkun in 2000 revealed the phylogenetic conservation of let-7
RNA, which really opened up the field (Pasquinelli et al. 2000). miRNAs are 22 nt
RNA guides that control gene expression in both plants and animals (Kim 2005;
Bartel 2009). All known plant miRNAs and some mammalian miRNAs guide
cleavage of the mRNAs they regulate, whereas animal miRNAs typically reduce
the stability or repress translation of the mRNAs they regulate. Plant miRNAs
usually cleave in open reading frames, whereas the binding sites of animal miRNAs
are most often, but not always, located in 30 untranslated regions (UTRs). In animals,
miRNAs bind their target RNAs largely through a small region at the 50 -end of the
miRNA (positions 2 through 8), known as the “seed” (Brennecke et al. 2005; Lewis
et al. 2005). Consequently, a large fraction of the protein coding genes in Drosophila
and humans is predicted to be regulated by miRNAs (Brennecke et al. 2005;
Friedman et al. 2009). miRNA primary precursors (pri-miRNAs) are mostly tran-
scribed by RNA polymerase II. pri-miRNAs contain stem–loop structures that harbor
the miRNA in the 50 or 30 half of the stem. In plant miRNA production, a single
RNaseIII protein, Dicer-like protein 1 (DCL1), generates miRNA–miRNA*
duplexes in the nucleus (miRNA* is the sequence in the hairpin that pairs opposite
the miRNA, equivalent to the passenger stand of siRNA duplexes). In animals,
miRNAs are sequentially processed in the nucleus and cytoplasm by endoribonu-
cleases in partnership with dsRNA-binding proteins. The nuclear localized RNase
III, Drosha, first defines one end of the miRNA–miRNA* duplex and releases
approximately 65 nt pre-miRNAs (Fig. 1a). The pre-miRNA hairpin is then exported
to the cytoplasm, where the second RNase III, Dicer, completes the processing.
Drosha is part of a large complex, known as the “Microprocessor,” which acts like a
molecular ruler to determine the cleavage site (Han et al. 2006). In this complex,
Drosha interacts with its dsRNA-binding domain (dsRBD) cofactor protein, DGCR8
(also known as Pasha) (Denli et al. 2004; Gregory et al. 2004). A typical metazoan
pri-miRNA contains areas of local snapback structure that consists of a 33 bp stem, a
terminal loop and flanking segments and can be “cropped” by the Drosha complex.
DGCR8 prefers the junction between flexible single-stranded RNA and a double-
stranded stem; indeed, the flanking single-stranded RNA (ssRNA) segments are vital
for binding to DGCR8, and the 33 bp stem is also required for efficient binding.
Drosha may not be in direct contact with RNA at this stage. Drosha may interact
transiently with the stem of this “pre-cleavage” complex, where the processing center
of the enzyme, located about 11 bp from the ssRNA–dsRNA (SD) junction, makes a
staggered pair of breaks in the RNA to create the 65 nt long pre-miRNA. Thus,
DGCR8 may function as the molecular anchor that measures the distance from the
SD junction (Han et al. 2006). The Microprocessor could recognize the terminal loop
as ssRNA and bind to the stem–loop in the opposite orientation. In this case, abortive
cleavage can occur at an alternative site at 11 bp from the terminal loop. However,
most pri-miRNAs contain internal bulges or weakly paired bases 11 bp from the
terminal loop (Han et al. 2006). The computational search for miRNAs may be
facilitated by seeking sequences capable of folding into a structure predicted to be
bound by DGCR8 and to promote Drosha cleavage 11 bp from the junction of a
stem with single-stranded RNA tails.
siRNA and miRNA duplexes derive from the processing of longer duplexes
and pre-miRNA hairpins, respectively. This is performed by Dicer, which forms
21–25-nt duplexes possessing 50 -monophosphates, 30 -hydroxyl groups, and 2 nt 30
overhangs, which are classic hallmarks of RNase III enzymes. Dicer binds dsRNA
and generates RNA products of specific lengths. The prototypical Dicer contains,
from N to C terminus, a PAZ domain, two tandem RNase III domains, and a dsRNA-
binding domain (dsRBD). The distance between the PAZ domain that binds the 30
end of dsRNA and the two catalytic RNase III domains matches the length spanned
by 25 base pairs of RNA (MacRae et al. 2007). Thus Dicer itself is a molecular ruler.
The examination of sequencing data of small RNAs from D. melanogaster led to
the identification of clusters of small RNAs originating from the outer edges of an
annotated small intron (Okamura et al. 2007; Ruby et al. 2007). The 30 end of the
stem–loop precursor structure of these intronic small RNAs coincides with the 30
splice site, and is cleaved by nuclear pre-mRNA splicing rather than by Drosha.
After being debranched, these small intronic RNAs mimic the structural features of
pre-miRNA hairpins and enter the miRNA-processing pathway without Drosha-
mediated cleavage. These pre-miRNAs/introns were termed “mirtrons.”
The imprecision of Drosha or Dicer cleavage can result in the production of an
miRNA:miRNA* duplex with different 50 and 30 ends. This population of miRNA
variants is termed isomiRs. Most miRNAs in animals form imperfect hybrids with
8 K. Saito et al.
sequences in the target mRNA, with the miRNA 50 -proximal “seed” region providing
most of the pairing specificity. Therefore, the imprecise cleavage either alters the seed
sequence or inverts the relative stabilities of the 50 end of the duplex (see next
section). Recent deep-sequencing results of small RNAs reveal that human cells
may take advantage of such imprecise cleavage to create a diverse set of miRNAs
from a single precursor, which could broaden the reach of the miRNA regulatory
network (Landgraf et al. 2007; Azuma-Mukai et al. 2008; Morin et al. 2008).
3.2 Dicer-Independent Pathways
3.2.1 piRNAs
Piwi-interacting RNAs (piRNAs) have recently been discovered (Girard and

Hannon 2008; Siomi and Siomi 2008). They bind to the Piwi subfamily of proteins,
which are important for germline development and for the suppression of transpo-
son activity in the germline cells of mammals and flies. piRNAs carry a 50 mono-
phosphate group and exhibit a preference for a 50 uridine residue (Fig. 1b). Unlike
mammalian miRNAs, but similar to plant miRNAs, piRNAs carry a 20 -O-methyl
(20 -O-Me) modification at their 30 ends, which is appended by a Hen-1-like single-
stranded-specific methyltransferase (Horwich et al. 2007; Kirino and Mourelatos
2007; Ohara et al. 2007; Saito et al. 2007). Mutations of Dicer in flies and zebrafish
do not affect piRNA production (Vagin et al. 2006; Houwing et al. 2007). These
findings indicate that the biogenesis of piRNAs is distinct from that of miRNAs and
siRNAs, and does not involve dsRNA precursors. Sequencing of small RNAs
associated with fly Piwi proteins (Piwi, Aub, and AGO3) (Saito et al. 2006;
Brennecke et al. 2007; Gunawardane et al. 2007; Yin and Lin 2007) revealed that
piRNAs associated with Aub and Piwi are derived mainly from the antisense strand
of retrotransposons, while AGO3-associated piRNAs arise mainly from the sense
strand. Aub- and Piwi-associated piRNAs show a strong preference for uracil at
their 50 ends, and AGO3-associated piRNAs show a preference for adenine at
nucleotide 10, with no 50 nucleotide preference (Fig. 1b). Intriguingly, the 50 ends
of the AGO3-associated sense piRNAs frequently share a 10 bp complementarity
with the 50 end of Aub-associated antisense piRNAs. Piwi proteins retain the Slicer
activity that allows them to cleave an RNA substrate across from position 10 of
their bound piRNA (Saito et al. 2006; Gunawardane et al. 2007). These observa-
tions suggest a self-amplifying loop for piRNAs, where sense piRNAs in AGO3
cleave long antisense transcripts and guide the formation of the 50 -end of antisense
piRNA in Aub, and antisense piRNAs in Aub cleave long sense transcripts and
guide the formation of the 50 -end of sense piRNA in AGO3. In the amplification
loop, therefore, transposons are both a source gene for piRNAs and a target of
piRNA-mediated silencing. In other words, the Slicer activities of Piwi proteins
serve a dual role; to degrade sense transposon transcripts and to produce sense and
antisense piRNAs. However, how AGO3, Aub, and Piwi are able to specifically
recognize the respective sense and antisense strands of transposons remains to be
elucidated. In addition, since piRNA biogenesis is a Slicer-mediated process, it is

conceivable that cellular mRNAs containing a stretch of sequences that are suffi-
ciently complementary to a piRNA could also be targets of piRNA. Nonetheless,
piRNA biogenesis presents a new concept in which transcription occurs not simply
to express the gene but also to regulate it.
After loading the resulting cleavage products onto another member of the Piwi
proteins, a second nuclease activity generates the 30 end of the piRNA with the
specific size determined by the footprint of the particular Piwi family member on the
RNA. This cleavage appears to precede the 20 -O-Me modification by the Drosophila
homolog of Arabidopsis Hen1 (Horwich et al. 2007; Saito et al. 2007). In each Piwi
protein, the PAZ domain may be positioned from the MID domain at a distance that
corresponds to the length of each piRNA. Thus, the PAZ domain may act as part of a
molecular ruler for processing piRNAs of a defined size. The accumulation of
Drosophila piRNAs requires two putative piRNA maturation nucleases, Zucchini
and Squash (Pane et al. 2007). Other genes involved in piRNA biogenesis include
armitage, maelstrom, and spindle-E, all of which have been shown to be required for
the silencing of oskar mRNA translation and for organized microtubule formation
(Seto et al. 2007). How these microtubule elements regulate activities in the piRNA
pathway remains to be explored. Signatures of this amplification cycle, called the
Ping-Pong amplification loop, are also apparent in zebrafish and in mammalian
prepachytene piRNAs (Aravin et al. 2007; Houwing et al. 2007). Mammalian
piRNAs are strongly expressed in the male germline, and their total number per
cell in tissue obtained from testis reaches up to two million, i.e., about tenfold higher
than the miRNA content of these cells (Farazi et al. 2008). Gene knockout in mice of
any of the three testis-expressed Piwi proteins (Mili, Miwi and Miwi2) abolishes
spermatogenesis (Siomi and Kuramochi-Miyagawa 2009). A key difference
between transposon management in Drosophila and mammals is the role of cytosine
methylation in maintaining stable repression. Recent evidence indicates that mam-
malian PIWI–piRNA complexes may function in an RNA-dependent DNA methyl-
ation pathway (Aravin et al. 2007; Kuramochi-Miyagawa et al. 2008). This finding
also suggests that mammalian PIWI–piRNA complexes may interact with DNA
methyltransferases, probably in the nucleus.
The amplification loop model requires initiator piRNAs, which are products of
the biogenesis pathway that generates primary piRNAs. The piRNA precursor
appears to be a long, single-stranded transcript that is cleaved, preferentially at
U residues (Brennecke et al. 2007). The maternal loading of Piwi proteins into
embryos is observed in flies and fish (Harris and MacDonald 2001; Megosh et al.
2006; Houwing et al. 2007). In addition, when Drosophila strains that differ in the
presence of a particular transposon were crossed, daughters show a markedly
different content of piRNAs, depending on the origin of their parents. This indicates
that piRNAs are maternally inherited and can be seen as a genetic reservoir of
transposon resistance (Brennecke et al. 2008). Thus, the maternal contribution of
PIWI proteins and presumably their associated piRNAs may imply that primary
piRNAs that initiate an amplification cycle of piRNA biogenesis are supplied
through germline transmission. However, the amplification cycle in flies engages
10 K. Saito et al.
mainly AGO3 and Aub (Brennecke et al. 2007; Gunawardane et al. 2007), and Piwi
is spatially separated from them at subcellular and cell-type levels (Brennecke et al.
2007; Gunawardane et al. 2007; Nishida et al. 2007). In addition, piRNAs derived
from a particular piRNA cluster (flamenco locus) associate almost exclusively
with Piwi (Brennecke et al. 2007). These observations indicate that some other
mechanisms of piRNA biogenesis must operate. It remains unclear how primary
Piwi-associated antisense piRNAs are produced. Perhaps the most crucial question
is how transposon antisense transcripts, among the sea of the transcriptome, are
distinguished from other cellular transcripts and are specifically recognized by the
piRNA biogenesis machinery as source transcripts for primary piRNAs.
Recently, a new class of abundant 21 nt small RNAs was discovered in
C. elegans (Fig. 1c). They are characterized by a 50 -U bias (and are thus termed
21U-RNAs) and have a characteristic sequence motif about 42 bp upstream of the
start of the small RNA (Ruby et al. 2006). They originate in only a few clusters that
are specific to chromosome IV. 21U-RNAs are not produced by processing dsRNA
precursors but may be derived from thousands of individual, autonomously
expressed loci, broadly scattered in two large regions of chromosome IV; each
21U-RNA could be the product of an individual RNA polymerase transcription
event. 21U-RNAs are expressed in the germline and interact with the PIWI protein
PRG-1 (Batista et al. 2008; Das et al. 2008). They depend on PRG-1 activity for
their accumulation but are independent of Dicer for their production. Mutations in
prg-1 exhibit a reduced brood size and a temperature-sensitive sterile phenotype,
consistent with the notion that Piwi proteins are linked to germline maintenance.
Thus, 21U-RNAs are the piRNAs of C. elegans. Like the abundant pachytene
piRNAs found in mammals, 21U-RNAs encode remarkable sequence diversity
and yet lack obvious targets. In addition, none of the 15,000 individual
21U-RNA sequences are conserved between C. elegans and Caenorhabditis brigg-
sae. Although only one transposon (Tc3)-directed 21U-RNA, out of more than
15,000 different 21U-RNAs encoded in C. elegans, was identified, PRG-1/21U-
RNA complexes still suppress the transposon, probably by the production of
secondary siRNAs by RNA-dependent RNA polymerase (RdRP) activity (see
below). The sequence diversity and lack of obvious targets may suggest that
21U-RNAs act in cis to regulate their own genomic loci.
Scan RNAs (scnRNAs) in the ciliate Tetrahymena thermophila direct elimina-
tion of transposon-like DNA sequences and associate with a Piwi protein, Twi1.
Thus, they are the piRNAs in Tetrahymena. However, they are produced by the
Dicer-dependent pathway (Mochizuki and Gorovsky 2005). These examples indi-
cate that, during evolution, the core Piwi and piRNA machinery may have adopted
different strategies for the production and silencing of targets.
3.2.2 Secondary siRNAs in Worm
In C. elegans, RNAi deficiency resulting from the knockdown of WAGO subfamily

proteins could not be rescued by RDE-1, the Argonaute protein involved in classical
RNAi in worms. Moreover, an RDE-1 mutant, deficient in RNAi, could not be

rescued by WAGOs. These observations led to a model in which RNAi in worms
occurs in a two-step pathway (Yigit et al. 2006), where functionally and structurally
distinct Argonautes perform RNAi to direct gene silencing in a sequential manner.
Initially, a primary Argonaute [such as RDE-1 for exogenous siRNAs (exo-siRNAs)
and ERGO-1 for endogenous siRNAs (endo-siRNAs)] is guided by primary siRNAs
generated by Dicer processing of long dsRNAs. Then amplification of the silencing
signal occurs by the production of secondary siRNAs by RdRP, which are then
discriminately bound to secondary Argonaute proteins (SAGOs, the WAGO sub-
family proteins), which mediate downstream silencing (Fig. 1a). Overexpression of
SAGOs enhances RNAi, and inactivation of SAGOs decreases RNAi. Therefore,
they are probably limiting for RNAi in C. elegans. The primary function of worm
primary siRNAs may be to recruit an RdRP to the target RNA. In plants, the aberrant
or unwanted RNA is copied into double-stranded forms by RdRPs and becomes
a substrate for Dicer, which converts it into siRNA duplexes (Chapman and
Carrington 2007). However, the C. elegans somatic RdRP predominantly produces
21 nt single-stranded and 50 triphosphorylated small RNAs directly from the target
mRNA in a primer-independent manner, without the need for dicing of dsRNA (Pak
and Fire 2007; Sijen et al. 2007; Aoki et al. 2007). Recruitment of RdRP directly to
the target mRNA would permit new dsRNA synthesis without consuming the
original trigger-derived siRNAs, although it is not yet clear how the 30 end of
secondary siRNAs is formed nor is the molecular ruler that determines the size.
3.3 Endo-siRNAs
Traditionally, siRNAs are thought to originate from exogenous sources; infection

by viruses or experimental introduction by the bench scientist. However, deep-
sequencing of small RNAs from somatic tissues and from cultured fly cells has
identified a class of 30 -end methylated endogenous 21 nt RNAs derived from
transposons and several other loci, including cis-natural antisense transcript pairs,
and long stem–loop structures containing many mismatch pairs in their stems (Kim
et al. 2009; Siomi and Siomi 2009). In Drosophila, miRNAs are processed by
Dicer-1 and loaded onto AGO1, whereas exo-siRNAs are generated by Dicer-2 and
loaded onto AGO2 (Fig. 1a). Like exo-siRNAs, these endogenous small RNAs are
also produced by the Dicer-2-dependent pathway and loaded onto AGO2 and are,
therefore, termed endogenous siRNAs (Fig. 1a) (endo-siRNAs or esiRNAs). How-
ever, many endo-siRNAs depend on Loqs (Czech et al. 2008; Okamura et al. 2008),
the canonical partner of Dicer-1 in the miRNA pathway (Förstemann et al. 2005;
Saito et al. 2005), but not on R2D2, the partner of Dicer-2 (Liu et al. 2003). Flies
deficient in Dicer-2 or AGO2 show increased expression of transposons and,
therefore, endo-siRNAs may be the primary mechanism for silencing selfish genetic
elements in somatic cells, which lack the piRNA pathway. Hence, endo-siRNAs
and piRNAs are fundamentally similar, in that they defend against nucleic acid-
based parasites. This also indicates that Drosophila has two RNA silencing
12 K. Saito et al.
pathways that repress transposons. Endo-siRNAs are also present in mouse oocytes
and are derived from a variety of sources, including transposable elements (Tam
et al. 2008; Watanabe et al. 2008). However, a fraction of endo-siRNAs are
processed from overlapping regions of functional genes and from cognate pseudo-
genes suggesting that pseudogenes, previously thought to be nonfunctional protein
fossils, may actually regulate the expression of their founder gene.
Although siRNAs and miRNAs are distinguished from one another, not by their
size or function, but rather by their origin (Tomari and Zamore 2005), with
miRNAs being processed from stem–loop structures and siRNAs arising from the
cleavage of long dsRNA precursors, the discovery of endo-siRNAs makes this
distinction more difficult. This blurring of boundaries among different types of
small RNA has interesting evolutionary implications. Long stem–loop structures
for endo-siRNAs are reminiscent of plant miRNA precursors. One hypothesis for
the evolutionary origin of plant miRNAs is that new plant miRNA loci may evolve
from the inverted duplication of founder loci, producing a hairpin RNA (Chapman
and Carrington 2007). The transcribed hairpin RNAs would exhibit near perfect
self-complementarity and may be processed by DCL enzymes other than DCL1, the
primary plant miRNA processing enzyme, since DCL1 has limited activity on such
substrates. Subsequent acquisition of mutations by genetic drift would produce a
hairpin with imperfect complementarity and channel its processing to DCL1. Thus,
stem–loop structures for endo-siRNAs could be evolutionary intermediates in the
gradual transformation into miRNAs. An adaptive switch from Dicer2- to Dicer1-
mediated processing during the course of miRNA gene evolution might occur in
Drosophila, as suggested for miRNAs in plants.
4 RISC Loading, Sorting, and Target-Sensing of Small RNAs
siRNA duplexes are converted into a single-stranded form as they assemble onto
the RISC, where they provide the sequence specificity or guide for mRNA degra-
dation. Thus, the key steps in converting pre-RISC to mature RISC (holo-RISC) are
siRNA strand unwinding and preferential strand retention or strand selection. The
prevalent view of RISC loading is that thermodynamic asymmetry along the siRNA
or miRNA duplex determines which RNA strand will be retained as the “guide” and
which RNA strand will be discarded as the “passenger.” Specifically, the RNA
strand with its 50 end at the thermodynamically less stable end of the siRNA duplex
is preferentially loaded onto the RISC as the guide strand, a phenomenon referred to
as the asymmetry rule (Khvorova et al. 2003; Schwarz et al. 2003).
Considering the interactions between Dicer and Argonaute proteins (Tabara
et al. 2002), siRNA production and RISC assembly with siRNAs might be physi-
cally coupled. In Drosophila, Dicer-2 does not simply transfer siRNAs to a distinct
RISC but rather assembles onto the RISC along with the siRNAs, indicating that its
role extends beyond the initiation phase. Loading of siRNA duplexes onto AGO2 is
facilitated by the RISC-loading complex (RLC), which includes Dicer-2 and its
dsRBD partner protein, R2D2 (Liu et al. 2003). Which strand of the siRNA duplex
is loaded onto AGO2 appears to be determined by the orientation of the Dicer-2/

R2D2 heterodimer on the siRNA duplex (Tomari et al. 2004). R2D2 is thought to
sense the thermodynamic stability of the siRNA duplex and to bind to the more
stably paired end, whereas Dicer-2 is recruited to the less stable end. This complex
probably recruits AGO2 through an interaction between Dicer-2 and AGO2 and
helps to determine the orientation of AGO2 loading. The transition from a double-
to a single-stranded silencing trigger has been frequently attributed to an unidenti-
fied ATP-dependent RNA helicase. However, siRNA duplex unwinding and RISC
loading is facilitated by cleavage of the unincorporated “passenger” strand by
AGO2, a step which does not require ATP (Matranga et al. 2005; Miyoshi et al.
2005; Rand et al. 2005). Cleavage in the middle of the passenger strand would be
expected to reduce the annealing temperature and the free energy of duplex
formation that has to be invested by the helicase to separate the guide siRNA
strands. These data support a model where siRNAs are initially loaded as duplexes
onto an AGO2-containing pre-RISC. Competition among siRNA duplexes has been
observed resulting in significantly reduced RNAi efficacy, suggesting that there
is competition among siRNA duplexes for being loaded onto RISC (Castanotto
et al. 2007). This also suggests that AGO2 is limiting. In Drosophila, the cleaved
passenger strand is removed by C3PO (Fig. 1a) (component 3 promoter of RISC).
CP3O is an Mg2+-dependent endoribonuclease consisting of Translin and Trax and
is thought to be a key activator of the core RNAi machinery (Liu et al. 2009).
Conversion of pre-RISC to holo-RISC also occurs by a slicing-independent
mechanism. Three of the four Argonaute proteins in humans (hAgo1, hAgo3, and
hAgo4) lack endonuclease activity (though the catalytic residues are conserved in
these Ago proteins) but are nonetheless loaded with single-stranded guide siRNAs
(Meister et al. 2004; Liu et al. 2004; Azuma-Mukai et al. 2008). Similarly, single-
stranded miRNAs associate with human Ago2 (hAgo2) despite the expectation that
mismatches within the unwound precursor should block the passenger-strand
cleavage activity of hAgo2. Thus, alternative cleavage-independent mechanisms
for Ago loading and, therefore, RISC assembly exist. The identity of such an
“unwindase” remains unknown.
In humans, the pre-miRNA is known to bind the preformed complex of hAgo2,
Dicer, and its dsRBD partner proteins TRBP (Chendrimada et al. 2005) or PACT
(Lee et al. 2006). This preformed trimeric complex is capable of generating small
RNAs from dsRNA precursors, transferring one strand to Ago2, cleaving target
RNAs using pre-miRNA and distinguishing miRNA from miRNA*, none of which
require ATP hydrolysis (Gregory et al. 2005; Maniataki and Mourelatos 2005). This
suggests that Dicer-cleavage and sensing of thermodynamic stability occur in series
in the Ago2-Dicer-TRBP/PACT complex. The enhancement of mature miRNA
expression by ectopically expressed Ago proteins (all Ago proteins) has been
consistently observed, suggesting that Ago proteins can bind and stabilize mature
miRNAs and, thereby, increase their abundance at a posttranscriptional level.
An impact of Ago2 on mature miRNA expression is independent of its Slicer
activity. This also indicates that Agos are rate-limiting factors in miRNA proces-
sing (Diederichs and Haber 2007; O’Carroll et al. 2007).
14 K. Saito et al.
The emerging picture suggests that miRNAs and siRNAs are initially bound to
the Argonaute-containing complex as duplexes, and are subsequently unwound,
resulting in bound single-stranded RNAs. Subsequently, how is a small RNA
directed to a specific silencing pathway? Small RNAs destined for the different
silencing pathways are often generated by the same Dicer proteins; therefore, the
most obvious solution would be to have different Argonaute proteins for different
silencing pathways, dependent on the source of the dsRNA. siRNA production by
Dicer may be directly coupled to RISC assembly as described. According to this
view, Dicer may pass siRNAs directly to RISC, without siRNAs diffusing freely in
the cytoplasm after their production. This would also aid the discrimination of bona
fide siRNAs from various RNA degradation products within the cell. However,
analyses of how different small RNAs are channeled to different AGO proteins
shows multiple variations.
4.1 Sorting by Precursor Structures
In Drosophila, AGO1 and AGO2 are responsible for miRNA and siRNA pathways,
respectively (Okamura et al. 2004). Although such restriction of each class of small
RNA to a distinct Argonaute complex could occur because miRNAs and siRNAs
are produced by different Dicer pathways, Dicer1 for miRNAs and Dicer2 for
siRNAs, it appears that in flies small RNA production and small RNA loading
onto Argonaute protein complexes are separate steps. Small RNAs are loaded into
either AGO1 or AGO2 based on the structure of a small intermediate RNA duplex
(Tomari et al. 2007). If the duplex has a bulge in the middle (frequently observed in
miRNA precursors), the RNA is routed onto AGO1. If the duplex is perfectly
matched, the small RNA is channeled onto AGO2. This is because the Dicer-2/
R2D2 heterodimer binds well to highly paired small RNA duplexes but poorly to
duplexes bearing central mismatches. Thus, the Dicer-2/R2D2 heterodimer not only
determines the polarity of siRNA loading based on thermodynamic stability rules
but also acts as a gatekeeper for AGO2 complex assembly, promoting the incor-
poration of siRNAs over miRNAs. These observations suggest that each siRNA
duplex dissociates from the Dicer active site after it is produced and is subsequently
recaptured by the Dicer-2/R2D2 heterodimer; sorting can occur postdicing in
Drosophila, highlighting the importance of central mismatches in precursor
duplexes. Indeed, a single central bulge involving positions 7–11 was sufficient to
redirect an otherwise fully-paired siRNA duplex to AGO1 (Tomari et al. 2007).
Thus, a central unpaired region serves as both an antideterminant for the AGO2-
loading pathway and a preferred binding substrate for the AGO1 pathway.
Although AGO1 favors small RNA duplexes with central mismatches, the large
fraction of an miRNA/miRNA* duplex whose central region is base paired still
associates with the AGO1 RISC (Okamura et al. 2004; Kawamura et al. 2008),
suggesting that the AGO1-loading pathway is inherently selective and not a default
pathway for small RNAs rejected by the AGO2 pathway. The proteins facilitating
AGO1 loading remain to be identified. Recent studies in C. elegans also support this
“precursor structure” model for small RNA sorting (Jannot et al. 2008).
4.2 Sorting by the 50 Ends
The 50 nucleotide identity or phosphorylation status at the 50 end of small RNAs has
been shown to influence Argonaute protein associations. Arabidopsis miRNAs and
ta-siRNA begin with a 50 -terminal uridine residue and preferentially associate with
AGO1 (Mi et al. 2008). By contrast, AGO2 associates preferentially with small
RNAs containing 50 -terminal adenosines, whereas AGO5 prefers small RNAs with
50 -terminal cytosines. Interestingly, the opposite strands of miRNAs (miRNA*) with
50 adenosines and 50 cytosines are also bound to AGO2 and AGO5, respectively. These
findings have led to the hypothesis that the binding affinity of Argonaute proteins for
small RNAs is determined by the nucleotide at the 50 end. This implies that some
miRNA variants are loaded more efficiently than others, according to the identity
of their 50 nucleotide. Although these 50 -terminal nucleotide preferences generally
hold true for these Argonautes in plants, miR172 has a 50 -terminal adenosine, but
the majority of miR172 molecules associate with AGO1. Also, AGO7 preferentially
associates with miR390, which possesses a 50 -terminal adenosine (Montgomery et al.
2008). Therefore, it does not appear to be the sole determinant influencing Argonaute
association. In Arabidopsis, Dicer processing may be uncoupled from association with
Argonaute because miRNAs are generated by DCL1.
Secondary siRNAs in C. elegans are specifically loaded onto SAGOs (Yigit et al.
2006). Secondary siRNAs carry a 50 -triphosphate modification (Pak and Fire 2007;
Sijen et al. 2007), the hallmark of RdRP products (Fig. 1d). This might serve as a
recognition element for specific SAGO binding, while excluding binding by pri-
mary Argonautes, such as RDE-1. These findings imply that Argonaute diversifica-
tion is a consequence of which small RNA they recruit. Whether the protein
conformation dictates its small RNA partners awaits the determination of eukary-
otic Argonaute protein structures.
The 50 nucleotide was found to be inserted into a pocket in the MID domain
(Faehnle and Joshua-Tor 2007). It can be envisioned that the residues constituting
the 50 end binding pocket may differ between Argonaute proteins to accommodate
particular 50 terminal nucleotides. It will be interesting and important to solve the
structures of different Argonaute proteins in plants and animals.
4.3 Sorting by Dicer Processing Polarity
Endogenous siRNAs in C. elegans and transgene siRNAs in fission yeast display a

striking strand bias in which only the antisense siRNA strand, corresponding to the
RNA strand synthesized by RdRP, is loaded onto Argonaute complexes. Worm
16 K. Saito et al.
RdRPs produce small RNAs directly from the target mRNA, in a primer-independent
manner. Therefore, all secondary siRNAs are of negative polarity and serve to
reinforce silencing of the RNA target (Yigit et al. 2006; Pak and Fire 2007; Sijen
et al. 2007). In fission yeast, the physical association of Dicer with a RdRP
complex, termed RNA-dependent RNA polymerase complex (RDRC), and an
Argonaute complex, named RNA-induced transcriptional silencing complex
(RITS), may facilitate loading of siRNAs onto Argonaute in a directional manner
as Dicer moves along and cuts the dsRNA products of RdRP, giving rise to
antisense strand bias (B€ uhler and Moazed 2007). Thus, Dicer processing polarity
defines siRNA strand polarity. RITS and RDRC are proposed to act in a self-
reinforcing loop in which DNA-interacting proteins and the siRNA in RITS guide
H3K9 methylation and heterochromatin formation, and also the RdRP-mediated
conversion of transcripts into siRNA precursors (Moazed 2009). Thus, active
transcription may be a prerequisite for the assembly of heterochromatin in fission
yeast. Interestingly, all of these processing events appear to occur in the nucleus of
fission yeast. By contrast, nuclear functions for mammalian AGO and Dicer
proteins are still a matter of debate, although nuclear functions of siRNAs have
been reported in humans (Meister 2008).
4.4 Target-Sensing
Once loaded with single-strand guide small RNA, how does RISC find its target
RNA? It is clear that most of the binding energy that tethers RISC to a target RNA
comes from bases in the 50 half (the “seed” region) of the small RNA (Haley and
Zamore 2004), which has led to the hypothesis of the “two-state” model for RISC
target binding (Filipowicz 2005; Tomari and Zamore 2005). In this model, the 50
part of the small RNA within RISC has a favorable structure for base pairing,
whereas the arrangement of the 30 part antagonizes base pairing with the target
RNA. After a base-pairing between the seed and the target is formed, the confor-
mational change of an Argonaute occurs so that the 30 half of the small RNA can
base-pair with the target RNA. However, it should be noted that some miRNA
sequences, including let-7 and miR-34, are perfectly conserved across a vast
evolutionary distance. And about 10% of the miRNA families identified in inverte-
brates are completely conserved in mammals. Therefore, the principle that base-
pairing to the 50 seed part of the miRNA is a dominant factor in miRNA target
recognition alone does not account for the perfect conservation of the full sequences
of these miRNAs. It appears that accessibility of the target site can be sensed by an
intrinsic, nonspecific affinity of RISC for single-stranded RNA, which follows the
initial specific RISC-target association via the 50 “seed” part of the siRNA (Ameres
et al. 2007). On the other hand, the accessibility of the target site correlates directly
with the efficiency of cleavage, indicating that RISC is unable to unfold structured
RNA. Since mRNA targets exist in the cell as RNPs (Dreyfuss et al. 2002),
accessibility can, therefore, also be controlled by a number of RNA-binding
proteins that either mask the target site or facilitate unfolding. Binding site accessi-
bility provides an additional layer of regulatory control.
5 Safeguards for RNA Silencing Pathways
By analogy with classic immunity, there are also mechanisms in RNA silencing that
manage the population of effector molecules involved in surveillance, both by
subtracting out specific components that are not engaging targets and in some
cases by amplifying specific components that are engaging their targets. In RNA
silencing, one nonsequence-specific RNA-binding protein, Argonaute, binds many
different small guide RNAs, resulting in effector RISC complexes. Because target
recognition uses complementary RNA sequences, once a particular element or gene
is recognized by the RNAi system, all copies within a cell will be targets for
inactivation. This system, therefore, requires gatekeepers to ensure that Argonaute
can bind small guide RNAs but cannot degrade small RNAs.
Small RNAs produced by the RNase III enzymes, Drosha and Dicer, character-
istically leave 50 -monophosphates and 2 nt 30 overhangs on the processed product.
Therefore, the PAZ domain of Argonautes can initially distinguish these small
RNAs, by binding to their characteristic 30 -overhangs, from degraded RNAs that
are derived from nonrelated pathways. In addition, to become incorporated into
RISC and mediate target cleavage, the guide strand of siRNA needs to display a
phosphate group at the 50 end (Pham and Sontheimer 2005). In humans, the enzyme
capable of phosphorylating the 50 end of siRNAs is hClp1 (Weitzer and Martinez
2007), which also has roles in the splicing of transfer RNAs (tRNAs) and in the
formation of 30 ends of mRNAs, suggesting a functional link between these fun-
damental processes of RNA metabolism. Interestingly, both tRNA splicing and
mRNA 30 -end formation occur in the nucleus (Paushkin et al. 2004; Danckwardt
et al. 2008), suggesting that synthetic siRNA duplexes with a 50 OH group and
dephosphorylated siRNA duplexes are transported (or diffuse) into the nucleus, and
after phosphorylation by hClp1, they are exported to the cytoplasm and become
assembled onto RISC.
The production of dsRNA must be carefully controlled to prevent inappropriate
silencing. While amplification of the silencing signal would have obvious benefits
for suppressing the expression of repetitive elements and viruses, this should be
balanced against the danger of amplifying off-target silencing. Although the Slicer-
mediated ping-pong mechanism for piRNA production does not lead to “transitive”
RNAi, but rather to conservative amplification of functional primary piRNA
sequences, any off-target events mediated by RdRPs could, conceivably, lead to a
chain reaction or transitive effect of silencing with deleterious consequences. Thus,
safeguards must exist to prevent pervasive use of RdRPs. One such safeguard is
built into the pathway itself. In C. elegans, the processing of trigger dsRNA and
loading of the primary siRNAs onto the RDE-1 complex appear to be inherently
inefficient to limit the initial round of target recognition by RDE-1 and thus to
18 K. Saito et al.
minimize the risk of amplifying off-target silencing reactions (Yigit et al. 2006). In
addition, each secondary siRNA appears to be made by RdRP in a nonprocessive,
self-termination manner, which restricts transitive effects (Aoki et al. 2007; Pak and
Fire 2007; Sijen et al. 2007). Furthermore, secondary siRNAs associate with worm-
specific Argonautes (SAGOs), which lack catalytic residues for mRNA cleavage,
suggesting that they may be unable to generate cleaved substrates for further
amplification, thereby preventing them from inducing the exponential generation
of secondary siRNAs (Yigit et al. 2006; but see also Aoki et al. 2007). SAGOs are
also present in limited supply and thus provide limited capacity to support multiple
simultaneous silencing reactions.
mRNA quality control will also be important. For instance, in fission yeast, the
RNAi machinery is negatively regulated by the conserved siRNA nuclease, Eri1
(Iida et al. 2006), and links transgene silencing to a protein complex resembling the
Trf4-Air1/Air2-Mtr4 polyadenylation (TRAMP) complex of budding yeast, a
nuclear surveillance mechanism that degrades aberrant transcripts via the exosome
(B€uhler et al. 2007). Thus, RNAi in fission yeast is actively restricted from exerting
its effects throughout the genome and appears to be subject to competition from an
RNA quality control machinery.
6 Effector Modes of RNA Silencing Pathways
Small RNA-Argonaute or RISC complexes exhibit a variety of functions. But the

functions of RISC appear to be context-dependent; specific structures of small
RNA-binding sites for the target mRNA or on associated proteins specific to each
Argonaute influence effector modes. For example, animal miRNAs silence gene
expression by at least three independent mechanisms via binding sites mostly in the
30 UTR of target mRNAs; by cleaving mRNAs, by repressing translation and/or by
promoting mRNA degradation (Filipowicz et al. 2008). The molecular mechanism
of how miRNA–Argonaute complexes or miRISCs inhibit the translation of target
mRNAs is highly controversial. It is likely that miRNA-mediated translational
repression can occur by multiple mechanisms. A possible mechanism is repression
of translation initiation resulting from the simple competition for the cap structure
of mRNAs between miRNA-associated Ago protein and eIF4E Kiriakidou et al.
2007). Another mechanism is cap-independent translational repression mediated by
eIF6, a RISC component, which is known to prevent the productive assembly of
80S ribosomes (Chendrimada et al. 2007). Argonaute proteins, miRNAs, and their
target mRNAs accumulate in cytoplasmic foci, usually known as P-bodies, or
processing bodies (Eulalio et al. 2007a). miRNA-mediated repression and P-body
localization of the target mRNA are sometimes correlated and are potentially
reversible processes (Bhattacharyya et al. 2006). Translational repression is also
mediated by proteins that directly interact with Argonaute proteins, such as the P-
body component GW182 (Behm-Ansmant et al. 2006). miRNAs accelerate mRNA
decay by two distinct mechanisms. Those that are fully complementary to their
mRNA targets (or nearly so) direct endonucleolytic cleavage within the base-paired
region (Tomari and Zamore 2005). However, mRNA decay by partially comple-
mentary miRNAs in animal cells may not occur via endonucleolytic cleavage but
rather by removal of the 30 poly(A) tail from mRNAs (Bagga et al. 2005; Lim et al.
2005; Giraldez et al. 2006). Deadenylation and the consequent loss of poly(A)-
binding protein triggers 50 decapping, thereby exposing the message to the general
mRNA degradation machinery. miRNA-mediated mRNA decay requires an Argo-
naute protein, GW182, the CCR4–CAF1–NOT deadenylase complex, the decap-
ping enzyme DCP2, and several decapping activators (Behm-Ansmant et al. 2006;
Eulalio et al. 2007b). However, the contribution to gene silencing of translational
repression or mRNA degradation appears to differ for each miRNA-target pair and
is likely to depend on the particular set of proteins bound to the 30 UTR of the
mRNA or on proteins that bind to miRNA RISC (miRISC) complexes (Kedde et al.
2007). In some cases, RNA-binding proteins may physically block access to an
miRNA target site. In other cases, RNA-binding proteins may change the subcellu-
lar localization of an mRNA, taking it out of reach of miRNAs. mRNA structure
could also restrict miRISC accessibility to the miRNA target site. Thus, the final
outcome of miRNA regulation is influenced by other proteins interacting with the
target mRNA or with RISC, thereby counteracting the effects of the miRNA, or by
mRNA structure influencing miRISC accessibility, resulting in differential, tissue-
specific regulation (Bhattacharyya and Filipowicz 2007; Brodersen and Voinnet
2009). These findings also predict that if the miRNA target site is close to a binding
site for a strong translational repressor, RISC might compete with the translational
repressor for access to the mRNA, resulting in translational activation by miRNAs
(Brodersen and Voinnet 2009).
7 Regulations of RNA Silencing Pathways
Many plant and animal viruses encode suppressor proteins that block host RNA
silencing at various stages of the siRNA and miRNA pathways (Mlotshwa et al.
2008). Some virus suppressors sequester siRNA duplexes and prevent them from
entering RISC. Other suppressors inhibit DCLs or Argonaute activities. Cellular
proteins can also regulate the RNA silencing pathway.
7.1 Processing
Specific pri-miRNAs can be abundantly expressed in mammalian cells but not

processed into mature miRNAs until a developmental signal releases the processing
block (Obernosterer et al. 2006). Deregulation of miRNA expression has also been
observed in human malignancies (Thomson et al. 2006). The processing of let-7
miRNA, a tumor suppressor and cell-cycle regulator, is posttranscriptionally
20 K. Saito et al.
a b
Control of miRNA homeostatic control of
biogenesis miRNA biogenesis
Cleave DGCR8 mRNA
pri-miRNA
Drosha DGCR8
cap AAAA
pre-miRNA DGCR8 mRNA
export protein
stabilization
translation
Dicer control of
Lin28 recognizes miRISC activity
miRNA loop
containinig
GGAG Argonaute
Lin28 recruits cap ORF AAAA
Argonaute TUT4
UUUU Terminal
uridylation
Argonaute
cap ORF AAAA
TRIM-NHL
Degradation
miRNA biogenesis is
repressed in stem cells
Fig. 2 Control of the miRNA-mediated silencing pathway. (a) A subset of miRNA biogenesis is
controlled posttranscriptionally. Lin28, mainly expressed in undifferentiated stem cells, recog-
nizes the miRNA loop containing GGAG. Lin28 recruits TUT-4, a noncanonical poly (A)
polymerase, to pre-miRNAs, which adds a uridine tail to the 30 end of pre-miRNAs. Uridylation
prevents processing by Dicer, resulting in degradation of the miRNA. (b) DGCR8 mRNA is
cleaved and downregulated by the Drosha/DGCR8 complex. Drosha is stabilized by DGCR8 via
protein–protein interaction. This feedback circuit may help maintain the homeostatic control of
miRNA production. (c) TRIM-NHL proteins increase miRNA activity by modulating the interac-
tion of miRNP with downstream effectors. The molecular mechanisms for this enhancement are
largely unknown. The expression of TRIM-NHL is regulated during development and differentia-
tion, indicating that miRISC activity might be controlled during development
inhibited in embryonic cells by the pluripotency factor Lin-28, which appears to

block Microprocessor-mediated cleavage of pri-let-7 miRNAs and processing of
pre-let-7 miRNAs by Dicer (Fig. 2) (Rybak et al. 2008; Viswanathan et al. 2008). In
the latter case, Lin-28 induces uridylation of the let-7 precursor and recruits
TUTase4 (TUT4), a noncanonical poly (A) polymerase, to pre-let-7, which inhibits
Dicer processing (Heo et al. 2009). In contrast, TGF-b and BMP signaling promotes
a rapid increase in the expression of mature miR-21, an oncogenic miRNA, through
the promotion of pri-miR-21 processing into pre-miR-21 by Drosha (Davis et al.
2008). TGF-b- and BMP-specific SMAD signal transducers are recruited to
pri-miR-21 in a complex with the RNA helicase, p68, a component of the Micro-
processor complex, to facilitate pre-miRNA accumulation.
miRNA processing factors are also regulated. For example, Drosha and
DGCR8 regulate each other posttranscriptionally (Fig. 2) (Han et al. 2009). The
Drosha–DGCR8 complex cleaves the hairpin structures embedded in the DGCR8

mRNA, whose folds are similar to the pri-miRNA structure, and thereby destabi-
lizes the mRNA. DGCR8 stabilizes the Drosha protein via protein–protein inter-
action. These regulations constitute a feedback control system between Drosha and
DGCR8, which may help maintain the homeostatic control of miRNA biogenesis.
siRNA loading onto Argonaute proteins also appears to be subject to regulation.
In Drosophila S2 cells persistently infected by Flock House Virus (FHV), the
replication intermediate of the virus is substantially diced by Dicer-2 to produce
FHV siRNAs. However, bulk FHV-derived siRNAs are not loaded into any Argo-
naute proteins (Flynt et al. 2009). These findings suggest that activities of the RNAi
pathway during virus infection are regulated at the siRNA-loading step.
7.2 Modification
In human cells, Ago2 is hydroxylated at proline 700 by the alpha-(P4H-alpha(I)) and

beta-(P4H-beta) subunits of the type I collagen prolyl-4-hydroxylase (C-P4H(I)).
Depletion of P4H-alpha (I) or P4H-beta resulted in destabilization of Ago2,
showing the importance of hydroxylation as a posttranslational modification for
effective RNA interference. P4H-alpha(I) is regulated by several pathways or
stimuli at transcriptional and posttranscriptional levels, indicating that the RNAi
machinery is regulated by the hydroxylation of Ago proteins (Qi et al. 2008).
7.3 RISC Activity
As mentioned above, miRNA-mediated silencing is tuned by RNA-binding proteins

that are not directly part of the miRISC. Members of The TRIM-NHL family,
originally identified as ubiquitin ligases, enhance activities of a subset of miRNAs
without affecting levels of those miRNAs (Fig. 2) (Schwamborn et al. 2009;
Hammell et al. 2009). However, the precise mechanism of this action remains
unclear and an important challenge is to understand why TRIM-NHL proteins
regulate the activity of only a subset of miRNAs. The activity of the RISC can
also be regulated by target-mimic RNAs. Noncleavable plant miRNA targets
sequester miRNAs, thus regulating miRNA availability; the nonprotein coding
gene IPS1 (Induced by Phosphate Starvation 1) from A. thaliana contains a motif
with sequence complementarity to the phosphate (Pi) starvation-induced miRNA,
miR-399, but the pairing is interrupted by a mismatched loop at the expected
miRNA cleavage site (Franco-Zorrilla et al. 2007). IPS1 RNA is not cleaved but
instead sequesters miR-399. Thus, IPS1 overexpression results in increased accu-
mulation of the miR-399 target, PHO2 mRNA. The idea of target mimicry intro-
duces unanticipated complexity into the network of RNA regulatory interactions
and raises the possibility that a large number of mRNA-like noncoding RNAs,
22 K. Saito et al.
recently identified in humans (ENCODE Project Consortium 2007), could function

as attenuators of small RNA–Argonaute complexes.
8 Perspective
Biogenesis pathways of small guide RNAs and the ways in which small
RNA–Argonaute complexes regulate gene expression are surprisingly diverse.
The future challenges are obvious. How many new classes of silencing processes
remain to be discovered? How are all of these pathways regulated? Does the
number of Argonaute family genes among species set an upper limit on the number
of classes of small regulatory RNAs that remain to be identified and on the number
of small RNA-guided regulatory processes? Given that many classes of small
RNAs are modified at their ends (Farazi et al. 2008), it is still uncertain whether
current cloning technologies (including ligation of linkers to 30 and 50 ends with
potential PCR amplification bias) are really sampling the entire spectrum of small
RNAs in cells. As the power of our observational tools, such as novel deep-
sequencing technology, increases (Mardis 2008), we need to develop new cloning
methods that are insensitive to specific modifications of small RNAs to reveal the
entire spectrum of small RNA molecules in cells.
It is important to note that small guide RNAs, their precursors, and their target
RNAs are not naked, but rather they are mostly associated with multiple proteins
that regulate many aspects of gene expression. A major challenge for future studies
will be to identify how specific RNA-binding proteins influence the final outcome
of small RNA regulation. For example, genome-wide in vivo approaches, with a
combination of immunoprecipitation methodologies and high-throughput sequenc-
ing, will be required to establish protein–RNA interactions or RNP occupancy at
certain regions of RNA transcripts in vivo (Chi et al. 2009). It is also important to
identify cellular activators and repressors of RNA silencing pathways. Identifica-
tion of such modifiers of the pathways will, in turn, help to identify compounds that
regulate RNAi for potential therapeutic use.
Finally, it is becoming apparent that changes in the activity of RNA silencing
pathways could create quantitative genetic variation in gene expression. Have such
changes contributed to human biology, such as the domestication of Homo sapiens?
There is relatively little correlation between morphological complexity and the
number and diversity of protein-coding genes. However, the number of miRNAs
correlates well with an organism’s total number of neurons (Grimson et al. 2008).
This suggests that miRNA diversity in higher eukaryotes may correlate with an
increase in their relative morphological complexity. Since miRNA seed matches
are often necessary and sufficient for target regulation, a single mutation in a
cellular gene or in the seed sequence of an miRNA gene is presumably enough to
inactivate an miRNA target sequence or to create a new miRNA target site and
might be a rapid way for evolution to fine tune gene expression. Polymorphic
sequence variants in human populations that create or destroy miRNA-binding sites

may also have significant effects on phenotype variation.
Acknowledgments We apologize to all colleagues whose relevant primary publications were not
included because of space limitations and the focus on recent results. The authors thank all
members of the Siomi laboratory for their comments and critical reading of the manuscript.
References
Ameres SL, Martinez J, Schroeder R (2007) Molecular basis for target RNA recognition and
cleavage by human RISC. Cell 130:101–112
Aoki K, Moriguchi H, Yoshioka T et al (2007) In vitro analyses of the production and activity of
secondary small interfering RNAs in C. elegans. EMBO J 26:5007–5019
Aravin AA, Sachidanandam R, Girard A et al (2007) Developmentally regulated piRNA clusters
implicate MILI in transposon control. Science 316:744–747
Azuma-Mukai A, Oguri H, Mituyama T et al (2008) Characterization of endogenous human
Argonautes and their miRNA partners in RNA silencing. Proc Natl Acad Sci USA
105:7964–7969
Bagga S, Bracht J, Hunter S et al (2005) Regulation by let-7 and lin-4 miRNAs results in target
mRNA degradation. Cell 122:553–563
Bartel DP (2009) MicroRNAs: target recognition and regulatory functions. Cell 136:215–233
Batista PJ, Ruby JG, Claycomb JM et al (2008) PRG-1 and 21U-RNAs interact to form the piRNA
complex required for fertility in C. elegans. Mol Cell 31:67–78
Behm-Ansmant I, Rehwinkel J, Doerks T et al (2006) mRNA degradation by miRNAs and GW182
requires both CCR4:NOT deadenylase and DCP1:DCP2 decapping complexes. Genes Dev
20:1885–1898
Bhattacharyya SN, Filipowicz W (2007) Argonautes and company: sailing against the wind. Cell
128:1027–1028
Bhattacharyya SN, Habermacher R, Martine U et al (2006) Relief of microRNA-mediated
translational repression in human cells subjected to stress. Cell 125:1111–1124
Brennecke J, Stark A, Russell RB et al (2005) Principles of microRNA-target recognition. PLoS
Biol 3:e85
Brennecke J, Aravin AA, Stark A et al (2007) Discrete small RNA-generating loci as master
regulators of transposon activity in Drosophila. Cell 128:1089–1103
Brennecke J, Malone CD, Aravin AA et al (2008) An epigenetic role for maternally inherited
piRNAs in transposon silencing. Science 322:1387–1392
Brodersen P, Voinnet O (2009) Revisiting the principles of microRNA target recognition and
mode of action. Nat Rev Mol Cell Biol 10:141–148
B€uhler M, Moazed D (2007) Transcription and RNAi in heterochromatic gene silencing. Nat
Struct Mol Biol 14:1041–1048
B€uhler M, Haas W, Gygi SP et al (2007) RNAi-dependent and -independent RNA turnover
mechanisms contribute to heterochromatic gene silencing. Cell 129:707–721
Castanotto D, Sakurai K, Lingeman R et al (2007) Combinatorial delivery of small interfering
RNAs reduces RNAi efficacy by selective incorporation into RISC. Nucleic Acids Res
35:5154–5164
Cerutti H, Casas-Mollano JA (2006) On the origin and functions of RNA-mediated silencing: from
protists to man. Curr Genet 50:81–99
Chapman EJ, Carrington JC (2007) Specialization and evolution of endogenous small RNA
pathways. Nat Rev Genet 8:884–896
24 K. Saito et al.
Chendrimada TP, Gregory RI, Kumaraswamy E et al (2005) TRBP recruits the Dicer complex to
Ago2 for microRNA processing and gene silencing. Nature 436:740–744
Chendrimada TP, Finn KJ, Ji X et al (2007) MicroRNA silencing through RISC recruitment of
eIF6. Nature 447:823–828
Chi SW, Zang JB, Mele A et al (2009) Argonaute HITS-CLIP decodes microRNA–mRNA
interaction maps. Nature 460:479–486
Czech B, Malone CD, Zhou R et al (2008) An endogenous small interfering RNA pathway in
Drosophila. Nature 453:798–802
Danckwardt S, Hentze MW, Kulozik AE (2008) 30 end mRNA processing: molecular mechanisms
and implications for health and disease. EMBO J 27:482–498
Das PP, Bagijn MP, Goldstein LD et al (2008) Piwi and piRNAs act upstream of an endogenous
siRNA pathway to suppress Tc3 transposon mobility in the Caenorhabditis elegans germline.
Mol Cell 31:79–90
Davis BN, Hilyard AC, Lagna G et al (2008) SMAD proteins control DROSHA-mediated
microRNA maturation. Nature 454:56–61
Denli AM, Tops BB, Plasterk RH et al (2004) Processing of primary microRNAs by the
Microprocessor complex. Nature 432:231–235
Diederichs S, Haber DA (2007) Dual role for argonautes in microRNA processing and posttran-
scriptional regulation of microRNA expression. Cell 131:1097–1108
Ding SW, Voinnet O (2007) Antiviral immunity directed by small RNAs. Cell 130:413–426
Dreyfuss G, Kim VN, Kataoka N (2002) Messenger-RNA-binding proteins and the messages they
carry. Nat Rev Mol Cell Biol 3:195–205
ENCODE Project Consortium (2007) Identification and analysis of functional elements in 1% of
the human genome by the ENCODE pilot project. Nature 447:799–816
Eulalio A, Behm-Ansmant I, Izaurralde E (2007a) P bodies: at the crossroads of post-
transcriptional pathways. Nat Rev Mol Cell Biol 8:9–22
Eulalio A, Rehwinkel J, Stricker M et al (2007b) Target-specific requirements for enhancers of
decapping in miRNA-mediated gene silencing. Genes Dev 21:2558–2570
Faehnle CR, Joshua-Tor L (2007) Argonautes confront new small RNAs. Curr Opin Chem Biol
11:569–577
Farazi TA, Juranek SA, Tuschl T (2008) The growing catalog of small RNAs and their association
with distinct Argonaute/Piwi family members. Development 135:1201–1214
Filipowicz W (2005) RNAi: the nuts and bolts of the RISC machine. Cell 122:17–20
Filipowicz W, Bhattacharyya SN, Sonenberg N (2008) Mechanisms of post-transcriptional
regulation by microRNAs: are the answers in sight? Nat Rev Genet 9:102–114
Fire A, Xu S, Montgomery MK et al (1998) Potent and specific genetic interference by double-
stranded RNA in Caenorhabditis elegans. Nature 391:806–811
Flynt A, Liu N, Martin R et al (2009) Dicing of viral replication intermediates during silencing of
latent Drosophila viruses. Proc Natl Acad Sci USA 106:5270–5275
Förstemann K, Tomari Y, Du T et al (2005) Normal microRNA maturation and germ-line stem cell
maintenance requires Loquacious, a double-stranded RNA-binding domain protein. PLoS Biol
3:e236
Franco-Zorrilla JM, Valli A, Todesco M et al (2007) Target mimicry provides a new mechanism
for regulation of microRNA activity. Nat Genet 39:1033–1037
Friedman RC, Farh KK, Burge CB et al (2009) Most mammalian mRNAs are conserved targets of
microRNAs. Genome Res 19:92–105
Giraldez AJ, Mishima Y, Rihel J et al (2006) Zebrafish MiR-430 promotes deadenylation and
clearance of maternal mRNAs. Science 312:75–79
Girard A, Hannon GJ (2008) Conserved themes in small-RNA-mediated transposon control.
Trends Cell Biol 18:136–148
Gregory RI, Yan KP, Amuthan G et al (2004) The Microprocessor complex mediates the genesis
of microRNAs. Nature 432:235–240
Gregory RI, Chendrimada TP, Cooch N et al (2005) Human RISC couples microRNA biogenesis
and posttranscriptional gene silencing. Cell 123:631–640
Grimson A, Srivastava M, Fahey B et al (2008) Early origins and evolution of microRNAs and
Piwi-interacting RNAs in animals. Nature 455:1193–1197
Gunawardane LS, Saito K, Nishida KM et al (2007) A slicer-mediated mechanism for repeat-
associated siRNA 50 end formation in Drosophila. Science 315:1587–1590
Haley B, Zamore PD (2004) Kinetic analysis of the RNAi enzyme complex. Nat Struct Mol Biol
11:599–606
Hamilton AJ, Baulcombe DC (1999) A species of small antisense RNA in posttranscriptional gene
silencing in plants. Science 286:950–952
Hammell CM, Lubin I, Boag PR et al (2009) nhl-2 Modulates microRNA activity in Caenorhab-
ditis elegans. Cell 136:926–938
Han J, Lee Y, Yeom KH et al (2006) Molecular basis for the recognition of primary microRNAs by
the Drosha-DGCR8 complex. Cell 125:887–901
Han J, Pedersen JS, Kwon SC et al (2009) Posttranscriptional crossregulation between Drosha and
DGCR8. Cell 136:75–84
Hannon GJ (2002) RNA interference. Nature 418:244–251
Harris AN, MacDonald PM (2001) Aubergine encodes a Drosophila polar granule component
required for pole cell formation and related to eIF2C. Development 128:2823–2832
Heo I, Joo C, Kim YK et al (2009) TUT4 in concert with Lin28 suppresses microRNA biogenesis
through pre-microRNA uridylation. Cell 138:696–708
Horwich MD, Li C, Matranga C et al (2007) The Drosophila RNA methyltransferase, DmHen1,
modifies germline piRNAs and single-stranded siRNAs in RISC. Curr Biol 17:1265–1272
Houwing S, Kamminga LM, Berezikov E et al (2007) A role for Piwi and piRNAs in germ cell
maintenance and transposon silencing in Zebrafish. Cell 129:69–82
H€uttenhofer A, Schattner P (2006) The principles of guiding by RNA: chimeric RNA-protein
enzymes. Nat Rev Genet 7:475–482
Hutvagner G, Simard MJ (2008) Argonaute proteins: key players in RNA silencing. Nat Rev Mol
Cell Biol 9:22–32
Iida T, Kawaguchi R, Nakayama J (2006) Conserved ribonuclease, Eri1, negatively regulates
heterochromatin assembly in fission yeast. Curr Biol 16:1459–1464
Jannot G, Boisvert ME, Banville IH et al (2008) Two molecular features contribute to the
Argonaute specificity for the microRNA and RNAi pathways in C. elegans. RNA 14:829–835
Kawamura Y, Saito K, Kin T et al (2008) Drosophila endogenous small RNAs bind to Argonaute 2
in somatic cells. Nature 453:793–797
Kedde M, Strasser MJ, Boldajipour B et al (2007) RNA-binding protein Dnd1 inhibits microRNA
access to target mRNA. Cell 131:1273–1286
Khvorova A, Reynolds A, Jayasena SD (2003) Functional siRNAs and miRNAs exhibit strand
bias. Cell 115:209–216
Kim VN (2005) MicroRNA biogenesis: coordinated cropping and dicing. Nat Rev Mol Cell Biol
6:376–385
Kim VN, Han J, Siomi MC (2009) Biogenesis of small RNAs in animals. Nat Rev Mol Cell Biol
10:126–139
Kiriakidou M, Tan GS, Lamprinaki S et al (2007) An mRNA m7G cap binding-like motif within
human Ago2 represses translation. Cell 129:1141–1151
Kirino Y, Mourelatos Z (2007) Mouse Piwi-interacting RNAs are 20 -O-methylated at their 30
termini. Nat Struct Mol Biol 14:347–348
Kuramochi-Miyagawa S, Watanabe T, Gotoh K et al (2008) DNA methylation of retrotransposon
genes is regulated by Piwi family members MILI and MIWI2 in murine fetal testes. Genes Dev
22:908–917
Landgraf P, Rusu M, Sheridan R et al (2007) A mammalian microRNA expression atlas based on
small RNA library sequencing. Cell 129:1401–1414
26 K. Saito et al.
Lee RC, Feinbaum RL, Ambros V (1993) The C. elegans heterochronic gene lin-4 encodes small
RNAs with antisense complementarity to lin-14. Cell 75:843–854
Lee Y, Hur I, Park SY et al (2006) The role of PACT in the RNA silencing pathway. EMBO J
25:522–532
Lewis BP, Burge CB, Bartel DP (2005) Conserved seed pairing, often flanked by adenosines,
indicates that thousands of human genes are microRNA targets. Cell 120:15–20
Lim LP, Lau NC, Garrett-Engele P et al (2005) Microarray analysis shows that some microRNAs
downregulate large numbers of target mRNAs. Nature 433:769–773
Liu Q, Rand TA, Kalidas S et al (2003) R2D2, a bridge between the initiation and effector steps of
the Drosophila RNAi pathway. Science 301:1921–1925
Liu J, Carmell MA, Rivas FV et al (2004) Argonaute2 is the catalytic engine of mammalian RNAi.
Science 305:1437–1441
Liu Y, Ye X, Jiang F et al (2009) C3PO, an endoribonuclease that promotes RNAi by facilitating
RISC activation. Science 325:750–753
MacRae IJ, Zhou K, Doudna JA (2007) Structural determinants of RNA recognition and cleavage
by Dicer. Nat Struct Mol Biol 14:934–940
Maniataki E, Mourelatos Z (2005) A human, ATP-independent, RISC assembly machine fueled by
pre-miRNA. Genes Dev 19:2979–2990
Mardis ER (2008) The impact of next-generation sequencing technology on genetics. Trends
Genet 24:133–141
Matranga C, Tomari Y, Shin C et al (2005) Passenger-strand cleavage facilitates assembly of
siRNA into Ago2-containing RNAi enzyme complexes. Cell 123:607–620
Megosh HB, Cox DN, Campbell C et al (2006) The role of PIWI and the miRNA machinery in
Drosophila germline determination. Curr Biol 16:1884–1894
Meister G (2008) Molecular biology. RNA interference in the nucleus. Science 321:496–497
Meister G, Tuschl T (2004) Mechanisms of gene silencing by double-stranded RNA. Nature
431:343–349
Meister G, Landthaler M, Patkaniowska A et al (2004) Human Argonaute2 mediates RNA
cleavage targeted by miRNAs and siRNAs. Mol Cell 15:185–197
Mi S, Cai T, Hu Y et al (2008) Sorting of small RNAs into Arabidopsis argonaute complexes is
directed by the 50 terminal nucleotide. Cell 133:116–127
Miyoshi K, Tsukumo H, Nagami T et al (2005) Slicer function of Drosophila Argonautes and its
involvement in RISC formation. Genes Dev 19:2837–2848
Mlotshwa S, Pruss GJ, Vance V (2008) Small RNAs in viral infection and host defense. Trends
Plant Sci 13:375–382
Moazed D (2009) Small RNAs in transcriptional gene silencing and genome defence. Nature
457:413–420
Mochizuki K, Gorovsky MA (2005) A Dicer-like protein in Tetrahymena has distinct functions in
genome rearrangement, chromosome segregation, and meiotic prophase. Genes Dev 19:77–89
Montgomery TA, Howell MD, Cuperus JT et al (2008) Specificity of ARGONAUTE7-miR390
interaction and dual functionality in TAS3 trans-acting siRNA formation. Cell 133:128–141
Morin RD, O’Connor MD, Griffith M et al (2008) Application of massively parallel sequencing to
microRNA profiling and discovery in human embryonic stem cells. Genome Res 18:610–621
Nishida KM, Saito K, Mori T et al (2007) Gene silencing mechanisms mediated by Aubergine
piRNA complexes in Drosophila male gonad. RNA 13:1911–1922
Novina CD, Sharp PA (2004) The RNAi revolution. Nature 430:161–164
Obernosterer G, Leuschner PJ, Alenius M et al (2006) Post-transcriptional regulation of micro-
RNA expression. RNA 12:1161–1167
O’Carroll D, Mecklenbrauker I, Das PP et al (2007) A Slicer-independent role for Argonaute 2 in
hematopoiesis and the microRNA pathway. Genes Dev 21:1999–2004
Ohara T, Sakaguchi Y, Suzuki T et al (2007) The 30 termini of mouse Piwi-interacting RNAs are
20 -O-methylated. Nat Struct Mol Biol 14:349–350
Okamura K, Ishizuka A, Siomi H et al (2004) Distinct roles for Argonaute proteins in small RNA-
directed RNA cleavage pathways. Genes Dev 18:1655–1666
Okamura K, Hagen JW, Duan H et al (2007) The mirtron pathway generates microRNA-class
regulatory RNAs in Drosophila. Cell 130:89–100
Okamura K, Chung WJ, Ruby JG et al (2008) The Drosophila hairpin RNA pathway generates
endogenous short interfering RNAs. Nature 453:803–806
Pak J, Fire A (2007) Distinct populations of primary and secondary effectors during RNAi in
C. elegans. Science 315:241–244
Pane A, Wehr K, Sch€ upbach T (2007) Zucchini and squash encode two putative nucleases required
for rasiRNA production in the Drosophila germline. Dev Cell 12:851–862
Pasquinelli AE, Reinhart BJ, Slack F et al (2000) Conservation of the sequence and temporal
expression of let-7 heterochronic regulatory RNA. Nature 408:86–89
Paushkin SV, Patel M, Furia BS et al (2004) Identification of a human endonuclease complex
reveals a link between tRNA splicing and pre-mRNA 30 end formation. Cell 117:311–321
Peters L, Meister G (2007) Argonaute proteins: mediators of RNA silencing. Mol Cell 26:611–623
Pham JW, Sontheimer EJ (2005) Molecular requirements for RNA-induced silencing complex
assembly in the Drosophila RNA interference pathway. J Biol Chem 280:39278–39283
Qi HH, Ongusaha PP, Myllyharju J et al (2008) Prolyl 4-hydroxylation regulates Argonaute 2
stability. Nature 455:421–424
Rand TA, Petersen S, Du F et al (2005) Argonaute2 cleaves the anti-guide strand of siRNA during
RISC activation. Cell 123:621–629
Reinhart BJ, Slack FJ, Basson M et al (2000) (2000) The 21-nucleotide let-7 RNA regulates
developmental timing in Caenorhabditis elegans. Nature 403:901–906
Rocheleau CE, Downs WD, Lin R et al (1997) Wnt signaling and an APC-related gene specify
endoderm in early C. elegans embryos. Cell 90:707–716
Ruby JG, Jan C, Player C et al (2006) Large-scale sequencing reveals 21U-RNAs and additional
microRNAs and endogenous siRNAs in C. elegans. Cell 127:1193–1207
Ruby JG, Jan CH, Bartel DP (2007) Intronic microRNA precursors that bypass Drosha processing.
Nature 448:83–86
Rybak A, Fuchs H, Smirnova L et al (2008) A feedback loop comprising lin-28 and let-7 controls
pre-let-7 maturation during neural stem-cell commitment. Nat Cell Biol 10:987–993
Saito K, Ishizuka A, Siomi H et al (2005) Processing of pre-microRNAs by the Dicer-1-
Loquacious complex in Drosophila cells. PLoS Biol 3:e235
Saito K, Nishida KM, Mori T et al (2006) Specific association of Piwi with rasiRNAs derived from
retrotransposon and heterochromatic regions in the Drosophila genome. Genes Dev
20:2214–2222
Saito K, Sakaguchi Y, Suzuki T et al (2007) Pimet, the Drosophila homolog of HEN1, mediates
20 -O-methylation of Piwi- interacting RNAs at their 30 ends. Genes Dev 21:1603–1608
Schwamborn JC, Berezikov E, Knoblich JA (2009) The TRIM-NHL protein TRIM32 activates
microRNAs and prevents self-renewal in mouse neural progenitors. Cell 136:913–925
Schwarz DS, Hutvágner G, Du T et al (2003) Asymmetry in the assembly of the RNAi enzyme
complex. Cell 115:199–208
Seto AG, Kingston RE, Lau NC (2007) The coming of age for Piwi proteins. Mol Cell 26:603–609
Sijen T, Steiner FA, Thijssen KL et al (2007) Secondary siRNAs result from unprimed RNA
synthesis and form a distinct class. Science 315:244–247
Siomi MC, Kuramochi-Miyagawa S (2009) RNA silencing in germlines–exquisite collaboration
of Argonaute proteins with small RNAs for germline survival. Curr Opin Cell Biol 21:426–434
Siomi H, Siomi MC (2008) Interactions between transposable elements and Argonautes have
(probably) been shaping the Drosophila genome throughout evolution. Curr Opin Genet Dev
18:181–187
Siomi H, Siomi MC (2009) On the road to reading the RNA-interference code. Nature
457:396–404
28 K. Saito et al.
Song JJ, Smith SK, Hannon GJ, Joshua-Tor L (2004) Crystal structure of Argonaute and its
implications for RISC slicer activity. Science 305:1434–1437
Stefani G, Slack FJ (2008) Small non-coding RNAs in animal development. Nat Rev Mol Cell Biol
9:219–230
Tabara H, Sarkissian M, Kelly WG et al (1999) The rde-1 gene, RNA interference, and transposon
silencing in C. elegans. Cell 99:123–132
Tabara H, Yigit E, Siomi H et al (2002) The dsRNA binding protein RDE-4 interacts with RDE-1,
DCR-1, and a DExH-box helicase to direct RNAi in C. elegans. Cell 109:861–871
Tam OH, Aravin AA, Stein P et al (2008) Pseudogene-derived small interfering RNAs regulate
gene expression in mouse oocytes. Nature 453:534–538
Thomson JM, Newman M, Parker JS et al (2006) Extensive post-transcriptional regulation of
microRNAs and its implications for cancer. Genes Dev 20:2202–2207
Tomari Y, Zamore PD (2005) Perspective: machines for RNAi. Genes Dev 19:517–529
Tomari Y, Matranga C, Haley B et al (2004) A protein sensor for siRNA asymmetry. Science
306:1377–1380
Tomari Y, Du T, Zamore PD (2007) Sorting of Drosophila small silencing RNAs. Cell
130:299–308
Vagin VV, Sigova A, Li C et al (2006) A distinct small RNA pathway silences selfish genetic
elements in the germline. Science 313:320–324
Viswanathan SR, Daley GQ, Gregory RI (2008) Selective blockade of microRNA processing by
Lin28. Science 320:97–100
Watanabe T, Totoki Y, Toyoda A et al (2008) Endogenous siRNAs from naturally formed dsRNAs
regulate transcripts in mouse oocytes. Nature 453:539–543
Weitzer S, Martinez J (2007) The human RNA kinase hClp1 is active on 30 transfer RNA exons
and short interfering RNAs. Nature 447:222–226
Yigit E, Batista PJ, Bei Y et al (2006) Analysis of the C. elegans Argonaute family reveals that
distinct Argonautes act sequentially during RNAi. Cell 127:747–757
Yin H, Lin H (2007) An epigenetic activation role of Piwi and a Piwi-associated piRNA in
Drosophila melanogaster. Nature 450:304–308
Selected Strategies for the Delivery of siRNA
In Vitro and In Vivo
Sandra D. Laufer, Anke Detzer, Georg Sczakiel, and Tobias Restle
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2 Mechanism of RNA Interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3 Naked Delivery of siRNA In Vitro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.1 Cellular Uptake of Naked Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2 The Phosphorothioate-Stimulated Cellular Delivery of siRNA . . . . . . . . . . . . . . . . . . . . . . 34
3.3 The siRNA-Peptide Conjugate Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.4 Intracellular Release of siRNA: A Major Hurdle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4 CPP-Mediated siRNA Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.1 Cell-Penetrating Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.2 Selected Examples of CPP-Mediated siRNA Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5 Selected Examples of siRNA Delivery In Vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6 Conclusions and Future Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Abstract RNA-based therapeutic strategies are considered as a highly promising

alternative to conventional drug development. Among the different classes of
oligonucleotide-derived prospective drugs, small interfering RNAs (siRNAs) are
of particular interest. However, cellular uptake and subsequent intracellular traf-
ficking to the effector complex (RNA-induced silencing complex; RISC) represent
major technical hurdles for the efficacy of these macromolecular drugs. Thus, the
development of appropriate delivery systems is an essential requirement to turn
these molecules into medicine. In this review, we will focus on two particular
auspicious aspects in this context, the phosphorothioate-stimulated uptake of naked
siRNA and the use of cell-penetrating peptides as shuttles for a controlled cellular
S.D. Laufer (*), A. Detzer, G. Sczakiel, and T. Restle

Institut f€ur Molekulare Medizin, ZMSZ, Universit€at zu L€ ubeck & Universit€atsklinikum
Schleswig-Holstein, Ratzeburger Allee 160, 23538 L€ubeck, Germany
e-mail: laufer@imm.uni-luebeck.de
30 S.D. Laufer et al.
uptake. Moreover, we will present some of the most promising recent approaches
for siRNA delivery in vivo, which may help to pave the road to drugs of the future.
Keywords Argonaute 2 Caveosomal endocytosis pathway Cell-penetrating
peptides Cellular uptake Clinical trials Delivery Endocytosis Endoplasmic
reticulum Endosomal escape Extracellular RNA Golgi apparatus Ilimaquinone
Microinjection Nanoparticles Non-covalent Nonviral delivery systems Oligo-
nucleotide-based drugs Phosphorothioate-stimulated uptake Polycations Protein
transduction domains RNAi Signal peptides siRNAs siRNA-peptide conjugate
Abbreviations
Ago2 Argonaute 2
AMD age-related macular degeneration
CPP cell-penetrating peptide
dsRNA double-stranded RNA
ER endoplasmic reticulum
exNA extracellular nucleic acids
exRNA extracellular RNA
GFP green fluorescent protein
gp41 glycoprotein 41
HA hemagglutinin
HIV human immunodeficiency virus
IL interleukin
IFN interferon
JEV Japanese encephalitis virus
LF2000 Lipofectamine™ 2000
MEND multifunctional envelope-type nano device
miRNA microRNA
NLS nuclear localization sequence
PCI photochemical internalization
PCR polymerase chain reaction
PEG polyethylene glycol
PEI polyethyleneimine
PLL Poly-L-Lysine
PS-ON phosphorothioate-modified oligonucleotides
PTD protein transduction domain
PTGS posttranscriptional gene silencing
R8/R9 oligoarginines
RBD RNA-binding domain
RISC RNA-induced silencing complex
RNAi RNA interference
Selected Strategies for the Delivery of siRNA In Vitro and In Vivo 31
RVG rabies virus glycoprotein

siRNA small interfering RNA
shRNA short hairpin RNA
ssDNA single-stranded DNA
STR-R8 stearyl-R8
TLR Toll-like receptor
TNF tumor necrosis factor
TP10 transportan 10
VEGF vascular endothelial growth factor
1 Introduction
In recent years, RNA interference (RNAi) has gained a lot of interest as a tool for
functional genomics and probably equally important as a promising therapeutic
approach for the treatment of various diseases (Bumcrot et al. 2006; Castanotto and
Rossi 2009; de Fougerolles et al. 2007). However, despite these bright prospects, a
major impediment to the development of siRNA-based strategies for treatment and
prevention of diseases is the relatively inefficient means to effectively deliver these
macromolecules into the desired target cells or tissues. Although viral vectors have
been widely used to transfer genetic material into cells (Kootstra and Verma 2003;
Verma and Weitzman 2005), they bear an inherent risk for the patient to encounter
severe immunological responses or even develop cancer (Check 2005; Hacein-Bey-
Abina et al. 2003; Raper et al. 2002, 2003). As a result of these problems, much
attention has been paid in recent years to the delivery of naked RNA into a target
organ such as the lung or eye and the development of nonviral delivery systems.
Accumulating experimental evidence suggests that naked oligonucleotide-based
drugs including siRNA may be taken up by specific cell types in cell culture and
in vivo where they exert suppression of their target gene expression. Those findings
warrant more detailed analyses of this mode of delivery. The conception of nonviral
delivery includes an assortment of fairly unrelated approaches yielding various
degrees of enhanced cellular uptake of nucleic acids. Currently, liposomes and
cationic polymers are used as a standard tool to transfect cells in vitro. However,
these procedures are characterized by a significant lack of efficiency accompanied by
a high level of toxicity rendering them mostly inadequate for in vivo applications. In
this context, cell-penetrating peptides (CPPs) represent an interesting alternative as
they generally are less toxic than liposomes or cationic polymers. Moreover, they are
commonly better suited to transfer cargo into different cell types such as nonadherent
cells and primary cells, which are hard to transfect using commercially available
standard protocols. The most advanced approaches in the field are complex carrier
systems combining vantages of assorted strategies to generate nanoparticles with
better defined properties aimed toward enhanced uptake as well as intracellular
trafficking in combination with cell-specific functionalities.
In this chapter, we will report about particular aspects of siRNA delivery in vitro
and in vivo, with special emphasis on naked and CPP-mediated cellular delivery of
these macromolecules. Additionally, we will present and briefly discuss selected
recent examples of promising siRNA delivery approaches in vivo.
2 Mechanism of RNA Interference
RNAi is a highly evolutionally conserved and specific process of posttranscriptional

gene silencing (PTGS) by which double-stranded RNA (dsRNA), when introduced
into a cell, causes sequence-specific degradation of homologous mRNA sequences
(Fire et al. 1998; Rana 2007). Mechanistically, the process can be divided into two
steps. In the initiator step, dsRNA is cleaved by Dicer, a member of the RNase III
family, into 21–25 nt long siRNA fragments (Bernstein et al. 2001). In a consecutive
step, these fragments are transferred to RISC where one of the strands, the so called
guide strand, serves as a molecular template to recognize homologous mRNA that is
cleaved by Argonaute 2 (Ago2) (Hammond et al. 2001; Hutvagner and Simard 2008),
a protein component of RISC (Fig. 1a). Ago 2 is a protein of ca. 100 kDa and contains
four defined domains, N-terminal, PAZ, Mid, and PIWI (Fig. 1b). Current structural
knowledge is mainly derived from crystal structures of archaebacterial proteins
(Jinek and Doudna 2009). In the binary complex of Thermus thermophilus Ago and
guide strand, the 30 end of single-stranded RNA is bound to the PAZ domain and the
50 -phosphate is anchored within a binding pocket in the Mid domain (Wang et al.
2008a). Enzymatic activity is mediated by the C-terminal PIWI domain, which
resembles the catalytic triad of three carboxylate groups of RNase H (Song et al.
2004). These amino acid residues coordinate the essential metal and activate water
molecules for nucleolytic attack (Wang et al. 2008b). Once the guide strand is bound
to RISC, this complex can undergo many rounds of mRNA binding and cleavage
(Haley and Zamore 2004). To circumvent application of long double-stranded RNAs,
which inevitably trigger an interferon response, it is sufficient to extracellularly
supply 21 nt long dsRNAs (Elbashir et al. 2001a, b). Alternatively, siRNAs can be
expressed endogenously using DNA vectors that code for short hairpin (sh) RNAs
(Leung and Whittaker 2005; Paddison et al. 2002; Yu et al. 2002). These shRNAs are
then cleaved by Dicer to siRNAs. Short hairpin RNA constructs have advantages
over siRNA because the effects of these constructs can lead to a more stable and long-
term result (Rao et al. 2009). However, evidently they can interfere with the endoge-
nous microRNA pathway, thus causing severe side effects (Grimm et al. 2006; Snøve
and Rossi 2006).
3 Naked Delivery of siRNA In Vitro
3.1 Cellular Uptake of Naked Nucleic Acids
In tissue cell culture of mammalian cells and in vivo, nucleic acids including single-
stranded RNA and double-stranded DNA can be isolated from the extracellular
Fig. 1 Mechanistic principles of RNAi and structure of Thermus thermophilus Argonaute protein.
(a) Details are given in the text. (b) X-ray structure of the ternary complex of T. thermophilus
Argonaute bound to a 21-nucleotide guide DNA and a 20-nucleotide target RNA (pdb file: 3F73).
The protein contains four defined domains, N-terminal, PAZ, Mid, and PIWI, which are color
coded blue, magenta, gold, and green, respectively. Additionally, two linker regions are shown in
grey. Guide DNA is shown in red and target RNA in blue. The coordinated Mg2+ within the active
site (amino acids: D478, D546 and D660) is shown in cyan
environment. In higher mammals, this includes different body fluids such as blood,
serum, and urine.
Extracellular nucleic acids (exNA) were shown to be released from normal cells
and also from tumor cells, which means that one could hypothesize on tumor
cell-specific DNA and RNA in blood. In fact, non-invasive methods of early tumor
diagnostics are increasingly based on the analysis of circulating DNA. Even though
RNA is thought to be highly instable in blood, when compared to DNA, it was
found that extracellular RNA (exRNA) circulates in humans at amounts and
integrity that allow isolation, reverse transcription, and quantification by polymer-
ase chain reaction (PCR). The existence of amplifiable tumor-specific RNA in the
plasma of melanoma patients (Kopreski et al. 1999) and breast cancer patients
(Chen et al. 2000) was discovered despite the fact that the activity of blood RNases
is increased in patients with malignancies (Reddi and Holland 1976). Possible
sources of cell-free DNA and RNA are apoptotic bodies resulting from somatic
cell death [summarized by Garcia-Olmo et al. (2000)] and nutrition (Doerfler et al.
2001). Another endogenous source of cell-free nucleic acids to which cells and
organs of mammals are exposed is blood. The circulating blood system contains
significant concentrations of cell-free DNA and RNA (Anker and Stroun 2002; Ng
et al. 2002). Uptake of exNA by individual cells seems to be possible and may be of
biological relevance (Garcia-Olmo et al. 2000).Thus, it is warranted to speculate on
a biological role of exNA, which implies their recognition by cell surface molecules
and it might even include their cellular uptake.
Little is known about the internalization of cell-free nucleic acids by cells.
However, over the past years, Doerfler and colleagues have shown that mice fed
with naked DNA may incorporate this DNA in specific subsets of mononuclear cell
populations in the bloodstream (Doerfler 1995; Schubbert et al. 1994, 1997).
Surprisingly, such DNA is not completely degraded or metabolized, as fragments
of 200–400 bp in length of exogenously introduced DNA could be unequivocally
detected (Schubbert et al. 1994). For single-stranded DNA (ssDNA), much more is
known about the pathways of their cellular uptake (de Diesbach et al. 2000;
Laktionov et al. 1999).
Conversely, almost nothing is known about the conceivable cellular uptake of
short-chain RNA. By co-incubating a mammalian cell culture set-up with various
classes of nucleic acids and short double-stranded DNA competition of uptake was
measured quantitatively (Lehmann and Sczakiel 2005). Firstly, these studies sug-
gest that higher mammalian cells do take up nucleic acids measurably. Secondly,
cells distinguish between DNA and RNA as well as their characteristics regarding
chain length, global structure, and single- versus double-stranded forms. Recent
studies have shown that simple co-incubation of certain human cell types with
naked siRNA at micromolar and sub-micromolar concentrations leads to their
spontaneous cellular uptake within a few hours (Overhoff et al. 2004). Under
certain conditions, this is related to siRNA-specific target suppression indicating
that critical amounts of siRNA are internalized by the exposed cells in a biologi-
cally functional fashion.
3.2 The Phosphorothioate-Stimulated Cellular Delivery of siRNA
Fully phosphorothioate-modified oligonucleotides (PS-ON) enhance the cellular

uptake of naked siRNA in trans by various mammalian cell types (Overhoff and
Fig. 2 Model of the PS-stimulated cellular uptake of naked siRNA. Upon stimulation of a yet
unknown cell surface molecule by PS-ON, siRNA is taken up via caveolae into caveosomes and
transported to the perinuclearly located smooth ER. Since RNAi is thought to be a cytoplasmic
process, internalized siRNA needs to be released from the perinuclear compartments to be able to
interact with the RISC machinery. This figure was produced using Servier Medical Art
Sczakiel 2005). This means that siRNA and PS-ON are neither complexed nor is
there any measurable co-uptake of the PS-ON by the cells. A schematic depiction of
the underlying model is shown in Fig. 2. Essentially, one hypothesizes that PS-ON
recognize an unknown cell surface molecule that induces a kind of cellular stimu-
lation cascade giving rise to increased apparent uptake of coincubated extracellular
naked siRNA. This process is critically dependent on a number of characteristics
including the chemistry of the stimulating nucleic acid, its chain length, and its
concentration (Fig. 3). More specifically, one hypothesizes that two characteristics
of the stimulating nucleic acid are important for its activity, the phosphorothioate
internucleotide phosphate and a certain structure of the sugar.
The cellular uptake pathway of the cargo, i.e., siRNA, seems to make use of the
caveosomal endocytosis pathway, which is supported by experimental constraints
using pathway-specific activators or inhibitors and by fluorescence microscopy
(Overhoff and Sczakiel 2005). Present experimental data suggest that siRNA
migrates via caveosomes to the smooth endoplasmic reticulum (ER) where it is
trapped and only small amounts of siRNA seem to be released, thereby giving rise
to suppression of target gene expression (Detzer et al. 2009).
3.3 The siRNA-Peptide Conjugate Approach
A bulk of observations concerning the PS-stimulated delivery of siRNA indicates

that siRNA needs to be released from intracellular compartments or vesicles in
Fig. 3 Key characteristics of the PS-stimulated uptake of siRNA by mammalian cells. (a) Only a
fully PS-modified DNA backbone is active. (b) There is a sharp dependency of the amount of
internalized siRNA on the length of the PS-ON. The dotted line represents the detection limit of
the used nuclease protection assay. (c) The PS-stimulated uptake of siRNA reaches a plateau above
a concentration of PS-modified 24 mer of 500 nM
order to become biologically active as a suppressor of target RNA via RNAi. This
includes microscopic studies in the use of fluorescently labeled siRNA after its PS-
stimulated delivery and the discrepancy between large amounts of intracellular
siRNA and surprisingly low effectiveness, i.e., target suppression (Overhoff and
Sczakiel 2005). This view is compatible with the finding that ilimaquinone, a
substance that transiently disrupts the Golgi apparatus and at higher concentrations
also the ER (Takizawa et al. 1993; Wang et al. 1997), is related to increased target
suppression (Fig. 4). In particular, the concentration-dependent disruption of these
two cellular compartments strongly indicates that capturing of siRNA mainly
occurs in the smooth ER (Detzer et al. 2008).
In case of intracellular sorting of proteins, signal peptides serve as promoters of
intracellular transport, a process that may include transmembrane translocation
steps. Similar transport signals on the level of nucleic acids are not known;
however, one might think of a covalent attachment of signal peptides derived
from intracellular protein sorting to siRNA in order to facilitate the intracellular
release and, hence, to enhance the biological activity of siRNA (Fig. 5). For this
reason, the signal peptide TQIENLKEKG, which is thought to facilitate transloca-
tion of the catalytic domains of several bacterial protein toxins from transport
vesicles into the cytoplasm, was used as a tool to be covalently conjugated to
siRNA, thereby bypassing its presumed capturing in the ER (Detzer et al. 2009).
This study showed increased RNAi, i.e., siRNA-mediated target suppression.
Fig. 4 The progressive disruption of the ER and the Golgi apparatus by ilimaquinone is related to
increased intracellular release and biological activity of siRNA. This figure was produced using
Servier Medical Art
Fig. 5 Signal peptides steer a transmembrane translocation step of polypeptides (upper panel).
The concept of covalently attaching signal peptides to siRNA in order to enhance its release from
capturing in intracellular compartments is depicted in the lower panel. This figure was produced
using Servier Medical Art
We hypothesize that this is due to increased intracellular release and bioavailability

of this siRNA–peptide conjugate.
3.4 Intracellular Release of siRNA: A Major Hurdle
In past years, cellular delivery of siRNA was regarded as one of the major technical
problems for the successful application of oligonucleotide-based drugs in therapeu-
tic settings including antisense oligonucleotides and siRNA. Progressively, it
became obvious that physical delivery to mammalian cells could be substantially
improved with regard to the percentage of transfected cells as well as the total
amounts of internalized oligonucleotide-based drugs. However, in many cases,
improved delivery was not reflected by the extent of target suppression. For exam-
ple, limited biological activity of siRNA was observed in the use of a number of
delivery peptides as well as in the use of the PS-stimulated pathway. Hence, it seems
to be reasonable to assume that a block of “functional delivery” exists intracellu-
larly. Further, those findings suggest that intracellular transport and intracellular
release are crucial for the effectiveness of a variety of delivery modes of siRNA
(Fig. 6). A comparison of the mode of delivery of siRNA, the amount of intracellular
Fig. 6 Schematic depiction of the dose–response relationship of siRNA delivered by different

delivery modes as indicated at the top panel. Target gene expression is indicated by the Y-axis,
where half maximal inhibition is indicated by the IC50 value (dotted line). The concentration of the
drug, i.e., the siRNA is shown here as siRNA molecules per cell rather than using the usual
dimensions. This figure shows, for example, that MPGa-mediated delivery gives rise to approxi-
mately 10,000 copies per cell at half maximal target suppression. Conversely, LF2000 or electro-
poration require the amount of approximately 300 or 400 siRNA molecules per cell in order to
achieve a similar extent of target suppression
siRNA molecules, and the extent of siRNA-mediated target suppression indicates

that intracellular transport and/or release is a major obstacle to the application of
siRNA as, for example, microinjection of less than 20 copies of siRNA per cell leads
to a similar extent of target inhibition as the delivery of several hundred copies and
even up to more than 10,000 copies of siRNA by using the delivery technologies
indicated in Fig. 6 (Laufer and Restle 2008; Mescalchin et al. 2007).
As a consequence, this suggests exploring new strategies for steered intracellular
trafficking and biological activation of siRNA via subcellular release in order to
increase its biological effectiveness. One of such approaches might be the use of
siRNA–peptide conjugates as described above.
4 CPP-Mediated siRNA Delivery
4.1 Cell-Penetrating Peptides
The idea of using peptides as carriers for a controlled cellular delivery of siRNA
represents a promising concept to bypass the problem of poor bioavailability and
clinical efficacy of these nucleic acids. Twenty years ago, it was discovered that the
HIV-1 transactivating protein Tat is taken up by mammalian cells (Frankel and
Pabo 1988; Green and Loewenstein 1988), and a few years later, the Antennapedia
homeodomain of Drosophila melanogaster was shown to act similarly (Joliot et al.
1991). Later on, it could be shown that peptides derived from Tat and Antennape-
dia, i.e., Tat48–60 and penetratin, as well as other proteins are capable of transporting
macromolecular cargo molecules into cells (Allinquant et al. 1995; Fawell et al.
1994; Schwarze et al. 1999). Based on such promising results, a rapidly expanding
field focusing on the so-called cell-penetrating peptides (CPPs), also referred to as
protein transduction domains (PTDs), began to develop. Since the first reports about
Tat, a large number of naturally occurring as well as engineered CPPs have been
described (Foged and Nielsen 2008; Heitz et al. 2009; Langel 2006; Lindgren et al.
2000; Morris et al. 2008; Patel et al. 2007; Veldhoen et al. 2008; Zorko and Langel
2005). In addition to Tat and penetratin, well-known examples include transportan,
a chimeric peptide composed of galanin and mastoparan (Pooga et al. 1998), and
oligoarginines (Futaki et al. 2001; Futaki 2006). Generally, CPPs are short poly-
cationic sequences of less than 30 amino acids that are able to translocate different
cargoes (e.g., nucleic acids, peptides, and even entire proteins) into cells. The only
common characteristic of these peptides appears to be that they are net positively
charged at physiological pH. Table 1 gives an overview of selected “classical”
CPPs. In the majority of cases, the cargo is covalently attached to the CPP, which
can be achieved by expression as a fusion construct or by chemical coupling [for a
review see, Zatsepin et al. (2005)]. In particular cases, cargo and carrier bind each
other non-covalently through mainly ionic interactions (Crombez et al. 2008;
Deshayes et al. 2008; Laufer and Restle 2008; Morris et al. 2008). Depending on
the nature of both binding partners, the assembly of nanoparticles may occur.
Table 1 Selected examples of “classical” CPPs

Peptide Sequence References
Tat48–60 GRKKRRQRRRPPQ Vives et al. (1997)
Penetratin (Antp43–58) RQIKIWFQNRRMKWKK Derossi et al. (1994)
Transportan GWTLNSAGYLLGKINLKALAALAKKIL Pooga et al. (1998)
TP10 AGYLLGKINLKALAALAKKIL Soomets et al. (2000)
Oligoarginine (R8) RRRRRRRR Futaki et al. (2001)
MAP KLALKLALKALKAALKLA Oehlke et al. (1998)
MPG GALFLGFLGAAGSTMGAWSQPKKKRKV Morris et al. (1997)
MPGa GALFLAFLAAALSLMGLWSQPKKKRKV Deshayes et al. (2004)
Despite the widespread interest in using peptides as carriers, the mechanisms

underlying the cellular translocation of CPPs are still not completely understood.
Early work relied upon fluorescence imaging or flow cytometry analysis of chemi-
cally fixed cells to examine the intracellular localization of fluorescently labeled
peptides in the absence or presence of cargo. From these experiments, it was
concluded that CPPs penetrate cell membranes by an energy-independent mecha-
nism as they appeared to be internalized very rapidly within minutes even at 4 C
(Derossi et al. 1996; Futaki et al. 2001; Morris et al. 1997; Schwarze and Dowdy
2000; Vives et al. 1997). Despite some reports that certain fixation procedures may
cause artifacts leading to an overestimation of cellular uptake rates (Lundberg and
Johansson 2001; Lundberg and Johansson 2002; Pichon et al. 1999), the whole
extent of this problem was not commonly recognized until a detailed side by side
comparison was performed by Richard et al. (2003). The authors demonstrated that
the distribution pattern of Tat48–60 and R9 as well as their conjugates was
completely different in living versus fixed cells and that an important source of
misinterpretation is caused by difficulties to distinguish cell surface-associated
CPPs from internalized CPPs. Above all, it could be shown that endocytotic
transport is significantly involved in the internalization process (Richard et al.
2003). Prior to endocytosis, CPPs interact electrostatically with the extracellular
matrix of the cell surface mostly through binding to negatively charged glycosa-
minoglycans, i.e., heparan sulfate proteoglycans (Console et al. 2003; Tyagi et al.
2001). Based on the findings described above, many groups reexamined their data,
and in most cases, endocytosis was suggested as the main route of internalization
(Fig. 7), although substantial difficulties are encountered in identifying the exact-
pathway of CPP uptake ((Veldhoen et al. 2006; Veldhoen et al. 2008) and
references therein). As a consequence, retention in endosomes is one of the major
rate-limiting steps for cellular delivery of macromolecules via cationic lipids,
polyplexes, and especially CPPs. Endosomal release can be increased by endo-
some-disrupting substances (i.e., chloroquine, calcium, or sucrose), by coadminis-
tration of photosensitive substances [so-called photochemical internalization (PCI)
(Berg et al. 2007; Bøe et al. 2007; Bonsted et al. 2008; Folini et al. 2007; Oliveira
et al. 2007a, 2008)] or viral fusogenic peptides (Epand 2003; Futaki et al. 2005;
Haque et al. 2005; Kwon et al. 2008; Michiue et al. 2005; Oliveira et al. 2007b;
Plank et al. 1998; Tu and Kim 2008).
Fig. 7 Principles of peptide-based nucleic acid delivery systems. Interaction of CPP and cargo is
either achieved by covalent attachment or by non-covalent complexation through mainly ionic
interactions. In case of non-covalent complex formation, a further assembly of cargo/carrier
complexes occurs, leading to the formation of large nan-oparticles. In case of covalently joined
molecules, a similar scenario is less likely, yet cannot be excluded. Prior to the translocation
process, the particles attach to the cell surface by ionic interactions of positively charged CPP
residues with negatively charged membrane components. Subsequently, complexes are taken up
via an endocytotic pathway. Although less likely, direct penetration cannot be excluded and may
occur simultaneously. Once inside the cell, the cargo has to escape from vesicular compartments;
otherwise, it eventually gets degraded in the lysosome. This figure was produced using Servier
Medical Art
In summary, the precise mechanism of internalization remains elusive and

strongly depends on the properties of both CPP and cargo as well as on the
transfection conditions and the cell lines used (De Coupade et al. 2005; Edenhofer
2008; El-Andaloussi et al. 2007; Fittipaldi et al. 2003; Maiolo et al. 2005; Mano
et al. 2005; Richard et al. 2003, 2005; Rothbard et al. 2004; Wadia et al. 2004).
4.2 Selected Examples of CPP-Mediated siRNA Delivery
As described above, siRNAs represent a valuable tool to inhibit the expression

of a target gene in a sequence-specific manner. In the following section, selected
examples of CPP-mediated siRNA delivery will be presented, which are summar-

ized in Table 2.
Only a few studies describe the covalent attachment of nucleic acid cargo and
peptide carrier (confer Table 2). In one approach, simple mixing of siRNA targeted
against GFP or CDK9 and Tat peptide did not generate any measurable RNAi effect,
whereas cross-linked siRNA-Tat47–57 led to a significant downregulation of the target
proteins (Chiu et al. 2004). Tat47–57-mediated transfection of siRNA resulted in a
perinuclear localization of the nucleic acid. In contrast, fluorescently labeled Tat47–57
without cargo was mainly found in the nucleolus, suggesting that interactions with
RISC influence subcellular localization. In another approach, significant uptake of
siRNAs targeted against luciferase or GFP could be observed after coupling the 50 -end
of the sense strand via a disulfide bond to penetratin or transportan (Muratovska and
Eccles 2004). Davidson et al. (2004) similarly used disulfide coupling to conjugate
siRNAs directed against different endogenous proteins, e.g., several caspases, to
penetratin. After transfection into neuronal cells, a remarkably strong downregulation
of the target proteins in these hard to transfect primary cells was observed, implying
that peptide-mediated siRNA delivery was far more effective in comparison to
LF2000. Concerning the in vivo delivery of Tat48–60- or penetratin-siRNA conjugates,
Moschos et al. showed that intratracheal administration did not lead to any intensifi-
cation of the knockdown of the target gene p38 mitogen-activated protein kinase in
mouse lungs in comparison to unmodified nonformulated siRNA (Moschos et al.
2007). Strikingly, it was found that the peptides alone triggered a detectable decrease
in target gene expression and that the penetratin-conjugate induced elevated levels of
the immune markers IFN-a, TNF-a, and IL-12p40 in lung tissue.
Although CPPs are able to deliver a wide variety of cargo into cells, technical
difficulties arise especially from the syntheses of conjugates consisting of short
cationic or hydrophobic peptides and highly negatively charged siRNAs. Moreover,
Dowdy and his group (Meade and Dowdy 2008) present a rather critical point of
view referring to previous studies with CPP-siRNA-conjugates. They claim that the
successful delivery described therein is solely the result of excess free peptide,
which leads to additional complexation, and thereby cellular import of the siRNA.
This is in accordance with Turner et al. (2005), who were the first to observe
that careful purification of CPP-antisense oligonucleotide-conjugates abrogates
their biological effect. Among other things, this might be the reason why most
of the studies reporting on successful peptide-mediated delivery of siRNAs use a
noncovalent complexation approach (confer Table 2).
As described above, direct conjugation of anionic siRNAs to cationic peptides,
in this paragraph called PTDs, results in charge neutralization followed by aggre-
gation and thereby inactivation of the PTD. To avoid this problem, Eguchi et al.
(2009) fused Tat with a double-stranded RNA-binding domain (DRBD), which
binds to siRNA and masks its negative charge. The resulting complex consists of a
single RNA surrounded by four PTD-DRBDs and was used to deliver siRNAs
against GFP or GAPDH. Efficient gene silencing without cytotoxicity or off-target
effects could be shown even in difficult-to-transfect primary cells as well as in a
reporter mouse model in vivo.
Table 2 Examples for peptide-mediated delivery of siRNA

CPP/delivery system Mode Target Cell line References
of
linkagea
47–57
Tat , Tat-derived c EGFP, HeLa Chiu et al. (2004)
oligocarbamate CDK9
Penetratin, transportan c luciferase, Cos-7, C166, Muratovska and Eccles
GFP EOMA, CHO- (2004)
AA8
Penetratin c Cu-ZN Primary rat Davidson et al. (2004)
SOD-1, hippocampal or
Caspase-3/ sympathetic
-8/-9 neurons
Tat48–60, penetratin c p38 MAP L929 (mouse Moschos et al. (2007)
kinase fibroblasts),
mouse lung
(intratracheal)
PTD-DRBD n-c dGFP, H1299, Jurkat, Eguchi et al. (2009)
GAPDH HUVEC,
human
embryonic stem
cells
MPG, MPGDNLS n-c Luciferase, HeLa, Cos-7, HS- Simeoni et al. (2003)
GAPDH 68
MPGa n-c Luciferase HeLa, ECV 304 Veldhoen et al. (2006)
CADY n-c GAPDH, U2OS, THP1, Crombez et al. (2009a)
p53 HUVEC, 3T3C
MPG-8 n-c cyclin B1 HeLa, HS68, MCF- Crombez et al. (2009b)
7, PC3, SKBr3-
HER, mouse
(intravenously,
intratumoral)
Chol-R9 n-c VEGF CT-26, mouse Kim et al. (2006)
(intratumoral)
H3K8b, H3K8b(+RGD) n-c b-Gal, SVR-bag4, MDA- Leng et al. (2005)
luciferase MB-435, C6
POD n-c EGFP HER 911 Johnson et al. (2008)
EB1, MPGDNLS, bPrPp n-c Luciferase HeLa, HepG2 Lundberg et al. (2007)
TatU1A n-c EGFP, CHO, A431, Endoh et al. (2008)
EGFR
stearyl-R8 n-c EGFP, Primary rat Tönges et al. (2006)
MAP2B hippocampal
neurons
R8-MEND (siRNA/ n-c Luciferase HeLa Nakamura et al. (2007)
stearyl-R8 core)
R8/GALA-MEND n-c Luciferase HeLa Sakurai et al. (2009)
Chol-R9 n-c VEGF CT-26, mouse Kim et al. (2006)
YSA-nanogel n-c EGFR Hey, BG-1 Blackburn et al. (2009)
DMMAn-Mel n-c Luciferase Neuro 2A- Meyer et al. (2008)
eGFPLuc
a
c ¼ covalent/n-c ¼ non-covalent
Simeoni et al. (2003) were the first who non-covalently complexed siRNA with
the peptide MPG. MPG is a 27 amino acid peptide composed of a hydrophobic
domain derived from the N-terminal fusion sequence of the HIV-1 glycoprotein 41
and a hydrophilic domain derived from the nuclear localization sequence (NLS) of
the SV40 large T-antigen, which are linked by a 3 amino acid spacer (Morris et al.
1997). At a 1:10 ratio of negative nucleic acid to positive peptide charges, a
decrease in luciferase activity of about 80% was detectable in HeLa or Cos-7
cells. This effect was further enhanced to about 90% downregulation by a mutation
in the NLS sequence of the carrier peptide (MPGDNLS), presumably due to an
increased delivery to the cytoplasm, where RISC is localized.
In the following, Veldhoen et al. (2006) used a derivative of the MPG peptide for
the delivery of siRNA. This variant, termed MPGa, differs from MPG by five
amino acids in the hydrophobic part. These changes result in an alteration of the
overall structure of the peptide towards a higher tendency of adopting a helical
conformation (Deshayes et al. 2004). MPGa forms highly stable non-covalent
complexes with nucleic acids through ionic interactions of the positively charged
NLS sequence and negative charges of the cargo. Furthermore, hydrophobic pep-
tide/peptide interactions lead to the formation of nanoparticles. Using a luciferase-
targeted siRNA as cargo, reporter gene activity could be inhibited up to 90% with
an IC50 value in the subnanomolar range. Confocal microscopy studies as well as
transfections in the presence of inhibitors of different endocytotic pathways
strongly indicate that endocytosis is involved in the cellular uptake of peptide/
siRNA complexes. As a key issue, the authors quantified the intracellular number of
siRNA molecules after MPGa-mediated transfection and compared it to the amount
of extracellularly applied RNA. Together with data from microinjection experi-
ments (Laufer and Restle 2008), this comparison yields the percentage of inter-
nalized molecules that are biologically active. In the case of MPGa-mediated
siRNA delivery, only 0.1% of internalized oligonucleotides are biologically active
whereas more than 99% are probably retained in endosomes (confer Fig. 6).
Recently, Crombez et al. (2009a) designed a similar secondary amphipathic
peptide, called CADY, which adopts a helical conformation within cell membranes,
exposing cationic arginine residues on one side and aromatic tryptophan groups on
the other. CADY forms stable complexes with siRNAs already at a molar ratio of
5:1–10:1 (peptide:siRNA), whereas for protection from serum nucleases, optimal
cellular uptake and significant target knockdown higher molar ratios (>20:1) are
required. Cellular uptake and the associated biological response were hardly
affected in the presence of different inhibitors of endocytosis; therefore, the authors
concluded that the entry mechanism of CADY/siRNA complexes is independent of
the endosomal pathway.
The same group (Crombez et al. 2009b) shortened the original MPG peptide by
six residues and mutated two residues to tryptophan, yielding a 21 amino acid
peptide called MPG-8. In addition to the cysteamide group at the C-terminus, a
b-alanine was added at the N-terminus to allow further functionalization of the
peptide. Concerning siRNA delivery, the optimal molar ratio was determined
to be 20:1 (peptide: siRNA), and under these conditions, MPG-8 exhibited a
significantly higher gene silencing activity than the parent peptide MPGDNLS. In
addition to target downregulation on the mRNA- as well as the protein-level, MPG-
8-mediated delivery of anticyclin B1 siRNA induced G2 arrest and blocked cell
proliferation specifically in cancer cells. Using a xenograft tumor mouse model,
local intratumoral administration but not intravenous injection of 0.25 mg/kg MPG-
8/siRNA particles prevented tumor growth completely. To improve systemic deliv-
ery, MPG-8/siRNA particles were functionalized with a cholesterol moiety through
activation of the N-terminal b-alanine group. This modification increased the
distribution level of anti-cyclin B1 siRNA and blocked tumor growth upon systemic
intravenous administration in a xenograft human prostate as well as human lung
cancer mouse model without activation of the innate immune system.
Similar synergistic effects had already been shown by Kim et al. (2006), who
combined oligoarginine with cholesterol (Chol-R9) for the non-covalent complex-
ation of an anti-VEGF siRNA. Chol-R9/siVEGF complexes suppressed VEGF
production in vitro in CT-26 cells as well as in an in vivo mouse model after
local administration to a subcutaneous tumor. Here, the lowered VEGF level was
accompanied by decreased tumor growth, which was probably due to the anti-
angiogenetic effect on tumor vascularization.
As briefly outlined above, the major rate-limiting step for most delivery
approaches is endosomal entrapment of the nucleic acids. Thus, many groups try
to improve their systems with the aim to increase endosomal escape of siRNA after
peptide-mediated delivery. Lundberg et al. (2007) rationally modified penetratin to
form a CPP (termed EB1) with improved endosomolytic properties. They achieved
a pH-dependent conformational change of the peptide to a higher degree of helicity
by the replacement of two basic amino acids with histidines and the N-terminal
addition of six amino acids. In this study, several CPPs were compared in a non-
covalent approach by measuring the overall cellular uptake via fluorescence and the
biological effect of siRNA targeted to luciferase mRNA. Penetratin- as well as
TP10-mediated transfection did not lead to any silencing of luciferase gene expres-
sion, despite high amounts of intracellular siRNA (Lundberg et al. 2007) in contrast
to previous reports using siRNA–penetratin-conjugates (Davidson et al. 2004) or
TP10/DNA-complexes (El-Andaloussi et al. 2005). EB1-mediated delivery of
100 nM siRNA led to approximately 50% reduction of luciferase activity. This
silencing effect was slightly better than for bPrPp and in the same range as for
MPGDNLS. As it was described earlier that addition of a pH-sensitive peptide
derived from hemagglutinin (HA2) can promote endosomal escape (Wadia et al.
2004), the authors linked HA2 to penetratin (Lundberg et al. 2007). It turned out
that although HA2-penetratin improved the silencing effect when coincubated with
penetratin, EB1 was more potent than this combination of peptides. Together with
confocal microscopy studies, the authors concluded that the lack of biological
effect after penetratin-mediated siRNA delivery is due to a lack of endosomal
escape and that EB1 has a superior endosomolytic activity in comparison to
HA2-penetratin.
Endoh et al. (2007, 2008) and Endoh and Ohtsuki (2009) recently presented
an innovative strategy, called CLIP-RNAi (i.e., CPP-linked RBP-mediated RNA
internalization and photoinduced RNAi), combining delivery of a specific RNA

sequence with enhanced photoinduced release of RNA from endosomes. This goal
was accomplished by fusing the U1A RNA-binding domain (RBD) to the Tat
peptide and extending the siRNA with a short stretch of nucleotides specifically
recognized by this RBD. These complexes were efficiently internalized but exhib-
ited a punctate cytoplasmic localization pattern, indicative of endosomal entrap-
ment. However, photostimulation of a fluorophore attached to the peptide led to a
redistribution of complexes into the cytosol followed by efficient RNAi-mediated
gene silencing.
In addition to “simple” CPP-based delivery systems composed of single peptides,
there is a trend to develop systems of higher complexity. A main goal of such
approaches is to generate nanoparticles with defined properties (e.g., size and charge
distribution) as well as to provide cell-specific functionalities, which are especially
important for in vivo use. Since a comprehensive description of recent developments
would be far beyond the scope of this article, we can only give a few examples with
respect to successful siRNA delivery. In general, there are attempts to combine
peptides with cationic liposomes (Futaki et al. 2005; Hyndman et al. 2004; Preuss
et al. 2003; Read et al. 2003, 2005; Torchilin et al. 2001, 2003) or polyethyleneimine
(PEI) (Kilk et al. 2005). Other applications are aimed towards the synthesis of high
or low molecular weight branched polymers and/or peptides (Chen et al. 2001; Fattal
and Barratt 2009; Leng et al. 2005; Liu et al. 2005; Midoux and Monsigny 1999;
Read et al. 2003; Ritter et al. 2003) or dendrimers (Bayele et al. 2005, 2006; Kang
et al. 2005). Recent developments of even more complex systems are particularly
promising with respect to in vivo delivery (Kale and Torchilin 2007a, b; Khalil et al.
2007; Rahbek et al. 2008; Soundara Manickam and Oupický 2006).
One example is a recently developed novel packaging approach, a multifunc-
tional envelope-type nano device (MEND), which allows the assembly of multiple
devices in a single delivery system (Kogure et al. 2008). In principle, nucleic acids
are condensed using a polycation to form a core particle, followed by encapsulation
in a lipid envelope. For the delivery of siRNA, further modifications were included
into MEND (Nakamura et al. 2007; Sakurai et al. 2009). First, the addition of
stearylated octaarginine (STR-R8) on the lipid envelope led to efficient cellular
uptake by macropinocytosis. Second, dioleoylphosphatidyl ethanolamine (DOPE)
and phosphatidic acid (PA) were added because of their high fusogenic activity,
which improved silencing activity through increased release of functional siRNA
into the cytosol. Third, the pH-sensitive fusogenic peptide GALA, additionally
conjugated with cholesterol, also enhanced endosomal release of encapsulated
nucleic acids. Finally, the STR-R8-MEND, prepared with a lipid composition of
DOPE/PA plus Chol-GALA, was modified with a cleavable PEG–peptide–DOPE
conjugate (PPD) to enhance in vivo tumor targeting. So far, this system has already
been used for efficient gene silencing in HeLa cells and appears to be a promising
new carrier for siRNA into tumor cells.
Several groups have developed drug delivery approaches using synthetic hydro-
gel nanoparticles (nanogels). These core/shell particles physically segregate the
function of cell and drug binding (¼shell) from the function of endosomal
disruption (¼core) (Hu et al. 2009). Furthermore, they assemble into stable and
well-defined complexes with a high payload capacity and can be selectively sur-
face-functionalized to enable cell type-specific targeting. For example, Blackburn
et al. (2009) used the 12 amino acid peptide YSA for the delivery of anti-EGFR
siRNA to ovarian cancer cells via ligand-receptor binding mediated endocytosis.
Polycations, like PEI or PLL alone, can promote significant plasmid DNA
transfer efficiency but show only modest siRNA delivery activity. Therefore,
Meyer et al. (2008, 2009) functionalized these polycations with polyethylene glycol
(PEG) and a pH-responsive endosomolytic melittin peptide from bee venom (Ogris
et al. 2001). To minimize lytic activity in the extracellular environment, melittin
was further modified with dimethylmaleic anhydride (DMMAn), which is cleaved
in the endosome and therefore restores lytic activity in the intracellular compart-
ment. Modification of PEI or PLL with DMMAn-Mel greatly enhanced siRNA-
mediated luciferase gene knockdown (Meyer et al. 2009).
5 Selected Examples of siRNA Delivery In Vivo
When it comes to in vivo delivery of siRNA, the situation gets much more
complicated than described above on a cellular level. In principle, the nucleic
acid molecules can be administered topically or locally to, for example, the eye,
skin, mucus membranes, and local tumors, or systemically through the blood
stream. Especially in the latter case, besides cellular uptake, there are many more
additional hurdles to consider like serum stability, aggregation with serum proteins,
uptake by phagocytes, and clearance by the kidneys (Alexis et al. 2008; Xie et al.
2006). Moreover, a significant challenge for siRNA delivery to many tissues
represents migration from the bloodstream across the vascular endothelium and
subsequently diffusion through the extracellular matrix, a dense network of poly-
saccharides and fibrous proteins.
There are a variety of techniques described to deliver siRNA in vivo. The
simplest option is the application of naked RNA into a target organ either non-
modified or chemically modified (e.g., 20 -O-methyl modifications). For systemic
delivery, siRNA can be conjugated for example with PEG, cholesterol, or small
peptides or alternatively complexed with peptides, lipids, polymers, polycations, or
even complex nanoparticles, in certain cases, in combination with antibodies or cell
surface-specific ligands for targeted delivery [for a review see, Jeong et al. (2009)].
To cover all of the different approaches described in the literature would be far
beyond the scope of this article. Therefore, we will focus on some of the most
promising examples described in recent years. A summary of these experiments can
be found in Table 3. For a more comprehensive coverage, the reader is referred to
these recent reviews (Aigner 2008; Castanotto and Rossi 2009; Whitehead et al.
2009) and references therein.
The first successful downregulation of a target mRNA by siRNA in mammals
was shown by McCaffrey et al. (2002). In this study, the authors showed that
Table 3 Selected examples of nonviral siRNA delivery in vivo

Formulation of siRNA Mode of Target protein Observed effect References
administration
Naked Hydrodynamic Firefly luciferase Reduction of target McCaffrey et al.
transfection gene expression (2002)
Naked Intravenous Fas (also known as Protected mice from Song et al. (2003)
injection TNFRSF6) liver fibrosis
Naked/TransIT-TKO Intranasal RSV-P, PIV-P Protection from Bitko et al. (2005)
(polyamine) respiratory
infection
Protamine–antibody Intratumoral or c-myc, MDM2, Inhibition of s.c. Song et al. (2005)
fusion protein intravenous VEGF melanoma
injection xenograft growth
Aptamer/siRNA Intratumoral PLK1 (A10-Plk1), Triggering of apoptosis, McNamara et al.
chimeras injection BCL2 (A10- growth inhibition, (2006)
Bcl2) and tumor
regression in mouse
xenograft model
Rabies virus Intravenous Japanese Protection against fatal Kumar et al. (2007)
glycoprotein injection encephalitis viral encephalitis
peptide/r9/siRNA virus (JEV)
b1,3-D-glucan- Oral gavage Tumor necrosis Protection from Aouadi et al. (2009)
encapsulated factor a (TNF- lipopolysaccharide
siRNA particles a), mitogen- (LPS)-induced
activated lethality
protein kinase
kinase kinase
kinase 4
(Map4k4)
transgene firefly luciferase expression can be suppressed in adult mice by synthetic

siRNAs injected to the liver. The first therapeutic application was reported just 1
year later by Song et al. (2003). In this study, mice could be protected from Fas-
mediated liver fibrosis by downregulation of Fas. In a fulminant hepatitis induced
by injecting agonistic Fas-specific antibody, 82% of mice treated with siRNA that
effectively silenced Fas survived for 10 days of observation, whereas all control
mice died within 3 days. This was a first promising example of the therapeutic
potential of RNAi in vivo. In 2005, Bitko et al. (2005) reported that individual as
well as joint infection by respiratory syncytial virus (RSV) and parainfluenza virus
(PIV) can be specifically prevented and inhibited by siRNAs, instilled intranasally
in the mouse, with or without transfection reagents. Their results suggested for the
first time that, if properly designed, low dosages of inhaled siRNA might offer a
fast, potent, and easily administrable antiviral regimen against respiratory viral
diseases in humans.
One of the first examples of targeted delivery, i.e., cell type-specific delivery, was
a study by Song et al. (2005). Here, the authors used a protamine/antibody fusion
protein to deliver siRNAs specifically to cells expressing the HIV-1 envelope protein.
The positively charged protamine served as binding partner for the negatively
charged siRNA, whereas a heavy-chain antigen-binding region (Fab) permitted
specific interaction with surface exposed gp160 molecules followed by internaliza-

tion and eventual release of the siRNA cargo. Intratumoral or intravenous injection of
Fab/protamine-complexed siRNAs into mice targeted HIV envelope-expressing B16
melanoma cells, but not normal tissue or envelope-negative B16 cells. Using siRNAs
against c-myc, MDM2 or VEGF envelope-expressing subcutaneous B16 tumors
could be inhibited. Another technology for cell type-specific delivery is based on
aptamer/siRNA chimeras (McNamara et al. 2006). Aptamers are small (25–60
nucleotides) oligonucleotide ligands (either DNA or RNA) derived from an in vitro
evolution process called SELEX (systematic evolution of ligands by exponential
enrichment) (Ellington and Szostak 1990; Robertson and Joyce 1990; Tuerk and
Gold 1990). Such nucleic acid ligands do bind with high affinity and specificity to
their target molecules. In the present case, an aptamer selected against the cell-surface
receptor PSMA (prostate specific membrane antigen) was used, linked to either polo-
like kinase 1 (PLK1) or B-cell lymphoma 2 (BCL2)-specific siRNAs. Intratumoral
injection of these conjugates into a mouse xenograft model resulted in triggering of
apoptosis, growth inhibition, and tumor regression. A third example of targeted
delivery is the use of a short peptide derived from rabies virus glycoprotein (RVG),
which enabled transvascular delivery of siRNAs directed against Japanese encepha-
litis virus (JEV) to the brain (Kumar et al. 2007). The 29-amino-acid RVG peptide
specifically binds to the acetylcholine receptor expressed by neuronal cells. This
peptide was fused with R9 to permit siRNA binding. Intravenous treatment with
RVG-9R-bound antiviral siRNA let to a robust protection against fatal viral enceph-
alitis in mice.
An interesting additional delivery route for siRNAs was recently published by
Aouadi et al. (2009). Here, orally delivered siRNA targeting macrophage mitogen-
activated protein kinase kinase kinase kinase 4 (Map4k4) suppressed systemic
inflammation in mice. As vehicle hollow, porous 2–4 mm-sized shells composed
primarily of b1,3-D-glucan were prepared by treating baker’s yeast with a series of
alkaline, acid and solvent extractions to remove cytoplasm and other cell wall
polysaccharides. The anionic siRNA is bound within these particles between cationic
polyethylenimine layers through electrostatic interactions. The orally administered
particles are then phagocytosed by macrophages and dendritic cells in the gut-
associated lymphatic tissue. Moreover, the authors speculate that these cells may
traffic away from the gut and infiltrate other reticuloendothelial system tissues, so
that, over time, total body macrophages contain siRNAs. Lipopolysaccharide (LPS)/
D-galactosamine (DGalN) challenged mice could be protected from inflammatory
cytokine toxicity by oral gavage of Map4k4-siRNA-containing particles through
inhibition of tumor necrosis factor a (TNF-a) and Interleukin-1b (IL-1b) production
in macrophages. Interestingly, in vivo potency of these siRNAs was 5–250 times
greater than that in previous studies reporting systemic delivery (Filleur et al. 2003;
McCaffrey et al. 2002; Peer et al. 2007; Song et al. 2005; Sorensen et al. 2003;
Soutschek et al. 2004; Wesche-Soldato et al. 2005; Zimmermann et al. 2006).
Currently, there are several ongoing clinical trials for siRNA therapeutics
(Table 4). Several of the more advanced trials are targeted at age-related macular
Table 4 Selected examples of current clinical trials for siRNA therapeutics

siRNA Company Disease Mode of administration Status
ALN- Alnylam Respiratory syncytial Local II
RSV01 Pharmaceuticals virus
CALAA-01 Calando Solid tumors Systemic/intravenous I
Pharmaceuticals
Sirna-027 Sirna Therapeutics Age-related macular Topical/intravitreal I
degeneration
TD101 TransDerm Pachyonychia Topical/foot I
congenita
I5NP Quark Pharmaceuticals Acute kidney injury Systemic/intravenous I
after cardiac bypass
surgery
Bevasiranib Opko Health Age related macular Topical/intravitreal III, discontinued
degeneration
degeneration (AMD), which is a leading cause of blindness. This disease arises

from excessive blood-vessel growth and rupture within the cornea. In terms of drug
delivery, a treatment of AMD is less challenging than other diseases since the
molecules of interest can be administered intravitreally, a procedure which avoids
many of the problems with in vivo delivery briefly described above. The siRNAs
under investigation are targeted to vascular endothelial growth factor (VEGF) and
its receptor (VEGFR). Although initially the results obtained were quite encourag-
ing, a recent study by Kleinman et al. (2008) reported about sequence-independent
angiogenesis suppression by siRNA via nonspecific stimulation of the Toll-like
receptor 3 (TLR3) pathway. While this study questions the AMD-related clinical
trials, it does not explain the therapeutic effects of other trials where appropriate
controls have been performed.
6 Conclusions and Future Prospects
Currently, the development of effective and safe delivery systems for therapeutic
oligonucleotides like siRNA is crucial to one day bring these molecules to the
clinic. Besides the development of viral vectors as delivery vehicles, there is a
highly diverse and constantly increasing number of nonviral systems evolving.
However, at present, even the most advanced systems either lack the efficiencies
required for downstream drug development or do show a substantial degree of
toxicity or both. Of the many factors that limit their use, cellular uptake of the
cargo/carrier complexes and particularly subsequent intracellular trafficking to
reach the target site are the most important. Moreover, for in vivo use, various
additional obstacles are to be taken into account like serum stability, pharmacoki-
netic considerations, and tissue barriers as well as target cell specificity. In spite of
these somewhat sobering insights, there is noticeable progress especially in recent
years. While most of the underlying problems are meanwhile identified, the
answers to these problems remain challenging. Most likely, there will be no
magic bullet but individual solutions for any given application.
Acknowledgments We apologize to those authors whose work was not cited directly owing to
space limitations. T.R. acknowledges funding by EC-grant LSHG-CT-2003-503480.
References
Aigner A (2008) Cellular delivery in vivo of siRNA-based therapeutics. Curr Pharm Des
14:3603–3619
Alexis F, Pridgen E, Molnar LK et al (2008) Factors affecting the clearance and biodistribution of
polymeric nanoparticles. Mol Pharm 5:505–515
Allinquant B, Hantraye P, Mailleux P et al (1995) Downregulation of amyloid precursor protein
inhibits neurite outgrowth in vitro. J Cell Biol 128:919–927
Anker P, Stroun M (2002) Progress in the knowledge of circulating nucleic acids: plasma RNA is
particle-associated. Can it become a general detection marker for a cancer blood test? Clin
Chem 48:1210–1211
Aouadi M, Tesz GJ, Nicoloro SM et al (2009) Orally delivered siRNA targeting macrophage
Map4k4 suppresses systemic inflammation. Nature 458:1180–1184
Bayele HK, Sakthivel T, O’Donell M et al (2005) Versatile peptide dendrimers for nucleic acid
delivery. J Pharm Sci 94:446–457
Bayele HK, Ramaswamy C, Wilderspin AF et al (2006) Protein transduction by lipidic peptide
dendrimers. J Pharm Sci 95:1227–1237
Berg K, Folini M, Prasmickaite L et al (2007) Photochemical internalization: a new tool for drug
delivery. Curr Pharm Biotechnol 8:362–372
Bernstein E, Caudy AA, Hammond SM et al (2001) Role for a bidentate ribonuclease in the
initiation step of RNA interference. Nature 409:363–366
Bitko V, Musiyenko A, Shulyayeva O et al (2005) Inhibition of respiratory viruses by nasally
administered siRNA. Nat Med 11:50–55
Blackburn WH, Dickerson EB, Smith MH et al (2009) Peptide-functionalized nanogels for
targeted siRNA delivery. Bioconjug Chem 20(5):960–968
Bøe S, Longva AS, Hovig E (2007) Photochemically induced gene silencing using small inter-
fering RNA molecules in combination with lipid carriers. Oligonucleotides 17:166–173
Bonsted A, Wagner E, Prasmickaite L et al (2008) Photochemical enhancement of DNA delivery
by EGF receptor targeted polyplexes. Methods Mol Biol 434:171–181
Bumcrot D, Manoharan M, Koteliansky V et al (2006) RNAi therapeutics: a potential new class of
pharmaceutical drugs. Nat Chem Biol 2:711–719
Castanotto D, Rossi JJ (2009) The promises and pitfalls of RNA-interference-based therapeutics.
Nature 457:426–433
Check E (2005) Gene therapy put on hold as third child develops cancer. Nature 433:561
Chen XQ, Bonnefoi H, Pelte MF et al (2000) Telomerase RNA as a detection marker in the serum
of breast cancer patients. Clin Cancer Res 6:3823–3826
Chen QR, Zhang L, Stass SA et al (2001) Branched co-polymers of histidine and lysine are
efficient carriers of plasmids. Nucleic Acids Res 29:1334–1340
Chiu YL, Ali A, Chu CY et al (2004) Visualizing a correlation between siRNA localization,
cellular uptake, and RNAi in living cells. Chem Biol 11:1165–1175
Console S, Marty C, Garcia-Echeverria C et al (2003) Antennapedia and HIV transactivator of
transcription (TAT) “protein transduction domains” promote endocytosis of high molecular
weight cargo upon binding to cell surface glycosaminoglycans. J Biol Chem 278:35109–35114
Crombez L, Morris MC, Deshayes S et al (2008) Peptide-based nanoparticle for ex vivo and
in vivo drug delivery. Curr Pharm Des 14:3656–3665
Crombez L, Aldrian-Herrada G, Konate K et al (2009a) A new potent secondary amphipathic cell-
penetrating peptide for siRNA delivery into mammalian cells. Mol Ther 17:95–103
Crombez L, Morris MC, Dufort S et al (2009b) Targeting cyclin B1 through peptide-based

delivery of siRNA prevents tumour growth. Nucleic Acids Res 37(14):4559–4569
Davidson TJ, Harel S, Arboleda VA et al (2004) Highly efficient small interfering RNA delivery to
primary mammalian neurons induces MicroRNA-like effects before mRNA degradation. J
Neurosci 24:10040–10046
De Coupade C, Fittipaldi A, Chagnas V et al (2005) Novel human-derived cell-penetrating
peptides for specific subcellular delivery of therapeutic biomolecules. Biochem J 390:407–418
de Diesbach P, Berens C, N’Kuli F et al (2000) Identification, purification and partial character-
isation of an oligonucleotide receptor in membranes of HepG2 cells. Nucleic Acids Res
28:868–874
de Fougerolles A, Vornlocher HP, Maraganore J et al (2007) Interfering with disease: a progress
report on siRNA-based therapeutics. Nat Rev Drug Discov 6:443–453
Derossi D, Joliot AH, Chassaing G et al (1994) The third helix of the Antennapedia homeodomain
translocates through biological membranes. J Biol Chem 269:10444–10450
Derossi D, Calvet S, Trembleau A et al (1996) Cell internalization of the third helix of the
Antennapedia homeodomain is receptor-independent. J Biol Chem 271:18188–18193
Deshayes S, Plenat T, Aldrian-Herrada G et al (2004) Primary amphipathic cell-penetrating
peptides: structural requirements and interactions with model membranes. Biochemistry
43:7698–7706
Deshayes S, Morris M, Heitz F et al (2008) Delivery of proteins and nucleic acids using a non-
covalent peptide-based strategy. Adv Drug Deliv Rev 60:537–547
Detzer A, Overhoff M, Mescalchin A et al (2008) Phosphorothioate-stimulated cellular uptake of
siRNA: a cell culture model for mechanistic studies. Curr Pharm Des 14:3666–3673
Detzer A, Overhoff M, W€ unsche W et al (2009) Increased RNAi is related to intracellular release
of siRNA via a covalently attached signal peptide. RNA 15:627–636
Doerfler W (1995) Uptake of foreign DNA by mammalian cells via the gastrointestinal tract in
mice: methylation of foreign DNA – a cellular defense mechanism. Curr Top Microbiol
Immunol 197:209–224
Doerfler W, Hohlweg U, Muller K et al (2001) Foreign DNA integration – perturbations of the
genome – oncogenesis. Ann N Y Acad Sci 945:276–288
Edenhofer F (2008) Protein transduction revisited: novel insights into the mechanism underlying
intracellular delivery of proteins. Curr Pharm Des 14:3628–3636
Eguchi A, Meade BR, Chang YC et al (2009) Efficient siRNA delivery into primary cells by a
peptide transduction domain-dsRNA binding domain fusion protein. Nat Biotechnol
27:567–571
El-Andaloussi S, Johansson H, Magnusdottir A et al (2005) TP10, a delivery vector for decoy
oligonucleotides targeting the Myc protein. J Control Release 110:189–201
El-Andaloussi S, Johansson HJ, Holm T et al (2007) A novel cell-penetrating peptide, M918, for
efficient delivery of proteins and peptide nucleic acids. Mol Ther 15:1820–1826
Elbashir SM, Harborth J, Lendeckel W et al (2001a) Duplexes of 21-nucleotide RNAs mediate
RNA interference in cultured mammalian cells. Nature 411:494–498
Elbashir SM, Lendeckel W, Tuschl T (2001b) RNA interference is mediated by 21- and 22-
nucleotide RNAs. Genes Dev 15:188–200
Ellington AD, Szostak JW (1990) In vitro selection of RNA molecules that bind specific ligands.
Nature 346:818–822
Endoh T, Sisido M, Ohtsuki T (2007) Photo inducible RNA interference using cell permeable
protein carrier. Nucleic Acids Symp Ser (Oxf) 51:127–128
Endoh T, Sisido M, Ohtsuki T (2008) Cellular siRNA delivery mediated by a cell-permeant RNA-
binding protein and photoinduced RNA interference. Bioconjug Chem 19:1017–1024
Endoh T, Ohtsuki T (2009) Cellular siRNA delivery using cell-penetrating peptides modified for
endosomal escape. Adv Drug Deliv Rev 61:704–709
Epand RM (2003) Fusion peptides and the mechanism of viral fusion. Biochim Biophys Acta
1614:116–121
Fattal E, Barratt G (2009) Nanotechnologies and controlled release systems for the delivery of
antisense oligonucleotides and small interfering RNA. Br J Pharmacol 157:179–194
Fawell S, Seery J, Daikh Y et al (1994) Tat-mediated delivery of heterologous proteins into cells.
Proc Natl Acad Sci USA 91:664–668
Filleur S, Courtin A, Ait-Si-Ali S et al (2003) SiRNA-mediated inhibition of vascular endothelial
growth factor severely limits tumor resistance to antiangiogenic thrombospondin-1 and slows
tumor vascularization and growth. Cancer Res 63:3919–3922
Fittipaldi A, Ferrari A, Zoppe M et al (2003) Cell membrane lipid rafts mediate caveolar
endocytosis of HIV-1 Tat fusion proteins. J Biol Chem 278:34141–34149
Foged C, Nielsen HM (2008) Cell-penetrating peptides for drug delivery across membrane
barriers. Expert Opin Drug Deliv 5:105–117
Folini M, Bandiera R, Millo E et al (2007) Photochemically enhanced delivery of a cell-
penetrating peptide nucleic acid conjugate targeting human telomerase reverse transcriptase:
effects on telomere status and proliferative potential of human prostate cancer cells. Cell Prolif
40:905–920
Frankel AD, Pabo CO (1988) Cellular uptake of the tat protein from human immunodeficiency
virus. Cell 55:1189–1193
Futaki S, Suzuki T, Ohashi W et al (2001) Arginine-rich peptides. An abundant source of
membrane-permeable peptides having potential as carriers for intracellular protein delivery.
J Biol Chem 276:5836–5840
Futaki S, Masui Y, Nakase I et al (2005) Unique features of a pH-sensitive fusogenic peptide that
improves the transfection efficiency of cationic liposomes. J Gene Med 7:1450–1458
Futaki S (2006) Oligoarginine vectors for intracellular delivery: design and cellular-uptake
mechanisms. Biopolymers 84:241–249
Garcia-Olmo D, Garcia-Olmo DC, Ontanon J et al (2000) Horizontal transfer of DNA and the
“genometastasis hypothesis”. Blood 95:724–725
Green M, Loewenstein PM (1988) Autonomous functional domains of chemically synthesized
human immunodeficiency virus tat trans-activator protein. Cell 55:1179–1188
Grimm D, Streetz KL, Jopling CL et al (2006) Fatality in mice due to oversaturation of cellular
microRNA/short hairpin RNA pathways. Nature 441:537–541
Hacein-Bey-Abina S, Von Kalle C, Schmidt M et al (2003) LMO2-associated clonal T cell
proliferation in two patients after gene therapy for SCID-X1. Science 302:415–419
11:599–606
Hammond SM, Boettcher S, Caudy AA et al (2001) Argonaute2, a link between genetic and
biochemical analyses of RNAi. Science 293:1146–1150
Haque ME, Koppaka V, Axelsen PH et al (2005) Properties and structures of the influenza and HIV
fusion peptides on lipid membranes: implications for a role in fusion. Biophys J 89:3183–3194
Heitz F, Morris MC, Divita G (2009) Twenty years of cell-penetrating peptides: from molecular
mechanisms to therapeutics. Br J Pharmacol 157:195–206
Hu Y, Atukorale PU, Lu JJ et al (2009) Cytosolic delivery mediated via electrostatic surface
binding of protein, virus, or siRNA cargos to pH-responsive core-shell gel particles. Bioma-
cromolecules 10:756–765
Hutvagner G, Simard MJ (2008) Argonaute proteins: key players in RNA silencing. Nat Rev Mol
Cell Biol 9:22–32
Hyndman L, Lemoine JL, Huang L et al (2004) HIV-1 Tat protein transduction domain peptide
facilitates gene transfer in combination with cationic liposomes. J Control Release 99:435–444
Jeong JH, Mok H, Oh YK et al (2009) siRNA conjugate delivery systems. Bioconjug Chem
20:5–14
Jinek M, Doudna JA (2009) A three-dimensional view of the molecular machinery of RNA
interference. Nature 457:405–412
Johnson LN, Cashman SM, Kumar-Singh R (2008) Cell-penetrating peptide for enhanced delivery
of nucleic acids and drugs to ocular tissues including retina and cornea. Mol Ther 16:107–114
Joliot AH, Triller A, Volovitch M et al (1991) Alpha-2, 8-Polysialic acid is the neuronal surface
receptor of antennapedia homeobox peptide. New Biol 3:1121–1134
Kale AA, Torchilin VP (2007a) “Smart” drug carriers: PEGylated TATp-modified pH-sensitive
liposomes. J Liposome Res 17:197–203
Kale AA, Torchilin VP (2007b) Design, synthesis, and characterization of pH-sensitive PEG-PE
conjugates for stimuli-sensitive pharmaceutical nanocarriers: the effect of substitutes at the
hydrazone linkage on the ph stability of PEG-PE conjugates. Bioconjug Chem 18:363–370
Kang H, Delong R, Fisher MH et al (2005) Tat-conjugated PAMAM dendrimers as delivery agents
for antisense and siRNA oligonucleotides. Pharm Res 22:2099–2106
Khalil IA, Kogure K, Futaki S et al (2007) Octaarginine-modified multifunctional envelope-type
nanoparticles for gene delivery. Gene Ther 14:682–689
Kilk K, El-Andaloussi S, J€arver P et al (2005) Evaluation of transportan 10 in PEI mediated
plasmid delivery assay. J Control Release 103:511–523
Kim WJ, Christensen LV, Jo S et al (2006) Cholesteryl oligoarginine delivering vascular endothe-
lial growth factor siRNA effectively inhibits tumor growth in colon adenocarcinoma. Mol Ther
14:343–350
Kleinman ME, Yamada K, Takeda A et al (2008) Sequence- and target-independent angiogenesis
suppression by siRNA via TLR3. Nature 452:591–597
Kogure K, Akita H, Yamada Y et al (2008) Multifunctional envelope-type nano device (MEND) as
a non-viral gene delivery system. Adv Drug Deliv Rev 60:559–571
Kootstra NA, Verma IM (2003) Gene therapy with viral vectors. Annu Rev Pharmacol Toxicol
43:413–439
Kopreski MS, Benko FA, Kwak LW et al (1999) Detection of tumor messenger RNA in the serum
of patients with malignant melanoma. Clin Cancer Res 5:1961–1965
Kumar P, Wu H, McBride JL et al (2007) Transvascular delivery of small interfering RNA to the
central nervous system. Nature 448:39–43
Kwon EJ, Bergen JM, Pun SH (2008) Application of an HIV gp41-derived peptide for enhanced
intracellular trafficking of synthetic gene and siRNA delivery vehicles. Bioconjug Chem
19:920–927
Laktionov PP, Dazard JE, Vives E et al (1999) Characterisation of membrane oligonucleotide-
binding proteins and oligonucleotide uptake in keratinocytes. Nucleic Acids Res 27:
2315–2324
Langel U € (ed) (2006) Handbook of cell-penetrating peptides. CRC Press, Boca Raton
Laufer SD, Restle T (2008) Peptide-mediated cellular delivery of oligonucleotide-based therapeu-
tics in vitro: quantitative evaluation of overall efficacy employing easy to handle reporter
systems. Curr Pharm Des 14:3637–3655
Lehmann MJ, Sczakiel G (2005) Spontaneous uptake of biologically active recombinant DNA by
mammalian cells via a selected DNA segment. Gene Ther 12:446–451
Leng Q, Scaria P, Zhu J et al (2005) Highly branched HK peptides are effective carriers of siRNA.
J Gene Med 7:977–986
Leung RK, Whittaker PA (2005) RNA interference: from gene silencing to gene-specific thera-
peutics. Pharmacol Ther 107:222–239
Lindgren M, H€allbrink M, Prochiantz A et al (2000) Cell-penetrating peptides. Trends Pharmacol
Sci 21:99–103
Liu Z, Li M, Cui D et al (2005) Macro-branched cell-penetrating peptide design for gene delivery.
J Control Release 102:699–710
Lundberg M, Johansson M (2001) Is VP22 nuclear homing an artifact? Nat Biotechnol 19:713–714
Lundberg M, Johansson M (2002) Positively charged DNA-binding proteins cause apparent cell
membrane translocation. Biochem Biophys Res Commun 291:367–371
Lundberg P, El-Andaloussi S, Sutlu T et al (2007) Delivery of short interfering RNA using
endosomolytic cell-penetrating peptides. FASEB J 21:2664–2671
Maiolo JR, Ferrer M, Ottinger EA (2005) Effects of cargo molecules on the cellular uptake of
arginine-rich cell-penetrating peptides. Biochim Biophys Acta 1712:161–172
Mano M, Teodosio C, Paiva A et al (2005) On the mechanisms of the internalization of S4(13)-PV
cell-penetrating peptide. Biochem J 390:603–612
McCaffrey AP, Meuse L, Pham TT et al (2002) RNA interference in adult mice. Nature 418:38–39
McNamara JO, Andrechek ER, Wang Y et al (2006) Cell type-specific delivery of siRNAs with
aptamer-siRNA chimeras. Nat Biotechnol 24:1005–1015
Meade BR, Dowdy SF (2008) Enhancing the cellular uptake of siRNA duplexes following
noncovalent packaging with protein transduction domain peptides. Adv Drug Deliv Rev
60:530–536
Mescalchin A, Detzer A, Wecke M et al (2007) Cellular uptake and intracellular release are major
obstacles to the therapeutic application of siRNA: novel options by phosphorothioate-stimu-
lated delivery. Expert Opin Biol Ther 7:1531–1538
Meyer M, Philipp A, Oskuee R et al (2008) Breathing life into polycations: functionalization with
pH-responsive endosomolytic peptides and polyethylene glycol enables siRNA delivery. J Am
Chem Soc 130:3272–3273
Meyer M, Dohmen C, Philipp A et al (2009) Synthesis and biological evaluation of a biorespon-
sive and endosomolytic siRNA-polymer conjugate. Mol Pharm 6:752–762
Michiue H, Tomizawa K, Wei FY et al (2005) The NH2 terminus of influenza virus hemaggluti-
nin-2 subunit peptides enhances the antitumor potency of polyarginine-mediated p53 protein
transduction. J Biol Chem 280:8285–8289
Midoux P, Monsigny M (1999) Efficient gene transfer by histidylated polylysine/pDNA com-
plexes. Bioconjug Chem 10:406–411
Morris MC, Vidal P, Chaloin L et al (1997) A new peptide vector for efficient delivery of
oligonucleotides into mammalian cells. Nucleic Acids Res 25:2730–2736
Morris MC, Deshayes S, Heitz F et al (2008) Cell-penetrating peptides: from molecular mechan-
isms to therapeutics. Biol Cell 100:201–217
Moschos SA, Jones SW, Perry MM et al (2007) Lung delivery studies using siRNA conjugated to
TAT(48–60) and penetratin reveal peptide induced reduction in gene expression and induction
of innate immunity. Bioconjug Chem 18:1450–1459
Muratovska A, Eccles MR (2004) Conjugate for efficient delivery of short interfering RNA
(siRNA) into mammalian cells. FEBS Lett 558:63–68
Nakamura Y, Kogure K, Futaki S et al (2007) Octaarginine-modified multifunctional envelope-
type nano device for siRNA. J Control Release 119:360–367
Ng EK, Tsui NB, Lam NY et al (2002) Presence of filterable and nonfilterable mRNA in the
plasma of cancer patients and healthy individuals. Clin Chem 48:1212–1217
Oehlke J, Scheller A, Wiesner B et al (1998) Cellular uptake of an alpha-helical amphipathic
model peptide with the potential to deliver polar compounds into the cell interior non-
endocytically. Biochim Biophys Acta 1414:127–139
Ogris M, Carlisle RC, Bettinger T et al (2001) Melittin enables efficient vesicular escape
and enhanced nuclear access of nonviral gene delivery vectors. J Biol Chem 276:
47550–47555
Oliveira S, Fretz MM, Hogset A et al (2007a) Photochemical internalization enhances silencing of
epidermal growth factor receptor through improved endosomal escape of siRNA. Biochim
Biophys Acta 1768:1211–1217
Oliveira S, van Rooy I, Kranenburg O et al (2007b) Fusogenic peptides enhance endosomal escape
improving siRNA-induced silencing of oncogenes. Int J Pharm 331:211–214
Oliveira S, Hogset A, Storm G et al (2008) Delivery of siRNA to the target cell cytoplasm:
photochemical internalization facilitates endosomal escape and improves silencing efficiency,
in vitro and in vivo. Curr Pharm Des 14:3686–3697
Overhoff M, W€unsche W, Sczakiel G (2004) Quantitative detection of siRNA and single-stranded
oligonucleotides: relationship between uptake and biological activity of siRNA. Nucleic Acids
Res 32:e170
Overhoff M, Sczakiel G (2005) Phosphorothioate-stimulated uptake of short interfering RNA by

human cells. EMBO Rep 6:1176–1181
Paddison PJ, Caudy AA, Bernstein E et al (2002) Short hairpin RNAs (shRNAs) induce sequence-
specific silencing in mammalian cells. Genes Dev 16:948–958
Patel LN, Zaro JL, Shen WC (2007) Cell penetrating peptides: intracellular pathways and
pharmaceutical perspectives. Pharm Res 24:1977–1992
Peer D, Zhu P, Carman CV et al (2007) Selective gene silencing in activated leukocytes by
targeting siRNAs to the integrin lymphocyte function-associated antigen-1. Proc Natl Acad Sci
USA 104:4095–4100
Pichon C, Monsigny M, Roche AC (1999) Intracellular localization of oligonucleotides: influence
of fixative protocols. Antisense Nucleic Acid Drug Dev 9:89–93
Plank C, Zauner W, Wagner E (1998) Application of membrane-active peptides for drug and gene
delivery across cellular membranes. Adv Drug Deliv Rev 34:21–35
Pooga M, H€allbrink M, Zorko M et al (1998) Cell penetration by transportan. FASEB J 12:67–77
Preuss M, Tecle M, Shah I et al (2003) Comparison between the interactions of adenovirus-derived
peptides with plasmid DNA and their role in gene delivery mediated by liposome-peptide-
DNA virus-like nanoparticles. Org Biomol Chem 1:2430–2438
Rahbek UL, Howard KA, Oupický D et al (2008) Intracellular siRNA and precursor miRNA
trafficking using bioresponsive copolypeptides. J Gene Med 10:81–93
Rana TM (2007) Illuminating the silence: understanding the structure and function of small RNAs.
Nat Rev Mol Cell Biol 8:23–36
Rao DD, Vorhies JS, Senzer N et al (2009) siRNA vs. shRNA: similarities and differences. Adv
Drug Deliv Rev 61:746–759
Raper SE, Yudkoff M, Chirmule N et al (2002) A pilot study of in vivo liver-directed gene transfer
with an adenoviral vector in partial ornithine transcarbamylase deficiency. Hum Gene Ther
13:163–175
Raper SE, Chirmule N, Lee FS et al (2003) Fatal systemic inflammatory response syndrome in a
ornithine transcarbamylase deficient patient following adenoviral gene transfer. Mol Genet
Metab 80:148–158
Read ML, Bremner KH, Oupicky D et al (2003) Vectors based on reducible polycations facilitate
intracellular release of nucleic acids. J Gene Med 5:232–245
Read ML, Singh S, Ahmed Z et al (2005) A versatile reducible polycation-based system for
efficient delivery of a broad range of nucleic acids. Nucleic Acids Res 33:e86
Reddi KK, Holland JF (1976) Elevated serum ribonuclease in patients with pancreatic cancer. Proc
Natl Acad Sci USA 73:2308–2310
Richard JP, Melikov K, Vives E et al (2003) Cell-penetrating peptides. A reevaluation of the
mechanism of cellular uptake. J Biol Chem 278:585–590
Richard JP, Melikov K, Brooks H et al (2005) Cellular uptake of unconjugated TAT peptide
involves clathrin-dependent endocytosis and heparan sulfate receptors. J Biol Chem
280:15300–15306
Ritter W, Plank C, Lausier J et al (2003) A novel transfecting peptide comprising a tetrameric
nuclear localization sequence. J Mol Med 81:708–717
Robertson DL, Joyce GF (1990) Selection in vitro of an RNA enzyme that specifically cleaves
single-stranded DNA. Nature 344:467–468
Rothbard JB, Jessop TC, Lewis RS et al (2004) Role of membrane potential and hydrogen bonding
in the mechanism of translocation of guanidinium-rich peptides into cells. J Am Chem Soc
126:9506–9507
Sakurai Y, Hatakeyama H, Akita H et al (2009) Efficient short interference RNA delivery to tumor
cells using a combination of octaarginine, GALA and tumor-specific, cleavable polyethylene
glycol system. Biol Pharm Bull 32:928–932
Schubbert R, Lettmann C, Doerfler W (1994) Ingested foreign (phage M13) DNA survives
transiently in the gastrointestinal tract and enters the bloodstream of mice. Mol Gen Genet
242:495–504
Schubbert R, Renz D, Schmitz B et al (1997) Foreign (M13) DNA ingested by mice reaches
peripheral leukocytes, spleen, and liver via the intestinal wall mucosa and can be covalently
linked to mouse DNA. Proc Natl Acad Sci USA 94:961–966
Schwarze SR, Ho A, Vocero-Akbani A et al (1999) In vivo protein transduction: delivery of a
biologically active protein into the mouse. Science 285:1569–1572
Schwarze SR, Dowdy SF (2000) In vivo protein transduction: intracellular delivery of biologically
active proteins, compounds and DNA. Trends Pharmacol Sci 21:45–48
Simeoni F, Morris MC, Heitz F et al (2003) Insight into the mechanism of the peptide-based gene
delivery system MPG: implications for delivery of siRNA into mammalian cells. Nucleic
Acids Res 31:2717–2724
Snøve O Jr, Rossi JJ (2006) Toxicity in mice expressing short hairpin RNAs gives new insight into
RNAi. Genome Biol 7:231
Song E, Lee SK, Wang J et al (2003) RNA interference targeting Fas protects mice from fulminant
hepatitis. Nat Med 9:347–351
Song E, Zhu P, Lee SK et al (2005) Antibody mediated in vivo delivery of small interfering RNAs
via cell-surface receptors. Nat Biotechnol 23:709–717
Song JJ, Smith SK, Hannon GJ et al (2004) Crystal structure of Argonaute and its implications for
RISC slicer activity. Science 305:1434–1437
Soomets U, Lindgren M, Gallet X et al (2000) Deletion analogues of transportan. Biochim
Biophys Acta 1467:165–176
Sorensen DR, Leirdal M, Sioud M (2003) Gene silencing by systemic delivery of synthetic
siRNAs in adult mice. J Mol Biol 327:761–766
Soundara Manickam D, Oupický D (2006) Multiblock reducible copolypeptides containing
histidine-rich and nuclear localization sequences for gene delivery. Bioconjug Chem 17:
1395–1403
Soutschek J, Akinc A, Bramlage B et al (2004) Therapeutic silencing of an endogenous gene by
systemic administration of modified siRNAs. Nature 432:173–178
Takizawa PA, Yucel JK, Veit B et al (1993) Complete vesiculation of Golgi membranes and
inhibition of protein transport by a novel sea sponge metabolite, ilimaquinone. Cell 73:1079–1090
Tönges L, Lingor P, Egle R et al (2006) Stearylated octaarginine and artificial virus-like particles
for transfection of siRNA into primary rat neurons. RNA 12:1431–1438
Torchilin VP, Rammohan R, Weissig V et al (2001) TAT peptide on the surface of liposomes
affords their efficient intracellular delivery even at low temperature and in the presence of
metabolic inhibitors. Proc Natl Acad Sci USA 98:8786–8791
Torchilin VP, Levchenko TS, Rammohan R et al (2003) Cell transfection in vitro and in vivo with
nontoxic TAT peptide-liposome-DNA complexes. Proc Natl Acad Sci USA 100:1972–1977
Tu Y, Kim JS (2008) A fusogenic segment of glycoprotein H from herpes simplex virus enhances
transfection efficiency of cationic liposomes. J Gene Med 10:646–654
Tuerk C, Gold L (1990) Systematic evolution of ligands by exponential enrichment: RNA ligands
to bacteriophage T4 DNA polymerase. Science 249:505–510
Turner JJ, Arzumanov AA, Gait MJ (2005) Synthesis, cellular uptake and HIV-1 Tat-dependent
trans-activation inhibition activity of oligonucleotide analogues disulphide-conjugated to cell-
penetrating peptides. Nucleic Acids Res 33:27–42
Tyagi M, Rusnati M, Presta M et al (2001) Internalization of HIV-1 tat requires cell surface
heparan sulfate proteoglycans. J Biol Chem 276:3254–3261
Veldhoen S, Laufer SD, Trampe A et al (2006) Cellular delivery of small interfering RNA by a
non-covalently attached cell-penetrating peptide: quantitative analysis of uptake and biological
effect. Nucleic Acids Res 34:6561–6573
Veldhoen S, Laufer SD, Restle T (2008) Recent developments in peptide-based nucleic acid
delivery. Int J Mol Sci 9:1276–1320
Verma IM, Weitzman MD (2005) Gene therapy: twenty-first century medicine. Annu Rev
Biochem 74:711–738
Vives E, Brodin P, Lebleu B (1997) A truncated HIV-1 Tat protein basic domain rapidly
translocates through the plasma membrane and accumulates in the cell nucleus. J Biol Chem
272:16010–16017
Wadia JS, Stan RV, Dowdy SF (2004) Transducible TAT-HA fusogenic peptide enhances escape
of TAT-fusion proteins after lipid raft macropinocytosis. Nat Med 10:310–315
Wang HJ, Benlimame N, Nabi I (1997) The AMF-R tubule is a smooth ilimaquinone-sensitive
subdomain of the endoplasmic reticulum. J Cell Sci 110(Pt 24):3043–3053
Wang Y, Sheng G, Juranek S et al (2008a) Structure of the guide-strand-containing argonaute
silencing complex. Nature 456:209–213
Wang Y, Juranek S, Li H et al (2008b) Structure of an argonaute silencing complex with a seed-
containing guide DNA and target RNA duplex. Nature 456:921–926
Wesche-Soldato DE, Chung CS, Lomas-Neira J et al (2005) In vivo delivery of caspase-8 or Fas
siRNA improves the survival of septic mice. Blood 106:2295–2301
Whitehead KA, Langer R, Anderson DG (2009) Knocking down barriers: advances in siRNA
delivery. Nat Rev Drug Discov 8:129–138
Xie FY, Woodle MC, Lu PY (2006) Harnessing in vivo siRNA delivery for drug discovery and
therapeutic development. Drug Discov Today 11:67–73
Yu JY, DeRuiter SL, Turner DL (2002) RNA interference by expression of short-interfering RNAs
and hairpin RNAs in mammalian cells. Proc Natl Acad Sci USA 99:6047–6052
Zatsepin TS, Turner JJ, Oretskaya TS et al (2005) Conjugates of oligonucleotides and analogues
with cell penetrating peptides as gene silencing agents. Curr Pharm Des 11:3639–3654
Zimmermann TS, Lee AC, Akinc A et al (2006) RNAi-mediated gene silencing in non-human
primates. Nature 441:111–114
Zorko M, Langel U € (2005) Cell-penetrating peptides: mechanism and kinetics of cargo delivery.
Adv Drug Deliv Rev 57:529–545
RNAi Suppression and Its Application
Xiaoping Yi and Rui Lu
Contents
1 RNA Interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2 RNAi-Directed Viral Immunity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3 Identification and Function Characterization of Viral RNAi Suppressors . . . . . . . . . . . . . . . . . . 68
3.1 Agrobacterium-Mediated Transient Suppression Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.2 Reversal of Transgene-Induced Gene Silencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.3 Grafting Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.4 Replication Rescue of Mutant Viruses Defective in RNAi Suppression . . . . . . . . . . . . . . 70
4 Function Mechanism of Viral RNAi Suppressors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.1 Viral Suppressors that Bind Viral dsRNA to Interfere with viRNA Biogenesis . . . . . . 73
4.2 Viral Suppressors that Target Virus-Derived siRNA for RNAi Suppression . . . . . . . . . 75
4.3 Viral Suppressors that Target RNAi Effectors for Suppression . . . . . . . . . . . . . . . . . . . . . . . 77
4.4 Viral Suppressors that Suppress Systemic RNAi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5 RNAi Suppressors of Nonviral Origin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.1 Suppression of RNAi by Bacterial Pathogens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.2 Cellular RNAi Suppressors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6 Biotechnological Application of RNAi Suppressors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.1 Enhance Gene Expression for Rapid Function Analysis and Mass
Production of Valuable Protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.2 Serve as Molecular Tools to Probe Various RNAi-Directed Functions . . . . . . . . . . . . . . . 84
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Abstract In eukaryotes, RNA interference (RNAi) is a gene silencing mechanism

mediated by small RNAs (sRNAs), currently classified as small interfering RNA
(siRNA), microRNA (miRNA), and piwi-interacting RNA (piRNA). These small
RNAs are produced in Dicer (a ribonuclease III enzyme)-dependent (siRNA and
miRNA) or Dicer-independent (piRNA) manner and are effected by a group of
X. Yi and R. Lu (*)
Department of Biological Sciences, Louisiana State University, Baton Rouge, LA 70803, USA
e-mail: ruilu@lsu.edu
60 X. Yi and R. Lu
Argonaut (AGO) proteins. These small RNAs mediate silencing of target genes
with complementary sequence at transcriptional or posttranscriptional level
thereby to control a wide variety of biological functions. In worms and plants,
RNA-dependent RNA polymerases (RdRPs) amplify RNAi by converting AGO
cleavage products into dsRNAs for the generation of secondary siRNAs. One of the
well characterized functions of RNAi is antiviral, which has been shown to serve as
major viral innate immunity in fungi, plants, and invertebrates. Typically, RNAi-
directed viral immunity (RDVI) is initiated with Dicer processing of viral dsRNAs,
usually the replication intermediates, into siRNAs. These virus-derived
siRNAs (viRNA) will then be used as sequence guide for target viral RNA
destruction. Host-encoded miRNAs also contribute to viral control in mammal or
bacterial control in plant. As a counterdefensive mechanism, many viruses and
some bacteria are found to encode RNAi suppressors, previously known as patho-
genicity factors. These RNAi antagonists target key components of RNAi for
suppression, which eventually leads to defects in viRNA biogenesis or function.
Since transgene expression in plants and invertebrates is often targeted by RNAi for
suppression but can be reversed by various RNAi suppressors, codelivery of a VSR
has been used to facilitate the isolation and biochemical characterization of a broad
range of proteins of interests. RNAi suppressors can also be used as genetic tool for
the study of biological functions controlled by certain class of endogenous sRNA
(siRNA or miRNA). This is because, when expressed as transgenes, some RNAi
suppressors can specifically target and interfere with the biogenesis or function of
certain class of endogenous sRNA but not the other.
Abbreviations

sRNA small RNA
siRNA small interfering RNA
miRNA microRNA
piRNA PIWI-interacting RNA
RdRP RNA-dependent RNA polymerase
AGO Argonaut
viRNA virus-derived siRNA
C. elegans Caenorhabditis elegans
PTGS posttranscriptional gene silencing
S. pombe Schizosaccharomyces pombe
endo-siRNA endogenous siRNA
RDVI RNAi-directed viral immunity
TEV Tobacco etch virus
PVY Potato virus Y
CP coat protein
RNAi Suppression and Its Application 61
N. benthamiana Nicotiana benthamiana

PDS phytoene desaturase
N. clevelandii Nicotiana clevelandii
PVX Potato virus X
VSR viral suppressor of RNAi
HC-Pro Helper component-proteinase
CMV Cucumber mosaic virus
FHV Flock house virus
PFV Primate foamy virus
CTV Citrus tristeza virus
GFP green fluorescent protein
TYLCV Tomato yellow leaf curl geminivirus
TBSV Tomato bushy stunt virus
CNV Cucumber necrosis virus
CRV Cymbidium ringspot virus
SPCSV Sweet potato chlorotic stunt virus
ssRNA single-stranded RNA
SPFMV Sweet potato feathery mottle virus
TMV Tobacco mosaic virus
ORMV Oilseed rape mosaic tobamovirus
TCV Turnip crinkle virus
BWYV Beet western yellows virus
SKP1 S-phase kinase-related protein 1
SCF Skp1-Cul1/Cdc53,-F-box protein
SYLV Sugarcane yellow leaf virus
shRNA short hairpin RNA
A. tumefaciens Agrobacterium tumefaciens
T-DNA transfer DNA
P. syringae Pseudomonas syringae
PAMP pathogen-associated molecular pattern
rgs-CaM regulator of gene silencing-calmodulin-like protein
SDN1 small RNA degrading nuclease 1
eri-1 enhanced RNAi-1
CHS chalcone synthase
aa amino acids
nt nucleotide
bp base pair
1 RNA Interference
Potent gene silencing triggered by double-stranded RNA (dsRNA) was first demon-
strated in an experimental setup in which delivery of artificial dsRNA into
nematode worm Caenorhabditis elegans (C. elegans) by microinjection initiated
62 X. Yi and R. Lu
Fig. 1 Silencing of chalcone synthase by a posttranscriptional gene silencing in petunia plants.

Left panel: Comparison of the flower pigmentation in between nontransgenic parental inbred V26
(left) and V26 plants transgenic for chalcone synthase (CHS). Right panel: RNase protection assay
of the chalcone synthase transcripts. (a) RNase protections of total RNA isolated from three
separate 40-mm-long violet revertant corollas. (b) RNase protections of RNA isolated from three
separate 40-mm-long white corollas. E indicates the position of the protected fragment for the
endogenous CHS transcripts and I indicates the position of the protected fragment for the
introduced CHS transcripts [adapted, with permission, from American Society of Plant Biologists:
The Plant Cell (Napoli et al. 1990), copyright 1990]
sequence-specific silencing of homologous target gene (Fire et al. 1998). This

phenomenon is often termed RNA interference (RNAi) to reflect the fact that
both the trigger and target are RNA molecules. Soon after this discovery, it
becomes clear that the mechanism involved in RNAi is also responsible for
posttranscriptional gene silencing (PTGS) that was observed at an earlier time in
plants. In 1990, Dr. Jorgensen and colleagues demonstrated that introduction of a
plant transgene triggered a posttranscriptional mechanism that silenced both the
transgene and endogenous gene (Fig. 1) (Napoli et al. 1990). Later on, in addition to
worm, PTGS phenomena were also observed in other eukaryotic organisms such as
fungus (Cogoni et al. 1996), fruitfly (Kennerdell and Carthew 1998), and mammal
(Elbashir et al. 2001). In plants, replicating virus can also trigger PTGS presumably
by the dsRNA form of viral replication intermediates (Kumagai et al. 1995; Ruiz
et al. 1998; Baulcombe 1999). Because of its enormous potential in biomedical and
biological research and application, mainly owing to its sequence-specific mode of
action, RNAi-related phenomena have been receiving extensive study aimed at
deciphering the underlying core mechanism.
A hallmark of an ongoing RNAi event is the production of a class of small RNA
species, termed small interfering RNA (siRNA), ranging from 21 to 26 nucleotides
long (Hamilton and Baulcombe 1999; Hammond et al. 2000; Sontheimer 2005). A
ubiquitous enzyme, called Dicer, is responsible for the production of siRNAs in
diverse organisms (Fig. 2) (Bernstein et al. 2001). Dicer belongs to the class 3 of
ribonuclease III, of which all members contain at least one ribonuclease domain
Fig. 2 Schematic
representation of dsRNA-
triggered RNAi. Dicer, Type
III ribonuclease that
processes double-stranded
RNA (dsRNA) into small
interfering RNA (siRNA).
AGO, Argonaut protein that
cleaves target transcript
complementary to siRNA.
RdRP, RNA-dependent RNA
polymerase found in plants
and worms, which amplifies
RNAi by converting AGO
cleavage product into
secondary dsRNA
commonly called “RNase III domain”. Cleavage by RNase III enzymes produces a
characteristic dsRNA structure consisting of a 50 phosphate group and a two base
overhang at the 30 end. Dicer enzymes typically contain two ribonuclease domains,
a dsRNA binding domain, and an N-terminal DExD/H-box helicase domain fol-
lowed by a small domain of unknown function (DUF283) and a PAZ domain. Dicer
specifically recognizes and processes long dsRNA molecules into siRNA duplexes
of discrete size in a sequence-independent manner (Bernstein et al. 2001). Crystal-
lographic study suggests that Dicer functions as a molecular ruler that binds to the
ends of dsRNA through the PAZ domain and cleaves a set distance away (Song
et al. 2004; Macrae et al. 2006).
During RNAi-mediated gene silencing, one strand of the siRNA duplex, usually
the strand with a 50 end which is less stable in base pairing, is incorporated into an
RNA-induced silencing complex (RISC) and provides the sequence specificity
through Watson–Crick base pairing for target RNA destruction or translational
arrest. An Argonaut (AGO) protein can be found in all RISCs and its role is to
bind the small RNA and position it in a conformation that facilitates cognate target
recognition (Fig. 2). AGO proteins are ubiquitous in eukaryotes and some archaea.
The AGO protein superfamily falls into two major clades: the AGO clade found in
fungi, plants, and animals and the PIWI clade that are found only in animals so far.
The number of AGO genes found in different species varies from one [as in the
fission yeast Schizosaccharomyces pombe (S. pombe)] to over two dozens (27 in
C. elegans) (Volpe et al. 2002; Yigit et al. 2006). Although, in some cases, multiple
copies of an AGO gene function redundantly, it is common that AGO genes in an
organism have specialized and nonoverlapping functions. Typically, a PAZ domain
64 X. Yi and R. Lu
at the N-terminal and a PIWI domain at the C-terminal can be found for an AGO
protein. The PAZ domains function to bind the 30 end of the small guide RNA,
whereas some PIWI domain adopts a RNase H-like fold and can hydrolyze target
RNAs using an RNase H-like mechanism (Song et al. 2004). This activity is usually
termed as “slicing” of target RNAs. In plants and worms, RNA-dependent RNA
polymerase (RdRP) amplifies RNAi by converting the cleavage products of initial
RNAi targets into secondary dsRNAs, which are further processed into secondary
siRNAs (Fig. 2). RdRPs are responsible for systemic spreading of a silencing status
in both plants and worms and are believed to prime an antiviral status prior to viral
arrival (Sijen et al. 2001; Baulcombe and Molnar 2004).
Although the RNAi phenomenon was first observed in model plant, function
mechanism studies suggested that the genes involved in RNAi actually are con-
served in almost all eukaryotes. In addition to their response to artificial dsRNAs
delivered through transgene and microinjection, key RNAi genes also play essential
roles in the biogenesis or function of a wide variety of small RNA species derived
from endogenous transcripts and transposons under natural conditions. Based on
their origin and the effector proteins they associate with, these small RNAs can be
classified as microRNA (miRNA), endogenous siRNA (endo-siRNA), and PIWI-
interacting RNA (piRNA) (Lee and Ambros 2001; Girard et al. 2006; Yigit et al.
2006; Ambros and Chen 2007; Aravin et al. 2007; Brennecke et al. 2007).
miRNAs are a class of abundant 20–24 nt RNAs that are processed by Dicer
enzymes from stem–loop precursors found in intergenic regions, introns, and
coding regions. miRNA silences gene expression by guiding RISC to target
mRNA with fully or partly complementary sequences (Carrington and Ambros
2003; Bartel 2004; Voinnet 2009). In plant, miRNAs base pair with target mRNAs
through fully complementary sequences, and the common outcome of repression is
manifested as target mRNA cleavage. Animal miRNAs base pair with target
transcripts through a “seed” sequence, the 2–8 nucleotides at the 50 end. In contrast
to plant miRNA, base-pairing between animal miRNA and target in most cases
leads to translational arrest. It is of interest to note, since the “seed” sequences of
animal miRNAs are relatively short, 7 nt long, a single animal miRNA could, in
principle, potentially target and regulate a large number of protein-coding genes,
making miRNA target prediction and identification very difficult. miRNAs have
received the most extensive studies among these three known classes of endoge-
nous small RNAs and are shown to play essential roles in development, viral and
bacterial resistance, tumorigenesis, and stress response.
Endo-siRNAs are produced as populations through processing of long, perfectly
complementary dsRNA precursors by Dicer enzymes. The dsRNA precursors
might have derived from inverted repeats or convergent transcription of transpo-
sons or might have been produced by RdRPs using aberrant RNAs as templates
(Herr et al. 2005). Endo-siRNAs in many species are found to regulate gene
expression in trans or to suppress transposon mobility (Duchaine et al. 2006;
Ghildiyal et al. 2008; Okamura and Lai 2008; Tam et al. 2008).
piRNAs are a class of animal noncoding small RNAs whose major function is to
suppress the mobility of transposons. Interestingly, unlike miRNAs and endo-
siRNAs, the biogenesis of piRNAs is Dicer-independent. It has been proposed that

piRNAs are generated by a “ping-pong” mechanism in which the target RNA of one
PIWI protein is cleaved and becomes the guide RNA of another PIWI protein
(Aravin et al. 2007). To date, study on the biogenesis and function of these small
RNAs has greatly expanded our knowledge about RNAi-mediated gene regulation
and, accordingly, RNAi has been frequently used as a generic term to describe
various gene silencing mechanisms that act at transcriptional or posttranscriptional
level.
2 RNAi-Directed Viral Immunity
As aforementioned, RNAi controls a wide variety of biological functions through

transcriptional and posttranscriptional mechanisms. One of the first natural func-
tions discovered for RNAi is antiviral. The earliest reports demonstrating an RNA-
directed viral immunity (RDVI) can be traced back to early 1990s. In 1992, Lindbo
and colleagues reported that transgenic tobacco plants expressing untranslatable
sense or antisense forms of coat protein (CP) gene from Tobacco etch virus (TEV)
acquired resistance against TEV. This report clearly showed that the observed viral
resistance is mediated by an RNA molecule and is TEV-specific (Lindbo and
Dougherty 1992). In agreement with this report, transgenic tobacco plants expres-
sing transcripts corresponding to the CP of Potato virus Y (PVY) were found to
become resistant to PVY (Van der Vlugt et al. 1992). In this report, although the
viral transgene contains intact open reading frame, the transgenic lines that are
resistant PVY did not produce detectable viral proteins, suggesting a nonprotein-
mediated viral immunity. Later on, it was found that this type of viral resistance can
be triggered in plants that carry transgene of nonviral origin (English et al. 1996).
These observations together strongly suggested that viral infection in plant can
trigger a novel antiviral defense as long as the invading virus shares sequence
homology with host transgene. However, at the time, it was unclear whether this
transgene mediated viral resistance represents a natural antiviral defense, and little
was known about the trigger of this antiviral immunity.
In an independent research, Kumagai and colleagues found that infection of
Nicotiana benthamiana (N. benthamiana) with a recombinant tobamovirus that
carries part of phytoene desaturase (PDS) caused silencing of endogenous PDS as
revealed in photobleaching phenotype (Kumagai et al. 1995). This study, for the
first time, demonstrated that viral infection in plants can trigger sequence homology-
dependent gene silencing. However, at the time it was unclear whether virus-triggered
gene silencing and the transgene-mediated viral immunity share the same mecha-
nism. In 1997, by examining the antiviral response triggered by an RNA virus
in nontransgenic Nicotiana clevelandii (N. clevelandii) plants, Baulcombe and
colleagues demonstrated that viral infection-triggered gene silencing represents a
natural viral immunity that is responsible for the transgene-mediated antiviral
response reported at earlier time. In this study, nepovirus infection of N. clevelandii
66 X. Yi and R. Lu
plants ended up in a rapid recovery from severe virus disease symptoms and the
“recovered” plant became resistant to secondary infection by the same virus but
remained susceptible to secondary infection by unrelated viruses, such as Potato
virus X (PVX). Importantly, this nepovirus-induced viral resistance can be targeted
against PVX in secondary infection if the PVX genome is modified to carry a piece
of nepovirus sequence (Ratcliff et al. 1997). These observations together suggested
that plant viral infection triggers a natural antiviral response that operates in a
sequence-dependent manner, a characteristic feature of transgene-induced gene
silencing as reported in many studies (Ratcliff et al. 1997). In an independent
research, another group also reported gene silencing-mediated viral immunity
against DNA virus (Covey et al. 1997). These findings together suggested that
homology-dependent gene silencing, a term frequently used to describe RNAi at
that time, represents a novel viral immunity that occurs under natural condition.
RNAi as a natural antiviral defense mechanism was reconfirmed by the discovery
of viral suppressor of RNAi (VSR). Most of VSRs were previously known to be
dispensable for viral replication but responsible for enhanced viral infection and
disease symptom, and thus were often referred to as viral pathogenicity factors. The
first two VSRs identified are the helper component-proteinase (HC-Pro) encoded by
Tobacco etch virus and the 2b protein encoded by Cucumber mosaic virus (CMV)
(Anandalakshmi et al. 1998; Brigneti et al. 1998; Kasschau and Carrington 1998).
When expressed as transgene, HC-Pro reversed RNAi targeted against a transgene
in tobacco plants. HC-Pro is also capable of suppressing RNAi triggered by viral
replication when expressed from recombinant virus (Anandalakshmi et al. 1998;
Brigneti et al. 1998; Kasschau and Carrington 1998) (Fig. 3c, d). Later on, the potent
suppression activity of HC-Pro was ascribed to its ability to bind and interfere with
the stability of virus-derived siRNAs (viRNAs) (Lakatos et al. 2006; Lozsa et al.
2008). The 2b of CMV apparently adopts a different mechanism for RNAi suppres-
sion. When expressed from recombinant PVX, CMV 2b did not reverse the silencing
that has already been established but instead prevented the initiation of gene
silencing at the growing points of the plant (Brigneti et al. 1998) (Fig. 3e, f). Now
we know that this is because CMV 2b is able to suppress systemic silencing (Guo
and Ding 2002). As a result, target gene in the new emerging leaves will remain
nonsilenced because of the lack of systemic silencing signal.
RDVI was also found to be conserved in animal through studying the replication
of Flock house virus (FHV), a member of the nodavirus family, in Drosophila cell
culture (Li et al. 2002). Like what has been found in plant system, infection by FHV
virions results in the rapid accumulation of FHV-specific siRNAs of both plus and
minus polarities in infected cells. Moreover, increased accumulation of FHV RNAs
was observed in the fly cells depleted of AGO2, which is known to be a core
component of the effecter complex RISC. This finding strongly suggested that viral
replication is checked by an RNAi-directed antiviral defense. Most importantly, the
B2 protein of FHV, a suppressor of RNAi in diverse systems, is essential for
accumulation of FHV in wild-type fly cells but becomes dispensable in cells
depleted of AGO2. This perfect complementation unequivocally demonstrated
that the essential role of B2 in FHV infection is to suppress the AGO2-dependent
Fig. 3 Suppression of transgene-induced RNAi by PVY HC-Pro and CMV 2b. (a) Nicotiana
benthamiana plant (line 16c) showing high levels of GFP expression under UV illumination.
(b) 16c plant showing GFP silencing triggered by infiltration of an A. tumefaciens strain carrying
GFP T-DNA. The bright red color is from the chlorophyll fluorescence under UV illumination.
(c) GFP-silenced 16c plant infected with PVY under UV illumination (15 days post-inoculation).
The green fluorescence showing the reversal of GFP silencing by PVY. (d) GFP-silenced 16c plant
infected with recombinant PVX carrying PVY HC-Pro under UV illumination. The green fluore-
scence showing the reversal of GFP silencing by HC-Pro. (e) GFP-silenced 16c plant infected with
CMV (21 days post inoculation). GFP expression was restored in the newly emerging tissue after
systemic CMV infection had been established. (f) GFP-silenced 16c plant infected with recombi-
nant PVX carrying CMV 2b under UV illumination. All images in this figure are adapted, with
permission, from MacMillan Publisher Ltd: EMBO J (Brigneti et al. 1998), copyright 1998
antiviral RNAi against FHV (Li et al. 2002). RNAi suppressors were also isolated
from DNA viruses. These suppressors presumably prevent RNAi from targeting
highly structured viral transcripts produced from read-through transcription of
repeated DNA sequence or convergent transcription of the same DNA sequence
(Voinnet et al. 1999; Guillaume and Olivier 2004; Fukunaga and Doudna 2009).
Recently, induction and suppression of antiviral RNAi has also been documented in
fungi and other invertebrates such as shrimp, mosquito, silk moth, and tick cells.
These findings together establish RDVI as a viral innate immunity in diverse
organism species (Isobe et al. 2004; Li et al. 2004; Garcia et al. 2006; Segers
et al. 2007; Su et al. 2008).
68 X. Yi and R. Lu
While it remains unclear whether siRNA-mediated RNAi serves as viral control

in mammal, some mammalian viruses were found to encode proteins capable of
RNAi suppression in heterologous system. Typical examples are the NS1 from
influenza virus, the E3L from Vaccinia virus, the VP35 from Ebola virus, the Tat
from HIV, and the Tas from Primate foamy virus (PFV) (Li et al. 2004; Bennasser
et al. 2005; Lecellier et al. 2005; Haasnoot et al. 2007). However, caution should be
taken to interpret the functional role of these proteins in natural virus infection and
antiviral defense. This is mainly because attempts to isolate virus-derived siRNAs
by several labs have been unsuccessful, questioning the possibility that mammalian
Dicer can access and process viral dsRNA to initiate antiviral RNAi (Pfeffer et al.
2005; Cai et al. 2006; Buck et al. 2007; Randall et al. 2007). It has also been
suggested that RNAi suppression by mammalian viral suppressors might be a
nonspecific effect since previously it has been shown that RNAi response can be
readily suppressed by dsRNA-binding proteins from prokaryotes, for which an
RNAi pathway seems not exist (Lichner et al. 2003). In fact, the RNAi suppression
activity of some mammalian viral suppressors was mapped to the dsRNA binding
domains, suggesting that suppression activity observed for those mammalian viral
proteins may be purely a nonspecific effect. Since these viral proteins have other
well-defined functions essential for viral replication, it is almost impossible to
access their function in RNAi suppression using corresponding loss-of-function
mutants. Nevertheless, cellular miRNA-mediated viral resistance has been docu-
mented for some mammalian viruses and has been shown to be targeted by
corresponding viral RNAi suppressors (Lecellier et al. 2005; Triboulet et al.
2007). miRNAs were also found to modulate disease resistance against plant
bacterial pathogens (Navarro et al. 2006). Accordingly, as a counterdefensive
mechanism, bacteria also encode proteins that actively suppress the antibacterial
response mediated by miRNAs (Navarro et al. 2008).
3 Identification and Function Characterization

of Viral RNAi Suppressors
Viruses are obligate intracellular parasites that must rely on host protein and nucleic
acid synthesis machinery to propagate. In order to survive in a largely hostile
environment, all viruses must have evolved a plethora of functions to suppress
host antiviral defense mechanisms such as RNAi. In agreement with this notion,
every single plant virus that has been closely examined is found to encode at least
one VSR. These include both single-stranded DNA viruses and RNA viruses with
positive-, negative-, or double-strand RNA genomes (see appendix, Table 1). Some
viruses, such as Citrus tristeza virus (CTV) and geminiviruses, encode multiple
VSRs with each of them appearing to have a distinct mode of action (Lu et al. 2004;
Vanitharani et al. 2004). So far, over thirty viral RNAi suppressors have been
identified, and this number is still growing. On prediction, there will be a very large
number of viral suppressors waiting to be identified for viruses infecting eukaryotic

hosts. Identification and function characterization of these novel suppressors will
not only give us a better picture about how pathogens counteract on host disease
control mechanisms but also improve our understanding of the evolutionary arm
race between hosts and parasitic pathogens. To date, many experimental strategies
have been developed for VSR identification in diverse systems. Below are the four
major strategies that are frequently used for VSR identification and preliminary
function characterization.
3.1 Agrobacterium-Mediated Transient Suppression Assay
Agrobacterium-mediated transient assay has been frequently used for VSR isola-
tion mainly because of its simplicity and rapidity. In this assay, the candidate
suppressor protein is often delivered into plant with a transgene construct that
triggers silencing of a stable reporter transgene, usually a gene encoding green
fluorescent protein (GFP). The delivery is usually done using agroinfiltration.
Agroinfiltration is a technique that allows transient expression of target gene in
plant leaves. Typically, Agroinfiltration is carried out by vacuum-infiltrating plant
leaves with a recombinant agrobacterial culture containing target gene T-DNA.
When a GFP coding sequence is used as inducer in a transient assay, the GFP
mRNAs produced from the stable GFP transgene in the infiltrated zone will be
destroyed by RNAi in the absence of a suppressor but will accumulate in the
presence of an RNAi suppressor manifested as enhanced green fluorescence
Fig. 4 Agrobacterium-mediated transient suppression assay for the identification of RNAi sup-
pressors encoded by CTV. Leaves of the 16c plants were infiltrated with an A. tumefaciens strain
carrying 35S-GFP together with an A. tumefaciens strain carrying the empty binary plasmid (35S:–),
35S:TAV 2b, 35S:CMV 2b, 35S:CTV p23, 35S:CTV p20, or 35S:CTV CP. The green fluores-
cence images of the coinfiltrated leaves with the abaxial-side up were taken 3 days postinfiltration
under a long-wave UV lamp. Image in this figure is adapted, with permission, from National
Academy of Sciences, USA (Lu et al. 2004), copyright 2004
70 X. Yi and R. Lu
(Fig. 4). Owing to its simplicity, agrobacterium-mediated transient assay allows

rapid characterization of a large number of VSR candidates (Voinnet et al. 1999;
Johansen and Carrington 2001; Guillaume and Olivier 2004).
3.2 Reversal of Transgene-Induced Gene Silencing
This strategy is based on the observation that many viral RNAi suppressors are able
to reverse transgene-induced silencing that occurs autonomously in some transgene
plants (Anandalakshmi et al. 1998; Brigneti et al. 1998; Kasschau and Carrington
1998). Thus, for identification purpose, candidate suppressors are assayed for their
ability to reverse an ongoing silencing targeting a transgene in a reporter plant.
Very often, the candidate suppressors are introduced into the reporter plants
through genetic crossing between reporter plants and transgenic plants expressing
candidate suppressors. The candidate suppressors can be also ectopically expressed
from a heterologous virus vector that is inoculated onto the reporter plants. How-
ever, if the latter strategy is used, an additive or synergistic effect resulted from
suppressors already carried by viral vector should be considered to access the
suppression activity of the candidate.
3.3 Grafting Assay
Grafting assay is so far the most reliable strategy used to assay VSR activity on
systemic silencing. This strategy is based on the observation that systemic silencing
signal generated from a silenced rootstock can spread into scion and cause
sequence-specific silencing targeting a scion transgene. Figure 5 illustrates a graft-
ing assay using a GUS-expressing tobacco line T19 as scion and a GUS-silenced
tobacco line 6b5 as rootstock. After introducing the candidate suppressor into line
6b5 through genetic crossing, whether or not the candidate suppressor is capable of
systemic silencing suppression is determined by assaying the GUS expression in the
T19 scion. Although a bit time-consuming, this strategy has not only allowed for
the identification of systemic suppression activity for VSRs that suppress local/
intracellular silencing but also allowed for the identification of VSR with specific
activity on systemic silencing (Guo and Ding 2002; Lu et al. 2004).
3.4 Replication Rescue of Mutant Viruses Defective

in RNAi Suppression
This strategy has been successfully used for identifying VSRs of animal viruses in
cell culture-based system (Li et al. 2004). The major component of this strategy is a
Fig. 5 Grafting assay of a

suppression activity on T19 T19
systemic silencing. T19,
tobacco line containing a
GUS transgene, which is
constitutively expressed. 6b5,
tobacco line containing GUS
transgenic, which is silenced.
(a) Suppression of GUS
expression in T19 scion 6b5 6b5
grafted onto 6b5 rootstock.
(b) Expression of CMV 2b b
T19 T19
transgene in 6b5 rootstock
leads to the suppression of
systemic silencing targeting
GUS transgene in T19 scion
6b5
6b5/2b 6b5/2b
mutant virus whose genome is modified to carry a GFP reporter gene in the place of
an RNAi suppressor. Thus, replication of this virus is suppressed by RNAi in wild-
type cells but will be restored to produce GFP if an RNAi suppressor is provided in
trans. This strategy has also been adopted by plant researchers in VSR identifica-
tion. In this case, the rescue of virulence and systemic accumulation is often used as
an indication of activity in RNAi suppression (Yelina et al. 2002).
It is important to note that, so far, a large number of VSRs have been identified
and characterized through transient expression together with a second reporter
transgene. This simple and fast approach may have limited application in charac-
terizing VSRs from viruses like CTV and geminiviruses, which are shown to
produce multiple VSRs with distinct modes of action (Lu et al. 2004; Vanitharani
et al. 2004). In the case of CTV, because the CP does not suppress local/intracellu-
lar silencing (see Sect. 4 for more details), its function in systemic silencing
suppression, as demonstrated in grafting assay, would have been overlooked in
transient expression assay. This transgenic approach may have its limitation in
identifying VSRs that exhibit temporal or spatial expression pattern during the
course of viral infection. For example, VSRs produced early during viral infection
and targeting viRNA biogenesis for suppression will fail to suppress siRNA-
induced RNAi, whereas those capable of inhibiting secondary viRNA synthesis
will not be identified as RNAi suppressors in the context of RNAi triggered by
inverted repeat transgenes. As we know, production of foldback RNAs from
inverted repeat transgenes does not require RdRP activities. Additionally, because
the level and timing of VSR and RNAi trigger expression in most transgenic or
transient approaches are usually set arbitrarily such that the results from indepen-
dent tests might not be comparable or inaccurate (Ding and Voinnet 2007).
Recently, a genetic rescue strategy has been demonstrated in a few reports, which
72 X. Yi and R. Lu
may serve as supplementary strategy for identification of novel VSRs (Deleris et al.
2006; Diaz-Pendon et al. 2007; Wang et al. 2006). In these reports, rescue of viral
replication was assayed in RNAi-defective hosts using mutant viruses with disabled or
modified VSRs. These assays not only reconfirmed the suppression activity of
known VSRs but also unequivocally implicate key RNAi genes in antiviral RNAi.
Thus, to fully appreciate the functionality of VSRs during authentic viral infections, a
combination of complementary strategies will be needed for function characteriza-
tion.
The fact that viruses are intracellular parasites that entirely rely on host macro-
molecule synthesis and metabolism machineries entails that viruses must overcome
various host antiviral mechanisms to survive and evolve. Since RNAi-based viral
immunity is not conserved in prokaryotes, it can be inferred that viral RNAi
suppressors must have emerged as RNAi antagonists after viruses expanded their
host range from prokaryotes to eukaryotes. The lack of any obvious sequence or
structure similarity in between known VSRs further suggests that they must have
emerged independently as a result of viral adaptation to RDVI. In agreement with
this hypothesis, many RNAi suppressors are produced from out-of-frame over-
lapped region of viral genome, and their function is dispensable for some basic viral
function such as viral replication and packaging. For example, both CMV 2b and
FHV B2 are produced from out-of-frame overlapped region of respective viral
genomes. Both will become dispensable for viral replication when host antiviral
RNAi is made defective (Wang et al. 2006; Diaz-Pendon et al. 2007). The over-
lapped suppressor genes might have been created through overprinting, a phenom-
enon in which a single coding sequence is translated in different reading frames
(Keese and Gibbs 1992). Very often, for each pair of genes created through over-
printing, one is more ancient and widespread whereas the other is novel and has a
confined lineage in the phylogeny of viruses (Li and Ding 2006).
In plants and worms, cellular proteins can also serve as negative regulators of
RNAi (Anandalakshmi et al. 2000; Kennedy et al. 2004; Ramachandran and Chen
2008). Some of these proteins target and destabilize siRNA or miRNA for RNAi
suppression (Kennedy et al. 2004; Ramachandran and Chen 2008). Presumably,
under natural conditions, these cellular proteins contribute to a fine-tuning mecha-
nism that ensures various RNAi directed functions to be carried out in a controlled
manner.
4 Function Mechanism of Viral RNAi Suppressors
RNAi-directed viral immunity involves a series of well-characterized molecular

events that culminate in viRNA production and target viral transcript cleavage.
Accumulating evidence suggests that viruses have evolved diverse counterdefen-
sive mechanisms that target and interfere with individual steps of RNAi pathway,
leading to defects in viRNA biogenesis or function. Probably as a reflection of the
complex nature of viral counterdefensive mechanisms, so far, only very few VSRs
are better characterized for their modes of action in RNAi suppression. Based on
their biological targets of RNAi machinery, these VSRs can be loosely classified
into four independent groups as discussed below.
4.1 Viral Suppressors that Bind Viral dsRNA to Interfere

with viRNA Biogenesis
Despite the extreme sequence diversity, many VSRs are in fact dsRNA-binding
proteins. This probably reflects the fact that all RNAi mediated antiviral responses
invariably begin with dicer processing of virus-derived dsRNAs. Thus, targeting
dicer substrate dsRNA for protection would serve as common strategy for many
VSRs. The B2 protein encoded by FHV is a VSR of this category that has been
subject to extensive functional and structural analyses. The full length FHV B2
protein is only 106 amino acids (aa) long. A dsRNA-binding domain can be found
at the N-terminal region. B2 is synthesized at high levels in the early stage of viral
replication and, like many other VSRs, suppresses RNAi across kingdoms (Li et al.
2002). B2 binds both siRNA and long dsRNA in a sequence-independent manner
(Chao et al. 2005; Lu et al. 2005). This finding suggests dual modes of action for B2
in RNAi suppression, in that long dsRNA binding would interfere with dicer
function in dsRNA processing whereas siRNA binding would inhibit active RISC
formation, leading to defect in target RNA cleavage. In agreement with this
hypothesis, B2 was found to inhibit dicer processing of long dsRNA in vitro
(Chao et al. 2005; Lu et al. 2005). Both NMR and crystallization structural analyses
have been performed for B2. These studies revealed an all-helix structure for the
N-terminal 72 aa (Chao et al. 2005; Lingel et al. 2005). A cocrystal structure of B2
and an 18-bp dsRNA shows that B2 recognizes dsRNA as a homodimer that forms a
four-helix bundle (Fig. 6a). B2 interacts exclusively with the ribose-phosphate
backbone of two successive minor grooves and the intervening major groove
such that when multiple B2 proteins bind to the same dsRNA molecule, they
would form a “coat” for the bound dsRNA, making the target dsRNA inaccessible
to Dicer (Fig. 6b, c). In consistence with structural data, replacement of Arg at
position 54, which is in the center of the dsRNA-binding surface, by Gln abolished
B2 activity in both dsRNA binding and dicing inhibition in vitro (Lu et al. 2005).
Despite sharing very limited sequence identity with FHV B2, a recent report shows
that the B2 protein from Nodamura virus (NoV), another member of the nodavirus
family, also forms dimmers with helical bundle structure. The crystal packing
places the RNA-binding residues along one face of symmetry-related molecules,
suggesting that NoV B2 suppresses RNAi using a similar mechanism (Shaik Syed
Ali and Chen 2009).
Compared to FHV B2, the V2 protein from Tomato yellow leaf curl geminivirus
(TYLCV) seems to have a unique taste in dsRNA binding in that it specifically
binds dsRNA with 50 end overhangs. Previously, it has been shown that V2 inhibits
74 X. Yi and R. Lu
Fig. 6 Structure basis of B2 and p19 function in RNAi suppression. Upper panel: Blue and green,
B2 dimer; Magenta, dsRNA. (a) B2 binds dsRNA as a dimmer and recognizes two successive
minor grooves and a major groove that separates them. (b) the dimmer is rotated by approximately
10 relative to the RNA to maximize contacts with both minor grooves. (c) the region of RNA that
is bound by B2 dimers is localized to one face of the duplex [adapted, with permission, from
MacMillan Publisher Ltd: Nat. Struct. Mol. Biol. (Chao et al. 2005), copyright 2005]. Lower
panel: Blue and magenta, individual monomers of the p19 dimer; Orange and pink, the two siRNA
strands. Two tryptophans from each monomer of the p19 dimer, which bracket the terminal base
pairs at either end of the siRNA duplex, are shown in stick representation. (d) A stereo view
perpendicular to the twofold axis. (e) and (f), Alternative mono views of the complex rotated by
90 along different axes (modified and adapted, with permission, from MacMillan Publisher Ltd:
Nature (Ye et al. 2003), copyright 2003)
silencing targeting a GFP transgene but does not affect the biogenesis of GFP-
specific siRNAs, suggesting that V2 targets a step in the RNAi pathway that is
downstream of Dicer processing of dsRNA precursor (Zrachya et al. 2007). RNA-
binding assay revealed that V2 binds to dsRNA of various lengths with 50 end
overhangs. Interestingly, dsRNA binding of V2 is inhibited by the presence of
phosphate group at the 30 ends, but the 30 end methylation modification does not
seem to affect V2 activity in dsRNA binding, indicating that V2 may recognize and
bind to the junction of a dsRNA region with a 50 end overhanging strand (Fukunaga
and Doudna 2009). dsRNA-binding competition assay suggests that V2 may
suppress RNAi through interfering with function of sgs3, an Arabidopsis gene

involved in posttranscriptional gene silencing and natural virus resistance (Mourrain
et al. 2000). SGS3 protein binds dsRNA with 50 end overhangs in vitro. Very much
like V2, SGS3-dsRNA binding is inhibited by the presence of phosphate group at
the 30 end but not by the presence of methylation modifications. In dsRNA-binding
competition assay, V2 can efficiently outcompete SGS3 in dsRNA binding (Fukunaga
and Doudna 2009). This finding implies that V2 may suppress RNAi through
depriving SGS3 of its natural substrate dsRNA, although the nature and
functional role of the SGS3 substrate remain unclear. It is possible that binding of
dsRNAs with 50 overhangs by SGS3 serve as molecular trigger of host RdRP
function in secondary siRNA production. If proved true, a functional role of V2
in RNAi suppression will be to interfere with the biogenesis of secondary siRNA
production.
4.2 Viral Suppressors that Target Virus-Derived siRNA

for RNAi Suppression
The first viral protein characterized in this category is the p19 protein from Tomato
bushy stunt virus (TBSV), a member of the tombusvirus family that also include
Cucumber necrosis virus (CNV) and Cymbidium ringspot virus (CRV). Initially, it
was found that the p19 proteins from these viruses are not essential for virus cell-
to-cell movement but, instead, are required for virus systemic spread and symptom
development. p19 was identified as an RNAi suppressor based on its ability to
reverse the silencing targeting GFP transgene in the systemic leaves of plants
infected with either TBSV or PVX carrying a p19 insert (Voinnet et al. 1999).
Subsequently, several groups independently demonstrated the suppression activity
of p19 proteins from a number of different tombusviruses using the agroinfiltration
assay (Qiu et al. 2002; Qu and Morris 2002). Through the study of p19-mediated
RNAi suppression in Drosophila embryo extracts, Burgyan and colleagues found
out that p19 inhibited siRNA-directed slicing of the target mRNA only when p19
were added to the embryo extracts at the same time as the siRNA duplex. The fact
that p19 failed to inhibit slicing when added 20 minutes later than the siRNA duplex
suggested that p19 may bind siRNA duplexes and prevent them from being
incorporated into RISC (Lakatos et al. 2004). In supporting this hypothesis, P19
binds 21-nt siRNA duplex with high affinity independent of the 2-nt overhangs at
the 30 end of siRNA, and its affinity is much weaker for dsRNAs of 22 nt or longer.
Interestingly, P14, a P19 homologue encode by aureusviruses of the same family,
binds dsRNA for RNAi suppression without a size preference (Merai et al. 2005).
p19 structure studies through crystallography reconfirmed the siRNA binding
specificity and established a structural explanation for how dimerization of p19
was essential for binding siRNA (Vargason et al. 2003; Ye et al. 2003). From these
studies, it is clear that p19 binds to a siRNA as a homodimer with two molecules of
p19 per siRNA duplex. In a crystal structure of p19 bound to a 21-nucleotide
76 X. Yi and R. Lu
siRNA, the 19-basepair RNA duplex is “cradled” within the concave face of a
continuous eight-stranded b-sheet, formed across the p19 homodimer interface
(Fig. 6d). In consistence with the sequence-independent siRNA recognition of
p19, the direct and water-mediated intermolecular contacts are restricted to the
backbone phosphates and sugar 20 -OH groups. Two a-helical “reading heads”
project from opposite ends of the p19 homodimer and position pairs of tryptophans
for stacking over the terminal base pairs, thereby measuring and bracketing both
ends of the siRNA duplex (Fig. 6e, f). These studies provide a perfect structural
explanation of siRNA sequestering by p19, making it one of the best characterized
VSRs.
Recently, siRNA binding capacity was also demonstrated for p21 of clostero-
viruses and AC4 protein from geminivirus, suggesting a common strategy used by
viral suppressors in counteracting on RDVI (Chellappan et al. 2005; Lakatos et al.
2006). Interestingly, p21 binds siRNA duplex thereby to inhibit the formation of
mature RISC, whereas AC4 preferentially binds the single-stranded siRNA or
miRNA in in vitro RNA binding assay presumably to inhibit the activity of mature
RISC.
The RNase3 protein of Sweet potato chlorotic stunt virus (SPCSV) suppresses
RNAi in a unique way, in that it targets viRNA for destruction. SPCSV is a single-
stranded RNA (ssRNA) crinivirus whose infection is phloem limited. SPCSV
synergizes Seet potato Fathery mottle virus (SPFMV), an HC-Pro producing
virus, in coinfected sweet potato plants as manifested in enhanced disease symp-
toms and elevated titer of SPFMV. SPCSV RNase3 is a class I endoribonuclease III
that specifically binds and cleaves dsRNA but not ssRNA of various length. When
expressed from a sweet potato transgene, SPCSV RNase3 alone is sufficient to
break down resistance to SPFMV and other unrelated viruses, leading to higher
accumulation of beneficiary viruses and severe disease symptoms (Cuellar et al.
2009). Notably, SPCSV RNase3 suppresses RNAi in an endonuclease activity-
dependent manner. In vitro, it cleaves synthetic siRNAs duplexes of 21, 22, and
24 bp in length, and the cleavage product is approximately 14 bp long. Thus, it
appears that SPCSV RNase3 suppresses antiviral defense by destroying infecting
virus-derived siRNAs. In agreement with this notion, SPCSV RNase3 treatment
caused a clear reduction of total siRNA isolated from SPFMV-infected sweet potato
plants (Cuellar et al. 2009). Interestingly, deep sequencing analysis suggested that
only a small margin of SPFMV-derived siRNAs, 3.95%, are siRNA duplexes. It is
thus possible that clearance of this small portion of siRNAs may allow beneficiary
viruses to get an upper hand on silencing. It is also possible that the siRNA duplexes
targeted by RNase3 may play an essential role in establishing systemic antiviral
silencing, which can attenuate/slow down viral infection in the presence of RNAi
suppressor encoded by beneficiary viruses (Voinnet 2005).
In Arabidopsis, miRNA and endo-siRNA are methylated at the 30 end by HEN1
for enhanced stability (Yu et al. 2005). Virus-derived siRNAs are also the target of
HEN1, suggesting a role of HEN1 in antiviral defense. As a counterdefensive
mechanism, some viral RNAi suppressors target HEN1-directed modification for
RNAi suppression. The HC-Pro from potyvirus (Ebhardt et al. 2005) is one of
such suppressors whose expression from transgene was found to be associated with
marked decrease of the 30 end modification of viral siRNAs. Interestingly, HC-Pro
does not seem to significantly affect the modification of endogenous miRNAs and
24-nt siRNAs. The silencing suppressor from tobamovirus, the 126-kDa protein,
appears to use similar strategy in RNAi suppression (Ding et al. 2004; Vogler et al.
2007). Previously, it was shown that the 126-kDa protein of TMV is dispensable for
viral replication in protoplast. In N. benthamiana plants, TMV infection leads to the
production of virus-specific siRNAs in the presence of 126-kDa protein, indicating
that this TMV suppressor targets a step downstream of viRNA biogenesis. b-
elimination assay confirmed that TMV infection indeed interferes with HEN1-
mediated methylation of viRNAs and that this interference and the formation of
virus-induced disease symptoms are experimentally linked to the silencing suppres-
sor activity of the 126-kDa protein. This RNAi suppression strategy seems to be
shared within the tobamovirus family as the infection by Olseed rape mosaic
tobamovirus (ORMV) also leads to interference with HEN1-mediated methylation
of siRNA and miRNA. Previously, it was shown that, for unknown reason, RNAi
suppression by the coat protein of Turnip crinkle virus (TCV) is often associated
with diminished level of target-derived siRNAs (Qu et al. 2003; Dunoyer et al.
2004). Now, it turns out that this is because TCV CP interferes with the methylation
of siRNA, leading to its instability. Interestingly, in contrast to suppressors of
tobamovirus origin, TCV CP does not interfere with the methylation of host
miRNA.
4.3 Viral Suppressors that Target RNAi Effectors for Suppression
The first viral protein identified in this category is the P0 protein from poleroviruses
(Pfeffer et al. 2002). P0 protein was first identified as RNAi suppressor based on an
agroinfiltration-based transient assay in which overexpression of P0 protein from
Beet western yellows virus (BWYV) suppressed the silencing targeting GFP
reporter in transgenic N. benthamiana plants. Later on, it was found that P0
physically interacts with Arabidopsis orthologs of S-phase kinase-related protein
1 (SKP1) through a conserved F-box-like motif (Pazhouhandeh et al. 2006). It
appears that P0 functions in an SKP1-like protein-dependent manner, in that down-
regulation of a SKP1 ortholog in N. benthamiana rendered the plants resistant to
polerovirus infection. In agreement with this observation, point mutations in the
F-box-like motif of P0 abolished not only the P0-SKP1 ortholog interaction but also
diminished virus pathogenicity and the RNAi suppression activity of P0. SKP1 is a
component of the SCF (Skp1, Cul1/Cdc53, F-box proteins) family of ubiquitin E3
ligases, which are known to add polyubiquitin tracts on selected lysine residues,
thereby marking a protein for proteasome-mediated degradation. These findings
therefore strongly suggested that P0 may target key RNAi components for degra-
dation. In supporting this hypothesis, later on, it was shown that P0 targets the PAZ
motif and its adjacent upstream sequence in Arabidopsis AGO1 and mediates its
78 X. Yi and R. Lu
degradation (Baumberger et al. 2007). Interestingly, P0-mediated AGO1 degrada-

tion does not seem to involve proteasome. It is also clear that P0 does not block a
mobile signal of silencing, which is believed to keep infection by poleroviruses in a
phloem restricted manner. However, the P0 gene of Sugarcane yellow leaf virus
(SYLV) seems to have evolved additional function, which allows this suppressor to
suppress systemic silencing. It remains unclear whether this extra function is
responsible for the devastating effect SYLV caused on sugarcane production
(Mangwende et al. 2009).
A recent study by Zhang and colleagues showed that the 2b protein of CMV
FNY strain also targets AGO1 for RNAi suppression (Zhang et al. 2006). The CMV
FNY 2b (FNY2b) shares 51% identity and 62% similarity with the 2b protein of
CMV Q strain (Q2b), which was one of the first two viral RNAi suppressors
identified. Interestingly, Arabidopsis plants expressing Q2b produced no visible
abnormalities compared to transgenic plants transformed with empty vector,
whereas FNY2b expressing Arabidopsis showed severe developmental defects
characteristic of that observed for Arabidopsis AGO1 mutants ago1-25 and ago1-27
(Morel et al. 2002; Zhang et al. 2006). Since AGO1 is known to be required for
miRNA function in Arabidopsis, this observation suggested that FNY2b is capable
of miRNA function suppression. In supporting this hypothesis, FNY2b was found
to physically interact with AGO1 in vivo. In vitro, FNY2b specifically binds to one
surface of the PAZ-containing module of AGO1. However, this specific binding
does not interfere with siRNA loading because physical association between AGO1
and endogenous siRNAs can still be detected in the presence of FNYb2. Instead,
RISC reconstitution assay suggested that 2b specifically inhibits AGO1 cleavage
activity. The Arabidopsis AGO1 also recruits virus-derived siRNAs in vivo, sug-
gesting a role of AGO1 in RNAi-directed defense against virus infection.
Virus-produced RNA molecules can also directly bind and inhibit the function of
key RNAi factors. One of the best studied examples is the VA1 RNA produced by
adenovirus (Lu and Cullen 2004; Andersson et al. 2005). In adenovirus-infected
cells, VA1 is expressed at very high levels and can potently inhibit RNAi induced
by short hairpin RNAs (shRNAs) or human miRNA precursors but not RNAi
induced by artificial siRNA duplexes, suggesting that VA1 targets a step down-
stream of dicer processing of dsRNA in the RNAi pathway (Lu and Cullen 2004).
The fact that VA1 is a highly structured RNA molecule that shares structural
features with miRNA precursors suggests that VA1 may compete with bona fide
Dicer substrate for Dicer binding and processing (Fig. 7, compare a and b). In
agreement with this hypothesis, VA1 was found to directly bind and inhibit the
function of Dicer in vitro (Lu and Cullen 2004). Cellular RNA molecules with
highly structured regions can also inhibit Dicer function using similar strategy.
Recently, a worm starvation-induced transcript, rncs-1, has been found to inhibit
Dicer function in vitro. Very much like VA1, rncs-1 transcript is a highly structured
RNA molecule of 800 nt long (Fig. 7c). The secondary structure of rncs-1 contains
an almost perfectly double-stranded helix of 300 bp, making rncs-1 a perfect
substrate of Dicer. However, because of the branched terminal structures, rncs-1
transcripts are resistant to Dicer processing as assayed in vitro.
a
GU
5' CU UGA
UAGCUUAUCAGACUGAUGUUGA pre-miR-21
A
GUCGGGUAG CUGACCACAACG U
CU GU UC
3' AC
b
GAUA
G A
5' U A
A G G U UUCGA
GGGC CUCUUCCGU GUCUG UOGCAAGGGUAUCAUGGOGGAOGAOOGGGG A
CUCG GAGGGGGCA CAGACUGCAGCGU OOCA AGUGOOGOCUGCCGGOOUA C
C C A G A GA U GGCCC
U UGG C C
U UG C
U VA1
UA
3' GU
CG
GC
CG
c CG
C U 800
G U 1
CCA
400 500 600 700
300 200 100
Fig. 7 Comparison of the structures of a representative human pre-miRNA, adenovirus VA1

RNA and worm rncs-1 transcript. (a) Proposed structure of the human pre-miR-21 miRNA
processing intermediate. The mature miR-21 sequence is in red. (b) Proposed structure of adeno-
virus VA1. Both human pre-miR-21 and adenovirus VA1 have a terminal stem of 16 base pairs
or more and a short 30 overhang, which are required for binding by exportin-5. Black arrowheads
indicate known Dicer cleavage sites in pre-miR-21 and hypothetical cleavage sites in VA1
(both images are modified and adapted, with permission, from MacMillan Publisher Ltd: Nature
Immunology (Cullen 2006), copyright 2006). (c) Predicted secondary structure of mature rncs-1
RNA. Asterisks, GU pairs; dots, mismatches, and bulges; gray arrowhead, exon–exon junction;
numbering, nucleotide relative to transcription start (adapted, with permission, from National
Academy of Sciences, USA: (Hellwig and Bass 2008), copyright 2008)
4.4 Viral Suppressors that Suppress Systemic RNAi
In plants and worms, a silencing status can spread out of the initial silencing site
(Voinnet and Baulcombe 1997; Tijsterman et al. 2004; Winston et al. 2002). The
target sequence-specific feature of systemic silencing suggests the involvement of a
class of RNA molecules. Systemic silencing in plants is believed to be responsible
for initiating an antiviral condition prior to virus arrival (Baulcombe 2002). Accord-
ingly, as a counter-defense mechanism, some viruses encode suppressors that are
capable of specific targeting of systemic silencing signal. This was first demon-
strated for the P25 protein of PVX. An earlier study suggested a role for P25 in PVX
systemic spread in that deletion of P25 coding sequence does not affect PVX
accumulation in inoculated protoplasts but abolishes spread of PVX out of the
initially infected cells (Angell et al. 1996). A functional role of P25 in systemic
80 X. Yi and R. Lu
silencing suppression was established in a study in which systemic silencing

targeting a transgene induced by a movement defective PVX would not occur in
N. benthamiana plants unless P25 was inactivated (Voinnet et al. 2000). Later on, it
was shown that expression of a P25 homolog prevented virus RNA and systemic
silencing from entering the meristem tissue, reconfirming the function of P25-like
viral proteins in systemic silencing suppression (Foster et al. 2002). In agreement
with these observations, P25 suppression of RNA silencing was also shown to be
required for the cell-to-cell movement of PVX (Bayne et al. 2005). Currently, the
function mechanism involved in RNAi suppression by PVX P25 remains unclear.
In addition to PVX P25, the 2b protein from CMV Q strain (Q2b) and the major
coat protein (CP) of Citrus tristeza virus (CTV) are another two VSRs which are well
characterized for their suppression activity on systemic silencing. Suppression of
systemic silencing by the Q2b was demonstrated in a series of grafting assays as
illustrated in Fig. 5. In these assays, systemic silencing did not occur when the Q2b
protein was expressed in the rootstock or only in the intergraft between rootstock and
scion, suggesting that Q2b inactivates systemic silencing signal or stop it from
spreading out. The former assumption seems more likely since systemic silencing
did not occur even when Q2b is only expressed scion (Guo and Ding 2002).
Interestingly, expression of 2b also reduced DNA methylation associated with
transgene silencing. Apparently, the CP of CTV suppresses systemic silencing
using a mechanism different from that of Q2b. Agroinfiltration-mediated transient
assay suggested that CTV CP does not suppress silencing in the initial silencing site,
nor the silencing signal capable of short distance movement (Fig. 4). However, when
expressed in rootstock used in grafting assay, CTV CP prevents silencing from spread
into scion, making it so far the only viral suppressor that specifically interferes with
systemic silencing (Lu et al. 2004). Interestingly, when expressed in scion that was
grafted onto a rootstock that keeps generating systemic silencing signal, CTV CP
could not prevent silencing from occurring in the scion (R. Lu, unpublished data).
Thus, unlike Q2b, CTV CP is unable to inactivate the silencing signal. These data
rather suggested CTV CP suppresses systemic silencing by intercepting the silencing
signal. CTV CP does not reverse DNA methylation-associated transgene silencing as
well, suggesting that, unlike Q2b, the CP of CTV has a dedicate role in intercepting
the systemic silencing signal capable of long distance movement. This unique feature
makes CTV CP an ideal molecular tool for the study of the nature and identity of
silencing signal that is capable of long distance movement.
5 RNAi Suppressors of Nonviral Origin
5.1 Suppression of RNAi by Bacterial Pathogens
Agrobacterium tumefaciens (A. tumefaciens)-mediated horizontal gene delivery

has long been used to characterize target gene function. However, A. tumefaciens
transfer DNA (T-DNA) mediated gene delivery often triggers an RNAi-based
response that silences the transgene at transcriptional or posttranscriptional level

(Napoli et al. 1990; Waterhouse et al. 1998; Matzke et al. 2000). This RNAi-
based response would have served as a disease-limiting mechanism against
virulent Agrobacterium in wild as we know that the bacterial growth requires
the proper function of the tumor-inducing genes and genes responsible for opine
synthesis. Thus, when these genes are targeted by the RNAi, the bacterial growth
will be affected due to lack of nutrients produced in tumor. The first experimental
evidence to support this hypothesis comes from the detection of oncogene-
derived siRNAs in virulent Agrobacterium induced tumors (Dunoyer et al.
2006). Despite this RNAi-based defense mechanism, successful infection, as
manifested in the formation of tumors, by virulent Agrobacterium is frequently
observed in diverse plant species. This observation suggests that A. tumefaciens
must have evolved a mechanism to suppress the RNAi-based defense response.
This hypothesis is supported by the fact that RNAi-deficient Arabidopsis becomes
hypersusceptible to A. tumefaciens pathogen. Thus, current model suggests that
successful infection by A. tumefaciens relies on a potent anti-RNAi state estab-
lished in tumors in which the synthesis of the T-DNA derived siRNA is inhibited.
Since silencing targeting a GFP transgenes is lost in developing tumors but partly
regains its strength in mature tumor, it is believed that the silencing suppression
mechanism is intrinsic to cell dedifferentiation and/or proliferation (Dunoyer
et al. 2006). In supporting this hypothesis, cell dedifferentiation and proliferation
induced by hormone treatments of A. tumefaciens-free tissues was also sufficient
to suppress silencing targeting GFP transgene.
miRNA pathway can also contribute to bacterial control in plants and this
phenomenon is best characterized for disease resistance against Pseudomonas
syringae (P. syringae) (Navarro et al. 2006). In Arabidopsis miRNA miR393
targets and negatively regulates the expression of F-box auxin receptors TIR1,
AFB2, and AFB3, leading to the repression of auxin signaling that naturally
restricts P. syringae growth. In an elegant setup, Navarro and colleagues demon-
strated that the expression of miR393 can be induced by a flagellin-derived
peptide, a type of pathogen-associated molecular patterns (PAMPs), thus estab-
lishing a role for plant miRNA in bacterial disease control (Navarro et al. 2006).
This conclusion was further supported by the observation in which Arabidopsis
mutants defective in miRNA pathway partially restored the growth of a type III
secretion-defective mutant of P. syringae. These Arabidopsis mutants also sus-
tained the growth of nonpathogenic bacteria, implicating miRNAs as key compo-
nents of plant basal defense. Notably, very much like that for viral pathogens,
P. syringae uses the effector proteins of type III secretion system to interfere
with bacterial innate immunity by modulating the biogenesis and activity of
PAMP-responsive miRNAs (Navarro et al. 2008). Two of the effector proteins,
AvrPtoB and AvrPto, were analyzed for their function in the suppression of
miRNA-mediated antibacterial defense. These analyses suggested that P. syringae
effectors proteins target key steps in the biogenesis or function of some PAMP-
responsive miRNAs for suppression, although the detailed function mechanism
remains to be elucidated.
82 X. Yi and R. Lu
5.2 Cellular RNAi Suppressors
Probably as fine-tuning mechanism, many organism species also encode cellular

proteins capable of RNAi suppression. In plants, rgs-CaM (regulator of gene
silencing-calmodulin-like protein) represents such a cellular protein that suppresses
RNAi targeting GFP transgene and replicating virus (Anandalakshmi et al. 2000).
rgs-CaM was first isolated as HC-Pro interactor in a yeast-two hybrid screen.
Overexpression of rgs-CaM led to tumor formation as that found for overexpression
of HC-Pro. The fact that the expression of rgs-CaM can be induced by HC-Pro
indicated that the RNAi suppression function of HC-Pro may be rgs-CaM-depen-
dent. It is yet to be tested whether HC-Pro suppresses RNAi through physical
interaction with rgs-CaM in vivo. A C-terminal domain of rgs-CaM contains
three EF-hand calcium-binding motifs with high similarity to plant calmodulins
and calmodulin-related proteins. An N-terminal domain of 40–50 amino acids may
specify the intracellular location or the regulatory properties of rgs-CaM. Interest-
ingly, rgs-CaM mRNA is present at low levels in leaves and flowers and at higher
levels in stem and root. Whether this reflects an active regulation of RNAi-mediated
function under natural condition remains to be confirmed.
While it remains unclear whether rgs-CaM negatively regulates gene silencing-
mediated miRNAs, SDN1, an Arabidopsis homolog of worm ERI-1 (see next
section for more details), has recently been identified as negative regulator of
miRNA function in Arabidopsis (Ramachandran and Chen 2008). SDN1 stands
for Small RNA DEGRADING NUCLEASE1. As indicated in its name, SDN1
is a 30 –50 exoribonuclease that specifically degrades single-stranded small
RNAs, miRNAs, and endo-siRNAs in a sequence-independent manner. SDN1 is
a multiple-turnover enzyme, which functions redundantly with other four members
of the same clade. Because SDN1 degrades miRNAs with a 20 -O-methyl group at
the 30 end inefficiently, it looks like that the 20 -O-methyl group present in all
plant small RNAs may deter SDN1 activities (Ramachandran and Chen 2008).
The fact that SDN1 acts less efficiently on uridylated miRNAs further suggests that
uridylation could have a protective role against SDN exonucleolytic degradation
in plants.
As aforementioned, eri-1(enhanced RNAi-1) of C. elegans worm represents a
cellular RNAi antagonist in animals. eri-1 was identified in a genetic screen aimed
to isolate mutants with enhanced sensitivity to RNAi triggered by dsRNA (Kennedy
et al. 2004). Sequence analysis suggests that eri-1 encodes an evolutionarily con-
served protein with domains homologous to nucleic-acid-binding and exonuclease
proteins. ERI-1 specifically degrades siRNA duplexes with 2-nt 30 overhangs
in vitro and reduces RNA interference efficiency in vivo (Kennedy et al. 2004;
Iida et al. 2006). In consistence with these results, worms with eri-1 mutations were
found to accumulate more siRNAs than do wild-type animals after being exposed to
dsRNA or siRNAs. The nematode ERI-1 is predominantly cytoplasmic and is
expressed most highly in the gonad and a subset of neurons. This is in agreement
with previous observations that RNAi activity in those tissues is relatively weaker.
6 Biotechnological Application of RNAi Suppressors
6.1 Enhance Gene Expression for Rapid Function Analysis

and Mass Production of Valuable Protein
Transient expression of target gene for rapid function assay has been frequently used
in diverse systems. However, horizontal gene delivery through transfection, electro-
poration, bombardment, microinjection, and agrobacterium-mediated transformation
often results in poor expression of target gene. Although other possibilities exist, a
major host factor that is responsible for suboptimal expression of target gene is an
RNAi-based gene surveillance system that guards against nucleic acid intruders.
Thus, as a circumventing strategy, codelivery of an RNAi suppressor has been used
to enhance target gene expression in diverse systems, and this strategy has been
shown to be particularly successful in plants (Johansen and Carrington 2001;
Voinnet et al. 2003).
In plants, transient expression of target gene can be easily achieved using
recombinant A. tumefaciens strain that contains the target gene in the T-DNA
region of the binary-Ti plasmid. When the bacterial culture is vacuum-infiltrated
into leaves, and, upon T-DNA transfer, the target gene will be ectopically
expressed in cells that received the T-DNA. However, very often, the ectopic
gene expression ceases after 2–3 days mainly due to RNAi triggered by transgene
corresponding to the target (see Sect. 5.1 for detailed discussion). By coinfiltration
of a viral RNAi suppressor, Johansen and Voinnet have shown that the target gene
expression can be significantly enhanced (Johansen and Carrington 2001; Voinnet
et al. 2003). In the case of p19 suppressor from TBSV, expression of a range of
proteins was enhanced 50-fold or more in the presence of p19. Quantitative
analysis using GFP as target protein suggested that in coinfiltrated tissues, GFP
proteins accumulated up to 7% of total soluble protein (Voinnet et al. 2003). Based
on these observations, many research labs have developed various transient gene
expression systems for fast, flexible, and reproducible function assay. In fact,
because of its simplicity and rapidity, the p19-enhanced expression system is
also used in industrial production for isolation of a broad range of proteins without
the need for the time-consuming regeneration of stably transformed plants.
Viruses multiply very efficiently and thereby drive their gene expression to
extremely high level. This has inspired researchers to utilize replicating virus for
target gene overexpression. To facilitate large-scale virus inoculation, the recom-
bined virus is often delivered into host as stable transgene. In principle, this
technology, often referred to “amplicon,” will allow for much higher yield of the
target gene products compared to conventional transgene approaches that utilize
strong promoters. However, in plants, the first attempt of this strategy failed, in that
all of the transformants consistently exhibited RNA silencing of the amplicon
transgene (Angell and Baulcombe 1997). Now we know that the dsRNA form of
viral replication intermediates produced in every single cell of the transgenic plants
84 X. Yi and R. Lu
would have served as potent triggers of the RNAi-based defense mechanism that
normally suppresses viral replication during natural infections. Conceivably, coex-
pression of viral suppressors would diminish the suppression effect of antiviral
RNAi, thereby allowing high-level target gene expression initially envisaged for
amplicons. This idea was once tested using transgenic tobacco plants expressing
TEV HC-Pro. When the HC-Pro expressing line is crossed with amplicon line that
is designed to express a GUS reporter gene from the PVX genome (Mallory et al.
2002), a dramatic increase in virus accumulation and GUS expression was observed
in progeny plants that contain both the suppressor transgene and the amplicon
locus. The GUS expression was so much enhanced that leaves of mature plants
accumulated the GUS protein up to 3% of total soluble proteins.
6.2 Serve as Molecular Tools to Probe Various RNAi-Directed

Functions
The target sequence-independent mode of action of all VSRs has limited their
application in function analysis of specific genes. However, the fact that some
of VSRs function through specific interaction with key components of RNAi
machinery suggests that we may be able to use VSRs as molecular tools to for
function and mechanism study of RNAi. For example, the CP of CTV has been
shown to specifically suppress systemic silencing, probably by interfering with the
long distance spread of systemic silencing signal (Lu et al. 2004). Current studies
support a role for 21 nt siRNAs in short distance spread of gene silencing but give
no clue about the identity of the signal molecules responsible for long distance
spread of silencing (Dunoyer et al. 2005). The unique feature of CTV CP makes it
an ideal molecular tool for identifying such a signal molecule. It can be anticipated
that detection and biochemical identity characterization of protein and/or RNA
molecules that are physically associated with CTV CP might allow us to gain
insight into the identity of this mysterious class of signal molecules. Similarly,
VSRs that function through specific interaction with key RNAi factors may allow
us to study RNAi-mediated function in organisms for which a genetic approach is
not available. For example, since VA1 is known to specifically bind and inhibit the
function of mammalian Dicer, overexpression of VA1 in a tissue-specific manner
may allow for the characterization of miRNA-mediated function in the target tissue.
In plants, some viral suppressors, when expressed as transgene, are able to
suppress miRNA function, leading to the accumulation of miRNA targets (Chapman
et al. 2004; Dunoyer et al. 2004). Although these observations were made on model
plant Arabidopsis, it can be inferred that transgenic expression of these VSRs would
also lead to accumulation of miRNA targets in plant species for which a genetic
approach for miRNA function study is still lacking or impossible. Thus, these VSRs
can be used as molecular probes to study miRNA-mediated function in economically
important crop plants. It is also possible that overexpression of these VSRs in some
animal species will compromise miRNA function, thereby leading to enrichment of
miRNA targets. Once proved true, this strategy may greatly facilitate miRNA target
identification in animal. Animal miRNA target identification has been difficult
mainly because only as few as 7 nt complementary sequence is needed for animal
miRNA to confer suppression effect.
endo-siRNAs of animal are a class of relatively new small RNAs isolated from
both germline and soma. In addition to transposon control, animal endo-siRNAs
may also modulate the gene expression in trans like that found for plant endo-
siRNAs (Czech et al. 2008; Okamura and Lai 2008; Watanabe et al. 2008).
Recently, it was shown that expression of various plant and insect VSRs in
transgenic flies led to no perturbation of miRNA pathway but instead inhibited
harpin RNA-triggered RNAi as well as transposon silencing conferred by endo-
siRNAs (Berry et al. 2009). These findings thus identify VSRs as genetic tools for
the study of endo-siRNA-mediated functions in animals.
Appendix
Table 1 Viral suppressors of RNAi

Virus genus Virus name VSR Motif
implicated in
VSR activity
Positive-strand RNA viruses of plant origin
Aureusvirus Pothos latent virus P14 dsRNA binding
Carmovirus Turnip crinkle virus CP
Melon necrotic spot virus P42
Hibiscus chlorotic ringspot virus CP
Closterovirus Beet yellows virus P21 dsRNA binding
Citrus tristeza virus P20
P23
CP
Grapevine leafroll-associated virus-2 P24
Beet yellow stunt virus P22
Crinivirus Sweet potato chlorotic stunt virus P22
RNase3 RNaseIII
Comovirus Cowpea mosaic virus Small CP
Cucumovirus Cucumber mosaic virus 2b dsRNA binding
Tomato aspermy virus
Furovirus Soil-borne wheat mosaic virus 19K Cysteine-rich
protein
Hordeivirus Barley stripe mosaic virus gb Cysteine-rich
protein
Pecluvirus Peanut clump virus P15 Cysteine-rich
protein
Polerovirus Beet western yellows virus P0
(continued)
86 X. Yi and R. Lu
Table 1 (continued)
implicated in
VSR activity
Cucurbit aphid-born yellows virus
Potexvirus Potato virus X P25 Helicase
Potyvirus Tobacco etch virus Hc-Pro
Potato virus Y
Turnip mosaic virus
Sobemovirus Rice yellow mottle virus P1
Tobamovirus Tobacco mosaic viruses P130
Tomato mosaic viruses
Tobravirus Tobacco rattle virus 16K Cysteine-rich
protein
Tombusvirus Tomato bushy stunt virus P19 dsRNA
bindinga
Cymbidium ringspot virus
Tymovirus Turnip yellow mosaic virus P69
Vitiviruses Grapevine virus A P10
Negative-strand RNA viruses of plant origin
Tenuivirus Rice hoja blanca virus NS3
Tospovirus Tomato spotted wilt virus NSs
Double-stranded RNA viruses of plant origin
Phytoreovirus Rice dwarf virus Pns10
DNA viruses of plant origin
Begomovirus Tomato leaf curl virus C2 DNA binding,
NLS
Tomato Yellow Leaf Curl Virus V2 dsRNA
bindingd
TYLCCNV-Y10 Y10b bC1 DNA binding,
NLS
African cassava mosaic virus (KE) AC2 DNA binding,
NLS, AD
EACMCV, ICMV, TGMV
Mungbean yellow mosaic virus
African cassava mosaic virus (CM) AC4 miRNA
bindingb
Curtovirus Beet curly top virus L2 Protein binding
Positive-strand RNA viruses of animal origin
Nodavirus Flock house virus, nodamura virus, B2 dsRNA binding
Striped jack nervous necrosis virus,
Greasy grouper nervous necrosis virus
Negative-strand RNA viruses of animal origin
Orthomyxovirus Influenza virus A NS1 dsRNA binding
Orthobunyavirus La Crosse virus NSs
Filovirus Ebola virus VP35
Double-stranded RNA viruses of animal origin
Orthoreovrivus s3 dsRNA
bindingc
(continued)
Table 1 (continued)
implicated in
VSR activity
Retroviruses of animal origin
Lentivirus HIV-1 Tat
Spumavirus PFV-1 Tas
DNA viruses of animal origin
Adenovirus Adenovirus VA1 RNA Dicer binding
Poxvirus Vaccinia virus E3L dsRNA binding
This table is adapted and modified, with permission, from Annual Reviews: Annu. Rev. Microbiol.
(Li and Ding 2006), copyright 2006
a
Prefer 19-nt RNA duplex
b
Single-strand mature miRNA
c
Prefer dsRNA longer than 30 nt
d
Prefer dsRNA with 50 end overhangs
References
Ambros V, Chen X (2007) The regulation of genes and genomes by small RNAs. Development
134:1635–1641
Anandalakshmi R, Marathe R, Ge X et al (2000) A calmodulin-related protein that suppresses
posttranscriptional gene silencing in plants. Science 290:142–144
Anandalakshmi R, Pruss GJ, Ge X et al (1998) A viral suppressor of gene silencing in plants. Proc
Andersson MG, Haasnoot PC, Xu N et al (2005) Suppression of RNA interference by adenovirus
virus-associated RNA. J Virol 79:9556–9565
Angell SM, Baulcombe DC (1997) Consistent gene silencing in transgenic plants expressing a
replicating potato virus X RNA. EMBO J 16:3675–3684
Angell SM, Davies C, Baulcombe DC (1996) Cell-to-cell movement of potato virus X is asso-
ciated with a change in the size-exclusion limit of plasmodesmata in trichome cells of
Nicotiana clevelandii. Virology 216:197–201
Aravin AA, Sachidanandam R, Girard A et al (2007) Developmentally regulated piRNA clusters
implicate MILI in transposon control. Science 316:744–747
Bartel DP (2004) MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116:281–297
Baulcombe D (2002) Viral suppression of systemic silencing. Trends Microbiol 10:306–308
Baulcombe DC (1999) Fast forward genetics based on virus-induced gene silencing. Curr Opin
Plant Biol 2:109–113
Baulcombe DC, Molnar A (2004) Crystal structure of p19 – a universal suppressor of RNA
silencing. Trends Biochem Sci 29:279–281
Baumberger N, Tsai CH, Lie M et al (2007) The Polerovirus silencing suppressor P0 targets
ARGONAUTE proteins for degradation. Curr Biol 17:1609–1614
Bayne EH, Rakitina DV, Morozov SY et al (2005) Cell-to-cell movement of potato potexvirus X is
dependent on suppression of RNA silencing. Plant J 44:471–482
Bennasser Y, Le SY, Benkirane M et al (2005) Evidence that HIV-1 encodes an siRNA and a
suppressor of RNA silencing. Immunity 22:607–619
Berry B, Deddouche S, Kirschner D et al (2009) Viral Suppressors of RNA silencing hinder
exogenous and endogenous small RNA pathways in Drosophila. PLoS ONE 4:e5866
88 X. Yi and R. Lu
Brennecke J, Aravin AA, Stark A et al (2007) Discrete small RNA-generating loci as master
regulators of transposon activity in Drosophila. Cell 128:1089–1103
Brigneti G, Voinnet O, Li WX et al (1998) Viral pathogenicity determinants are suppressors of
transgene silencing in Nicotiana benthamiana. EMBO J 17:6739–6746
Buck AH, Santoyo-Lopez J, Robertson KA et al (2007) Discrete clusters of virus-encoded
microRNAs are associated with complementary strands of the genome and the 7.2-kilobase
stable intron in murine cytomegalovirus. J Virol 81:13761–13770
Cai X, Schafer A, Lu S et al (2006) Epstein-Barr virus microRNAs are evolutionarily conserved
and differentially expressed. PLoS Pathog 2:e23
Carrington JC, Ambros V (2003) Role of microRNAs in plant and animal development. Science
301:336–338
Chao JA, Lee JH, Chapados BR et al (2005) Dual modes of RNA-silencing suppression by Flock
House virus protein B2. Nat Struct Mol Biol 12:952–957
Chapman EJ, Prokhnevsky AI, Gopinath K et al (2004) Viral RNA silencing suppressors inhibit
the microRNA pathway at an intermediate step. Genes Dev 18:1179–1186, Erratum in: Genes
Dev. 2004, 18:1510
Chellappan P, Vanitharani R, Fauquet CM (2005) MicroRNA-binding viral protein interferes with
Arabidopsis development. Proc Natl Acad Sci USA 102:10381–10386
Cogoni C, Irelan JT, Schumacher M et al (1996) Transgene silencing of the Al-1 gene in vegetative
cells of Neurospora is mediated by a cytoplasmic effector and does not depend on DNA-DNA
interactions or DNA methylation. EMBO J 15:3153–3163
Covey SN, Al-Kaff NS, Langara A et al (1997) Plants combat infection by gene silencing. Nature
385:781–782
Cuellar WJ, Kreuze JF, Rajam€aki ML et al (2009) Elimination of antiviral defense by viral RNase
III. Proc Natl Acad Sci USA 106:10354–10358
Cullen BR (2006) Is RNA interference involved in intrinsic antiviral immunity in mammals? Nat
Immunol 7:563–567
Czech B, Malone CD, Zhou R et al (2008) An endogenous small interfering RNA pathway in
Drosophila. Nature 453:798–802
Deleris A, Gallego-Bartolome J, Bao J et al (2006) Hierarchical action and inhibition of plant
Dicer-like proteins in antiviral defense. Science 313:68–71
Diaz-Pendon JA, Li F, Li WX et al (2007) Suppression of antiviral silencing by cucumber mosaic
virus 2b protein in Arabidopsis is associated with drastically reduced accumulation of three
classes of viral small interfering RNAs. Plant Cell 19:2053–2063
Ding SW, Voinnet O (2007) Antiviral immunity directed by small RNAs. Cell 130:413–426
Ding XS, Liu J, Cheng NH et al (2004) The Tobacco mosaic virus 126-kDa protein associated with
virus replication and movement suppresses RNA silencing. Mol Plant Microbe Interact
17:583–592
Duchaine TF, Wohlschlegel JA, Kennedy S et al (2006) Functional proteomics reveals the
biochemical niche of C. elegans DCR-1 in multiple small-RNA-mediated pathways. Cell
124:343–354
Dunoyer P, Himber C, Voinnet O (2005) DICER-LIKE 4 is required for RNA interference and
produces the 21-nucleotide small interfering RNA component of the plant cell-to-cell silencing
signal. Nat Genet 37:1356–1360
Dunoyer P, Himber C, Voinnet O (2006) Induction, suppression and requirement of RNA
silencing pathways in virulent Agrobacterium tumefaciens infections. Nat Genet 38:258–263
Dunoyer P, Lecellier CH, Parizotto EA et al (2004) Probing the microRNA and small interfering
RNA pathways with virus-encoded suppressors of RNA silencing. Plant Cell 16:1235–1250
Ebhardt HA, Thi EP, Wang MB et al (2005) Extensive 30 modification of plant small RNAs is
modulated by helper component-proteinase expression. Proc Natl Acad Sci USA
102:13398–13403
Elbashir SM, Martinez J, Patkaniowska A et al (2001) Functional anatomy of siRNAs for
mediating efficient RNAi in Drosophila melanogaster embryo lysate. EMBO J 20:6877–6888
English JJ, Mueller E, Baulcombe DC (1996) Suppression of virus accumulation in transgenic

plants exhibiting silencing of nuclear genes. Plant Cell 8:179–188
Foster TM, Lough TJ, Emerson SJ et al (2002) A surveillance system regulates selective entry of
RNA into the shoot apex. Plant Cell 14:1497–1508
Fukunaga R, Doudna JA (2009) dsRNA with 50 overhangs contributes to endogenous and antiviral
RNA silencing pathways in plants. EMBO J 28:545–555
Garcia S, Billecocq A, Crance JM et al (2006) Viral suppressors of RNA interference impair RNA
silencing induced by a Semliki Forest virus replicon in tick cells. J Gen Virol 87:1985–1989
Ghildiyal M, Seitz H, Horwich MD et al (2008) Endogenous siRNAs derived from transposons
and mRNAs in Drosophila somatic cells. Science 320:1077–1081
Girard A, Sachidanandam R, Hannon GJ et al (2006) A germline-specific class of small RNAs
binds mammalian Piwi proteins. Nature 442:199–202
Guillaume M, Olivier V (2004) Viral suppression of RNA silencing in plants. Mol Plant Pathol
5:71–82
Guo HS, Ding SW (2002) A viral protein inhibits the long range signaling activity of the gene
silencing signal. EMBO J 21:398–407
Haasnoot J, de Vries W, Geutjes E-J et al (2007) The Ebola virus VP35 protein is a suppressor of
RNA silencing. PLoS Pathog 3:e86
Hamilton AJ, Baulcombe DC (1999) A species of small antisense RNA in posttranscriptional gene
Hammond SM, Bernstein E, Beach D et al (2000) An RNA-directed nuclease mediates post-
transcriptional gene silencing in Drosophila cells. Nature 404:293–296
Hellwig S, Bass BL (2008) A starvation-induced noncoding RNA modulates expression of Dicer-
regulated genes. Proc Natl Acad Sci USA 105:12897–12902
Herr AJ, Jensen MB, Dalmay T et al (2005) RNA polymerase IV directs silencing of endogenous
DNA. Science 308:118–120
Iida T, Kawaguchi R, Nakayama J-I (2006) Conserved ribonuclease, Eri1, negatively regulates
heterochromatin assembly in fission yeast. Curr Biol 16:1459–1464
Isobe R, Kojima K, Matsuyama T et al (2004) Use of RNAi technology to confer enhanced
resistance to BmNPV on transgenic silkworms. Arch Virol 149:1931–1940
Johansen LK, Carrington JC (2001) Silencing on the spot. Induction and suppression of RNA
silencing in the Agrobacterium-mediated transient expression system. Plant Physiol
126:930–938
Kasschau KD, Carrington JC (1998) A counterdefensive strategy of plant viruses: suppression of
post-transcriptional gene silencing. Cell 95:461–470
Keese PK, Gibbs A (1992) Origins of genes: “big bang” or continuous creation? Proc Natl Acad
Sci USA 89:9489–9493
Kennedy S, Wang D, Ruvkun G (2004) A conserved siRNA-degrading RNase negatively regulates
RNA interference in C. elegans. Nature 427:645–649
Kennerdell JR, Carthew RW (1998) Use of dsRNA-mediated genetic interference to demonstrate
that frizzled and frizzled 2 act in the wingless pathway. Cell 95:1017–1026
Kumagai MH, Donson J, Della-Cioppa G et al (1995) Cytoplasmic inhibition of carotenoid
biosynthesis with virus-derived RNA. Proc Natl Acad Sci USA 92:1679–1683
Lakatos L, Csorba T, Pantaleo V et al (2006) Small RNA binding is a common strategy to suppress
RNA silencing by several viral suppressors. EMBO J 25:2768–2780
Lakatos L, Szittya G, Silhavy D et al (2004) Molecular mechanism of RNA silencing suppression
mediated by p19 protein of tombusviruses. EMBO J 23:876–884
Lecellier CH, Dunoyer P, Arar K et al (2005) A cellular microRNA mediates antiviral defense in
human cells. Science 308:557–560
Lee RC, Ambros V (2001) An extensive class of small RNAs in Caenorhabditis elegans. Science
294:862–864
90 X. Yi and R. Lu
Li F, Ding SW (2006) Virus counterdefense: Diverse strategies for evading the RNA-silencing
immunity. Annu Rev Microbiol 60:503–531
Li H, Li WX, Ding SW (2002) Induction and suppression of RNA silencing by an animal virus.
Science 296:1319–1321
Li WX, Li H, Lu R et al (2004) Interferon antagonist proteins of influenza and vaccinia viruses are
suppressors of RNA silencing. Proc Natl Acad Sci USA 101:1350–1355
Lichner Z, Silhavy D, Burgyan J (2003) Double-stranded RNA-binding proteins could suppress
RNA interference-mediated antiviral defences. J Gen Virol 84:975–980
Lindbo JA, Dougherty WG (1992) Pathogen-derived resistance to a potyvirus: immune and
resistant phenotypes in transgenic tobacco expressing altered forms of a potyvirus coat protein
nucleotide sequence. Mol Plant Microbe Interact 5:144–153
Lingel A, Simon B, Izaurralde E et al (2005) The structure of the flock house virus B2 protein, a
viral suppressor of RNA interference, shows a novel mode of double-stranded RNA recogni-
tion. EMBO Rep 6:1149–1155
Lozsa R, Csorba T, Lakatos L et al (2008) Inhibition of 30 modification of small RNAs in virus-
infected plants require spatial and temporal co-expression of small RNAs and viral silencing-
suppressor proteins. Nucleic Acids Res 36:4099–4107
Lu R, Folimonov A, Shintaku M et al (2004) Three distinct suppressors of RNA silencing encoded
by a 20-kb viral RNA genome. Proc Natl Acad Sci USA 101:15742–15747
Lu R, Maduro M, Li F et al (2005) Animal virus replication and RNAi-mediated antiviral silencing
in Caenorhabditis elegans. Nature 436:1040–1043
Lu S, Cullen BR (2004) Adenovirus VA1 noncoding RNA can inhibit small interfering RNA and
MicroRNA biogenesis. J Virol 78:12868–12876
Macrae IJ, Zhou K, Li F et al (2006) Structural basis for double-stranded RNA processing by
Dicer. Science 311:195–198
Mallory AC, Parks G, Endres MW et al (2002) The amplicon-plus system for high-level expres-
sion of transgenes in plants. Nat Biotechnol 20:622–625
Mangwende T, Wang M-L, Borth W et al (2009) The P0 gene of sugarcane yellow leaf virus
encodes an RNA silencing suppressor with unique activities. Virology 384:38–50
Matzke MA, Mette MF, Matzke AJ (2000) Transgene silencing by the host genome defense:
implications for the evolution of epigenetic control mechanisms in plants and vertebrates. Plant
Mol Biol 43:401–415
Merai Z, Kerenyi Z, Molnar A et al (2005) Aureusvirus P14 is an efficient RNA silencing
suppressor that binds double-stranded RNAs without size specificity. J Virol 79:7217–7226
Morel JB, Godon C, Mourrain P et al (2002) Fertile hypomorphic ARGONAUTE (ago1)
mutants impaired in post-transcriptional gene silencing and virus resistance. Plant Cell
14:629–639
Mourrain P, Beclin C, Elmayan T et al (2000) Arabidopsis SGS2 and SGS3 genes are required for
posttranscriptional gene silencing and natural virus resistance. Cell 101:533–542
Napoli C, Lemieux C, Jorgensen RA (1990) Introduction of a chimeric chalcone synthase gene
into Petunia results in reversible co-suppression of homologous genes in trans. Plant Cell
2:279–289
Navarro L, Dunoyer P, Jay F et al (2006) A plant miRNA contributes to antibacterial resistance by
repressing auxin signaling. Science 312:436–439
Navarro L, Jay F, Nomura K et al (2008) Suppression of the microRNA pathway by bacterial
effector proteins. Science 321:964–967
Okamura K, Lai EC (2008) Endogenous small interfering RNAs in animals. Nat Rev Mol Cell Biol
9:673–678
Pazhouhandeh M, Dieterle M, Marrocco K et al (2006) F-box-like domain in the polerovirus
protein P0 is required for silencing suppressor function. Proc Natl Acad Sci USA
103:1994–1999
Pfeffer S, Dunoyer P, Heim F et al (2002) P0 of beet Western yellows virus is a suppressor of
posttranscriptional gene silencing. J Virol 76:6815–6824
Pfeffer S, Sewer A, Lagos-Quintana M et al (2005) Identification of microRNAs of the herpesvirus

family. Nat Methods 2:269–276
Qiu W, Park JW, Scholthof HB (2002) Tombusvirus P19-mediated suppression of virus-induced
gene silencing is controlled by genetic and dosage features that influence pathogenicity. Mol
Plant Microbe Interact 15:269–280
Qu F, Morris TJ (2002) Efficient infection of Nicotiana benthamiana by tomato bushy stunt virus
is facilitated by the coat protein and maintained by p19 through suppression of gene silencing.
Mol Plant Microbe Interact 15:193–202
Qu F, Ren T, Morris TJ (2003) The coat protein of turnip crinkle virus suppresses posttranscrip-
tional gene silencing at an early initiation step. J Virol 77:511–522
Ramachandran V, Chen X (2008) Degradation of microRNAs by a family of exoribonucleases in
Arabidopsis. Science 321:1490–1492
Randall G, Panis M, Cooper JD et al (2007) Cellular cofactors affecting hepatitis C virus infection
and replication. Proc Natl Acad Sci USA 104:12884–12889
Ratcliff F, Harrison BD, Baulcombe DC (1997) A similarity between viral defense and gene
Ruiz MT, Voinnet O, Baulcombe DC (1998) Initiation and maintenance of virus-induced gene
silencing. Plant Cell 10:937–946
Segers GC, Zhang X, Deng F et al (2007) Evidence that RNA silencing functions as an antiviral
defense mechanism in fungi. Proc Natl Acad Sci USA 104:12902–12906
Shaik Syed Ali P, Chen JCH (2009) Structure of the RNA-binding domain of nodamura virus
protein B2, a suppressor of RNA interference. Biochemistry 48:2307–2309
Sijen T, Fleenor J, Simmer F et al (2001) On the role of RNA amplification in dsRNA-triggered
gene silencing. Cell 107:465–476
Sontheimer EJ (2005) Assembly and function of RNA silencing complexes. Nat Rev Mol Cell Biol
6:127–138
Su J, Oanh DT, Lyons RE et al (2008) A key gene of the RNA interference pathway in the black
tiger shrimp, Penaeus monodon: identification and functional characterisation of Dicer-1. Fish
Shellfish Immunol 24:223–233
Tijsterman M, May RC, Simmer F et al (2004) Genes required for systemic RNA interference in
Caenorhabditis elegans. Curr Biol 14:111–116
Triboulet R, Mari B, Lin YL et al (2007) Suppression of microRNA-silencing pathway by HIV-1
during virus replication. Science 315:1579–1582
Van der Vlugt RAA, Ruiter RK, Goldbach RW (1992) Evidence for sense RNA-mediated
protection to PVYN in tobacco plants transformed with the viral coat protein cistron. Plant
Mol Biol 20:631–639
Vanitharani R, Chellappan P, Pita JS et al (2004) Differential roles of AC2 and AC4 of cassava
geminiviruses in mediating synergism and suppression of posttranscriptional gene silencing.
J Virol 78:9487–9498
Vargason JM, Szittya G, Burgyan J et al (2003) Size selective recognition of siRNA by an RNA
silencing suppressor. Cell 115:799–811
Vogler H, Akbergenov R, Shivaprasad PV et al (2007) Modification of small RNAs associated
with suppression of RNA silencing by tobamovirus replicase protein. J Virol 81:10379–10388
Voinnet O (2005) Non-cell autonomous RNA silencing. FEBS Lett 579:5858–5871
Voinnet O (2009) Origin, biogenesis, and activity of plant microRNAs. Cell 136:669–687
Voinnet O, Baulcombe DC (1997) Systemic signalling in gene silencing. Nature 389:553
Voinnet O, Lederer C, Baulcombe DC (2000) A viral movement protein prevents spread of the
gene silencing signal in Nicotiana benthamiana. Cell 103:157–167
92 X. Yi and R. Lu
Voinnet O, Pinto YM, Baulcombe DC (1999) Suppression of gene silencing: a general

strategy used by diverse DNA and RNA viruses of plants. Proc Natl Acad Sci USA 96:
14147–14152
Voinnet O, Rivas S, Mestre P et al (2003) An enhanced transient expression system in plants based
on suppression of gene silencing by the p19 protein of tomato bushy stunt virus. Plant J
33:949–956
Volpe TA, Kidner C, Hall IM et al (2002) Regulation of heterochromatic silencing and histone H3
lysine-9 methylation by RNAi. Science 297:1833–1837
Wang XH, Aliyari R, Li WX et al (2006) RNA interference directs innate immunity against viruses
in adult Drosophila. Science 312:452–454
Waterhouse PM, Graham HW, Wang MB (1998) Virus resistance and gene silencing in plants can
be induced by simultaneous expression of sense and antisense RNA. Proc Natl Acad Sci USA
95:13959–13964
Winston WM, Molodowitch C, Hunter CP (2002) Systemic RNAi in C. elegans requires the
putative transmembrane protein SID-1. Science 295:2456–2459
Ye K, Malinina L, Patel DJ (2003) Recognition of small interfering RNA by a viral suppressor of
RNA silencing. Nature 426:874–878
Yelina NE, Savenkov EI, Solovyev AG et al (2002) Long-distance movement, virulence, and RNA
silencing suppression controlled by a single protein in hordei- and potyviruses: complementary
functions between virus families. J Virol 76:12981–12991
Yigit E, Batista PJ, Bei Y et al (2006) Analysis of the C. elegans Argonaute family reveals that
distinct Argonautes act sequentially during RNAi. Cell 127:747–757
Yu B, Yang Z, Li J et al (2005) Methylation as a crucial step in plant microRNA biogenesis.
Science 307:932–935
Zhang X, Yuan Y-R, Pei Y et al (2006) Cucumber mosaic virus-encoded 2b suppressor inhibits
Arabidopsis Argonaute1 cleavage activity to counter plant defense. Genes Dev 20:3255–3268
Zrachya A, Glick E, Levy Y et al (2007) Suppressor of RNA silencing encoded by tomato yellow
leaf curl virus-Israel. Virology 358:159–165
Strategies to Prevent siRNA-Triggered
Cellular Toxicity
Matthias Bauer
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2 Cellular Sensors of siRNA-Triggered Innate Immune Response . . . . . . . . . . . . . . . . . . . . . . . . . . 96
2.1 TLR-Mediated Innate Immunity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
2.2 Non-TLR-Mediated Innate Immunity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3 Cellular Sequels After siRNA-Triggered Innate Immune System Activation . . . . . . . . . . . . . . 98
4 Overcoming Synthetic siRNA-Triggered Innate Immune Response . . . . . . . . . . . . . . . . . . . . . . . 99
5 Overcoming shRNA-Triggered Innate Immune Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6 shRNA-Mediated Disruption of the Endogenous miRNA Machinery . . . . . . . . . . . . . . . . . . . . 102
7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Abstract RNA interference (RNAi) allows selective gene silencing, is widely used
for functional analysis of individual genes in mammalian cells, and represents an
attractive therapeutic option for treating various diseases. However, growing evi-
dence exists that chemically synthesized small interfering RNAs (siRNAs) as well
as promoter-expressed short-hairpin RNAs (shRNAs) may cause cellular toxicity
resulting in unspecific cellular phenotypes and, in case of therapeutic interventions,
severe side effects. Various mechanisms have been identified including the induc-
tion of interferon-stimulated gene (ISG) expression as well as the disruption of
natural microRNA biogenesis and function by competition of exogenous shRNAs
for the endogenous RNAi machinery. This review highlights recent progress in
the understanding of siRNA-triggered toxicity and outlines strategies to prevent
undesirable side effects.
M. Bauer
Department of Neurology, Hertie Institute for Clinical Brain Research, University of T€
ubingen,
T€ubingen, Germany
Department of Protein Sciences, Helmholtz Center M€ unchen, German Research Center for
Environmental Health, M€unchen-Neuherberg, Germany
Institute for Human Genetics, Klinikum rechts der Isar, TU M€unchen, M€unchen, Germany
e-mail: matthias.bauer@helmholtz-muenchen.de
94 M. Bauer
Keywords siRNA shRNA Off-target effects Innate immune response
Abbreviations

IFN interferon
IPS-1 IFNb promoter stimulator 1
ISG interferon-stimulated gene
PKR dsRNA-dependent protein kinase
OAS oligo adenylate synthetase enzymes
PAMPs pathogen-associated molecular patterns
RIG-I retinoic acid inducible gene I
shRNAs short-hairpin RNAs
siRNAs small interfering RNAs
TLRs Toll-like receptors
1 Introduction
The discovery of RNA interference (RNAi) in 2001 in mammalian cells (Elbashir

et al. 2001a) provided a tool unimagined before, to silence any gene with a known
sequence and opened exciting new possibilities for basic science research as well as
for therapeutic purposes. RNAi is an evolutionarily conserved cellular process that
regulates gene expression (Carthew and Sontheimer 2009) and is thought to be part
of the innate viral defence machinery (Lecellier et al. 2005). Initially, chemically
synthesized small interfering RNAs (siRNA) with a length of 21–22 nucleotides
and 30 dinucleotide overhangs have been used to target mammalian genes by RNAi
(Elbashir et al. 2001b) (Fig. 1a). Later on, it has been demonstrated that endoge-
nously expressed short-hairpin RNAs (shRNAs), which bear a fold-back stem–loop
structure, can also mediate gene silencing (Brummelkamp et al. 2002; Paddison
et al. 2002) (Fig. 1b). Thus, siRNA delivery methods to cells fall into two major
categories: (1) promotor-driven expression of shRNAs from a viral or a plasmidal
vector, which are then cleaved into small 21–25 nucleotide double-strand RNAs
(dsRNAs) by the enzymes Drosha and/or Dicer and (2) exogenous delivery of 21
dsRNAs with 30 dinucleotide overhangs that mimic Dicer cleavage products.
Exogenously delivered negatively charged dsRNAs do not cross the negatively
charged cellular membrane in sufficient quantities; thus, carriers including cationic
lipids (Tseng et al. 2009), polycationic nanoparticles (Bartlett et al. 2007), or other
polymers have been tested and successfully applied to shuttle synthetic siRNAs into
Strategies to Prevent siRNA-Triggered Cellular Toxicity 95
a
Conventional siRNA
N(21)
5′- NNNNNNNNNNNNNNNNNNNNN U U -3′
3′- U U NNNNNNNNNNNNNNNNNNNNN -5′
Antisense strand
b
“1stGeneration” shRNA
N(21) C
U A
5′- NNNNNNNNNNNNNNNNNNNNNU
A
3′- U U NNNNNNNNNNNN NNNNNNNNN A
G G
Antisense strand A
c
miR-30
A
5′- G C G C U G U A A A C A U C C G U C A C U GG A A G C U GU G A A
3′- C GU C G A C G U U U G U A G G C U G A C U U U C GG CAC C G
G UA GA
Drosha Dicer
d
“2ndGeneration” miRNA-designed shRNA
N(13) N(8)
A UC A
5′- G C G N N N N N N N N N N N N N NNNNNNN N C U GUG A
3′- C GU C NNNNNNNNNNNN N NNNNNNN N GG CA C C G
GU AGA
Antisense strand
Fig. 1 Schematic structure of a synthetic siRNA, first and second generation miRNA-designed
shRNA. (a) Structure of a conventional siRNA with a length of 21 nucleotides and 30 dinucleotide
overhangs. (b) first generation shRNA construct with a stem–loop–stem structure (Brummelkamp
et al. 2002). The sequence of the target site (sense orientation) is shown in blue (passenger strand)
and the antisense strand (guide strand) is shown in red. (c) Endogenous miR-30 primary transcript;
putative cleavage sites for Dicer and Drosha are indicated with arrows. (d) second generation
miR-30-designed shRNA construct
the cytosol [for comprehensive overview see Juliano et al. (2008)]. In the case of
widely used cationic lipid reagents, siRNA-containing lipoplexes are internalized
by endocytosis followed by entering the endosomal/lysosomal pathway after cellu-
lar uptake (Doherty and McMahon 2009; Zuhorn et al. 2002). After destabilization
of the endosomal membrane by carrier cationic lipids and subsequent dissociation
from the carrier, dsRNA is released into the cytosol and enters the RNAi pathway
by its incorporation into the RNA-induced silencing complex (RISC) (Carthew and
Sontheimer 2009).
96 M. Bauer
In contrast, intracellularly expressed shRNAs from either plasmidal or viral

vectors under the control of polymerase II or polymerase III promoters (e.g., H1
or U6) are exported into the cytoplasm by exportin-5 (Yi et al. 2003) and processed
by Dicer (Provost et al. 2002), which produces the mature 19–25 nucleotide double-
stranded siRNAs. As in exogenously applied synthetic siRNAs, the shRNA-derived
dsRNA antisense strand incorporates together with its complementary target
mRNA into RISC resulting in Argonaute-2-mediated mRNA cleavage and
sequence-specific knockdown of target gene expression (Carthew and Sontheimer
2009).
Over the past 7 years, a huge number of in vitro as well as in vivo studies have
provided strong evidence that almost every gene with a known sequence can be a
target for RNAi. However, results obtained from unbiased gene expression profiling
of siRNA studies in vitro disclosed that RNAi technology in general might be
compromised by off-target effects (Jackson et al. 2003; Sledz et al. 2003). Since
siRNAs can also function through a miRNA-like mechanism, imperfect comple-
mentarity to 30 UTRs of other than the target mRNA might lead to translational
repression and/or degradation of nontarget mRNAs (Carthew and Sontheimer 2009;
Birmingham et al. 2006). In addition, the so-called “nonspecific” off-target effects
(Jackson and Linsley 2004) caused by siRNA-induced interferon response or the
disturbance of the endogenous miRNA pathway may lead to global changes in gene
expression, unspecific effects, and cellular toxicity, which profoundly compromise
the safety of RNAi-based therapeutic approaches. The purpose of the current article
is to highlight some of the recently emerging insights into nonspecific off-target
effects and to outline strategies to prevent those by modifications of siRNA/
shRNA-construct design.
2 Cellular Sensors of siRNA-Triggered Innate

Immune Response
Two years after the first demonstration of RNAi in mammalian cells, two groups
have independently shown that chemically synthesized 21 nucleotide siRNAs
(Sledz et al. 2003) as well as lentivirally expressed shRNAs (Bridge et al. 2003)
resulted in a global upregulation of interferon-stimulated genes (ISG) in targeted
cells. In mammalian cells, several membrane-bound as well as cytosolic receptors
exist for the detection of dsRNAs and ssRNAs, which belong to the cellular
antiviral defence machinery (de Veer et al. 2005; Garcı́a-Sastre and Biron 2006).
Innate immunity to siRNA can be classified as (1) Toll-like receptor (TLR)-
mediated and (2) non-TLR-mediated (Robbins et al. 2009). As a general rule, it is
believed that endogenously expressed shRNAs predominantly activate non-TLR-
mediated dsRNA sensors, whereas exogenously delivered siRNAs, which have
entered the endosomal/lysosomal pathway, may activate both the TLR-mediated
as well as the non-TLR-mediated innate immune system.
2.1 TLR-Mediated Innate Immunity
Thirteen TLR receptors have been identified in human and mouse cells and three –
TLR3, 7, and 8 – have been linked to siRNA-mediated activation of the innate
immune system so far. In humans, TLR3 is expressed in myeloid dendritic cells
(Muzio et al. 2000) and primary endothelial cells (Kleinman et al. 2008), whereas
its expression in mice is found in a wider range of cell types (Applequist et al.
2002). At the subcellular level, TLR3 is expressed at the cell surface as well as in
the endosome and might therefore detect siRNAs in the blood stream and cationic
lipid-complexed siRNAs after endocytosis by leukocytes. After in vivo delivery of
21 nucleotide siRNAs, TLR3 on endothelial cells has been activated followed by
the expression of several cytokines including IFNg and IL-12 (Kleinman et al.
2008), which then can eventually trigger a more widespread immune response.
TLR7 and TLR8 are expressed in dendritic cells and other immune cell types in
humans and are located in the endosomal membrane (Zarember and Godowski
2002). siRNA-triggered TLR7 activation in primary immune cells resulted in the
induction of type 1 interferon expression and expression of other proinflammatory
cytokines (Karikó et al. 2004; Hornung et al. 2005; Gorden et al. 2005). Unlike
TLR3, TLR7 activation seems to be dependent on the presence of immunostimu-
latory motifs including GU-rich sequences (Judge et al. 2005). Interestingly, TLR8
does not respond to conventional TLR7/8 ligands in murine immune cells (Gorden
et al. 2006). In fact, TLR8 expression has been detected in the embryonic mouse
brain localizing in axonal structures, which points towards a function of TLRs in
murine neuronal development (Ma et al. 2007) and reminds us of the importance to
consider interspecies differences in innate immune activation. Even more complex,
different culture conditions as well as cellular stress may by itself modulate or
induce components of the TLR pathway and may prime cells to respond to siRNAs
in an unintended way (Robbins et al. 2009).
2.2 Non-TLR-Mediated Innate Immunity
Non-TLR, cytoplasmic RNA sensors include dsRNA-dependent protein kinase

(PKR) (Meurs et al. 1990), oligo adenylate synthetase enzymes (OAS) catalyzing
the conversion of ATP to 20 –50 oligoadenylates (Merlin et al. 1983), and the RNA
helicase retinoic acid inducible gene I (RIG-I) (Yoneyama et al. 2004). Unlike
TLR7/8, these molecules are not located in the endosome compartment and can
therefore detect shRNA-derived dsRNAs, which have not entered the endosomal/
lysosomal pathway. In addition, the expression of non-TLR sensors goes far beyond
immune cells and its activation might therefore represent the only mechanism by
which foreign RNA is detected in solid tissue organs including brain or liver.
PKR is a serin–threonine kinase and is expressed in most mammalian cells (Meurs
et al. 1990). At the subcellular level, PKR protein is localized predominately in
98 M. Bauer
the cytoplasm and is associated with ribosomes (Thomis et al. 1992). Although it
has been previously thought that activation of PKR requires >30 base pair dsRNA,
(Manche et al. 1992), kinase assays have disclosed that synthetic 21 nucleotide
siRNAs can induce PKR to some extent (Zhang et al. 2006; Lemaire et al. 2008).
Activated PKR leads to the phosphorylation of elongation/initiation factor eIF-
2alpha and blocks overall cellular translation (de Veer et al. 2005). siRNA-acti-
vated PKR also activates transcription factors including NFkB, IRFs, and ATF2,
which then trigger interferon (IFN) expression within an affected cell (Garcı́a-
Sastre and Biron 2006). IFN then binds to cell surface receptors in an auto- and/
or paracrine fashion and confers a more global antiviral state by inducing a complex
array of additional IFN-stimulated genes (Garcı́a-Sastre and Biron 2006).
Various pathogen-associated molecular patterns (PAMPs) helicases including
RIG-I have been identified as cytoplasmic sensors of foreign RNAs in human cells.
RIG-I acts as a sensor for either single-stranded or double-stranded blunt-ended
RNAs containing uncapped 50 -triphosphates, which is a characteristic of some viral
RNAs (Hornung et al. 2006) and is present in phage polymerase-transcribed
siRNAs (Kim et al. 2004). When activated, RIG-I interacts with IFNb promoter
stimulator 1 (IPS-1) followed by signal transmission to IRF transcription factors
and NFkB, which then leads to the synthesis of proinflammatory cytokines and type
1 IFNs (Yoneyama et al. 2005).
So far, three members of the OAS family, OAS1 to OAS3, have been identified
(Hovnanian et al. 1998). dsRNA-activated OAS enzymes convert ATP to 20 –50 -
oligoadenylates, which then activate the cellular endoribonuclease RNase L. Acti-
vated RNase L unselectively cleaves mRNAs leading to a general inhibition of
protein synthesis (de Veer et al. 2005) and apoptosis (Li et al. 2004). Synthetical 21
nucleotide dsRNAs (Sledz et al. 2003) as well as shRNA-derived siRNAs (Bridge
et al. 2003) are capable of activating OAS proteins, and it has been shown
previously that even dsRNAs below a length of 20 nucleotides can activate OAS-
mediated immune response (Sarkar et al. 1999).
3 Cellular Sequels After siRNA-Triggered Innate Immune

System Activation
Depending on the origin and developmental stage, cells are differentially sensitive
to siRNA-triggered IFN response. We have tested over 30 different shRNA-expressing
(Fig. 1b) lentiviral vectors in various cell lines as well as in primary embryonic
mouse neuronal cultures. Whereas none of the shRNA constructs induced an IFN
response in the tested cell lines (i.e., NIH3t3, 293T, HELA, and COS cells), shRNA
expression in primary cultures almost regularly resulted in elevated Oas1 expres-
sion accompanied by significant adverse cellular phenotypes (own unpublished
observations). Moreover, different neuronal cell populations from different parts
of the embryonic brain seem differentially susceptible and even cells derived from
the same anatomical brain region have shown different degrees of IFN response
depending on the embryonic age of the donor (own unpublished observations).
The induction of interferon-stimulated genes (ISGs) by synthetic siRNAs as well
as shRNA-derived siRNAs can result in a multitude of morphological changes and
cellular phenotypes in vitro and in vivo. For example, we have shown previously
that shRNA-triggered activation of innate antiviral response in primary cortical
neurons resulted in neurite retraction and apoptotic cell death (Bauer et al. 2009).
In another study with primary hippocampal neurons, shRNA-induced IFN-response
resulted in more subtle morphological changes with rarefication of dendritic spines
and electrophysiological perturbations (Alvarez et al. 2006). In vivo experiments
with transgenic mice expressing shRNAs directed against arylamine N-acetyltransferase
showed increased early lethality, which was – at least in part – associated with
elevated Oas1 mRNA levels (Cao et al. 2005) and adenovirus-mediated shRNA
expression in the striatum of adult mice resulted in loss of striatal neuron identity
and microglial activation (McBride et al. 2008). However, in the latter study, ISG
expression has not been monitored and therefore other shRNA or vector-related
immunological sequels than IFN response may be involved.
Due to the aforementioned findings and in consideration of the fact that siRNA-
based technologies have been and will be applied extensively in basic research as
well as for therapeutic approaches, there is a strong need for the identification of
determinants triggering IFN response and to invent strategies to prevent cellular
immune response-mediated toxicity.
4 Overcoming Synthetic siRNA-Triggered Innate

Immune Response
Over the past years, extensive research has been conducted to identify IFN inducing
determinants of siRNAs, and it became clear that RNAi mediated by different RNA
sources – i.e., synthetic siRNA and shRNA-derived siRNAs – are associated with
different modes of innate immune system activation. One of the obvious advantages
in using synthetic siRNAs is that they are clearly defined in length and structure and
that chemical modifications can be introduced to abrogate or at least reduce
unintended immunostimulatory effects.
Undoubtedly, the size of applied dsRNAs is one of the major factors for the
induction of innate immune response, whereby duplexes shorter than 20 base pairs –
at least in HeLa S3 cells – do not activate PKR and TLR3 even at high RNA
concentrations (100 nM). With increased siRNA length, the RNA concentration
threshold for IFN response induction continuously declined, showing toxic effects
with 23 base pair siRNA at concentrations as low as 10 nM (Reynolds et al. 2006).
In another study, OAS protein family sensors have been activated by 21 base
pair siRNAs at even lower concentrations in RCC1 cells (Sledz et al. 2003),
which strongly suggests the use of 19–21 base pair siRNAs at the lowest effective
concentration.
100 M. Bauer
To circumvent activation of RIG-I and other helicases, blunt-ended dsRNA and

50 -triphosphates should be avoided and 2 nucleotide 30 overhangs should be intro-
duced instead.
The avoidance of immunostimulatory GU-rich sequences seems to be the “conditio
sine qua non” in the design of synthetic siRNAs to lower the risk for TLR7/8 activation
in immune cells. However, it has been shown previously that TLR7 might be
generally activated in the presence of ribose sugar backbone and uridine, two
biochemical features characteristic of RNA molecules (Diebold et al. 2006). There-
fore, the introduction of 20 -O-methyl-modified nucleotides in the siRNA ribose
sugar backbone is recommended, since this modification inhibits TLR7/8 activation
and prevents toxicity due to the activation of type I interferon pathway gene
expression (Judge et al. 2006; Robbins et al. 2007).
5 Overcoming shRNA-Triggered Innate Immune Response
shRNAs, which are endogenously expressed from plasmidal or viral vectors differ
from exogenously administered siRNAs regarding the involvement of cellular
mechanisms responsible for the activation of innate immune response, and there-
fore different strategies must be taken into consideration. In general, it is believed
that endogenously expressed shRNAs are more toxic to cells than synthetic siRNA
delivery and that some IFN inducing determinants cannot be readily controlled due
to prior endogenous processing of shRNAs before they become functional siRNAs.
For example, due to variable Dicer cleavage and heterogenous 30 -ends caused by
the polyT transcription termination signal, siRNAs derived from polymerase III
promoter-transcribed shRNAs have shown a marked heterogeneity in length
(21–25 nucleotides) (Olejniczak et al. 2009). However, polymerase II and III
driven expression of shRNA is still the method of choice when long-lasting
silencing effects are required and considerable progress has been made to make
shRNA-mediated RNAi a safer and more predictable gene silencing tool.
After the first report in 2003 showing IFN induction after lentivirus-mediated
overexpression of shRNAs in human lung fibroblasts (Bridge et al. 2003), the same
group published a paper in which they systematically modified the U6 promoter
transcription start site and 50 shRNA sequences showing that an AA-dinucleotide
near the transcription start site has been the prime determinant for ISG induction
(Pebernard and Iggo 2004). In contrast, removal of the AA-dinucleotide at the
transcription initiation site in a H1-driven shRNA construct capable to induce ISG
expression in primary cortical neurons did not abolish activation of innate immune
response (Bauer et al. 2009). This finding indicates that U6 promoter-driven shRNA
expression differs from H1-driven constructs with respect to stimulate dsRNA-
triggered immune response and general rules regarding polymerase III promoter
sequences may not have been deduced so far. However, it seems more favorable to
express shRNAs under the control of the H1 polymerase III promoter, since it has
been repeatedly reported that U6 vectors give a higher frequency of ISG induction
(Bridge et al. 2003; Pebernard and Iggo 2004).
In contrast to exogenously administered siRNAs, activation of innate immune
response by specific sequences of the guide and/or passenger strand of shRNA
constructs is less clear. Although chemically synthesized shRNAs may theoreti-
cally activate TLRs within the endosome in the presence of GU-rich immunosti-
mulatory sequence motifs (Karikó et al. 2004), endogenously vector-mediated
expression of shRNAs is rather unlikely to do so, since the resulting dsRNAs
normally do not enter the endosomes, therefore bypassing TLR recognition.
Thus, the presence of immunostimulatory sequence motifs in shRNAs expressed
from either plasmidal or viral vectors may not be the cause and its avoidance
might not be sufficient to circumvent interferon response, when present (Robbins
et al. 2006).
Due to the fact that a multitude of endogenously expressed, cell-derived short
dsRNA species, including miRNAs, are present in the cytosol (H€uttenhofer et al.
2005; Carthew and Sontheimer 2009), it is crucial that cellular sensors, known as
pattern-recognition receptors, can distinguish between self and nonself dsRNAs. It
is therefore reasonable to think that the implementation of design features of
naturally occurring short dsRNAs in shRNA constructs lowers the risk of induction
of innate immune response. A proof of principal study conducted by the Cullen
laboratory in 2002 has shown that the introduction of target gene-derived sequences
within a miRNA-30 backbone was capable of decreasing the complementary
mRNA target in human cells, providing first experimental evidence that artificial
miRNAs can be used as siRNA shuttles (Zeng et al. 2002). Following studies
demonstrated that miRNA-designed or “second generation” shRNAs are (1) more
efficient in gene silencing than “first generation” shRNA constructs (Boden et al.
2004; Bauer et al. 2009) and (2) seem to circumvent induction of ISG expression at
least in primary neurons in vitro (Bauer et al. 2009 and unpublished data). Interest-
ingly, only an exact implementation of all design features of naturally occurring
miRNA-30 precursors, including the introduction of a dinucleotide bulge structure
within the stem duplex was capable of avoiding induction of ISGs (Bauer et al.
2009) (Fig. 1c, d). In addition, switching from first generation to second generation
miRNA-designed shRNA constructs in a study targeting huntingtin expression in
the CNS resulted in reduced microglia activation and inflammation in vivo
(McBride et al. 2008). These studies give a strong indication that the use of
miRNA-based shRNA constructs is favorable with respect to the avoidance of
toxic, immune response-related side-effects in target cells.
In general, it seems advantageous when sh/siRNAs comprise natural features of
endogenous small RNA species (i.e., O-methylation of the ribose sugar backbone of
synthetic siRNAs, inclusion of bulges in the passenger strand of shRNAs, and usage
of miRNA-designed shRNAs) to avoid activation of innate viral defense machin-
ery. Therefore, ongoing research to understand the components and mechanisms of
the endogenous si/miRNA machinery as well as innate viral defense mechanisms
will further help to improve and rationalize the design of synthetic siRNA and
shRNA constructs.
102 M. Bauer
6 shRNA-Mediated Disruption of the Endogenous

miRNA Machinery
The safety of first generation shRNA constructs is not only compromised by

activation of innate immune response but also by its competition with endogenous
miRNA biogenesis. Referring to this, Grimm and coworks have shown for the first
time that sustained AAV8 vector-driven shRNA expression in livers of adult mice
caused hepatocellular toxicity and death in a significant proportion of animals.
Liver cell toxicity has been correlated with (1) shRNA expression levels, (2)
shRNA stem length and (3) the decrease of endogenous liver miRNA levels
(Grimm et al. 2006). In addition, indirect experimental evidence has been provided
that shRNA-related saturation of the exportin-5-mediated nuclear export (Yi et al.
2003) is one of the molecular mechanisms for miRNA inhibition (Grimm et al.
2006). However, it has been shown in another study that exogenously delivered
shRNA and siRNAs may compete with miRNA biosynthesis in vitro, suggesting
that additional components of the endogenous siRNA/miRNA pathway down-
stream of exportin-5 can also be saturated (Castanotto et al. 2007).
Interestingly, miRNA-designed shRNA constructs seem to avoid interference
with endogenous miRNA biogenesis. Boudreau and coworkers have shown that
although first generation shRNAs are expressed at higher levels than miRNA-
designed shRNAs, the latter are processed more efficiently (Boudreau et al. 2008).
As one possible consequence, they then showed in an shRNA/miRNA coexpression
assay that miRNA processing is compromised by forced first generation shRNA
expression. Moreover, first generation shRNA expression in C2C12 myoblasts
entirely inhibited endogenous miR-1 target silencing, whereas miRNA-designed
shRNAs did not compromise miR-1 function (Boudreau et al. 2009). Using the same
shRNA constructs in vivo, miRNA-designed shRNAs targeting ataxin-1 in SCA1
mice resulted in gene silencing with preserved morphologic integrity of the cerebel-
lum, whereas a loss of Purkinje neurons has been observed with AAV-mediated
expression of non miRNA-designed shRNA constructs (Boudreau et al. 2009).
Although combining the two aforementioned in vitro and in vivo observations
may lead to the conclusion that the disturbance of the endogenous miRNA pathway
highlighted by decreased miR-1 levels in a nonneuronal cell line is connected to
cerebral neuron toxicity in first generation shRNAs, the authors did neither provide
data on endogenous miRNA expression in transduced cerebellar tissue, nor did they
monitor for increased expression of ISG. Thus, this particular study raises an
important issue, since the disruption of endogenous miRNA biogenesis in CNS
tissue results in apoptosis and neuronal death [for detailed review see Hébert and De
Strooper (2009)], but did not provide a sufficient experimental link between first
generation shRNA expression, disruption of miRNA pathway, and neurotoxicity
in vivo and even more important, the capability of second generation shRNAs to
avoid these side-effects. Therefore, further in vivo studies are needed to clearly
identify the nature of cytotoxic effects after intracerebral shRNA applications.
7 Conclusions
So far, therapeutic applications in clinical trials based on RNAi technology used

direct introduction of synthetic siRNAs. The advantage of chemical-synthesized
siRNAs is that they are well-defined molecules, which are accessible to chemical
modifications to increase silencing efficacy/specificity and decrease the risk for
toxic side effects. However, medical conditions may demand long-lasting gene
silencing and may exclude repeated application of synthetic siRNAs – for example
when gene expression should be targeted in specific areas of the brain. Therefore,
single application of vector-expressed shRNAs might fill this therapeutic gap.
Although several other features related to shRNA expression by viral or plasmidal
vectors are still under intensive investigation and negotiation, marked progress has
been achieved in optimizing shRNA design in recent years. For example, highly
efficient miRNA-designed shRNAs capable of circumventing the induction of
innate immune response and preserving the integrity of the endogenous miRNA
machinery in target cells represent a big step towards a safer molecular tool for gene
silencing. In addition, miRNA-based vectors are amenable to polymerase II-driven
transcription, which allows regulated and/or cell-specific expression of inhibitory
RNAs. Thus, the combination of these safety features may further decrease the risk
of cellular toxicity in therapeutic approaches and provide a more reliable molecular
tool for basic research as well as for therapeutic applications.
References
Alvarez VA, Ridenour DA, Sabatini BL (2006) Retraction of synapses and dendritic spines
induced by off-target effects of RNA interference. J Neurosci 26:7820–7825
Applequist SE, Wallin RP, Ljunggren HG (2002) Variable expression of Toll-like receptor in
murine innate and adaptive immune cell lines. Int Immunol 14:1065–1074
Bartlett DW, Su H, Hildebrandt IJ et al (2007) Impact of tumor-specific targeting on the
biodistribution and efficacy of siRNA nanoparticles measured by multimodality in vivo imag-
ing. Proc Natl Acad Sci USA 104:15549–15554
Bauer M, Kinkl N, Meixner A et al (2009) Prevention of interferon-stimulated gene expression
using microRNA-designed hairpins. Gene Ther 16:142–147
Birmingham A, Anderson EM, Reynolds A et al (2006) 30 UTR seed matches, but not overall
identity, are associated with RNAi off-targets. Nat Methods 3:199–204
Boden D, Pusch O, Silbermann R et al (2004) Enhanced gene silencing of HIV-1 specific siRNA
using microRNA designed hairpins. Nucleic Acids Res 32:1154–1158
Boudreau RL, Monteys AM, Davidson BL (2008) Minimizing variables among hairpin-based
RNAi vectors reveals the potency of shRNAs. RNA 14:1834–1844
Boudreau RL, Martins I, Davidson BL (2009) Artificial microRNAs as siRNA shuttles: improved
safety as compared to shRNAs in vitro and in vivo. Mol Ther 17:169–175
Bridge AJ, Pebernard S, Ducraux A et al (2003) Induction of an interferon response by RNAi
vectors in mammalian cells. Nat Genet 34:263–264
Brummelkamp TR, Bernards R, Agami R (2002) A system for stable expression of short interfer-
ing RNAs in mammalian cells. Science 296:550–553
104 M. Bauer
Cao W, Hunter R, Strnatka D et al (2005) DNA constructs designed to produce short hairpin,
interfering RNAs in transgenic mice sometimes show early lethality and an interferon
response. J Appl Genet 46:217–225
Carthew RW, Sontheimer EJ (2009) Origins and mechanisms of miRNAs and siRNAs. Cell
136:642–655
35:5154–5164
de Veer MJ, Sledz CA, Williams BR (2005) Detection of foreign RNA: implications for RNAi.
Immunol Cell Biol 83:224–228
Diebold SS, Massacrier C, Akira S et al (2006) Nucleic acid agonists for Toll-like receptor 7 are
defined by the presence of uridine ribonucleotides. Eur J Immunol 36:3256–3267
Doherty GJ, McMahon HT (2009) Mechanisms of endocytosis. Annu Rev Biochem 78:
857–902
Elbashir SM, Lendeckel W, Tuschl T (2001b) RNA interference is mediated by 21- and 22-
nucleotide RNAs. Genes Dev 15:188–200
Garcı́a-Sastre A, Biron CA (2006) Type 1 interferons and the virus-host relationship: a lesson in
détente. Science 312:879–882
Gorden KB, Gorski KS, Gibson SJ et al (2005) Synthetic TLR agonists reveal functional
differences between human TLR7 and TLR8. J Immunol 174:1259–1268
Gorden KK, Qiu XX, Binsfeld CC et al (2006) Cutting edge: activation of murine TLR8 by a
combination of imidazoquinoline immune response modifiers and polyT oligodeoxynucleo-
tides. J Immunol 177:6584–6587
Hébert SS, De Strooper B (2009) Alterations of the microRNA network cause neurodegenerative
disease. Trends Neurosci 32:199–206
Hornung V, Guenthner-Biller M, Bourquin C et al (2005) Sequence-specific potent induction of
IFN-alpha by short interfering RNA in plasmacytoid dendritic cells through TLR7. Nat Med
11:263–270
Hornung V, Ellegast J, Kim S et al (2006) 50 -Triphosphate RNA is the ligand for RIG-I. Science
314:994–997
Hovnanian A, Rebouillat D, Mattei MG et al (1998) The human 20 ,50 -oligoadenylate synthetase
locus is composed of three distinct genes clustered on chromosome 12q24.2 encoding the 100-,
69-, and 40-kDa forms. Genomics 52:267–277
H€uttenhofer A, Schattner P, Polacek N (2005) Non-coding RNAs: hope or hype? Trends Genet
21:289–297
Jackson AL, Linsley PS (2004) Noise amidst the silence: off-target effects of siRNAs? Trends
Genet 20:521–524
Jackson AL, Bartz SR, Schelter J et al (2003) Expression profiling reveals off-target gene
regulation by RNAi. Nat Biotechnol 21:635–637
Judge AD, Sood V, Shaw JR et al (2005) Sequence-dependent stimulation of the mammalian
innate immune response by synthetic siRNA. Nat Biotechnol 23:457–462
Judge AD, Bola G, Lee AC et al (2006) Design of noninflammatory synthetic siRNA mediating
potent gene silencing in vivo. Mol Ther 13:494–505
Juliano R, Alam MR, Dixit V et al (2008) Mechanisms and strategies for effective delivery of
antisense and siRNA oligonucleotides. Nucleic Acids Res 36:4158–4171
Karikó K, Bhuyan P, Capodici J et al (2004) Small interfering RNAs mediate sequence-indepen-
dent gene suppression and induce immune activation by signaling through toll-like receptor 3.
J Immunol 172:6545–6549
Kim DH, Longo M, Han Y et al (2004) Interferon induction by siRNAs and ssRNAs synthesized
by phage polymerase. Nat Biotechnol 22:321–325
Lecellier CH, Dunoyer P, Arar K et al (2005) A cellular microRNA mediates antiviral defense in
human cells. Science 308:557–560
Lemaire PA, Anderson E, Lary J et al (2008) Mechanism of PKR Activation by dsRNA. J Mol Biol
381:351–360
Li G, Xiang Y, Sabapathy K et al (2004) An apoptotic signaling pathway in the interferon antiviral
response mediated by RNase L and c-Jun NH2-terminal kinase. J Biol Chem 279:1123–1131
Ma Y, Haynes RL, Sidman RL et al (2007) TLR8: an innate immune receptor in brain, neurons and
axons. Cell Cycle 6:2859–2868
Manche L, Green SR, Schmedt C et al (1992) Interactions between double-stranded RNA
regulators and the protein kinase DAI. Mol Cell Biol 12:5238–5248
McBride JL, Boudreau RL, Harper SQ et al (2008) Artificial miRNAs mitigate shRNA-mediated
toxicity in the brain: implications for the therapeutic development of RNAi. Proc Natl Acad Sci
USA 105:5868–5873
Merlin G, Chebath J, Benech P et al (1983) Molecular cloning and sequence of partial cDNA for
interferon-induced (20 -50 )oligo(A) synthetase mRNA from human cells. Proc Natl Acad Sci
USA 80:4904–4908
Meurs E, Chong K, Galabru J (1990) Molecular cloning and characterization of the human double-
stranded RNA-activated protein kinase induced by interferon. Cell 62:379–390
Muzio M, Bosisio D, Polentarutti N et al (2000) Differential expression and regulation of toll-like
receptors (TLR) in human leukocytes: selective expression of TLR3 in dendritic cells.
J Immunol 164:5998–6004
Olejniczak M, Galka P, Krzyzosiak WJ (2010) Sequence-non-specific effects of RNA interference
triggers and microRNA regulators. Nucleic Acids Res. 38:1–16
Pebernard S, Iggo RD (2004) Determinants of interferon-stimulated gene induction by RNAi
vectors. Differentiation 72:103–111
Provost P, Dishart D, Doucet J (2002) Ribonuclease activity and RNA binding of recombinant
human Dicer. EMBO J 21:5864–5874
Reynolds A, Anderson EM, Vermeulen A et al (2006) Induction of the interferon response by
siRNA is cell type- and duplex length-dependent. RNA 12:988–993
Robbins MA, Li M, Leung I, Li H et al (2006) Stable expression of shRNAs in human CD34þ
progenitor cells can avoid induction of interferon responses to siRNAs in vitro. Nat Biotechnol
24:566–571
Robbins M, Judge A, Liang L et al (2007) 20 -O-methyl-modified RNAs act as TLR7 antagonists.
Mol Ther 15:1663–1669
Robbins M, Judge A, MacLachlan I (2009) siRNA and innate immunity. Oligonucleotides
19:89–102
Sarkar SN, Bandyopadhyay S, Ghosh A et al (1999) Enzymatic characteristics of recombinant
medium isozyme of 20 -50 oligoadenylate synthetase. J Biol Chem 274:1848–1855
Sledz CA, Holko M, de Veer MJ et al (2003) Activation of the interferon system by short-
interfering RNAs. Nat Cell Biol 5:834–839
Thomis DC, Doohan JP, Samuel CE (1992) Mechanism of interferon action: cDNA structure,
expression, and regulation of the interferon-induced, RNA-dependent P1/eIF-2 alpha protein
kinase from human cells. Virology 188:33–46
Tseng YC, Mozumdar S, Huang L (2009) Lipid-based systemic delivery of siRNA. Adv Drug
Deliv Rev 61:721–731
Yi R, Qin Y, Macara IG, Cullen BR (2003) Exportin-5 mediates the nuclear export of pre-
microRNAs and short hairpin RNAs. Genes Dev 17:3011–3026
106 M. Bauer
Yoneyama M, Kikuchi M, Natsukawa T et al (2004) The RNA helicase RIG-I has an essential
function in double-stranded RNA-induced innate antiviral responses. Nat Immunol 5:730–737
Yoneyama M, Kikuchi M, Matsumoto K et al (2005) Shared and unique functions of the DExD/
H-box helicases RIG-I, MDA5, and LGP2 in antiviral innate immunity. J Immunol 175:
2851–2858
Zarember KA, Godowski PJ (2002) Tissue expression of human Toll-like receptors and differen-
tial regulation of Toll-like receptor mRNAs in leukocytes in response to microbes, their
products, and cytokines. J Immunol 168:554–561
Zeng Y, Wagner EJ, Cullen BR (2002) Both natural and designed micro RNAs can inhibit the
expression of cognate mRNAs when expressed in human cells. Mol Cell 9:1327–1333
Zhang Z, Weinschenk T, Guo K et al (2006) siRNA binding proteins of microglial cells: PKR is an
unanticipated ligand. J Cell Biochem 97:1217–1229
Zuhorn IS, Kalicharan R, Hoekstra D (2002) Lipoplex-mediated transfection of mammalian cells
occurs through the cholesterol-dependent clathrin-mediated pathway of endocytosis. J Biol
Chem 277:18021–18028
RNAi in Malignant Brain Tumors: Relevance
to Molecular and Translational Research
Mitsutoshi Nakada, Daisuke Kita, Yutaka Hayashi, Kazuyuki Kawakami,

Jun-ichiro Hamada, and Toshinari Minamoto
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
1.1 Diagnostic Characteristics of Diffuse Astrocytic Tumor . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
1.2 Clinical Course . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
1.3 Obstacles in Treatment of Diffuse Astrocytic Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
2 RNAi for Glioma: from Bench to Clinic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
2.1 Preclinical Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
2.2 Target Genes for Silencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
2.3 Problems in Clinical Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3 miRNAs in Glioma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
3.1 Molecular Pathology of Aberrant miRNAs in Glioblastoma . . . . . . . . . . . . . . . . . . . . . . . 120
4 Conclusions and Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Abstract Gliomas are the most frequent malignant intracranial tumors arising from
the brain or spinal cord tissue. The most malignant among them is glioblastoma
multiforme (GBM), which constitutes approximately 20–25% of all primary intra-
cranial neoplasms with an incidence of 3–4/100,000. This type of tumor is character-
ized by progressive overgrowth of neoplastic glial cells with widespread and
relentless invasion, resulting in acquisition of resistance to treatment and a poor
prognosis due to recurrence. These features hamper efficient surgical intervention of
M. Nakada (*), D. Kita, Y. Hayashi, and J.-i. Hamada

Department of Neurosurgery, Graduate School of Medical Science, Kanazawa University, 13-1
Takara-machi, Kanazawa 920-8640, Japan
e-mail: nakada@ns.m.kanazawa-u.ac.jp
K. Kawakami
Division of Translational and Clinical Oncology, Cancer Research Institute, Kanazawa University,
13-1 Takara-machi, Kanazawa 920-0934, Japan
T. Minamoto (*)
Divisions of Translational and Clinical Oncology and Surgical Oncology, Cancer Research Institute,
Kanazawa University, 13-1 Takara-machi, Kanazawa 920-0934, Japan
e-mail: minamoto@staff.kanazawa-u.ac.jp
108 M. Nakada et al.
the disease. Current chemo/radiotherapy conditions act sublethally but cannot effec-
tively suppress the proliferation of glioma cells. Despite some progress, the thera-
peutic options are yet limited, and novel therapeutic strategies are clearly needed.
Targeting of disease-specific molecules involved in the proliferation, apoptosis, and
invasion of the tumor cells as well as in tumor angiogenesis may offer a high potential
for the development of more effective therapies for GBM. RNA interference (RNAi)
has been emerging not only for in vitro target validation, but also for a novel
therapeutic strategy based on the highly specific and efficient silencing of a target
gene. Indeed, RNAi has shown to act against GBM efficiently in numerous preclini-
cal studies. Many efforts have been devoted to overcome the three major obstacles
in use of RNAi in vivo; their specificity, instability, and poor cellular delivery
of bioactive small interfering RNA (siRNA) across the blood–brain barrier. The
identification of effective target and the establishment of novel siRNA delivery
systems are needed for the clinical applications of RNAi for treatment of GBM.
Keywords Glioma Glioblastoma RNAi microRNA
1 Introduction
1.1 Diagnostic Characteristics of Diffuse Astrocytic Tumor
Diffuse astrocytic tumors are the most frequently observed intracranial neoplasms;
they account for more than 60% of all primary brain tumors. A signature character-
istic of astrocytic tumor cells is their histological resemblance to astrocytes.
Therefore, the presence of cells showing fine fibrillary processes and expression
of glial fibrillary acidic protein (GFAP) is a major diagnostic feature for these
tumors. These findings in a certain proportion of tumor cells are important for
astrocytic tumor cell identification, even in the more undifferentiated and anaplastic
form of the tumor. Since the tumor cells typically exhibit diffuse infiltration of the
adjacent brain parenchyma, they are termed diffuse astrocytic tumors.
Although many classifications and grading systems for diffuse astrocytic tumors
have been introduced, the World Health Organization (WHO) classification is
adopted in this chapter because of its widespread recognition (Louis et al. 2007).
According to the WHO classification and grading system, diffuse astrocytic tumors
are divided into three categories: (1) diffuse astrocytoma (WHO grade II); (2) ana-
plastic astrocytoma (WHO grade III); and (3) the “glioblastoma” (GBM; WHO
grade IV), which is the most undifferentiated. GBM is the most malignant among
the astrocytic tumors and is composed of poorly differentiated neoplastic astro-
cytes. The presence of microvascular proliferation and/or necrosis is essential for
histopathological diagnosis of GBM (Fig. 1).
RNAi in Malignant Brain Tumors: Relevance to Molecular and Translational Research 109
Fig. 1 Representative histological finding of glioblastoma. The tumor consists of pleomorphic,

poorly differentiated tumor cells concomitant with multinucleated giant cells and anaplastic glial
cells (H & E, 200). Inset: prominent capillary endothelial proliferation is noted (H & E, 100)
1.2 Clinical Course
1.2.1 Symptoms
Clinical signs and symptoms induced by diffuse astrocytic tumors are dependent on
the site where the tumors locate within the central nervous system (Buckner et al.
2007; Chandana et al. 2008). Diffuse astrocytomas usually develop slowly; how-
ever, some of them may present suddenly with an onset of epilepsy. Frontal lobe
lesions are associated with intellectual deterioration. When the tumor locates more
posteriorly in the hemisphere (into the motor cortex or pyramidal tract), progressive
hemiparesis tends to occur. A more centrally located tumor involving the thalamus
frequently presents with raised intracranial pressure due to hydrocephalus caused
by occlusion of the cerebrospinal-fluid pathway. If located within the brain stem,
the tumor may induce cranial nerve palsies resulting from invasion of the tumor into
the cranial nerve nuclei.
After a varying time course after the diagnosis of diffuse astrocytoma, anaplastic
change occurs in the major proportion of anaplastic astrocytomas. Some cases of
anaplastic astrocytoma are diagnosed as tumors arising de novo after the onset of
symptoms due to increased intracranial pressure, such as headache (typically in the
morning), drowsiness, and vomiting.
As is the case for other grades of diffuse astrocytic tumors, the pattern of GBM
clinical presentation depends on the site of the tumor. In addition, GBM’s rapid
growth pattern is an important sign for preoperative diagnosis. Complications such
as hemorrhage and infarction of the tumor, which are caused by its rapid growth,
may result in sudden and rapid enlargement of the lesion and an increase in
intracranial pressure.
1.2.2 Prognosis
Anaplastic change occurring in diffuse astrocytoma is usually accompanied by

rapid deterioration of the patient’s condition and by the appearance of enhancing
lesions on computed tomography (CT) or magnetic resonance imaging (MRI). The
prognosis of patients with diffuse astrocytoma is determined by the neurological
state of the patient as scored preoperatively by Karnofsky Performance Status
(KPS), age of onset, and the location of the tumor (Kita et al. 2009; Ohgaki
2009). The location of the tumor influences radical surgical removal and is an
important factor with regard to tumor management. Many patients with diffuse
astrocytoma survive for several years. However, predicting the prognosis of
patients with diffuse astrocytoma is difficult, because anaplastic change can occur
after varying time courses.
The prognosis of patients with anaplastic astrocytoma and GBM is worse than
that of diffuse astrocytoma. Most clinical studies suggest that 50% of patients with
GBM survive no longer than 1 year and that almost no patients survive more than
5 years. The prognosis of anaplastic astrocytoma is only slightly better than that of
GBM.
Surgical removal, radiotherapy, and chemotherapy with the recently introduced
agent temozolomide have some beneficial effects on overall survival and may
improve the quality of life (QOL) for patients. Nevertheless, despite advances in
combination treatment with surgical resection and adjuvant chemo/radiotherapy,
overall survival time still does not exceed more than 2 years in most GBM patients.
1.3 Obstacles in Treatment of Diffuse Astrocytic Tumors
Due to the diffuse infiltrating nature of diffuse astrocytic tumors, complete removal
of the tumor is quite difficult (Nakada et al. 2007). In addition, there are many
regions, such as eloquent cortices (motor cortex, sensory cortex, language cortex,
etc.), thalamus, internal capsule, brain stem, and spinal cord, where tumor excision
damages critical neural structures and functions. Damages to these neural structures
are associated with irreversible neurological deficits and impairment of the patient’s
QOL. Due to the infiltrating nature of these tumors, the risk of recurrence or
regrowth from a residual tumor cannot be eliminated. Thus, one of the most
important strategies for adjuvant treatment is the control of residual tumor cells
to minimize invasion and regrowth. In terms of such a tumor control strategy,
conventional treatment with radiation and/or chemotherapies has so far failed to
suppress tumor cell activity completely.
Recent advances in molecular biology and genetic engineering have provided
new insights into tumor cell biology and into the distinct genetic alterations
underlying formation and progression of diffuse astrocytic tumors. Several genetic
alterations have been revealed in diffuse astrocytic tumors (Hayashi et al. 2004;
Ohgaki 2005; Ohgaki and Kleihues 2007; Watanabe et al. 2009; Yan et al. 2009).
The TP53, PDGFRA, and IDH1 alterations and/or their overexpression have been
observed in early stages of diffuse astrocytic tumor development. In addition to the
alteration of these genes, GBM frequently carries the loss of heterozygosity (LOH)
in chromosome 10 and exhibits EGFR, PTEN, and CDKN2A alterations. Moreover,
many studies have demonstrated several other genetic alterations occurring in
diffuse astrocytic tumors. Along with a deeper understanding of the genetic features
of tumor cells, newly developed treatments such as immunotherapy, gene therapy,
and molecular-targeted therapy may challenge the inherent difficulties of diffuse
astrocytic tumor treatment. However, until now, most of these challenges have
failed to achieve their clinical goal of controlling these tumors. In the future, new
treatment directions for diffuse astrocytic tumors will include unique and specific
therapies targeting molecules responsible for tumor cell proliferation, invasion, and
angiogenesis with the assistance of newly developed techniques, including RNA
interference (RNAi) and RNAi-related mechanisms.
2 RNAi for Glioma: from Bench to Clinic
RNAi describes a conserved biological response to double-stranded RNA (dsRNA),

resulting in the degradation of homologous messenger RNA (mRNA). In the last
decade, this process of sequence-specific, posttranscriptional gene silencing has
become a key technique for rapidly evaluating gene function in species ranging
from plants to mammals. Additionally, the discovery of RNAi has resparked
interest in developing nucleic acid drugs based on targeting mRNAs. This field of
research offers the potential to construct low-dose, nontoxic RNAi agents to treat
human diseases. Malignant tumors, such as GBM, are one of the major targets of
RNAi-based therapy, because oncogenes or other genes contributing to tumor
progression can be targeted by RNAi for silencing. Accumulating numbers of
reports have confirmed the efficacy of RNAi in GBM. Multiple cellular pathways
in GBM have been successfully targeted with RNAi by numerous research groups
in order to test for novel therapeutic strategies. Although most studies have been
performed in vitro, evolving strategies for in vivo application of RNAi will lead to
the efficient use of this targeted strategy for future therapeutic paradigms.
2.1 Preclinical Studies
RNAi represents a useful experimental approach for use in future clinical trials,
including RNAi-mediated targeting in vitro and in vivo for functional studies of
various genes whose expressions are shown to be upregulated in tumor tissues.
Preclinical studies, mostly in in vitro models, confirm that RNAi techniques can be
used to silence glioma-related target transcripts (Miyashita et al. 2009). In vivo
studies have also shown favorable outcomes using RNAi targeting of gene products
critical for glioma cell growth (Purow et al. 2005; Saydam et al. 2005; Grzelinski
et al. 2006; Pallini et al. 2006; Gillespie et al. 2007), invasion (Gondi et al. 2004a;
Gondi et al. 2004b; Lakka et al. 2004; Nakada et al. 2006; Gondi et al. 2007),
and angiogenesis (Zhen et al. 2007). These studies show the promising possibility of
developing novel therapeutic approaches against GBM based on RNAi gene targeting.
The RNAi phenomenon consists of a multistep intracellular process that can be
divided into two phases. In the first phase, endogenous or exogenous dsRNA
molecules that are present in the cell are processed through the cleavage activity
of RNase III-type (Dicer) into short 20–30-nucleotide fragments called small
interfering RNAs (siRNAs). In the second step, siRNAs as well as many proteins,
including nucleases and helicase, form the RNA-induced silencing complex
(RISC). Through unwinding of the double-stranded siRNA, this complex becomes
activated with single-stranded, noncoding siRNA, which guides the RISC to its
complementary target mRNA, resulting in its endonucleolytic cleavage (Fig. 2)
(Mathupala et al. 2006).
The applications of RNAi can be mediated by two types of molecules – the
chemically synthesized double-stranded siRNA and the vector-based short hairpin
RNA (shRNA). Effective RNAi was initially demonstrated by the application of
synthetic siRNA (Fire et al. 1998). Although siRNA and shRNA are used to achieve
similar functional outcomes, they are intrinsically different molecules. siRNAs
comprise 21–23-nucleotide dsRNA molecules. Once incorporated into the RISC,
they facilitate the cleavage and degradation of its recognized mRNA. shRNA is an
RNA molecule that contains a sense strand, an antisense strand, and a short-loop
sequence that intervenes between both strands. The complementarity of the sense
and antisense fragments in their sequence allows these RNA molecules to form
hairpin-shaped dsRNA structures. The cloning of shRNA into a vector allows for its
expression by an RNA polymerase III promoter. The expressed shRNA is then
exported into the cytoplasm, where it is processed by Dicer into siRNA and
incorporated into the RISC. The use of shRNA allows for more efficient gene
silencing using endogenous processing machinery as compared with siRNA. The
feasibility of siRNA manufacturing and the transient nature of the effect per dose
optimally meet the treatment requirement of certain diseases in which high vector
doses are required, e.g., viral injections.
dsRNA RNAi machinery
Dicer
1st step
Processing
cleavage
siRNA
2nd step
Unwinding
Target mRNA
RISC
Fig. 2 Molecular process of RNAi machinery. In the first step, long dsRNA molecules of
exogenous or endogenous origin are cleaved to produce siRNA by the enzyme, Dicer. In the
second step, siRNA molecules are incorporated into a RISC. The duplex RNA is unwound leaving
the antisense strand to guide RISC to complementary mRNA for subsequent endonucleolytic
cleavage. dsRNA: double-stranded RNA, siRNA: small interfering RNA, RISC: RNA-inducing
silencing complex
RNAi-based therapeutics have been shown to be well-tolerated in numerous

animal models, allowing further progress toward clinical application. At least one
RNAi-based drug targeting tenascin-C is currently being prescribed in early phase
clinical trials for GBM (Zukiel et al. 2006; Wyszko et al. 2008), and the result
seems to be promising. Both siRNA and shRNA are attractive for knocking down
specific genes but each have their respective advantages and disadvantages from a
mechanistic point of view. However, safety issues with this new therapeutic
paradigm are of utmost importance. Although no serious adverse events involving
initial RNAi-based clinical trials have been reported, there are concerns over the
potential “off-target effect” of RNAi-based agents, which are discussed below.
2.2 Target Genes for Silencing
The target molecules for RNAi therapeutics in glioma should meet the following
characteristics.
1. Target genes are overexpressed in glioma cells.
2. Target genes are not expressed in normal cells in the brain or behave as
bystanders in physiological cells even if the genes are expressed.
3. Target genes are associated with the malignant potential of glioma.
To date, there have been numerous publications on targeting glioma using

RNAi. The target molecules are classified into the following six categories.
1. Genes involved in cell-adhesion, motility, and invasiveness of glioma (Chuang
et al. 2004; Gondi et al. 2004b; Lakka et al. 2004; Salhia et al. 2005; Grzelinski
et al. 2006; Nakada et al. 2006; Tran et al. 2006; Bates et al. 2007; Gondi et al.
2007; Gondi et al. 2008; Hoelzinger et al. 2008; Salhia et al. 2008; Wesolowska
et al. 2008; Wyszko et al. 2008; Cai et al. 2009).
2. Genes involved in cell proliferation (Pallini et al. 2006; Gillespie et al. 2007;
Zhen et al. 2007; Jiang et al. 2008; Lin et al. 2008; Chu et al. 2009).
3. Genes involved in apoptotic pathways (Brown et al. 2005; Stegh et al. 2007;
Zhen et al. 2007).
4. Genes involved in angiogenesis (Zhen et al. 2007.)
5. Genes involved in epidermal growth factor receptor (EGFR)-related signal
transduction (Boado 2005; Saydam et al. 2005; Vollmann et al. 2006; Karpel-
Massler et al. 2009).
6. Genes involved in glioma chemoresistance (Zhao et al. 2008b; Wang et al.
2009).
Many groups have attempted to apply RNAi therapeutics against GBM that
inhibit EGFR (Boado 2005; Saydam et al. 2005; Vollmann et al. 2006; Karpel-
Massler et al. 2009), since EGFR signaling is believed to be critical for regulating
proliferation and differentiation of GBM. The EGFR gene is frequently amplified in
human GBMs but is undetectable or weakly expressed in normal cells in brain
(Bigner et al. 1990). In addition to amplification of the wild-type gene in human
glioma, there are mutant forms of EGFR that may be constitutively active, and the
EGFR variant III (vIII), the most common form of mutant, has an in-frame deletion
of exons 2–7 (Kuan et al. 2001). This mutant enables the tyrosine kinase domain of
the EGFR to be constitutively hyperactive, resulting in enhanced proliferation of
glioma cells (Kuan et al. 2001). On the basis of biochemical characteristics of the
EGFR and its oncogenic role, potential therapeutics for glioma should target both
wild-type EGFR and EGFRvIII.
2.3 Problems in Clinical Translation
Several issues must first be overcome in order to achieve successful clinical trials;
these issues are discussed herein. The specific knockdown of the target genes is
certainly important. Related undesirable consequences caused by off-target effects
due to lack of specificity and by “interferon response” should be prevented. The
stability of RNAi chemicals in brain is also critical for the continuous silencing of
target genes. The delivery method is the most crucial issue for glioma therapy,
because entry into glioma cells is essential if RNAi chemicals are to exert their
effects.
2.3.1 Specificity
RNAi is a natural cellular process through which expression of a targeted gene can
be knocked down with high specificity and selectivity. In order to knockdown target
genes, the single-stranded antisense oligonucleotides were first developed for anti-
mRNA strategies. RNAi has since been proven to target mRNA more efficiently
than antisense techniques. It works more specifically than the antisense oligonucle-
otide to decrease expression of a gene or to eliminate it entirely (Bertrand et al.
2002; Chi et al. 2003; Semizarov et al. 2003). Thus, siRNAs are effective at
concentrations that are several orders of magnitude below the concentrations
typically used in antisense experiments (Meister et al. 2004). Since RNAi is
considered an effective strategy to knockdown gene expression in mammalian
cells, several technical limitations relating to the specificity of the RNAi response
are highlighted below.
Off-Target Effect
Some reports have documented unintended effects of gene expression mediated by

RNAi, namely, the so-called off-target effect. Distinct off-target effects are
mediated by partial sequence complementarity of the RNAi construct to mRNAs
other than the primary target. Several microarray-based studies provide conflicting
results on the specificity of siRNA and its potential for off-target mRNA cleavage
(Chi et al. 2003; Jackson et al. 2003; Semizarov et al. 2003). However, it has been
reported that a single base-pair replacement in a siRNA duplex can block its
specific RNAi effect (Elbashir et al. 2001), and many other reports support that
siRNA activity is highly sequence specific (Rao et al. 2009). On the other hand,
other groups have demonstrated off-target effects occurring specifically through a
microRNA (miRNA)-like suppression of protein translation rather than by mRNA
degradation (Saxena et al. 2003; Scacheri et al. 2004). These issues have been
resolved through careful design of siRNA templates by initially comparing the
primary target (mRNA) sequence with all other messages in a given organism using
available public databases. In addition, algorithms for selection of optimal siRNA
templates and prescreened siRNA directed against a specific gene transcript are
currently available from various commercial sources. In most cases, prescreened or
in silico-selected siRNAs that are chemically modified to resist cellular nuclease
can be obtained from commercial resources for initial in vitro analysis of target-
specific gene knockdown. Optimized shRNA constructs allow for powerful and
stable effects using low copy numbers, resulting in fewer off-target effects as
compared with siRNA if embedded in an miRNA scaffold.
Interferon Response
Another obstacle against specific RNAi found during in vitro studies is interferon
response in the targeted cells. This undesirable effect was discovered in studies
showing that transfection of certain 21-nucleotide siRNAs (Sledz et al. 2003) or

lentiviral shRNA vectors (Bridge et al. 2003) resulted in upregulation of compo-
nents of the interferon pathway. This effect is thought to depend upon the siRNA
sequence, dosage, and delivery method. Interestingly, the in vivo studies using
either synthetic siRNA or vector-based shRNA delivery methods have found an
absence of active interferon response in targeted tissues or organs. Thus, when
siRNA techniques are employed in vivo, interferon pathway activation is most
likely suppressed by the physiological controls in mammals. Therefore, while the
presence of an interferon response is undesirable in studies in vitro, it may not
hamper validation of preclinical data nor will it affect the use of RNAi chemicals as
a future therapeutic tool.
2.3.2 Instability
siRNA is fairly unstable in vivo and its half-life in peripheral blood is only a few
minutes after intravenous administration. The short half-lives of siRNA in vivo,
even with chemical modification, prohibit long-term application of RNAi therapeu-
tics (Braasch et al. 2003). In order to obtain a continuous RNAi effect, technical
improvements are necessary to stabilize synthetic siRNA in vivo and to maintain
long-term in vivo expression of shRNA in the targeted cell. One of these improve-
ments is the development of an expression vector-based system, which is the most
effective method available at present for prolonged application of RNAi in vivo.
Many groups have examined the use of both plasmid- and virus-based delivery of
shRNA-generating gene therapy constructs in the mammalian brain and have found
varying shRNA delivery efficiencies. Obviously, once applied in situ, regulation of
the expression of the RNAi within the targeted glioma will be important. Moreover,
the influence of RNAi on the adjacent normal brain tissue should be minimized in
long-term survivors with glioma to prevent any side effects. Tumor cell-specific
control of RNAi is exemplified by a tetracycline-regulatable system of gene
expression (Gossen et al. 1995) that is routinely used in molecular biology research
and has been tested as a potential strategy for controlled expression of siRNA in the
lesion (Szulc et al. 2006).
2.3.3 Delivery
The delivery of si/shRNA to target cells in vivo is a limiting factor that determines
the therapeutic effects of RNAi (Zeng and Cullen 2002). The main obstacle to
delivery is the presence of the blood–brain barrier (BBB) in the cerebral vascula-
ture. The BBB is a physiological mechanism that controls the permeability of brain
capillaries so that some substances, such as certain drugs, are prevented from
entering the brain tissue while other substances are allowed to enter freely. This
barrier functions in the following two ways. First, the adjoining capillary endothe-
lial cells are sealed together at their edges by tight junctions that form a mechanical
Systemic route Local route
Vessels
BBB
RNAi
Viral
vehicle
Non- viral
Glioma cells
endocytosis
Fig. 3 Schematic illustration of the routs for delivery of RNAi chemicals to glioma cells in vivo
barrier. These barriers prevent water-soluble substances in the blood from extra-
vasating and entering the extra- and intracellular fluid in the brain. Such excluded
substances include siRNA (where the average molecular weight is around 14 kDa)
and large plasmid vector-based shRNA-generating constructs. Second, these capil-
laries are enclosed by the flattened “end-feet” of astrocytes, which also act as a
functional barrier. These types of barriers need to be bypassed by any therapeutic
strategy, including siRNA therapeutics. Numerous efforts have been made to
bypass the BBB for efficient delivery of si/shRNA to the brain (Pardridge 2004;
Mathupala et al. 2006; Pardridge 2007). These delivery systems are divided mainly
into systemic and local routes (Fig. 3).
Systemic Route
The best method for selective delivery of RNAi-based agents to glioma tissue is
intravenous injection of them by using a virus or nonvirus system as a vehicle. Viral
vectors are often used in a laboratory setting for delivery of shRNA to cells and
rodents because of their high transfection efficiency and effective integration of
exogenous DNA into the host genome (Saydam et al. 2005). However, concerns
over safety issues and recent reports about immunogenicity (Check 2002; Nguyen
et al. 2008) have slowed the clinical application of viral vectors. Nonviral poly-
meric delivery, in particular delivery using biodegradable vehicles, is much safer
than a viral delivery system although its transfection efficiency is generally lower
(Aigner 2007; Luten et al. 2008). There are four major classes of vehicles for
nonviral delivery systems: liposomes (Zhang et al. 2003; Pardridge 2004; Zhang
et al. 2004), biopolymers (Lesniak et al. 2005), nanoparticles (Lesniak 2005;
Moreira et al. 2008), and avidin–biotin complexes (Xia et al. 2007). Many of the
vehicles that have recently shown promise are actually a combination of these
classes.
Liposomes are commonly used vehicles in which an aqueous substance is
entirely enclosed by a membrane composed of phospholipid molecules. They can
entrap materials for delivery in their aqueous compartment (water-soluble materials
such as RNAi chemicals) as well as within the membrane (organic solvent-soluble
materials). The presence of hydrophilic polymers on the surface of the liposomes
gives rise to a steric barrier that inhibits the adsorption of blood components.
Studies using immune liposomes in vivo have demonstrated the feasibility of this
approach for glioma therapy (Zhang et al. 2003; Pardridge 2004; Zhang et al. 2004).
Lipid-based nanoparticles are also potential vehicles for the delivery of shRNA and
siRNA. These nanoparticles mimic low-density lipoproteins (LDLs) and thus
interact with the LDL receptors on endothelial cells, resulting in their uptake across
the BBB (Lesniak 2005).
One strategy that is applicable to delivery of vehicles containing RNAi chemi-
cals to the brain via a systemic route is the hydrodynamic approach, which enables
distribution of the RNAi chemicals in various organs by rapid high-dose injections
of these chemicals (McCaffrey et al. 2002; Campbell et al. 2008). Since the new
vessels formed via angiogenesis in GBM are usually immature and have a less
functional BBB, “naked” RNAi chemicals may pass through the pathological
vascular endothelial layer.
Local Route
The best way to allow water-soluble substances to cross the BBB is to bypass the
walls of cerebral capillaries. Local route application obviates any concerns about
the BBB and makes it possible to achieve very high concentrations of chemothera-
peutic agents at the tumor site and to avoid the side effects associated with systemic
high-dose chemotherapy. Direct delivery of RNAi chemicals to the residual tumor
bed and/or dead space formed by tumor resection may be an effective option. The
two approaches that have yielded promising results are polymerically controlled
release and convection-enhanced delivery (CED). Controlled-release polymers are
implanted directly at the resection site to allow for restricted slow release of
chemotherapeutic agents into the residual tumor bed (Sampath and Brem 1998).
Similar to the Gliadel wafers routinely used for localized delivery of the chemo-
therapeutic agent carmustine (BCNU), this strategy involves the entrapment of
siRNA in biodegradable polymers for effective delivery (Valtonen et al. 1997).
CED is an alternative method of controlled local drug release that has been
developed to deliver compounds throughout the brain. This method overcomes the
diffusion barrier seen with polymerically controlled release systems (Bobo et al.
1994) by using an applied external positive pressure infusion to generate fluid
convection in the brain, thus distributing chemotherapeutic drugs throughout the
parenchyma via interstitial spaces. Currently, the CED system is actively employed
in clinical trials (Stukel and Caplan 2009).
Even if RNAi chemicals are successfully distributed to the glioma, these che-
micals need to penetrate glioma cell membranes selectively in order to knockdown
the target genes. Efficiency of tumor targeting and cell entry are enhanced by
modification of the vehicle with targeting moieties, such as monoclonal antibodies,
peptides, small molecule ligands, and aptamers, which recognize tumor cell-specific
surface markers (Hughes and Rao 2005; Vorhies and Nemunaitis 2007). The
positive charge of nonviral delivery vehicles facilitates complex formation with
negatively charged nucleic acids and binding to the negatively charged glycocalyx
on external cell membranes, thereby promoting endocytosis. Within cells, the
vehicle’s positive charge facilitates immediate escape from the endosome (Godbey
et al. 2000; Thomas and Klibanov 2002). However, although the positive charge of
these vehicles improves their transfection efficiency, it is also associated with
increased toxicity (Zhang et al. 2007; Kim et al. 2009).
3 miRNAs in Glioma
miRNAs (also called miRs) are a recently discovered class of small (18–25
nucleotides in length), noncoding RNAs that modulate gene expression posttran-
scriptionally. Recently, alterations in miRNAs and their relevant targets have been
identified in various types of cancers, including GBM (Nicoloso and Calin 2008;
Lawler and Chiocca 2009). The global expression profile of GBM-related miRNAs
was first determined by Ciafre et al. (Ciafre et al. 2005) using a microarray
technique. They found upregulation of miR-221 and downregulation of miR-128,
miR-181a, miR-181b, and miR-181c. Since then, aberrant expression of miRNAs in
GBM has been reported, including upregulation of miR-21 (Chan et al. 2005), miR-26a
(Huse et al. 2009), miR-125b (Xia et al. 2009c), miR-221-222 cluster (le Sage et al.
2007), and miR-296 (Wurdinger et al. 2008), and downregulation of miR-7 (Kefas
et al. 2008), miR-124/137 (Silber et al. 2008), and miR-128 (Godlewski et al. 2008).
In tumors, downregulated miRNAs target oncogenes while upregulated miRNAs
target tumor-suppressor genes.
Each miRNA is identified and annotated according to the guidelines proposed by
Ambros et al. (2003). The Rfam database (http://rfam.janelia.org/) is provided as an
online clearinghouse for annotation of RNA families, including miRNAs. In the
miRNA database and registry miRBase (http://www.mirbase.org/), 10,883 entries
have been registered as of September, 2009.
3.1 Molecular Pathology of Aberrant miRNAs in Glioblastoma
To date (September, 2009), aberrant expressions of a number of miRNAs have been

identified in GBM, as shown in Table 1.
3.1.1 Upregulated miRNAs
miR-21
One of the most frequently and highly expressed miRNAs in various types of
malignancies, miR-21 plays important roles in tumor cell proliferation, invasion,
angiogenesis, and metastasis (Calin et al. 2005; Chan et al. 2005; Landgraf et al.
2007; Si et al. 2007; Zhu et al. 2008). Chan et al. (Chan et al. 2005) found marked
elevation of miR-21 levels in GBM tissues and cell lines as compared with non-
neoplastic human brain tissues and nonneoplastic cultured glial cells, respectively.
This miRNA acts as an antiapoptotic factor; pathway analyses of computer-
generated lists of miR-21 target genes have shown that miR-21 targets a network
of p53, TGF-b, and mitochondrial apoptosis factors in GBM (Papagiannakopoulos
et al. 2008). miR-21 promotes glioma invasion by targeting matrix metalloprotei-
nase (MMP) regulators and by suppressing RECK (a reversion-inducing cysteine-
rich protein with Kazal motifs, a tumor suppressor gene) and TIMP3 (a tissue
inhibitor of metalloproteinase 3) (Gabriely et al. 2008). Although their functions are
not sufficiently known in GBM, current studies have identified PTEN (phosphatase
and tensin homologue) (Meng et al. 2007), TPM1 (tropomyosin 1) (Zhu et al. 2007),
PDCD4 (programmed cell death 4) (Frankel et al. 2008), and Bcl-2 (Si et al. 2007)
as direct targets of miR-21. Thus, overexpression of miR-21 is pathological in
GBM, which provides a significant survival advantage to tumor cells.
miR-26a
Huse et al. (Huse et al. 2009) showed that miR-26a was overexpressed in a subset of
high-grade gliomas and that it directly targets the PTEN transcript. They also
demonstrated that overexpression of miR-26a in glioma primarily resulted from
amplification at the DNA level, a genomic event strongly associated with mono-
allelic PTEN loss. This association suggests that amplification of miR-26a may
contribute to silencing residual PTEN transcript in PTEN+/ tumors, which is
analogous to a loss-of-heterozygosity event (Huse et al. 2009).
miR-125b
miR-125b has been found to be a brain-enriched miRNA that is evenly distributed

between neurons and astrocytes (Smirnova et al. 2005). High expression of miR-125b
Table 1 microRNAs and hallmarks of GBM
MicroRNA Locus OMIM (Entrez) Target genes in Function in GBM cell References
GBM line
Upregulated
0
miR-21 17q23.2 611020 RECK, TIMP3, Invasion, (Gabriely et al. 2008), (Papagiannakopoulos
TGF-b, TP53 antiapoptosis et al. 2008), (Chan et al. 2005)
miR-26a1, 2 3p22.2, 12q14.1 612151 PTEN Invasion, (Huse et al. 2009)
antiapoptosis
miR-125b1, 2 11q24.2, 21q21.1 610104, 610105 BMF Proliferation, (Ciafre et al. 2005), (Xia et al. 2009c)
antiapoptosis
miR-221-222 Xp11.3 300568, 300569 p27 (Kip1), Anti-G1 arrest, (le Sage et al. 2007), (Medina et al. 2008),
cluster p57 (Kip2) proliferation (Ciafre et al. 2005)
miR-296 20q13.3 610945 HGS Angiogenesis (Wurdinger et al. 2008)
Downregulated
miR-7-1, 2, 3 9q21.32, 15q26.21, (407043, 407044, (Kefas et al. 2008), (Webster et al. 2009)
19q13.3 407045)
0 0
miR-15b 3q.25.33 (408949) (Xia et al. 2009a)
0
miR-124/137 8q23.1/1p21.3 611774/610567 (Silber et al. 2008)
0
miR-128 2q21 611774 (Ciafre et al. 2005), (Godlewski et al. 2008)
0
miR-146b 10q24.32 610567 (Xia et al. 2009b)
miR-181 Unknown Growth, (Ciafre et al. 2005), (Shi et al. 2008)
family antiapoptosis,
invasion
miR181a1, 2 1q31.3, 9q33.3 612742, 612743
miR181b1, 2 1q31.3, 9q33.3 612744, 612745
RNAi in Malignant Brain Tumors: Relevance to Molecular and Translational Research
miR181c 19q13.3 612746

121
has also been observed in oligodendroglial tumors (Nelson et al. 2006). The
miRNAs – miR-125b1 and miR-125b2 – are precursors of a mature miRNA
sequence designated as miR125b, but the sequences flanking the mature miRNA
are different. The expression of miR-125b1 was significantly upregulated in GBMs
as compared with normal brain tissues (Ciafre et al. 2005); however, Lee et al. (Lee
et al. 2005) revealed that miR-125b2 was the source of most of the miR-125b they
detected in human cell lines. In glioma U373 cells treated by all-trans-retinoic acid
(ATRA), an inverse correlation between the expression of miR-125b and the cell
apoptosis-related protein Bcl-2 modifying factor (Bmf) was shown, and both miR-
125b1 and miR-125b2 restored cell viability and inhibited cell apoptosis (Xia et al.
2009c).
miR-221-222 Cluster
miR-221 and miR-222, both of which are encoded by their respective genes on
chromosome Xp11.3, were found to be upregulated in GBM (Ciafre et al. 2005) as
well as in papillary thyroid carcinomas (He et al. 2005), hepatocellular carcinomas
(Gramantieri et al. 2007), breast cancers (Miller et al. 2008; Zhao et al. 2008a), and
prostate cancers (Galardi et al. 2007). The miR-221-222 cluster downregulates
p27/Kip1 (Zhang et al. 2009) and p57/Kip2 – well-known cyclin-dependent kinase
inhibitors (Medina et al. 2008). In a large proportion of GBM, elevated levels of
miR-221 and miR-222 are correlated with low levels of its target p27/Kip1 tran-
script (le Sage et al. 2007).
miR-296
miR-296 contributes to angiogenesis by directly targeting the hepatocyte growth

factor-regulated tyrosine kinase substrate (HGS) mRNA, leading to decreased levels
of HGS and thereby reducing HGS-mediated degradation of the growth factor
receptors VEGFR-2 (vascular endothelial growth factor receptor-2) and PDGFR-b
(platelet-derived growth factor receptor-b) (Wurdinger et al. 2008). The expression
level of miR-296 is elevated in primary tumor endothelial cells isolated from GBMs
compared with normal brain endothelial cells (Wurdinger et al. 2008).
3.1.2 Downregulated miRNAs
miR-7
Expression of miR-7 was downregulated in GBM as compared with surrounding

brain tissue (Kefas et al. 2008). miR-7 directly binds to two of the three binding
sites in the 30 -UTR of EGFR mRNA and reduces glioma cell proliferation and
invasion (Webster et al. 2009). miR-7 also regulates Raf1 expression via specific
binding to its mRNA 30 -UTR, resulting in modulation of the EGFR–Raf–MEK
(mitogen-activated protein kinase)–ERK (extracellular signal-regulated kinase)

cascade, which is a major signaling pathway for glioma malignancy (Webster
et al. 2009).
miR-15
Deregulation of miR-15 in GBM tissue samples obtained from Chinese patients was
investigated by Xia et al. (Xia et al. 2009a). They observed that overexpression of
miR-15b in glioma cell lines U87 and U118 resulted in cell cycle arrest at the
G0/G1 phase, while suppression of miR-15b expression resulted in a decrease of
cell populations in the G0/G1 phase and a proportional increase of cell populations
in the S phase. They also showed that CCNE1 (encoding cyclin E1) is one of the
downstream targets of miR-15b. These findings indicate that miR-15b regulates
cell cycle progression in glioma cells by targeting cell cycle-related molecules
(Xia et al. 2009a).
miR-124 and miR-137
miR-124 and miR-137 are known to belong to the family of upregulated miRNAs in
developing neuronal cells (Sempere et al. 2004; Visvanathan et al. 2007). They are
markedly downregulated in high-grade glioma and are associated with upregulation
of CDK6 (cyclin-dependent kinase 6) (Silber et al. 2008). miR-124 and miR-137
collectively inhibit proliferation of GBM cells and induce differentiation of brain
tumor stem cells by inhibiting G1 cell cycle arrest (Silber et al. 2008).
miR-128
miR-128 causes a striking decrease in expression of the Bmi-1 oncogene by direct

regulation of the Bmi-1 mRNA 30 -UTR via a single miR-128 binding site, resulting
in inhibition of glioma stem-like cell proliferation and self-renewal. In a panel of
patient GBM specimens, Bmi-1 expression was significantly upregulated and miR-128
was downregulated as compared with the normal brain (Godlewski et al. 2008).
miR-146b
A recent study using an miRNA microarray revealed that expression of miR-146b is

decreased in human GBM tissue and that transfection of the precursor miR-146b
into the U373 GBM cell line reduced tumor cell migration and invasion but had no
effect on their growth (Xia et al. 2009b). The study also identified that MMP16,
known as MT (membrane-type) 3-MMP and associated with GBM cell migration
and invasion, is one of the downstream targets of miR-146b.
miR-181 Family
Using microarray and northern blot analyses, Ciafre et al. found that expression of
miR-181a, b, and c were significantly downregulated in primary GBMs and in
human GBM cell lines as compared with normal brain tissue (Ciafre et al. 2005).
Shi et al. showed that miR-181a and miR-181b function as tumor suppressors that
trigger growth inhibition, induce apoptosis, and inhibit invasion of glioma cells (Shi
et al. 2008). They also revealed that the tumor-suppressive effect of miR-181b was
stronger than miR-181a. While their target genes in GBM are not yet sufficiently
understood, the miR-181 family has been shown to be one of the regulators of the
Tcl-1 oncogene in B-cell chronic lymphocytic leukemia (Pekarsky et al. 2006).
Distinct genes, miR-181a1 and miR-181a2, and miR-181b1 and miR-181b2 encode
mature miR-181a and miR-181b miRNAs, respectively (Table 1).
4 Conclusions and Perspective
Our understanding and application of RNAi has dramatically advanced since the
RNAi phenomenon was discovered in 1998. The promise of RNAi as a future
clinical strategy in targeting glioma has been sufficiently demonstrated at bench side.
As with any novel therapeutic tool, it is evident that several issues, such as target
selection, effector potency, off-target effects, and delivery vehicle design, must be
addressed and resolved before future application of these RNA technologies for
treatment of refractory diseases, as represented by GBM, can be implemented.
References
Aigner A (2007) Nonviral in vivo delivery of therapeutic small interfering RNAs. Curr Opin Mol
Ther 9:345–352
Ambros V, Bartel B, Bartel DP et al (2003) A uniform system for microRNA annotation. RNA
9:277–279
Bates DC, Sin WC, Aftab Q et al (2007) Connexin43 enhances glioma invasion by a mechanism
involving the carboxy terminus. Glia 55:1554–1564
Bertrand JR, Pottier M, Vekris A et al (2002) Comparison of antisense oligonucleotides and
siRNAs in cell culture and in vivo. Biochem Biophys Res Commun 296:1000–1004
Bigner SH, Humphrey PA, Wong AJ et al (1990) Characterization of the epidermal growth factor
receptor in human glioma cell lines and xenografts. Cancer Res 50:8017–8022
Boado RJ (2005) RNA interference and nonviral targeted gene therapy of experimental brain
cancer. NeuroRx 2:139–150
Bobo RH, Laske DW, Akbasak A et al (1994) Convection-enhanced delivery of macromolecules
in the brain. Proc Natl Acad Sci USA 91:2076–2080
Braasch DA, Jensen S, Liu Y et al (2003) RNA interference in mammalian cells by chemically-
modified RNA. Biochemistry 42:7967–7975
Brown R, Morash B, Ur E et al (2005) RNAi-mediated silencing of leptin gene expression

increases cell death in C6 glioblastoma cells. Brain Res Mol Brain Res 139:357–360
Buckner JC, Brown PD, O’Neill BP et al (2007) Central nervous system tumors. Mayo Clin Proc
82:1271–1286
Cai M, Han L, Chen R et al (2009) Inhibition of LRIG3 gene expression via RNA interference
modulates the proliferation, cell cycle, cell apoptosis, adhesion and invasion of glioblastoma
cell (GL15). Cancer Lett 278:104–112
Calin GA, Ferracin M, Cimmino A et al (2005) A MicroRNA signature associated with prognosis
and progression in chronic lymphocytic leukemia. N Engl J Med 353:1793–1801
Campbell M, Kiang AS, Kenna PF et al (2008) RNAi-mediated reversible opening of the blood–brain
barrier. J Gene Med 10:930–947
Chan JA, Krichevsky AM, Kosik KS (2005) MicroRNA-21 is an antiapoptotic factor in human
glioblastoma cells. Cancer Res 65:6029–6033
Chandana SR, Movva S, Arora M et al (2008) Primary brain tumors in adults. Am Fam Physician
77:1423–1430
Check E (2002) A tragic setback. Nature 420:16–118
Chi JT, Chang HY, Wang NN et al (2003) Genomewide view of gene silencing by small
interfering RNAs. Proc Natl Acad Sci USA 100:6343–6346
Chu SH, Feng DF, Zhang H et al (2009) c-Met-targeted RNA interference inhibits growth and
metastasis of glioma U251 cells in vitro. J Neurooncol 93:183–189
Chuang YY, Tran NL, Rusk N et al (2004) Role of synaptojanin 2 in glioma cell migration and
invasion. Cancer Res 64:8271–8275
Ciafre SA, Galardi S, Mangiola A et al (2005) Extensive modulation of a set of microRNAs in
primary glioblastoma. Biochem Biophys Res Commun 334:1351–1358
Elbashir SM, Martinez J, Patkaniowska A et al (2001) Functional anatomy of siRNAs for
Frankel LB, Christoffersen NR, Jacobsen A et al (2008) Programmed cell death 4 (PDCD4) is an
important functional target of the microRNA miR-21 in breast cancer cells. J Biol Chem
283:1026–1033
Gabriely G, Wurdinger T, Kesari S et al (2008) MicroRNA 21 promotes glioma invasion by
targeting matrix metalloproteinase regulators. Mol Cell Biol 28:5369–5380
Galardi S, Mercatelli N, Giorda E et al (2007) miR-221 and miR-222 expression affects the
proliferation potential of human prostate carcinoma cell lines by targeting p27Kip1. J Biol
Chem 282:23716–23724
Gillespie DL, Whang K, Ragel BT et al (2007) Silencing of hypoxia inducible factor-1a by RNA
interference attenuates human glioma cell growth in vivo. Clin Cancer Res 13:2441–2448
Godbey WT, Barry MA, Saggau P et al (2000) Poly(ethylenimine)-mediated transfection: a new
paradigm for gene delivery. J Biomed Mater Res 51:321–328
Godlewski J, Nowicki MO, Bronisz A et al (2008) Targeting of the Bmi-1 oncogene/stem cell
renewal factor by microRNA-128 inhibits glioma proliferation and self-renewal. Cancer Res
68:9125–9130
Gondi CS, Dinh DH, Gujrati M et al (2008) Simultaneous downregulation of uPAR and MMP-9
induces overexpression of the FADD-associated protein RIP and activates caspase 9-mediated
apoptosis in gliomas. Int J Oncol 33:783–790
Gondi CS, Lakka SS, Dinh DH et al (2004a) Downregulation of uPA, uPAR and MMP-9 using
small, interfering, hairpin RNA (siRNA) inhibits glioma cell invasion, angiogenesis and tumor
growth. Neuron Glia Biol 1:165–176
Gondi CS, Lakka SS, Dinh DH et al (2004b) RNAi-mediated inhibition of cathepsin B and uPAR
leads to decreased cell invasion, angiogenesis and tumor growth in gliomas. Oncogene
23:8486–8496
Gondi CS, Lakka SS, Dinh DH et al (2007) Intraperitoneal injection of a hairpin RNA-expressing
plasmid targeting urokinase-type plasminogen activator (uPA) receptor and uPA retards
angiogenesis and inhibits intracranial tumor growth in nude mice. Clin Cancer Res
13:4051–4060
Gossen M, Freundlieb S, Bender G et al (1995) Transcriptional activation by tetracyclines in
mammalian cells. Science 268:1766–1769
Gramantieri L, Ferracin M, Fornari F et al (2007) Cyclin G1 is a target of miR-122a, a microRNA
frequently down-regulated in human hepatocellular carcinoma. Cancer Res 67:6092–6099
Grzelinski M, Urban-Klein B, Martens T et al (2006) RNA interference-mediated gene silencing
of pleiotrophin through polyethylenimine-complexed small interfering RNAs in vivo exerts
antitumoral effects in glioblastoma xenografts. Hum Gene Ther 17:751–766
Hayashi Y, Yamashita J, Watanabe T (2004) Molecular genetic analysis of deep-seated glioblastomas.
Cancer Genet Cytogenet 153:64–68
He H, Jazdzewski K, Li W et al (2005) The role of microRNA genes in papillary thyroid
carcinoma. Proc Natl Acad Sci U S A 102:19075–19080
Hoelzinger DB, Nakada M, Demuth T et al (2008) Autotaxin: a secreted autocrine/paracrine factor
that promotes glioma invasion. J Neurooncol 86:297–309
Hughes JA, Rao GA (2005) Targeted polymers for gene delivery. Expert Opin Drug Deliv
2:145–157
Huse JT, Brennan C, Hambardzumyan D et al (2009) The PTEN-regulating microRNA miR-26a is
amplified in high-grade glioma and facilitates gliomagenesis in vivo. Genes Dev 23:1327–1337
Jiang W, Xiang C, Cazacu S et al (2008) Insulin-like growth factor binding protein 7 mediates
glioma cell growth and migration. Neoplasia 10:1335–1342
Karpel-Massler G, Schmidt U, Unterberg A et al (2009) Therapeutic inhibition of the epidermal
growth factor receptor in high-grade gliomas: where do we stand? Mol Cancer Res
7:1000–1012
Kefas B, Godlewski J, Comeau L et al (2008) MicroRNA-7 inhibits the epidermal growth factor
receptor and the Akt pathway and is down-regulated in glioblastoma. Cancer Res
68:3566–3572
Kim Y, Tewari M, Pajerowski JD et al (2009) Polymersome delivery of siRNA and antisense
oligonucleotides. J Control Release 134:132–140
Kita D, Ciernik IF, Vaccarella S et al (2009) Age as a predictive factor in glioblastomas:
population-based study. Neuroepidemiology 33:17–22
Kuan CT, Wikstrand CJ, Bigner DD (2001) EGF mutant receptor vIII as a molecular target in
cancer therapy. Endocr Relat Cancer 8:83–96
Lakka SS, Gondi CS, Yanamandra N et al (2004) Inhibition of cathepsin B and MMP-9 gene
expression in glioblastoma cell line via RNA interference reduces tumor cell invasion, tumor
growth and angiogenesis. Oncogene 23:4681–4689
Landgraf P, Rusu M, Sheridan R et al (2007) A mammalian microRNA expression atlas based on
small RNA library sequencing. Cell 129:1401–1414
Lawler S, Chiocca EA (2009) Emerging functions of microRNAs in glioblastoma. J Neurooncol
92:297–306
le Sage C, Nagel R, Egan DA et al (2007) Regulation of the p27(Kip1) tumor suppressor by
miR-221 and miR-222 promotes cancer cell proliferation. EMBO J 26:3699–3708
Lee YS, Kim HK, Chung S et al (2005) Depletion of human micro-RNA miR-125b reveals that it
is critical for the proliferation of differentiated cells but not for the down-regulation of putative
targets during differentiation. J Biol Chem 280:16635–16641
Lesniak MS (2005) Novel advances in drug delivery to brain cancer. Technol Cancer Res Treat
4:417–428
Lesniak MS, Upadhyay U, Goodwin R et al (2005) Local delivery of doxorubicin for the treatment
of malignant brain tumors in rats. Anticancer Res 25:3825–3831
Lin W, Zhang J, Zhang J et al (2008) RNAi-mediated inhibition of MSP58 decreases tumor growth,
migration and invasion in a human glioma cell line. J Cell Mol Med 13(11–12):4608–4622
Louis DN, Ohgaki H, Wiestler OD et al (2007) The 2007 WHO classification of tumours of the
central nervous system. Acta Neuropathol 114:97–109
Luten J, van Nostrum CF, De Smedt SC et al (2008) Biodegradable polymers as non-viral carriers
for plasmid DNA delivery. J Control Release 126:97–110
Mathupala SP, Guthikonda M, Sloan AE (2006) RNAi based approaches to the treatment of
malignant glioma. Technol Cancer Res Treat 5:261–269
McCaffrey AP, Meuse L, Pham TT et al (2002) RNA interference in adult mice. Nature 418:38–39
Medina R, Zaidi SK, Liu CG et al (2008) MicroRNAs 221 and 222 bypass quiescence and
compromise cell survival. Cancer Res 68:2773–2780
Meister G, Landthaler M, Dorsett Y et al (2004) Sequence-specific inhibition of microRNA- and
siRNA-induced RNA silencing. RNA 10:544–550
Meng F, Henson R, Wehbe-Janek H et al (2007) MicroRNA-21 regulates expression of the PTEN
tumor suppressor gene in human hepatocellular cancer. Gastroenterology 133:647–658
Miller TE, Ghoshal K, Ramaswamy B et al (2008) MicroRNA-221/222 confers tamoxifen
resistance in breast cancer by targeting p27Kip1. J Biol Chem 283:29897–29903
Miyashita K, Kawakami K, Nakada M et al (2009) Potential therapeutic effect of glycogen
synthase kinase 3b inhibition against human glioblastoma. Clin Cancer Res 15:887–897
Moreira JN, Santos A, Moura V et al (2008) Non-viral lipid-based nanoparticles for targeted
cancer systemic gene silencing. J Nanosci Nanotechnol 8:2187–2204
Nakada M, Drake KL, Nakada S et al (2006) Ephrin-B3 ligand promotes glioma invasion through
activation of Rac1. Cancer Res 66:8492–8500
Nakada M, Nakada S, Demuth T et al (2007) Molecular targets of glioma invasion. Cell Mol Life
Sci 64:458–478
Nelson PT, Baldwin DA, Kloosterman WP et al (2006) RAKE and LNA-ISH reveal microRNA
expression and localization in archival human brain. RNA 12:187–191
Nguyen T, Menocal EM, Harborth J et al (2008) RNAi therapeutics: an update on delivery. Curr
Opin Mol Ther 10:158–167
Nicoloso MS, Calin GA (2008) MicroRNA involvement in brain tumors: from bench to bedside.
Brain Pathol 18:122–129
Ohgaki H (2005) Genetic pathways to glioblastomas. Neuropathology 25:1–7
Ohgaki H (2009) Epidemiology of brain tumors. Methods Mol Biol 472:323–342
Ohgaki H, Kleihues P (2007) Genetic pathways to primary and secondary glioblastoma. Am J
Pathol 170:1445–1453
Pallini R, Sorrentino A, Pierconti F et al (2006) Telomerase inhibition by stable RNA interference
impairs tumor growth and angiogenesis in glioblastoma xenografts. Int J Cancer
118:2158–2167
Papagiannakopoulos T, Shapiro A, Kosik KS (2008) MicroRNA-21 targets a network of key
tumor-suppressive pathways in glioblastoma cells. Cancer Res 68:8164–8172
Pardridge WM (2004) Intravenous, non-viral RNAi gene therapy of brain cancer. Expert Opin Biol
Ther 4:1103–1113
Pardridge WM (2007) shRNA and siRNA delivery to the brain. Adv Drug Deliv Rev 59:141–152
Pekarsky Y, Santanam U, Cimmino A et al (2006) Tcl1 expression in chronic lymphocytic
leukemia is regulated by miR-29 and miR-181. Cancer Res 66:11590–11593
Purow BW, Haque RM, Noel MW et al (2005) Expression of notch-1 and its ligands, delta-like-1
and jagged-1, is critical for glioma cell survival and proliferation. Cancer Res 65:2353–2363
Rao DD, Vorhies JS, Senzer N et al (2009) siRNA vs. shRNA: similarities and differences. Adv
Drug Deliv Rev 61:746–759
Salhia B, Rutten F, Nakada M et al (2005) Inhibition of Rho-kinase affects astrocytoma morphol-
ogy, motility, and invasion through activation of Rac1. Cancer Res 65:8792–8800
Salhia B, Tran NL, Chan A et al (2008) The guanine nucleotide exchange factors trio, Ect2, and
Vav3 mediate the invasive behavior of glioblastoma. Am J Pathol 173:1828–1838
Sampath P, Brem H (1998) Implantable slow-release chemotherapeutic polymers for the treatment
of malignant brain tumors. Cancer Control 5:130–137
Saxena S, Jonsson ZO, Dutta A (2003) Small RNAs with imperfect match to endogenous mRNA
repress translation. Implications for off-target activity of small inhibitory RNA in mammalian
cells. J Biol Chem 278:44312–44319
Saydam O, Glauser DL, Heid I et al (2005) Herpes simplex virus 1 amplicon vector-mediated
siRNA targeting epidermal growth factor receptor inhibits growth of human glioma cells
in vivo. Mol Ther 12:803–812
Scacheri PC, Rozenblatt-Rosen O, Caplen NJ et al (2004) Short interfering RNAs can induce
unexpected and divergent changes in the levels of untargeted proteins in mammalian cells.
Semizarov D, Frost L, Sarthy A et al (2003) Specificity of short interfering RNA determined
through gene expression signatures. Proc Natl Acad Sci USA 100:6347–6352
Sempere LF, Freemantle S, Pitha-Rowe I et al (2004) Expression profiling of mammalian micro-
RNAs uncovers a subset of brain-expressed microRNAs with possible roles in murine and
human neuronal differentiation. Genome Biol 5:R13
Shi L, Cheng Z, Zhang J et al (2008) hsa-mir-181a and hsa-mir-181b function as tumor suppres-
sors in human glioma cells. Brain Res 1236:185–193
Si ML, Zhu S, Wu H et al (2007) miR-21-mediated tumor growth. Oncogene 26:2799–2803
Silber J, Lim DA, Petritsch C et al (2008) miR-124 and miR-137 inhibit proliferation of glioblas-
toma multiforme cells and induce differentiation of brain tumor stem cells. BMC Med 6:14
Smirnova L, Grafe A, Seiler A et al (2005) Regulation of miRNA expression during neural cell
specification. Eur J Neurosci 21:1469–1477
Stegh AH, Kim H, Bachoo RM et al (2007) Bcl2L12 inhibits post-mitochondrial apoptosis
signaling in glioblastoma. Genes Dev 21:98–111
Stukel JM, Caplan MR (2009) Targeted drug delivery for treatment and imaging of glioblastoma
multiforme. Expert Opin Drug Deliv 6:705–718
Szulc J, Wiznerowicz M, Sauvain MO et al (2006) A versatile tool for conditional gene expression
and knockdown. Nat Methods 3:109–116
Thomas M, Klibanov AM (2002) Enhancing polyethylenimine’s delivery of plasmid DNA into
mammalian cells. Proc Natl Acad Sci USA 99:14640–14645
Tran NL, McDonough WS, Savitch BA et al (2006) Increased fibroblast growth factor-inducible
14 expression levels promote glioma cell invasion via Rac1 and nuclear factor-kappaB and
correlate with poor patient outcome. Cancer Res 66:9535–9542
Valtonen S, Timonen U, Toivanen P et al (1997) Interstitial chemotherapy with carmustine-loaded
polymers for high-grade gliomas: a randomized double-blind study. Neurosurgery 41:44–48,
discussion 48–49
Visvanathan J, Lee S, Lee B et al (2007) The microRNA miR-124 antagonizes the anti-neural
REST/SCP1 pathway during embryonic CNS development. Genes Dev 21:744–749
Vollmann A, Vornlocher HP, Stempfl T et al (2006) Effective silencing of EGFR with RNAi
demonstrates non-EGFR dependent proliferation of glioma cells. Int J Oncol 28:1531–1542
Vorhies JS, Nemunaitis J (2007) Nonviral delivery vehicles for use in short hairpin RNA-based
cancer therapies. Expert Rev Anticancer Ther 7:373–382
Wang HB, Zhang SY, Wang S et al (2009) REV3L confers chemoresistance to cisplatin in human
gliomas: the potential of its RNAi for synergistic therapy. Neuro Oncol 11(6):790–802
Watanabe T, Nobusawa S, Kleihues P et al (2009) IDH1 mutations are early events in the
development of astrocytomas and oligodendrogliomas. Am J Pathol 174:1149–1153
Webster RJ, Giles KM, Price KJ et al (2009) Regulation of epidermal growth factor receptor
signaling in human cancer cells by microRNA-7. J Biol Chem 284:5731–5741
Wesolowska A, Kwiatkowska A, Slomnicki L et al (2008) Microglia-derived TGF-beta as an

important regulator of glioblastoma invasion – an inhibition of TGF-beta-dependent effects by
shRNA against human TGF-beta type II receptor. Oncogene 27:918–930
Wurdinger T, Tannous BA, Saydam O et al (2008) miR-296 regulates growth factor receptor
overexpression in angiogenic endothelial cells. Cancer Cell 14:382–393
Wyszko E, Rolle K, Nowak S et al (2008) A multivariate analysis of patients with brain tumors
treated with ATN–RNA. Acta Pol Pharm 65:677–684
Xia CF, Zhang Y, Zhang Y et al (2007) Intravenous siRNA of brain cancer with receptor targeting
and avidin–biotin technology. Pharm Res 24:2309–2316
Xia H, Qi Y, Ng SS et al (2009a) MicroRNA-15b regulates cell cycle progression by targeting
cyclins in glioma cells. Biochem Biophys Res Commun 380:205–210
Xia H, Qi Y, Ng SS et al (2009b) MicroRNA-146b inhibits glioma cell migration and invasion by
targeting MMPs. Brain Res 1269:158–165
Xia HF, He TZ, Liu CM et al (2009c) MiR-125b expression affects the proliferation and apoptosis
of human glioma cells by targeting Bmf. Cell Physiol Biochem 23:347–358
Yan H, Parsons DW, Jin G et al (2009) IDH1 and IDH2 mutations in gliomas. N Engl J Med
360:765–773
Zeng Y, Cullen BR (2002) RNA interference in human cells is restricted to the cytoplasm. RNA
8:855–860
Zhang C, Kang C, You Y et al (2009) Co-suppression of miR-221/222 cluster suppresses human
glioma cell growth by targeting p27kip1 in vitro and in vivo. Int J Oncol 34:1653–1660
Zhang S, Zhao B, Jiang H et al (2007) Cationic lipids and polymers mediated vectors for delivery
of siRNA. J Control Release 123:1–10
Zhang Y, Boado RJ, Pardridge WM (2003) in vivo knockdown of gene expression in brain cancer
with intravenous RNAi in adult rats. J Gene Med 5:1039–1045
Zhang Y, Zhang YF, Bryant J et al (2004) Intravenous RNA interference gene therapy targeting
the human epidermal growth factor receptor prolongs survival in intracranial brain cancer. Clin
Cancer Res 10:3667–3677
Zhao JJ, Lin J, Yang H et al (2008a) MicroRNA-221/222 negatively regulates estrogen receptor
alpha and is associated with tamoxifen resistance in breast cancer. J Biol Chem
283:31079–31086
Zhao P, Zhang Y, Sun M et al (2008b) Reversion of multidrug resistance in human glioma by RNA
interference. Neurol Res 30:562–566
Zhen HN, Li LW, Zhang W et al (2007) Short hairpin RNA targeting survivin inhibits growth and
angiogenesis of glioma U251 cells. Int J Oncol 31:1111–1117
Zhu S, Si ML, Wu H et al (2007) MicroRNA-21 targets the tumor suppressor gene tropomyosin 1
(TPM1). J Biol Chem 282:14328–14336
Zhu S, Wu H, Wu F et al (2008) MicroRNA-21 targets tumor suppressor genes in invasion and
metastasis. Cell Res 18:350–359
Zukiel R, Nowak S, Wyszko E et al (2006) Suppression of human brain tumor with interference
RNA specific for tenascin-C. Cancer Biol Ther 5:1002–1007
Silencing Huntington’s Disease Gene with RNAi
Yu Zhang and Robert M. Friedlander
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
2 A New Approach to Understanding Normal Function of Wild-Type Huntingtin . . . . . . . . . 133
3 Nonallele-Specific Silencing of Huntingtin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
4 Allele-Specific Silencing of Mutant Huntingtin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
5 Current Challenges and Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Abstract Huntington’s disease (HD), a hereditary condition afflicting 30,000

Americans, cannot be treated by existing therapies and it is universally fatal. It is
characterized by movement disorder (Huntington’s chorea), emotional distress, and
dementia. HD is caused by a highly penetrant, autosomal-dominant mutation in the
HD gene at chromosomal locus 4p16.3. Expansion of the CAG repeat at the 50 -end
of this gene increases the number of tandem glutamine residues in the encoded
protein (huntingtin) from under 30 to 36–100 (or more). Most HD patients are
heterozygotes, carrying one allele for the polyQ-expanded mutant huntingtin and
one for the wild-type protein. The former protein is harmful, particularly to striatal
neurons, whereas the latter is essential to neuronal survival. Experiments using
RNAi are expanding our understanding of the functions of wild-type huntingtin.
Using this technique, RNAi-based therapies for HD are being developed.
Biotechnologies using both allele-specific and allele-nonspecific RNAi have
proven effective at countering disease progression in multiple transgenic animal
models for HD. RNAi against the transgene (mutant human HD transcript)
Y. Zhang and R.M. Friedlander (*)

Neuroapoptosis Laboratory, Department of Neurosurgery, Brigham and Women’s Hospital,
Harvard Medical School, Boston, MA 02115, USA
e-mail: rfriedlander@rics.bwh.harvard.edu, friedlanderr@upmc.edu
University of Pittsburgh School of Medicine, Department of Neurological Surgery, UPMC
Presbyterian Hospital, SuiteB449, USA
132 Y. Zhang and R.M. Friedlander
decreases expression of the pathogenic protein and slows neurodegeneration. RNAi

can be directed at a polymorphism linked to the polyQ-expansion mutation in HD.
Consequently, the mutant allele is silenced while the wild-type one remains
expressed. Such allele-specific silencing has been achieved in fibroblasts from
HD patients. Technical improvements in local and systemic delivery combined
with chemical modifications to the RNAi should improve efficiency and specificity
thereby making RNAi-based therapy a successful treatment for HD.
Keywords Huntington’s disease (HD) RNAi Wild-type huntingtin Mutant

huntingtin siRNA shRNA Allele-specific silencing Allelenon-specific
silencing Nucleotide polymorphism
1 Introduction
Huntington’s disease (HD) is an inherited progressive neurodegenerative disease

caused by a highly penetrant autosomal-dominant mutation in the HD gene located
at position 4p16.3 (The Huntington’s Disease Collaborative Research Group 1993).
The syndrome is named for George Huntington, an American physician who first
thoroughly described the illness in 1872 and called it “an heirloom from generations
way back in the dim past.” The affliction was known long before its scientific
characterization, being described as “chorea”, the Greek word for dance. Indeed,
HD is characterized by movement disorder (Huntington’s chorea) as well as emotional
distress and dementia. Most people with the disease first show symptoms between the
ages of 30 and 50. Symptoms may occur earlier, however, even before age 20 among
persons with “juvenile HD” (a.k.a. the Westphal variant or a kinetic-rigid HD). The
latter syndrome, a form of HD accounting for one sixth of all cases, progresses quite
rapidly and leads to muscle rigidity rather than uncontrolled movement. The selec-
tively sensitive neural population consists of medium spiny GABAergic neurons
(MSNs) of the striatum, particularly those in the caudate nucleus and globus pallidus.
At later stages of the disease, cortical neurons are also affected. About 30,000
Americans suffer from the disease, and approximately 200,000 others are at risk of
inheriting it from an affected parent. HD progresses without remission for 10–25 years
before causing death. Moreover, there is currently no treatment.
HD is caused by expansion of the CAG repeat at the 50 -end of the “HD gene” that
encodes a protein (“huntingtin”) with an unusually long stretch of tandem gluta-
mine residues near its N-terminal. Whereas healthy individuals have fewer than 35
such glutamines, HD patients have anywhere from 36 to 100, occasionally more.
Patients with HD are usually heterozygous, carrying one allele of wild-type HD and
another allele of mutant HD. An effective gene therapy must silence the deleterious
allele without eliminating expression of the normal one. One rational strategy
is to make use of RNA interference (RNAi) technology, currently undergoing
rapid development. RNAi is an evolutionarily conserved process mediated by
double-stranded RNAs of 19–23 nucleotides. These short small interference
RNAs (siRNAs) efficiently and specifically cleave or degrade targeted mRNAs,
Silencing Huntington’s Disease Gene with RNAi 133
reducing or even preventing their translation to proteins. Unlike traditional drug

development, which targets disease-associated proteins, the great advantage of
RNAi is its potential to block expression of the mutant protein at the RNA level,
thereby preventing physiological changes at their source rather than via the
activated pathological pathways or simply by countering symptoms. Because HD
is caused by a single, highly penetrant mutation at a known locus, the disease is a
promising target for RNAi therapy. In addition, the specificity and effectiveness of
RNAi technology makes it a valuable and promising tool for investigating the
physiological functions of wild-type huntingtin and the pathological impact of
mutant huntingtin in disease progression. Most importantly, this technology may
enable us to develop new RNAi-based therapy for HD.
2 A New Approach to Understanding Normal Function

of Wild-Type Huntingtin
Mutant huntingtin is toxic to neurons, but wild-type huntingtin is important to such

physiological functions as vesicle trafficking, axonal transport, and transcriptional
regulation (DiFiglia et al. 1995; Velier et al. 1998; Steffan et al. 2000; Waelter
et al. 2001; Zuccato et al. 2001; Gunawardena et al. 2003; Goehler et al. 2004; Qin
et al. 2004; Cattaneo et al. 2005; Woda et al. 2005; Leavitt et al. 2006). Moreover,
Hdh-null mice die in utero during early embryonic development (Duyao et al. 1995;
Nasir et al. 1995; Zeitlin et al. 1995). In mice with a conditionally inactive Hdh gene,
termination of huntingtin expression early in the postnatal period results in neurode-
generation (Dragatsis et al. 2000). Such loss-of-function experiments demonstrate that
the huntingtin protein is critical to neuronal development and survival, but they tell us
little about its mechanism of action. In addition to experiments with overexpression
of huntingtin, investigations using RNAi technology in primary cultured cells,
immortalized cell lines, and in vivo injection have revealed new functions of wild-
type huntingtin and elucidated the pathophysiology caused by mutant huntingtin.
Inhibition of huntingtin in Drosophila resulted in axonal transport defects,
indicating that it also plays an important role in neuronal survival in the mature
brain. The protective role of wild-type huntingtin in neuronal survival may come
from its inhibitory effect on a key cell death executioner, caspase-3 (Kuida et al.
1996; Zhang et al. 2006), and/or huntingtin’s interaction with mitochondrial protein
Omi/HtrA2 (Inagaki et al. 2008). These observations were made in a variety of
in vitro systems. Knocking down endogenous wild-type huntingtin by synthetic
siRNA in both neuroblastoma (N2a) and striatal neuronal lines (ST14A) sensitized
cells to death accompanied (and probably caused) by a significant activation of
caspase-3 (Zhang et al. 2006). Huntingtin interacted with active caspase-3 and
inhibited caspase-3 activation in a cell-free system (Zhang et al. 2006). Microarray
analysis with cortical, striatal, and cerebellar primary neurons expressing mutant
huntingtin showed that heat shock protein 70 (Hsp70) is unregulated in granule
cells of the cerebellum (whose neurons are unaffected by HD) (Tagawa et al. 2007)
and that mitochondrial-membrane protein Omi/HtrA2 is specifically downregulated
in striatal neurons (Inagaki et al. 2008). Granule cells lost their resistance to the
toxicity of mutant HD when Hsp70 was inhibited by siRNA (Tagawa et al. 2007).
SiRNA-mediated suppression of Omi-HtrA accelerated mutant htt-induced cell
death of primary neurons (Inagaki et al. 2008). Hsp70 is a chaperone protein that
protects neurons from the toxicity of mutant HD (Chai et al. 1999; Warrick et al.
1999; Zhou et al. 2001; Wacker et al. 2004). Omi/HtrA2, a mitochondrial protein
essential to the maintenance of mitochondrial homeostasis, suppresses apoptosis of
neurons by preventing the accumulation of activated Bax in the mitochondrial outer
membrane (Chao et al. 2008). Neuron type – selective dysregulation of Hsp70 and
Omi-HtrA2 might be related to striatal neuron-specific pathology in HD (Tagawa
et al. 2007; Inagaki et al. 2008).
Besides mitochondria, the functions of huntingtin have also been related to other
organelles such as the endoplasmic reticulum (ER) and Golgi apparatus and
they are cell-type specific (Omi et al. 2005; del Toro et al. 2009). Depletion of
endogenous wild-type huntingtin by siRNA in N2a and T98G cells resulted in a
congregated ER network as shown by staining with anticalnexin, while other
organelles showed no significant changes when immunostained with their specific
markers (Omi et al. 2005). However, depletion of huntingtin by RNAi in HeLa cells
results in partial Golgi disruption (Caviston et al. 2007). Some huntingtin might
associate with dynein and facilitate dynein-based vesicle motility (Caviston et al.
2007). Using full-length unspliced genomic HD it was found that huntingtin coloca-
lized in the cytoplasm with trans-Golgi and clathrin-coated vesicles (Strehlow et al.
2007). Another important observation regarding Hdh-null mouse embryonic stem
cells and neurons is that they have fewer transcripts destined for the extracellular
space and show less lysosomal activity and apoptosis than their normal counterparts.
It has been suggested that huntingtin plays a role in trafficking a discrete set of
proteins between Golgi and the extracellular space (Strehlow et al. 2007). Optineurin
is a huntingtin-interacting protein colocalized with huntingtin in the Golgi apparatus.
Knocking down wild-type huntingtin by siRNA resulted in a more diffuse distribu-
tion of optineurin and GT-YFP (a special label for the Golgi apparatus) and decreased
mannose-6-phosphate receptor postGolgi transport (del Toro et al. 2009). Optineurin
and Rab8 form a complex that regulates postGolgi transport. While huntingtin is
crucial for localization of optineurin, disruption of huntingtin by RNAi results in
partial Golgi disruption and impaired optineurin/Rab8-dependent postGolgi traffick-
ing to lysosomes (del Toro et al. 2009). These reports indicate an important role for
huntingtin in the regulation of postGolgi trafficking.
RNAi was also used in a high-throughput screen for modifiers of aggregate
formation in Drosophila larval CNS-derived cells expressing mutant htt exon 1
with 62 CAG repeats. The 68 candidates from this screen were further analyzed in
degenerative Drosophila eyes in in vivo models of HD. The degeneration of fly eyes
is characterized by loss of pigment cells and increased formation of dark necrotic
spots on the external eye. In a second functional screen, 21 genes were found to
have effects on mutant huntingtin-induced toxicity in fly eyes. These newly identi-
fied modifiers are related to nuclear transport, nucleotide processing, and the
ubiquitin-proteasome system (Doumanis et al. 2009).
Using a virus-carried shRNA, huntingtin expression was decreased 3.5-fold in

the striatum of C57BL/6 mice. Microarray evaluation showed that several hundred
transcripts were up- or down-regulated by 2.0-fold. Compared with the array
results on cells from Hdh-null mice, 18 genes were down or up-regulated in the
same direction (Strehlow et al. 2007; Zhang et al. 2008; Boudreau et al. 2009).
Among them are Hdbp, whose protein is involved in cholesterol homeostasis, and
Foxp1, Sox11, and Sox21, all transcription factors characteristic of postmitotic
neurons. Also modulated is the Mbp gene, which yields a protein associated with
free radical-induced tissue damage. In addition, some genes identified in these
investigations of Hdh-knockdown mice were related to the physiological processes
and structures known to be affected in HD. They include lipid metabolism, the
microtubules of the cytoskeleton, cellular localization, protein transport/secretion,
GTPase signaling, apoptosis, nervous system development, and synaptic transmission
(Strehlow et al. 2007; Zhang et al. 2008; Boudreau et al. 2009). Lentivirus-delivered
shRNA targeting both HD alleles (i.e., the mutant human HD transgene and the
endogenous mouse HD) also affected molecular pathways of known importance to
the pathophysiology of HD (Drouet et al. 2009). Among them are G protein-coupled
pathways of receptor signaling, synaptic long-term potentiation/depression, axonal
guidance, cyclic adenosine monophosphate-mediated signaling, and calcium and
glutamate signaling (Drouet et al. 2009). Evidently, significant transcriptional
changes accompanied silencing of the HD alleles (Drouet et al. 2009).
Although RNAi is a good tool with which to study huntingtin function, it has its
limitations. Unlike gene knockout, RNAi does not completely abolish synthesis
of the target protein. The fraction of HD mRNA that escapes degradation or
cleavage varies among cell types and depends on transfection efficiency and the
concentration and structure of the RNAi. Some functions of huntingtin may require
the protein to be expressed at its normal level. Other functions may need only a low
level of protein expression. The latter activities may escape detection in partial
loss-of-function studies using RNAi. The occurrence of off-target effects is another
major concern in this part of study. While a high concentration of RNAi efficiently
silences the gene of interest, it also increases problems due to off-target effects. The
optimum dosage of the interfering RNA is difficult to estimate. Not only is it
different for different target genes and different cell types, it varies with the nature
of the interfering species, i.e., siRNA or shRNA, unmodified or modified siRNA.
Designing multiple interfering RNAs that target various portions of the huntingtin
sequence will improve the chances of making one that causes effective and, more
importantly, specific gene silencing.
3 Nonallele-Specific Silencing of Huntingtin
HD is one of several dominant hereditary diseases (which also include spinocere-

bellar ataxia or SCA and prion disorders) that have been studied intensively in
preclinical testing with animal models. To determine the potential clinical efficacy
of experimental drugs and to establish proof of therapeutic principle, mice have
been genetically modified to express full-length human mutant huntingtin or a

fragment thereof (Menalled and Chesselet 2002). The former type of transgenic
animals, e.g., YAC46, 72, and 128 (Hodgson et al. 1999; Slow et al. 2003), has the
entire human HD gene with an expanded CAG repeat. The latter type of animals
carries just the N-terminal of the human HD fragment with CAG expansion in their
germ line, as is the case for HD-N171-82Q, R6/1, and R6/2 mice (Mangiarini et al.
1996; Schilling et al. 1999). Mice with the larger transgene develop a neurologic
syndrome that progresses slowly; those with the small fragment degenerate much
faster and more closely resemble the human juvenile onset of HD. In all of these
transgenic constructs, animals exhibit symptoms reminiscent of HD in humans:
nuclear inclusion bodies in neurons, loss of striatal neurons, and motor dysfunction.
Because of the short life span of the latter transgenic animals, they were the first to
be tested for the effect of synthetic siRNA and virus-borne shRNA. These in vivo
procedures were conducted to provide proof of principle for RNAi-based therapeutic
interventions (Boudreau and Davidson 2006; Denovan-Wright and Davidson 2006;
Harper 2009). Several potentially beneficial RNAi delivery strategies use adeno-
associated virus (AAV), adenovirus, and lentivirus to deliver shRNA or artificial
microRNA vectors. Another method uses lipofectamine2000 or cholesterol-assisted
delivery of synthetic siRNA. Both shRNA and siRNA were injected once in all these
studies; most injected RNAi into the striatum, and one injected siRNA into the lateral
ventricle (Wang et al. 2005). Some of these RNAs are “nonallele-specific,” meaning
that they are designed to target transcripts for both human HD transgene and mouse
endogenous HD. Some of these early studies using RNAi targeted the mutant human
HD transgene, which would not silence endogenous mouse HD. However, because
these sequences will silence both human wild-type and mutant HD alleles when
eventually undergo human clinical trials, they are also summarized in this “nonallele-
specific” category. In these proof-of-principle animal studies, RNAi against the 50
portion of the transgene (mutant human HD transcript) silenced mutant HD and
slowed neurodegeneration in these animals (Table 1).
The transgenic mouse HD-N171-82Q contains exon1–3 of human HD mRNA
(Schilling et al. 1999). This animal model for HD displays motor deficits including
shortened stride length and impaired performance on an accelerating rotarod, as
well as such neuropathological abnormalities as E48-positive neuronal inclusions.
The typical lifespan for these animals is approximately 5.5 months (Schilling et al.
1999). In one study, 4-week-old mice were injected with an AAV1-based construct
that expresses U6 promoter-transcribed shHD2.1 (AAV-sh2.1) targeting nucleotides
416–436 of human HD mRNA (exon 2) (Harper et al. 2005). This striatal delivery of
AAV.sh2.1 reduced intracellular levels of huntingtin protein in mice 5 weeks of age
and mutant HD mRNA by 51–55% in mice 5.5 months of age. At the cellular level,
this treatment decreased the number of inclusion bodies. Most importantly, host mice
showed improved performance on the rotarod and had longer stride length than
HD-N171-82Q animals injected with AAV.shLacZ (Harper et al. 2005). Although
it was also noticed that AAV.sh2.1 injection impaired rotarod performance of
wild-type mice at 10 weeks, this effect resolved by 18 weeks (Harper et al. 2005).
It was suggested that AAV-shRNA might cause toxicity and result in short-term
Table 1 Animal trials of therapeutic applications of RNAi to HD
Animal model RNAi target Delivery RNAi Reduction of Reduction of Histological Behavioral Prolonged Ref.
methods injection sites mRNA level protein level observations performance after lifespan
following RNAi as RNAi as compared to
compared to controls controls w/o RNAi
w/o RNAi
Transgenic HD-N171-82Q Mutant human AAV1-shRNA Bilateral, into 51–55% at 5.5 Reduction Absent of huntingtin Greater stride length (Harper et al.
mice transgenic HD transgene striatum, at months appears at reactive inclusions at 4 months; 2005)
mouse model 4 weeks 5 weeks in transduced cells Improved rotarod
and fewer mEM48- performance at 10,
positive inclusions 18 weeks; No
in cerebellum after weight
cerebellar injection normalization
HD-N171-82Q Both mutant AAV1-miRNA Bilateral, into 60% at 4 weeks, Improved rotarod Yes, 75% (Boudreau et al.
transgenic human and or shRNA striatum at 75% at 13 performance at 14 alive 2009)
mouse model mouse HD 7 weeks weeks and 18 weeks compare
gene (unilaterally postinjection with 45% of
in QPCR control at 20
assay) weeks
CAG140 Both mutant AAV2/1- Bilateral, into 70% at 4 AAV-miRNA cassette (McBride et al.
human and shRNA or striatum at months reduced cytotoxicity 2008)
mouse HD miRNA 5 weeks postinjection and microglial
gene (unilaterally activation from
in miRNA/ AAV-shRNA
shRNA (CD11b mRNA,
comparison antiDARPP-32 and
study) antiIbal) 4 months
postinjection
R6/2 Mutant human siRNA- Lateral 50% at 4–7 Fewer nuclear Less weight loss at Yes (Wang et al.
HD transgene lipofectamine ventricle at days inclusion and 13, 14 and 15 2005)
2000 postnatal postinjection protein aggregates weeks; delayed
day 2 stained by anti- and reduced feet-
huntingtin clasping response
antibodies. at 8 weeks;
At the anatomical level: improved
not as much performance on
ventricular rotarod at 6–8
enlargement and week; improved
less striatal atrophy open field activity
at 12+ weeks
(continued)
Table 1 (continued)
Animal model RNAi target Delivery RNAi Reduction of Reduction of Histological Behavioral Prolonged Ref.
methods injection sites mRNA level protein level observations performance after lifespan
following RNAi as RNAi as compared to
compared to controls controls w/o RNAi
w/o RNAi
R6/2 Adenovirus- Bilateral, into Fewer aggregates (Huang et al.
shRNA striatum at 5 analyzed with 2007)
weeks huntingtin antibody
(2B4) at 4 weeks
postinjection
R6/1 Mutant human rAAV5-shRNA Unilateral or 75–78% 28% Reduced abundance of Delayed the onset of (Rodriguez-
HD transgene bilateral, NIIs at 10 weeks rear paw clasping Lebron et al.
into postinjection for 2 weeks (from 2005)
striatum at (unilateral injection) 20 to 22 weeks;
6–8 weeks bilateral injection)
HD190QG EGFP rAAV5-shRNA Unilateral into 24 weeks, fewer (Machida et al.
striatum at aggregates and 2006)
12 weeks protein
accumulation
(detected by
antiGFP, -htt, -
ubiquitin and filter
trap); restoration of
DARPP-32- and
enkephalin
Mice AAV1/8-HD- Mutant human Cholesterol- Unilateral, 66% at 2.7 days Reduced abundance of Less severe feet- (DiFiglia et al.
expressing 400aa-100Q HD transgene siRNA intrastriatal, postinjection inclusion (EM48, clasping response; 2007)
mutant HD coinjected antihtt) and neuropil less frequent foot-
from virus with AAV- aggregates slipping when
injection HD100Q beam walking
AAV1/2-HD70 AAV-shRNA 2 weeks after 80% 50% Greater abundance of Less frequent (Franich et al.
injection w/ cells staining for spontaneous 2008)
AAV-HD- NeuN and exploratory
70 calbindin-28 and extension of
less for Fluoro-Jade. forepaw
HD-171-128Q Mutant human Adenovirus- Unilateral into 4 weeks postinjection, (Huang et al.
HD transgene shRNA striatum at fewer aggregates 2007)
5 weeks, analyzed with
Coinjected antihuntingtin in
w/Adenoviral right striatum
vector HD- compared with left
171-128Q striatum injected
with control vector
Lentivirus HD- Mutant human Lenti-shRNA Striatum 2 months after (Drouet et al.
N171-82Q HD transgene injection: More 2009)
injected rat DARPP-32 and
NeuN positive cells;
reduction of EM48-/
Ubiquitin-positive
inclusions;
preserved succinate
dehydrogenase
(SDH) function;
greater rate of
cerebral glucose
metabolism
(CMRGlu)
Lentivirus Both human and Lenti-shRNA Striatum 85.7% in Laser Increased number of (Drouet et al.
HD-N171-82Q mouse HD micro- DARPP-32- and 2009)
injected rat gene dissected NeuN-positive
GFP positive cells; reduction of
cells EM48-/ubiquitin-
positive inclusions;
at 1, 3 and 9 months
postinjection
deficits in motor behavior. Later studies also found that AAV-shRNA causes
sequence-specific neurotoxicity to the mouse striatum (McBride et al. 2008;
Boudreau et al. 2009). A particularly successful strategy was to infect HD transgenic
mice with AAV-microRNA instead of AAV-shRNA (McBride et al. 2008; Boudreau
et al. 2009). AAV-mi2.4 is an artificial microRNA-based cassette driven by RNA
polymerase II promoter that can generate siRNA that targets the same sequence
within exon 2 of both introduced and endogenous HD transcripts (McBride et al.
2008; Boudreau et al. 2009). AAV-mir2.4 produced less antisense RNA and less
toxicity but achieved similar gene silencing effects as AAV-sh2.4 (McBride et al.
2008; Boudreau et al. 2009). In both CAG140 heterozygous knock-in mice and
HD-171-82Q mice, AAV-mi2.4 injection resulted in more cells exhibiting immuno-
reactivity toward DARPP-32 compared with AAV-sh2.4 injection. Moreover,
AAV-mi2.4 did not increase Iba1 staining as much as AAV-sh2.4 and is unlikely to
have elicited the same extent of microglial activation. Although there is controversy
regarding the immunoreaction as judged by real-time analysis of CD11b mRNA in
these two transgenic mice (McBride et al. 2008; Boudreau et al. 2009), AAV-mi2.4
clearly exhibited less neuronal toxicity than AAV-sh2.4. In addition, RNA polymer-
ase II promoter has the potential to be used in cell-specific expression. This artificial
AAV-microRNA-based RNAi delivery system might provide a new tool to efficiently
silence the mutant HD gene and improve the safety of AAV-shRNA.
The R6/2 transgenic mouse is the most thoroughly studied animal model for HD.
The insertion in its germ line of exon 1 of the human HD gene with 105 CAG
repeats causes rapid neurologic decline with death after about 3 months (Mangiarini
et al. 1996). To decrease expression of the mutant human HD-derived transgene,
postnatal day 2 mice were injected with synthetic siRNA targeting human HD exon
1 in the brain’s lateral ventricle (siRNA-HDexon1) (Wang et al. 2005). This
treatment caused a drop in expression of mutant HD mRNA in the animals’ brain
4–7 days later. Staining with antibodies against huntingtin protein revealed less
protein aggregation in the striatum of 8-week-old R6/2 mice treated with siRNA-
HDexon1 than in mock-treated or untreated R6/2 mice brains of the same age.
Moreover, siRNA-HDexon1-treated R6/2 mice retained greater motor control than
untreated animals (Wang et al. 2005). Up to the age of 8 or 9 weeks, the animals’
body weight remained relatively stable compared with untreated R6/2s. In addition,
the onset of the feet-clasping response was delayed by the synthetic siRNA, and
it was repeated at a lower frequency once it became manifest. At 12 weeks, an
age when symptoms were already advanced, test animals exhibited more activity in
the open field test than untreated controls. Finally, intraventricular injection with
synthetic siRNA extended the transgenic animals’ lifespan by 15 days (Wang et al.
2005). Although siRNA has been reported to degrade rapidly in vivo, intraven-
tricular injection on the day after birth had unexpectedly long-lived effects (Wang
et al. 2005). Apparently, the molecule’s ability to cause physiological changes goes
beyond simple silencing of gene expression. It seems that the loss of mutant HD
expression early on has long-lasting effects that persist after protein synthesis
resumes. Besides the experiments using synthetic siRNA, R6/2 mice were treated
with a high-capacity adenoviral vector expressing shRNA against HD exon1
(Huang et al. 2007). Immunohistological analysis of huntingtin immunoreactive

aggregates in brain slices from R6/2 animals revealed that this adenoviral shRNA
also reduced the abundance of protein aggregates (Huang et al. 2007).
R6/1 is another R6- series mice carrying exon-1 of HD with 115-CAG expansion
CAG expansion. These mice have a slightly longer life span than R6/2 mice. AAV-
hiHUNT-1 (a shRNA whose target is within 50 UTR of human mutant huntingtin)
was injected into the striatum of R6/1 mice 6–8 weeks old. Ten weeks later, a 78%
reduction of mutant HD mRNA and a 28% reduction of protein were found in the
transduced area compared with contralateral injection of rAAV-shRNA control
vector (Rodriguez-Lebron et al. 2005). The size and number of neuronal inclusion
bodies are also decreased with normalization for levels of preproenkephalin and
DARRP-32 mRNA, two transcripts that are themselves significantly decreased in the
R6/1 mice brain. Bilateral injection of rAAV-siHUNT-1 into the striatum delayed the
onset of rear paw clasping (Rodriguez-Lebron et al. 2005).
The HD190QG transgenic mouse expresses N-terminal huntingtin with 190
CAG repeats fused with EGFP (Kotliarova et al. 2005). This animal also showed
progressive motor abnormality and neuropathology in the striatum. rAAV-shRNA
targeting the EGFP (rAAV5-shEGFP) was injected into the striatum of this transgenic
mouse at 12 weeks (Machida et al. 2006). ShEGFP reduced aggregate formation,
suppressed insoluble protein accumulation, and restored DARPP-32 expression in the
striatum at 24 weeks (Machida et al. 2006).
Although there is striatal neuron loss, the loss of neurons is not restricted to the
striatum; it is observed globally throughout the transgenic HD mice brain. This
drawback was addressed by generating acute models for HD by intrastriatal injec-
tion of virus-borne mutant HD (Senut et al. 2000; de Almeida et al. 2002). The
resulting neuronal death occurred with a distribution similar to that seen in human
HD, i.e., it was more closely confined to the striatum than is the case in transgenic
HD mice. These rapid-onset HD models have progressive and robust neuropathol-
ogy and motor impairment, which facilitate experimental testing (DiFiglia et al.
2007). In addition, these methods can be developed into preclinical studies in
nonhuman primates (Palfi et al. 2007). Acute HD model animals were treated
with test RNAi as a new therapeutic approach (DiFiglia et al. 2007; Franich et al.
2008; Drouet et al. 2009). One study used mice injected in the right striatum with
AAV expressing 365 N-terminal amino acids from the human HD with 100 CAG
repeats (AAV-htt100Q) (DiFiglia et al. 2007). These animals suffered a debilitating
rapid-onset syndrome in which striatal neurodegeneration and behavioral
impairment were evident within 2 weeks. Cholesterol-conjugated synthetic
siRNA against the HD transgene prolonged survival, ameliorated deficits in neuron
function, and reduced aggregates and inclusion bodies in mice injected with AAV-
htt100Q (DiFiglia et al. 2007). In another acute rat model of HD, injection of
AAV1/2-HD-70 into the striatum resulted in substantial loss of projection neurons,
cholinergic interneurons, and NAPDH-d interneurons (Franich et al. 2008). In that
model, the level of mRNA for mutant huntingtin increased 150-fold within 2 weeks
but dropped to near endogenous levels by 5 weeks. AAV-shHD2 targeting human
huntingtin mRNA was injected 2 weeks after the first injection of AAV1/2-HD-70.
ShHD2 demonstrated substantial neuroprotection with increased numbers of NeuN

and calbindin-positive cells and decreased numbers of Fluoro jade B-positive cells.
In addition, AAV1/2-shHD2 injection reduced impairment of spontaneous explor-
atory forepaw in this HD model (Franich et al. 2008). Another model of HD was
created by giving rats an intrastriatal injection of a lentivirus vector expressing HD-
171aa-82Q (Drouet et al. 2009). Many of the resulting molecular changes were
reversed by inducing the expression of lentivirus-shRNA targeting the HD171aa-
82Q transgene, including preservation of DARPP-32 and NeuN-positive cells,
decreasing EM48-/Ubiquitin-positive inclusions, and recovering succinate dehydro-
genase (SDH) function; the rate of cerebral glucose metabolism (CMRGlu) was also
increased (Drouet et al. 2009). In addition, high-capacity adenoviral vector
(HC-AdHB07) was injected into the striatum to generate an acute HD model expres-
sing HD-171aa-128Q (Huang et al. 2007). Huntingtin aggregates in neuronal cells
were noted in the vicinity of injection sites. HC-AdHB04 (expressing shRNA target
HD-171-12Q) or HC-adHB05 (expressing antiEGFP shRNA) was co-injected with
HC-AdHB07. Mutant huntingtin aggregates were efficiently reduced by coinjection
with HC-AdHB04, but not with HC-adHB05. This was similar to the effect of
HC-AdHB04 in R6/2 mice (Huang et al. 2007).
Introducing the transgene by virus injection caused great pathological changes in
the striatum, but these effects were not confined to medium spiny neurons. In some
cases, it affected neuron types that are largely unaffected in HD. Injection of
AAV1/2-HD70 into rat striatum resulted in quite rapid expression of HD70 in
large interneurons and weaker immunoreactivity in medium-sized neurons (Franich
et al. 2008). AAV-HD70 caused significant striatal interneuron death, including
neuropeptide-Y, parvalbumin, and choline acetyltransferase immunoreactive inter-
neurons (Franich et al. 2008). Whether this lack of neuron-type specificity will
compromise the technique’s potential for testing an RNAi-based therapy must be
further investigated in animal models, especially nonhuman primate models.
Some studies have addressed whether postsymptomatic injection of RNAi is also
beneficial to HD. Ten years ago, data from conditional mutant HD knockout mice
demonstrated disappearance of inclusions and behavioral improvement when
mutant HD gene expression was blocked (Yamamoto et al. 2000). For conditional
RNAi application, one study developed a lentiviral vector allowing shRNA levels
to be manipulated by addition/withdrawal of doxycycline. Two months after
lentivirus-sihtt1.1 (which targets mutant HD only) was injected into the striatum
of HD rats, when pathology begins to appear, doxycycline was administered,
turning on sihtt1.1 expression. Inducing shRNA synthesis rescued these rats from
depletion of DARPP-32-positive cells, and partially cleared EM48 or ubiquitin-
positive inclusions (Drouet et al. 2009). In another study, 2 weeks after intrastriatal
injection of AAV1/2-HD70, substantial loss of neural immunoreactivity was evident.
However, AAV-shHD2 injected at that time prevented HD-70-induced neurode-
generation and behavioral impairment (Franich et al. 2008). It was suggested that
RNAi-based therapy could reverse HD pathology even once disease symptoms had
become obvious (Franich et al. 2008; Drouet et al. 2009).
While allele-specific silencing is the preferred treatment strategy in animal models

of HD, nonallele-specific silencing is also being investigated. As described above,
intrastriatal injection of lentivirus carrying the harmful htt171-82Q caused test rats to
lose DARPP-32 and NeuN-positive cells and to accumulate EM-48 and ubiquitin-
positive inclusions (Drouet et al. 2009). Lentivirus-mediated shRNAs (sihtt 3 and 6),
which target mRNA of both mutant human HD and endogenous mouse HD, reduced
the number of EM48/ubiquitin-positive inclusion bodies and preserved more DARPP-
32 and NeuN-positive cells (Drouet et al. 2009). Lentivirus-mediated sihtt 13, which
targets only the mRNA of endogenous mouse HD, has no detrimental effects on
GABAergic neuron survival or huntingtin inclusions for up to 2 months (Drouet et al.
2009). Moreover, lentivirus-mediated sihtt6 (silencing both endogenous mouse HD
and mutant human HD) has long-term effects similar to those of sihtt1 (silencing only
mutant human HD) in reducing HD pathology. sihtt6 did not cause striatal toxicity or
striatal vulnerability up to 9 months after injection. DARPP-32-positive cells are well
preserved, and ubiquitin staining is absent at 1, 3, and 9 months in the striatum of
sihtt6-treated htt171-82Q rats. Similar nonallele-specific silencing effects were
obtained in experiments on HD transgenic mice expressing human HD-N171-82Q
(Boudreau et al. 2009). In these investigations, test mice were injected with an
artificial microRNA that targets a sequence common to both the foreign and endoge-
nous HD gene (McBride et al. 2008; Boudreau et al. 2009). At 7 weeks of age, animals
treated with AAV-mi2.4 performed better on the rotarod and survived longer than did
control-treated HD-N171-82Q mice. The endogenous mouse HD gene and the mutant
human HD gene responded to the AAV-mir.2.4 silencing effort similarly. At 4 and 13
weeks after bilateral intrastriatal injection with AAV-mi2.4, transcripts of both HD
genes were 60% (4 weeks) and 75% (13 weeks) lower in injected mice than in control
animals (Boudreau et al. 2009). When assessed at 10, 14, and 18 weeks of age, test
animals’ rotarod performance had not declined as much as that of control-treated HD-
N171-82Q mice at 14 and 18 weeks (Boudreau et al. 2009). In addition, AAV-mi2.4-
treated mice displayed a trend of improved survival. At 20 weeks, while 45% of
control-treated HD mice survived, over 75% of AAV-mi2.4 treated mice were still
alive. The results of both these experiments demonstrate that partial inactivation of
wild-type huntingtin is merely tolerated, while concomitant silencing of its mutant
counterpart is beneficial (Boudreau et al. 2009; Drouet et al. 2009). Since these studies
are based on silencing the endogenous wild-type HD gene and mutant transgene, the
silencing effect might be different when both HD genes are endogenous. In the
previous case in rat, the transgene introduced is expressed at a level 25-fold higher
than in the endogenous species (Drouet et al. 2009). In human HD, both mutant and
wild-type HD alleles are equally expressed. Moreover, major depletion of wild-type
huntingtin could still interfere with neuronal function, especially under stress condi-
tions or simply over the long term. We can expect to see future experiments designed
to address how effectively and how long nonallele-specific targeting sufficiently
silences mutant HD gene effects without causing detrimental toxicity from partially
losing wild-type HD, and to what extent this loss can be tolerated.
The potential off-target effects of AAV-mediated HD silencing are illustrated
by the different responses of R6/1 mice to unilateral intrastriatal injection with
rAAV-siHUNT-1 and rAAV-siHUNT-2. rAAV-siHUNT-1 did not strongly silence

transcripts besides it of HD. By contrast, rAA-siHUNT-2 caused depletion of
the mRNAs encoding striatum-specific proteins preproenkephalin (ppENK) and
DARPP-32 (Rodriguez-Lebron et al. 2005; Denovan-Wright et al. 2008). This
negative impact on transcripts in R6/1 mice was also reported with rAAV-HD6
and HD7 (two catalytically active hammerhead ribozymes) against the same region
of human HD mRNA targeted by siHUNT-2 (Denovan-Wright et al. 2008). This
suggests that the reduction in these two transcript mRNAs is an off-target effect
from the targeting sequence and inhibited some unidentified transcripts, not those
from rAAV- or ribozyme-mediated transduction (Denovan-Wright et al. 2008). In
contrast, AAV-shHD2, which targets a region of human HD mRNA similar to that
targeted by siHUNT-2, was not found to cause off-target reduction of proenkepha-
lin in a rat HD model by AAV-HD70 injection (Franich et al. 2008). The different
off-target effects might come from the responses of different strains of mice to the
same interfering RNA. Off-target effects and RNAi-induced toxicity that are
problematic in some animals may be unimportant in others (Rodriguez-Lebron
et al. 2005; Wang et al. 2005; Denovan-Wright et al. 2008; Franich et al. 2008;
McBride et al. 2008; Boudreau et al. 2009). All these findings should remind us that
in the design of future clinical trials there may be differences in off-target effects
among individual patients. In anticipation of this complexity, preclinical studies on
RNAi therapy should be conducted in HD mice with various genetic backgrounds.
In addition, HD is a chronic neurodegenerative disease; the transgenic animal
model, with its quick disease onset and progression, resembles some features of
this disease, but not the slow disease progression of HD. We hope to determine
whether virus-borne shRNA or synthetic siRNA with single or multiple injections
can elicit effects in the transgenic mouse model expressing full-length HD similar
to those we have seen in rapid-onset transgenic models using HD fragments.
4 Allele-Specific Silencing of Mutant Huntingtin
Most HD patients are heterozygotic at the HD locus, carrying one allele for the wild-
type protein and one for its mutant counterpart. The former protein is essential to cell
survival; the latter is harmful, with greatest toxicity to striatal neurons. Developing
RNAi therapy for HD faces a dilemma: how to silence expression of one protein
while maintaining expression of the other? The technology of RNA interference
offers the requisite efficiency and specificity to satisfy both imperatives. The critical
difference between wild-type and mutant HD genes is the number of tandem CAG
triplets at the 50 -end of the sequence – over 36 in deleterious alleles but generally
below 30 in healthy ones. Though this variable sequence is an obvious target for
RNAi, it is of minimal practical use. RNAi molecules must have 19–23 bases to act
efficiently yet not elicit a strong immune response. Such molecules are too short to
span the CAG repeat, making them incapable of distinguishing between mutant and
wild-type alleles. An alternative approach is to target a polymorphic site that is
genetically linked to the CAG expansion. There are many candidate sequences.
A total of 190 SNPs (single-nucleotide polymorphisms) are known in the introns and
exons at the HD locus at chromosome 4p16.3. Analysis of HD alleles from 65 patients
of European origin revealed that disease-associated SNPs form a cluster of similar
haplotypes (haplogroup A) found on 95% of CAG-expanded chromosomes. Two
variants of haplogroup A (A1 and A2) are dramatically and specifically enriched on
HD chromosomes and are therefore markers for persons at risk of the disease (Warby
et al. 2009). In a study of 327 European Caucasian HD patients, 86% were heterozy-
gous at one or more of 26 SNPs analyzed (Lombardi et al. 2009). In these cases,
allele-specific RNAi is potentially applicable should the shRNA target polymorphic
sites linked to CAG expansion in a downstream part of the gene. Many questions
remain about methods to approach allele-specific RNAi therapy for HD and how
successful it will be. The following issues are important: (1) Is it possible to find one
or a few SNPs that are common and have a sufficiently strong linkage to the CAG
expansion to be generally useful for a relative larger HD population and can be tested
in large-scale clinical trials to obtain permits from drug regulatory agencies? Alter-
natively, can one use allele-specific RNAi therapy only for personal medicine? (2) On
an individual basis, can one rapidly identify SNPs with sufficiently tight linkage to
CAG expansion? (3) When wild type and mutant alleles only differ by one to three
nucleic acids, can one design a RNAi that silences the mutant allele enough to render
it innocuous but preserves sufficient expression of the normal species to execute
essential functions? While there is a long road ahead, researchers have already made
a significant start. Searches of the human-genome database have revealed many
polymorphic sites within the HD gene. Some of these SNPs occur at a high frequency
among test groups of substantial size. The most promising results come from several
studies on fibroblasts from HD patients. In cultures of these cells, RNAis have been
used to silence the mutant HD allele while preserving expression of its wild-type
counterpart (Table 2).
A noteworthy polymorphism that is linked to CAG expansion is the D2642
triplet deletion in exon 58 of the HD gene (Ambrose et al. 1994; Novelletto et al.
1994; Almqvist et al. 1995; Rubinsztein et al. 1995; Vuillaume et al. 1998). The
typical sequence beginning at this position is GAG.GAG.GAG.GAG; that for
D2642 is GAG.GAG.GAG. The deletion of a codon causes the loss of one of four
tandem glutamate residues in the huntingtin protein. The three-glutamate species
occur substantially more frequently among HD alleles than in those without CAG
expansion. According to previous study, the D2642 deletion occurred in 38% of HD
alleles but in only 7% of healthy ones (Ambrose et al. 1994). Efficient and specific
silencing of a D2642-tagged HD allele has been achieved in an in vitro system.
Immortalized cells were transfected with the two HD alleles, each of which had
been fused to the 30 end of a luciferase reporter gene. Monitoring the intensity of
fluorescence from these cells showed that the appropriate siRNA specifically
silenced expression of the mutant copy. One skin fibroblast from an HD patient
was identified that carried both an HD allele marked by the D2642 deletion and its
wild-type counterpart. These cells were transfected with each of four siRNAs
designed to target the polymorphic site. One of these molecules efficiently and
specifically silenced the D2642-marked, CAG-expanded allele in this HD fibroblast
Table 2 List of experimentally tested polymorphic sites linked to HD mutant allele with CAG
expansion
Reference SNP sites HD Location Identified HD Maxim References
number population fibroblast to reduction of
validate mutant
isoform-specific allele (%)
siRNA
rs363125 A/C 11% Exon 39 HD fibroblast 80% (van Bilsen
GM04022 et al.
2008)
D2642 Deletion 38% Exon 58 HD fibroblast 43% (Zhang
deletion of one GM09197 et al.
GAG 2009)
Rs362307 C/T 48% 30 UTR Luciferase reporter (Palfi et al.
for each isoform 2007)
Rs362273 G/A 35% Exon 57 Luciferase reporter (Palfi et al.
for each isoform 2007)
Rs362331 C/T 39% Exon 57 HD fibroblast 58% (Lombardi
GM09197 et al.
2009)
Rs2276881 A/G 8.6% Exon 60 Luciferase reporter (Lombardi
for each isoform et al.
2009)
CAG Expanded 100% Exon 1 HD fibroblasts: 100% (with (Hu et al.
repeats CAG GM09197, 30% 2009)
GM04281, reduction
GM04869, in wt-
GM04719, huntingtin)
GM04717
(Zhang et al. 2009). Similar allele-specific silencing results were obtained from a
later study of this HD fibroblast (Hu et al. 2009).
Only a minority of mutant HD alleles are marked by the D2642 deletion
mutation. Ongoing research aims to identify additional SNP sites in the HD gene
that could be used for RNAi silencing directed specifically to alleles with the CAG
expansion (Liu et al. 2008; van Bilsen et al. 2008). SiRNAs can be designed to
discriminate between two alleles that differ at a single nucleotide, i.e., a SNP
(Schwarz et al. 2006). It was found that the difference between half maximal
silencing (IC50) of mismatch and match reached the maximum discrimination is
from siRNA-target at position 16 (start from 50 guide strand of siRNA) of purine:
purine mismatch (Schwarz et al. 2006). Because the HD locus is large and both
isoforms are over 10 kb, SNPs in that gene are usually distant from the CAG repeat
at the 50 terminal. For this reason, it is difficult to identify SNPs that are linked to the
harmful triplet expansion in mutated HD alleles. Two methods have been used to
identify allele-specific SNPs (Fig. 1). One is to use SNP-specific RT primers to
selectively generate cDNA from a single allele of HD. A second round of PCR is
conducted on the resulting cDNA using primers spanning the CAG repeat sequence
(van Bilsen et al. 2008). Using this method, 11 known SNP heterozygote sites in
fibroblast cells from 21 different HD patients were allele-specifically determined
Fig. 1 Schematic image of two methods to identify allele-specific SNPs in HD
(van Bilsen et al. 2008). For example, fibroblasts from one patient had the rs363125
SNP with cytosine at that position in the mutant HD allele but adenine in its wild-
type counterpart. Using the appropriate RNAi molecule, it was possible to silence
the mutant allele while preserving expression of the wild-type one (van Bilsen et al.
2008). Another method used the circularization method to link the SNP with CAG
repeats (Liu et al. 2008). It was reverse-transcribed as usual. A primer flanking the
SNP with a KasI site at each end was used to amplify DNA ranging from SNP to the
CAG expansion. The key is to circularize this PCR product by intramolecular KasI
ligation. A PCR spanning SNP and CAG repeats was then conducted on two wild-
type and mutant individual ligation products. The length of CAG repeats was
evaluated on the PCR products by sequencing. Because the allele-specific SNP
site and CAG repeats are adjacent, it was easy for direct sequencing to identify the
SNP–(CAG)n linkage (Liu et al. 2008). This method was used in later studies with
large sample sizes (Palfi et al. 2007). Evaluation of 225 human HD and control
samples from American and European carriers revealed that over 48% of patients
were heterozygous at one of 24 identified SNP sites. Incidence of the U isoform of
rs362307 SNP at exon 67 on HD alleles much exceeded that on control samples
(Palfi et al. 2007). Seven out of 16 HD blood samples evaluated have expanded
CAG-linked heterozygous U isoform at the rs362307 site, an approximate 50%
linkage. Moreover, disease-associated SNPs at sites rs363125, rs362273, and
rs362307 are so frequent within this sample that five siRNAs that specifically
silence the harmful allele have the potential to treat three-quarters of these HD
patients (Palfi et al. 2007). In addition, a computational and experimental analysis

was conducted in 327 unrelated European Caucasian HD patients at 26 SNP sites in
the HD gene. It predicted that over 85% of HD patients could benefit from shRNA-
based therapies that target 6–7 SNP sites (Lombardi et al. 2009).
Researchers are trying to adapt RNAi technology to target the expanded CAG
repeat that makes the mutant HD allele harmful. As discussed earlier, standard
techniques using shRNA and siRNA are not applicable to a sequence as long and
regular as (CAG)35. However, longer CAG trinucleotide repeat sequences have been
predicted to form a hairpin structure (Gacy et al. 1995; Sobczak et al. 2003). The
stability of this hairpin conformation might differ between the wild type and mutant
alleles. Pursuing this reasoning, antisense oligoribonucleotides were designed
to silence HD alleles with unusually long CAG repeats. Two types of chemical
modifications to antisense oligomers (peptide nucleic acid conjugation and locked
nucleic acids) gave them the capability to potently silence the mutant but not the
wild-type HD allele in fibroblasts from HD patients (Hu et al. 2009). Targeting the
CAG repeats is an alternative strategy for specific HD allele silencing.
There is no transgenic mouse in which to test the feasibility of allele-specific
RNAi as a therapy for HD. Unlike the situation with nonallele-specific RNA
silencing, researchers remain unable to obtain proof of principle in an in vivo
system. One goal of current research is to construct an animal in which to identify
RNAi molecules that efficiently and specifically silence the mutant HD allele.
Although from previous observations in nonallele-specific silencing in mice, it
seems that partial knock-down of wild-type HD for a period could be tolerated in
animals, wild-type huntingtin is essential for the survival and function of neurons.
It is difficult to construct the necessary transgenic animal to test the on-target and
off-target effects of allele-specific silencing. Transgenic mice can express both
alleles linked with different CAG length with or without the HD fragment. How-
ever, this human HD wild-type allele transgene will not function as endogenous
wild-type mouse huntingtin. It will work only when the polymorphism site on
the mutant allele targeted by the RNAi has the same isoform sequence in human
wild-type allele as it in mice HD. If off-target effects are seen, not only will that be
revealed from the wild-type allele construct, but there will also be a functional
readout from mouse endogenous HD. Experimental approaches exploring the
allele-specific silencing efficiency and therapeutic potential need to be well
designed and evaluated in vivo. Targeting mutant HD alleles linked to SNPs is a
promising strategy. However, SNPs associated with HD alleles differ from patient
to patient. Consequently, allele-specific RNAi-based therapies may be more useful
in personalized medicine than as general remedies.
5 Current Challenges and Future Perspectives
The greatest challenge to advancing an RNAi-based therapy to clinical trials among

HD patients is to achieve efficient silencing of the harmful HD allele with an
interfering RNA that is not itself toxic to cells and model animals. These constraints
are particularly stringent for allele-specific silencing, where on- and off-target
alleles differing by only 1–3 nucleotides. The technique of nonallele-specific
silencing is beset with a different problem, i.e., preserving sufficient expression
of endogenous huntingtin to prevent molecular changes that cause cell death,
especially over the long term.
The foremost difficulty when designing a safe and effective RNAi-based
therapy for HD comes from off-target effects of the RNAi. The first step toward
understanding (and thus solving) this problem is to identify the off-target mRNAs
that are partially or fully silenced. The current approach uses microarrays to
determine the global pattern of mRNA expression with and without changes
effected by RNAi. As is to be expected, the mRNAs responsible for off-target
effects have sequences resembling those of the target species. A “BLAST search”
of the gene database enables the investigator to identify additional mRNAs with
close sequence homology to the target mRNA. Even mRNAs that are perfectly
complementary to the siRNA’s seed region are not always silenced by that
interfering molecule (Birmingham et al. 2006). Moreover, it is difficult to deter-
mine whether a gene is silenced because of an siRNA:mRNA interaction or
simply from nonspecific changes due to systemic perturbations. Another con-
founding phenomenon arises when an siRNA mimics the function of a microRNA
or simply causes its cleavage. MicroRNAs are themselves regulatory elements
that inhibit translation of target messages but do not necessarily degrade them.
In that situation, the cell’s mRNA profile remains the same, whereas the spectrum
of expressed proteins changes. Analysis by proteomics reveals effects due to altered
translation, but the subsequent procedures necessary to evaluate sequence-specific
and concentration-related off-target effects in individual mRNA and translated
proteins are labor-intensive. In addition, the nature of off-target effects varies
among both cell lines and mouse strains (McBride et al. 2008; Boudreau et al.
2009). An example of this confounding variable is the different changes in CD11b
regulation when AAV1-mi2.4 is injected into B6C3F1/J and C57BL/6 mice. Upon
treatment with this RNAi, the former mouse exhibits increased levels of CD11, a
change indicative of microglial activation. By contrast, the latter animals are not
affected by introducing that very RNAi in a microRNA cassette (McBride et al. 2008;
Boudreau et al. 2009).
The above discussion underlines the complexity expected when developing
RNAi therapies for HD. As noted earlier, the clinical responses to RNAi-mediated
gene silencing are likely to vary among patients. The extent and nature of off-target
effects will largely determine who tolerates the therapy and who does not. In
anticipation of such variability, preclinical studies should be conducted using a
number of mouse models. As described above, one variable in the biology of HD
mice is the nature of the transgene. Mice that express a fragment of the HD gene
(i.e., R6/2) suffer rapidly progressing neurodegeneration. By contrast, animals that
carry a full-length HD gene (i.e., YAC128) exhibit more gradual disease progres-
sion that is reminiscent of the chronic symptoms of human HD patients. Conse-
quently, while mice of the former sort are easier to study, the latter are more reliable
models of the human disease.
A major hurdle to developing an RNAi therapy – or any other sort of genetic

medicine – is constructing a system that efficiently delivers the transgenic DNA.
Currently, there are two major strategies with RNAi for HD: (1) virus vectors with
DNA encoding a shRNA; (2) synthetic siRNA. Each approach has distinct advan-
tages and disadvantages, as discussed below in the context of therapies for HD. One
property that recommends the virus-vector system is its long-term effectiveness
following a single injection of DNA. Lasting gene silencing has been demonstrated
using a transgene encoding a fragment of mutant huntingtin (Harper et al. 2005;
Rodriguez-Lebron et al. 2005; Machida et al. 2006; Huang et al. 2007; Franich et al.
2008; McBride et al. 2008; Boudreau et al. 2009; Drouet et al. 2009). Because its
immunogenicity is relatively low, AAV is the most widely used vector for making
RNAi constructs for HD. After a single injection, shRNA expressed from an AAV
vector was shown to cause long-term silencing of a mutant huntingtin transgene
(Harper et al. 2005; Rodriguez-Lebron et al. 2005; Machida et al. 2006; Huang et al.
2007; Franich et al. 2008; McBride et al. 2008; Boudreau et al. 2009; Drouet et al.
2009). The low immunogenicity of AAV as compared with lentivirus and its small
size make it an attractive vector for investigations on RNAi. Another advantage of
AAV is its ability to exist in a stable episomal form, resulting in lasting gene
expression. In addition, transduction selectivity in different neuron populations
from different AAV serotypes’ tropism will improve the AAV transduction speci-
ficity (Davidson et al. 2000; Burger et al. 2004; Paterna et al. 2004; Taymans et al.
2007). Applications of the AAV system are limited, however, due to the modest
size of allowable inserts (4.7 kb). Moreover, AAV-borne shRNA constructs are
relatively difficult to create. In some cases, lentivirus may be a better choice for
inducible RNAi expression (Drouet et al. 2009). In particular, the lentivirus vector
proved to be effective in nondividing neurons. In this and other systems, lentivirus-
based constructs effect long-lasting gene silencing, presumably a result of their
integration into the human genome. This phenomenon is a mixed blessing, how-
ever. As integration of the foreign DNA into the host’s chromosomes may cause
harmful insertion mutations, it also poses a safety issue in clinical trials. Further
constraints on the use of shRNA stem from the molecular mechanism of its action.
Before a shRNA can effect physiological changes, it must be processed by enzymes
for endogenous microRNA biogenesis, e.g., nuclear exportin-5. Because injected
AAV-shRNA competes with and overloads the enzymes that process microRNA, it
disrupts this essential molecular pathway. In one study on mice, injected shRNA
depressed levels of mature liver-specific microRNAs so much as to kill the host
animal (Grimm et al. 2006). Significant buildup of unprocessed shRNA was
detected, suggesting that the processing machinery was largely saturated. Reducing
the dosage of this shRNA vector diminished toxicity to the host. In one study, HD
mice were given intrastriatal injections with AAV-shRNAs that target HD. There
was significant neural cell death in the animals’ striatum both from an shRNA that
targets the endogenous Hdh mRNA and from one targeting the mutant human HD
transgene. Death of striatal neurons was also observed when the shRNA was
designed to have a few mismatches relative to both alleles (McBride et al. 2008;
Boudreau et al. 2009). Moreover, this AAV-shRNA remained toxic to striatal cells
even as its dosage was reduced. Unlike the situation for liver tissue, unprocessed
shRNA molecules did not accumulate to high levels in the striatum of these test
animals. The character and vulnerability of brain cells in the one study contrasts
with those qualities of liver cells in the other (Grimm et al. 2006; McBride et al.
2008; Boudreau et al. 2009). A reasonable explanation is that the affected micro-
RNA-processing machinery differs between liver cells and striatal cells (Grimm
et al. 2006; McBride et al. 2008; Boudreau et al. 2009).
In summary, HD therapy must use AAV-shRNA or AAV-miRNA at a dosage
that silences target genes but does not overwhelm the cell’s ability to process
endogenous small RNAs. Complicating this task is the difficulty of controlling
expression levels of virus-borne shRNA in vivo. A compliant system must allow for
fine adjustment of shRNA expression. The effectiveness of an inducible shRNA
vector for HD has been demonstrated in vivo using a lentivirus (Drouet et al. 2009).
Transcription from a doxycycline/tetracycline-responsive promoter may lend itself
to reversible repression, though such sensitivity and reversibility in processed
RNAi has yet to undergo sophisticated evaluation in vivo (Drouet et al. 2009).
The other important strategy for introducing RNAi into the host cell or animal is
to use naked siRNA. This technique has a major disadvantage when compared with
virus-based methods: unlike virus constructs, synthetic siRNA is easily degraded
in vivo. Nevertheless, naked siRNA has several properties that recommend it.
Simple nucleic acids are less immunogenic than viral complexes. While this
property is not relevant to in vitro systems, it removes a major barrier to application
in vivo (i.e., the ultimate goal – a clinical therapy). At the molecular level, naked
siRNA interferes less with endogenous pathways of microRNA biogenesis. The
central challenge when developing therapies built on this strategy is to identify
nucleic acids sequences with high specificity for the harmful HD. Any attempt to
solve this puzzle must be based on the molecular biology of RNAi. The first step
is to optimize sequence design. Within the molecule’s total length of 19–23
nucleotide pairs, positions 2 through 8 at the 50 -end of the guide strands of RNAi
are of exceptional importance to the recognition of and binding to target mRNAs
(Doench and Sharp 2004; Haley and Zamore 2004). In this way, it is similar to
microRNA, in which the corresponding portion also determines target mRNA
selection. Nucleotides at the 30 -end of the guide strands dictate cleavage efficiency
(Doench and Sharp 2004; Haley and Zamore 2004). In general, the affinity of
a siRNA for an mRNA increases as the number of mismatches decreases. Never-
theless, there can be partial gene silencing of mRNAs that match the introduced
siRNA at as few as 11 nucleotides (Jackson et al. 2003). Bioinformatic analysis
reveals that off-target effects most often arise from binding of the siRNA to a
sequence in an mRNA’s 30 UTR (Jackson et al. 2003; Lin et al. 2005; Qiu et al.
2005). Moreover, it is usually the siRNA’s seed region that matches the sequence
within the off-targeted mRNA (Lin et al. 2005; Birmingham et al. 2006; Jackson
et al. 2006b). The rules of mRNA-target selection are too complex to be fully
determined by the 16,348 possible combinations for nucleotides in positions 2–8.
Evidently, nucleotides beyond the seed region – some of them part of the siRNA’s
guest strand – contribute to target selection by siRNAs. Consequently, the next
step is increasing the likelihood of the guide strand to be recognized by RISC

(RNA-Induced Silencing Complex). As an empirical rule, high specificity is
observed when there is relatively weak base pairing between the 50 portion of the
siRNA and the mRNA in question (Khvorova et al. 2003; Schwarz et al. 2003;
Reynolds et al. 2004). Applying this design strategy and others will result in a
stronger preference for the guide strand over the guest strand in determining which
mRNAs are silenced by the RISC machinery (Khvorova et al. 2003; Schwarz et al.
2003; Reynolds et al. 2004).
Chemical modification of small interfering RNAs adds a new dimension to
optimizing their utility. Various covalent modifications to siRNAs increase their
stability, target specificity, and the efficiency with which they silence the target
mRNA. Moreover, these chemical derivatives of oligoribonucleotides produce
lesser off-target effects and prove less toxic to the host cells. Their large size
(molecular weight of 13,000 or more), strongly negative charge, and hydrophilic
nature make siRNAs different from other small-molecule pharmaceuticals. As
noted earlier, siRNA is easily degraded by nucleases in serum. Chemical changes
to the sugars, backbone, or bases of siRNAs make them more like conventional
drugs (Bumcrot et al. 2006). Ribonucleic acids can be rendered resistant to exonu-
cleases by replacing the phosphodiester bond to the most 30 sugar with a phosphor-
othioate (P¼S) backbone linkage. RNA molecules can also be protected from
degradation by blocking the free hydroxyl on the ribose. When RNA is hydrolyzed,
the 20 OH acts as a nucleophile to attack the 30 hydroxyl and displace the phosphate
group. This reaction mechanism is not possible if the 20 oxygen is blocked with a
methyl, methoxyethyl, or fluoro group. For this reason, ribonucleic acids with 20 - O-
methyl (20 -O-Me), 20 -O-methoxyethyl (20 -O-MOE), and 20 -fluoro (20 -F) sugars all
resist degradation by nucleases. Other chemical analogs of RNA made from
nonstandard sugars (i.e., 40 -thioribose; the bridged ribose in locked nucleic acids)
are also more stable than RNA. Because of these chemical changes, the above
nucleic acid analogs exhibit significantly greater stability and in vitro potency than
siRNA without modification. In addition, they are less immunostimulatory than
species used in conventional siRNA (Hoshika et al. 2004; Allerson et al. 2005;
Bumcrot et al. 2006; Dande et al. 2006). Replacing the second sugar (or sugars at
other internal positions) in the guide strand with 20 O-methyl ribosyl causes less
severe off-target effects than those of conventional siRNA without losing its
efficiency at silencing the target mRNA (Czauderna et al. 2003; Jackson et al.
2006a). 20 O-Allylated small interfering RNAs and phosphorothioated species have
similar advantages (Amarzguioui et al. 2003).
In addition to optimizing siRNA design algorithms and chemical modifications,
cell lines, transfection methods, and mice strains all affect innate immune response
and off-target toxicity. The most prudent and robust strategy is to synthesize and
screen a substantial library of siRNA duplexes for multiple sites in HD mRNA
targets to identify the most promising one. Finally, the FDA has no established code
to regulate potential RNAi-based biological products. Technologies using naked
siRNA are likely to be classified as “antisense therapies”. Virus-borne shRNA is
apt to be classified as “gene therapy”. Because of several unfortunate events in past
years, the latter branch of experimental medicine is strictly regulated. For this reason,
siRNA-based therapies are more likely to obtain FDA approval. BevasiranibTM, the
first such drug, has already passed phase II and entered phase III clinical trials for
treating age-related macular degeneration (AMD).
Besides off-target effects, another big challenge in the development of RNAi-
based therapy for HD concerns delivery. In previous in vivo studies, viral and
nonviral application and local intrastriatal or intraventricular administration were
successfully used to silence HD in vivo. While virus-carried shRNA showed high
efficiency in transducing cells in the striatum, coupling siRNA with lipids or
cholesterol are other strategies for introducing HD RNAi into a living organism.
In one study, siRNA was coupled to liposomes (lipofectamine2000) to form amor-
phous lipoplex particles. Intraventricular injection with lipofectamine2000-siRNA
into day-2 (P2) postnatal mice countered both the generation of striatal nuclear
inclusion and brain atrophy. The molecular and anatomical changes were correlated
with a reduction in abnormal behavioral and increased lifespan (Wang et al. 2005).
Because this cationic lipid-based transfection reagent is toxic to primary neurons,
novel delivery strategies have been used. Lipophilic polypeptides such as stearylated
octaarginine (steary1-R8), MPG-based particles, and lipid-based artificial virus-like
particles (AVPs) are all peptide-based particles with potential applications in siRNA
therapy. They achieve transfection efficiency comparable to liposomal reagents, but
are less detrimental to primary neurons and embryonic stem cells (Futaki et al. 2001;
Simeoni et al. 2005; Tonges et al. 2006; Crombez et al. 2007). Another experimental
system for siRNA-based therapy has been explored in mice transduced with the
AAV-htt 100Q viral construct. Intrastriatal injection with a cholesterol-coupled HD
siRNA silenced the mutant transgene for 3 days and relieved neuropathological
symptoms for 1 week (DiFiglia et al. 2007). In addition, siRNA coupled to cell-
penetrating peptides provides a potential tool for cell type-specific siRNA targeting.
SiRNA coupled with the penetrating peptide via a disulfide bond is far more
effective than lipofectamine2000 in entering and silencing genes in primary neurons
(Davidson et al. 2004; Muratovska and Eccles 2004; Jankowski et al. 2009). siRNA
can also be conjugated to aptamers for cell-specific delivery (Chu et al. 2006).
Moreover, electroporation by a nucleofector enables siRNA to be introduced into
over 50% of cells in a population of mouse striatal primary neurons (Jin et al. 2005).
This technology also enables siRNA to reach 70% and shRNA to reach 59% of cells
within a culture of primary neurons (Dail et al. 2006; Hood et al. 2006; Seng et al.
2006; Yang et al. 2006). These new delivery methods will increase the silencing
efficiency of RNAi against HD and its related proteins to further elucidate its
physiological and pathological roles in HD and guide us in developing new drug
targets (Schwartz 2009). It would be interesting to compare the effects of siRNA
with virus-carried shRNA targeting the same HD sequence in the same transgenic
mouse model.
For local delivery, in collaboration with Dr. Don M. Gash of the University of
Kentucky, Alnylam Pharmaceuticals and Medtronic are developing an implantable
pump to infuse ALN-HTT (an RNAi-based drug against HD) into the brain. Using
an experimental catheter and the commercially available Medtronic SynchroMed II
pump, this technology permits the local delivery of siRNA to striatum and
subsequent transport to distant regions of the brain. Early studies have shown that
continuous infusion of an appropriate siRNA over 7 days caused an approximately
45% drop in the level of HD mRNA in the putamen. One big advantage of
implanted pumps is their ability to deliver drugs with much greater efficiency
than other techniques. However, the long-term efficacy of gene silencing is not
yet known. Another critical issue is the safety of the procedure. While no clinical
abnormalities were observed following 28 consecutive days of siRNA infusion into
a nonhuman primate, a comprehensive study must still be conducted. Ongoing
research on the promising technology of implanted pumps is likely to reach the
level of clinical trials in the not-so-distant future.
Whereas the most desirable way to get a drug into the CNS is by injection into
the bloodstream (or by ingestion), such systemic administration works only for
compounds that cross the blood–brain barrier. Unfortunately, this protective barrier
prevents RNAi molecules from entering the brain and spinal cord. However, the
situation is different in the liver. ApoB mRNA in liver has been silenced by 90% for
at least 11 days after systemic administration of one dose of siRNA encapsulated in
stable nucleic acid-lipid particles (SNALPS) (a bilayer consisting of cationic and
neutral lipids and coated with hydrophilic polyethylene) (Morrissey et al. 2005;
Zimmermann et al. 2006). Recently, investigators have developed a method to
circumvent this physiological barrier to the siRNA’s entrance into the CNS (Kumar
et al. 2007). In this technique, the siRNA is conjugated to a nona-D-arginine
polypeptide, which associates with a portion of a rabies virus glycoprotein
(RVG). Due to its affinity for the acetylcholine receptor (AChR), the molecular
complex binds to and is endocytosed by cells expressing that surface protein. The
internalized siRNA has been shown to silence the target gene (Kumar et al. 2007).
Moreover, the AChR is expressed in medium spiny projection (MSP) neurons of the
striatum and cerebral cortex (Weiner et al. 1990; Bernard et al. 1992; Ince et al.
1997). Theoretically, this technology should enable investigators to silence HD
mRNAs by targeting siRNA to the very cells with selective sensitivity to HD.
Practical considerations, i.e., the efficiency with which the molecular complex can
be delivered into the CNS and its toxicity once there, must still be determined.
Ultimately, it may be possible to target siRNAs to specific classes of cells by
coupling them to fragments of the appropriate viral glycoproteins. This technology
could be used to target other cell types besides MSNs involved in the pathological
mechanism of HD. For example, besides damage from mutant huntingtin’s toxicity
to MSNs, it may diminish the glia’s protective ability against excitotoxicity caused
by glutamate (Shin et al. 2005). Should the above technology be adapted to target
many cell types, it will prove a more effective tool with which to counter HD.
The rapidly developing technology of RNA interference has already proven
useful in experimental therapies for HD. Studies using RNAi in cellular and animal
systems have increased our understanding of the function of wild-type huntingtin.
In addition, they have already been used in preclinical trials in animals. Findings
from these investigations suggest that targeted gene silencing may be used in therapy
for what is now an untreatable disease. As a rule, difficult and time-consuming
investigations are necessary to develop conventional drugs. By comparison, siRNAs

can be produced easily and their sequence designed to efficiently and selectively
silence target genes. The US Patent and Trademark Office has already received
more than ten applications concerning therapeutics for HD based on AAV-shRNA
for allele-specific RNAi (To date, no patent has been issued). Sirna Therapeutics
(acquired by Merck & Co., Inc. in December of 2006) has been working with
Dr. Beverly Davison (University of Iowa) and Alnylam Pharmaceuticals with
Dr. Nein Aronin (University of Massachusetts) to develop RNAi-based treatments
for HD, which indicates the commercial potential of RNAi-based HD therapeutics.
Many technical hurdles remain: optimization of local and systemic delivery,
improvement of target specificity and efficacy, ensuring host safety, etc. We should
expect to see this rapidly developing technique become a major therapeutic modal-
ity for HD once these concerns are successfully addressed. Overall, the promise of
RNAi technology as a tool against HD is great, and this discipline certainly
deserves future research and development. We hope that patients who suffer from
this currently untreatable disease will someday benefit from novel RNAi therapies
that spring from the abundant enterprise and innovation in this new field.
References
Allerson CR, Sioufi N, Jarres R et al (2005) Fully 20 -modified oligonucleotide duplexes with
improved in vitro potency and stability compared to unmodified small interfering RNA. J Med
Chem 48:901–904
Almqvist E, Spence N, Nichol K et al (1995) Ancestral differences in the distribution of the delta
2642 glutamic acid polymorphism is associated with varying CAG repeat lengths on normal
chromosomes: insights into the genetic evolution of Huntington disease. Hum Mol Genet
4:207–214
Amarzguioui M, Holen T, Babaie E et al (2003) Tolerance for mutations and chemical modifica-
tions in a siRNA. Nucleic Acids Res 31:589–595
Ambrose CM, Duyao MP, Barnes G et al (1994) Structure and expression of the Huntington’s
disease gene: evidence against simple inactivation due to an expanded CAG repeat. Somat Cell
Mol Genet 20:27–38
Bernard V, Normand E, Bloch B (1992) Phenotypical characterization of the rat striatal neurons
expressing muscarinic receptor genes. J Neurosci 12:3591–3600
Birmingham A, Anderson EM, Reynolds A et al (2006) 30 UTR seed matches, but not overall
identity, are associated with RNAi off-targets. Nat Methods 3:199–204
Boudreau RL, Davidson BL (2006) RNAi therapy for neurodegenerative diseases. Curr Top Dev
Biol 75:73–92
Boudreau RL, McBride JL, Martins I et al (2009) Nonallele-specific silencing of mutant and wild-
type huntingtin demonstrates therapeutic efficacy in Huntington’s disease mice. Mol Ther
17:1053–1063
Bumcrot D, Manoharan M, Koteliansky V et al (2006) RNAi therapeutics: a potential new class of
pharmaceutical drugs. Nat Chem Biol 2:711–719
Burger C, Gorbatyuk OS, Velardo MJ et al (2004) Recombinant AAV viral vectors pseudotyped
with viral capsids from serotypes 1, 2, and 5 display differential efficiency and cell tropism
after delivery to different regions of the central nervous system. Mol Ther 10:302–317
Cattaneo E, Zuccato C, Tartari M (2005) Normal huntingtin function: an alternative approach to

Huntington’s disease. Nat Rev Neurosci 6:919–930
Caviston JP, Ross JL, Antony SM et al (2007) Huntingtin facilitates dynein/dynactin-mediated
vesicle transport. Proc Natl Acad Sci USA 104:10045–10050
Chai Y, Koppenhafer SL, Bonini NM et al (1999) Analysis of the role of heat shock protein (Hsp)
molecular chaperones in polyglutamine disease. J Neurosci 19:10338–10347
Chao JR, Parganas E, Boyd K et al (2008) Hax1-mediated processing of HtrA2 by Parl allows
survival of lymphocytes and neurons. Nature 452:98–102
Chu TC, Twu KY, Ellington AD et al (2006) Aptamer mediated siRNA delivery. Nucleic Acids
Res 34:e73
Crombez L, Charnet A, Morris MC et al (2007) A non-covalent peptide-based strategy for siRNA
delivery. Biochem Soc Trans 35:44–46
Czauderna F, Fechtner M, Dames S et al (2003) Structural variations and stabilising modifications
of synthetic siRNAs in mammalian cells. Nucleic Acids Res 31:2705–2716
Dail M, Richter M, Godement P et al (2006) Eph receptors inactivate R-Ras through different
mechanisms to achieve cell repulsion. J Cell Sci 119:1244–1254
Dande P, Prakash TP, Sioufi N et al (2006) Improving RNA interference in mammalian cells by
40 -thio-modified small interfering RNA (siRNA): effect on siRNA activity and nuclease
stability when used in combination with 20 -O-alkyl modifications. J Med Chem 49:1624–1634
Davidson BL, Stein CS, Heth JA et al (2000) Recombinant adeno-associated virus type 2, 4, and 5
vectors: transduction of variant cell types and regions in the mammalian central nervous
system. Proc Natl Acad Sci USA 97:3428–3432
Davidson TJ, Harel S, Arboleda VA et al (2004) Highly efficient small interfering RNA delivery to
primary mammalian neurons induces microRNA-like effects before mRNA degradation.
J Neurosci 24:10040–10046
de Almeida LP, Ross CA, Zala D et al (2002) Lentiviral-mediated delivery of mutant huntingtin in
the striatum of rats induces a selective neuropathology modulated by polyglutamine repeat
size, huntingtin expression levels, and protein length. J Neurosci 22:3473–3483
del Toro D, Alberch J, Lazaro-Dieguez F et al (2009) Mutant huntingtin impairs post-Golgi
trafficking to lysosomes by delocalizing optineurin/Rab8 complex from the Golgi apparatus.
Mol Biol Cell 20:1478–1492
Denovan-Wright EM, Davidson BL (2006) RNAi: a potential therapy for the dominantly inherited
nucleotide repeat diseases. Gene Ther 13:525–531
Denovan-Wright EM, Rodriguez-Lebron E, Lewin AS et al (2008) Unexpected off-targeting
effects of anti-huntingtin ribozymes and siRNA in vivo. Neurobiol Dis 29:446–455
DiFiglia M, Sapp E, Chase K et al (1995) Huntingtin is a cytoplasmic protein associated with
vesicles in human and rat brain neurons. Neuron 14:1075–1081
DiFiglia M, Sena-Esteves M, Chase K et al (2007) Therapeutic silencing of mutant huntingtin with
siRNA attenuates striatal and cortical neuropathology and behavioral deficits. Proc Natl Acad
Sci USA 104:17204–17209
Doench JG, Sharp PA (2004) Specificity of microRNA target selection in translational repression.
Genes Dev 18:504–511
Doumanis J, Wada K, Kino Y et al (2009) RNAi screening in Drosophila cells identifies new
modifiers of mutant huntingtin aggregation. PLoS One 4:e7275
Dragatsis I, Levine MS, Zeitlin S (2000) Inactivation of Hdh in the brain and testis results in
progressive neurodegeneration and sterility in mice. Nat Genet 26:300–306
Drouet V, Perrin V, Hassig R et al (2009) Sustained effects of nonallele-specific Huntingtin
silencing. Ann Neurol 65:276–285
Duyao MP, Auerbach AB, Ryan A et al (1995) Inactivation of the mouse Huntington’s disease
gene homolog Hdh. Science 269:407–410
Franich NR, Fitzsimons HL, Fong DM et al (2008) AAV vector-mediated RNAi of mutant
huntingtin expression is neuroprotective in a novel genetic rat model of Huntington’s disease.
Mol Ther 16:947–956
Futaki S, Ohashi W, Suzuki T et al (2001) Stearylated arginine-rich peptides: a new class of

transfection systems. Bioconjug Chem 12:1005–1011
Gacy AM, Goellner G, Juranic N et al (1995) Trinucleotide repeats that expand in human disease
form hairpin structures in vitro. Cell 81:533–540
Goehler H, Lalowski M, Stelzl U et al (2004) A protein interaction network links GIT1, an
enhancer of huntingtin aggregation, to Huntington’s disease. Mol Cell 15:853–865
Gunawardena S, Her LS, Brusch RG et al (2003) Disruption of axonal transport by loss of
huntingtin or expression of pathogenic polyQ proteins in Drosophila. Neuron 40:25–40
11:599–606
Harper SQ (2009) Progress and challenges in RNA interference therapy for Huntington disease.
Arch Neurol 66:933–938
Harper SQ, Staber PD, He X et al (2005) RNA interference improves motor and neuropathological
abnormalities in a Huntington’s disease mouse model. Proc Natl Acad Sci USA 102:
5820–5825
Hodgson JG, Agopyan N, Gutekunst CA et al (1999) A YAC mouse model for Huntington’s
disease with full-length mutant huntingtin, cytoplasmic toxicity, and selective striatal neuro-
degeneration. Neuron 23:181–192
Hood C, Cunningham AL, Slobedman B et al (2006) Varicella-zoster virus ORF63 inhibits
apoptosis of primary human neurons. J Virol 80:1025–1031
Hoshika S, Minakawa N, Matsuda A (2004) Synthesis and physical and physiological properties of
40 -thioRNA: application to post-modification of RNA aptamer toward NF-kappaB. Nucleic
Acids Res 32:3815–3825
Hu J, Matsui M, Corey DR (2009) Allele-selective inhibition of mutant huntingtin by peptide
nucleic acid-peptide conjugates, locked nucleic acid, and small interfering RNA. Ann NY
Acad Sci 1175:24–31
Huang B, Schiefer J, Sass C et al (2007) High-capacity adenoviral vector-mediated reduction of
huntingtin aggregate load in vitro and in vivo. Hum Gene Ther 18:303–311
Inagaki R, Tagawa K, Qi ML et al (2008) Omi/HtrA2 is relevant to the selective vulnerability of
striatal neurons in Huntington’s disease. Eur J Neurosci 28:30–40
Ince E, Ciliax BJ, Levey AI (1997) Differential expression of D1 and D2 dopamine and m4
muscarinic acetylcholine receptor proteins in identified striatonigral neurons. Synapse
27:357–366
Jackson AL, Burchard J, Leake D et al (2006a) Position-specific chemical modification of siRNAs
reduces “off-target” transcript silencing. RNA 12:1197–1205
Jackson AL, Burchard J, Schelter J et al (2006b) Widespread siRNA “off-target” transcript
silencing mediated by seed region sequence complementarity. RNA 12:1179–1187
Jankowski MP, McIlwrath SL, Jing X et al (2009) Sox11 transcription factor modulates peripheral
nerve regeneration in adult mice. Brain Res 1256:43–54
Jin K, LaFevre-Bernt M, Sun Y et al (2005) FGF-2 promotes neurogenesis and neuroprotection
and prolongs survival in a transgenic mouse model of Huntington’s disease. Proc Natl Acad Sci
USA 102:18189–18194
bias. Cell 115:209–216
Kotliarova S, Jana NR, Sakamoto N et al (2005) Decreased expression of hypothalamic neuropep-
tides in Huntington disease transgenic mice with expanded polyglutamine-EGFP fluorescent
aggregates. J Neurochem 93:641–653
Kuida K, Zheng TS, Na S et al (1996) Decreased apoptosis in the brain and premature lethality in
CPP32-deficient mice. Nature 384:368–372
Kumar P, Wu H, McBride JL et al (2007) Transvascular delivery of small interfering RNA to the

central nervous system. Nature 448:39–43
Leavitt BR, van Raamsdonk JM, Shehadeh J et al (2006) Wild-type huntingtin protects neurons
from excitotoxicity. J Neurochem 96:1121–1129
Lin X, Ruan X, Anderson MG et al (2005) siRNA-mediated off-target gene silencing triggered by
a 7 nt complementation. Nucleic Acids Res 33:4527–4535
Liu W, Kennington LA, Rosas HD et al (2008) Linking SNPs to CAG repeat length in Huntington’s
disease patients. Nat Methods 5:951–953
Lombardi MS, Jaspers L, Spronkmans C et al (2009) A majority of Huntington’s disease patients
may be treatable by individualized allele-specific RNA interference. Exp Neurol 217:312–319
Machida Y, Okada T, Kurosawa M et al (2006) rAAV-mediated shRNA ameliorated neuropathology
in Huntington disease model mouse. Biochem Biophys Res Commun 343:190–197
Mangiarini L, Sathasivam K, Seller M et al (1996) Exon 1 of the HD gene with an expanded CAG
repeat is sufficient to cause a progressive neurological phenotype in transgenic mice. Cell
87:493–506
USA 105:5868–5873
Menalled LB, Chesselet MF (2002) Mouse models of Huntington’s disease. Trends Pharmacol Sci
23:32–39
Morrissey DV, Lockridge JA, Shaw L et al (2005) Potent and persistent in vivo anti-HBV activity
of chemically modified siRNAs. Nat Biotechnol 23:1002–1007
Muratovska A, Eccles MR (2004) Conjugate for efficient delivery of short interfering RNA
(siRNA) into mammalian cells. FEBS Lett 558:63–68
Nasir J, Floresco SB, O’Kusky JR et al (1995) Targeted disruption of the Huntington’s disease
gene results in embryonic lethality and behavioral and morphological changes in heterozygotes.
Cell 81:811–823
Novelletto A, Persichetti F, Sabbadini G et al (1994) Polymorphism analysis of the huntingtin gene
in Italian families affected with Huntington disease. Hum Mol Genet 3:1129–1132
Omi K, Hachiya NS, Tokunaga K et al (2005) siRNA-mediated inhibition of endogenous
Huntington disease gene expression induces an aberrant configuration of the ER network
in vitro. Biochem Biophys Res Commun 338:1229–1235
Palfi S, Brouillet E, Jarraya B et al (2007) Expression of mutated huntingtin fragment in the
putamen is sufficient to produce abnormal movement in non-human primates. Mol Ther
15:1444–1451
Paterna JC, Feldon J, Bueler H (2004) Transduction profiles of recombinant adeno-associated
virus vectors derived from serotypes 2 and 5 in the nigrostriatal system of rats. J Virol
78:6808–6817
Qin ZH, Wang Y, Sapp E et al (2004) Huntingtin bodies sequester vesicle-associated proteins by a
polyproline-dependent interaction. J Neurosci 24:269–281
Qiu S, Adema CM, Lane T (2005) A computational study of off-target effects of RNA interference.
Nucleic Acids Res 33:1834–1847
Reynolds A, Leake D, Boese Q et al (2004) Rational siRNA design for RNA interference. Nat
Biotechnol 22:326–330
Rodriguez-Lebron E, Denovan-Wright EM, Nash K et al (2005) Intrastriatal rAAV-mediated
delivery of anti-huntingtin shRNAs induces partial reversal of disease progression in R6/1
Huntington’s disease transgenic mice. Mol Ther 12:618–633
Rubinsztein DC, Leggo J, Goodburn S et al (1995) Haplotype analysis of the delta 2642 and
(CAG)n polymorphisms in the Huntington’s disease (HD) gene provides an explanation for an
apparent ‘founder’ HD haplotype. Hum Mol Genet 4:203–206
Schilling G, Becher MW, Sharp AH et al (1999) Intranuclear inclusions and neuritic aggregates in
transgenic mice expressing a mutant N-terminal fragment of huntingtin. Hum Mol Genet
8:397–407
Schwartz EI (2009) Potential application of RNAi for understanding and therapy of neurodegenerative
diseases. Front Biosci 14:297–320
Schwarz DS, Ding H, Kennington L et al (2006) Designing siRNA that distinguish between genes
that differ by a single nucleotide. PLoS Genet 2:e140
Schwarz DS, Hutvagner G, Du T et al (2003) Asymmetry in the assembly of the RNAi enzyme
Seng S, Avraham HK, Jiang S et al (2006) KLHL1/MRP2 mediates neurite outgrowth in a
glycogen synthase kinase 3beta-dependent manner. Mol Cell Biol 26:8371–8384
Senut MC, Suhr ST, Kaspar B et al (2000) Intraneuronal aggregate formation and cell death after
viral expression of expanded polyglutamine tracts in the adult rat brain. J Neurosci 20:219–229
Shin JY, Fang ZH, Yu ZX et al (2005) Expression of mutant huntingtin in glial cells contributes to
neuronal excitotoxicity. J Cell Biol 171:1001–1012
Simeoni F, Morris MC, Heitz F et al (2005) Peptide-based strategy for siRNA delivery into
mammalian cells. Methods Mol Biol 309:251–260
Slow EJ, van Raamsdonk J, Rogers D et al (2003) Selective striatal neuronal loss in a YAC128
mouse model of Huntington disease. Hum Mol Genet 12:1555–1567
Sobczak K, de Mezer M, Michlewski G et al (2003) RNA structure of trinucleotide repeats
associated with human neurological diseases. Nucleic Acids Res 31:5469–5482
Steffan JS, Kazantsev A, Spasic-Boskovic O et al (2000) The Huntington’s disease protein
interacts with p53 and CREB-binding protein and represses transcription. Proc Natl Acad Sci
USA 97:6763–6768
Strehlow AN, Li JZ, Myers RM (2007) Wild-type huntingtin participates in protein trafficking
between the Golgi and the extracellular space. Hum Mol Genet 16:391–409
Tagawa K, Marubuchi S, Qi ML et al (2007) The induction levels of heat shock protein 70
differentiate the vulnerabilities to mutant huntingtin among neuronal subtypes. J Neurosci
27:868–880
Taymans JM, Vandenberghe LH, Haute CV et al (2007) Comparative analysis of adeno-associated
viral vector serotypes 1, 2, 5, 7, and 8 in mouse brain. Hum Gene Ther 18:195–206
The Huntington’s Disease Collaborative Research Group (1993) A novel gene containing a
trinucleotide repeat that is expanded and unstable on Huntington’s disease chromosomes.
The Huntington’s Disease Collaborative Research Group. Cell 72:971–983
Tonges L, Lingor P, Egle R et al (2006) Stearylated octaarginine and artificial virus-like particles
for transfection of siRNA into primary rat neurons. RNA 12:1431–1438
van Bilsen PH, Jaspers L, Lombardi MS et al (2008) Identification and allele-specific silencing of
the mutant huntingtin allele in Huntington’s disease patient-derived fibroblasts. Hum Gene
Ther 19:710–719
Velier J, Kim M, Schwarz C et al (1998) Wild-type and mutant huntingtins function in vesicle
trafficking in the secretory and endocytic pathways. Exp Neurol 152:34–40
Vuillaume I, Vermersch P, Destee A et al (1998) Genetic polymorphisms adjacent to the CAG
repeat influence clinical features at onset in Huntington’s disease. J Neurol Neurosurg
Psychiatry 64:758–762
Wacker JL, Zareie MH, Fong H et al (2004) Hsp70 and Hsp40 attenuate formation of spherical and
annular polyglutamine oligomers by partitioning monomer. Nat Struct Mol Biol 11:1215–1222
Waelter S, Scherzinger E, Hasenbank R et al (2001) The huntingtin interacting protein HIP1 is a
clathrin and alpha-adaptin-binding protein involved in receptor-mediated endocytosis. Hum
Mol Genet 10:1807–1817
Wang YL, Liu W, Wada E et al (2005) Clinico-pathological rescue of a model mouse of
Huntington’s disease by siRNA. Neurosci Res 53:241–249
Warby SC, Montpetit A, Hayden AR et al (2009) CAG expansion in the Huntington disease gene is
associated with a specific and targetable predisposing haplogroup. Am J Hum Genet
84:351–366
Warrick JM, Chan HY, Gray-Board GL et al (1999) Suppression of polyglutamine-mediated
neurodegeneration in Drosophila by the molecular chaperone HSP70. Nat Genet 23:425–428
Weiner DM, Levey AI, Brann MR (1990) Expression of muscarinic acetylcholine and dopamine
receptor mRNAs in rat basal ganglia. Proc Natl Acad Sci USA 87:7050–7054
Woda JM, Calzonetti T, Hilditch-Maquire P et al (2005) Inactivation of the Huntington’s disease
gene (Hdh) impairs anterior streak formation and early patterning of the mouse embryo. BMC
Dev Biol 5:17
Yamamoto A, Lucas JJ, Hen R (2000) Reversal of neuropathology and motor dysfunction in a
conditional model of Huntington’s disease. Cell 101:57–66
Yang YC, Lin CH, Lee EH (2006) Serum- and glucocorticoid-inducible kinase 1 (SGK1) increases
neurite formation through microtubule depolymerization by SGK1 and by SGK1 phosphory-
lation of tau. Mol Cell Biol 26:8357–8370
Zeitlin S, Liu JP, Chapman DL et al (1995) Increased apoptosis and early embryonic lethality in
mice nullizygous for the Huntington’s disease gene homologue. Nat Genet 11:155–163
Zhang H, Das S, Li QZ et al (2008) Elucidating a normal function of huntingtin by functional and
microarray analysis of huntingtin-null mouse embryonic fibroblasts. BMC Neurosci 9:38
Zhang Y, Engelman J, Friedlander RM (2009) Allele-specific silencing of mutant Huntington’s
disease gene. J Neurochem 108:82–90
Zhang Y, Leavitt BR, van Raamsdonk JM et al (2006) Huntingtin inhibits caspase-3 activation.
Embo J 25:5896–5906
Zhou H, Li SH, Li XJ (2001) Chaperone suppression of cellular toxicity of huntingtin is independent
of polyglutamine aggregation. J Biol Chem 276:48417–48424
Zimmermann TS, Lee AC, Akinc A et al (2006) RNAi-mediated gene silencing in non-human
Zuccato C, Ciammola A, Rigamonti D et al (2001) Loss of huntingtin-mediated BDNF gene
transcription in Huntington’s disease. Science 293:493–498
Application of Dicer-Substrate siRNA
in Pain Research
Philippe Sarret, Louis Doré-Savard, Pascal Tétreault, Valérie Bégin-Lavallée,

and Nicolas Beaudet
Contents
1 RNAi in Pain Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
2 Potential Applications of RNA Interference for Pain Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . 163
2.1 Ion Channels as Therapeutic Targets for Pain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
2.2 G-Protein-Coupled Receptors as Pain Targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
3 Advantages of DsiRNA Over Conventional siRNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
3.1 Inherent Character of DsiRNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
3.2 Molecular Mechanism of Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
4 DsiRNA for Efficient Silencing In Vitro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
4.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
4.2 Validation of Knockdown Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
4.3 Targets In Vitro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
5 DsiRNA In Vivo: One Step Forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
5.1 Working Evidence of DsiRNA In Vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
5.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
6 Conclusion and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Abstract Chronic pain is a pathology affecting a large proportion of the worldwide

population. The socio-economical stakes involved in its treatment are huge. Clini-
cal management of chronic pain relies mainly on opioids and nonsteroidal anti-
inflammatory drugs. However, these pharmacological treatments remain inefficient
for several patients coping with chronic pain conditions, such as neuropathic or
cancer related pain. Genetic approaches are reported as an innovative solution to
circumvent pharmacological issues and also to uncover pain-related mechanisms or
P. Sarret (*), L. Doré-Savard, P. Tétreault, V. Bégin-Lavallée, and N. Beaudet

Department of Physiology and Biophysics, Faculty of Medicine and Health Sciences, Université
de Sherbrooke, 3001, 12Ème Avenue Nord, Sherbrooke, Québec J1H 5N4, Canada
Centre des Neurosciences de Sherbrooke, Université de Sherbrooke, 3001, 12Ème Avenue Nord,
Sherbrooke, Québec, Canada J1H 5N4
e-mail: philippe.sarret@usherbrooke.ca
162 P. Sarret et al.
identify new therapeutic targets. Indeed, RNA interference (RNAi) technology

represents a straightforward and efficient tool for genetic therapy. The present chapter
describes the major recent advances realized with RNAi in pain research, with an
emphasis on pain-related ion channels and GPCR. Furthermore, we address the
mechanism of action of a novel RNAi tool, namely dicer-substrate siRNA (DsiRNA).
This cutting edge technology facilitated recent progress in pain research because of its
high efficiency allowing the use of lower doses in the central nervous system, a
major downfall encountered with conventional siRNAs. Finally, in this chapter, we
review several concepts leading toward the successful use of DsiRNAs in vitro and
in vivo in the context of pain research. In brief, the precepts of RNAi gene therapy in
pain research should be more accessible to any scientist by the end of the chapter.
Keywords Chronic pain Dicer-subtrate siRNA GPCR knockdown Ion Channel

knockdown RNA interference Silencing methodology
1 RNAi in Pain Research
Pain is an unpleasant sensory and emotional experience but indispensable sensation

for healthy survival. Chronic pain, on the other hand, serves no protective
biological function, can persist for years after the initial injury, and has a major
impact on individual quality of life. Indeed, inadequate treatment of persistent pain
is highly associated with the development of comorbid conditions, including
chronic anxiety, depression, sleeplessness, and impairment of social interaction.
Pain is a major clinical problem worldwide. Rather than being the symptom of a
disease, chronic pain is a self-perpetuating disease process that is not the simple
result of continuous activation of the nociceptive pathways. Physiologic and mor-
phologic changes occur in peripheral (nerves, roots) and/or central (spinal cord,
supraspinal cerebral structures) brain regions involved in the pain transmission.
Different in any chronic pain syndrome, these neuroplastic alterations observed
during pain chronicisation result in receptor and ion-channel reorganization, neu-
rotransmitter synthesis and release pattern modification, variations in transcription
and protein translation in both neurons and glial cells, changes in the electrophysi-
ological properties of neuronal cells, as well as in structural reorganization.
Chronic pain is now recognized as a substantial social and economic burden.
Despite significant advances in our understanding of the pathophysiology of pain,
the management of chronic pain remains a clinical challenge. Chronic painful
states, such as neuropathic pain, are particularly resistant to opioids or nonsteroidal
anti-inflammatory drugs (NSAIDs). Besides the lack of efficacy, prolonged admini-
stration of these analgesic substances results for some patients in somnolence,
confusion, constipation, nausea, renal toxicity, respiratory impairment, and ulti-
mately cardiovascular side effects. In addition to these drawbacks, continuous
infusion of opioids leads to tolerance and the requirement of escalating doses of
Application of Dicer-Substrate siRNA in Pain Research 163
medication to maintain the same level of pain relief. Although a large number of
pharmacological treatments are currently available, there is still a pressing need for
the development of new alternative therapeutic approaches to alleviate pain and to
increase patient comfort. With this perspective, RNA interference (RNAi) repre-
sents a major tool for discovery and validation of targets implicated in pain
disorders and may be useful therapeutically to treat some intractable conditions
where conventional therapy is ineffective or inexistent. In this chapter, we first
provide a brief overview of the use of RNAi in pain research, with emphasis on ion
channels and G-protein-coupled receptors (GPCRs) as potential targets for thera-
peutic intervention. We then discuss the advantages of designing longer exogenous
siRNA to serve as Dicer substrates (DsiRNAs). Finally, we focus on the potency of
these new synthetic 27-mer RNA duplexes to efficiently knockdown pain targets in
both in vitro and in vivo studies.
2 Potential Applications of RNA Interference for

Pain Treatment
In the last 5 years, siRNA strategies have been widely used to study pain mechani-
sms in vivo, providing a straightforward approach to validate new pain targets. Due
to space limitations, however, only a limited number of reports can be described
here. For further details, the reader can refer to recent reviews (Ganju and Hall
2004; Kurreck 2004; Sah 2006; Jain 2008).
2.1 Ion Channels as Therapeutic Targets for Pain
Dorsal root ganglia (DRG) of the spinal cord represent an attractive site for gene
therapy designed to treat chronic pain. The first successful in vivo proof of concept
of the use of siRNA in a pain model was conducted through the intrathecal infusion
of siRNA targeting the cation-channel subunit P2X3 (Dorn et al. 2004). This P2X
purinergic receptor is highly expressed in nociceptive sensory neurons and plays a
crucial role in ATP-induced inflammatory and neuropathic pain. It has been
demonstrated that 21-nt siRNA sequences can specifically knockdown the rat
P2X3 receptor in vitro and in vivo (Hemmings-Mieszczak et al. 2003; Dorn et al.
2004). Continuously administered via osmotic mini-pump into neuropathic rats at a
high dose of 0.4 mg/day for up to 6 days, intrathecal naked P2X3 siRNA reduced
both mechanical allodynia and thermal hyperalgesia induced by partial sciatic
nerve ligation. In line with these results, siRNA-treated rats showed diminished
pain responses following injection of the P2X3 receptor agonist, ab-methylene ATP
(stable analog of ATP). At a molecular level, siRNA administration resulted in
down-regulation of P2X3 mRNA and protein pools in DRG and spinal cord.
Members of the transient receptor potential (TRP) family of ion channels also
represent attractive targets for pain therapy. TRPV1 is a cation channel that is
predominantly expressed by small- to medium-diameter primary sensory neurons

and is activated by vanilloid compounds such as capsaicin, protons, or thermal
stimuli greater than 43 C. Studies by Christoph and coworkers contributed to the
increase of knowledge about the function of TRPV1 in the development and
maintenance of pain (Christoph et al. 2006, 2008). They demonstrated that intrathe-
cal bolus administration of two different TRPV1-specific siRNA reduced cold
allodynia of mononeuropathic rats by more than 50% over a time period of approxi-
mately 5 days (Christoph et al. 2006). In a second set of experiments, they also found
that spinal administration of siRNA against TRPV1 diminished spontaneous vis-
ceral pain behaviors induced by TRPV1 agonist capsaicin application. The analgesic
effects reached by silencing TRPV1 by RNAi were similar to the reduction of pain
sensitivity observed with the TRPV1 antagonist BCTC. More recently, they also
provided further evidence for the relevance of TRPV1 in neuropathic pain. They
decided to achieve long-term inhibition of the TRPV1 gene by generating transgenic
mice expressing self-complementary short hairpin RNA (shRNAs) (Christoph et al.
2008). ShRNAs are processed by the RNase Dicer to give siRNA-type molecules.
Interestingly, continuously expressed shRNA blocked the development of mechani-
cal allodynia and hypersensitivity following spinal nerve injury. Furthermore,
TRPV1 shRNA transgenic mice were less sensitive to intraplantar injection of
capsaicin as compared with wild-type animals. Likewise, sensitivity toward noxious
heat was significantly reduced in animals permanently inhibiting TRPV1 expression
by means of RNAi. TRPV1 also plays an essential role in the establishment of
inflammatory thermal hyperalgesia after tissue injury. Forepaw inflammation
induced by intraplantar injection of complete Freund’s adjuvant (CFA) significantly
increases sensitivity to noxious thermal stimuli. Kasama et al. (2007) demonstrated
that paratracheal injection of TRPV1 short interfering RNA blocked TRPV1
up-regulation in cervical DRG and prevented inflammatory thermal hyperalgesia
after forepaw injuries. The TRPV4 mechanosensor calcium channel is also an
essential mediator of spinal nociceptive transmission in DRG. Accordingly, inhibi-
tion of TRPV4 activity by siRNA treatment blocked the release of the pronocicep-
tive neuropeptide substance P in primary cultures of DRG neurons (Liu et al. 2009).
In vivo, siRNAs were used to investigate the implication of TRPV4 in visceral
nociception (Cenac et al. 2008). Intravertebral injections of TRPV4 siRNA were
effective in reducing basal visceral nociception. Similarly, following TRPV4 siRNA
treatments, abdominal muscle contractions were decreased in response to colorectal
distension in mice receiving intracolonically either the TRPV4 agonist 4aPDD or the
protease-activated receptor 2 (PAR2)-activating peptide (Cenac et al. 2008). These
results suggest that TRPV4 could be a new potential analgesic target in patients
suffering from the irritable bowel syndrome.
Acid-sensing ion channels (ASICs) represent another family of cationic channels
involved in pain transmission. Activated by protons, ASICs have been proposed
to sense extracellular acidification occurring in pathological conditions such as
inflammation, ischemia, fractures, lesions as well as in postoperative states. RNA
interference has recently allowed to evaluate the contribution of ASIC3 to the
development of CFA-induced primary inflammatory pain (Deval et al. 2008).
Intrathecal delivery of an siRNA targeting the ASIC3 channel exerted a potent

analgesic effects against cutaneous inflammation-induced hyperalgesia in rats.
These data are consistent with those obtained by a classical pharmacological
approach; concomitant administration of the ASIC3 blocker APETx2 with CFA
inhibiting pain settling (Deval et al. 2008).
There is considerable evidence that pain associated with peripheral tissue or
nerve injury involves N-methyl-D-aspartate (NMDA) receptor activation. Consistent
with this, NMDAR antagonists have been shown to effectively alleviate pain-related
behaviors in animal models as well as in clinical situations. However, NMDA
receptors are important for normal brain functions, and the use of NMDA receptor
antagonists can often be limited by serious side effects, such as memory impairment,
motor incoordination, ataxia, and hallucinations. Newly developed siRNAs target-
ing different NMDA receptor subunits may therefore offer several advantages over
current approaches. NMDA-R2B receptor subunit protein (NR2B) is localized in the
superficial dorsal horn of the spinal cord. Thus, Tan et al. (2005) investigated the
effect of intrathecal injection of 5 mg of siRNA–NR2B complexed with the cationic
polymer PEI, for the modulation of pain. They found that the formalin-induced
flinching response was decreased by approximately 50% in the tonic phase. Con-
comitantly, the expression of NR2B mRNA and its associated protein were reduced
until day 14 after siRNA–NR2B injection. In the same line, spinal delivery of an
NR2B shRNA expression plasmid resulted in a dose-dependent blockade of hyper-
algesia induced by activation of the metabotropic glutamate receptor mGluR1/5 in
mGluR1/5 agonist-treated mice (Gabra et al. 2007). Additional sites of intervention
with siRNA may lie in supraspinal structures, although the regulation of pain
perception by higher centers of the brain has not been explored extensively using
gene silencing strategies. Effective administration of siRNA to this site in humans is
likely to be technically challenging and less attractive, especially from a therapeutic
point of view. Nevertheless, Fan et al. (2009) recently demonstrated that specific
down-regulation of the NR2B subunit by microinjection of siRNA into the anterior
cingulated cortex reduced the visceromotor pain responses to colorectal distension
in viscerally hypersensitive rats. Functional NMDA receptors are tetramers com-
posed of 2 NR1 and 2 NR2 subunits. Therefore, specific silencing of the NR1 subunit
may also disrupt the NMDA receptor function. Indeed, recombinant adeno-
associated virus (rAAV) vectors expressing an active siRNA targeting the NR1
subunit attenuated both formalin-induced pain behaviors in adult rats (Garraway
et al. 2009) and also prevented the mechanical allodynia measured 48 h after CFA
injection into the paw (Garraway et al. 2007). The authors also reported that the
knockdown of NR1 expression in spinal cord dorsal horn persisted for at least 6
months after a single administration of the rAAV vectors. Long-lasting and stable
expression can be especially important when exploring the therapeutic opportunities
for the use of RNAi to silence the expression of genes contributing to persistent
conditions such as chronic pain.
Voltage-gated sodium channels have long been recognized as being critically
important for driving neuronal excitability in both central and peripheral nervous
system. Nine sodium channel isoforms (NaV1.1–NaV1.9) have been identified so far.
From then on, the development of small molecules to specifically block each sodium
channel subtype has presented significant challenges for conventional medicinal
chemistry. As a therapeutic class, sodium channel blockers such as lidocaine,
amitriptyline, and lamotrigine have been widely used to treat disorders involving
neuronal hyperexcitability, including management of pain. However, none of the
currently used sodium channel blockers is capable to distinguish between the various
subtypes that have been identified to date. This lack of selectivity is thought to
contribute to side effects of these drugs such as motor dysfunction, therefore
compromising seriously the use of these medications in a clinical pain setting. A
critical role of NaV1.8 in mediating pathologic pain has been suggested by the
observation that tetrodotoxin-resistant sodium channel a-subunit knockout mice
showed strong analgesia to noxious mechanical stimuli and delayed development
of inflammatory hyperalgesia (Akopian et al. 1999). The restricted expression of
NaV1.8 to the peripheral sensory neurons suggests that the blockade of this channel
has therapeutic potential in various pain paradigms and may improve the compliance
in patients actually treated with existing sodium channel blockers. In 2006, Alnylam
Pharmaceutical Inc. and collaborators were the first to release preclinical data
demonstrating that intrathecal injection of siRNA targeting NaV1.8 provided nearly
complete pain relief in an animal model of chronic neuropathic pain (Sah et al. 2006).
Since then, Dong et al. (2007) demonstrated that NaV1.8 siRNA delivered to the
lumbar DRG via an indwelling epidural cannula caused a significant reduction of
NaV1.8 mRNA expression in L4–L5 DRG neurons and consequently reversed
mechanical allodynia in the rat chronic constriction nerve-injury model.
Inward rectifying K+ channels and voltage-gated Ca2+ channels, which regulate
either cellular excitability or synaptic transmission, also play an important role in
the detection and transmission of nociceptive stimuli in primary sensory neurons.
The Kir4.1 potassium channel subunit and the N-type calcium channel Cav2.2 were
validated as pain targets in vivo by RNAi. Kir4.1 expressed in the trigeminal
ganglion is reduced following chronic constriction injury of the infraorbital nerve
and specific silencing of Kir4.1 by RNAi produces a facial pain-like behavior in
freely moving rats (Vit et al. 2008). Two Cav2.2 N-type channel isoforms (e37a and
e37b) can be generated by alternative splicing. Altier et al. (2007) demonstrated by
selective down-regulation of each Cav2.2 isoform by RNAi that channels contain-
ing exon 37a are specifically required for developing thermal and mechanical
hyperalgesia during inflammatory and neuropathic pain states.
2.2 G-Protein-Coupled Receptors as Pain Targets
GPCRs represent the largest and most diverse family of cell surface proteins that act
as sensors for extracellular stimuli, including hormones, neuropeptides, neurotrans-
mitters, or sensory stimuli such as light and smell. Despite their molecular and
function diversity, all GPCRs share a similar structure, which consists of seven
transmembrane domains linked by alternating intracellular and extracellular loops.
Extracellular domains, which vary among the different classes of GPCR, contribute
to ligand recognition and binding, whereas coupling to G-proteins is determined

mainly by interactions with intracellular regions. GPCRs are widely distributed in
the peripheral and central nervous system and are actually considered as the best
family of drug targets. They account for 40–50% of all pharmaceutical prescrip-
tions on the market, representing more than USD 30 billion in sales annually. The
potential of this family remains tremendous as currently marketed GPCR drugs
only target one-third of the “druggable” GPCRs. There is a broad consensus that
GPCRs will remain a hub for drug development activities and clearly one of the
most important therapeutic targets in pain medicine. To tackle this challenge, there
is a need for new technologies to address orphan and peptidic GPCRs that are often
considered unnatainable. In this context, siRNA candidates may represent a new
class of therapeutics to target GPCRs.
The first GPCR involved in pain modulation validated by RNA interference is the
delta opioid receptor (DOR) (Luo et al. 2005). An siRNA directed against DOR
mixed with the transfection reagent i-FectTM was delivered via an implanted intra-
thecal catheter to the lumbar spinal cord of rats. Daily boluses of 2 mg for 3 days
reduced the levels of DOR mRNA by 38% in DRG and 62% in the spinal cord dorsal
horn. This decrease in DOR expression was paralleled by a reduction in [3H]
naltrindole (DOR agonist) binding to lumbar spinal cord as well as in DOR immu-
noreactivity. Silencing was transient and returned to normal after 72 h after the last
siRNA injection. Inhibition of DOR function resulted in the blockade of antinocicep-
tion evoked by the DOR selective agonist [D-Ala2, Glu4] deltorphin II, as demon-
strated by the reduction in the response latency in the radiant heat paw withdrawal
test. Furthermore, RNAi-mediated silencing of DOR did not alter signaling through
the mu opioid receptor, a closely related receptor (Luo et al. 2005).
Subsequently, several siRNA targeting nonopioidergic GPCRs have been vali-
dated for pain (Lin et al. 2006; Ndong et al. 2009; Zhang et al. 2009). Indeed, the
role of the neurotensin receptor family (NTS1 and NTS2) in central regulation of
pain perception is covered in one of the following section. NTS1 and NTS2 are two
GPCR receptors presenting an important distribution pattern in DRG and spinal
cord neurons (Sarret et al. 2005; Roussy et al. 2008, 2009). Both receptors were
shown to induce an opioid-independent type analgesia at the spinal and supraspinal
levels (Dobner 2006). To this day, there is no selective pharmacological antagonist
available to inhibit NTS2 signaling to study its in-depth function and role in chronic
pain paradigms. For that purpose, a new generation of siRNA were evaluated for
their improved in vitro and in vivo stability and potency to silence NTS2; namely,
dicer-substrate siRNA (DsiRNA).
3 Advantages of DsiRNA Over Conventional siRNA
As detailed in the previous section, RNAi is nowadays an important tool to

elucidate gene function or serve as a promising therapeutic approach in pain
research. However, the last 5 years also abound in publications reporting adverse
effects and consequently misinterpretations following siRNA treatments. We will

review, in the next pages, the improvements DsiRNA introduces in RNAi techno-
logies to circumvent the existing drawbacks.
3.1 Inherent Character of DsiRNA
In the last decade, RNA species of varying lengths have been tested for their
potency to silence target mRNA. Double-stranded siRNA (i.e., duplexes) was
shown more stable than single-stranded RNA, but additionally, those shorter than
30 base pairs were reported to be devoid of severe immune stimulation. Among
those tested, Kim and colleagues demonstrated that 25–30 bp duplexes were more
potent than corresponding 21-mers; 27-mers dsRNA presenting a 100-fold amelio-
ration (Kim et al. 2005). The hypothesis proposed implies a role for Dicer in
promoting the processing of the double stranded RNA (dsRNA) in an appropriate
size and its loading in the RISC complex. Classic 21-mers RNAs bypass this step
explaining the poor odds for their RISC loading. Although DsiRNA are intrinsically
more potent, optimal in vivo RNAi can only be achieved if the duplexes reach the
cells of interest, are properly up-taken into the cytoplasm and loaded in the
silencing machinery. To this day, no universal delivery system is yet available to
assist siRNA to cross all those barriers while remaining an inert carrier. Thus,
modifications made to the duplex structure are the best solution to circumvent
stability issues and specific/nonspecific off-target effects (OTEs). Evading from
nucleases action, decreasing the activation of the innate immune system, and
limiting the interference with the endogenous micro RNA pathway while optimiz-
ing the cellular uptake are challenges to reach a proper level of RNAi.
3.1.1 Dodging Nucleases Attack
Inherently, 27-nt dsRNA such as DsiRNAs are somewhat more stable in 10% FBS
serum than siRNA (t1/2 ¼ 83 h vs. 1.6 h) or in active serum like in mammals
(30 min vs. 5 min) (Kubo et al. 2007). Though, their stability can still be enhanced
to optimize RNAi. The most straightforward approach to increase siRNAs stability
is to modify their internucleotide phosphate linkage. Importantly, modifications
added to protect from nuclease degradation should not interfere with the gene-
silencing capacity of the duplex. In other words, the cellular proteins of the RNAi
machinery should maintain an adequate access to the guide strand. Indeed, for
DsiRNAs, attention is required when adding modifications to protect from serum
nucleases, since some degree of nuclease activity is still desired for Dicer endoribo-
nuclease to perform its cleavage of the substrate (Behlke 2008). Chemical
substitutions of a nonbridging oxygen are multiple to confer stability (nitrogen,
sulfur, boron, methyl). The most secure, easy to perform and studied substitution
remains the phosphorothioate. However, in the case of DsiRNA, the retained
strategy involves the introduction of 20 -O-Methyl RNA (20 OMe) in the strands. This
variant is naturally occurring in mammalian rRNA and tRNA (Behlke 2008) and it
was demonstrated that its inclusion in DsiRNA had minimal impact on potency at
multiples sites (Collingwood et al. 2008). Duplexes presenting 20 OMe RNA on both
sense (S) and antisense (AS) strands were however found inactive, just as too heavy
modification protocols did. The “evader” pattern was retained, introducing ten
20 OMe RNA bases in alternation in the AS strand. Modified DsiRNA were shown
to remain intact after 24 h of incubation in serum (Collingwood et al. 2008). This
pattern was exploited for research purposes until recently when experimenters
observed an in vitro loss of potency for certain sequences. Again, no rule could
help to define which targeted sequence would be affected or not. Behlke’s team thus
switched to a “m7 pattern”, with fewer 20 OMe bases in the middle of the AS strand
(seven in total), a hotspot recognized as sensitive to sequence substitutions. This
M7 modification pattern confers adequate stability while maintaining potency in
every sequence tested. Moreover, 20 -F modifications could be introduced in addi-
tion to 20 OMe, since this type of add-on is tolerated at the Ago2 site of cleavage.
However, 20 -F sugars are artificial and once available in the cells, they can incor-
porate in the endogenous transcriptome or genome (Behlke 2008). Toxicity sur-
veillance should be performed if a modified DsiRNA presenting both substitutions
is used in in vivo sustained release. Since, 20 OMe bases are present in the nature, no
apparent toxicity is anticipated if used as the sole modification.
3.1.2 Keeping “On-Target”
Specific “off-target” effects (OTEs) arise from the undesired recognition of a

mRNA different from the one coding for the protein of interest. This phenomenon
has been characterized by wide microarray studies showing the lack of specificity of
siRNA (Jackson et al. 2003). Off-target events occur even if the design was
performed accordingly to very stringent criteria. The main reason for unexpected
effects can initiate from a partial pairing of the nucleotides 2–8, sufficient to lead to
undesired mRNA degradation (Lin et al. 2005). However, these side effects can be
reduced either by decreasing the concentration of siRNA administered or by
inserting chemical modifications in the interfering RNA structure. The 20 OMe
modifications made on DsiRNA to divert nucleases degradation have proven
efficient in reducing OTEs as well. However, Rose et al. (2005) demonstrated
that blunt 27-mers lead to an array of dicing patterns, limiting the specificity of
the silencing toward a target sequence. Up to five dicing species could be detected.
Thus, they proposed the inclusion of a 30 -overhang on (AS) strand to confer more
homogeneity in the dicing products. Additionally, they substituted two DNA bases
at the blunt 30 end of the (S) strand, which is more tolerant to this kind of
modification. Indeed, DNA residues at this site orient the PAZ domain of Dicer
toward the 30 -overhang of the AS strand, yielding a predictable cleavage product of
21 bases (Rose et al. 2005). This improved design of DsiRNAs gives a higher rate of
success for precise silencing of a target sequence.
3.1.3 Eluding the Unpredictable
Nonspecific off-target effects are related to unexpected immune response and

toxicity that is intrinsic to siRNA or attributable to the delivery formulation (see
Sect. 4). Indeed, interferon (IFN) response is typically induced by long double-
stranded RNA molecules (30 nucleotides or more), but was also detected following
classic 21-mers siRNA administration. This activation is mediated in parts by
protein kinase R and toll-like receptors (TLRs) (Kurreck 2009), which recognize
siRNAs and activate multiple signaling pathways involving the nuclear factor kappa
B (NFkB), mitogen-activated protein kinase (MAPK), or interferon regulatory
factor (IRF). The major effect of such IFN activation is a massive degradation of the
mRNA pool and a global inhibition of transcription, often leading to cell death and
apoptosis in vitro. It has been demonstrated that TLRs are particularly sensitive to
specific motives that should be avoided in siRNA design. Indeed, single-stranded
RNAs activate TLR7-8, CpG motives trigger TLR9 while certain GU-rich
sequences have a high affinity for TLR3 (Sledz et al. 2003; Kariko et al. 2004).
Interestingly, DNA substitutions were shown to block immune detection (Behlke
2008). 20 OMe was reported to decrease PKR and TLR7 through a competitive
inhibition. Although longer siRNA are susceptible to raise OTEs, modified
DsiRNAs with 20 OMe on at least two G or U residues and DNA bases greatly
decrease the risk of an immune response. Collingwood et al. (2008) demonstrated
that 20 OMe modifications on a single DsiRNA strand did not raise any pro-
inflammatory cytokines out of a broad panel (IL-1b, 2, 4, 6, 8, 10, 12, TNF-a,
IFN-a, IFN-g and GM-CSF). Keep in mind that TLRs are very present in endosomal
compartments, a route of entry for cationic lipid and liposome formulations. Thus,
carriers are not inert and modifying the sequence is not the only aim to focus on in
order to optimize gene silencing.
3.1.4 Sharing Is Not Always Best
A less characterized undesirable effect of siRNAs is the crosstalk with endogenous

interference pathways, i.e., micro RNA (miRNA) translational suppression path-
way. The interaction of siRNAs with 30 UTR segments of mRNAs can block their
expression without their degradation, just as miRNAs do (Saxena et al. 2003).
Consequently, shRNAs have been reported to cause an accumulation of Exportin-5,
which is responsible for the transport of miRNA precursors. It has also been
recently demonstrated that standard exogenous siRNAs can cross-react directly
with miRNA signaling (John et al. 2007). A 6–7 nt base “seed region” at the 50 end
of the AS strand is sufficient to lead to undesired gene suppression. Since the
probability of a 6–7 bp interaction with the whole transcriptome is high, even
with a careful cross-hybridization analysis while designing, participation of
siRNA in miRNA pathways remains a reality. Interference with the miRNA
pathway has been associated with severe toxicity in vivo under high siRNA dose
administration (Grimm et al. 2006). Replacing the seed region by DNA bases or
adding a 20 OMe in position +2 are strategies to reduce the interference with

miRNA. Another approach consists in using very low doses of siRNA, the down-
side being the risk of loosing potency.
In summary, the fact that DsiRNAs provide six times more stability in active
serum and three times higher long-term RNAi activity than other siRNA species
(Kubo et al. 2007) is also an asset in decreasing effective doses and consequently
OTEs. Add that 20 -F also decrease nonspecific OTEs and increase stability, a new
generation of DsiRNAs could lead to a very potent and durable silencing. Finally,
DNA substitutions in the 30 -blunt end and use of RNA bases at the 30 -overhang
complementary to the target infers functional polarity to favor antisense strand
loading into the RISC complex by Dicer (Rose et al. 2005).
3.2 Molecular Mechanism of Action
We herein briefly describe in three main steps an overview of Dicer-substrate

siRNA processing; the cleavage of DsiRNA in 21-mers molecules by Dicer, the
loading into RISC complex and the recognition and degradation of target mRNA.
3.2.1 Initiation of RNAi Pathways: Role of DICER
As showed in step 1 of Fig. 1, DsiRNAs are first processed by the endoribo-

nuclease Dicer in the cytoplasm after their cellular uptake. This 200 kDa protein
belongs to the RNAIII family of enzymes consisting of ATPase/RNA helicase and
PAZ domains, two catalytic endonuclease domains (endoND), and a C-terminal
dsRNA binding domain (dsRBD) (Ji 2008). The two endoND domains are
conserved in all RNase III proteins and form a dimer responsible of the hydroly-
sis of the DsiRNA opposing strands into 21–23 nucleotides length double
stranded (dsRNA) left with 2-nucleotide-long 30 -overhangs. These features of
dicer-processed siRNA ensure their efficient incorporation into the RNA-
induced-silencing-complex called RISC (Jinek and Doudna 2009). The PAZ
domain is also highly conserved in Dicer and Ago proteins. The hydrophobic
pocket of the PAZ domain is responsible for binding to the 30 -dinucleotide
overhang Dicer substrate and serves as a siRNA-end binding module for siRNA
within the RISC complex (Ji 2008). PAZ and the first endoND domain are
separated by an a-helix that functions as a molecular ruler to generate the 20-
base pair dsRNA products (Macrae et al. 2006). Vertebrates contain single Dicer
genes, while Drosophila and some other organisms express multiple Dicer iso-
forms with specialized functions (Tang 2005; Lima et al. 2009). The observations
made for Dicer variants propose the existence of modulatory mechanisms for
siRNA transfer into RISC.
Cell membrane
25/27mer DsiRNA
transfection Sites of
Dicer processing Degraded target mRNA
5’ pCCUACUGACCCAGUGCAUACUGC a g
3’ CUGGAUGACUGGGUCACGUAUGACGUC
5
1 Ago 2 mRNA
NNNNNNNNNNNGACCUACUGACCCAGUGCAUACUANNNNNAAAA
3’ CUGGAUGACUGGGUCACGUAU
Dicer
5’ pCCUACUGACCCAGUGCAUACUGCa g
3’ CUGGAUGACUGGGUCACGUAUGACGUC
Target recognition
2 4
CU
A
Ago 2 GCAU
5’ CCUACUGACCCAGUGC AUAtt 5’ pCCUACUGACCC AGUGCAUACU U
ACCCAG
3’ ttGGAUGACUGGGUCACGUAU 3’ CUGGAUGACUGGGUC ACGUAUp 5’ pCCUACUG
3’ CUGGAUGACUGGGUCACGUAUp
5’ pCCUACUGACCC AGUGCAUACUG
Dicer product 3’ CUGGAUGACUGGGUC ACGUAUGp
(21-mer and 22-mer)
3
Standard synthetic
Transfection Ago 2
21-mer siRNA
5’ pCCUACUGACCCAGUGC AUACU
3’ CUGGAUGACUGGGUCACGUAUp
Cytoplasm RISC
Fig. 1 Mechanism of Dicer-substrate small-interfering RNA (DsirNA) processing in mammalian

systems. (1) After cellular uptake, Dicer endoribonuclease takes in charge the DsiRNAs released
from the delivery carrier. The introduction of DNA bases (red lowercases) on the (S) strand forces
the positioning of the DsiRNA within the Dicer. (2) The modified DsiRNA is processed into
homogenous 21/22-bases product and benefits from the preassociation with Dicer to facilitate
loading into the RNA-induced silencing complex (RISC). (3) Asymmetric 27-mer duplexes
(DsiRNA) are incorporated in the RNA interference processing machinery one step earlier than
conventional 21-bases siRNA. Thus, the odds are in favor of DsiRNA for success of loading. (4)
During RISC assembly, one the two siRNA strands, referred as the passenger strand, is cleaved and
released, whereas the other strand (guide strand) is incorporated into the argonaute protein (Ago2)
component of RISC. (5) The remaining single-stranded siRNA then guides the RISC complex to
its complementary target mRNA. In general, perfect matching directs the cleavage of target
mRNA by the endonucleolytical activity located in the Ago2 protein, whereas bulged siRNA
directs translational suppression [Adapted from Dore-Savard et al. (2008)]
3.2.2 Formation of RISC Loading Complex and Its Activation
Many proteins have been identified as essential in the multiproteic complex of

RISC but are not well characterized yet. The most documented components of this
complex are Dicer and Argonautes proteins (Ago). The preassociation of DsiRNA
with Dicer occurring prior to the cleavage facilitates their loading onto Ago at the
core of the RISC complex and therefore leads to higher levels of processing in
comparison to classic siRNA (Kim et al. 2005). During RISC assembly, one of the
two siRNA strand referred to as the guide strand (or AS) is loaded onto Ago
(step 2). Afterwards, RISC is activated by the cleavage of the other strand called
the passenger strand (or S) and the guide strand directs the sequence specificity of
RISC and therefore the RNAi silencing (step 3). The incorporated strand that serves
as the guide strand is generally the one which 50 terminus is at the thermodynami-
cally less stable end of the duplex. For instance, in the case of asymmetric siRNAs,
it is easier to unwind the duplex from the less stable end, thus preferentially
facilitating the incorporation into the RISC complex. In contrast, a symmetric
siRNA possesses two equally stable ends, thus both strands of the siRNA are
assembled into the RISC complex with an equal efficiency (Schwarz et al. 2003).
Ago is a multidomain protein of 100 kDa comprising the PAZ domain and the
C-terminal PIWI domain. Cleavage catalysis is mediated by the PIWI domain
showing extensive homology to RNase H, an endonuclease that nicks RNA strands
in 10 nucleotides segments from the 50 end siRNA, leaving the siRNA intact for
another round of cleavage (Ameres et al. 2007; Hock and Meister 2008). A third
functionally important domain that resides between the PAZ and the PIWI domains
is termed the middle domain (MID), containing a high basic pocket that binds the
characteristic 50 phosphate of small RNAs (Hock and Meister 2008). Thus, evidence
supports that the 20 -nucleotide-long 30 overhangs and the 50 phosphate of siRNA are
anchored onto Ago proteins by the PAZ and MID domains, respectively. The
number of Ago genes is highly variable between species. Eight Ago proteins are
encoded by the human genome and of these, only hAgo2 shows an endonucleolytic
activity (Ameres et al. 2007). Therefore, several forms of RISC have been reported.
Indeed in human cells, a third player called TRBP, a RNA-binding protein, is
necessary for efficient transfer of siRNA from Dicer to Ago2, and thus essential
for the RISC loading complex formation (Wang et al. 2009).
3.2.3 Target Recognition by RISC and Gene Silencing
Depending on both nature of Ago protein and the degree of complementary between
the DsiRNA and the mRNA-target sequence, gene silencing results as the cleavage
or translational repression of mRNA. If the base-pairing complementary between
siRNA and its target mRNA is not optimal, the target may be physically unreachable
by the active center of the endonuclease domain of RISC complex. Consequently,
this will result in translational repression instead of efficient cleavage of target
mRNA. By contrast, a noncleaving RISC complex resulting in a lack of slicing
activity of Ago can direct the target mRNA only for translational repression regard-
less of high complementarity between siRNA and target (van den Berg et al. 2008).
The way that RISC complex efficiently recognizes its target remains unclear.
However, evidences suggest that a RISC-mediated siRNA-target interaction is
more complex than simple nucleic acid hybridization and some factors within
RISC may facilitate target recognition (Hutvagner et al. 2004).
Overall, DsiRNAs demonstrate a higher efficiency in RISC loading. This is
imputable to the fact that DsiRNAs are a substrate of Dicer enzyme instead of a
product (i.e., siRNAs) and that the complex formed binds directly to Ago in the core
of the RISC complex. The advanced design elaborated from experimental observa-
tions done on silencing efficiency allowed to further modify the duplex to increase
its potency. DsiRNAs are far more stable and their uptake and processing surpasses
the standards in in vitro knockdown research.
4 DsiRNA for Efficient Silencing In Vitro
In order to obtain a potent and efficient knockdown when using RNAi technologies,
the first step consists in a sturdy design. Thereafter, rigorous controls and the choice
of a representative cell assay are the basis to optimize the validation toward a
further use in in vivo research. In the next section, we will describe the preparation
process of DsiRNA candidates suitable for in vivo pain research.
4.1 Methodology
4.1.1 Design Rules
Various design methods exist, but for all cases, there are basic rules to follow.
Because it has become more convincing in the literature that chemically modified
asymmetric 27-mers siRNA are more potent than unmodified blunt 27-mers, we
will place more emphasis on general guidelines for this particular kind of mole-
cules. Indeed, as reported in the previous section, processing of unmodified 27-mers
duplexes by Dicer is unpredictable and can result in the generation of siRNAs of
poor activity, sometimes below that of an optimal 21-mer siRNA (Amarzguioui and
Rossi 2008). Indeed, the chemically modified asymmetric 27-mers DsiRNA pro-
posed by Rose and colleagues or similar species with different modifications
suggested by Kubo et al. are more stable than classical 21-nt dsRNAs and possess
higher long-term RNAi activity (Rose et al. 2005; Kubo et al. 2007). However,
because a range of potencies is frequently seen among different target sites within a
same gene, it is highly recommended to design multiple siRNA candidates for the
same gene of interest.
By analyzing the relevant structural features of endogenous pre-miRNAs, the
problem of making DsiRNA processing predictable has been solved (Amarzguioui
et al. 2006). There are now many siRNA design algorithms that are freely accessible
on the web. The aim here is to help the reader to access a list of algorithms to choose
suitable criteria for optimal designing. Charité’s University, in Germany, has listed
many of the currently available Web-based siRNA design tools (http://itb.biologie.
hu-berlin.de/nebulus/sirna/siDesign.htm) separated in commercial- or academic-
supported Web site. Fully automated DsiRNA design is also available through
Integrated DNA Technology (IDT) (http://www.idtdna.com/Scitools/Applications/
RNAi/RNAi.aspx). But whatever design algorithm is used, to increase the chance of
success, it is suggested to blast more than one and choose candidates with high
scores in multiple or all of the algorithms (Amarzguioui et al. 2006).
4.1.2 Cell Lineage Selection
Once the DsiRNA candidates are designed and manufactured, in vitro testing is
prime. The choice of the cell lineage is important in regard to the targeted gene. The
choice should be made toward cells natively expressing the gene of interest and
ideally in a cell line specific to the pathology or tissue aimed for future in vivo
studies or clinical therapy (e.g., neuronal cell line in the event of in vivo pain
research). However, expression levels of a gene of interest can be very low or
absent in commercially available cell lines. In such a case, the use of a transgenic
expression cell line is suggested. Nevertheless, overexpression of a gene can bias
the expected silencing outcomes of a DsiRNA. Indeed, silencing efficiency can be
high due to an imposing “artificial” pool of target mRNA available in the cytoplasm
of transgenic cells, but results could be rather disappointing in ex vivo or in vivo
systems attributable to more realistic stable levels of various mRNAs. Therefore,
consideration of more than one potent in vitro candidate is essential to act as a
backup in case of a weak hit in vivo.
4.1.3 Extensive Formulation Screening
Once the cell line is chosen, the selection of an optimal formulation agent is
required to facilitate cell uptake. Indeed, naked RNA usually gives poor silencing
rates at low doses and is toxic at high doses at which cellular uptake is though
greater. Formulation is a complex research domain and there is no universal recipe
for an effective delivery. Yet, in vitro research allows performing extensive evalu-
ation assays in parallel conversely to in vivo research for which costly and time-
consuming experiments limit the margin for error. Thus, because there is no
universal reagent, every cell line, tested for DsiRNA uptake, should be screened
with a battery of transfection reagents to identify the most potent. The final goal
remains to find the best balance between an efficient knockdown and the least
toxicity and variability between assays. The best protocol consists in using a
“control” DsiRNA (e.g., HPRT is a constitutively expressed housekeeping gene)
to screen the optimal reagent formulation for a specific cell line. Once the reagent is
chosen, then the assays using the target DsiRNA at study can be performed.
The most frequently employed class or reagents for classical 21-mers siRNA,
i.e., cationic lipids, can work as well for DsiRNA. Note that an adjustment of the
N:P ratio should be considered to optimize the formulation (i.e., the equilibrium
cationic lipid nitrogen/anionic RNA phosphate charge ratio to insure optimal
formulation but also appropriate release once in the cell). Nevertheless, polymer
or peptide-based reagents are also conceivable alternatives for delivery where
cationic lipids failed to yield high transfection levels. Hard-to-transfect cells such
as neurons are ideal candidates for new formulations such as the protein-based
transfection reagent Transductin™. Indeed, this novel transfection reagent, com-
posed of a peptide transduction domain (PTD) and a double-stranded RNA-binding
domain (DRBD), has recently shown to be able to efficiently transfect siRNA into
primary cultures of hard-to-transfect cells (Eguchi and Dowdy 2009; Eguchi et al.
2009). New avenues are frequently exposed in specialized journals such as
Advanced Drug Delivery Reviews to which the reader can refer for updates.
4.1.4 Control Your Assay
As previously stated, controls must be carefully chosen throughout the validation

process. At least four controls should be performed while evaluating the gene of
interest levels: (1) cells only for basal gene expression level; (2) cells and transfec-
tion reagent per se to assess possible toxicity and/or undesired effects of the reagent
on gene expression; (3) cells, transfection reagent, and negative control DsiRNA
(i.e., scrambled DsiRNA) to assess off-target effect; (4) cells, transfection reagent,
and positive control DsiRNA (e.g., HPRT) to ensure the efficiency of the silencing
and the assay performance. The scramble is also used to set the remaining mRNA
levels to 100% to allow normalization of the assay and determine knockdown
efficiency of the target mRNA. The two first controls should show similar levels
to the scramble negative control (approximately 100% when normalized) thus
allowing the confirmation of the absence of interference.
Besides expression level controls, a cellular uptake control could be performed.
For instance, it can be a fluorescent control (i.e., scramble DsiRNA coupled to a
fluorescent dye) to determine if the formulation efficiently incorporated the cells
without apparent cellular toxicity. Recordings can be performed on fixed cells or in
live acquisition with a confocal spinning disk microscope (see Sect. 5). This
technique should always be performed second to molecular validation since the
entry of the DsiRNA in the cytoplasm does not reflect its potency. Performing this
set of controls is a prerequisite for validating functional DsiRNA and increasing the
chances of success for further in vivo applications.
4.2 Validation of Knockdown Efficiency
A dose-response curve is paramount to discriminate between DsiRNA candidates.

It will allow the experimenter to find the right balance between reaching an efficient
functional knockdown while keeping the dose to a minimum to avoid activation of
off-target effects. Also, the best candidates are the ones showing an improvement in
silencing efficiency in respect to increasing doses. Note that the concentration of
DsiRNA is dependent of the type of reagent to be used. Typically, for standard
cationic lipid, a maximum of 10 nM is generally enough to induce a near maximal
knockdown effect. In this case, dose response will vary as follow: 10 nM, 1 nM, and
0.1 nM. However, other carrier reagents such as peptides or liposomes often require
greater concentrations of DsiRNA. Positive and negative controls should be set at
the maximal concentration used in the dose-response curve. Keep in mind that
when looking at the knockdown efficiency, the results greatly depend upon the cell
type used and the targeted gene. For instance, easy to transfect HeLa cells are in
general expected to provide a knockdown of about 95% with the maximal chosen
dose for positive control DsiRNA. However, weaker silencing might be achieved if
targeting a gene with a strong compensation mechanism. On the other hand, with
hard-to-transfect cells (e.g., RAW), reaching 70–80% of knockdown with a high
dose of DsiRNA is considered satisfactory.
Whichever cells or transfection agent are at use, the key step for obtaining
optimal DsiRNA silencing remains to extensively optimize each cell line with a
series of reagents using a housekeeping positive control DsiRNA. The number of
DsiRNA candidates to be tested is user-dependent. An experimenter that requests
only one or two good DsiRNAs will approximately analyze a dozen issued
from appropriate algorithms. However, for a researcher who wants several potent
DsiRNAs, tiling the entire gene with DsiRNA may be the only way to find the
optimal one.
4.3 Targets In Vitro
Throughout the whole literature referring to dicer-substrate siRNA use in vitro

(including unmodified and modified 27-nt duplex), no reports point at a gene known
or suggested as implicated directly in pain mechanisms. Consequently, we will
focus on the few articles relating the use of DsiRNA in an in vitro application.
Dong-Ho Kim and colleagues, in 2005, published one of the first articles discussing
the increased potency of DsiRNA over classical 21-nt siRNA (Kim et al. 2005). In
this innovative study, the authors utilize increasing concentration of DsiRNA from
50 pM to 50 nM, formulated to the cationic lipid reagent Lipofectamine 2000
(Invitrogen) in two different cell lines, namely HEK293 and NIH3T3. Cells trans-
fected with the enhanced green fluorescent protein (EGFP) were incubated with a
target DsiRNA and fluorescence was quantified to assess silencing efficiency. With
this methodology, they demonstrated a more potent and longer lasting knockdown
effect of 27-nt DsiRNA over conventional 21-nt siRNA. In 2006, it has been
demonstrated the possibility to inhibit TNF-a expression when transfecting
RAW 264.7 macrophage cell line with TransIT-TKO reagent from Mirrus Bio
(Amarzguioui et al. 2006). According to the authors, this reagent was preferred
in vitro because of its high transfection efficiency and low toxicity for those cells.
They also reported applying successfully this method to the nonadherent cell line
HL60. This study is indirectly interesting for pain research because of the known
role of TNF-a in inflammation mechanism that can participate in exacerbating pain.
More recently, in 2008, DsiRNAs targeting five different genes, which are CDK2,
TP53, RAF1, ACTB, and AKT1 were tested (Hefner et al. 2008). Using the HeLa
cell line and siLentFect reagent from Bio-Rad, they managed to obtain successful
knockdown for all genes with a range of DsiRNA concentration from 100 pM to
50 nM. IDT R&D team further demonstrated the potency of different chemical
modifications on DsiRNA, including 20 -O-methyl (20 OMe) RNA, 20 -fluoro (20 -F)
RNA and DNA bases (Collingwood et al. 2008). Based on their study, a modifica-
tion pattern that includes a 20 OMe bases was identified as the one having minimal
impact on potency, did not trigger immune responses in human immune cells
in vitro and showed improved serum stability. This study was performed either in
HeLa cells with TriFECT reagent (IDT), HCT116 cells with Lipofectamine 2000,
human peripheral blood mononuclear cells (PBMCs), again with Lipofectamine
2000, or T98 cells with siLentFect reagent. In an interesting study recently pub-
lished by Takanori Kubo and colleagues, this group demonstrated the efficacy of
amino-modified DsiRNA by using a luciferase gene reporter assay in HeLa cell line
using the transfection reagent Lipofectamine 2000 (Kubo et al. 2008). The authors
further proposed two different routes (Fig. 2) by which DsiRNA can be processed.
Depending on the modification made to the DsiRNA structure, the second route of
activation, the only one leading to induction of gene silencing, would be favored
thus increasing the knockdown potency. This group also used an unconventional
transfection method represented by the conjugation of 27-mers DsiRNA with
cholesterol bound at the 50 -sense end. This construct possesses high membrane
permeability in the absence of delivery reagent. These results were in agreement
with the data previously demonstrating this phenomenon with classical 21-nt
siRNAs (Soutschek et al. 2004). Another recent study evoked the use of DsiRNA
without a conventional transfection method (Zhou et al. 2009). They created a
chimera composed of DsiRNA linked to an anti-gp120 aptamer, which specifically
binds to and internalizes into cells expressing HIV gp160. Once inside the cells,
DsiRNAs can be processed by dicer and further inhibit HIV-1 replication and
infectivity in CEM T-cells or PBMCs. This study demonstrated the increasing
possibilities of DsiRNA for future therapies. DsiRNAs complexed to chitosan
nanoparticles is also a method proposed to transfect primary peritoneal macro-
phages (Howard et al. 2009). With this technique, they managed to obtain a
knockdown of approximately 66% of TNF-a mRNA pools. This study supports
the one of Amarzguioui in 2006, and both demonstrate an effective knockdown of
TNF-a and a potential for those DsiRNA to be tested in a pain paradigm.
The first studies directly targeting pain mechanisms with DsiRNAs were
published by our group in 2008 (Dore-Savard et al. 2008) and more recently by
others (LaCroix-Fralish et al. 2009). In the first study, in vitro validation of DsiRNAs
targeting the NTS2 receptor, a GPCR known to be involved in nociceptive modula-
tion, was performed (Roussy et al. 2009). Indeed, six DsiRNA and an appropriate
mismatch control were evaluated in vitro for their ability to specifically reduce
NTS2 mRNA levels in an NTS2-stable Chinese Hamster Ovary (CHO) cell line. The
cells were treated with a dose-response curve (0.1–10 nM) of DsiRNA formulated to
RNAiMAX reagent. The capacity for DsiRNA to silence NTS2 expression was
further analyzed using quantitative reverse-transcription PCR. On the six DsiRNA
tested, three resulted in nearly complete inhibition of the targeted gene at 10 nM and
a Non-modified asymmetric duplex RNA

Dicer
Access at 3'-sense
Asymmetric RNA dangling end- easier
5' Sense 3'
Access at 3'-antisense
bland end 3' Antisense 5'
Dicer
Sense
5' 3'
Sense 3' 5'
5' 3' Antisense
3' Sense 5'
Route A Route B
Sense
RISC 5' 3' RISC
Sense 3' 5'
3' RISC interacts
3' 5'
Sense with sense strand
Antisense
3' 5' RISC interacts with
5' 3'
Sense antisense strand
Cleavage! Induction of gene silencing
mRNA
Rate of Route A < Rate of Route B
b 5'-Sense amino-modified asymmetric duplex RNA

Dicer
Access at 3'-sense
dangling end –easier
Access at 3'-antisense bland end 5' NH Sense 3'
2
– impeded by the NH2 -group 3' Antisense 5'
Dicer
Sense
5' 3'
5' Sense 3' 3' 5'
Antisense
3' Sense 5'
Route A Route B
Sense
RISC 5' 3' RISC
Sense 3' 3' 5'
RISC interacts
3' Sense 5' with sense strand
Antisense
3' 5' RISC interacts with
5' 3'
Sense antisense strand
Cleavage! Induction of gene silencing
mRNA
Rate of Route A << Rate of Route B
Fig. 2 Relationship between dicing and RNAi activity of asymmetric duplex RNA and 50 -sense
amino-modification. (a) Asymmetric duplex increases the interaction of Dicer with the 30 -sense
dangling end, thus promoting route B, the only one leading to efficient gene silencing, over
route A. (b) Amino-modified 50 -sense strand impedes the access of Dicer to the 30 -antisense
end, thus blocking route A and leaving route B as the only one active. This figure represents an
hypothetical mechanism of the higher efficacy of modified asymmetric DsiRNA (Kubo et al. 2008)
remained effective at a very low dose (0.1 nM). In vitro validation was assessed for
functional potency in NIH3T3 cell line (LaCroix-Fralish et al. 2009). In this case, 0.1
or 10 nM DsiRNAs were transfected with siLentFect reagent and incubated for 24 h
on cells before RNA isolation and subsequent quantitative real-time PCR analysis.
Atp1b3, a gene coding for the b3 subunit of the Na+, K+-ATPase pump that
contributes to interstrain differences in the early phase of the formalin test, was
the targeted gene in this study. When testing five different Atp1b3 DsiRNAs, they
managed to obtain more than 90% knockdown at the highest dose, and two of the
candidates remained potent even at 0.1 nM. These data are expected to stimulate
in vivo pain research in the next few years.
5 DsiRNA In Vivo: One Step Forward
RNAi studies rapidly reached the step of applicability to animal models and clinical
trials. This transfer from the cell to the organism is impressively fast considering
that this mechanism was discovered in the late 1990s. The first experiments using
siRNA have been performed in peripheral organs so was the first ex vivo assay
with DsiRNA (Amarzguioui et al. 2006). Only when the application was proven
functional, the technology moved onward the central nervous system. The CNS
environment being hermetic, difficult to access without invasive procedures and
having low tolerance for exogenous contaminants contributes to making this system
less targeted. The situation is nowadays different as technologies evolve. As dis-
cussed earlier in this chapter, DsiRNAs are particularly suitable for such studies
since the effective dose used is very low (ng to low mg), thus lowering side effects.
The formulations used are also more potent in delivering their content in hard-
to-transfect neuronal cells. In the following section, we will describe the application
of DsiRNA in vivo at large, with a final focus on the CNS and pain research.
5.1 Working Evidence of DsiRNA In Vivo
5.1.1 Peripheral Organs
A recent study showed efficient delivery of DsiRNA to the lungs by intratracheal

administration (Merkel et al. 2009). They characterized the delivery of EGFP-
targeting DsiRNA into the lungs using PEG-PEI polyplexes. DsiRNAs were first
radiolabeled for future visualization and formulated with either one of the PEG-PEI
polyplexes tested. Using a 24 gauge catheter, 50 ml of PBS containing 50 mg of
EGFP DsiRNA was instillated through the trachea of actin-EGFP mice. EGFP
fluorescence significantly decreased in treated-mice bronchioles, 5 days after
instillation. Additionally, flow cytometry quantification of the knockdown was
performed on lung homogenates. The residual fluorescence in the cells was quanti-
fied on a FACScan. The DsiRNA-treated animals showed a 42% decrease of the
level of fluorescence compared with controls, validating the efficiency of DsiRNAs

and their delivery method.
Another group reported the uptake of DsiRNAs injected intra-peritoneally by
macrophages in mice (Howard et al. 2009), using the same novel delivery method
as in the in vitro section (chitosan particles/DsiRNA). DsiRNAs targeting TNF-a
were formulated in vitro and prepared for in vivo administration. 200 ml of the
formulation (2.5 mg or 5 mg of DsiRNA) were injected intra-peritonealy. Macro-
phages were then isolated and plated for control Cy3-labeled DsiRNA visualization
or TNF-a quantification via immunoassay. The efficacy of the DsiRNA treatment
was verified using a mouse model of collagen type II-induced arthritis. The
same formulation was administered five times over a 9-day period to arthritic
mice and the development of inflammation in the paw was followed for 14 days.
DsiRNA-treated animals had their arthritis score significantly reduced at the end of
the experiment, showing an inflammation level between that of scramble- and
dexamethasone-treated (negative and positive control, respectively) mice.
Cancer being one of the most investigated pathologies, studies using DsiRNAs
to specifically target tumor cells in vivo have started being published (Kortylewski
et al. 2009). In that study, the authors used DsiRNAs to inhibit the transcription
factor stat3 to potentiate the antitumor immune response in vivo. An original
method to target preferentially immune cells was employed by conjugating a
TLR9-antagonist sequence to the stat3 DsiRNA. The delivery method consisted
in a single peri-tumoral injection of 0.78 nmol of the DsiRNA. Tumor growth was
monitored every other day. A systemic approach was also used for a 3-week
treatment. The same dose of DsiRNA was injected (this time intravenously) every
2 days. The level of stat3 was assessed by qPCR on lymph nodes or subcutaneous
tumor extracts. A high expression of various immune cytokines was observed and
correlated with decreased tumor proliferation and metastatic activity.
5.1.2 Central Organs
Two recent studies also demonstrated the potency of DsiRNAs to knockdown a

target in the spinal cord in the context of pain research. Both groups used behavi-
oral, pharmacological, and biochemical observations to confirm the knockdown
efficacy. The following subsection will take benefit of those two significant pub-
lications to describe the methodology for the use of DsiRNAs in vivo.
5.2 Methodology
5.2.1 Formulation
The first choice the experimenter has to make in order to use RNAi in vivo is the
administration technique in combination with the best carrier to favor an optimal
penetration in the targeted tissue and cells. Recently, high efficiency has been
reached with cationic lipids. Their relative toxicity limited the amount used for
systemic studies but the lipid iFect proved to be very efficient in the central nervous
system when locally injected (Luo et al. 2005). DsiRNAs are first diluted in sterile
physiological saline and then formulated with the iFect at a ratio of 1:4–1:5 to avoid
precipitation impairments. A 10 ml volume of the formulation is injected in the
intrathecal space of the rat (5 ml for mice) between L5 and L6 vertebrae under brief
(but deep) anesthesia. It is important to mention that, despite its obvious efficiency,
some concern has been raised with the repeated administration of iFect. Indeed,
transient ataxia on the rotarod was observed (LaCroix-Fralish et al. 2009) while
Sarret’s group noticed a certain level of interference with the pain response at the
formalin test (Dore-Savard et al. 2008). Sarret’s group managed this issue by testing
different cationic lipids, polymers and finally obtained a satisfactory result with the
peptide agent Transducin. As opposed to the other reagents tested, no adverse
effects were identified in tonic pain tests.
When it comes to the knockdown of a targeted protein, many strategies could be
used in terms of dose and timing of injection. However, it is now accepted that
repeated daily injections (generally 2 or 3) in the low mg range are enough to induce a
significant decrease in the expression of the selected molecule. For instance, 1 mg
of DsiRNA daily injected intrathecally for two consecutive days resulted in a
knockdown for three subsequent days (Dore-Savard et al. 2008). In the mouse, a
dose of 0.5 mg has been administered for three consecutive days. The effectiveness of
the knockdown was verified 24 h after the last injection (LaCroix-Fralish et al. 2009).
5.2.2 Uptake Validation
The next step before running any knockdown experiment is to ensure that delivery
of a fluorophore-coupled DsiRNA is observable in extracted tissue. This strategy was
used by Dore-Savard et al. (2008) with a Texas Red-coupled scramble duplex injected
i.t. and then observed in DRG neurons and superficial layers of the spinal cord
24 h after the last injection (Fig. 3a, b). LaCroix-Fralish and colleagues performed a
similar validation using a Cy3-labeled scrambled DsiRNA to visualize its entry
in dorsal root ganglia (DRG) neurons after an intrathecal injection (Fig. 3c, d)
(LaCroix-Fralish et al. 2009).
5.2.3 Molecular Silencing Confirmation
The level of remaining expression of the target mRNA after the knockdown repre-
sents crucial information. Real-time quantitative PCR is the approach of choice to
confirm the knockdown at the molecular level. However, to avoid misinterpretation
of the phenotypic behavioral observations, protein expression levels should not be
neglected. If the target mRNA content is decreasing but the protein of interest still
present (e.g., secretion vesicles still loaded), the conclusions observed at the behav-
ioral level will not be the same. Immunohistochemical studies using an antibody for
a b
c d
Fig. 3 Distribution of fluorescence in dorsal root ganglia and spinal cord following intrathecal
injection of a fluorophore-tagged dicer-substrate siRNA. (a and b) Texas Red-labeled control
DsiRNA distribution is shown in rat DRG and lumbar spinal cord after two spinal injections (1 mg).
The fluorescence is heterogeneous and expressed in every neuronal subpopulation. (c and d) The
same observations are made with a Cy3-labeled DsiRNA (red) injected in the mouse intrathecal
space (0.5 mg). Nuclei were counterstained with DAPI for visualization purposes (Blue). Scale bar:
(100 mm)
the targeted protein or a western blot can be performed to ensure that the mRNA
silencing resulted in an effective knockdown of the protein as well.
5.2.4 Assessing Specificity
As it is routinely done for in vitro experiments, it is important to eliminate the

possibility of an off-target action induced by in vivo DsiRNA administration. First,
a sequence-dependent effect has to be eliminated using one or several control
DsiRNAs with no affinity for any mRNA of the transcriptome. Further, at least
one additional DsiRNA designed for the same messenger, but aiming at a different
region of the target gene, must induce a similar molecular and physiological effect.
This way, the experimenter can confirm the expected knockdown. In order to
support the elimination of a sequence-dependent off-target effect, the functional
activity of a close gene from the same family can be assessed. It was verified that
the activity of the neurotensin receptor NTS1 was not modified by NTS2 knock-
down (Dore-Savard et al. 2008). Finally, the most frequent sequence-independent
side effect of RNAi being the induction of an immune response, the evaluation of
the expression levels of pro-inflammatory genes like IFN-g, TNF-a, Il-6, or COX
can be determined. The latter can be detected in tissue lysates using qPCR or in
cerebrospinal fluid using large-scale immunoassays (e.g., Bioplex: Bio-Rad). Once
all the important controls are performed, the functional observations regarding
the effects of a knockdown can be analyzed without any bias, even if we start
with the concept that low doses of DsiRNA make these events presumably rare.
5.2.5 Pain Assessment
In order to study the role of a gene in pain perception, the effect of the deletion
needs to be observed at the behavioral or electrophysiological level. Acute pain
tests such as the tail flick, hot plate, von Frey hair, or Hargreaves radiant plantar test
are adequate to detect any modification of the pain threshold at a basal state. A step
further, the formalin tonic pain test is a well-characterized biphasic inflammatory
test to evaluate persistent pain in rodents. The acute phase (0–9 min), induced by
direct stimulation of the nociceptors by the chemical substance, is followed by an
interphase of active inhibition (10–20 min) and ended by a longer tonic inflamma-
tory phase (20–60 min) consisting in secondary hyperalgesia, inflammation, and
central sensitization. The above tests are ultimately performed to predict the
behavior of a new treatment in chronic pain.
In order to provide a reliable “proof of concept” for DsiRNA in the central
nervous system, Sarret’s team chose to aim at a GPCR involved in pain modulation
mechanisms. In 2008, they proved that the intrathecal injection of DsiRNAs could
result in an efficient knockdown of the neurotensin NTS2 receptor known for its
central analgesic effects (Dore-Savard et al. 2008). First, they exposed that the
central delivery of a texas-red DsiRNAs formulated to iFect efficiently penetrated
DRG and spinal cord neurons. Then, they started by determining the efficacy of this
technology in acute pain. Two different duplexes targeting the mRNA coding for
the third intracellullar loop of the receptor were formulated to the cationic lipid
iFect. The complex induced a decrease in the amount of NTS2 mRNA up to 70.4%
in the spinal cord and 85.5% in the DRG following daily intrathecal injections
of only 1 mg for two consecutive days. The behavioral validation of the down-
regulation consisted in the observation of the analgesic effect of a selective NTS2
pharmacological agonist at the tail-flick test, in control and treated rats. The
scrambled-treated group displayed increased latency to tail withdrawal as expected,
while DsiRNA-treated animals showed no analgesia following the intrathecal
injection of the pharmacological agonist, demonstrating the functional deletion of
the receptor. The knockdown lasted for 3 days following the last injection of
DsiRNA (Fig. 4a, b).
a ** * * iFect + mismatch
14
iFect + DsiRNAv1-1
12 iFect + JMV-431
10 iFect + DsiRNAv1-1
Latency (sec)
+ JMV-431
8
Day 1 Day 2 Day 3 Day 4
b *** iFect
14
iFect + mismatch
12 **
iFect + DsiRNAv2-5
Latency (sec)
10 iFect + DsiRNAv1-1
8 iFect + JMV-431
iFect + mismatch
6 + JMV-431
c Ctl saline n = 6
3
DsiRNAv2-5 +
JMV-431 n = 6
DsiRNA_NC1 + JMV-431 n=6
2
Pain Score
0
0 10 20 30 40 50 60
Time (min)
Fig. 4 Effects of NTS2 receptor knockdown using dicer-substrate siRNA on pain perception in
rats. (a) Two daily injections of NTS2-targeting DsiRNA induced a marked decrease in the
In the process to add data to this “proof of concept”, knockdown of NTS2 was
confirmed using the formalin test (Tetreault et al. 2009). This step toward chronic
pain assessment was capital to validate sustained pain in a context of RNAi
technology. The exercise was appropriate since Sarret’s team uncovered that the
initial transfection reagent induced undesired side effects in the formalin test. As
previously described, Transducin was selected to perform the formalin test ade-
quately without adverse effects. Using a similar injection protocol and a low dose of
5 mg, the NTS2 receptor was appropriately silenced and the selective NTS2
pharmacological agonist lost its efficiency to inhibit pain behavior during the late
phase of the formalin test (Fig. 4c).
This proof of concept was also reported later by another team looking at the
functional role of the b3 subunit of the Na+, K+-ATPase in the modulation of the
response to formalin in different mice strains (LaCroix-Fralish et al. 2009). In this
study, the hypothesis was that this pump was responsible for the differential pain
responses of two mouse strains to the formalin test. After the uptake of the
DsiRNAs was validated in DRG neurons, it was determined that the intrathecal
infusion of two different DsiRNAs targeting the b3 subunit abolished the difference
between the strains (Fig. 5a). The electrophysiological activity of extracted DRG
neurons in response to an inhibitor of this pump (ouabain) also confirmed the loss of
phenotype in the knockdown mice (Fig. 5b).
6 Conclusion and Perspectives
Management of chronic pain is a challenge, as the currently available drugs do

not target the multiple mechanisms underlying the generation and propagation of
pain. In this regard, RNA interference represents a promising approach to identify,
validate, and finally serve as a pain treatment. Use of this technology has rapidly
moved from in vitro studies to in vivo drug development programs. RNAi-based
drugs are already in clinical trials and it is hopeful that siRNA will receive FDA
approval in the future for pain therapy. However, there are a number of drawbacks
that we need to overcome before using siRNA candidates as a new class of
therapeutic drugs for pain disorders. These critical technical issues include
improvement of siRNA stability, consistency and predictability of expression
<
Fig. 4 (continued) analgesic effect of the selective agonist JMV-431. Indeed, the latency to
withdrawal at the tail flick test was not increased in the treated group, as opposed to the scramble
DsiRNA group, where the analgesic effect of the agonist is conserved. (b) Both DsiRNAs tested did
not modify the tail withdrawal latency compared with the vehicle i-Fect or scramble (mismatch)
DsiRNA treatments. Additionally, mismatch duplexes did not change the antinociceptive effects of
the JMV-431. (c) The same observations were made with the tonic pain formalin test. NTS2-
targeting DsiRNA treatment abolished the analgesic effect (decrease of the pain score in the late
phase of the test) of JMV-431, this time with the Transducin as a carrier (From: Dore-Savard 2008)
Fig. 5 Effects of b3 subunit a

of Na+, K+ ATPase on Early Phase A /J
formalin-induced licking/ C57BL /6J
biting behavioral differences 100
Time Spent Licking (s)

between two mice strains.
80 *
(a) Intrathecal treatment with *
two different DsiRNA
(siRNA-1 and siRNA-2) 60
duplexes targeting the b3
subunit abolished the 40
difference between strains by
increasing significantly the 20
licking/biting behaviors of A/
J mice compared with control 0
(saline), vehicle (i-Fect),
2
ec
lin
A-
A-
N
and scramble DsiRNA
i-F
R
Sa
N
si
R
(scrm siRNA) group.
si
si
rm
sc
(b) Consequently, Na+, K+
pump density of current is
modified in A/J mice
b
4.0
following b3 knockdown
(From: Lacroix-Fralish 2009) 3.5
Na+,K+ Pump Current
3.0
2.5
(pA/pF)
2.0
**
1.5
1.0
0.5 7 8
0.0
scrm siRNA siRNA-2
levels, confirmation of target gene knockdown, safe and efficient delivery as well as
elimination of toxicity and immunogenicity. DsiRNAs represent a technological
improvement in regards to their ability to circumvent the drawbacks while yielding
a high silencing potency for pain research. DsiRNA bioavailability to supraspinal
sites remains the most significant barrier to their use in a clinical pain setting. At
present, considerable effort is being made to develop appropriate vehicle to opti-
mize the delivery and stability for siRNA (Whitehead et al. 2009). Alternatively,
recent advances in blood–brain barrier disruption for delivery of chemotherapy in
the treatment of brain tumor patient may also be applied as a brain delivery strategy
to facilitate siRNA penetration (Bellavance et al. 2008). In summary, carefully
planned preclinical analyses of therapeutic RNAi are required to prevent premature
failures in human trials that could delay or dampen maturation of the field.
References
Akopian AN, Souslova V, England S et al (1999) The tetrodotoxin-resistant sodium channel SNS
has a specialized function in pain pathways. Nat Neurosci 2:541–548
Altier C, Dale CS, Kisilevsky AE et al (2007) Differential role of N-type calcium channel splice
isoforms in pain. J Neurosci 27:6363–6373
Amarzguioui M, Rossi J (2008) Principles of Dicer substrate (D-siRNA) design and function.
Methods Mol Biol 442:3–10
Amarzguioui M, Lundberg P, Cantin E et al (2006) Rational design and in vitro and in vivo
delivery of Dicer substrate siRNA. Nat Protoc 1:508–517
Ameres SL, Martinez J, Schroeder R (2007) Molecular basis for target RNA recognition and
cleavage by human RISC. Cell 130:101–112
Behlke MA (2008) Chemical modification of siRNAs for in vivo use. Oligonucleotides
18:305–319
Bellavance MA, Blanchette M, Fortin D (2008) Recent advances in blood-brain barrier disruption
as a CNS delivery strategy. AAPS J 10:166–177
Cenac N, Altier C, Chapman K et al (2008) Transient receptor potential vanilloid-4 has a major
role in visceral hypersensitivity symptoms. Gastroenterology 135:937–946, e931–932
Christoph T, Grunweller A, Mika J et al (2006) Silencing of vanilloid receptor TRPV1 by RNAi
reduces neuropathic and visceral pain in vivo. Biochem Biophys Res Commun 350:238–243
Christoph T, Bahrenberg G, De Vry J et al (2008) Investigation of TRPV1 loss-of-function
phenotypes in transgenic shRNA expressing and knockout mice. Mol Cell Neurosci
37:579–589
Collingwood MA, Rose SD, Huang L et al (2008) Chemical modification patterns compatible with
high potency dicer-substrate small interfering RNAs. Oligonucleotides 18:187–200
Deval E, Noel J, Lay N et al (2008) ASIC3, a sensor of acidic and primary inflammatory pain.
EMBO J 27:3047–3055
Dobner PR (2006) Neurotensin and pain modulation. Peptides 27:2405–2414
Dong X-W, Goregoaker S, Engler H et al (2007) Small interfering RNA-mediated selective
knockdown of Na(V)1.8 tetrodotoxin-resistant sodium channel reverses mechanical allodynia
in neuropathic rats. Neuroscience 146:812–821
Dore-Savard L, Roussy G, Dansereau MA et al (2008) Central delivery of Dicer-substrate siRNA:
a direct application for pain research. Mol Ther 16:1331–1339
Dorn G, Patel S, Wotherspoon G et al (2004) siRNA relieves chronic neuropathic pain. Nucleic
Acids Res 32:e49
Eguchi A, Dowdy SF (2009) siRNA delivery using peptide transduction domains. Trends Phar-
macol Sci 30:341–345
Eguchi A, Meade BR, Chang Y-C et al (2009) Efficient siRNA delivery into primary cells by a
peptide transduction domain-dsRNA binding domain fusion protein. Nat Biotechnol
27:567–571
Fan J, Wu X, Cao Z et al (2009) Up-regulation of anterior cingulate cortex NR2B receptors
contributes to visceral pain responses in rats. Gastroenterology 136(1732–1740):e1733
Gabra BH, Kessler FK, Ritter JK et al (2007) Decrease in N-methyl-D-aspartic acid receptor-NR2B
subunit levels by intrathecal short-hairpin RNA blocks group I metabotropic glutamate
receptor-mediated hyperalgesia. J Pharmacol Exp Ther 322:186–194
Ganju P, Hall J (2004) Potential applications of siRNA for pain therapy. Expert Opin Biol Ther
4:531–542
Garraway SM, Xu Q, Inturrisi CE (2007) Design and evaluation of small interfering RNAs that
target expression of the N-methyl-D-aspartate receptor NR1 subunit gene in the spinal cord
dorsal horn. J Pharmacol Exp Ther 322:982–988
Garraway SM, Xu Q, Inturrisi CE (2009) siRNA-mediated knockdown of the NR1 subunit gene of
the NMDA receptor attenuates formalin-induced pain behaviors in adult rats. J Pain
10:380–390
Hefner E, Clark K, Whitman C et al (2008) Increased potency and longevity of gene silencing
using validated Dicer substrates. J Biomol Tech 19:231–237
Hemmings-Mieszczak M, Dorn G, Natt FJ et al (2003) Independent combinatorial effect of
antisense oligonucleotides and RNAi-mediated specific inhibition of the recombinant rat
P2X3 receptor. Nucleic Acids Res 31:2117–2126
Hock J, Meister G (2008) The Argonaute protein family. Genome Biol 9:210
Howard KA, Paludan SR, Behlke MA et al (2009) Chitosan/siRNA nanoparticle-mediated TNF-
alpha knockdown in peritoneal macrophages for anti-inflammatory treatment in a murine
arthritis model. Mol Ther 17:162–168
Hutvagner G, Simard MJ, Mello CC et al (2004) Sequence-specific inhibition of small RNA
function. PLoS Biol 2:E98
Jain KK (2008) Gene therapy for pain. Expert Opin Biol Ther 8:1855–1866
Ji X (2008) The mechanism of RNase III action: how dicer dices. Curr Top Microbiol Immunol
320:99–116
Jinek M, Doudna JA (2009) A three-dimensional view of the molecular machinery of RNA
interference. Nature 457:405–412
John M, Constien R, Akinc A et al (2007) Effective RNAi-mediated gene silencing without
interruption of the endogenous microRNA pathway. Nature 449:745–747
Kariko K, Bhuyan P, Capodici J et al (2004) Small interfering RNAs mediate sequence-
independent gene suppression and induce immune activation by signaling through toll-like
receptor 3. J Immunol 172:6545–6549
Kasama S, Kawakubo M, Suzuki T et al (2007) RNA interference-mediated knock-down of
transient receptor potential vanilloid 1 prevents forepaw inflammatory hyperalgesia in rat.
Eur J Neurosci 25:2956–2963
Kim D-H, Behlke MA, Rose SD et al (2005) Synthetic dsRNA Dicer substrates enhance RNAi
potency and efficacy. Nat Biotechnol 23:222–226
Kortylewski M, Swiderski P, Herrmann A et al (2009) In vivo delivery of siRNA to immune cells
by conjugation to a TLR9 agonist enhances antitumor immune responses. Nat Biotechnol
27:925–932
Kubo T, Zhelev Z, Ohba H et al (2007) Modified 27-nt dsRNAs with dramatically enhanced
stability in serum and long-term RNAi activity. Oligonucleotides 17:445–464
Kubo T, Zhelev Z, Ohba H et al (2008) Chemically modified symmetric and asymmetric duplex
RNAs: an enhanced stability to nuclease degradation and gene silencing effect. Biochem
Biophys Res Commun 365:54–61
Kurreck J (2004) Antisense and RNA interference approaches to target validation in pain research.
Curr Opin Drug Discov Devel 7:179–187
Kurreck J (2009) RNA interference: from basic research to therapeutic applications. Angew Chem
Int Ed Engl 48:1378–1398
LaCroix-Fralish ML, Mo G, Smith SB et al (2009) The beta3 subunit of the Na+, K+-ATPase
mediates variable nociceptive sensitivity in the formalin test. Pain 144:294–302
Lima WF, Murray H, Nichols JG et al (2009) Human Dicer binds short single-strand and double-
strand RNA with high affinity and interacts with different regions of the nucleic acids. J Biol
Chem 284:2535–2548
Lin X, Ruan X, Anderson MG et al (2005) siRNA-mediated off-target gene silencing triggered
by a 7 nt complementation. Nucleic Acids Res 33:4527–4535
Lin CR, Amaya F, Barrett L et al (2006) Prostaglandin E2 receptor EP4 contributes to inflammatory
pain hypersensitivity. J Pharmacol Exp Ther 319:1096–1103
Liu TT, Bi HS, Lv SY, Wang XR, Yue SW (2010) Inhibition of the expression and function of
TRPV4 by RNA interference in dorsal root ganglion. Neurol Res 32(5):466–471
Luo MC, Zhang DQ, Ma SW et al (2005) An efficient intrathecal delivery of small interfering
RNA to the spinal cord and peripheral neurons. Mol Pain 1:29
Merkel OM, Beyerle A, Librizzi D et al (2009) Nonviral siRNA delivery to the lung: investigation
of PEG-PEI polyplexes and their in vivo performance. Mol Pharm 6:1246–1260
Ndong C, Pradhan A, Puma C et al (2009) Role of rat sensory neuron-specific receptor (rSNSR1)
in inflammatory pain: contribution of TRPV1 to SNSR signaling in the pain pathway. Pain
143:130–137
Rose SD, Kim D-H, Amarzguioui M et al (2005) Functional polarity is introduced by Dicer
processing of short substrate RNAs. Nucleic Acids Res 33:4140–4156
Roussy G, Dansereau MA, Dore-Savard L et al (2008) Spinal NTS1 receptors regulate nociceptive
signaling in a rat formalin tonic pain model. J Neurochem 105:1100–1114
Roussy G, Dansereau MA, Baudisson S et al (2009) Evidence for a role of NTS2 receptors in the
modulation of tonic pain sensitivity. Mol Pain 5:38
Sah DW (2006) Therapeutic potential of RNA interference for neurological disorders. Life Sci
79:1773–1780
Sah D, Guo W, Luo M et al (2006) Small interfering RNA targeting NaV1.8 alleviates experimen-
tally-induced chronic pain. Abstr Soc Neurosci 245.11/N10
Sarret P, Esdaile MJ, Perron A et al (2005) Potent spinal analgesia elicited through stimulation of
NTS2 neurotensin receptors. J Neurosci 25:8188
Saxena S, Jonsson ZO, Dutta A (2003) Small RNAs with imperfect match to endogenous mRNA
repress translation. Implications for off-target activity of small inhibitory RNA in mammalian
Soutschek J, Akinc A, Bramlage B et al (2004) Therapeutic silencing of an endogenous gene by
systemic administration of modified siRNAs. Nature 432:173–178
Tan PH, Yang LC, Shih HC et al (2005) Gene knockdown with intrathecal siRNA of NMDA
receptor NR2B subunit reduces formalin-induced nociception in the rat. Gene Ther 12
(1):59–66
Tang G (2005) siRNA and miRNA: an insight into RISCs. Trends Biochem Sci 30:106–114
Tetreault PG, Beaudet N, Dore-Savard L et al (2009). Functional inhibition of NTS2 receptors by
Dicer-substrate siRNA modulates tonic pain perception. 39th annual meeting of the Society for
Neuroscience, Chicago
van den Berg A, Mols J, Han J (2008) RISC-target interaction: cleavage and translational
suppression. Biochim Biophys Acta 1779:668–677
Vit JP, Ohara PT, Bhargava A et al (2008) Silencing the Kir4.1 potassium channel subunit in
satellite glial cells of the rat trigeminal ganglion results in pain-like behavior in the absence of
nerve injury. J Neurosci 28:4161–4171
Wang HW, Noland C, Siridechadilok B et al (2009) Structural insights into RNA processing by the
human RISC-loading complex. Nat Struct Mol Biol 16:1148–1153
Whitehead KA, Langer R, Anderson DG (2009) Knocking down barriers: advances in siRNA
delivery. Nat Rev Drug Discov 8:129–138
Zhang HM, Chen SR, Cai YQ et al (2009) Signaling mechanisms mediating muscarinic
enhancement of GABAergic synaptic transmission in the spinal cord. Neuroscience
158:1577–1588
Zhou J, Swiderski P, Li H et al (2009) Selection, characterization and application of new RNA
HIV gp 120 aptamers for facile delivery of Dicer substrate siRNAs into HIV infected cells.
RNAi Treatment of HIV-1 Infection
Karin J. von Eije and Ben Berkhout
Contents
1 The RNAi Pathway for Expression of Therapeutic siRNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
2 RNAi Gene Therapy Against HIV-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
3 Finding the Optimal Viral Target Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
4 Combinatorial RNAi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
5 Targeting Cellular Cofactors of HIV-1 Replication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6 Alternative Inhibitory RNA Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
7 Preclinical Test Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
8 Sequence-Specificity of RNAi Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
9 Safety Issues Raised in Clinical Trials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
10 Clinical Trials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Abstract RNA interference (RNAi) is a cellular mechanism that mediates sequence-

specific gene silencing by cleavage or translational inhibition of the targeted
mRNA. RNAi can be used as an antiviral approach to silence the human immuno-
deficiency virus type 1 (HIV-1). The first clinical trial using RNAi against HIV-1
in a lentiviral gene therapy setting was initiated in early 2008. In this chapter, we
will focus on the basic principles of such an RNAi-based gene therapy against
HIV-1. Subjects that will be covered include target site selection within the viral
RNA genome, viral escape, and therapeutic strategies to prevent this, such as
combinatorial RNAi approaches, systems available for multiplexing of RNAi
inhibitors, methods to deliver the antiviral RNAi molecules and gene therapy
protocols to achieve durable HIV-1 inhibition. We will also discuss several
K.J. von Eije and B. Berkhout (*)

Center for Infection and Immunity Amsterdam (CINIMA), Department of Medical Microbiology,
Laboratory of Experimental Virology, Academic Medical Center of the University of Amsterdam,
Amsterdam, The Netherlands
e-mail: b.berkhout@amc.uva.nl
192 K.J. von Eije and B. Berkhout
in vitro and in vivo test systems to evaluate the efficacy and safety of an RNAi
gene therapy.
Keywords HIV-1 RNAi Antiviral therapy Viral resistance Combinatorial

RNAi Lentiviral vector Gene therapy
1 The RNAi Pathway for Expression of Therapeutic siRNAs
The discovery of the RNAi mechanism and RNAi-mediated gene silencing in

mammalian cells triggered the development of RNAi-based therapies against a
wide variety of diseases, including cancer, neurological, autoimmune, and infec-
tious diseases (McCaffrey et al. 2003; Kapadia et al. 2003; Banerjea et al. 2003;
Ter Brake et al. 2009; Ding et al. 2003; Takeshita and Ochiya 2006; Davidson and
Paulson 2004). RNAi also holds promise for antiviral therapy or intracellular
vaccination against pathogenic viruses such as HIV-1. To properly evaluate
RNAi therapeutic approaches, including the risks involved by using the cellular
RNAi machinery, it is essential to understand the function of the RNAi mechanism.
We will first describe the natural microRNA (miRNA) pathway in some detail. It is
estimated that human cells express more than 500 miRNAs (http://microrna.sanger.
ac.uk). These miRNAs are important in the process of cell lineage development
(Ambros 2001; Carrington and Ambros 2003; Bartel 2004; Baehrecke 2003;
McManus 2004) by inhibiting gene expression at the posttranscriptional level
(translational repression or mRNA cleavage and degradation) (Ambros 2004).
The natural miRNA pathway uses RNA polymerase II or III to produce a
primary transcript or pri-miRNA that encodes the miRNA (Lee et al. 2002b).
Some miRNA genes are clustered and transcribed as a single polycistronic tran-
script. The pri-miRNA is processed into the pre-miRNA with 50 -monophospate and
30 -hydroxyl 2-nucleotide (nt) overhang (Bernstein et al. 2001) by the microproces-
sor complex that contains the RNase III-like enzyme Drosha and the dsRNA-
binding protein DGCR8/Pasha (Denli et al. 2004; Gregory et al. 2004; Han et al.
2004; Landthaler et al. 2004; Lee et al. 2003). The miRNAs encoded within introns
(mirtrons) are processed into pre-miRNAs through a distinct route that uses the
splicing machinery (Ruby et al. 2007; Chan and Slack 2007). The pre-miRNA is
formed in the nucleus and exported to the cytoplasm by Exportin-5 (Exp-5)
(Bohnsack et al. 2004; Lund et al. 2004; Yi et al. 2003). The RNAse III-like
endonuclease Dicer subsequently cleaves the base-paired stem approximately
22-basepairs (bp) away from its base, generating a two-nt overhang at the 30 -end
(Zhang et al. 2004). Dicer is associated with the TAR RNA binding protein
(TRBP), which is required to recruit Argonaute 2 (Ago2) (Chendrimada et al.
2005). The Ago2–RNA complex forms the minimal core of the RNA-induced
silencing complex (RISC) (Gregory et al. 2005; Maniataki and Mourelatos 2005).
RISC unwinds the miRNA and loads one RNA strand (guide strand) in the complex,
and the other strand gets degraded (passenger strand) (Hammond et al. 2001).
RNAi Treatment of HIV-1 Infection 193
One RNA strand gets preferentially incorporated into the complex (Tomari et al.
2004; Tomari and Zamore 2005).
In mammals, RNAi-mediated gene silencing is mainly elicited by translational
repression of the targeted mRNA (Ambros 2004). An important determinant is the
level of base pairing complementarity between the miRNA and the mRNA target,
leading to mRNA cleavage (perfect complementarity) or translational repression
(near-perfect complementarity) (Brennecke et al. 2003; Doench and Sharp 2004;
Lewis et al. 2003; Kiriakidou et al. 2004; Lai 2002). RISC typically forms com-
plexes when the “seed” region of the miRNA (50 end) that finds multiple target
sequences in the 30 untranslated region (30 UTR) of the mRNA. The number
of 30 UTR targets and their distance determines the silencing efficiency (Saetrom
et al. 2007). Most mammalian miRNAs anneal through imperfect base pairing
complementarity with the mRNA to cause translational repression, but at least
one case of perfect complementarity and mRNA cleavage is known in humans
(Yekta et al. 2004). Endonucleolytic cleavage of the targeted mRNA occurs
opposite of nucleotide position 10–11 in the miRNA, and the cleaved mRNA is
subsequently degraded.
In contrast to natural miRNAs, small interfering RNA (siRNA) with full base
pairing complementarity can direct mRNA cleavage with only a single target site
that can be located anywhere within the mRNA. Such artificial double-stranded (ds)
RNA can be produced by several methods. Synthetic mature siRNAs can be
transfected into cells (Elbashir et al. 2001), but short hairpin RNAs (shRNAs)
(Brummelkamp et al. 2002; Paddison et al. 2002) and artificial miRNAs are
expressed intracellularly from a transgene construct (Zeng et al. 2002). The natural
miRNA pathway can be instructed with the man-made inhibitors for therapeutic
downregulation of a specific mRNA. This therapeutic approach is relevant for
diseases caused by overexpression of a specific mRNA or to specifically target
the RNA genomes of invading microbes such as HIV-1.
Instruction of the cellular miRNA pathway with new siRNA specificity is
associated with certain risks. One potential problem is direct competition of the
artificial siRNAs with the endogenous siRNAs and/or saturation of the miRNA
pathway. Since the miRNA pathway is important in the control of cellular gene
expression, this can have unwanted side effects such as cell death, disturbances
in cell differentiation programs, or even cancer. Saturation of the miRNA pathway
can occur and lead to death when high doses of shRNAs were delivered by an
adeno-associated virus (AAV) vector in mice (Grimm et al. 2006; McBride et al.
2008; Boudreau et al. 2008; Vickers et al. 2007; Castanotto et al. 2007; Ter Brake
et al. 2009). Another potential problem is targeting of other mRNAs as the miRNAs
require only a seed sequence complementarity of 7–8 base pairs within the 30 UTR
of a given mRNA for function (Brennecke et al. 2005). Such “off-target” effects can
be elicited by the passenger or guide strand (Jackson et al. 2003; Fedorov et al.
2006; Jackson et al. 2006). Yet another problem relates to the induction of an
immune response by siRNAs and shRNAs (Bridge et al. 2003; Sledz et al. 2003),
which can be avoided by optimal design of the si/shRNA molecule (Marques and
Williams 2005).
2 RNAi Gene Therapy Against HIV-1
The goal of an RNAi-based gene therapy approach against HIV-1 is to protect the
cells of the immune system that are susceptible to HIV-1 infection. This includes
the CD4+ T cells, monocytes, macrophages, and dendritic cells. Such “intracellular
immunization” will prevent the depletion of these immune cells during disease
progression. Maintenance of the immune function should prevent opportunistic
infections and progression towards AIDS.
HIV-1 causes a chronic infection without the possibility of viral clearance. Thus,
a continuously active treatment regime is required. Repeated delivery of exogenous
siRNAs as anti-HIV therapy has been described (Kumar et al. 2008) in a humanized
immune system (HIS) mouse model. Effective virus inhibition was obtained in this
animal model, with a concomitant prevention of the loss of CD4+ T cells. However,
it is doubtful whether such an approach would be suitable in a patient setting, where
the prevention of viral escape requires the continuous presence of an effective
siRNA dose in all infected human cells that are located in many different body
compartments. Instead, we advocate the concept of continuous expression of anti-
HIV molecules obtained by a single transduction of susceptible cells with a
lentiviral vector. A lentiviral vector is based on the genome of HIV-1 itself. The
pathogenic genes are replaced by novel control and therapeutic sequences. The
lentiviral vector does infect the target cell and deposits the transgene, but cannot
replicate. A benefit of the lentiviral vector compared to other viral delivery methods
is that dividing and nondividing cell types can be transduced efficiently. Further-
more, the lentiviral vector is stably integrated into the host cell genome, thus
yielding permanent transduction (Naldini et al. 1996).
Some problems can be encountered when using lentiviruses to target HIV-1 with
RNAi, such as self-targeting of the transgene RNA by the shRNA expressed in the
producer cell, and targeting of HIV-derived sequences in the vector genome. These
problems and solutions have previously been discussed in detail by Ter Brake et al.
(Ter Brake and Berkhout 2007). There are other viral vector systems available for
delivering a therapeutic RNAi transgene, and these have been extensively discussed
by others (Nguyen et al. 2008; de Fougerolles 2008).
We depicted a possible gene therapy procedure for HIV-infected individuals
using the lentiviral vector in Fig. 1. Hematopoietic stem cells seed the different
lineages of immune cells in the blood and are therefore interesting candidates for an
ex vivo gene therapy followed by autolog transplantation. The lentiviral vector will
durably equip all hematopoietic stem cell-derived immune cells with the antiviral
arsenal. In the presence of HIV-1, one expects the preferential survival of the
shRNA-expressing immune cells over untreated cells, which will result in a gradual
increase in the percentage of protected cells. The treatment should result in partial
or complete reconstitution of the immune system, thus preventing HIV-1 infection
to progress towards AIDS. Ideally, a single gene therapy treatment should achieve a
durable effect because the transduced stem cells continue to generate immune cells
of the different lineages. Hematopoietic stem cells transduced with a retroviral
vector encoding an anti-HIV-1 ribozyme have already been used in clinical trials
Fig. 1 RNAi gene therapy for HIV-1. The HIV-1 infected patient that fails on regular antiretro-
viral therapy (1) could be offered the RNAi-based gene therapy with a lentiviral vector. The
lentiviral vector is produced in 293T cells (2) transfected with the lentiviral vector (e.g., JS1) and a
standard set of packaging plasmids (pRSV-Rev, pVSV-g and pSYNGP). The lentiviral vector will
produce viral genomes and the packaging plasmids will produce the proteins required to assemble
new viral particles. pVSV-g produces the Vesicular Stomatitis Virus glycoprotein that is used for
virus pseudotyping. Virus particles are collected after 2 or 3 days. The patient will undergo an
apheresis for the collection of hematopoietic stem cells after pre-treatment with granulocyte
colony stimulatory factor (G-CSF) that mobilizes these cells from the bone marrow into the
periphery (3). The hematopoietic stem cells will be purified and transduced with the therapeutic
lentiviral construct (4). This “intracellular immunization” with the antiviral shRNA will protect
these cells against HIV-1. Transduced cells will be infused back into the patient (5) and the
HIV-resistant immune cells will hopefully prevent disease progression towards AIDS (6)
(Amado et al. 2004; Mitsuyasu et al. 2009). These trials demonstrate the feasibility
and safety of the proposed stem cell approach. Another option is the treatment of
the mature CD4+ T-cell population, in which case repetitive gene therapy should be
applied because T cells have only a limited life span (Dropulic 2001).
Potent and sequence-specific HIV-1 inhibition has been reported with RNAi-
inducing reagents in cell culture infections. It soon became apparent that HIV-1 is
prone to viral escape in a mono-shRNA therapy (Boden et al. 2003; Das et al. 2004;
Nishitsuji et al. 2006; Sabariegos et al. 2006; Ter Brake et al. 2006; Unwalla et al.
2006; Westerhout et al. 2005), similar to single antiretroviral drug regimens. The
therapeutic vector used in a clinical trial should therefore always tackle the virus with
multiple inhibitors at the same time. Such a combinatorial RNAi attack can target the
virus (Ter Brake et al. 2008), host-encoded cofactors, or both. One could also
combine RNAi molecules with other RNA effector molecules such as decoys and
ribozymes (Li et al. 2005). Another elegant solution to avoid viral escape is the use of
the second generation shRNAs that specifically target viral escape variants (Brake
and Berkhout 2005). However, the relatively high number of viral escape routes
available to HIV-1 reduces the feasibility of this approach (von Eije et al. 2008).
3 Finding the Optimal Viral Target Sites
How to identify the optimal target sites for RNAi attack on the 9 kb HIV-1 RNA
genome? One option is the selection of targets in the early spliced mRNAs that
encode the early viral proteins Tat, Rev, and Nef. An early block in viral gene
expression will severely impact the expression of structural viral proteins and
consequently virion assembly. Alternatively, one could target HIV-1 genome
regions that are represented in all spliced viral mRNAs in the untranslated 50 -leader
and 30 -trailer domains (Muesing et al. 1985). Target RNA structure can block an
RNAi attack (Westerhout et al. 2005; Westerhout and Berkhout 2007). Thus,
targeting of “open” RNA domains is beneficial. In addition, the selection of targets
with highly conserved sequences seems important to allow inhibition of as many
virus strains as possible. Targeting of highly conserved genome regions may also
restrict the evolution of viral escape mutants because sequence variation may have
an impact on the viral fitness. An extensive shRNA screen against highly conserved
sequences of the HIV-1 genome has been performed, yielding approximately
20 candidate shRNAs (Ter Brake et al. 2006). Stable shRNA-expressing T cell
lines were infected with HIV-1, which yielded four durable shRNA inhibitors that
restricted virus replication for more than 100 days (von Eije et al. 2009). We and
others have identified effective shRNAs and siRNAs targeting regulatory HIV-1
sequences, e.g., in the long terminal repeat (LTR) and untranslated leader
RNA (Jacque et al. 2002; Ter Brake et al. 2006) and most viral genes: gag
(Chang et al. 2005; Novina et al. 2002; Park et al. 2002; Ter Brake et al. 2006),
pol (Chang et al. 2005; Berkhout and Brake 2008; Surabhi and Gaynor 2002), vif
(Jacque et al. 2002), tat (Ter Brake et al. 2006; Coburn and Cullen 2002; Lee et al.
2002a; Surabhi and Gaynor 2002), rev (Ter Brake et al. 2006; Coburn and Cullen
2002; Lee et al. 2002a), vpu (Chang et al. 2005), env (Park et al. 2002), and nef
(Jacque et al. 2002). Follow-up analyses should include prolonged culturing of
stably transduced T cells to score the impact on cell viability. Prolonged culturing
in the presence of HIV-1 can result in viral escape. The appearance of mutations in
the viral target demonstrates the exquisite sequence-specificity of RNAi action.
When only wild-type sequences are observed in break-through viruses, this is in

fact an indication of suboptimal inhibition (von Eije et al. 2008; von Eije et al.
2009). To address safety, off-targeting of human mRNAs by the antiviral shRNAs
can be evaluated (Jackson et al. 2003). Guidelines for the identification of shRNA
inhibitors against HIV-1 have recently been described (von Eije et al. 2009).
4 Combinatorial RNAi
A variety of strategies have been described for multiplexing of shRNA cassettes in

a single therapeutic vector. The multiple shRNA cassettes should use separate
polymerase III promoters or a combination of polymerase II and III promoters, as
repeat sequences should be avoided in the lentiviral vector (Ter Brake et al. 2008).
Multiplexed siRNAs can also be expressed from a single transcript. We developed
extended-shRNAs that are processed into two or maximally three functional
siRNAs (Liu et al. 2007). Another strategy uses long hairpin RNAs (lhRNAs) to
encode numerous siRNAs (Barichievy et al. 2007; Sano et al. 2008; Konstantinova
et al. 2006). A disadvantage of the lhRNA approach is that it is unknown whether the
produced siRNAs will be active inhibitors (Brake and Berkhout 2005). Polycistronic
miRNA transcripts have also been developed (Liu et al. 2008; Aagaard et al. 2008).
While various groups have reported toxicity of shRNAs (Grimm et al. 2006;
McBride et al. 2008; Castanotto et al. 2007; Vickers et al. 2007), this could be solved
by inserting the siRNA into a natural miRNA backbone (McBride et al. 2008).
Conditional expression of the siRNA molecules will increase the safety of a
therapeutic vector. For instance, one would like to avoid shRNA expression in
transduced hematopoietic stem cells that still have to undergo hematopoiesis, a
process that will be particularly sensitive to an altered RNAi machinery. Tissue-
specific miRNA expression has been described for the liver (Snyder et al. 2008).
Another option is the design of constructs that are induced by HIV-1 infection
(Unwalla et al. 2006). Selective expression in HIV-1 susceptible cells would be an
elegant way to restrict putative saturation and off-target effects. Another option is
the use of inducible gene expression systems such as the doxycycline-controlled
Tet system (Wiznerowicz and Trono 2003; Zhou et al. 2007). While shRNAs are
generally expressed from polymerase III promoters, miRNAs are expressed from
polymerase II promoters. These polymerase II systems are better equipped for
tissue-specific or drug-regulated expression.
5 Targeting Cellular Cofactors of HIV-1 Replication
An advantage of targeting host cell cofactors that are important for HIV-1
replication is the reduced chance of viral escape. Silencing of several cofactors
resulted in HIV-1 inhibition: nuclear factor kappa B (Surabhi and Gaynor 2002),
CD4 (Novina et al. 2002; Anderson et al. 2003a), CXCR4 (Martinez et al. 2002;
Anderson et al. 2003a; Anderson and Akkina 2005; Anderson et al. 2003b),
DDX-3 (Ishaq et al. 2008), LEDGF/p75 (Vandekerckhove et al. 2006), and
CCR5 (An et al. 2007a; Anderson et al. 2003a; Anderson and Akkina 2005).
CCR5 is a critical receptor for HIV-1 entry and a promising and well-studied
target. Individuals with the D-32 mutation in CCR5 are resistant to HIV-1 infec-
tion, yet are healthy. Only an increased risk for infection with the West Nile virus
has been reported (Lim et al. 2006). A potent shRNA targeting this host cell
factor has been developed (Anderson et al. 2003a; An et al. 2007a). CCR5-tropic
viruses are generally responsible for HIV-1 transmission to other individuals, but
the virus can also use the alternative CXCR4 receptor. When the CCR5 receptor
is downregulated, this will potentially set the stage for selection of CXCR4-tropic
HIV-1 variants, but this remains to be tested. Many cellular targets will obviously
not be good candidates for a gene therapy because they are essential for the cell
and the host. For example, CXCR4 is required for homing of hematopoietic stem
cells to the bone marrow and subsequent T-cell differentiation (Lapidot 2001).
High-throughput RNAi gene knockdown screens have recently identified many
candidate host cofactors of HIV-1 replication (Brass et al. 2008; Konig et al.
2008; Zhou et al. 2008). Identified cofactors should first be validated using
alternative assay systems to exclude false positives. Such studies should increase
the arsenal of cellular targets available for a combinatorial RNAi approach.
6 Alternative Inhibitory RNA Molecules
Other types of inhibitory RNA molecules can be used to target the virus. The
currently ongoing phase I clinical trial at the City of Hope uses a lentiviral vector
that encodes a TAR-decoy, CCR5-ribozyme, and a shRNA targeting the HIV-1
genome in the tat–rev region (Li et al. 2005). The TAR-decoy is a small nucleolar
RNA molecule that binds Tat, which will prevent the Tat–TAR interaction that is
essential for enhanced viral promoter activity (Lisziewicz et al. 1993). The CCR5
ribozyme cleaves the CCR5 mRNA to cause reduced CCR5 expression on the
cell surface (Sarver et al. 1990). Other anti-HIV-1 RNA molecules include
antisense transcripts (Chatterjee et al. 1992; Levine et al. 2006), decoys (Kohn
et al. 1999), ribozymes (Sarver et al. 1990), and aptamers (Symensma et al.
1996). A new addition to this arsenal is an antisense molecule that can elicit
transcriptional gene silencing of the viral LTR promoter (Weinberg et al. 2006).
The novel RNAu method is based on the expression of a modified U1 small
nuclear RNA that blocks polyadenylation of the targeted mRNA, which is
subsequently degraded (Abad et al. 2008). Comprehensive reviews on combina-
torial RNA approaches are available (Grimm and Kay 2007; Liu and Berkhout
2008).
7 Preclinical Test Systems
When potent antiviral shRNAs are successfully expressed from a single vector, one
can move to relevant preclinical models to critically assess the safety and efficacy.
A simple and efficient in vitro test system to measure the impact on cell viability is
to perform a co-culture of GFP+ transduced cells and non-treated cells (unpublished
results). A reduction over time in the percentage of GFP+ cells is an indication of
delayed cell growth and RNAi toxicity. Outgrowth of the transduced cells should
occur in the presence of HIV-1, again using simple FACS analysis to screen the
mixed cell culture.
The SIV-macaque model (Lackner and Veazey 2007), which has been used
extensively for vaccination studies, can be considered for testing of an anti-HIV-1
RNAi gene therapy, although it has several limitations. First, anti-HIV shRNAs
cannot easily be tested against SIV because of sequence dissimilarity, and the same
likely holds for the genes of cofactors in man versus macaque. Thus, the anti-HIV
shRNAs should either be converted into anti-SIV shRNAs, which may affect their
inhibitory power, or HIV-1 target sequences can be incorporated into the SIV
genome. Second, transduction of the HIV-based lentiviral vector is restricted by
TRIM5a in macaque cells (Stremlau et al. 2004). Third, macaque experiments are
rather expensive, and the number of animals that can be used is restricted. A
minimally modified simian-tropic (st) HIV-1 strain has recently been developed
that produces an acute viraemie and persistent infection in pig-tailed macaques
(Hatziioannou et al. 2009). In contrast to most infected humans, stHIV-1 infection
is controlled in the macaque model after several months.
Most of these limitations do not apply to HIS mouse models (Traggiai et al.
2004; Gimeno et al. 2004). All major human myeloid and lymphoid cellular
compartments develop and mature from input human stem cells in the most recent
HIS mouse (Legrand et al. 2006b; Shultz et al. 2007; Manz 2007). This model
provides access to in vivo and ex vivo experimentation on human T cells (Legrand
et al. 2006a). HIS mice can be infected by injection of the virus but also via rectal
and vaginal transmission routes. Infection results in viremia and the depletion of
human CD4+ cells as seen in the disease course of infected patients (Baenziger
et al. 2006; Zhang et al. 2007; Watanabe et al. 2007; Berges et al. 2006; Berges et al.
2008; An et al. 2007b). We used this model to test safety and efficacy of a lentiviral-
based gene therapy of human hematopoietic stem cells (Ter Brake et al. 2009).
These and other animal models have recently been reviewed (Goldstein 2008).
8 Sequence-Specificity of RNAi Action
An important lesson to be learned from various siRNA tests concerns the inclusion
of appropriate control experiments. Several studies on the inhibition of infections
and inflammation used a control siRNA that targets GFP. Results were in favor of a
therapeutic effect, but it turned out that the GFP control is a particular siRNA of low
immunogenicity compared to other shRNAs, including the therapeutic ones. Most
siRNAs trigger the TLR-7/8 interferon pathway, but the GFP siRNA control does
not (Robbins et al. 2008). Anyhow, the therapeutic effect was not elicited by
specific downregulation of the targeted mRNA.
Another lesson comes from a study on an siRNA therapeutic designed for the
eye against age-related macular degeneration. The siRNA exhibited a therapeutic
effect, but this was not elicited by the RNAi mechanism since the effector molecule
cannot penetrate cells. Instead, the clinical effect was obtained through TLR-3
signaling (Kleinman et al. 2008). Both examples illustrate the importance of
selecting the correct controls to ensure that one is looking at RNAi-specific effects.
For HIV-1 therapies that target the viral genome, exclusive specificity can be
demonstrated with escape variants with one or more point mutations in the target
sequence. The sequence-specificity of a particular RNAi effector molecule must be
demonstrated in vitro and in the animal model before moving to the clinical phase
(von Eije et al. 2008; Ter Brake et al. 2008; Ter Brake et al. 2009).
9 Safety Issues Raised in Clinical Trials
The first patient treated with a gene therapy in 1990 suffered from adenosine
deaminase deficiency, a form of severe combined immunodeficiency (SCID)
(Culliton 1990). In 1999, a patient died due to the administration of a gene therapy.
This patient was treated for a genetic liver disease – ornithine transcarbamylase
deficiency – and received an adenovirus treatment with the wild-type gene. He died
4 days later of a massive immune response, most likely triggered by the use of
the viral vector (Marshall 1999). A trial where patients with SCID received a
g-retroviral gene transfer with the wild-type interleukin-2 gene started in 2000.
Although this procedure improved the condition of all patients – a true success
(Hacein-Bey-Abina et al. 2002) – two patients developed a leukemia-like condition
of clonal lymphocyte proliferation. Both cases were caused by integration of the
retroviral vector near the promoter of the LMO2 proto-oncogene, leading to
enhanced expression of the LMO2 protein, which has a crucial role in hematopoie-
tic development (Hacein-Bey-Abina et al. 2003). More patients in this and a similar
trial subsequently developed leukemia-like conditions due to insertional oncogene-
sis. By now, more than 1,300 clinical trials involving a gene therapy have been
performed (Edelstein et al. 2007). From these clinical trials, lessons can be learned
for future improvement of gene therapies. For instance, retroviral vectors have been
replaced by lentiviral vectors, which are much more safe because all transcriptional
enhancer motifs have been removed (“self-inactivating” design (Bushman et al.
2005)) and because these vectors tend to integrate in genes, and not near the
promoter region. In addition, experiments with a lentiviral vector and hematopoie-
tic stem cells in tumor-prone mice did not, in contrast to the retroviral vector, show
signs of insertional oncogenesis (Montini et al. 2006). Other safety and regulatory
issues concerning lentiviral vectors are addressed in a comprehensive review based

on the expertise gained in the first lentiviral trial (Manilla et al. 2005).
10 Clinical Trials
An overview of gene therapy trials for HIV-1 is available (Rossi et al. 2007). While
positive in vitro results were obtained for these antiviral gene therapies, the clinical
trials failed to demonstrate a therapeutic benefit. In studies where T cells or
hematopoietic stem cells were treated with the original retroviral vectors, one of
the bottlenecks was effective gene delivery to a clinically relevant number of cells
(Mitsuyasu et al. 2009). This problem has disappeared with the use of lentiviral
vectors that have a much higher transduction efficiency in a variety of cell types. In
addition, many of the previously used inhibitory RNA molecules seem suboptimal
when compared to antiviral shRNAs. RNAi is therefore a promising candidate for
development of a future anti-HIV-1 gene therapy.
The first clinical trial with a lentiviral vector was in fact directed against HIV-1
by expression of an extended antisense transcript against the viral RNA genome.
Persistent in vivo expression of the therapeutic antisense molecule was documented
by the VirXsys company (Levine et al. 2006). In addition, vector integration sites in
blood cells revealed a preference for gene-rich regions, which is typical for a
lentivirus, and no signs of insertional oncogenesis were observed. Another lenti-
viral vector trial was initiated recently in children with X-linked adrenoleukody-
strophy. Promising results were presented by Nathalie Cartier-Lacave (Necker
Hospital, Paris) at the 2008 meeting of the European Society for Gene and Cell
Therapy (ESGCT). Hematopoietic stem cells were effectively transduced ex vivo,
persistence of the modified cells was documented upon reinjection, and clinical
improvement was reported. Another anti-HIV gene therapy trial with a triple RNA
payload (ribozyme, decoy, shRNA) was initiated at the City of Hope by the team of
John Rossi. Persistent gene marking and consistent shRNA expression was docu-
mented in blood cells (John Rossi, personal communication).
11 Conclusion
We reviewed the current status of the development of an RNAi-based gene therapy

to control HIV-1 infection and AIDS disease progression. Overall, an RNAi-based
gene therapy against HIV-1 seems to be a promising candidate for a durable
antiviral treatment, especially for a minority patient group for which the treatment
options are exhausted. The potent and sequence-specific inhibition of HIV-1 with
RNAi forms the corner stone for such a therapy. The superior transduction of
hematopoietic stem cells with lentiviral vectors provides the means to deliver the
transgene. The availability of several lentiviral production facilities is another
promising development in the field. We are currently testing a candidate clinical
vector that encodes four antiviral shRNAs to evaluate its safety and efficacy.
This vector yielded very potent antiviral effects in prolonged in vitro cell cultures
(Ter Brake et al. 2008). Safety and efficacy are currently addressed in a humanized
mouse model (Ter Brake et al. 2009), and we expect to initiate a clinical trial within
2 years.
Acknowledgments RNAi research in the Berkhout lab is sponsored by ZonMw (Translational

gene therapy program).
References
Aagaard LA, Zhang J, von Eije KJ et al (2008) Engineering and optimization of the miR-106b
cluster for ectopic expression of multiplexed anti-HIV RNAs. Gene Ther 15:1536–1549
Abad X, Vera M, Jung SP et al (2008) Requirements for gene silencing mediated by U1 snRNA
binding to a target sequence. Nucleic Acids Res 36:2338–2352
Amado RG, Mitsuyasu RT, Rosenblatt JD et al (2004) Anti-human immunodeficiency virus
hematopoietic progenitor cell-delivered ribozyme in a phase I study: myeloid and lymphoid
reconstitution in human immunodeficiency virus type-1-infected patients. Hum Gene Ther
15:251–262
Ambros V (2001) microRNAs: tiny regulators with great potential. Cell 107:823–826
Ambros V (2004) The functions of animal microRNAs. Nature 431:350–355
An DS, Donahue RE, Kamata M et al (2007a) Stable reduction of CCR5 by RNAi through
hematopoietic stem cell transplant in non-human primates. Proc Natl Acad Sci USA
104:13110–13115
An DS, Poon B, Ho Tsong FR et al (2007b) Use of a novel chimeric mouse model with a
functionally active human immune system to study human immunodeficiency virus type 1
infection. Clin Vaccine Immunol 14:391–396
Anderson J, Akkina R (2005) CXCR4 and CCR5 shRNA transgenic CD34+ cell derived macro-
phages are functionally normal and resist HIV-1 infection. Retrovirology 2:53
Anderson J, Banerjea A, Akkina R (2003a) Bispecific short hairpin siRNA constructs targeted to
CD4, CXCR4, and CCR5 confer HIV-1 resistance. Oligonucleotides 13:303–312
Anderson J, Banerjea A, Planelles V, Akkina R (2003b) Potent suppression of HIV type 1 infection
by a short hairpin anti-CXCR4 siRNA. AIDS Res Hum Retrovir 19:699–706
Baehrecke EH (2003) miRNAs: micro managers of programmed cell death. Curr Biol
13:R473–R475
Baenziger S, Tussiwand R, Schlaepfer E et al (2006) Disseminated and sustained HIV infection in
CD34+ cord blood cell-transplanted Rag2-/-gamma c-/-mice. Proc Natl Acad Sci USA
103:15951–15956
Banerjea A, Li MJ, Bauer G et al (2003) Inhibition of HIV-1 by lentiviral vector-transduced
siRNAs in T lymphocytes differentiated in SCID-hu mice and CD34+ progenitor cell-derived
macrophages. Mol Ther 8:62–71
Barichievy S, Saayman S, von Eije KJ et al (2007) The inhibitory efficacy of RNA POL
III-expressed long hairpin RNAs targeted to untranslated regions of the HIV-1 50 long terminal
repeat. Oligonucleotides 17:419–431
Berges BK, Akkina SR, Folkvord JM et al (2008) Mucosal transmission of R5 and X4 tropic
HIV-1 via vaginal and rectal routes in humanized Rag2(-/-)gammac(-/-) (RAG-hu) mice.
Virology 373:342–351
Berges BK, Wheat WH, Palmer BE et al (2006) HIV-1 infection and CD4 T cell depletion in the
humanized Rag2-/-gamma c-/-(RAG-hu) mouse model. Retrovirology 3:76
Berkhout B, Brake Ot (2008) Towards an RNAi-based gene therapy. BIOforum Europe 4:35–37
Boden D, Pusch O, Lee F et al (2003) Human immunodeficiency virus type 1 escape from RNA
interference. J Virol 77:11531–11535
Bohnsack MT, Czaplinski K, Gorlich D (2004) Exportin 5 is a RanGTP-dependent dsRNA-
binding protein that mediates nuclear export of pre-miRNAs. RNA 10:185–191
Boudreau RL, Monteys AM, Davidson BL (2008) Minimizing variables among hairpin-based
RNAi vectors reveals the potency of shRNAs. RNA 14:1834–1844
Brake Ot, Berkhout B (2005) A novel approach for inhibition of HIV-1 by RNA interference:
counteracting viral escape with a second generation of siRNAs. J RNAi Gene Silencing
1:56–65
Brass AL, Dykxhoorn DM, Benita Y et al (2008) Identification of host proteins required for HIV
infection through a functional genomic screen. Science 319:921–926
Brennecke J, Hipfner DR, Stark A et al (2003) Bantam encodes a developmentally regulated
microRNA that controls cell proliferation and regulates the proapoptotic gene hid in
Drosophila. Cell 113:25–36
Brennecke J, Stark A, Russell RB, et al (2005) Principles of microRNA-target recognition. PLoS
Biol 3:e85
Brummelkamp TR, Bernards R, Agami R (2002) A system for stable expression of short
interfering RNAs in mammalian cells. Science 296:550–553
Bushman F, Lewinski M, Ciuffi A et al (2005) Genome-wide analysis of retroviral DNA
integration. Nat Rev Microbiol 3:848–858
Carrington JC, Ambros V (2003) Role of microRNAs in plant and animal development. Science
301:336–338
35:5154–5164
Chan SP, Slack FJ (2007) And now introducing mammalian mirtrons. Dev Cell 13:605–607
Chang LJ, Liu X, He J (2005) Lentiviral siRNAs targeting multiple highly conserved RNA
sequences of human immunodeficiency virus type 1. Gene Ther 12:1133–1144
Chatterjee S, Johnson PR, Wong KK Jr (1992) Dual-target inhibition of HIV-1 in vitro by means of
an adeno- associated virus antisense vector. Science 258:1485–1488
Chendrimada TP, Gregory RI, Kumaraswamy E et al (2005) TRBP recruits the dicer complex to
Coburn GA, Cullen BR (2002) Potent and specific inhibition of human immunodeficiency virus
type 1 replication by RNA interference. J Virol 76:9225–9231
Culliton BJ (1990) Gene therapy begins. Science 249:1372
Das AT, Brummelkamp TR, Westerhout EM et al (2004) Human immunodeficiency virus type 1
escapes from RNA interference-mediated inhibition. J Virol 78:2601–2605
Davidson BL, Paulson HL (2004) Molecular medicine for the brain: silencing of disease genes
with RNA interference. Lancet Neurol 3:145–149
de Fougerolles AR (2008) Delivery vehicles for small interfering RNA in vivo. Hum Gene Ther
19:125–132
microprocessor complex. Nature 432:231–235
Ding H, Schwarz DS, Keene A et al (2003) Selective silencing by RNAi of a dominant allele that
causes amyotrophic lateral sclerosis. Aging Cell 2:209–217
Dropulic B (2001) Lentivirus in the clinic. Mol Ther 4:511–512
Edelstein ML, Abedi MR, Wixon J (2007) Gene therapy clinical trials worldwide to 2007 – an
update. J Gene Med 9:833–842
Elbashir SM, Harborth J, Lendeckel W et al (2001) Duplexes of 21-nucleotide RNAs mediate
Fedorov Y, Anderson EM, Birmingham A et al (2006) Off-target effects by siRNA can induce
toxic phenotype. RNA 12:1188–1196
Gimeno R, Weijer K, Voordouw A et al (2004) Monitoring the effect of gene silencing by
RNA interference in human CD34+ cells injected into newborn RAG2-/- gammac-/- mice:
functional inactivation of p53 in developing T cells. Blood 104:3886–3893
Goldstein H (2008) Summary of presentations at the NIH/NIAID new humanized rodent models
2007 workshop. AIDS Res Ther 5:3
Grimm D, Kay MA (2007) Combinatorial RNAi: a winning strategy for the race against evolving
targets? Mol Ther 15:878–888
Hacein-Bey-Abina S, Le Deist F, Carlier F et al (2002) Sustained correction of X-linked severe
combined immunodeficiency by ex vivo gene therapy. N Engl J Med 346:1185–1193
Hacein-Bey-Abina S, von Kalle C, Schmidt M et al (2003) LMO2-associated clonal T cell
proliferation in two patients after gene therapy for SCID-X1. Science 302:415–419
Hammond SM, Boettcher S, Caudy AA et al (2001) Argonaute2, a link between genetic and
biochemical analyses of RNAi. Science 293:1146–1150
Han J, Lee Y, Yeom KH et al (2004) The Drosha–DGCR8 complex in primary microRNA
processing. Genes Dev 18:3016–3027
Hatziioannou T, Ambrose Z, Chung NP et al (2009) A macaque model of HIV-1 infection. Proc
Ishaq M, Hu J, Wu X et al (2008) Knockdown of cellular RNA helicase DDX3 by short hairpin
RNAs suppresses HIV-1 viral replication without inducing apoptosis. Mol Biotechnol
39:231–238
Jackson AL, Burchard J, Schelter J et al (2006) Widespread siRNA “off-target” transcript
silencing mediated by seed region sequence complementarity. RNA 12:1179–1187
Jacque JM, Triques K, Stevenson M (2002) Modulation of HIV-1 replication by RNA interference.
Nature 418:435–438
Kapadia SB, Brideau-Andersen A, Chisari FV (2003) Interference of hepatitis C virus RNA
replication by short interfering RNAs. Proc Natl Acad Sci USA 100:2014–2018
Kiriakidou M, Nelson PT, Kouranov A et al (2004) A combined computational-experimental
approach predicts human microRNA targets. Genes Dev 18:1165–1178
Kohn DB, Bauer G, Rice CR et al (1999) A clinical trial of retroviral-mediated transfer of a
rev-responsive element decoy gene into CD34(+) cells from the bone marrow of human
immunodeficiency virus-1-infected children. Blood 94:368–371
Konig R, Zhou Y, Elleder D et al (2008) Global analysis of host-pathogen interactions that regulate
early-stage HIV-1 replication. Cell 135:49–60
Konstantinova P, de Vries W, Haasnoot J et al (2006) Inhibition of human immunodeficiency virus

type 1 by RNA interference using long-hairpin RNA. Gene Ther 13:1403–1413
Kumar P, Ban HS, Kim SS et al (2008) T cell-specific siRNA delivery suppresses HIV-1 infection
in humanized mice. Cell 134:577–586
Lackner AA, Veazey RS (2007) Current concepts in AIDS pathogenesis: insights from the
SIV/macaque model. Annu Rev Med 58:461–476
Lai EC (2002) Micro RNAs are complementary to 30 UTR sequence motifs that mediate negative
post-transcriptional regulation. Nat Genet 30:363–364
Landthaler M, Yalcin A, Tuschl T (2004) The human DiGeorge syndrome critical region gene 8
and Its D. melanogaster homolog are required for miRNA biogenesis. Curr Biol 14:2162–2167
Lapidot T (2001) Mechanism of human stem cell migration and repopulation of NOD/SCID and
B2mnull NOD/SCID mice. The role of SDF-1/CXCR4 interactions. Ann N Y Acad Sci
938:83–95
Lee NS, Dohjima T, Bauer G et al (2002a) Expression of small interfering RNAs targeted against
HIV-1 rev transcripts in human cells. Nat Biotechnol 20:500–505
Lee Y, Ahn C, Han J et al (2003) The nuclear RNase III Drosha initiates microRNA processing.
Nature 425:415–419
Lee Y, Jeon K, Lee JT et al (2002b) MicroRNA maturation: stepwise processing and subcellular
localization. EMBO J 21:4663–4670
Legrand N, Cupedo T, van Lent AU et al (2006a) Transient accumulation of human mature
thymocytes and regulatory T cells with CD28 superagonist in “human immune system” Rag2
(-/-)gammac(-/-) mice. Blood 108:238–245
Legrand N, Weijer K, Spits H (2006b) Experimental models to study development and function of
the human immune system in vivo. J Immunol 176:2053–2058
Levine BL, Humeau LM, Boyer J et al (2006) Gene transfer in humans using a conditionally
replicating lentiviral vector. Proc Natl Acad Sci USA 103:17372–17377
Lewis BP, Shih IH, Jones-Rhoades MW et al (2003) Prediction of mammalian microRNA targets.
Cell 115:787–798
Li MJ, Kim J, Li S et al (2005) Long-term inhibition of HIV-1 infection in primary hematopoietic
cells by lentiviral vector delivery of a triple combination of anti-HIV shRNA, anti-CCR5
ribozyme, and a nucleolar-localizing TAR decoy. Mol Ther 12:900–909
Lim JK, Glass WG, McDermott DH et al (2006) CCR5: no longer a “good for nothing”
gene–chemokine control of West Nile virus infection. Trends Immunol 27:308–312
Lisziewicz J, Sun D, Smythe J et al (1993) Inhibition of human immunodeficiency virus type 1
replication by regulated expression of a polymeric Tat activation response RNA decoy as a
strategy for gene therapy in AIDS. Proc Natl Acad Sci USA 90:8000–8004
Liu YP, Berkhout B (2008) Combinatorial RNAi strategies against HIV-1 and other escape-prone
viruses. Int J BioSci Technol 1:1–10
Liu YP, Haasnoot J, Berkhout B (2007) Design of extended short hairpin RNAs for HIV-1
inhibition. Nucleic Acids Res 35:5683–5693
Liu YP, Haasnoot J, Ter Brake O et al (2008) Inhibition of HIV-1 by multiple siRNAs expressed
from a single microRNA polycistron. Nucleic Acids Res 36:2811–2824
Lund E, Guttinger S, Calado A et al (2004) Nuclear export of microRNA precursors. Science
303:95–98
Manilla P, Rebello T, Afable C et al (2005) Regulatory considerations for novel gene therapy
products: a review of the process leading to the first clinical lentiviral vector. Hum Gene Ther
16:17–25
Manz MG (2007) Human-hemato-lymphoid-system mice: opportunities and challenges. Immunity
26:537–541
Marques JT, Williams BR (2005) Activation of the mammalian immune system by siRNAs.
Nat Biotechnol 23:1399–1405
Marshall E (1999) Gene therapy death prompts review of adenovirus vector. Science
286:2244–2245
Martinez J, Patkaniowska A, Urlaub H et al (2002) Single-stranded antisense siRNAs guide target
RNA cleavage in RNAi. Cell 110:563–574
USA 105:5868–5873
McCaffrey AP, Nakai H, Pandey K et al (2003) Inhibition of hepatitis B virus in mice by RNA
interference. Nat Biotechnol 21:639–644
McManus MT (2004) Small RNAs and immunity. Immunity 21:747–756
Mitsuyasu RT, Merigan TC, Carr A et al (2009) Phase 2 gene therapy trial of an anti-HIV
ribozyme in autologous CD34(+) cells. Nat Med 15:285–292
Montini E, Cesana D, Schmidt M et al (2006) Hematopoietic stem cell gene transfer in a tumor-
prone mouse model uncovers low genotoxicity of lentiviral vector integration. Nat Biotechnol
24:687–696
Muesing MA, Smith DH, Cabradilla CD et al (1985) Nucleic acid structure and expression of the
human AIDS/lymphadenopathy retrovirus. Nature 313:450–458
Naldini L, Blomer U, Gallay P et al (1996) In vivo gene delivery and stable transduction of
nondividing cells by a lentiviral vector. Science 272:263–267
Nguyen T, Menocal EM, Harborth J et al (2008) RNAi therapeutics: an update on delivery. Curr
Opin Mol Ther 10:158–167
Nishitsuji H, Kohara M, Kannagi M et al (2006) Effective suppression of human immunodefi-
ciency virus type 1 through a combination of short- or long-hairpin RNAs targeting essential
sequences for retroviral integration. J Virol 80:7658–7666
Novina CD, Murray MF, Dykxhoorn DM et al (2002) siRNA-directed inhibition of HIV-1
infection. Nat Med 8:681–686
Park WS, Miyano-Kurosaki N, Hayafune M et al (2002) Prevention of HIV-1 infection in human
peripheral blood mononuclear cells by specific RNA interference. Nucleic Acids Res
30:4830–4835
Robbins M, Judge A, Ambegia E et al (2008) Misinterpreting the therapeutic effects of siRNA
caused by immune stimulation. Hum Gene Ther 19(10):991–999
Rossi JJ, June CH, Kohn DB (2007) Genetic therapies against HIV. Nat Biotechnol 25:1444–1454
Nature 448:83–86
Sabariegos R, Gimenez-Barcons M, Tapia N et al (2006) Sequence homology required by human
immunodeficiency virus type 1 to escape from short interfering RNAs. J Virol 80:571–577
Saetrom P, Heale BS, Snove O Jr et al (2007) Distance constraints between microRNA target sites
dictate efficacy and cooperativity. Nucleic Acids Res 35:2333–2342
Sano M, Li H, Nakanishi M et al (2008) Expression of long anti-HIV-1 hairpin RNAs for the
generation of multiple siRNAs: advantages and limitations. Mol Ther 16:170–177
Sarver N, Cantin EM, Chang PS et al (1990) Ribozymes as potential anti-HIV-1 therapeutic
agents. Science 247:1222–1225
Shultz LD, Ishikawa F, Greiner DL (2007) Humanized mice in translational biomedical research.
Nat Rev Immunol 7:118–130
Snyder LL, Esser JM, Pachuk CJ et al (2008) Vector design for liver-specific expression of
multiple interfering RNAs that target hepatitis B virus transcripts. Antiviral Res 80:36–44
Stremlau M, Owens CM, Perron MJ et al (2004) The cytoplasmic body component TRIM5alpha
restricts HIV-1 infection in old world monkeys. Nature 427:848–853
Surabhi RM, Gaynor RB (2002) RNA interference directed against viral and cellular targets
inhibits human immunodeficiency virus type 1 replication. J Virol 76:12963–12973
Symensma TL, Giver L, Zapp M et al (1996) RNA aptamers selected to bind human immunodefi-
ciency virus type 1 Rev in vitro are Rev responsive in vivo. J Virol 70:179–187
Takeshita F, Ochiya T (2006) Therapeutic potential of RNA interference against cancer. Cancer
Sci 97:689–696
Ter Brake O, ’t Hooft K, Liu YP et al (2008) Lentiviral vector design for multiple shRNA
expression and durable HIV-1 Inhibition. Mol Ther 16:557–564
Ter Brake O, Berkhout B (2007) Lentiviral vectors that carry anti-HIV shRNAs: problems and
solutions. J Gene Med 9:743–750
Ter Brake O, Konstantinova P, Ceylan M et al (2006) Silencing of HIV-1 with RNA interference: a
multiple shRNA approach. Mol Ther 14:883–892
Ter Brake O, Legrand N, von Eije KJ et al (2009) Evaluation of safety and efficacy of RNAi
against HIV-1 in the human immune system (Rag-2(-/-)(c)(-/-)) mouse model. Gene Ther
16:148–153
Tomari Y, Matranga C, Haley B et al (2004) A protein sensor for siRNA asymmetry. Science
306:1377–1380
Tomari Y, Zamore PD (2005) Perspective: machines for RNAi. Genes Dev 19:517–529
Traggiai E, Chicha L, Mazzucchelli L et al (2004) Development of a human adaptive immune
system in cord blood cell-transplanted mice. Science 304:104–107
Unwalla HJ, Li HT, Bahner I et al (2006) Novel Pol II fusion promoter directs human immunode-
ficiency virus type 1-inducible coexpression of a short hairpin RNA and protein. J Virol
80:1863–1873
Vandekerckhove L, Christ F, Van Maele B et al (2006) Transient and stable knockdown of
the integrase cofactor LEDGF/p75 reveals its role in the replication cycle of human immuno-
deficiency virus. J Virol 80:1886–1896
Vickers TA, Lima WF, Nichols JG et al (2007) Reduced levels of Ago2 expression result in
increased siRNA competition in mammalian cells. Nucleic Acids Res 35:6598–6610
von Eije KJ, Ter Brake O, Berkhout B (2008) Human immunodeficiency virus type 1 escape is
restricted when conserved genome sequences are targeted by RNA interference. J Virol
82:2895–2903
von Eije KJ, Ter BO, Berkhout B (2009) Stringent testing identifies highly potent and escape-proof
anti-HIV short hairpin RNAs. J Gene Med 11:459–467
Watanabe S, Terashima K, Ohta S et al (2007) Hematopoietic stem cell-engrafted NOD/SCID/
IL2Rgamma null mice develop human lymphoid systems and induce long-lasting HIV-1
infection with specific humoral immune responses. Blood 109:212–218
Weinberg MS, Villeneuve LM, Ehsani A et al (2006) The antisense strand of small interfering
RNAs directs histone methylation and transcriptional gene silencing in human cells. RNA
12:256–262
Westerhout EM, Berkhout B (2007) A systematic analysis of the effect of target RNA structure on
RNA interference. Nucleic Acids Res 35:4322–4330
Westerhout EM, Ooms M, Vink M et al (2005) HIV-1 can escape from RNA interference by
evolving an alternative structure in its RNA genome. Nucleic Acids Res 33:796–804
Wiznerowicz M, Trono D (2003) Conditional suppression of cellular genes: lentivirus vector-
mediated drug-inducible RNA interference. J Virol 77:8957–8961
Yekta S, Shih IH, Bartel DP (2004) MicroRNA-directed cleavage of HOXB8 mRNA. Science
304:594–596
Yi R, Qin Y, Macara IG et al (2003) Exportin-5 mediates the nuclear export of pre-microRNAs
and short hairpin RNAs. Genes Dev 17:3011–3016
Zhang H, Kolb FA, Jaskiewicz L et al (2004) Single processing center models for human Dicer and
bacterial RNase III. Cell 118:57–68
Zhang L, Kovalev GI, Su L (2007) HIV-1 infection and pathogenesis in a novel humanized mouse
model. Blood 109:2978–2981
Zhou H, Xu M, Huang Q et al (2008) Genome-scale RNAi screen for host factors required for HIV
replication. Cell Host Microbe 4:495–504
Zhou X, Symons J, Hoppes R et al (2007) Improved single-chain transactivators of the Tet-On
gene expression system. BMC Biotechnol 7:6
Application of RNA Interference to Treat
Conditions Associated with Dysregulation
of Transient Receptor Potential Vanilloid 1
Channel
Vickram Ramkumar, Debashree Mukherjea, Sarvesh Jajoo, Tejbeer Kaur,

and Leonard P. Rybak
Contents
1 TRPV1 Ion Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
2 Neuropathic Pain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
2.1 Current Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
2.2 TRPV1 and Neuropathic Pain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
2.3 Diabetic Peripheral Neuropathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
3 Drug-Induced Hearing Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
3.1 TRPV1 and Cisplatin Ototoxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
3.2 Utility of RNAi in Treating Cisplatin Ototoxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
4 Other Potential Uses of RNAi Targeting TRPV1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
4.1 Inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
4.2 Arthritis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
4.3 Cystitis and Bladder Hyperactivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
4.4 Cancer Pain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
4.5 Obesity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Abstract Transient receptor potential vanilloid 1 (TRPV1) is a member of the

transient receptor potential (TRP) family of proteins and is most notably the target
of capsaicin, the active ingredient of “hot” pepper such as jalapeño and habanero. The
V. Ramkumar (*), S. Jajoo, and T. Kaur

Department of Pharmacology, Southern Illinois University School of Medicine, PO Box 19629,
e-mail: vramkumar@siumed.edu
D. Mukherjea
Department of Surgery, Southern Illinois University School of Medicine, PO Box 19629, Spring-
field, IL 62794-9629, USA
L.P. Rybak
Department of Pharmacology, Southern Illinois University School of Medicine, PO Box 19629,
Department of Surgery, Southern Illinois University School of Medicine, PO Box 19629, Spring-
field, IL 62794-9629, USA
210 V. Ramkumar et al.
channel is expressed primarily in small diameter neurons (Ad and C fibers) within
sensory ganglia comprising the pain pathway but expression is observed in larger
diameter neurons under conditions of inflammation. TRPV1 is a nonselective cation
channel, which responds to heat (activation threshold 43 C) and transmits pain
sensations in response to noxious heat. It is also expressed in nonneuronal tissue such
as keratinocytes, bladder uroepithelium, knee joints, the gastrointestinal tract and the
cochlea, suggesting additional physiological roles of channel activation. Several
disease states associated with dysregulation of TRPV1 include visceral and peripheral
inflammatory pain observed in irritable bowel disease, bone cancers, arthritis and
bladder inflammation. Other conditions include diabetic peripheral neuropathy, obe-
sity, and drug-induced hearing loss. Animal studies have shown beneficial effects of
targeting TRPV1, using both agonist and antagonist drugs, in the treatment of some of
these conditions. However, the propensity of TRPV1 antagonists to produce hyper-
thermia could limit their future application. Such problems might be resolved by the
focal administration of short interfering (si) RNA to the affected organ in order to
reduce TRPV1 expression. While demonstration of TRPV1 knockdown by RNA
interference (RNAi) is widely used in vitro, results obtained from a limited number
of in vivo studies suggest that RNAi could be applied to the treatment of diseases
associated with TRPV1 hyperactivity. To date, RNAi has proven beneficial in reduc-
ing inflammatory pain and in treating hearing loss associated with cisplatin chemo-
therapy. This review will focus on the current progress of RNAi in the treatment of
diseases associated with TRPV1 dysfunction, discuss potential future applications of
this technology, and highlight factors that could affect its use clinically.
Keywords TRPV1 siRNA Neuropathic pain Inflammation Hearing loss

Neuropathic pain Obesity Cystitis
Abbreviations
TRPV1 Transient receptor potential vanilloid 1

CGRP Calcitonin gene-related peptide
BCTC N-(4-Tertiarybutylphenyl)-4-(3-chloropyridin-2-yl)
tetrahydropyrazine-1(2H)-carbox amide
shRNA Short hairpin RNA
ROS Reactive oxygen species
siRNA Short interfering RNA
IL Interleukin
TNF-a Tumor necrosis factor-a
NK Neurokinin 1
NF-kB Nuclear factor kB
AP-1 Activator protein 1
SB366791 N-(3-Methoxyphenyl)-4-chlorocinnamide
Application of RNA Interference to Treat Conditions Associated with Dysregulation 211
1 TRPV1 Ion Channels
Transient receptor potential (TRP) channels are sensors of temperature and pressure
stimuli and are therefore the frontline indicators of excessive stimulation from these
modalities. These receptors are localized primarily to peripheral sensory and spinal
cord neurons where they can participate fully in pain perception. A subset of these
channels detect a wide range of temperatures, spanning the spectrum from hot to
cold, with each channel subtype exhibiting a discrete temperature range for activa-
tion. TRPV1 (also known as VR1), which is activated by capsaicin, the active
ingredient of “hot” chili pepper (Szallasi and Blumberg 1999) shows activation
threshold of 43 C. TRPV2 (also known as VRL-1) responds to extremely hot
temperatures, and TRPV3 and TRPV4 (also known as VROAC or OTRPC4) are
activated by warm temperatures (Tominaga and Caterina 2004). Two additional
members of this group, which are activated by cold stimuli, include TRPV8 (also
known as CMR1) and TRPA1 (also known as ANKTM1) (Patapoutian et al. 2003;
Jordt et al. 2004).
In addition to its function as a thermal sensor, TRPV1 is activated by vanilloids
(capsaicin, resiniferatoxin) and nonvanilloids (protons, anandamide) and serves as a
primary mediator of thermal and inflammatory pain. This receptor is highly
expressed on sensory afferents (C and Ad fibers), which terminate in laminae
I and II of the spinal cord. These fibers might also serve an efferent role since
they release substance P and calcitonin gene-related peptide (CGRP) upon activa-
tion. TRPV1 is also highly expressed in the nucleus of the solitary tract, an area that
receives vagal projections from the nodose ganglion. Other central locations of
TRPV1 include the hypothalamus (Hori et al. 1988), substantia nigra (Marinelli
et al. 2003), locus coeruleus (Marinelli et al. 2002), and hippocampus (Al-Hayani
and Davies 2002), where its functions are still undefined. Nonneuronal localization
of TRPV1 include epithelial cells of the skin and bladder (Birder et al. 2001),
kidney, GI mucosa, and the inner ear (Mukherjea et al. 2008).
Studies using TRPV1-deficient mice have been critical to demonstrating the
involvement of this receptor in thermal and inflammatory pain. This mouse model
exhibited diminished responses to noxious stimuli, such as capsaicin, proton, or
heat, in addition to decreased thermal hyperalgesia induced by inflammation
(Caterina et al. 2000; Davis et al. 2000). Surprisingly, these mice also showed
deficits in neuropathic pain, mechanical allodynia, and mechanical hyperalgesia,
suggesting a role of TRPV1 in integrating multiple pain stimuli.
Treatment of conditions associated with increased expression or dysregulation
of TRPV1 have focused primarily on the use of agonist and antagonist drugs. The
usefulness of agonists derives from the observation that they desensitize TRPV1 or
destroy TRPV1 containing nerve fibers and thereby provide different degrees of
pain relief. Antagonists have shown efficacy in the treatment of inflammation,
arthritis, and cancer pain in animal models. The ability of most TRPV1 antagonists
to produce hyperthermia in animal models (Gavva et al. 2007) and humans (Gavva
et al. 2008) has led to concerns as to the future clinical utility of these compounds.
However, since the hyperthermia is transient and disappears with repeated dosing,
these agents might still play a significant role in pain management. The utility of
RNAi technology for effecting TRPV1 knockdown has been shown in a limited
number of disease models in animals. These include animal models of neuropathic
pain and drug-induced hearing loss. Information detailing the involvement of
TRPV1 in these and other conditions and the potential utility of RNAi as a
treatment option are provided below.
2 Neuropathic Pain
2.1 Current Treatment
Neuropathic pain significantly reduces the quality of life of affected individuals

since it impedes both their physical as well as emotional functions. It is also
associated with a significant cost to society since affected individuals may require
costly medical treatment and may be unable to perform effectively at work. Neuro-
pathic pain is produced from damage to the peripheral or central nervous system and
represents aberrant signal generation from the damaged nerves. It is associated with
a number of surgical procedures, injuries, and diseases including diabetes mellitus,
cancer metastasis to the bone, herpes zoster infection, and HIV infection. Current
treatment guidelines for neuropathic pain include the use of tricyclic antidepres-
sants, selective serotonin and norepinephrine reuptake inhibitors, gabapentin, preg-
abalin, or topical lidocaine as first line agents (O’Connor and Dworkin 2009). Opioid
analgesics have also provided substantial relief from neuropathic pain, but concerns
about their long term safety have resulted in their being classified as second line
agents. Third line agents, which include capsaicin (administered topically), dextro-
methorphan, memantine, and mexiletine, have shown inconsistent results in the
management of neuropathic pain (O’Connor and Dworkin 2009).
2.2 TRPV1 and Neuropathic Pain
A role of TRPV1 in mediating neuropathic pain in animal experiments has been

controversial. TRPV1 knockout mice exhibit reduced pain sensitivity (Caterina et al.
2000; Davis et al. 2000) but did not show any difference in mononeuropathic pain
induced by partial sciatic nerve ligation as compared to wild type mice (Caterina
et al. 2000), suggesting a lack of TRPV1 involvement in this model of neuropathic
pain. However, other studies indicate that TRPV1 knockout mice exhibit increased
mechanical hyperalgesia to polyneuropathic pain produced by chronic streptozoto-
cin and cisplatin (Bölcskei et al. 2005), suggesting an antinociceptive action of
TRPV1 in this chronic pain model. In a rat model of neuropathic pain, capsazepine
blocked A-fibers evoked responses in the dorsal horn, and selective TRPV1 antago-
nists attenuated mechanical allodynia and hyperalgesia (Christoph et al. 2006;
Honore et al. 2005; Kanai et al. 2005; Pomonis et al. 2003). These studies implicate
TRPV1 in mediating mechanical allodynia, at least in these in vivo models. In a
recent study, Christoph et al. (2006) demonstrated the utility of intrathecal adminis-
tration of siRNA against TRPV1. These investigators showed that knockdown of
TRPV1 by RNAi reduced cold allodynia of mononeuropathic rats and spontaneous
visceral pain. The effectiveness of RNAi was comparable to the TRPV1 antagonist,
N-(4-tertiarybutylphenyl)-4-(3-chloropyridin-2-yl)tetrahydropyrazine-1(2H)-carbox
amide (BCTC), and persisted over a period of 5 days. In a more recent study, Christoph
et al. (2008) showed that a transgenic mouse expressing short hairpin RNA (shRNA)
for TRPV1 exhibited decreased response to capsaicin-induced hypothermia,
decreased heat sensitivity on hot plates, and decreased mechanical allodynia. These
data will add further support to the contention that TRPV1 is a mediator of neuropathic
pain and are in stark contrast to data showing development of mechanical allodynia
and hypersensitivity to spinal nerve injury in TRPV1 knockout mice (Caterina et al.
2000). While the reason for this difference is unclear, it might reflect differences in the
animal models used and/or differences in compensatory measures produced by
knockout versus knockdown of TRPV1 by RNAi. One unexpected finding between
these two models is an increase in TRPV3 expression in dorsal root ganglions (DRGs)
of transgenic mice expressing TRPV1 shRNA but a decrease in TRPV3 expression in
TRPV1 knockout mice (Christoph et al. 2008). Overall, these data would support the
utility of TRPV1 RNAi in the treatment of neuropathic pain but might be limited by
the inability to effectively target the “drug” to the affected area.
2.3 Diabetic Peripheral Neuropathy
Diabetic peripheral neuropathy is one of the complications associated with uncon-

trolled or poorly controlled diabetes mellitus. It is believed to result from a loss of
peripheral nerve terminals innervating the extremities. Symptoms include pain,
tingling, and loss of feeling in the extremities. Damage to nerves may extend to
those innervating end organs such as the digestive tract, heart, and sex organs.
TRPV1 has also been implicated in diabetic peripheral neuropathy in animal (Hong
and Wiley 2005; Pabbidi et al. 2008) and some human studies (Wilder-Smith et al.
2007). In their study, Hong and Wiley (2005) showed that dorsal root ganglion cells
(DRGs) isolated from streptozocin-induced diabetic rats possessed increased TRPV1
channel activity, increased plasma membrane TRPV1 levels, and increased levels
of TRPV1 tetramers. These investigators also indicated that the increased TRPV1
expression was localized to large myelinated A fibers, while the normally expres-
sing C fibers showed reduced expression. Since no analgesic testing was performed
on these rats, it is difficult to determine whether the receptor changes observed were
associated with a diabetic hyperalgesic response. Using two mouse model of
diabetes, Pabbidi et al. (2008) demonstrated a direct correlation between TRPV1

expression with thermal sensitivity. In the streptozocin model, these investigators
showed an early hyperalgesic phase, followed by a later phase of hypoalgesia,
which correlated with TRPV1 function and expression. Similar findings were
demonstrated in a double transgenic model of diabetes (Pabbidi et al. 2008). In
addition, the investigators failed to show enhanced thermal nociception by strepto-
zocin in TRPV1 knockout mice. These data would suggest that changes in TRPV1
dictates the thermal sensitivity profile of diabetic animals. In addition to altered
neuronal expression of TRPV1 levels in DRGs of diabetic animals, coordinate
changes in TRPV1 levels were observed in the paw skin (Pabbidi et al. 2008).
Interestingly, changes in TRPV1 levels in the skin of humans is associated with
diabetic neuropathy (Wilder-Smith et al. 2007; Facer et al. 2007). Whether these
TRPV1 channels in skin play a role in pain transmission is not clear.
One possible interpretation of the data showing a biphasic thermal sensitivity
response following induction of diabetes is that the increases in neuronal TRPV1
channel and activity observed early in the course of diabetes initiates damage and/
or death of these neurons, which is manifested in hypoalgesia. In such a situation,
measures to reduce neuronal channel activity might reduce or prevent neuronal
damage and the ensuing hypoalgesia. As such, the application of TRPV1 antago-
nists or RNAi early in the course of diabetes might prove beneficial in reducing or
delaying the onset of peripheral neuropathy.
3 Drug-Induced Hearing Loss
3.1 TRPV1 and Cisplatin Ototoxicity
Platinum containing drugs have been successfully used in the treatment of various
solid tumors of the head and neck. One such drug, cisplatin, is an important
component of chemotherapeutic regimen for treating solid tumors. This drug
produces ototoxicity, in part, through the generation of reactive oxygen species
(ROS) (Rybak and Ramkumar 2007). One target of ROS include the organ of Corti
(Rybak 1999), where it destroys outer hair cells (Kopke et al. 1997). Current
treatment strategy involves the use of antioxidant (Rybak 1999). However, con-
comitant antioxidant use could interfere with the anticancer efficacy of cisplatin and
limit its usefulness in chemotherapy. As such, other treatment targets have been
sought after. One such target that we have identified is TRPV1, expressed in the
organ of Corti and spiral ganglion cells. In a recent study (Mukherjea et al. 2008),
we showed that TRPV1 is a target of ROS generated by cisplatin. ROS promote
activation and induction of TRPV1 and the NOX3 isoform of NADPH oxidase
(a major source of ROS generation in the cochlea) in the rat organ of Corti and
spiral ganglion cells. Generation of ROS via NOX3 was shown to be crucial to the
activation and induction of TRPV1.
3.2 Utility of RNAi in Treating Cisplatin Ototoxicity
The importance of TRPV1 in mediating cisplatin ototoxicity was shown by addi-

tional studies, which targeted this protein for knockdown by RNAi. Reduction in
TRPV1 expression in the cochlea by short interfering RNA (siRNA) decreased
cisplatin-induced damage to outer hair cells in the organ of Corti as evidenced from
scanning electron microscopy (Fig. 1a). Administration of siRNAs was performed
by round window application (to limit systemic distribution and maintain adequate
concentrations), and cisplatin was administered systemically by intraperitoneal
infusion. In presence of scrambled siRNA, cisplatin significantly reduced outer
hair cell number in the hook, basal, and middle turns of the cochlea. Round window
application of TRPV1 siRNA prior to cisplatin administration significantly reduced
the percentage of hair cell loss observed (Fig. 1b). Western blotting studies and
real time PCR performed to determine the level of TRPV1 protein (Fig. 2a) and
mRNA in the cochlea showed reductions in the level of TRPV1 protein (Fig. 2b)
and mRNA, following administration of TRPV1 siRNA and examining the
cochlea on day 2. To assess hearing loss, auditory brainstem responses (ABRs)
a b
Untreated Scrambled siRNA siTRPV1
HOOK
Percentage of hair cell damage
100
90
80 Scrambled siRNA
70 siTRPV1
BASE
60
50
40
30 *
20 *
MIDDLE
10
TURN
0 *
Tu ddle
se
ok
Ho
Ba
rn
i
M
Fig. 1 TRPV1 siRNA protects against cisplatin-induced outer hair cell damage. (a), Rats were
pretreated with either a scrambled siRNA sequence or with siRNA against TRPV1 by round
window application for 48 h. This was followed by cisplatin administration (13 mg/kg, i.p.), the
cochleae was collected 72 h later and were processed from SEM. TRPV1 siRNA protects against
cisplatin-mediated damage and loss of cochlear outer hair cells. SEM of samples collected 72 h
following cisplatin administration indicate damage to or loss of outer hair cells in cochleae
pretreated with scrambled siRNA sequence, with greatest effects obtained in the hook, followed
by the base and the middle turn (see arrows). Cochleae obtained from rats pretreated with siRNA
against TRPV1 showed statistically significant reductions in hair cell loss in all three regions
examined. (b) The percentage of outer hair cell damage in (a) is presented in graphical format.
Asterisk (*) indicate statistically significant reductions in the cochleae treated with TRPV1 siRNA
versus those treated with a scrambled siRNA (n ¼ 5; p < 0.05). (Reprinted with the permission
from Society of Neuroscience)
a b
Percentage of TRPV1 mRNA

250
**
TRPV1 200
expression
150
β-Actin 100
Cisplain – + – + 50
TRPV1 siRNA – – + + *
Scrambled + + – – 0
l
siRNA
ntro P V1 la
tin
Co TR isp
si C
c
50
siTRPV1 (0.9μg /3μl)
45
Scrambled siRNA
Threshold change (dB)
40
35
30
25
20
15
* *
10
5
0
8k 16k 32k clix
Fig. 2 siRNAs against TRPV1 reduced cisplatin-induced ototoxicity in rats. (a) siRNA against
TRPV1 suppressed the basal and cisplatin-stimulated TRPV1 protein levels in the cochlea
assessed 24 h following cisplatin administration. This is a representative of three independent
experiments showing similar responses. (b) Cochlear mRNA obtained from rats administered
TRPV1 siRNA and assessed 48 h later indicated a 85% reduction in basal TRPV1 mRNA.
Asterisk (*) indicate statistically significant reductions in the cochleae treated with TRPV1 siRNA
versus those treated with a scrambled siRNA (n ¼ 5; p < 0.05). (c) Pretreatment ABRs were
determined in the rats, which were then administered either a scrambled siRNA sequence in one
ear or 0.9 mg siRNA against TRPV1 by round window application in the other ear. Cisplatin
(13 mg/kg i.p) was administered 48 h later and posttreatment ABRs were determined after an
additional 72 h period. Cisplatin produced a shift in ABR thresholds (determined by comparing
pre- and posttreatment ABR thresholds) of 25–40 dB over an 8–32 kHz frequency range in the
ears of animals pretreated with the scrambled siRNA. However, statistically significant reductions
in ABR thresholds (p < 0.05; n ¼ 5) were obtained at 8 and 16 kHz frequency range and a trend
for protection observed at the 32 kHz range and for clicks (clix) in rats pretreated with siRNA
against TRPV1. (Reprinted with the permission from Society of Neuroscience)
were measured in rats pretreated with scrambled or TRPV1 siRNA and treated
with cisplatin 48 h later. ABRs performed 72 h following administration of cisplatin
(in rats pretreated with scrambled siRNA) showed increased ABR thresholds
by 38 6, 30 9, 24 4 and 30 7 dB at testing frequencies of 8, 16,
32 kHz or clicks, respectively. However, rats pretreated with TRPV1 siRNA
showed significantly reduced cisplatin-induced shifts in ABR thresholds (Fig. 2c).

Overall, these data provide compelling evidence that inhibiting TRPV1 expression
by siRNA could prove a useful strategy for protection against cisplatin-induced
hearing loss. In addition, since NOX3 is a major contributor to the induction
of TRPV1 by cisplatin in the cochlea (Mukherjea et al. 2008), RNAi approach
targeting this protein could also prove beneficial.
Antioxidants are also useful in the treatment of aminoglycoside ototoxicity due
to their ability to reduce oxidative damage to the cochlea induced by these agents
(Sha and Schacht 2000; Campbell et al. 2007). Similarly, antioxidants provide
protection against noise-induced hearing loss (Campbell et al. 2007). The latter
effect could be explained from the observation that noise exposure increased
the generation of ROS via the NADPH oxidase system in the cochlea (Ramkumar
et al. 2004). It is likely that the expression of TRPV1 is induced under both of these
conditions and could contribute to the hearing deficits observed. As such, we
propose that, like cisplatin ototoxicity, RNAi could be useful in treating ototoxicity
produced by both aminoglycoside antibiotics and noise.
One interesting aspect of TRP channels in the outer hair cells is that they may
also serve as entry ports for chemotherapeutic agents such as cisplatin and amino-
glycosides (Rybak and Ramkumar 2007). Thus, regulation of channel expression
through the use of siRNA would lead to reduced drug entry into the hair cells and
thereby reduce toxicity and cell death.
Limitations of the RNAi approach to treating ototoxicity are the method of
delivery and the duration of knockdown of the target protein. With experience,
we can quickly administrate siRNA via the round window and, more recently, a
trans-tympanic route (Mukherjea et al. 2010). Cyanine-3 labeled scrambled siRNA
administered via the round window persisted up to 10 days. As such, siRNA should
provide a longer term beneficial effect as compared to a TRPV1 antagonist, which
is expected to show a more limited duration of action with round window applica-
tion. However, it is not yet clear whether knockdown of the protein by RNAi would
persist much longer than the 3 days needed to assess the rats for outer hair cell
damage. An additional benefit of this RNAi approach is that local administration of
siRNA via the round window or the trans-tympanic routes would be expected to
decrease significant systemic distribution and side effects.
4 Other Potential Uses of RNAi Targeting TRPV1
Increased expression and/or dysregulation of TRPV1 is observed in a number of

disease processes associated with pain. These include inflammation, arthritis,
diabetic peripheral neuropathy, and cystisis. TRPV1 has also been implicated in
the control of body weight and obesity. These conditions are potential targets
for drugs that regulate TRPV1 and also potential therapeutic targets for TRPV1
siRNAs.
4.1 Inflammation
Inflammation is initiated by tissue injury and/or infection that sensitizes the tissue to
stimuli, which normally do not produce pain (allodynia) or increase pain to normally
noxious stimuli (hyperalgesia). This hypersensitivity is believed to result from
increased release of inflammatory mediators, such as prostaglandins, adenosine,
serotonin, bradykinin, and ATP, which act on Gq coupled receptors to sensitize
TRPV1 to endogenous or exogenous activators. These mediators also lower the
temperature threshold for activation of TRPV1 to below normal body temperature,
leading to chronic pain at the site of inflammation. A role of TRPV1 in mediating this
thermal hyperalgesia is provided by the observation that TRPV1 knockout mice
failed to show thermal hypersensitivity from inflammation induced by carrageenan
(Davis et al. 2000) or complete Freund’s adjuvant (Caterina et al. 2000).
One form of inflammation, termed neurogenic inflammation, is produced by
TRPV1-mediated release of mediators such as CGRP and substance P from nerve
terminals. The close proximity of these terminals to mast cells can trigger mast cell
degranulation by the mediators, leading to the release of additional inflammatory
mediators such as histamine, proteoglycans, serotonin, interleukin (IL), and tumor
necrosis-a (TNF-a). Mast cell mediators can positively regulate sensory neuron to
release CGRP and substance P and thereby serve as a positive feedback loop to
enhance neurogenic inflammation (Bı́ró et al. 1998). TRPV1 activation on mast
cells promotes release of inflammatory cytokines, such as IL-4 (Bı́ró et al. 1998).
As such, activation of TRPV1 can contribute indirectly (via stimulation of nerve
terminal) or directly (via mast cells) to neurogenic inflammation. TRPV1 also
appears to mediate a portion of the inflammatory response observed in caerulein-
induced acute pancreatitis. Activation of sensory fibers innervating the pancreas
leads to increased release of inflammatory neuropeptides (such as substance P),
which promote edema and the inflammatory response via the neurokinin 1 (NK1)
receptor. Blockade of NK1 receptor or TRPV1 reduces experimental pancreatitis
(Liddle 2007). These findings would support the clinical utility of TRPV1 blockade
or knockdown by RNAi in suppressing pancreatitis.
4.2 Arthritis
One aspect of inflammatory pain where TRPV1 plays a role is arthritis. TRPV1 has
been studied for its potential role in arthritic conditions such as osteoarthritis and
rheumatoid arthritis. TRPV1 levels were higher in the iodoacetate osteoarthritis
model (Fernihough et al. 2005) and in the complete Freund’s adjuvant-model of
inflammation (Amaya et al. 2003; Carlton and Coggeshall 2001), suggesting that it
mediates inflammation or is induced by inflammation. In a model of knee joint
inflammation, TRPV1 knockout mice showed reduced knee swelling in comparison
to the wild-type TRPV1 mice (Keeble et al. 2005), implicating TRPV1 in the
inflammatory process. These animals show similar levels of TNF-a, but TNF-a
induced a greater degree of knee swelling in the wild-type mice as compared to the
TRPV1-knockout mice (Keeble et al. 2005). These findings suggest an indirect
rather than a direct role of TRPV1 in arthritis. In a separate study by Szabó et al.
(2005), TRPV1 activation was shown to mediate adjuvant-induced chronic arthritis
in mice. The proinflammatory effect of adjuvant was further enhanced by bradyki-
nin and lipoxygenase products (Szabó et al. 2005), which are able to sensitize
TRPV1 through protein kinase C activation (Premkumar and Ahern 2000). Activa-
tion of TRPV1 under these conditions promote the release of substance P and
CGRP from peripheral nerve terminals (Szolcsanyi 1996), which mediate local
arteriolar vasodilation and increased plasma extravasation and inflammation in the
synovium (Lam and Ferrell 1991). As the inflammation progresses, these neuropep-
tides also increase the proliferation of synoviocytes (Lambert et al. 1998) and
increase secretion of inflammatory cytokines and hyperplasia of synovial tissues.
TRPV1 was also detected in the human synoviocytes, where it increased intracel-
lular calcium release (Kochukov et al. 2006). The increased intracellular calcium
release promoted ROS generation and death of synovial fibroblasts (Hu et al. 2008).
Recent studies have shown that the expression of TRPV1 is higher in the synovial
fibroblasts from patients with osteoarthritis and rheumatoid arthritis (Engler et al.
2007), implicating these receptors in the pathophysiology of osteoarthritis and
rheumatoid arthritis. In addition, studies have implicated TRPV1 not only in the
propagation of the inflammation but also in the generation of the pain associated
with the disease process. Hence, targeting the local TRPV1 receptors for inhibition
or knockdown by RNAi strategies could provide relief from arthritic pain and
inflammation. Direct injections of TRPV1 siRNA into the joint cavity might enable
localized knockdown of TRPV1 without significant systemic side effects. In this
respect, capsaicin creams (which are expected to destroy sensory nerve terminals in
the area) have been used topically on joints to relieve joint swelling and pain of
arthritis (Hautkappe et al. 1998).
In contrast to a purported role of TRPV1 in mediating the pain and inflammation
of arthritis, antiinflammatory role of capsaicin has also been demonstrated in the
treatment of arthritis (Brand et al. 1990; Jarreau et al. 1994; Joe and Lokesh 1994),
mediated presumably via inhibition of nuclear factor (NF)-kB and activator protein-1
(AP-1) (Singh et al. 1996, Surh et al. 2000). However, this response is observed at
higher doses of capsaicin and appears to be independent of TRPV1 activation.
4.3 Cystitis and Bladder Hyperactivity
TRPV1 is expressed on small myelinated Ad and unmyelinated C afferent fibers

innervations to the bladder and also in epithelial cells lining the lumen of the
bladder (Birder et al. 2001). The former is important for initiating the voiding
reflex by providing the brain with information concerning bladder fullness. C fibers
activity is also observed in neuropathic and inflammatory conditions and is
manifested by bladder hyperactivity (Chancellor and de Groat 1999). TRPV1

mediates the burning sensation and reflex action of capsaicin in the bladder
(Szolcsanyi 1977; Ishizuka et al. 1994). Activation of this receptor mediates the
hyperalgesia and bladder hyperreflexia produced in the cyclophosphamide model of
cystitis in the rat and increased c-fos expression in dorsal horn cells in the spinal
cord (Lecci et al. 1994; Dinis et al. 2004a, b). These events are mediated via TRPV1
activation by endogenous anandamide, the bladder content of which is increased by
cyclophosphamide (Dinis et al. 2004b). Pain from cystitis is relieved by intravesical
administration of vanilloids, which desensitize or destroy TRPV1-containing neu-
rons or by antagonists of TRPV1. It is believed that anandamide synergies with
inflammation to increase the hyperalgesia and bladder hyperreflexia (Hunt et al.
1987; Dinis et al. 2004b). A recent study has indicated a similar beneficial effect of
the TRPV1 antagonist, GRC-6211, against bladder hyperactivity induced by
capsaicin, acetic acid, and following lipopolysaccharide-mediated inflammation
(Charrua et al. 2009). Since this compound is orally active, these data may provide
a rationale alternative in the treatment of bladder hyperactivity induced by these
various models of cystitis.
Targeting TRPV1 might also prove an important strategy in treating urinary
retention in patients with spinal cord injuries. These patients normally show
increased intravesical pressures due to a lack of coordination between the activities
of the detrusor and sphincter muscles of the bladder. In addition, these patients
show increased expression of TRPV1 in urothelial cells and nerve fibers (Brady
et al. 2004; Apostolidis et al. 2005), which might serve as a basis of bladder
dysfunction. Intravesical administration of capsaicin or RTX (Apostolidis et al.
2005) and intrathecal RTX (Cruz et al. 2008) reduce intravesical pressures by
desensitizing TRPV1 or by reducing local TRPV1-expressing neurons, which
transmit afferent sensory stimuli to the spinal cord and brain, respectively.
The studies described above would support the utility of RNAi for localized
knockdown of TRPV1 in the bladder or spinal cord for the treatment of bladder
hyperactivity and hyperalgesia observed in cystitis and spinal cord injuries.
Because of its selectivity, RNAi could provide longer term control of TRPV1
expression without interfering with other sensory modalities associated with neu-
ronal or nerve terminal loss produced by capsaicin and RTX.
4.4 Cancer Pain
Cancer pain can significantly reduce the quality of life in patients who might also
have reduced life expectancy. This type of pain has historically been poorly
treated by analgesics. The pain likely results from increased pressure of the
tumor on nerves in the vicinity, increased inflammation and nerve damage or
altered expression of receptors involved in pain transmission. For example, the
levels of TRPV1 in DRGs were increased in an animal model of bone cancer pain
(Shinoda et al. 2008) and in squamous cell carcinoma of the human tongue
(Marincsák et al. 2009). In an effort to better treat cancer pain, the World Health
Organization has developed a stepwise ladder, which includes a stepwise use of
nonopioid drugs followed by opioid drugs (Zech et al. 1995). Effective control of
pain was reported by 70–90% of patients on this regimen. However, this treatment
protocol does not produce adequate pain control in 10–30% of individuals,
especially those suffering from bone cancer pain. While morphine is generally
used for these patients, higher doses are generally prescribed and are associated
with significant side effects such as sedation, constipation, and drug abuse liabil-
ity. Several reports have indicated that TRPV1 antagonists could be useful as
adjuncts to morphine in the treatment of animal models of bone cancer pain
(Shinoda et al. 2008; Niiyama et al. 2009). Administration of the antagonist
N-(3-methoxyphenyl)-4-chlorocinnamide (SB366791) with morphine decreased
pain induced by pressure on the affected limb or by movement (Niiyama et al.
2009). Interestingly, no significant pain relief was observed in these animals when
the drugs were administered individually, reinforcing the need for combined drug
administration. These findings, if confirmed, would support the clinical utility of
targeting TRPV1 in the treatment of cancer pain. Such treatment could include
TRPV1 antagonists or localized delivery of siRNA directed against TRPV1.
4.5 Obesity
Hot pepper has been alluded to be a thermogenic compound in Ayurvedic, Unani,

and ancient Chinese writings. Scientific documentation of this started as early as
1956 (Stevenel 1956). Early reports showed that capsaicin alters the levels of
cholecystokinin, a peripheral satiety molecule, which suppresses food intake
(MacLean 1985; Ritter and Ladenheim 1985) by activating capsaicin sensitive
nerve fibers, presumably localized in the gastrointestinal tract. Ingestion of red
chili peppers with a meal has been reported to decrease appetite and food intake in
human subjects (Yoshioka et al. 1999), suggesting a role of TRPV1 in appetite
suppression. Two recent studies reported that the endogenous fatty acid oleoyletha-
nolamide, a peripheral satiety molecule, activates TRPV1 on vagal sensory affer-
ents and thereby suppresses food intake. (Ahern 2003; Wang et al. 2005). A study
by Zhang et al. (2007) showed that capsaicin prevents adipogenesis in stimulated
3T3-L1 preadipocytes. 3T3-L1 cells in which TRPV1 was knockdown by RNAi
showed reduced capsaicin-mediated suppression of adipogenesis. These investiga-
tors also showed that visceral adipose and subcutaneous fat tissue from obese
human subjects express lower levels of TRPV1 protein and transcripts compared
to similar tissues from lean individuals (Zhang et al. 2007). Mice fed diet rich in fat
along with capsaicin had smaller adipocyte sizes and a significant decrease in
TRPV1 receptor expression in the visceral fat as compared to mice on high fat
diet alone. TRPV1 knockout mice showed similar weight gains whether capsaicin
was included in the high fat diet or not. Oral administration of capsinoids to humans
over a 12 week period increased energy metabolism, fat oxidation, and abdominal
fat loss (Snitker et al. 2009). Overall, these studies support a role of TRPV1
activation in controlling obesity and would obviate the need to the use of TRPV1
antagonist or knockdown by RNAi to treat this condition.
In contrast to the aforementioned study by Zhang et al. (2007), it was reported
that TRPV1 knockout mice on a low fat diet (4.5%) gained similar amount of
weight as their wild-type counterparts (Motter and Ahern 2008). However, on a
high fat diet (11% fat), the wild type mice showed increased weight gain after
17 weeks, compared to the TRPV1 knockout mice. When these mice were tested
over a 44 week period, wild type mice showed a significant increase in the mean
body mass compared to the knockouts. Thus, a deficiency in TRPV1 rendered these
mice more resistant to diet-induced weight gain. It was also reported that TRPV1-
null mice were able to withstand 1 h of cold exposure (0–2 C) without significant
change in their core body temperature, suggesting a greater capacity for thermo-
genesis. The discrepancy in results between Zhang et al. (2007) and Motter and
Ahern (2008) may be due to the difference in the amount of fat in the diets as well as
time period for which the animals were observed. However, these findings shed
some doubt as to whether activation of TRPV1 is indeed important for controlling
obesity and would suggest that knockdown of TRPV1 by RNAi and capsaicin in a
high fat diet could control obesity.
5 Conclusions
Over the past few years, TRPV1 has expanded its reach into a number of physio-
logical and pathological processes beyond its initially characterized role as a
mediator of thermal and inflammatory pain and a target of vanilloids. This receptor
is implicated in the mediation of pain in various disease processes including
neuropathic pain, arthritis, cystitis, and diabetic neuropathy and is implicated in
conditions such as hearing loss and possibly obesity. As such, there is increasing
interest at research institutions and pharmaceutical companies to produce drugs that
would eliminate or block TRPV1 activation. Since the clinical utility of TRPV1
antagonist has been shadowed by their induction of hyperthermia, it is expected that
RNAi would contribute to the overall available treatment options for these and
other conditions associated with TRPV1 dysregulation.
Conditions such as ototoxicity, arthritis, and cystitis, which would be alleviated
by localized administration of drugs could be specifically targeted for TRPV1
knockdown by RNAi. This technology is expected to provide target selectivity
and longer duration of action than existing therapeutics (especially if DNA-based
RNAi drugs and viral vectors are utilized) and provide limited side effects espe-
cially if delivered locally. As such, RNAi would be ideal for the long-term
management of chronic pain conditions. Unlike TRPV1 agonists, which could
induce pain relief through destroying TRPV1 containing nerve terminals, RNAi
is expected to selectivly target TRPV1 for knockdown without affecting sensory
inputs by different receptors present on those terminals. One concern is that the
localized administration of siRNA would be best performed in a clinical setting and

would thereby limit the use of these agents to such facilities. Another concern is that
plasmid or viral vectors used for RNAi or the naked DNA used could themselves
contribute to inflammation through immune activation. Certainly, these considera-
tions will have to be addressed before considering application of this technology in
humans.
Acknowledgments This work was supported by a grants from the National Institute of Health
(DC02396) to LPR and (CA135494) to VR, a grant from the National Organization for Hearing
Research and SIU School of Medicine Excellence in Academic Medicine Award to VR.
References
Ahern GP (2003) Activation of TRPV1 by the satiety factor oleoylethanolamide. J Biol Chem
278:30429–30434
Al-Hayani A, Davies SN (2002) Effect of cannabinoids on synaptic transmission in the rat
hippocampal slice is temperature-dependent. Eur J Pharmacol 442:47–54
Amaya F, Oh-hashi K, Naruse Y et al (2003) Local inflammation increases vanilloid receptor 1
expression within distinct subgroups of DRG neurons. Brain Res 963:190–196
Apostolidis A, Brady CM, Yiangou Y et al (2005) Capsaicin receptor TRPV1 in urothelium of
neurogenic human bladders and effect of intravesical resiniferatoxin. Urology 65:400–405
Birder LA, Kanai AJ, de Groat WC et al (2001) Vanilloid receptor expression suggests a sensory
role for urinary bladder epithelial cells. Proc Natl Acad Sci 98:13396–13401
Bı́ró T, Maurer M, Modarres S et al (1998) Characterization of functional vanilloid receptors
expressed by mast cells. Blood 9:1332–1340
Bölcskei K, Helyes Z, Szabó A et al (2005) Investigation of the role of TRPV1 receptors in acute
and chronic nociceptive processes using gene-deficient mice. Pain 117:368–376
Brady CM, Apostolidis AN, Harper M et al (2004) Parallel changes in bladder suburothelial
vanilloid receptor TRPV1 and pan-neuronal marker PGP9.5 immunoreactivity in patients with
neurogenic detrusor overactivity after intravesical resiniferatoxin treatment. BJU Int
93:770–776
Brand LM, Skare KL, Loomans MF et al (1990) Anti-inflammatory pharmacology and mechanism
of the orally active capsaicin analogs, NE-19550 and NE-28345. Agents Actions 31:329–340
Campbell KC, Meech RP, Klemens JJ et al (2007) Prevention of noise- and drug-induced hearing
loss with D-methionine. Hear Res 226:92–103
Carlton SM, Coggeshall RE (2001) Peripheral capsaicin receptors increase in the inflamed rat
hindpaw: a possible mechanism for peripheral sensitization. Neurosci Lett 310:53–56
Caterina MJ, Leffler A, Malmberg AB et al (2000) Impaired nociception and pain sensation in
mice lacking the capsaicin receptor. Science 288:306–313
Chancellor MB, de Groat WC (1999) Intravesical capsaicin and resiniferatoxin therapy: spicing up
the ways to treat the overactive bladder. J Urol 162:3–11
Charrua A, Cruz CD, Narayanan S et al (2009) GRC-6211, a new oral specific TRPV1 antagonist,
decreases bladder overactivity and noxious bladder input in cystitis animal models. J Urol
181:379–386
Christoph T, Gr€unweller A, Mika J et al (2006) Silencing of vanilloid receptor TRPV1 by RNAi
reduces neuropathic and visceral pain in vivo. Biochem Biophys Res Commun 350:238–243
Christoph T, Bahrenberg G, De Vry J et al (2008) Investigation of TRPV1 loss-of-function
phenotypes in transgenic shRNA expressing and knockout mice. Mol Cell Neurosci
37:579–589
Cruz CD, Charrua A, Vieira E et al (2008) Intrathecal delivery of resiniferatoxin (RTX) reduces
detrusor overactivity and spinal expression of TRPV1 in spinal cord injured animals. Exp
Neurol 214:301–308
Davis JB, Gray J, Gunthorpe MJ et al (2000) Vanilloid receptor-1 is essential for inflammatory
thermal hyperalgesia. Nature 405:183–187
Dinis P, Charrua A, Avelino A et al (2004a) Intravesical resiniferatoxin decreases spinal c-foc
expression and increases bladder volume to reflex micturition in rats with chronic inflamed
urinary bladder. BJU Int 94:153–157
Dinis P, Charrua A, Avelino A et al (2004b) Anandamide-evoked activation of vanilloid receptor 1
contributes to the development of bladder hyperreflexia and nociceptive transmission to spinal
dorsal horn neurons in cystitis. J Neurosci 24:11253–11263
Engler A, Aeschlimann A, Simmen BR et al (2007) Expression of transient receptor potential
vanilloid 1 (TRPV1) in synovial fibroblasts from patients with osteoarthritis and rheumatoid
arthritis. Biochem Biophys Res Commun 359:884–888
Facer P, Casula MA, Smith GD et al (2007) Differential expression of the capsaicin receptor
TRPV1 and related novel receptors TRPV3, TRPV4 and TRPM8 in normal human tissues and
changes in traumatic and diabetic neuropathy. BMC Neurol 7:11
Fernihough J, Gentry C, Bevan S et al (2005) Regulation of calcitonin gene-related peptide and
TRPV1 in a rat model of osteoarthritis. Neurosci Lett 388:75–80
Gavva NR, Bannon AW, Hovland DN Jr et al (2007) Repeated administration of vanilloid receptor
TRPV1 antagonists attenuates hyperthermia elicited by TRPV1 blockade. J Pharmacol Exp
Ther 323:128–137
Gavva NR, Treanor JJ, Garami A et al (2008) Pharmacological blockade of the vanilloid receptor
TRPV1 elicits marked hyperthermia in humans. Pain 136:202–210
Hautkappe M, Roizen MF, Toledano A et al (1998) Review of the effectiveness of capsaicin for
painful cutaneous disorders and neural dysfunction. Clin J Pain 14:97–106
Hong S, Wiley JW (2005) Early painful diabetic neuropathy is associated with differential changes
in the expression and function of vanilloid receptor 1. J Biol Chem 280:618–627
Honore P, Wismer CT, Mikusa J et al (2005) A-425619 [1-isoquinolin-5-yl-3-(4-trifluoromethyl-
benzyl)-urea], a novel transient receptor potential type V1 receptor antagonist, relieves patho-
physiological pain associated with inflammation and tissue injury in rats. J Pharmacol Exp
Ther 314:410–421
Hori T, Shibata M, Kiyohara T et al (1988) Responses of anterior hypothalamic-preoptic thermo-
sensitive neurons to locally applied capsaicin. Neuropharmacology 27:135–142
Hu F, Sun WW, Zhao XT et al (2008) TRPV1 mediates cell death in rat synovial fibroblasts
through calcium entry-dependent ROS production and mitochondrial depolarization. Biochem
Biophys Res Commun 369:989–993
Hunt SP, Pini A, Evan G (1987) Induction of c-fos like protein in spinal cord neurons following
sensory stimulation. Nature 328:632–634
Ishizuka O, Igawa Y, Mattiasson A et al (1994) Capsaicin-induced bladder hyperactivity in normal
conscious rats. J Urol 152:525–530
Jarreau PH, D’Ortho MP, Boyer V et al (1994) Effects of capsaicin on the airway responses to
inhaled endotoxin in the guinea pig. Am J Crit Care Med 149:128–133
Joe B, Lokesh BR (1994) Role of capsaicin, curcumin and dietary n-3 fatty acids in lowering the
generation of reactive oxygen species in rat peritoneal macrophages. Biochem Biophys Acta
1224:255–263
Jordt SE, Bautista DM, Chuang HH et al (2004) Mustard oils and cannabinoids excite sensory
nerve fibres through the TRP channel ANKTM1. Nature 427:260–265
Kanai Y, Nakazato E, Fujiuchi A (2005) Involvement of an increased spinal TRPV1 sensitization
through its up-regulation in mechanical allodynia of CCI rats. Neuropharmacology
49:977–984
Keeble J, Russell F, Curtis B et al (2005) Involvement of transient receptor potential vanilloid 1
in the vascular and hyperalgesic components of joint inflammation. Arthritis Rheum
52:3248–3256
Kochukov MY, McNearney TA, Fu Y et al (2006) Thermosensitive TRP ion channels mediate
cytosolic calcium response in human synoviocytes. Am J Physiol Cell Physiol 291:C424–C432
Kopke RD, Liu W, Gabaizadeh R et al (1997) Use of organotypic cultures of Corti’s organ to study
the protective effects of antioxidant molecules on cisplatin-induced damage of auditory hair
cells. Am J Otol 18:559–571
Lam FY, Ferrell WR (1991) Specific neurokinin receptors mediate plasma extravasation in the rat
knee joint. Br J Pharmacol 103:1263–1267
Lambert N, Lescoulie PL, Yassine-Diab B et al (1998) Substance P enhances cytokine-induced
vascular cell adhesion molecule-1 (VCAM-1) expression on cultured rheumatoid fibroblast-
like synoviocytes. Clin Exp Immunol 113:269–275
Lecci A, Giuliani S, Santicioli P et al (1994) Involvement of spinal tachykinin NK1 and NK2
receptors in detrusor hyperreflexia during chemical cystitis in anaesthetized rats. Eur J Pharmacol
259:129–135
Liddle RA (2007) The role of transient receptor potential vanilloid 1 (TRPV1) channels in
pancreatitis. Biochim Biophys Acta 1772:869–878
MacLean DB (1985) Abrogation of peripheral cholecystokinin-satiety in the capsaicin treated rat.
Regul Pept 11:321–333
Marincsák R, Tóth BI, Czifra G et al (2009) Increased expression of TRPV1 in squamous cell
carcinoma of the human tongue. Oral Dis 15:328–335
Marinelli S, Vaughan CW, Christie MJ et al (2002) Capsaicin activation of glutamatergic synaptic
transmission in the rat locus coeruleus in vitro. J Physiol 543:531–540
Marinelli S, Di Marzo V, Berretta N et al (2003) Presynaptic facilitation of glutamatergic synapses
to dopaminergic neurons of the rat substantia nigra by endogenous stimulation of vanilloid
receptors. J Neurosci 23:3136–3144
Motter AL, Ahern GP (2008) TRPV1-null mice are protected from diet-induced obesity. FEBS
Lett 582:2257–2262
Mukherjea D, Jajoo S, Whitworth C et al (2008) Short interfering RNA against transient receptor
potential vanilloid 1 attenuates cisplatin-induced hearing loss in the rat. J Neurosci
28:13056–13065
Mukherjea D, Jajoo S, Kaur T et al (2010) Transtympanic administration of short interfering (si)
RNA for NOX3 isoform of NADPH oxidase protects against cisplatin-induced hearing loss in
the rat. Antiox. Redox Signal. [Epub June 2010]
Niiyama Y, Kawamata T, Yamamoto J et al (2009) SB366791, a TRPV1 antagonist, potentiates
analgesic effects of systemic morphine in a murine model of bone cancer pain. Br J Anaesth
102:251–258
O’Connor AB, Dworkin RH (2009) Treatment of neuropathic pain: an overview of recent guide-
lines. Am J Med 122:S22–32
Pabbidi RM, Yu SQ, Peng S et al (2008) Influence of TRPV1 on diabetes-induced alterations in
thermal pain sensitivity. Mol Pain 14:9
Patapoutian A, Peier AM, Story GM et al (2003) ThermoTRP channels and beyond: mechanisms
of temperature sensation. Nat Rev Neurosci 4:529–539
Pomonis JD, Harrison JE, Mark L (2003) N-(4-Tertiarybutylphenyl)-4-(3-cholorphyridin-2-yl)
tetrahydropyrazine -1(2H)-carbox-amide (BCTC), a novel, orally effective vanilloid receptor 1
antagonist with analgesic properties: II. in vivo characterization in rat models of inflammatory
and neuropathic pain. J Pharmacol Exp Ther 306:387–393
Premkumar LS, Ahern GP (2000) Induction of vanilloid receptor channel activity by protein
kinase C. Nature 408:985–990
Ramkumar V, Whitworth CA, Pingle SC et al (2004) Noise induces A1 adenosine receptor
expression in the chinchilla cochlea. Hear Res 188:47–56
Ritter RC, Ladenheim EE (1985) Capsaicin pretreatment attenuates suppression of food intake by
cholecystokinin. Am J Physiol 248:R501–504
Rybak LP (1999) Ototoxicity: bioprotective mechanisms. Curr Opin Otolaryngol Head Neck Surg
11:328–333
Rybak LP, Ramkumar V (2007) Ototoxicity. Kidney Int 72:931–935

Sha SH, Schacht J (2000) Antioxidants attenuate gentamicin-induced free radical formation
in vitro and ototoxicity in vivo: D-methionine is a potential protectant. Hear Res 142:34–40
Shinoda M, Ogino A, Ozaki N, Urano H, Hironaka K, Yasui M, Sugiura Y (2008) Involvement of
TRPV1 in nociceptive behavior in a rat model of cancer pain. J Pain 9:687–699
Singh S, Natarajan K, Aggarwal BB (1996) Capsaicin (8-methyl-N-vanillyl-6-nonenamide) is a
potent inhibitor of nuclear transcription factor-kappa B activation by diverse agents. J Immunol
157:4412–4420
Stevenel L (1956) Red pepper, a too little used therapeutic agent for anorexia, liver congestion and
various vascular disorders (hemorrhoids, varices). Bull Soc Pathol Exot Filiales 49:841–843
Surh YJ, Han SS, Keum YS et al (2000) Inhibitory effects of curcumin and capsaicin on phorbol
ester-induced activation of eukaryotic transcription factors, NF-kB and AP-1. Biofactors
12:107–112
Snitker S, Fujishima Y, Shen H et al (2009) Effects of novel capsinoid treatment on fatness and
energy metabolism in humans: possible pharmacogenetic implications. Am J Clin Nutr
89:45–50
Szabó A, Helyes Z, Sándor K et al (2005) Role of transient receptor potential vanilloid 1 receptors
in adjuvant-induced chronic arthritis: in vivo study using gene-deficient mice. J Pharmacol Exp
Ther 314:111–119
Szallasi A, Blumberg PM (1999) Vanilloid (Capsaicin) receptors and mechanisms. Pharmacol Rev
51:159–212
Szolcsanyi J (1977) A pharmacological approach to elucidation of the role of different nerve fibres
and receptor endings in mediation of pain. J Physiol (Paris) 73:251–279
Szolcsanyi J (1996) Capsaicin-sensitive sensory nerve terminals with local and systemic efferent
functions: facts and scopes of an unorthodox neuroregulatory mechanism. Prog Brain Res
113:343–359
Tominaga M, Caterina MJ (2004) Thermosensation and pain. J Neurobiol 61:3–12
Wang X, Miyares RL, Ahern GP (2005) Oleoylethanolamide excites vagal sensory neurones,
induces visceral pain and reduces short-term food intake in mice via capsaicin receptor
TRPV1. J Physiol 564:541–547
Wilder-Smith EP, Ong WY, Guo Y et al (2007) Epidermal transient receptor potential vanilloid 1
in idiopathic small nerve fibre disease, diabetic neuropathy and healthy human subjects.
Histopathology 51:674–680
Yoshioka M, St-Pierre S, Drapeau V et al (1999) Effects of red pepper on appetite and energy
intake. Br J Nutr 82:115–123
Zech DFJ, Grond G, Lynch J et al (1995) Validation of World Health Organization guidelines for
cancer pain relief: a 10-year prospective study. Pain 63:65–76
Zhang LL, Yan Liu D, Ma LQ et al (2007) Activation of transient receptor potential vanilloid
type-1 channel prevents adipogenesis and obesity. Circ Res 100:1063–1070
Harnessing RNAi-Based Functional Genomics
to Unravel the Molecular Complexity
Underlying Skin Pigment Variation
Hsiang Ho, Jayavani Aruri, Safoora Ahmed, and Anand K. Ganesan
Contents
1 Melanin: A Ubiquitous Pigment with a Multitude of Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 228
2 Melanogenesis: Insights Uncovered by the Detailed Analysis of Genetic Disorders of
Pigmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
3 An RNAi-Based Functional Genomics Approach to Identify
Novel Regulators of Melanogenesis in Human Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
4 Integration of Multiple Systems-Level Approaches to Uncover
Additional Regulators of Melanogenesis in Our Functional Genomics Dataset . . . . . . . . . . 237
5 Identification of Novel Pathways that Regulate the Transcription
of Melanogenic Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
6 Identification of Novel Pathways that Regulate Melanosome Biogenesis . . . . . . . . . . . . . . . . 244
7 Identification of Regulators of Human Pigment Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
8 Identification of Pharmacologic Agents that Impact
Melanin Accumulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
9 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Abstract Melanin, the primary chromophore in human skin, protects the skin and
eyes from the harmful effects of UV irradiation, protects neural cells from toxic
insults, and is required for sound conduction in the inner ear. Aberrant regulation of
melanogenesis underlies skin disorders (melasma and vitiligo), neurologic disor-
ders (Parkinson’s disease), auditory disorders (Waardenburg’s syndrome), and
ophthalmologic disorders (age-related macular degeneration). Extensive studies
have identified over 150 genes that regulate the production of melanin pigment in
human cells. Despite this extensive investigation, the phenotypic variation in
human skin color cannot simply be explained by the aberrant regulation of these
150 genes. We recently utilized RNAi-based functional genomics to unravel the
H. Ho, J. Aruri, S. Ahmed, and A.K. Ganesan (*)

Department of Dermatology and Biological Chemistry, University of California, 324 Sprague
Hall, Irvine, CA, 92697-2400, USA
e-mail: aganesan@uci.edu
228 H. Ho et al.
molecular complexity underlying skin pigment variation. These studies identified

92 novel genes that regulate melanogenesis in human cells, novel pharmacologic
inhibitors of melanin production, and novel pathways that regulate melanogenesis.
In this chapter, we will discuss an integrative approach to identify novel genes and
pathways that regulate melanogenesis from our dataset. First, we will discuss
strategies to integrate protein–protein interaction network algorithms with our
dataset to better define pathways that regulate melanogenesis in human cells.
Two aspects of our RNAi screen dataset were unique: our screen failed to identify
many of the known regulators of melanogenesis among our top tier of hits, indicating
that the approach had an appreciable false negative rate. Instead, our approach
identified a large number of genes that regulate melanogenesis by uncharacterized
mechanisms. Therefore, we sought to define our targets by measuring their impact
on known melanogenesis regulatory machinery. Initial studies segregated these
genes into functional classes by examining their impact on known regulators of
melanogenesis, tyrosinase and MITF. An annotation-based approach was used to
identify targets that potentially regulate melanosome trafficking, uncovering an
unexpected link between autophagy and melanogenesis. In vivo studies validated
that autophagy impacts melanogenesis. Recent studies have developed an integrative
functional genomics strategy utilizing both siRNA and shRNA platforms to define
the impact of individual genes on melanogenesis. In this chapter, we present a model
to integrate protein–protein interaction network datasets, functional genomics data,
siRNA and shRNA loss of function platforms, and pharmacologic databases to
identify functional gene networks that govern human phenotypes and novel treat-
ments for clinical disorders.
Keywords Melanogenesis siRNA Autophagy Functional genomics
1 Melanin: A Ubiquitous Pigment with a Multitude of Functions
Melanin plays a critical role in protecting the skin, eyes, and brain from toxic
insults. Melanin is produced from tyrosine via a highly regulated enzymatic path-
way (Slominski et al. 2004). It absorbs light over a broad spectrum, encompassing
both the UV and visible spectrum (Stein 1955). This property of melanin serves to
protect the skin from the harmful effects of UV irradiation (Slominski et al. 2004).
There is an inverse correlation between epidermal melanin content and skin cancer
incidence (Kadekaro et al. 2003), demonstrating that melanin is the key pigment
that protects the epidermis from UV damage. In addition to its protective effects
against UV irradiation, melanin, particularly neuromelanin, is a chelator of heavy
metals (Double 2006). The protective role of melanin in the brain has been linked
with its ability to block the toxic effects of environmental metals (Zecca et al.
2008). Loss of striatal melanin in the inner ear correlates with age-associated
hearing loss, suggesting that melanin may have protective effects in this organ
Harnessing RNAi-Based Functional Genomics to Unravel the Molecular Complexity 229
(Ohlemiller et al. 2009). The protective effect of melanin in the ear is thought to be
secondary to its ability to bind both trace metals and ototoxic drugs (Breathnach
1988). Melanin also protects against high-intensity noise damage by an unknown
mechanism (Breathnach 1988). In the eye, melanin protects choroidal blood vessels
from the toxic effects of UV radiation (Ohlemiller et al. 2009). The absorptive
properties of retinal pigment epithelium melanin prevents reflection and protects
the photoreceptors from harmful effects of UV (Breathnach 1988). The critical
protective role of melanin in both epithelial and mesenchymal tissues has spawned
interest in further understanding the normal cellular mechanisms that control the
production and metabolism of melanin.
In addition to its protective effects, melanin plays a key role in the etiology of
human disease. Several studies have suggested that aberrant production of melanin
may be an initial event in melanoma formation (Meyskens et al. 2007). Melanoma
cells accumulate abnormal melanosome structures both in human and rodent
melanoma models (Bomirski et al. 1987). Aberrant regulation of melanin in the
skin underlies the pigmentary disorders vitiligo and melasma (Grimes et al. 2006).
Vitiligo melanocytes are also characterized by abnormal accumulation of melano-
somal structures (Boissy et al. 1983, 1991a, b). Several Waardenburg’s syndrome
variants have both skin pigment defects and hearing loss, indicating that defects in
melanin production directly impact sound conduction (Price et al. 1998). Interest-
ingly, these variants have different patterns of epidermal pigment loss, demonstrat-
ing that there may be differences in the regulation of melanogenesis in different
body segments (Delezoide and Vekemans 1994). Age-related macular degeneration
is characterized by irregular deposition of melanin in the macula, suggesting a link
between melanin dysfunction and the most common cause of blindness in the
United States (Sarangarajan and Apte 2005). Additional studies have indicated
that melanin plays a role in the etiology of Parkinson’s disease (Zecca et al.
2006), a disorder characterized by the loss of both dopaminergic neurons and
melanin in the substantia nigra and locus ceruleus, a part of the brain that produces
the majority of neural melanin (Double and Halliday 2006). While the current
treatment of Parkinson’s disease involves the administration of agents that stimu-
late dopaminergic neurons, these agents are also known precursors of melanin
(L-dopa in particular) (Lewitt 2008). The impact of L-dopa treatment on melanin
production in the brain is currently unclear.
Melanin also plays a critical role in an organism’s adaptation to its environment
(Hoekstra 2006). Color adaptation can provide camouflage for certain animal
species, protecting them from predators or toxic insult. Variations in melanin
pigment define adaptive pigment variation in wolves (Anderson et al. 2009),
adaptive pigment variation in mice (Nachman 2005), and coat color variation in
cattle (Seo et al. 2007). Variation in plumage can provide a protective advantage for
bird species (Palleroni et al. 2005). Similarly, several species of fish can selectively
change color to adapt to their environment (Sugimoto 2002) to help them avoid
predators. In humans, skin color varies with latitude, suggesting an adaptive
relationship between variation in skin pigmentation and relative UV exposure
(Parra 2007). The mechanisms that underlie this differential regulation are
230 H. Ho et al.
unknown. In summary, the critical function of melanin in protecting tissues from

toxic insult, the link between melanin dysfunction and human disease, and the role
of melanin in phenotypic variation underscores the importance of elucidating the
genetic and biochemical mechanisms that regulate melanin production.
2 Melanogenesis: Insights Uncovered by the Detailed

Analysis of Genetic Disorders of Pigmentation
Extensive studies have identified over 150 genes that regulate melanin production
and have centered on characterizing inherited pigmentary disorders (Bennett and
Lamoreux 2003). Melanin production is precisely regulated within the cell, involv-
ing specific mechanisms of transcriptional, enzymatic, and spatial regulation
(Yamaguchi et al. 2007). Recent studies have identified several different pathways
that upregulate melanin production in the melanocyte. UV irradiation is the primary
carcinogen in the epidermis, and can induce DNA crosslinks (Gorner 1994).
A central role of the melanocyte in the skin is to protect adjacent epidermal cells
from the harmful effects of UV irradiation (Costin and Hearing 2007). To do this,
the melanocyte must be able to sense UV damage and stimulate the production of
melanin in response to ultraviolet light (Costin and Hearing 2007). Recent studies
have demonstrated that the keratinocyte play a vital role in inducing melanogenesis.
UV irradiation of keratinocytes in the epidermis leads DNA crosslinks, which
eventually result in p53 activation (Cui et al. 2007). Recent studies have demon-
strated that p53 activation in keratinocytes is critical for inducing melanogenesis in
adjacent melanocytes (Cui et al. 2007). This phenomenon can occur by one of the
two mechanisms. Some studies have determined that p53 activation in keratino-
cytes leads to increased MSH production, which can then bind to the melanocortin
receptor on melanocytes and stimulate pigment production (Cui et al. 2007). Other
studies have determined that p53 activation leads to increased Kit ligand expression
in keratinocytes, which can then bind to the Kit receptor on adjacent melanocytes
and induce pigment production (McGowan et al. 2008). The precise role of each of
these pathways in UV-induced melanogenesis is currently unclear.
While UV irradiation can stimulate melanin production, other mechanisms exist
to regulate the baseline levels of intracellular melanin. One such mechanism
involves the EDNRB pathway. This pathway is initiated when endothelin binds
to the endothelin receptor, EDNRB. EDNRB can then stimulate an intracellular
signaling cascade involving the G-protein GNA11, ultimately leading to an increase
in MITF transcriptional activity (Hou et al. 2004). Other ligands can regulate MC1R
activity in a UV-independent manner. In wolves, coat color variation is partially
regulated by b-defensin, a molecule that also regulates innate immunity (Anderson
et al. 2009). b-defensin binds to the melanocortin 1 receptor, stimulating melanin
production via a mechanism similar to MSH (Anderson et al. 2009). The precise
regulation of b-defensin and endothelin in human skin and the roles of these
molecules in pigment regulation is poorly understood.
Melanin production is controlled by the key transcriptional regulator of the

melanocyte lineage, MITF (Levy et al. 2006). Mutations in MITF result both in
defects in melanocyte development and in pigmentary defects. MITF, a b-HLH
transcription factor, regulates the expression of both melanogenic enzymes and
proteins required for melanocyte survival. MITF expression is regulated by three
different transcription factors, b-catenin, SOX10, and PAX3. SOX10 and PAX3
also independently regulate the production of other melanogenic enzymes (Murisier
et al. 2007). Genetic mutations in PAX3, SOX10, and MITF have similar clinical
phenotypes, further illustrating that they coordinately regulate the expression of key
pigmentation genes (Etchevers et al. 2006). Recent studies have focused on char-
acterizing the interactions of these transcriptional regulators in melanogenesis and
determining how they differentially regulate skin pigmentation.
In addition to the transcriptional control of melanogenesis, melanin is also
subject to enzymatic regulation. Tyrosinase, the rate-limiting enzyme in melanin
biosynthesis, converts tyrosine to L-dopa and then to dopaquinone and is mutated in
oculocutaneous albinism (Slominski et al. 2004). Mutations in tyrosinase impact
mouse hair color, demonstrating that variation in tyrosinase activity plays a role
in controlling pigment production (Lamoreux et al. 2001). Tyrp2 (dopachrome
tautomerase) and Tyrp1 (dopachrome oxidase) convert dopaquinone to melanin
(Slominski et al. 2004). Tyrp1 mutation causes another type of oculocutaneous
albinism (Oetting and King 1999) characterized by partial pigment loss. Together,
these studies indicate that alterations in the activities of these enzymes is sufficient
to impact skin color. Once melanin is synthesized, it is polymerized into higher
order structures on an amyloid fibril template composed of the protein Pmel17
(Fowler et al. 2006). Preliminary studies suggest that Pmel17 plays a role in
catalyzing melanin polymerization. This protein is mutated in Silver mice, a
mouse characterized by pigment dilution (Kwon et al. 1994). Other studies have
demonstrated that other melanosomal proteins also play a role in melanin polymer-
ization. It is not yet clear whether melanin polymerization is enzymatically
regulated.
Once melanin is synthesized and polymerized, it must be transferred to kera-
tinocytes. Melanin synthesis is accompanied by the generation of free radicals
(Sarangarajan and Apte 2006). Melanocytes have evolved complex mechanisms to
compartmentalize melanogenesis as a mechanism to protect the cell from the
harmful effects of free radicals, and much work has been completed to character-
ize the interplay between transcriptional regulation of melanogenesis and the
trafficking of melanin within the cell (Goding 2007). Electron microscopy mor-
phologic analysis has characterized the organelle responsible for melanin synthe-
sis, the melanosome, a specialized lysosome-related organelle that is exclusively
produced in the melanocyte or retinal pigment epithelial cell (Raposo et al. 2007).
Melanosomes mature via four distinct stages (Raposo and Marks 2007). The first
two stages lack pigment but do contain amyloid proteinaceous fibrils composed of
Pmel17 (Raposo et al. 2001), which organize to form an ellipsoidal structure
(Raposo et al. 2001). Subsequently, melanin synthesis begins and polymerizes on
an amyloid template, resulting in blackening of the melanosome (stage III) until
232 H. Ho et al.
ultimately the organelle is completely opaque as viewed by electron microscopy

(stage IV) (Setaluri 2000). The terminal melanosome is then transferred to
the periphery of the cell via microtubules in a process requiring both dynein
and kinesin motors (Aspengren et al. 2009). Ultimately, the melanosome is
transferred to adjacent keratinocytes in an actin-dependent manner via a process
mediated by Rab27a and melanophilin (Hume et al. 2001; Manneville et al. 2003;
Wu et al. 2001).
As a mechanism to control free radical generation, distinct transport events are
utilized to deliver cargo to the melanosome at different stages during melanosome
development. Morphologically, the stage I melanosome resembles multivesicular
bodies, a component of the early endosomal transport machinery (Berson et al.
2001). As Pmel17 is localized to both the multivesicular body and the plasma
membrane, it is unclear whether this protein is transported to the melanosome via a
circuitous route involving the plasma membrane and endosomal recycling (Valen-
cia et al. 2006), or whether it is directly transported to the melanosome (Raposo
et al. 2001). Other melanosome components (MART-1) are transported to the early
melanosome (stage I–II) via a mechanism requiring ubiquitin conjugation (Raposo
et al. 2001). While these studies have led to a rudimentary understanding of sorting
events required for directing proteins to the stage I melanosome, a complete
understanding of the pathways that regulate early melanosome sorting remain
elusive (Raposo and Marks 2007).
Much more is known about the molecular transport machinery used to generate
the terminal melanosome. Hermansky–Pudlak syndrome is a human disorder that
impacts the formation of lysosome-related organelles, which include the melano-
some (Shotelersuk and Gahl 1998). Mapping and functional characterization of
Hermansky–Pudlak syndrome mutants has broadened our understanding of the
trafficking pathways that deliver the melanin biosynthetic enzymes tyrosinase and
tyrp1 to the melanosome (Raposo and Marks 2007). Two distinct transport path-
ways are required to transport the melanogenic enzymes, tyrp1 and tyrosinase
(Raposo and Marks 2007), and the copper transporter ATP7A (Setty et al. 2008),
to the late melanosome. Mouse and human melanocytes with mutations in the AP-3
adaptor protein have defects in tyrosinase sorting but not tyrp1 sorting, indicating
that AP-3 is important in sorting tyrosinase to the melanosome compartment
(Huizing et al. 2001). In contrast, Tyrp1 binds to AP1 but not AP3, indicating
that a different set of adaptor proteins is required for the sorting of tyrp1 and
tyrosinase (Raposo and Marks 2007). Additional studies demonstrate that proper
AP1-mediated sorting requires a distinct set of kinesin motors (Delevoye et al.
2009). Once these two proteins are sorted into the appropriate adaptor protein-
containing vesicle, they utilize similar vesicle trafficking machinery to transport
them to the melanosome. Sorting of proteins from the endosome to the melanosome
requires the BLOC (biogenesis of lysosomal organelles) multiprotein complex
(Raposo et al. 2007). Genetic studies have defined two distinct intracellular traf-
ficking complexes (BLOC-1 and BLOC-2) that are required for proper sorting of
melanosomal proteins (Raposo et al. 2007). Deficiencies in the BLOC-1 complex
result in a profound inhibition of pigment production and gross defects in Tyrp1
trafficking (Setty et al. 2007). This pathway is specifically required for the transport
of ATP7A, a copper transporter that loads copper onto tyrosinase, to the melano-
some (Setty et al. 2008). BLOC-2 mutations have more subtle pigment dilution
phenotypes, suggesting that these two complexes are components of the same
trafficking pathway.
Despite the significant advances gained from studying genetic disorders of
pigmentation (Bennett and Lamoreux 2003), the complex regulation of melanogen-
esis in human cells still remains a mystery. Two aspects of melanin regulation have
been particularly difficult to decipher. The regulation of the transcription of mela-
nogenic enzymes is extremely complex, involving the interaction of multiple
different signaling pathways and signaling intermediates. Similarly, the transport
of melanosomal enzymes is a complex process involving the interaction of several
different transport pathways. RNAi-based functional genomics is a systems-level
approach to identify key regulators of biological phenotypes. We sought to use this
technology to uncover the underappreciated molecular complexity that governs
melanin production. Specifically, we hoped to use this approach to identify cellular
components that regulate the transcription of melanogenic enzymes and the trans-
port of melanosomal enzymes.
3 An RNAi-Based Functional Genomics Approach to Identify

Novel Regulators of Melanogenesis in Human Cells
Detailed characterization of genetic disorders of pigmentation in human, mouse,

and zebrafish models have led to a major advance in our understanding of pigment
regulation. While these studies are likely to identify genes that specifically impact
pigment production, they are unlikely to identify pathways that regulate pigment
production that also have pleiotropic effects on other pathways that impact viability
or development. To uncover a more complete spectrum of genes that impact
melanogenesis in human cells, we utilized a systems-level functional genomics
approach. Identification of regulators of pigment production in cultured cells can be
problematic: melanocytes are fragile in culture and produce scant amount of
pigment (Smit et al. 1997; Smit et al. 1998). To overcome these limitations, we
utilized MNT-1 melanoma cells to screen a genome-wide synthetic siRNA library
for single-gene loci that support melanocyte pigmentation (Ganesan et al. 2008).
These cells were specifically selected because these cells produce abundant melanin
in culture. Melanogenesis regulators identified in this system are more likely to
impact pigment production in melanocytes as MNT-1 cells are highly similar to
normal melanocytes (Hoek et al. 2004) and have been used by others to define
pigment regulatory mechanisms (Di Pietro et al. 2006; Kushimoto et al. 2001;
Theos et al. 2005, 2006). As previous studies have demonstrated that many mela-
nogenesis regulators also have impacts on cellular survival pathways, we developed
a strategy to identify genes that impact pigment production without impacting
cellular survival. A Dharmacon siRNA library of 84,508 siRNAs corresponding
234 H. Ho et al.
to four unique siRNA duplexes, targeting each of the 21,127 unique human genes
arrayed in a one-gene/one-well format on 96 well microtiter plates was used to
identify siRNAs that inhibit pigment production. A spectrophotometric melanin
quantitation assay was coupled with a luminescence cell viability assay (CellTiter-
Glo) to identify siRNAs that impact melanin production but do not have impacts on
cell survival (Ganesan et al. 2008). To identify genes that impact both pheomelanin
and eumelanin production, we measured melanin content at 405 nm (Ozeki et al.
1996), a wavelength at which both pheomelanin and eumelanin absorb light.
Previous studies have demonstrated that MNT-1 cells readily secrete melanin into
the media, facilitating the identification of genes that regulate all stages of melanin
secretion. Optimization studies using tyrosinase siRNAs determined the optimal
time to measure the impact of siRNA treatment on pigment production in MNT-1
cells (Ganesan et al. 2008). Other preliminary studies developed protocols to
normalize our dataset to control for variations in pigment due to plate or position
effects. The entire screen was completed in duplicate (Ganesan et al. 2008). The
mean and standard deviation for each siRNA pool was calculated, and the
corresponding log 2 value or Z-score (Fig. 1) was used to plot the data distribution.
2
Z-score
0
0 5000 10000 15000 20000
–1
–2
–3
–4
siRNAs
Fig. 1 Identification of positive regulators of melanogenesis in human cells. MNT-1 pigmented

melanoma cells were transfected with 84,920 siRNA duplexes targeting 21,230 genes in a one-
well, one-gene reverse transfection format as we have previously described (Whitehurst et al.
2007). 120 h posttransfection, Raw A405nm absorbance values were collected for each well and
normalized to internal reference samples on each plate, followed by normalization to the experi-
mental mean for each well calculated from the full dataset to control for variations in pigment due
to plate and position effects. Similarly adjusted luminescence values from a multiplexed viability
assay (CellTiter-Glo) were used to control for cell number, generating “normalized absorbance
ratios” for each well. The Z-score of the average normalized absorbance ratios among replicates is
shown for each gene from lowest (hypopigmentation) to highest (hyperpigmentation). Values
below 2 standard deviations from the mean are colored black
Table 1 Candidate pigmentation genes

Category Symbol Comments Motifs
Autophagy MAP1LC3C MAP1_LC3
WIPI1 Expressed in melanoma WD40
cell autophagosomes
GPSM1 GoLoco
GPCR GNG2 GGL
GPR113 GPS, 7tm_2
EDNRA Linked to migraine 7tm_1
resistance
OR4F15 7tm_1
EDG7 7tm_1
GPR92 7tm_1
AGTR2 Linked to mental 7tm_1
retardation
GRM7 ANF_receptor, NCD3G,
7tm_3
GPR84 7tm_1
P2RY1 7tm_1
Transcription PLAGL1 Mutation causes Znf_C2H2
Beckwith–Wiedeman
syndrome
EZH1 SANT, SET
TEF Maps to pigment BRLZ
mutations in mice
GATAD2A
ILF2 DZF
SMARCC2 CHROMO, SWIRM, SANT
Pigment TYR Albinism Tyrosinase
BMP1 ZnMc, CUB, EGF_CA
Phospholipid signaling PNPLA4 Patatin
ZFYVE1 FYVE
ITPK1 Maps near SNPs linked Ins134_P3_kin
to pigmentation
PLCXD1 PLCc
NRGN IQ
PLEKHA1 Linked to age related PH
macular degeneration
Ras family RAB4A RAB
GTPase HRASLS NC
ARL4A ARF, small_GTPase
ZDHHC9 Linked to mental zf-DHHC
retardation
C5ORF5 RhoGAP
ARHGEF11 PDZ, RGS, PH, RhoGEF
KLC4 Rab5-bind, TPR
Protease inhibitor SERPINB2 SERPIN
WFDC8 WAP, KU
SERPINE1 SERPIN
SERPINB1 SERPIN
Metabolism NT5E Metallophos, 5_nucleotid_C
G6PC3 AcidPPc
UROD URO-D
(continued)
236 H. Ho et al.
Table 1 (continued)
Mutation causes
porphyria cutanea
tarda
HPD Mutation causes Glyoxalase
tyrosinemia type III
ALDH9A1 Aldedh
PLTP BPI1, BPI2
MSRA Downregulated in vitiligo PMSR
(hypopigmentation)
SMOX Amino_oxidase, DAO
UEVLD UBCc, Ldh_1_N, Ldh_1_C
GMPPB NTP_transferase, Hexapep
ALDH1A1 Expression lost in Aldedh
Parkinson’s disease
MGC4172 adh_short, Epimerase, KR
ENO2 Enolase_N, Enolase_C
Protein phosphorylation NLK S_TKc
PKN2 Hr1, C2, S_TKc, S_TK_X
RIOK1 RIO
PPP1R15A Expression lost in
melanoma
transformation
PPP2CB PP2Ac
Helicase RTEL1 DEXDc, HELICc
LOC389901 Ku, SAP DNA bd
Peptidase ARTS-1 Peptidase_M1
KLK13 Tyrp_SPc
LYZ Amyloidosis LYZ1
ADAM19 Pep_M12B_propep,
Reprolysin, DISIN,
ACR, EGF_2
CPZ FRI, Zn_pept
TRY1 Tryp_SPc
SENP1 DSS1_SEM1
SHFM1 Split hand/foot Peptdiase_C48
malformation
Translation EEF1A1 GTP_EFTU,
GTP_EFTU_D2,
GTP_EFTU_D3
VARS2 tRNA-synt_1, Anticodon_1
Other unknown NPM3 Nucleoplasmin
STX18 Syntaxin
KRTAP4- Keratin_B2
11
FGF23 Overexpressed in FGF
hyperpigmentation
syndrome
SFRS2 RRM
SLC17A5 Mutation causes Salla MFS_1
disease
USHBP1
(continued)
Table 1 (continued)
UBE2V1 UBCc
TEX11 TPR_2
TANC2 ANK, TPR
FATE1
LRRC1 LRR
RTN3 Reticulon
SPATA22
ETAA1 Tumor antigen,
melanoma of soft
parts
c12orf49
FAM125B
HSPC049 WD40
AFAP1L2 PH
FLJ41423
MAGEA6 Melanoma antigen MAGE
MUC3b EGF, SEA
C1orf194 NuA4
FAM89B
Our siRNA-based screening approach identified 98 siRNAs that significantly inhibited pigment
production. Four of these 98 genes did not retest, while two of the 98 genes were removed from the
Refseq database. Gene Ontology databases were used to segregate the 92 remaining genes into
function classes and identify conserved domains within the corresponding proteins. Gene ontology
databases, OMIM, and Pubmed searching was utilized to identify associations between our genes
and human diseases. This table was taken with permission from Ganesan et al. (2008)
Initial studies utilized tyrosinase siRNA to set a threshold for hit determination.
Using this approach, 96 genes were identified that significantly impacted pigment
production in MNT-1 cells (Table 1).
4 Integration of Multiple Systems-Level Approaches to Uncover

Additional Regulators of Melanogenesis in Our Functional
Genomics Dataset
Upon completion of the screen, our first task was to identify those genes that
significantly impacted melanin production. Determination of hit threshold in
RNAi screens is a somewhat arbitrary process relying on a combination of statisti-
cal parameters and information about known genes that regulate the pathway under
study. Examination of our RNAi dataset suggested that our approach had a high
false negative rate, as many known regulators of melanogenesis did not score as
“hits” using our scoring threshold although they did have some impacts on pigment
production (Z-score less than 1) (Ganesan et al. 2008). While our screen had an
apparent high false negative rate, it also likely had a high false positive rate similar
238 H. Ho et al.
to other RNAi screens. Our first challenge was to identify those targets that most
likely impacted melanogenesis.
To adjust our hit threshold, we initially examined our target list to see if our hits
were components of known melanogenesis pathways or were components of other
novel pathways that regulate melanogenesis. While annotation analysis of our hits
did define a novel melanogenesis regulatory pathway (autophagy) that impacts
melanogenesis, the majority of our hits had no apparent connection with known
melanogenesis regulatory pathways, nor were they components of discrete biological
machines that could be linked to melanogenesis. Therefore, we sought to develop
better methods to identify novel melanogenesis regulators and novel melanogenesis
regulatory pathways hidden in our screen dataset.
To uncover novel melanogenesis regulatory pathways hidden within our dataset,
we integrated multiple systems-level approaches to identify additional genes in our
screen that may have impacts on melanogenesis. For this purpose, we utilized two
different PPI network algorithms to better identify “hits” in our RNAi screen.
Recent studies have developed algorithms (RNAicut) (Kaplow et al. 2009) to adjust
the score threshold in RNAi screens by integrating protein–protein interaction
network connectivity algorithms with functional genomics datasets. These
approaches rely on the contention that true hits in RNAi screens are more densely
connected within the PPI network. RNAiCut computes the edge count of the
induced subgraph for a given list of genes and estimates the p-value for obtaining
a PPI subgraph of at least that size within the PPI network. The program then
computes a cutoff based on the distribution of p-values. To better define the
spectrum of biological processes that regulate melanogenesis identified in our
screen, we utilized the RNAicut algorithm to identify GO biological process that
were enriched in our RNAi screen (Fig. 2). For this analysis, we entered all genes
that had a Z score less than 1 into the RNAicut algorithm (2,583 genes) and
identified 2,497 genes that met the cutoff. This cutoff (Z-score < 1 or >1) was
selected because this list contained other known regulators of melanogenesis
(Ganesan et al. 2008). We examined the GO biological process data for each of
these genes and sorted the genes into respective categories. While the majority of
genes did not sort into discrete functional categories, we did identify a relative
enrichment of genes that are components of intracellular signaling pathways (7.7%)
(Fig. 2). The finding that a large proportion of these genes are G-protein-coupled
receptors (4.3%) is consistent with other studies showing that G-protein-coupled
receptors and their signaling intermediates play a critical role in regulating mela-
nogenesis (Van Raamsdonk et al. 2004). Our hits were also enriched in transcrip-
tional regulators, further illustrating the complexity of transcriptional networks
that impact melanogenesis (Steingrimsson et al. 2004). Additionally, we identified
components of protein transport machinery and metabolic machinery that do
impact melanogenesis, consistent with the established role of metabolic pathways
(Scislowski et al. 1985) and protein transport (Raposo and Marks 2007) in melanin
synthesis. While the RNAicut algorithm did identify some biological pathways that
were enriched in our screen, it was not able to significantly pare down our list of
targets. Future validation studies will determine if the biological pathways
Angiogenesis
Apoptosis
Ion transport
Cell adhesion
Cell cycle
Cell differentiation
Cell proliferation
Cell-Cell signaling
Chromatin metabolism
Development/ Morphogenesis
DNA Metabolism
Electron transport
Epidermis development
Intracellular metabolism
Microtubule-based movement
Neurogenesis
Protein folding
Protein modification
Protein transport
Proteolysis and peptidolysis
Transcription regulation
RNA processing
RNA splicing
Signaling pathways
Ubiquitin
Other
Fig. 2 Identification of novel regulators of melanogenesis by coupling our RNAi dataset with PPI
network connectivity algorithms. To further identify novel regulators of melanogenesis, siRNAs
that negatively impacted melanogenesis (Z-score less than 1, 2,583 genes) were inputted into the
RNAicut Web-based program. Using the established p-value cutoff, 2,497 genes were identified
that were highly connected in the PPI network. GO biological process annotation data was utilized
to sort the genes into functional categories. A large number of genes were involved in either
transcription or intracellular signaling pathways
identified by RNAicut are relevant to melanogenesis. However, the identification of

additional hits and novel components of known pathways from our screen data
would likely require other methods.
Next, we sought to see if we could sort our RNAi targets into pathways using
other mechanisms. While some studies have indicated that proteins that are more
densely connected in PPI networks are more likely to have the same function, other
studies have demonstrated that proteins in topologically similar node neighbor-
hoods are more likely to have similar function. Przulj and colleagues have previ-
ously developed a sensitive graph theoretic method for comparing local structures
of node neighborhoods and demonstrated that in PPI networks, biological function
of a node and its local network structure are closely related (Milenkovic and Przulj
2008). The method summarizes a protein’s local topology in a PPI network into its
“signature.” Then, signature similarities between all protein pairs are computed,
measuring topological resemblance of their neighborhoods. Initial studies deter-
mined whether this approach could identify genes contained within our siRNA
functional genomics datasets that were topologically similar to known cancer
240 H. Ho et al.
genes. Using this topology-based approach, we were able to specifically identify

cancer genes in our functional genomics dataset that were negative regulators of
melanogenesis (Milenkovic et al. 2009), consistent with previous findings demon-
strating that cancer genes are negative regulators of melanogenesis (Halaban 2000).
More recent collaborations between our laboratory and the Przulj laboratory have
attempted to determine if topologically similar proteins are more likely to be
components of the same molecular pathways. In this analysis, we sought to deter-
mine whether we could use this topology-based approach to identify novel compo-
nents of known melanin regulatory pathways in our functional genomics dataset.
We applied the topologic methodology described above to the entire human PPI
network of 47,303 interactions amongst 10,282 proteins obtained by the union of
physical PPIs from BIOGRID (Stark et al. 2006), HPRD (Peri et al. 2004), OPHID
(Brown and Jurisica 2005), and (Rual et al. 2005) to examine the global network
properties (degrees, clustering coefficients, and eccentricities) of novel gene targets
that regulate pigmentation. 2,178 novel pigment regulators (Z-score less than 1 or
greater than 1) identified in our screen were present in the PPI network, and 1,244
genes were in a dense part of the network (graphlet degree signature of at least
four), allowing us to examine their topological similarity. Future studies will
examine topologically similar clusters within these genes to determine if they are
components of the same pathway and determine how these pathways regulate
melanogenesis. To validate that we had identified components of pathways that
regulated melanogenesis, we sought to determine if any of the clusters identified by
our integrative approach contained known regulators of melanogenesis.
One surprising result of our RNAi screen was that we identified only two known
regulators of melanogenesis in our top tier of candidate genes, making it difficult to
determine the mechanism by which these new pigment regulators impacted pig-
ment production (Ganesan et al. 2008). Our screen did identify several known
pigment regulators that had some impact on pigment production (Z-score less than
1) (Ganesan et al. 2008), suggesting that known regulators of pigment production
did impact pigmentation in our screen. To determine which of our novel melano-
genesis regulators were components of known melanin regulatory pathways, we
examined network properties of known melanogenesis regulators and identified
those that were located in a dense part of the PPI network (graphlet degree signature
of at least four). For each of these 38 genes, we identified a cluster containing the
target node and all other nodes in the network that are topologically similar to it, i.
e., that have signature similarities above a certain threshold. Thirteen of these 38
clusters contain targets identified in our siRNA screen; 2 of these 13 have a low p-
value for enrichment in putative targets. A large cluster containing 59 putative
target proteins clustering around ADAM-17 (Bennett and Lamoreux 2003) was
identified, as well as a smaller cluster containing 12 putative target proteins
clustering around EDNRB (Hou et al. 2004). ADAM-17, or TACE, is an extracel-
lular enzyme that cleaves m-KIT to s-KIT, inhibiting signaling through the
SCF–KIT pathway (Kasamatsu et al. 2008). Inhibitors of TACE increase pigment
production, indicating that TACE is a negative regulator of pigmentation (Kasamatsu
et al. 2008). EDNRB, a G-protein-coupled receptor that is mutated in a human
pigmentary disorder, binds to EDN3 (Hofstra et al. 1996) and activates an intracel-
lular signaling cascade via the G-protein GNA11 (Van Raamsdonk et al. 2004).
Mouse knockout studies revealed that EDNRB impacts MITF expression and
function (Hou and Pavan 2008). Recent studies have determined that the novel
gene targets identified in our screen that are topologically similar to EDNRB impact
MITF expression and are components of pathways downstream of EDNRB
(Ho et al. 2010). Complementation strategies validated that these novel genes
were components of the EDNRB pathway in MNT-1 melanoma cells and normal
melanocytes (Ho et al. 2010). Together, these studies provide compelling evidence
that topologically similar proteins are more likely components of the same molecu-
lar pathway, providing an additional method to identify pathway components from
RNAi datasets.
5 Identification of Novel Pathways that Regulate

the Transcription of Melanogenic Enzymes
The studies outlined in chapter ‘RNAi Suppression and its Application’ by authors
Xiaoping Yi and Rui Lu sought to use novel approaches to improve the identifica-
tion of score cutoffs in high-throughput RNAi screens. Before we could investigate
the utility of these approaches in the context of our screen, we first needed to
validate hits identified in our screen experimentally to determine the false positive
rate of our approach. To facilitate the identification of genes that significantly
impacted melanogenesis within our dataset, we utilized the values obtained for
tyrosinase siRNA, the enzyme that is the rate-limiting step in melanogenesis, to
determine a cutoff to identify hits in our analysis. In our initial analysis, tyrosinase
siRNA had a normalized pigment ratio 2.5 standard deviations below the mean.
Therefore, an arbitrary threshold of 2 standard deviations below the mean was
utilized to identify siRNAs that significantly impact pigment production. This
cutoff identified 96 novel genes that impact pigment production in MNT-1 cells
(Table 1). Extensive validation studies revealed that a significant proportion of
these genes were true regulators of pigment production. Initial retesting of 35 of the
96 siRNA pools identified in the screen revealed a false positive rate of 12.1%. The
high true positive rate of these validation studies indicate that our cutoff threshold
was too stringent, indicating that many other valid targets exist between Z-score
2 and Z-score 1. Indeed, siRNAs directed towards several known regulators of
melanogenesis had Z-scores between 2 and 1.
A pool deconvolution strategy was utilized to identify siRNA phenotypes that
were a consequence of the off-target effects of an individual siRNA. Briefly, we
retested each of the four siRNAs from each siRNA pool to determine if it had a
significant impact on pigment production. Relative pigmentation was assessed
spectrophotometrically, and the relative inhibition of pigment production was deter-
mined relative to tyrosinase siRNA (normalized percent inhibition). These studies
revealed that for each gene, two or more siRNAs were able to recapitulate the
242 H. Ho et al.
phenotype of the pooled siRNA. These findings led us to conclude that none of the
phenotypes observed in our screen were a consequence of siRNA off-target effects.
The low off-target rate of our screen is likely secondary to the specificity of our assay,
which was able to effectively exclude siRNAs that would have had off-target effects
on cellular survival pathways. To lend further credence to the specificity of our
approach, we determined if the pooled siRNAs visually inhibited pigment produc-
tion. Many of our pooled siRNAs inhibited pigment production both macroscopically
and microscopically (Ganesan et al. unpublished observations), further demonstrat-
ing the specificity of our results. In summary, while our results clearly demonstrated
that the threshold utilized in this screen to identify hits was enriched in true-positives
(Ganesan et al. 2008), the cutoff also had a high false negative rate given the fact that
many known regulators of pigment production had Z-scores between 1 and 2 and
were not identified as hits. Future studies will develop additional methods to identify
relevant false negatives from our screen dataset.
Analysis of GO annotation data for the list of pigment regulators exposed a wide
variety of cellular processes represented by the validated and candidate hits, but it
did not give a clue as to the direct mechanism by which these genes impact pigment
production. Two areas of pigment regulation are less characterized – the signaling
pathways that lead to upregulation of melanin synthesis and the molecular compo-
nents that regulate the early phases of melanosome biogenesis. Therefore, we
sought to identify genes that impact either melanosome biogenesis or that upregu-
late the transcription of melanogenic enzymes. We first employed a focused
approach to identify signaling intermediates that impact the production of tyrosi-
nase, the rate-limiting enzyme specifying melanogenesis (Kim and Uyama 2005)
among novel validated genes. Relative accumulation of tyrosinase, the key tran-
scription factor MITF, and the melanosomal marker protein Melan-A were exam-
ined 96 h post siRNA transfection. Remarkably, over half of the validated pigment
genes appear to be required for tyrosinase protein accumulation. Of those pigment
genes impacting tyrosinase accumulation, approximately half appear to act at the
level of tyrosinase mRNA accumulation, and most of these also impaired MITF
mRNA accumulation (Ganesan et al. 2008). Given that tyrosinase is an MITF target
gene, the pigmentation genes in this later class likely represent action at the level of
MITF mRNA.
Seven genes were identified with significant impacts on MITF and tyrosinase
expression (Table 2) – Aldh1a1, Aldh9a1, Wipi1, SerpinB2, Plekha1, Itpk1, and
tyrosinase. Aldehyde dehydrogenases are a class of enzymes that detoxify lipid
aldehydes produced as a result of UV stress (Downes et al. 1997). The finding that
these genes impact pigment production provides an additional functional link
between UV stress and melanin production. Mutations in Plekha1 have been
linked to age-related macular degeneration, an ocular condition that is character-
ized by defective melanogenesis in the macula (Leveziel et al. 2007). SerpinB2 is
known to protect retinoblastoma protein, a known regulator of melanogenesis,
from degradation (Tonnetti et al. 2008). Itpk1 is a key integrator of inositol
phosphate pathways, and inositol signaling intermediates are known to impact
melanogenesis (Shears 2009).
Table 2 Genome-wide siRNA screening identifies targets that differentially impact tyrosinase
and MITF expression
Phenotype Symbol MITF Tyrosinase Melan A
▾ TYR and MITF TYR ▾ RNA ▾ RNA NO CHANGE
WIPI1 ▾ RNA ▾ RNA ▾ PROTEIN
ALDH1A1 ▾ RNA ▾ RNA NO CHANGE
ALDH9A1 ▾ RNA ▾ RNA NO CHANGE
PLEKHA1 ▾ RNA ▾ RNA ▾ PROTEIN
RAB4A ▾ RNA ▾ RNA ~PROTEIN
SERPINB2 ▾ RNA ▾ RNA ~PROTEIN
MSRA ▾ RNA ▾ RNA NO CHANGE
NPM3 ▾ RNA ▾ RNA ~PROTEIN
▾ MITF mRNA ARHGEF11 ▾ RNA ▾ PROTEIN NO CHANGE
▾ TYR protein ZDHHC9 ▾ RNA ▾ PROTEIN NO CHANGE
ITPK1 ▾ RNA ▾ PROTEIN ▾ PROTEIN
▾ MITF mRNA AGTR2 ▾ RNA NO CHANGE NO CHANGE
▾ TYR protein PPP1R15A ▾ PROTEIN ▾ PROTEIN NO CHANGE
ZFYVE1 NO CHANGE ▾ PROTEIN NO CHANGE
▾ MITF protein GNG2 ▾ PROTEIN NO CHANGE NO CHANGE
No change in MITF EDNRA NO CHANGE NO CHANGE ~PROTEIN
or TYR SMARCC2 NO CHANGE NO CHANGE NO CHANGE
FLJ1123 NO CHANGE NO CHANGE NO CHANGE
UROD NO CHANGE NO CHANGE NO CHANGE
UEV3 NO CHANGE NO CHANGE NO CHANGE
ARL4A NO CHANGE NO CHANGE NO CHANGE
P66A NO CHANGE NO CHANGE NO CHANGE
OR4F15 NO CHANGE NO CHANGE NO CHANGE
Western blotting and quantitative RT-PCR was used to identify siRNAs that impact tyrosinase,
MITF, and Melan-A protein levels or impact tyrosinase and MITF mRNA levels in MNT-1 cells.
siRNAs that significantly impacted the expression of MITF and tyrosinase mRNA as determined
by quantitative RT-PCR (p < 0.05 by Student’s t-test) and protein as determined by western
blotting, or siRNAs that only impacted protein accumulation as determined by western blotting
(densitometry values less than 50%) are shown. Genes are sorted into several phenotypes: genes
that regulate tyrosinase and MITF protein and mRNA accumulation, genes that regulate MITF
mRNA accumulation but only tyrosinase protein accumulation, genes that regulate MITF mRNA
accumulation but not tyrosinase or melan-a protein accumulation, genes that regulate protein but
not mRNA accumulation of tyrosinase or MITF, and genes that did not impact protein accumula-
tion of tyrosinase or MITF. This table was taken with permission from Ganesan et al. (2008)
An interesting observation from our screen was the identification of potential

feedback loops that regulate melanogenesis. Wipi1 is a known component of
autophagosomes (Proikas-Cezanne et al. 2004), and our published validation stud-
ies provide convincing evidence that autophagy is a novel pathway that regulates
melanogenesis (see below for a more complete description). Recent studies indicate
that Wipi1 impacts both melanosome formation (Ho et al. unpublished observa-
tions) and the transcription of MITF (Ganesan et al. 2008). Taken together, these
studies are consistent with the hypothesis that inhibition of melanosome formation
has consequences on the transcription of melanogenic enzymes. Similarly, our
screen determined that tyrosinase siRNA impacts both MITF and tyrosinase expres-
sion, suggesting that tyrosinase itself may regulate its own transcription via MITF.
244 H. Ho et al.
Pmel17 GFP Phase Contrast
Tyr shRNA
Ctl shRNA
Tyr
Actin
DAPI Merge
Fig. 3 Tyrosinase shRNA inhibits melanosome formation and pigment production in pigmented
melanoma cells. MNT-1 melanoma cells were infected with a pGIPZ lentivirions encoding a
tyrosinase shRNA downstream of a IRES-GFP to generate a mixed population of cells that express
or do not express tyrosinase shRNA. In this system, green cells express tyrosinase shRNA while
cells that are not infected. Subsequent to infection, cells were fixed and stained with Pmel17
antibody (a melanosome marker, red) and DAPI and visualized by phase contrast and fluorescence
microscopy at high magnification (63) using equivalent exposure times. shRNA-expressing cells
(green cells, top central panel) express low levels of Pmel17 staining (top left panel, bottom
middle panel) when compared to cells that do not express the shRNA. Additionally, green cells
lack melanosome granules (dark dots on phase contrast microscopy, top right panel). MNT-1 cells
infected with tyrosinase shRNA lentivirions or control shRNA containing lentivirions were
harvested and subjected to immunoblotting with tyrosinase and actin antibodies. While tyrosinase
shRNA samples have a higher amount of protein loaded (see actin loading control), they do not
contain any tyrosinase signal
To validate these findings using another experimental system, we sought to deter-

mine if tyrosinase shRNA can inhibit the melanin biogenesis pathway in MNT-1
cells (Fig. 3). We observed that tyrosinase shRNA-expressing cells are melano-
some-deficient as indicated by Pmel17 staining, indicating that tyrosinase itself
plays a role in regulating the pigment cell pathway. Future studies will focus on
characterizing this mechanism of feedback regulation.
6 Identification of Novel Pathways that Regulate

Melanosome Biogenesis
Four genes characterized in our initial publication impacted tyrosinase accumula-

tion without impacting tyrosinase or MITF transcription, suggesting that inhibition
of the function of these genes induced tyrosinase degradation. Melanosomes are
distinct lysosome-related organelles dependent upon appropriate post-Golgi sorting
events for delivery of functionalizing “cargo” including tyrosinase (Raposo et al.
2007). Therefore, impaired accumulation of tyrosinase observed with these siRNAs
could be a consequence of missorting of tyrosinase to lysosomes and subsequent
hydrolysis in that organelle, or a consequence of increased proteasomal degradation
Zdhhc9
Control
Rab4a
Msra
Tyr
siRNA:
bafilomycin: – + – + – + – + – +
TYR
ERK 1/2
Fig. 4 Identification of melanogenesis regulators that impact protein transport. MNT-1 cells
transfected with siRNAs that impact tyrosinase turnover (50 nM) were incubated in the presence
and absence of bafilomycin A1 for 24 h prior to lyses and analyses of tyrosinase protein accumu-
lation. All results shown are representative of a minimum of three independent experiments. This
figure was taken with permission from Ganesan et al. (2008)
of tyrosinase. Initial studies determined that proteasome inhibitors were unable to

rescue the impact of these siRNAs on tyrosinase accumulation (data not shown). To
identify genes that may direct tyrosinase to lysosomal degradation, lysosome
acidification was inhibited by bafilomycin A2 exposure subsequent to target gene
depletion (Watabe et al. 2004). As shown in Fig. 4, a 24-h inhibition of lysosome
acidification rescued tyrosinase accumulation upon depletion of the small G-protein
RAB4A and the small G-protein palmitoyltransferase ZDHHC9. By contrast,
bafilomycin did not restore tyrosinase accumulation upon depletion of MSRA, a
protein that can protect against oxidative damage through reduction of methionine
sulfoxide. Other studies have demonstrated that this protein is expressed at a lower
level in vitiligo melanocytes (Zhou et al. 2009). Rab4a is a known regulator of
recycling endosomes (Roosterman et al. 2004). Others have postulated that Pmel17
is expressed on the plasma membrane, suggesting that endosomal recycling of this
protein may be required to deliver it to the melanosome (Valencia et al. 2006). It is
less clear how Zdhhc9 impacts melanosomal trafficking, as these proteins act to
prenylate Ras proteins and have no direct connection with melanosomal trafficking
(Nadolski and Linder 2007). Nevertheless, these findings provide evidence that our
screen did identify molecular regulators of melanosome trafficking.
To more specifically identify other regulators of melanosome trafficking in our
target list, we utilized gene annotation data to identify other genes involved in
protein sorting/vesicle trafficking pathways. Among the panel of validated pigment
regulatory genes was WIPI1. WIPI1 has been implicated as a human homolog of
the yeast autophagy protein ATG18 and is localized to starvation-induced autop-
hagosomes in human cell culture (Proikas-Cezanne et al. 2004). As WIPI1 is a
known component of the autophagosome, we sought to determine whether other
autophagosome components impacted pigment production in our screen. Initial
evaluation of gene annotation data of our targets revealed that two other targets,
LC3-C and GPSM1/AGS3, were putative components of the autophagy pathway.
Autophagy, or cellular self-degradation, is a highly conserved cellular pathway that
has been associated with cancer formation, neurodegeneration, and aging (Levine
and Kroemer 2008; Mizushima et al. 2008; Huang and Klionsky 2007). Classically,
this pathway functions to transport vesicle cargo (autophagosomes) to the lysosome
for degradation (Mizushima et al. 2008). Recent studies have demonstrated that
246 H. Ho et al.
components of this pathway may also regulate endosomal trafficking, a trafficking

pathway that is critically important in melanosome formation (Liang et al. 2008).
Previous studies have documented an abundance of autophagosomes in cells
obtained from patients with a disorder of pigmentation (HPS-1) but have hypothe-
sized that their presence is a consequence of the degradation of immature melano-
somes within these cells (Smith et al. 2005). Similar autophagosome like structures
have been observed in vitiligo melanocytes (Boissy et al. 1983, 1991a, b) and
melanoma cells (Bomirski et al. 1987), suggesting that autophagosomes may be
required for the removal of abnormal melanosomes. On the other hand, other
studies suggest that autophagy may be central to the overall process of melanogen-
esis. Autophagosome components are present in the stage II melanosome, suggest-
ing that parts of the melanosome originate from the autophagosome (Basrur et al.
2003). Pharmacologic inhibitors of autophagy (choloroquine) inhibit melanin pro-
duction (Ni-Komatsu et al. 2008), suggesting that autophagy is required for normal
melanogenesis. Additionally, pharmacologic activators of autophagy stimulate
melanin production. Our studies suggest that depletion of autophagy regulators
inhibit melanogenesis, suggesting that autophagy is required for normal melano-
genesis. Therefore, we sought to investigate the link between autophagy and
melanogenesis.
Initial studies sought to determine if other regulators of autophagy also impacted
melanogenesis. Although they were not identified in our screen, siRNA-mediated
depletion of two additional components required to trigger autophagosome forma-
tion, BECN1 or LC3-A, severely impaired pigment accumulation. Failure to
recover these genes in the primary screen is indicative of the false negative rate
associated with RNAi screens. However, the fact that we were able to identify these
regulators by integrating available gene annotation data illustrates the utility of
integrating available bioinformatics databases with functional genomics datasets.
The finding that Becn1 impacted pigment accumulation of MNT-1 cells led to the
close examination of the impact of Becn1 on hair pigmentation in mouse models.
Heterozygous deletion of the autophagy protein Beclin 1 (Qu et al. 2003) resulted in
a dramatic coat color defect in mice. Interestingly, these mice also had a chimeric
phenotype with normal and hypopigmented hair follicles (Ganesan et al. 2008).
Other studies indicated that the observed pigmentary defects were not the
consequence of the impact of autophagy depletion on the survival of mealnocyte
precursors (melanoblasts). Taken together, our siRNA data and histologic analysis
suggests that the phenotype observed in the Beclin1 haploinsufficient mice is not a
consequence of impacts of Beclin1 on melanocyte survival but is more likely
secondary to the impact of beclin1 on melanosome number or melanin content
within the hair follicle (Ganesan et al. 2008). Consistent with these findings, we
found that autophagy proteins LC3 and APG5 colocalize with the melanosome
marker PMEL17 in mature melanosomes in both MNT-1 cells and normal mela-
nocytes. When coupled with previous data demonstrating that autophagosomes
accumulate in cells defective in melanosome maturation, these results indicate
that autophagy likely impacts melanosome maturation (Smith et al. 2005).
The most logical conclusion from our data is that melanocytes utilize the same
machinery required for autophagosome formation to form melanosomes. These

autophagosomes may function to segregate stage II melanosomes from the early
endosomes or may play a role in regulating multivesicular bodies, vesicles that
traffic mature proteins to melanosomes and also regulate the formation of stage II
melanosomes. Alternatively, the targets identified in the screen could function in a
process that is distinct from autophagy. WIPI1, a key regulators of melanogenesis
identified in our screen, is one of the few proteins that bind to PtdIns (3,5) P2 and is
important for retrieval of membrane from the vacuole in yeast (Obara et al. 2008).
PtdIns (3,5) P2 metabolism is altered in Charcot–Marie Tooth disease, a disorder
with hypopigmentation (Chow et al. 2007). Similarly, WIPI1 may regulate mela-
nosome trafficking by interacting with a specific PtdIns (3,5) P2-specific vesicle
that is required for melanosome biogenesis. Studies are currently underway to
determine whether these novel autophagy regulators impact melanogenesis via an
impact on autophagy or impact melanogenesis via a mechanism that is autophagy-
independent. Translational studies are underway to determine whether defects in
autophagy underlie the depigmenting disorders HPS1 and vitiligo.
7 Identification of Regulators of Human Pigment Variation
Pigment shade varies widely between and among human ethnic populations.
Genome-wide association studies have determined that only a fraction of these
differences appear to be regulated by differences in the expression of known
melanogenesis regulators (Sulem et al. 2008). Therefore, it is likely that the key
molecular regulators of differences in pigment shade have yet to be identified. To
determine if the novel genes identified in our screen may have impacts on
pigment shade, we sought to examine the impact of novel pigmentation genes,
identified in MNT-1 cells, on pigment production in melanocytes of diverse
genetic backgrounds. Remarkably, the majority of targets that regulated tyrosi-
nase expression in MNT-1 cells also impacted tyrosinase expression when
depleted from darkly pigmented primary melanocytes (see Fig. 5). Approximately
half of these targets also inhibited tyrosinase expression when depleted from
moderately pigmented melanocytes. These results indicate that the primary screen
identified a number of genes that impact pigment production in several different
genetic backgrounds. Selective activity of some of these targets in different
genetic backgrounds suggests that some of these novel regulators of melanogene-
sis may play a role in human phenotypic variation. This phenomenon could be
either secondary to varying activities of these pigment regulators in different
pigment backgrounds or differential expression of these genes in different pig-
ment backgrounds. Future large-scale studies are required to determine how these
genes differentially regulate pigment shade. Additional study is also required to
determine the impact of natural selection and evolution on these novel pigment
regulatory genes.
248 H. Ho et al.
MNT-1 DP-PHM
MSRA
NPM3
ALDH9A1 RAB4A ATGR2
SERPINB2 PLEKHA1 ARHGEF11
ITPK1
TYR
WIPI1
ALDH1A1
ZDHHC9
MP-PHM
Fig. 5 Identification of putative regulators of human pigment variation. The indicated siRNAs,
targeting novel pigmentation genes identified in the MNT-1 screen, were tested for consequences
on tyrosinase accumulation in darkly pigmented primary human melanocyte (DP-PHM) and
moderately pigmented primary human melanocyte (MP-PHM) cultures 6 days posttransfection.
The results presented here is a Venn diagram of the data presented in Ganesan et al. (2008),
demonstrating that we have identified pigment regulators that differentially impact pigment
production in different genetic backgrounds. This figure was taken with permission from Ganesan
et al. (2008)
8 Identification of Pharmacologic Agents that Impact

Melanin Accumulation
A fundamental goal of our RNAi screen was to identify gene targets to facilitate the
rationale design of novel depigmenting agents. As a first step, we sought to
determine if our gene list contained “druggable” targets for which known inhibitors
are available. Among our gene targets were two isoforms of aldehyde dehydroge-
nase, ALDH1A1 and ALDH9A1, enzymes that regulate ethanol detoxification
(Edenberg 2007) and response to UV stress (Downes et al. 1997). A number of
chemical inhibitors of these enzymes have been identified (DeMaster et al. 1998),
and several of these agents are clinically utilized to treat alcoholism, presenting an
opportunity to pharmacologically validate our screen findings. Two Aldh inhibitors,
cyanamide and Angeli’s salt (DeMaster et al. 1998), inhibited pigmentation and
tyrosinase protein accumulation in MNT-1 cells at doses that are equivalent to those
required to inhibit Aldh activity in culture (Ganesan et al. 2008). In addition, these
compounds impaired UV-induced tyrosinase expression when tested in primary
melanocytes (Ganesan et al. 2008). Cyanamide has been clinically used for alcohol
detoxification in Europe for decades, and abundant clinical safety data is readily
available for this drug. By coupling existing pharmacologic databases with our
RNAi dataset, we have been able to identify important novel depigmenting agents
that can be rapidly translated from the bench to the bedside. Together, these studies
demonstrate the utility of coupling pharmacologic inhibitor databases and RNAi
datasets to identify novel treatments for clinical disorders and illustrate the potential
translational nature of RNAi technology.
9 Concluding Remarks
RNAi-based functional genomics is a systems-level approach to identify genes that

impact human disease phenotypes. We have utilized RNAi technology to uncover
the underappreciated molecular complexity that regulates melanogenesis in human
cells. These initial studies offered a great challenge – how do we utilize this
information to identify the spectrum of pathways that regulate melanogenesis? In
this chapter, we discuss the integration of multiple systems-level approaches to
answer these questions. We demonstrate that integrating protein–protein interaction
networks with our RNAi dataset is a potential mechanism to uncover additional
melanin regulatory components within our dataset. We utilize both siRNA and
shRNA technology to validate novel feedback mechanisms that regulate melano-
genesis. Coupling our target list with pharmacologic databases, we have identified
novel inhibitors of pigmentation ready for clinical application. Through these
studies, we present a framework of how to integrate other systems-level datasets
with RNAi datasets to create a more detailed molecular landscape of factors that
regulate individual biological phenotypes. Future studies will validate these
approaches utilizing siRNA and shRNA technology to further elucidate novel
pathways that regulate melanogenesis.
References
Anderson TM, vonHoldt BM, Candille SI et al (2009) Molecular and evolutionary history of
melanism in North American gray wolves. Science 323:1339–1343
Aspengren S, Hedberg D, Skold HN et al (2009) New insights into melanosome transport in
vertebrate pigment cells. Int Rev Cell Mol Biol 272:245–302
Basrur V, Yang F, Kushimoto T et al (2003) Proteomic analysis of early melanosomes: identifica-
tion of novel melanosomal proteins. J Proteome Res 2:69–79
Bennett DC, Lamoreux ML (2003) The color loci of mice – a genetic century. Pigm Cell Res
16:333–344, Sponsored by the European Society for Pigment Cell Research and the Interna-
tional Pigment Cell Society
Berson JF, Harper DC, Tenza D et al (2001) Pmel17 initiates premelanosome morphogenesis
within multivesicular bodies. Mol Biol Cell 12:3451–3464
Boissy RE, Beato KE, Nordlund JJ (1991a) Dilated rough endoplasmic reticulum and premature
death in melanocytes cultured from the vitiligo mouse. Am J Pathol 138:1511–1525
250 H. Ho et al.
Boissy RE, Liu YY, Medrano EE et al (1991b) Structural aberration of the rough endoplasmic
reticulum and melanosome compartmentalization in long-term cultures of melanocytes from
vitiligo patients. J Invest Dermatol 97:395–404
Boissy RE, Smyth JR Jr, Fite KV (1983) Progressive cytologic changes during the development
of delayed feather amelanosis and associated choroidal defects in the DAM chicken line.
A vitiligo model. Am J Pathol 111:197–212
Bomirski A, Wrzolkowa T, Arendarczyk M et al (1987) Pathology and ultrastructural character-
istics of a hypomelanotic variant of transplantable hamster melanoma with elevated tyrosinase
activity. J Invest Dermatol 89:469–473
Breathnach AS (1988) Extra-cutaneous melanin. Pigment Cell Res 1:234–237
Brown KR, Jurisica I (2005) Online predicted human interaction database. Bioinformatics
21:2076–2082
Chow CY, Zhang Y, Dowling JJ et al (2007) Mutation of FIG4 causes neurodegeneration in the
pale tremor mouse and patients with CMT4J. Nature 448:68–72
Costin GE, Hearing VJ (2007) Human skin pigmentation: melanocytes modulate skin color in
response to stress. FASEB J 21:976–994
Cui R, Widlund HR, Feige E et al (2007) Central role of p53 in the suntan response and pathologic
hyperpigmentation. Cell 128:853–864
Delevoye C, Hurbain I, Tenza D et al (2009) AP-1 and KIF13A coordinate endosomal sorting and
positioning during melanosome biogenesis. J Cell Biol 187:247–264
Delezoide AL, Vekemans M (1994) Waardenburg syndrome in man and splotch mutants in
the mouse: a paradigm of the usefulness of linkage and synteny homologies in mouse and
man for the genetic analysis of human congenital malformations. Biomed Pharmacother 48:
335–339
DeMaster EG, Redfern B, Nagasawa HT (1998) Mechanisms of inhibition of aldehyde dehydro-
genase by nitroxyl, the active metabolite of the alcohol deterrent agent cyanamide. Biochem
Pharmacol 55:2007–2015
Di Pietro SM, Falcon-Perez JM, Tenza D et al (2006) BLOC-1 interacts with BLOC-2 and the
AP-3 complex to facilitate protein trafficking on endosomes. Mol Biol Cell 17:4027–4038
Double KL (2006) Functional effects of neuromelanin and synthetic melanin in model systems.
J Neural Transm 113:751–756
Double KL, Halliday GM (2006) New face of neuromelanin. J Neural Transm 70:119–123
Downes JE, Swann PG, Holmes RS (1997) A genetic basis for corneal sensitivity to ultraviolet
light among recombinant SWXJ inbred strains of mice. Curr Eye Res 16:539–546
Edenberg HJ (2007) The genetics of alcohol metabolism: role of alcohol dehydrogenase and
aldehyde dehydrogenase variants. Alcohol Res Health 30:5–13
Etchevers HC, Amiel J, Lyonnet S (2006) Molecular bases of human neurocristopathies. Adv Exp
Med Biol 589:213–234
Fowler DM, Koulov AV, Alory-Jost C et al (2006) Functional amyloid formation within mammalian
tissue. PLoS Biol 4:e6
Ganesan AK, Ho H, Bodemann B et al (2008) Genome-wide siRNA-based functional genomics of
pigmentation identifies novel genes and pathways that impact melanogenesis in human cells.
PLoS Genet 4:e1000298
Goding CR (2007) Melanocytes: the new black. Int J Biochem Cell Biol 39:275–279
Gorner H (1994) Photochemistry of DNA and related biomolecules: quantum yields and con-
sequences of photoionization. J Photochem Photobiol B 26:117–139
Grimes P, Nordlund JJ, Pandya AG et al (2006) Increasing our understanding of pigmentary
disorders. J Am Acad Dermatol 54:S255–S261
Halaban R (2000) The regulation of normal melanocyte proliferation. Pigment Cell Res 13:4–14
Ho H, Milenkovic T, Memisevic V et al (2010) Protein interaction network topology uncovers
melanogenesis regulatory network components within functional genomics datasets. BMC Sys
Bio, in press
Hoek K, Rimm DL, Williams KR et al (2004) Expression profiling reveals novel pathways in the
transformation of melanocytes to melanomas. Cancer Res 64:5270–5282
Hoekstra HE (2006) Genetics, development and evolution of adaptive pigmentation in vertebrates.
Heredity 97:222–234
Hofstra RM, Osinga J, Tan-Sindhunata G et al (1996) A homozygous mutation in the endothelin-3
gene associated with a combined Waardenburg type 2 and Hirschsprung phenotype (Shah–
Waardenburg syndrome). Nat Genet 12:445–447
Hou L, Pavan WJ (2008) Transcriptional and signaling regulation in neural crest stem cell-derived
melanocyte development: do all roads lead to Mitf? Cell Res 18:1163–1176
Hou L, Pavan WJ, Shin MK et al (2004) Cell-autonomous and cell non-autonomous signaling
through endothelin receptor B during melanocyte development. Development (Cambridge,
England) 131:3239–3247
Huang J, Klionsky DJ (2007) Autophagy and human disease. Cell Cycle 6:1837–1849
Huizing M, Sarangarajan R, Strovel E et al (2001) AP-3 mediates tyrosinase but not TRP-1
trafficking in human melanocytes. Mol Biol Cell 12:2075–2085
Hume AN, Collinson LM, Rapak A et al (2001) Rab27a regulates the peripheral distribution of
melanosomes in melanocytes. J Cell Biol 152:795–808
Kadekaro AL, Kavanagh RJ, Wakamatsu K et al (2003) Cutaneous photobiology. The melanocyte
vs. the sun: who will win the final round? Pigment Cell Res 16:434–447
Kaplow IM, Singh R, Friedman A et al (2009) RNAiCut: automated detection of significant genes
from functional genomic screens. Nat Methods 6:476–477
Kasamatsu S, Hachiya A, Higuchi K et al (2008) Production of the soluble form of KIT, s-KIT,
abolishes stem cell factor-induced melanogenesis in human melanocytes. J Invest Dermatol
128:1763–1772
Kim YJ, Uyama H (2005) Tyrosinase inhibitors from natural and synthetic sources: structure,
inhibition mechanism and perspective for the future. Cell Mol Life Sci 62:1707–1723
Kushimoto T, Basrur V, Valencia J et al (2001) A model for melanosome biogenesis based on the
purification and analysis of early melanosomes. Proc Natl Acad Sci USA 98:10698–10703
Kwon BS, Kim KK, Halaban R et al (1994) Characterization of mouse Pmel 17 gene and silver
locus. Pigment Cell Res 7:94–397
Lamoreux ML, Wakamatsu K, Ito S (2001) Interaction of major coat color gene functions in mice
as studied by chemical analysis of eumelanin and pheomelanin. Pigment Cell Res 14:23–31
Leveziel N, Souied EH, Richard F et al (2007) PLEKHA1-LOC387715-HTRA1 polymorphisms
and exudative age-related macular degeneration in the French population. Mol Vis
13:2153–2159
Levine B, Kroemer G (2008) Autophagy in the pathogenesis of disease. Cell 132:27–42
Levy C, Khaled M, Fisher DE (2006) MITF: master regulator of melanocyte development and
melanoma oncogene. Trends Mol Med 12:406–414
Lewitt PA (2008) Levodopa for the treatment of Parkinson’s disease. N Engl J Med 359:2468–2476
Liang C, Lee JS, Inn KS et al (2008) Beclin1-binding UVRAG targets the class C Vps complex to
coordinate autophagosome maturation and endocytic trafficking. Nat Cell Biol 10:776–787
Manneville JB, Etienne-Manneville S, Skehel P et al (2003) Interaction of the actin cytoskeleton
with microtubules regulates secretory organelle movement near the plasma membrane in
human endothelial cells. J Cell Sci 116:3927–3938
McGowan KA, Li JZ, Park CY et al (2008) Ribosomal mutations cause p53-mediated dark skin
and pleiotropic effects. Nat Genet 40:963–970
Meyskens FL Jr, Farmer PJ, Yang S et al (2007) New perspectives on melanoma pathogenesis and
chemoprevention. Recent Results Cancer Res 174:191–195
Milenkovic T, Memisevic V, Ganesan AK et al (2010) Systems-level cancer gene identification
from protein interaction network topology applied to melanogenesis-related functional geno-
mics data. J R Soc Interface 7(44):423–437
Milenkovic T, Przulj N (2008) Uncovering biological network function via graphlet degree
signatures. Cancer Inform 4:257–273
252 H. Ho et al.
Mizushima N, Levine B, Cuervo AM et al (2008) Autophagy fights disease through cellular

self-digestion. Nature 451:1069–1075
Murisier F, Guichard S, Beermann F (2007) The tyrosinase enhancer is activated by Sox10 and
Mitf in mouse melanocytes. Pigment Cell Res 20:173–184
Nachman MW (2005) The genetic basis of adaptation: lessons from concealing coloration in
pocket mice. Genetica 123:125–136
Nadolski MJ, Linder ME (2007) Protein lipidation. FEBS J 274:5202–5210
Ni-Komatsu L, Tong C, Chen G et al (2008) Identification of quinolines that inhibit melanogenesis
by altering tyrosinase family trafficking. Mol Pharmacol 74:1576–1586
Obara K, Sekito T, Niimi K et al (2008) The Atg18–Atg2 complex is recruited to autophagic
membranes via phosphatidylinositol 3-phosphate and exerts an essential function. J Biol Chem
283:23972–23980
Oetting WS, King RA (1999) Molecular basis of albinism: mutations and polymorphisms of
pigmentation genes associated with albinism. Hum Mutat 13:99–115
Ohlemiller KK, Rice ME, Lett JM et al (2009) Absence of strial melanin coincides with age-
associated marginal cell loss and endocochlear potential decline. Hear Res 249:1–14
Ozeki H, Ito S, Wakamatsu K et al (1996) Spectrophotometric characterization of eumelanin and
pheomelanin in hair. Pigment Cell Res 9:265–270
Palleroni A, Miller CT, Hauser M et al (2005) Predation: prey plumage adaptation against falcon
attack. Nature 434:973–974
Parra EJ (2007) Human pigmentation variation: evolution, genetic basis, and implications for
public health. Am J Phys Anthropol 45:85–105
Peri S, Navarro JD, Kristiansen TZ et al (2004) Human protein reference database as a discovery
resource for proteomics. Nucleic Acids Res 32:D497–501
Price ER, Horstmann MA, Wells AG et al (1998) Alpha-melanocyte-stimulating hormone
signaling regulates expression of microphthalmia, a gene deficient in Waardenburg syndrome.
J Biol Chem 273:33042–33047
Proikas-Cezanne T, Waddell S, Gaugel A et al (2004) WIPI-1alpha (WIPI49), a member of the
novel 7-bladed WIPI protein family, is aberrantly expressed in human cancer and is linked to
starvation-induced autophagy. Oncogene 23:9314–9325
Qu X, Yu J, Bhagat G et al (2003) Promotion of tumorigenesis by heterozygous disruption of the
beclin 1 autophagy gene. J Clin Invest 112:1809–1820
Raposo G, Marks MS (2007) Melanosomes – dark organelles enlighten endosomal membrane
transport. Nat Rev Mol Cell Biol 8:786–797
Raposo G, Marks MS, Cutler DF (2007) Lysosome-related organelles: driving post-Golgi com-
partments into specialisation. Curr Opin Cell Biol 19:394–401
Raposo G, Tenza D, Murphy DM et al (2001) Distinct protein sorting and localization to
premelanosomes, melanosomes, and lysosomes in pigmented melanocytic cells. J Cell Biol
152:809–824
Roosterman D, Cottrell GS, Schmidlin F et al (2004) Recycling and resensitization of the
neurokinin 1 receptor. influence of agonist concentration and Rab GTPases. J Biol Chem
279:30670–30679
Rual JF, Venkatesan K, Hao T et al (2005) Towards a proteome-scale map of the human
protein–protein interaction network. Nature 437:1173–1178
Sarangarajan R, Apte SP (2005) Melanization and phagocytosis: implications for age related
macular degeneration. Mol Vis 11:482–490
Sarangarajan R, Apte SP (2006) The polymerization of melanin: a poorly understood phenomenon
with egregious biological implications. Melanoma Res 16:3–10
Scislowski PW, Slominski A, Bomirski A et al (1985) Metabolic characterization of three hamster
melanoma variants. Neoplasma 32:593–598
Seo K, Mohanty TR, Choi T et al (2007) Biology of epidermal and hair pigmentation in cattle:
a mini-review. Vet Dermatol 18:392–400
Setaluri V (2000) Sorting and targeting of melanosomal membrane proteins: signals, pathways,
and mechanisms. Pigment Cell Res 13:128–134
Setty SR, Tenza D, Sviderskaya EV et al (2008) Cell-specific ATP7A transport sustains copper-
dependent tyrosinase activity in melanosomes. Nature 454:1142–1146
Setty SR, Tenza D, Truschel ST et al (2007) BLOC-1 is required for cargo-specific sorting from
vacuolar early endosomes toward lysosome-related organelles. Mol Biol Cell 18:768–780
Shears SB (2009) Molecular basis for the integration of inositol phosphate signaling pathways via
human ITPK1. Adv Enzyme Regul 49:87–96
Shotelersuk V, Gahl WA (1998) Hermansky–Pudlak syndrome: models for intracellular vesicle
formation. Mol Genet Metab 65:85–96
Slominski A, Tobin DJ, Shibahara S et al (2004) Melanin pigmentation in mammalian skin and its
hormonal regulation. Physiol Rev 84:1155–1228
Smit NP, Kolb RM, Lentjes EG et al (1998) Variations in melanin formation by cultured
melanocytes from different skin types. Arch Dermatol Res 290:342–349
Smit NP, Van der Meulen H, Koerten HK et al (1997) Melanogenesis in cultured melanocytes can
be substantially influenced by L-tyrosine and L-cysteine. J Invest Dermatol 109:796–800
Smith JW, Koshoffer A, Morris RE et al (2005) Membranous complexes characteristic of
melanocytes derived from patients with Hermansky–Pudlak syndrome type 1 are macroauto-
phagosomal entities of the lysosomal compartment. Pigment Cell Res 18:417–426
Stark C, Breitkreutz BJ, Reguly T et al (2006) BioGRID: a general repository for interaction
datasets. Nucleic Acids Res 34:D535–539
Stein WD (1955) Ultra-violet absorption investigation of melanins. Nature 175(4454):472
Steingrimsson E, Copeland NG, Jenkins NA (2004) Melanocytes and the microphthalmia
transcription factor network. Annu Rev Genet 38:365–411
Sugimoto M (2002) Morphological color changes in fish: regulation of pigment cell density and
morphology. Microsc Res Tech 58:496–503
Sulem P, Gudbjartsson DF, Stacey SN et al (2008) Two newly identified genetic determinants of
pigmentation in Europeans. Nat Genet 40:835–837
Theos AC, Tenza D, Martina JA et al (2005) Functions of adaptor protein (AP)-3 and AP-1 in
tyrosinase sorting from endosomes to melanosomes. Mol Biol Cell 16:5356–5372
Theos AC, Truschel ST, Tenza D et al (2006) A lumenal domain-dependent pathway for sorting to
intralumenal vesicles of multivesicular endosomes involved in organelle morphogenesis. Dev
Cell 10:343–354
Tonnetti L, Netzel-Arnett S, Darnell GA et al (2008) SerpinB2 protection of retinoblastoma
protein from calpain enhances tumor cell survival. Cancer Res 68:5648–5657
Valencia JC, Watabe H, Chi A et al (2006) Sorting of Pmel17 to melanosomes through the plasma
membrane by AP1 and AP2: evidence for the polarized nature of melanocytes. J Cell Sci
119:1080–1091
Van Raamsdonk CD, Fitch KR, Fuchs H et al (2004) Effects of G-protein mutations on skin color.
Nat Genet 36(9):961–968
Watabe H, Valencia JC, Yasumoto K et al (2004) Regulation of tyrosinase processing and
trafficking by organellar pH and by proteasome activity. J Biol Chem 279:7971–7981
Whitehurst AW, Bodemann BO, Cardenas J (2007) Synthetic lethal screen identification of
chemosensitizer loci in cancer cells. Nature 446:815–819
Wu X, Rao K, Bowers MB et al (2001) Rab27a enables myosin Va-dependent melanosome
capture by recruiting the myosin to the organelle. J Cell Sci 114:1091–1100
Yamaguchi Y, Brenner M, Hearing VJ (2007) The regulation of skin pigmentation. J Biol Chem
282:27557–27561
Zecca L, Bellei C, Costi P et al (2008) New melanic pigments in the human brain that accumulate
in aging and block environmental toxic metals. Proc Natl Acad Sci USA 105:17567–17572
Zecca L, Zucca FA, Albertini A et al (2006) A proposed dual role of neuromelanin in the
pathogenesis of Parkinson’s disease. Neurology 67:S8–S11
Zhou Z, Li CY, Li K et al (2009) Decreased methionine sulphoxide reductase A expression renders
melanocytes more sensitive to oxidative stress: a possible cause for melanocyte loss in vitiligo.
Br J Dermatol 161(3):504–509
mRNA Structure and its Effects
on Posttranscriptional Gene Silencing
Stephen I. Rudnick, Veenu Aishwarya, and Alan M.Gewirtz
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
2 A Structured Target Site Reduces AON and siRNA Activity In Vitro . . . . . . . . . . . . . . . . . . . 257
3 Analysis of Binding Affinity to mRNA and Rate Dependencies
on Concentration for AON and siRNA Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
4 AON and siRNA Guide Strand Have Equal Affinity for the Target mRNA . . . . . . . . . . . . . 261
5 AON and siRNA Display Apparent First and Zero Order Kinetics . . . . . . . . . . . . . . . . . . . . . . 261
6 For Full In Vitro Activity, siRNA Require Greater
Target Site Accessibility Than AON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
7 A Double-Stranded Target Site Greatly Reduces In Vitro PTGS Activity . . . . . . . . . . . . . . . 265
8 An AON That is More Effective Than the siRNA Against an Identical
Target In Vitro is Less Effective Against the Same Target In Vivo . . . . . . . . . . . . . . . . . . . . . . 265
9 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
10 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
Abstract Posttranscriptional gene silencing (PTGS) is a process by which a

protein’s synthesis is impaired by targeting its messenger RNA (mRNA). The
antisense oligodeoxynucleotide (AON) and RNA interference (RNAi) pathways
can both accomplish PTGS by hybridization of a reverse complementary oligonucle-
otide and subsequent enzymatic degradation of the mRNA by an RNase mediated
mechanism. We have investigated the influence of specific mRNA structural
elements on short interfering RNA (siRNA) and AON targeting efficiency. Under
identical conditions, both PTGS pathways are significantly inhibited by target mRNA
S.I. Rudnick
Division of Hematology/Oncology, Department of Medicine, University of Pennsylvania School
of Medicine, Philadelphia, PA, USA
Fox Chase Cancer Center, Philadelphia, PA, USA
V. Aishwarya and A.M. Gewirtz (*)
Division of Hematology/Oncology, Department of Medicine, University of Pennsylvania School
of Medicine, Philadelphia, PA, USA
e-mail: gewirtz@mail.med.upenn.edu
256 S.I. Rudnick et al.
secondary structure. Surprisingly, we found that an AON was less stringent in its
requirement for mRNA target accessibility than the corresponding siRNA. By deter-
mining that the AON and siRNA guide strand have the same apparent KD in the
absence of protein, we show that nucleic acid binding affinity does not explain their
difference in in vivo silencing activity. Rather, it appears that RISC must increase the
binding affinity of the siRNA for the target. Furthermore, RNA binding proteins are
also potent inhibitors of AON activity. We conclude that mRNA secondary and
quaternary structure play important roles in PTGS by significantly affecting the
ability of a siRNA or AON to hybridize with their intended target. Recognition of
these effects will facilitate the design of more efficient antisense molecules for
therapeutically motivated gene silencing and argue for continued mechanistic studies
on AON and siRNA mediated mRNA destruction.
Keywords RNA structure siRNA Antisense Oligodeoxynucleotide Gene

silencing RNA cleavage
1 Introduction
Posttranscriptional gene silencing (PTGS) is a process by which a protein’s synthesis

is impaired by targeting its messenger RNA (mRNA). The two common methods to
achieve PTGS are through the use of antisense oligodeoxynucleotides (AON) or short
interfering RNAs (siRNA), both of which elicit mRNA degradation through an RNase
(oligodeoxynucleotides), or RNase H like (siRNA), mechanism after sequence-
specific hybridization to the mRNA. These approaches to PTGS are widely applied
as tools to study a gene’s function in the laboratory, but they are also being developed
as highly specific gene targeted therapies (Opalinska and Gewirtz 2002; Gewirtz
2007; Rayburn and Zhang 2008). AON and siRNA have both been employed to
screen thousands of genes in high-throughput loss of function assays in order to
identify novel drug targets, genes that affect drug efficacy, and components of cellular
pathways (Berns et al. 2004; Bartz et al. 2006; Koller et al. 2006; Ricke et al. 2006).
AON- and siRNA-initiated PTGS is dependent upon hybridization to the target
mRNA prior to degradation. It is well established that mRNA secondary and tertiary
structure affect heteroduplex formation of an oligonucleotide and RNA (Lima et al.
1992; Mir and Southern 1999) as well as antisense activity (Matveeva et al. 1998;
Vickers et al. 2000). Furthermore, in a study targeting the HIV-1 transactivation
response element (TAR), it was found that disruption of the characteristic TAR
hairpin resulted in greatly increased siRNA activity (Brown et al. 2005). In other
studies aimed at correlating siRNA activity to the structure of the mRNA, it
has generally been concluded that target accessibility increases siRNA efficacy
(Yoshinari et al. 2004; Overhoff et al. 2005; Schubert et al. 2005). These studies
were conducted in cell culture and monitored gene expression by RT-PCR, western
blot, or activity of a reporter gene. Since it is difficult to predict the structure of a
mRNA in a cell, conclusions of how RISC and siRNA recognize their target or
mRNA Structure and its Effects on Posttranscriptional Gene Silencing 257
define mRNA structural characteristics that inhibit siRNA and RISC activity have
yet to be made. Numerous chemical modifications have been made to increase the
binding affinity of AON (Pradeepkumar et al. 2003) and siRNA (Braasch et al. 2003;
Elmen et al. 2005) in order to overcome the energetic barriers imposed by mRNA
self structure, but no study has compared the in vitro activity of the RNase H-based
mechanisms of AON and siRNA by monitoring mRNA cleavage.
In order to determine the intramolecular structures in mRNA that affect PTGS by
AON and siRNA, we have employed an in vitro assay to monitor the sequence-
specific cleavage of mRNA (Rudnick et al. 2008). While maintaining a constant
target site and sequence, we introduced mRNA structure as the only variable in our
experiments by preannealing 20 -O-methyl oligonucleotides (20 OMe ON) at various
sites in the target mRNA. Since it is known that siRNA design can dictate activity
(Ding et al. 2003; Schwarz et al. 2003), we chose to use a single sequence for our
AON and siRNA. Therefore, fluctuations in cleavage activity could not be attributed
to the targeting molecule but instead will be a direct result of manipulating the
mRNA structure. This strategy, targeting a 182 nt segment of firefly luciferase,
reveals that secondary structure upstream or downstream of the target site is almost
inconsequential on AON and siRNA activity. Inducing secondary structure directly
on the target site revealed that AON can target smaller accessible regions in the
mRNA than siRNA, while a fully double-stranded target significantly hinders both
pathways. AON activity was much less pronounced than siRNA activity in a dual
luciferase assay. Perhaps not surprisingly, our data showed that in vitro AON and
siRNA activity does not necessarily correspond to cellular activity. This suggests that
AON activity is suppressed and siRNA activity enhanced when in the proper cellular
context and a further mechanistic understanding will allow for the more effective
development of these molecules as therapeutic agents and investigative tools.
2 A Structured Target Site Reduces AON and siRNA

Activity In Vitro
In order to evaluate the importance of secondary structure in PTGS-mediated by

either siRNA or AON, a method was needed to eliminate all variables other than
structure, such as siRNA incorporation into RISC, transfection efficiency, GC
content, and TM, without relying upon computational structural predictions. This
was accomplished by annealing 20 -O-methyl oligonucleotides (20 OMe ON) at vari-
ous positions along the mRNA relative to the target site prior to in vitro cleavage
reactions. 20 OMe ON were chosen for this study because they have a high affinity to
mRNA and do not elicit the activity of RNase H upon heteroduplex formation
(Monia et al. 1993). In this manner, short regions of double-stranded character are
induced at specific sites as a model of intramolecular structure in an mRNA of
otherwise unknown structure. By analyzing the resulting amount of cleavage prod-
uct, very specific insights into the importance of mRNA structure were gained.
In order to measure the effect of secondary structure at multiple positions on AON

and siRNA activity, the 20 OMe ON of Group I (Table 1) were annealed to the mRNA
target individually. As shown in Fig. 1a, where each bar represents the relative
position of a complementary 20 OMe ON, short double-stranded regions upstream of
the 19 nt target site are induced by annealing 20 OMe ON A and B. Similarly, 20 OMe
ON E and F create short double-stranded regions downstream of the 19 nt target site.
Oligos C and D each induce half of the target site to be double-stranded, while the
annealing of Target Blocker creates a short double-stranded target.
In 30 s AON and 8.5 min siRNA cleavage reactions where 20 OMe ON A, B, E, or F
were preannealed to the mRNA, cleavage activity was enhanced (Fig. 1b–d). This
finding is consistent with a previous report that shows enhanced siRNA activity after
disruption of local mRNA structure (Brown et al. 2005). In reactions where half of the
target site was complexed with 20 OMe ON C or D, siRNA activity was hindered more
than AON. When the target site was fully obscured through the use of 20 OMe ON Target
Blocker, siRNA activity was reduced to the same extent as in reactions with 20 OMe ON
C and D while AON activity was reduced to the level of the negative controls (Fig. 1d).
Despite the structure induced by preannealing 20 OMe ON, the siRNA-generated
cleavage product was consistently the same molecular weight. However, when one
side of the target site was obscured by 20 OMe ON C or D, the RNase H cleavage site
shifted in the 50 or 30 direction accordingly. Our data reinforce previous observa-
tions that RISC is a site-specific nuclease (Schwarz et al. 2004) and that RNase H
has multiple cleavage sites (Lima et al. 2006). Furthermore, these observations
emphasize that even when processing the identical substrate, these enzymes, with
similar nuclease domains, recognize and process their substrates quite differently.
3 Analysis of Binding Affinity to mRNA and Rate Dependencies

on Concentration for AON and siRNA Activity
AON and siRNA are both thought to bind their substrate in a diffusion mediated
manner, as opposed to having a protein guided mechanism (Stein 1999; Brown
et al. 2005; Yuan et al. 2005). Therefore, two properties closely tied to binding, and
presumably cleavage activity, should be the relative stoichiometry of the AON and
siRNA to the mRNA target and the relative stability of the resultant AON:mRNA or
siRNA:mRNA duplex.
In general, RNA duplexes are known to have relatively high melting tempera-
tures (TM) and a correspondingly low free energy compared to RNA:DNA and
DNA:DNA duplexes (Lesnik and Freier 1995). The TM for nucleic acid duplexes is
defined as the temperature where half of the molecules are double-stranded and half
are single-stranded. TM is related to the free energy of a duplex because the higher
the TM, the more the energy that is required to dissociate one nucleic acid strand
from another. Therefore, if a potential duplex has a high predicted TM, one expects
the two single strands to have a high affinity, or low KD, for each other. In order to
Table 1 Sequences of 20 -O-methyl oligonucleotides used to model intramolecular mRNA structure. Each group of 20 OMe ON was used in experiments where short
regions of double-stranded character were induced over a large area of the 182 nt luciferase target (Group I), in pairs on the target site with a variable number of free
target bases (Group II), or used to make the entire target site double-stranded with a variable internal loop (Group III). Between the sequence of each Group II 20 OMe
ON, the number of unhybridized target bases is indicated. In Group III 20 OMe ON, the lower case letters indicate bases that are not complementary with the mRNA
target
Group I 20 -O-methyl oligonucleotides Group II 20 -O-methyl oligonucleotides Group III 20 -O-methyl oligonucleotides
Name Sequence (50 –30 ) Sequence 50 –30 – Gap(nt) – 50 –30 Mismatches Sequence (50 –30 )
(nt)
A CCGAACGGACAUUUCGAAG GCCCAUAUCGUUUCA – 2 – 2 UCUGUGAUUUGUAUUacGCCCAUAUCGUUUCA
UCUGUGAU UUGUAUU
B GUUUCAUAGCUUCUGCCAA CCCAUAUCGUUUCAU – 4 – 4 UUCUGUGAUUUGUAUcacuCCCAUAUCGUUUCAU
UUCUGUGA UUUGUAU
C AGCCCAUAUCGUUUCAUAG CCAUAUCGUUUCAUA – 6 – 6 AUUCUGUGAUUUGUAgcacuaCCAUAUCGUUUCAUA
AUUCUGUG AUUUGUA
Target blocker AUUUGUAUUCAGCCCAUAU CAUAUCGUUUCAUAG – 8 – GAUUCUGU GAUUUGU
8 GAUUCUGUGAUUUGUcgcacuaaCAUAUCGUUUCAUAG
D CGAUUCUGUGAUUUGUAUU AUAUCGUUUCAUAGC – 10 – 10 CGAUUCUGUGAUUUGgcgcacuaaaAUAUCGUUUCAUAGC
CGAUUCU GUGAUUUG
E UGCAUACGACGAUUCUGUG UAUCGUUUCAUAGCU – 12 – 12 ACGAUUCUGUGAUUUagcgcacuaaacUAUCGUUUCAUAGCU
ACGAUUC UGUGAUUU
F AAUUGAAGAGAGUUUUCAC AUCGUUUCAUAGCUU – 14 – 14 GACGAUUCUGUGAUUcagcacuaaacgAUCGUUUCAUAGCUU
GACGAUU CUGUGAUU
UCGUUUCAUAGCUUC – 16 – 16 CGACGAUUCUGUGAUccagcgcacuaaacgcUCGUUUCAUAGCUUC
CGACGAUU CUGUGAU
mRNA Structure and its Effects on Posttranscriptional Gene Silencing
259
a
A C D F
5'...cuucgaaauguccguucgguuggcagaagcuaugaaacgAUAUGGGCUGAAUACAAAUcacagaaucgucguagcagugaaaacucucuucaauu...3'
B Target Blocker E
b c
r
ke
ke
oc
oc
Bl
Bl
et
et
rg
rg
Ta
- + A B C D E F
Ta
- + A B C D E F
d
0.4
30s AON Reactions
8.5min siRNA Reactions
5' Cleavage Product (pmol)
0.3
0.2
0.1
0.0
ol ol A B C er D E F
o ntr ontr l o ck
e C C t B
e
tiv tiv rge
ga Posi Ta
Ne
Preannealed 2'-O-Methyl Oligonucleotide
Fig. 1 Analysis of inducing structure over a large area of the 182 nt luciferase mRNA target.
Partial sequence of mRNA target is shown (a) with the relative position of each individual
preannealed 20 OMe ON. The target sequence of the AON and siRNA guide strand is indicated
in all capital letters. Each induced structure of a was used in cleavage assays for 30 s in the
presence of 2.5 mM AON (b) or for 8.5 min in the presence of 2.5 mM siRNA (c) and analyzed on
10% sequencing gels. Negative controls () have no siRNA or AON and positive controls (+) are
cleavage reactions against the unknown, native mRNA structure in the absence of 20 OMe ON.
Cleavage product formation (d) for each reaction was calculated by normalization to the amount of
target in the negative control lanes of each gel, and the mean of at least three independent reactions
is plotted (1 standard deviation)
Source: Rudnick et al, PNAS, 2008
critically compare the activity of the AON and siRNA, their relative affinities for
the mRNA target in the absence of protein must be determined. Then, in order to
evaluate a potential correlation between TM and substrate turnover, AON and
siRNA activities are measured above and below the apparent KD.
4 AON and siRNA Guide Strand Have Equal Affinity

for the Target mRNA
The AON and siRNA guide strand used for these studies were predicted to have
TM’s of 47 C and 59 C, respectively, with the RNA target (www.idtdna.com).
Therefore, in order to determine if oligonucleotide binding affinity and TM play a
role in AON or siRNA activity, we had to determine the apparent binding affinity
under the conditions used to detect mRNA cleavage. Since the mRNA target was
182 nt and the siRNA guide strand and AON are 21 nt, analyzing the mRNA and
the 203 nt duplexes can be accomplished easily by electrophoresis. 100 nM
mRNA was incubated in increasing concentrations of AON or siRNA guide strand
(Fig. 2). From 100 to 500 nM oligonucleotide, the duplexes can clearly be resolved
from the mRNA alone (Fig. 2a). At concentrations greater than or equal to 1 mM,
the mRNA is saturated with oligonucleotide and almost no free mRNA is detect-
able. The percent duplex in each lane of four independent experiments was plotted
against concentration of the oligonucleotide (Fig. 2b) with nearly identical results
for the AON and siRNA guide strand. The apparent KD for the AON and siRNA
under cleavage assay conditions, but in the absence of protein, is 378 37 nM
and 397 28 nM, respectively, suggesting that TM does not affect their affinity
in vitro.
5 AON and siRNA Display Apparent First and Zero

Order Kinetics
To determine if the apparent KD has impact on AON and siRNA activity, they were
screened in a dose-response fashion while keeping the mRNA at 100 nM. Since the
binding affinity of the oligonucleotides in Fig. 2 had no significant difference, the
average of the two, 388 nM, was used as a center concentration for the activity
screen. AON and siRNA were tested in the in vitro cleavage assay at 25 nM,
194 nM, 388 nM, 587 nM, and 776 nM, which correspond to 0.06, 0.5, 1.0,
1.5, and 2 the average apparent KD.
AON cleavage reactions showed a broad range of activity (Fig. 3a). The
samples at 25 nM (filled circle) and 194 nM (empty triangle) had barely distin-
guishable amount of product formation. Increasing the AON concentration to
388 nM and above gave significant stepwise increases in initial rate and total
amount of product formed. However, there was no significant difference in initial
rates or total product formed for the siRNA when tested at 25 nM and above
(Fig. 3b). The only concentration that displayed multiple turnovers was the 25 nM
(0.125 pmol) siRNA (filled circle) since it generated greater than 0.2 pmol
50 cleavage product.
Co ol
l
ro
ive ontr
nt
P o ve C
15 M
nM
nM
M
5n
M
ti
n
nM
nM
ga
si t
0n
0n
5n
0n
00
00
00
7.
Ne
25
50
10
18
25
37
50
10
20
AON:mRNA
AON
mRNA
ss-siRNA:mRNA
ss-siRNA
mRNA
b 120
siRNA
100 AON
80
Percent Duplex
60
40
20
0
10 100 1000 10000
Oligonucleotide(nM)
Fig. 2 Determination of apparent KD with mRNA for AON and siRNA in vitro. (a) mRNA:AON
duplex (top panel) or single-stranded siRNA:mRNA duplex (bottom panel) separated from mRNA
alone (bottom band) at indicated concentrations of oligonucleotide in a cell free system. Negative
controls are mRNA alone. Positive control has 2 mM oligonucleotide annealed to mRNA by heat
cool. (b) Quantification of plots in a. Points are average standard deviation of four independent
experiments. Data fit to four parameter logistic function
Calculating the initial velocity (Vo) of each reaction and plotting against concen-
tration of oligonucleotide used yielded relationships for both the AON and siRNA.
As expected from the data in Fig. 3, the AON reactions became faster and siRNA
reaction velocity was static, as the concentration of each oligonucleotide increased.
Therefore, in this concentration range, the AON and siRNA behave under apparent
first and zero order kinetics, respectively.
a b
0.25 0.40

25nm AON
194nm AON
388nm AON 0.35
582nm AON
0.20 776nm AON
0.30
0.15 0.25
0.20
0.10 0.15
25nm siRNA
0.10 194nm siRNA
0.05 388nm siRNA
0.05 582nm siRNA
776nm siRNA
0.00
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Time (min) Time (min)
Fig. 3 In vitro activity of AON and siRNA at varying concentrations. AON (a) and siRNA (b)
cleavage activity at increasing concentrations. The average standard deviation of three inde-
pendent experiments is plotted against time. The scale of the Y-axis used is different in order to
better illustrate differences in AON reactions
6 For Full In Vitro Activity, siRNA Require Greater

Target Site Accessibility Than AON
Since structure directly at the site of siRNA and AON targeting proved to be a
critical factor in mRNA cleavage, the minimum number of accessible bases for
activity was next examined. In order to create accessible targets of varying sizes
(Fig. 4a), two double-stranded regions were made equidistantly from the expected
siRNA cleavage site by annealing two Group II 20 OMe ON (Table 1). This tiling of
the target site was always done symmetrically, leaving an equal number of accessi-
ble bases on either side of the cleavage site.
After a 30 s cleavage reaction, the AON activity was greater than the positive
control when 8–16 nt of the target site remained accessible (Fig. 4b and d). Multiple
RNase H generated cleavage products were detectable when 14 nt and 16 nt were
free, while fewer products were observed when 8, 10, or 12 nt of the target were free
(Fig 4b). When 20 OMe ON were used to reduce target accessibility to 6 nt or less,
AON activity ranged from 56 to 27% of the positive control where no 20 OMe ON
were annealed.
In contrast to the AON, all siRNA cleavage reactions with a tiled target site generated
a single cleavage product (Fig. 4c). After an 8.5 min cleavage reaction, the siRNA had
enhanced activity with respect to the positive control when 16 nt of the target were left
accessible (Fig. 4c and d). However, when the target site was any size less than 16 nt,
siRNA activity ranged from 52 to 29% of the positive control (Fig. 4c and d).
These data again reveal differences in the way each PTGS pathway processes its
substrate in vitro. Surprisingly, the siRNA proved to be quite sensitive to any
addition of double-stranded character to the target site as demonstrated by the
need of 16 accessible nucleotides for full activity. Equally surprising, the AON
a
5'...uuggcagaagcuaugaaacgAUAUGGGCU*GAAUACAAAUcacagaaucgucguaugca...3'
Unhybridized Nucleotides
b c
Gap in 2'OMe Oligo (nt) Gap in 2'OMe Oligo (nt)
- + 2 4 6 8 10 12 14 16 - + 2 4 6 8 10 12 14 16
d 0.4
30s AON Reactions
0.3
0.2
0.1
0.0
ol ol 2 4 6 8 10 12 14 16
o ntr ontr
e C ive C
tiv t
e ga Posi
N
Gap Between Preannealed
2'-O-Methyl Oligonucleotides (nt)
Fig. 4 Determination of the minimum number of accessible bases for AON and siRNA guided
cleavage in vitro. (a) Partial sequence of the 182 nt firefly luciferase target mRNA. Two 20 OMe
ON were annealed at the outer edges of the target site (CAPS) and positioned incrementally closer
to the expected siRNA cleavage site (*). Cleavage activity is reported as a function of the number
of unhybridized bases between the two preannealed 20 OME ON. Analysis of the 30 s AON (b) and
8.5 min siRNA (c) reactions was performed on 10% sequencing gels where negative controls ()
have no AON or siRNA and positive controls (þ) target the unknown, native structure of the
mRNA in the absence of 20 OMe ON. The amount of cleavage product (d) was calculated by
normalization to the amount of target in the negative controls of each gel, and the mean of at least
three independent reactions is plotted (1 standard deviation)
revealed robust activity so long as 8 nt or more were not bound by 20 OMe ON.
Based on these data, one might anticipate that siRNA would have considerably less
activity than AON in vivo though clearly this is not the case.
7 A Double-Stranded Target Site Greatly Reduces In Vitro

PTGS Activity
As it is unlikely that the structure depicted in Fig. 4a exists in vivo, a double-

stranded target with internal loops or bulges was generated to better model “real
life” mRNA intramolecular structure. The intended size of the bulges ranged from
2 to 16 nt in an otherwise double-stranded target by annealing the 20 OMe ON of
Group III (Table 1). These 20 OMe ON each have two 15 nt arms that are comple-
mentary to the mRNA. Between the two arms are a variable number of bases not
complementary to the target. These mismatches are centered on the bond where the
siRNA is expected to induce cleavage (Fig. 5a).
In comparison to the previous structures examined, all of the double-stranded
targets greatly reduced AON and siRNA activity (Fig. 5b–c). For both AON and
siRNA, activity was maximal with the largest bulge and decreased as the number of
mismatches was reduced. When an internal loop of 14 nt was generated, the
cleavage activity was lower than expected based on the amount of product gener-
ated from targets with a 12 nt or 16 nt loop. The 14 nt loop may have adopted its
own intramolecular structure that blocked binding of the AON or siRNA.
The extent by which the double-stranded targets reduced cleavage activity was
unexpected, especially for the siRNA. In order to determine if either the AON or
siRNA could eventually infiltrate the double-stranded target, cleavage reactions
with 6 nt bulged targets were conducted for 2 h (Fig. 6a and b). The siRNA generated
a single primary cleavage product (Fig. 6b). The primary cleavage product generated
by the AON is initially a single band, but after 15 min, multiple cleavage sites are
resolved (Fig. 6a). Additionally, the AON generates the secondary cleavage product
seen previously (Fig. 6a) at a similar rate as the primary cleavage product.
In comparison to reactions targeting the native mRNA structure (Fig. 6c), the
total amount and rate of primary cleavage product formation is decreased for
both AON and siRNA (Fig. 6c). After 30 min, the AON reaction is essentially
complete and no further significant product formation is seen. However, the rate and
amount of primary cleavage product formation by the AON may be limited due the
competing reaction generating secondary cleavage product. The siRNA-mediated
digestion of the 6 nt bulged mRNA was slow, maintained a nearly constant rate, and
does not reach completion in 2 h. Therefore, resolving the barrier posed by intramo-
lecular mRNA structure is a rate-limiting step for in vitro mRNA targeting.
8 An AON That is More Effective Than the siRNA Against

an Identical Target In Vitro is Less Effective Against
the Same Target In Vivo
Our in vitro observations of rapid onset of AON-mediated mRNA cleavage,

combined with more relaxed target accessibility requirements, suggested that an
AON directed against this specific mRNA target ought to perform better in vivo
a
5'...uuggcagaagcuaugaaacgAUAUGGGCUGAAUACAAAUcacagaaucgucguaugca...3'
3' GAUACUUUGCUAUAC UGUUUAGUGUCUUAG 5'
AAUCACGC
b c
Mismatches in Bulge (nt) Mismatches in Bulge (nt)
- + 2 4 6 8 10 12 14 16 - + 2 4 6 8 10 12 14 16
d
0.4
30s AON Reactions
0.3
0.2
0.1
0.0
l l 2 4 6 8 10 12 14 16
tro tro
C on Con
e e
tiv tiv
e ga Posi
N
Bulge Created by Preannealed
2'-O-Methyl Oligonucleotides (nt)
Fig. 5 Analysis of cleavage activity when targeting a double-stranded mRNA with a variable
internal loop. (a) Partial sequence of the 182 nt firefly luciferase target. 20 OMe ON (CAPS) of
differing length was preannealed over the center of the target site (CAPS). Two 15 nt reverse
complementary arms were always annealed to the target to maintain duplex stability. Cleavage
activity is reported as a function of the number of mismatches (CAPS) used to vary the size of the
internal loop. Analysis of the 30 s AON (b) and 8.5 min siRNA (c) reactions was performed on 10%
sequencing gels where negative controls () have no AON or siRNA and positive controls (þ)
target the unknown, native structure of the mRNA in the absence of 20 OMe oligos. The amount of
cleavage product (d) was calculated by normalization to the amount of target in the negative controls
of each gel, and the mean of at least three independent reactions is plotted (1 standard deviation)
than the corresponding siRNA. We tested this hypothesis using K562 human
leukemia cells because similar AON and siRNA nucleofection efficiency was
obtained when codelivered with the reporter vectors. In three independent experi-
ments, scrambled AON and scrambled siRNA showed no effect on luciferase
a b
0 2.5 5 15 30 60 90 120min 0 2.5 5 15 30 60 90120min
T T
100 PCP 100 PCP

90 90
80 80
70 70
60 60
SCP
50 50
40 40
c
0.5
2.5µM AON
2.5µM siRNA
0.4
0.3
0.2
0.1
0.0
0 20 40 60 80 100 120
Time (min)
Fig. 6 Cleavage activity as a function of time with double-stranded target. Two hour cleavage
reactions in the presence of 2.5 mM AON (a) or 2.5 mM siRNA (b) are analyzed on 10%
sequencing gels. The double-stranded mRNA target with 6 nt mismatched bulge, primary cleavage
product, and secondary cleavage products are indicated by T, PCP, and SCP, respectively. Primary
cleavage product formation as a function of time (c) where values were calculated by normalizing
product at each time to the amount of target at the initial time point. In the AON reaction (filled
circles), all bands migrating near 100 nt were summed after normalization and plotted while only
one product was seen for the siRNA reaction (empty circles). The mean of at least three indepen-
dent reactions is plotted (1 standard deviation)
activity when compared to vector alone (Fig. 7). When firefly luciferase was
targeted, a reduction in the relative luminescence of 35 and 78% for AON and
siRNA was observed, respectively.
These data show that in a cellular context, an AON is not necessarily as robust
with respect to target cleavage as it is in vitro. We found the obverse true as well,
1.4
1.2
1.0
Relative Luminesence
0.8
0.6
0.4
0.2
0.0
Scrambled Scrambled Luciferase Luciferase
AON siRNA AON siRNA
Fig. 7 Dual luciferase assay in K562 cells. Reduction in luminescence was compared for the same
AON and siRNA used in in vitro experiments. For both the scrambled controls and knockdown
experiments, 0.8 nmol of total oligonucleotide was nucleofected into K562 cells. Twenty four hour
postnucleofection, cells were lysed and the ratio of firefly to renilla luciferase was determined and
normalized to that of the samples with luciferase vectors alone. The mean of at least three
independent experiments is plotted (1 standard deviation)
i.e., a siRNA with more modest in vitro cleaving activity can be very efficient when
employed in vitro.
9 Discussion
Sequence-specific targeting is dependent on the ability of the AON or siRNA

guide strand to hybridize via Watson–Crick base pairs to the mRNA. The cellular
structure of mRNA, characterized by base pairing with itself, base stacking, and
protein binding, is very complex and difficult to predict. The intra- and intermole-
cular interactions are not only complex but also dynamic depending on a given
mRNA’s regulation and cellular compartmentalization. It has long been a challenge
to the PTGS field to develop methods of increasing the likelihood of an AON or
siRNA overcoming structure and hybridizing to an accessible region of its target
mRNA. Means of overcoming the mRNA structural obstacle to hybridization have
often come in the form of chemical modifications to the guide strand nucleic acid
or in the analysis of the mRNA for an appropriate binding site. The success of
siRNA applications suggests that this technology has a higher frequency of
hybridization events than AON, translating to less effort in designing an efficacious

targeting molecule. Nonetheless, knockdown of any given mRNA target with
siRNA is not guaranteed, and the degree of knockdown achieved is often modest.
Therefore, it became our goal to decipher the specific effects of mRNA intramo-
lecular structure on the ability of these nucleic acids to degrade a target mRNA. A
single AON and siRNA sequence was used to target a mRNA whose structure was
manipulated by annealing 20 OMe ON (Table 1). In an in vitro assay, cleavage
product formation was directly monitored (Fig. 1). AON activity was observed to
have a more rapid rate than siRNA when targeting the native mRNA structure. In
vitro rates of reaction were directly related to the endogenous levels of each
enzyme. Since this is not adjustable without varying the total amount of protein,
all AON and siRNA cleavage assays with modified structure were instead com-
pared after 30 s and 8.5 min, respectively (Fig. 1c). When short double-stranded
regions were made at numerous positions over the majority of the mRNA, only
those on the 19 nt target site reduced product formation. Furthermore, structure
induced upstream or downstream of the target site enhanced product formation
(Fig. 1d). Presumably, this is due to the disruption of the native mRNA structure,
which makes the target site more accessible (Brown et al. 2005). Although it is not
directly evident, this data suggests a lack of mRNA unwinding and base pair
scanning as structure would likely slow this process.
The involvement of half or the entire target site in a double strand revealed that
each pathway’s activity responds differently to the number of accessible bases. This
was explored by annealing 20 OMe ON successively closer to the center of the target
site, creating a variable number of accessible target bases (Fig. 4a). The siRNA
demonstrated a high sensitivity to target site structure, because when less than 16 nt
were left open, activity dramatically fell (Fig. 4c and d). In contrast, the AON was
active as long as eight or more target bases were available (Fig. 4b and d).
Accessible bases in mRNA are more often thought to be involved in internal
loops and bulges rather than existing as individual single strands. When double-
stranded targets were tested in our assay, both the AON and siRNA pathways
were significantly inhibited despite loops with a maximum of 16 accessible
bases (Fig. 5). However, when given 2 h to digest the mRNA instead of 30 s or
8.5 min, both pathways demonstrated the ability to process a double-stranded
target with a 6 nt internal loop. When compared to rates observed when
targeting the native mRNA structure (Fig. 5), rates targeting the 6 nt bulged
target were significantly slower. The AON reaction retained its rapid initial
burst, while the siRNA displayed a slow and steady approach to cleaving the
mRNA target (Fig. 6). Finally, when tested in K562 cells, the siRNA demon-
strated a more robust reduction of luciferase activity when compared to the
AON (Fig. 7).
While it is clear from the data presented here that mRNA structure reduces the
rate of AON- and siRNA-mediated cleavage, it remains unclear whether
structured targets are simply poor substrates, if overcoming structure is a rate-
limiting step, or if structure itself is a type of inhibitor to gene silencing. The
differences in these classifications reflect both catalytic activity and substrate
a b
0 2.5 5 15 30 60 90 120min 0 2.5 5 15 30 60 90 120min
T T
100 PCP 100 PCP

90 90
80 80
70
70
60
60
50 SCP
50
40
40
c
0.5
0.4
0.3
0.2
0.1
2.5µM siRNA
2.5µM AON
0.0
0 20 40 60 80 100 120
Time (min)
Fig. 8 Comparison of AON and siRNA activity in Drosophila embryo whole cell lysate. A 182 nt
50 cap labeled segment of firefly luciferase mRNA was incubated in Drosophila embryo whole cell
lysate with 2.5 mM AON (a) or 2.5 mM siRNA (b). Aliquots were removed at given time points and
analyzed on 10% sequencing gels where the target mRNA, 104 nt primary cleavage product, and
51 nt secondary cleavage product are indicated by T, PCP, and SCP, respectively. (c) Plot of 50
cleavage product formation as a function of time. The amount of cleavage product was calculated
by normalizing the signal to that of the target at the initial time point, and the mean of at least three
independent reactions is plotted (1 standard deviation). The dotted line at 0.22 pmol product
indicates single time points used in later experiments in order to compare AON and siRNA activity
against structure targets
binding (Fig. 8). Clearly, our results suggest that further PTGS mechanistic
studies are warranted.
Since RNA:RNA duplexes are known to have higher TM’s than DNA:RNA
duplexes, it is reasonable to think that this would give an advantage to the siRNA
in binding the mRNA target. However in our in vitro cleavage assay conditions, the
AON and siRNA displayed the same apparent KD in the absence of their cognate
nuclease. When tested for cleavage activity, the AON cleavage activity reflected
the measured affinity. In other words, the AON generated significant product at the
concentration that should yield half of the mRNA bound, and half free. Consistent
with a simple two reagent binding equilibrium,
KD k
(1) AON + mRNA Duplex Fragments + AON
ð1Þ
RNase H
adding more AON generated more AON:mRNA duplex and an associated increase
cleavage product (Fig. 3). As illustrated in (1) below, it is hypothesized that the rate
of the forward reaction catalyzed by RNase H is dependent on duplex concentra-
tion. However, duplex formation is dependent on the binding affinity of the AON
for the mRNA. Furthermore, protein binding of the mRNA may provide as steric
barrier to AON hybridization. If so, this could certainly pose a mechanism of
inhibition that would block some, but not all AON.
While having the same affinity for the mRNA, the siRNA did not show the same
pattern of activity as the AON. Instead, every concentration tested between 0.06
and 2 the apparent KD gave the same rate of reaction and total product turnover.
Since the siRNA was most efficient at a concentration more than ten times lower
than the apparent KD measured, we hypothesize that RISC increases the binding
affinity of the siRNA for the target mRNA (Fig. 9). Since R2D2 and Dicer are
already known to interact with Ago2 and aid in loading the siRNA, as shown in (2),
perhaps they also aid in association with the mRNA.
Dicer, R2D2
siRNA + Ago2 siAgo2
(2) +
mRNA siAGo2 • mRNA siAGo2 + fragments
ð2Þ
D ic er, R 2D 2,or
un kn ow n protein
Since, as seen in Figs. 2 and 3, adding extra siRNA did not increase mRNA
degradation, we suspect that the in vitro RNAi machinery was saturated. Similar
observations have been reported in vivo and are thought to be from limiting
amounts of exportin-5 (Yi et al. 2003; Grimm et al. 2006). It is unlikely however
that exportin-5 activity is important in a WCL system, and it is probably Ago2
binding that is saturated in vitro.
Based on the in vitro results presented here, one would expect that AON and
RNase H can generally act more rapidly and efficaciously than siRNA when targeting
structured mRNA. Clearly, the body of evidence from in vivo studies does not support
this conclusion. In order for AON technology to be developed to its maximum
therapeutic potential, studies should be conducted to determine the true cellular
barrier to their activity. Such barriers may include poor cellular localization with
0.030
0.025
Reaction Velocity (pmol /min)
0.020
0.015
0.010
0.005
0.000 AON
siRNA
–0.005
0 200 400 600 800 1000
Oligonucleotide Concentration (nM)
Fig. 9 In vitro cleavage rates as a function of oligonucleotide concentration. Initial velocities

calculated from the amount of product formed between 15 s and 5 min for reactions varying AON
and siRNA concentration
respect to mRNA and RNase H, or may be as simple as steric hindrance from proteins
bound to mRNA in the nucleus not encountered in the cytoplasm by RNAi.
Our work also demonstrates that siRNA activity was unexpectedly sensitive to
mRNA structure in vitro. Due to the high success and relative ease by which an
effective siRNA can be designed for in vivo studies, it is suggestive that there exists
an unidentified means of overcoming mRNA structure in the RNAi pathway. Since
many successful studies have been conducted in vitro with purified proteins, any
target recognition mechanism involved in RNAi, however, would not be fully
critical to the basic activity of the minimal RISC complex.
10 Conclusions
It is difficult not to see some irony in the fact that the first, and at least thus far, the only
clinically approved AON was delivered intraocularly for treatment of HIV-associated
CMV retinitis (Roehr 1998). Whether this antisense oligodeoxynucleotide worked by
hybridization with its target is uncertain but its efficacy is clear (Group TVS 2002). A
siRNA molecule targeting VEGF, which was also delivered by intraocular injection
and designed to treat macular degeneration, appeared well on its way to becoming the
first approved siRNA therapeutic but stalled for lack of clear cut efficacy and strong
indications that it worked not by inhibition of VEGF mRNA but by stimulation of
Toll-like receptor 3 (Rossi et al. 2008; Yang et al. 2008). Still, the fact that nucleic acid
drugs continue to be developed shows that interest in this class of therapeutics remains
strong. Later generation of AON and siRNA are being tested in clinical trials of all
phases to treat a wide range of diseases (de Fougerolles et al. 2007; Graham et al. 2007;
Kamada et al. 2007; Prakash and Bhat 2007). Both technologies face similar, and by
now familiar, challenges of achieving specific, high efficiency gene knockdowns in
patients in the absence of toxic side effects. mRNA structure has long been thought to
act as a barrier to Watson–Crick base-pairing and heteroduplex formation (Mir and
Southern 1999), but most studies investigating this in PTGS applications never clearly
separate structure from other assay variables and rarely know the structures involved
in their target (Holen et al. 2002; Vickers et al. 2003). The work reported in this chapter
is a direct response to this challenge. Having said this, it is also important to state that
even the most perfectly targeted oligonucleotide will be of little use if it cannot be
delivered in a biologically relevant manner, meaning, not only introduced into cells
but also in a form that is or becomes bioavailable for hybridization. So, while
considerable work remains to be done, the prize is well worth the effort to those on
the hunt for effective oligonucleotide drugs.
References
Bartz SR, Zhang Z, Burchard J et al (2006) Small interfering RNA screens reveal enhanced
cisplatin cytotoxicity in tumor cells having both BRCA network and TP53 disruptions. Mol
Cell Biol 26:9377–9386
Berns K, Hijmans EM, Mullenders J et al (2004) A large-scale RNAi screen in human cells
identifies new components of the p53 pathway. Nature 428:431–437
Braasch DA, Jensen S, Liu Y et al (2003) RNA interference in mammalian cells by chemically-
modified RNA. Biochemistry 42:7967–7975
Brown KM, Chu CY, Rana TM (2005) Target accessibility dictates the potency of human RISC.
Nat Struct Mol Biol 12:469–470
de Fougerolles A, Vornlocher HP, Maraganore J et al (2007) Interfering with disease: a progress
report on siRNA-based therapeutics. Nat Rev Drug Discov 6:443–453
Ding H, Schwarz DS, Kenne A et al (2003) Selective silencing by RNAi of a dominant allele that
causes amyotrophic lateral sclerosis. Aging Cell 2:209–217
Elmen J, Thonberg H, Ljungberg K et al (2005) Locked nucleic acid (LNA) mediated improve-
ments in siRNA stability and functionality. Nucleic Acids Res 33:439–447
Gewirtz AM (2007) On future’s doorstep: RNA interference and the pharmacopia of tomorrow.
J Clin Invest 117:1–3
Graham MJ, Lemonidis KM, Whipple CP et al (2007) Antisense inhibition of proprotein
convertase subtilisin/kexin type 9 reduces serum LDL in hyperlipidemic mice. J Lipid Res
48:763–767
Group TVS (2002) A randomized controlled clinical trial of intravitreous fomivirsen for
treatment of newly diagnosed peripheral cytomegalovirus retinitis in patients with AIDS.
Am J Ophthalmol 133:467–474
Holen T, Amarzguioui M, Wiiger MT et al (2002) Positional effects of short interfering RNAs
targeting the human coagulation trigger tissue factor. Nucleic Acids Res 30:1757–1766
Kamada M, So A, Muramaki M et al (2007) Hsp27 knockdown using nucleotide-based therapies

inhibit tumor growth and enhance chemotherapy in human bladder cancer cells. Mol Cancer
Ther 6:299–308
Koller E, Propp S, Zhang H et al (2006) Use of a chemically modified antisense oligonucleotide
library to identify and validate Eg5 (kinesin-like 1) as a target for antineoplastic drug develop-
ment. Cancer Res 66:2059–2066
Lesnik EA, Freier SM (1995) Relative thermodynamic stability of DNA, RNA, and DNA:RNA
hybrid duplexes: relationship with base composition and structure. Biochemistry 34:
10807–10815
Lima W, Rose JB, Nichols JG et al (2006) The positional influence of the helical geometry of the
heteroduplex substrate on human RNase H1 catalysis. Mol Pharmacol 71:73–82
Lima WF, Monia BP, Ecker DJ et al (1992) Implication of RNA structure on antisense oligonu-
cleotide hybridization kinetics. Biochemistry 31:12055–12061
Matveeva O, Felden B, Tsodikov A et al (1998) Prediction of antisense oligonucleotide efficacy by
in vitro methods. Nat Biotechnol 16:1374–1375
Mir KU, Southern EM (1999) Determining the influence of structure on hybridization using
oligonucleotide arrays. Nat Biotechnol 17:788–792
Monia BP, Lesnik EA, Gonzalez C et al (1993) Evaluation of 20 -modified oligonucleotides
containing 20 -deoxy gaps as antisense inhibitors of gene expression. J Biol Chem
268:14514–14522
Opalinska JB, Gewirtz AM (2002) Nucleic-acid therapeutics: basic principles and recent appli-
cations. Nat Rev Drug Discov 1:503–514
Overhoff M, Alken M, Far RK et al (2005) Local RNA target structure influences siRNA efficacy:
a systematic global analysis. J Mol Biol 348:871–881
Pradeepkumar PI, Amirkhanov NV, Chattopadhyaya J (2003) Antisense oligonuclotides with
oxetane-constrained cytidine enhance heteroduplex stability, and elicit satisfactory RNase H
response as well as showing improved resistance to both exo and endonucleases. Org Biomol
Chem 1:81–92
Prakash TP, Bhat B (2007) 20 -Modified oligonucleotides for antisense therapeutics. Curr Top Med
Chem 7:641–649
Rayburn ER, Zhang R (2008) Antisense, RNAi, and gene silencing strategies for therapy: mission
possible or impossible? Drug Discov Today 13:513–521
Ricke DO, Wang S, Cai R et al (2006) Genomic approaches to drug discovery. Curr Opin Chem
Biol 10:303–308
Roehr B (1998) Fomivirsen approved for CMV retinitis. J Int Assoc Physicians AIDS Care
4:14–16
Rossi J, Zamore P, Kay MA (2008) Wandering eye for RNAi. Nat Med 14:611
Rudnick SI, Swaminathan J, Sumaroka M et al (2008) Effects of local mRNA structure on
posttranscriptional gene silencing. Proc Natl Acad Sci U S A 105:13787–13792
Schubert S, Grunweller A, Erdmann VA et al (2005) Local RNA target structure influences siRNA
efficacy: systematic analysis of intentionally designed binding regions. J Mol Biol 348:883–893
Schwarz DS, Tomari Y, Zamore PD (2004) The RNA-induced silencing complex is a Mg2+-
dependent endonuclease. Curr Biol 14:787–791
Stein CA (1999) Hybridization prediction gets to first base. Nat Biotechnol 17:751–752
Vickers TA, Koo S, Bennett CF et al (2003) Efficient reduction of target RNAs by small
interfering RNA and RNase H-dependent antisense agents. A comparative analysis. J Biol
Chem 278:7108–7118
Vickers TA, Wyatt JR, Freier SM (2000) Effects of RNA secondary structure on cellular antisense
activity. Nucleic Acids Res 28:1340–1347
Yang Z, Stratton C, Francis PJ et al (2008) Toll-like receptor 3 and geographic atrophy in
age-related macular degeneration. N Engl J Med 359:1456–1463

Yoshinari K, Miyagishi M, Taira K (2004) Effects on RNAi of the tight structure, sequence and
position of the targeted region. Nucleic Acids Res 32:691–699
Yuan YR, Pei Y, Ma JB et al (2005) Crystal structure of A. aeolicus argonaute, a site-specific
DNA-guided endoribonuclease, provides insights into RISC-mediated mRNA cleavage.
Mol Cell 19:405–419
Antisense RNA-Mediated Regulation of the p53
Tumor Suppressor
Marianne Farnebo and Klas G. Wiman
Contents
1 Antisense RNAs as Regulators of Gene Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
2 Regulation of p53 at the mRNA Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
3 Wrap53 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
4 Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Abstract The tumor suppressor p53 triggers cell death by apoptosis in response to
cellular stress. p53 is regulated at the protein level by various posttranslational
modifications, such as phosphorylation and acetylation. However, recent studies
have revealed a critical regulation of p53 at the RNA level. A natural antisense
gene, designated Wrap53, is localized in a head-to-head fashion with p53 on human
chromosome 17p13. Wrap53 mRNA positively regulates steady-state levels of p53
mRNA and p53 protein by targeting the 50 untranslated region of p53 mRNA.
Knockdown of Wrap53 by siRNA results in a significant decrease in p53 mRNA
and suppression of p53 induction upon DNA damage, whereas overexpression of
Wrap53 transcripts containing the antisense overlap region enhances p53 mRNA
and protein levels and sensitizes cells to p53-dependent apoptosis. Antisense
transcription, which occurs widely in mammalian genomes, is thought to play an
important role in regulation of gene expression. Wrap53 antisense RNA is a novel
mechanism for controlling p53 activity and an interesting example of antisense-
mediated gene regulation in human cells.
M. Farnebo (*) and K.G. Wiman

Department of Oncology-Pathology, Cancer Center Karolinska (CCK), Karolinska Institutet,
SE-171 76, Stockholm, Sweden
e-mail: Marianne.Farnebo@ki.se
278 M. Farnebo and K.G. Wiman
Keywords Antisense transcript Cancer DNA damage NAT p53 RNA

regulation Wrap53
1 Antisense RNAs as Regulators of Gene Expression
Natural antisense transcripts (NATs) are a group of regulatory RNAs with sequence
complementarity to other cellular RNAs referred to as sense RNAs. The existence
of antisense RNAs has been known for a long time, but their functional relevance
is still relatively unknown. Nevertheless, studies have shown that antisense
RNAs have the ability to modulate expression of their sense RNA. Considering
the widespread occurrence of antisense transcription in mammalian cells, this
mechanism may have a central role in gene regulation. Antisense RNAs can act
in cis or trans, depending on if the antisense RNA is transcribed from the same or a
distant locus with respect to the sense RNA. Around 20% of all human genes
overlap in a cis-antisense fashion, giving rise to cis-antisense RNAs with perfect
complementarity to their sense RNA (Chen et al. 2004; Yelin et al. 2003). The
orientation and length of the overlap varies between pairs. Most commonly, the
pairs overlap in a head-to-head or tail-to-tail orientation, reflecting overlap between
50 or 30 ends of both transcripts, respectively. However, complete overlap between
pairs is also found. In contrast, trans-antisense RNAs generally display imperfect
complementarity to its sense partner. One example of trans-antisense RNAs is
microRNAs.
The mechanisms of cis-antisense-mediated gene regulation are not fully under-
stood. Several modes of action have been proposed:
1. Regulation at the transcriptional level independently of the antisense transcript.
In this model, transcription of the sense RNA is repressed due to a competition
between the sense and antisense genes for transcription factors. Also, RNA
polymerases might collide if both genes are transcribed simultaneously, result-
ing in transcription termination.
2. Epigenetic alterations induced by the antisense transcript. Here, the antisense
transcript regulates expression of the sense RNA through epigenetic alterations
of the sense promoter region, such as DNA methylation or chromatin remodel-
ing. These processes shut down expression of the sense transcript. This type of
regulation has been found to control inactivation of the X chromosome and
genomic imprinting.
3. RNA–RNA interaction between the sense and antisense transcripts. This mode
of action occurs at the posttranscriptional level and involves base-paring
between the sense and antisense RNAs. RNA–RNA interaction can activate
the RNA interference pathway resulting in sense RNA degradation and endo-
siRNA formation. Alternatively, translation of the sense transcript is blocked.
Conversely, RNA–RNA interaction or duplex formation has been found to
Antisense RNA-Mediated Regulation of the p53 Tumor Suppressor 279
enhance sense RNA stability, possibly through masking sequences within the
sense transcript that otherwise could be recognized by RNA degradation factors.
Transcriptome studies show that antisense transcription occurs for up to 40–70%
of all mammalian genes (Katayama et al. 2005; Engstrom et al. 2006). However,
only a minor fraction of all putative sense–antisense pairs have been verified
and even fewer have been characterized. The identification of Wrap53 does not
only reveal a regulatory pathway for p53 but also demonstrates that RNA–RNA
interaction is a mechanism for antisense-mediated gene regulation in human cells
(Mahmoudi et al. 2009).
2 Regulation of p53 at the mRNA Level
The p53 tumor suppressor triggers cell cycle arrest, senescence, or apoptosis in
response to DNA damage, oncogene activation, and other stress stimuli (Vogelstein
et al. 2000). This allows elimination of incipient tumor cells. p53 is a transcription
factor that upregulates apoptosis-promoting genes such as Bax, Puma, and Noxa.
Several studies have also highlighted the miR-34a microRNA as a p53 target
(Raver-Shapira and Oren 2007). p53 mutation occurs frequently in human tumors,
resulting in evasion of apoptosis and progression towards more malignant pheno-
types. Most p53 mutations give rise to single amino acid substitutions in the DNA-
binding core domain, resulting in disruption of DNA binding and transcriptional
transactivation of target genes (Soussi and Wiman 2007). Accumulating evidence
suggests that mutant p53 proteins can also acquire novel functions that contribute
to tumor growth, e.g., illegitimate transactivation of tumor-promoting genes such
as c-myc.
Under normal conditions, p53 is expressed at very low levels due to rapid protein
turnover. Cellular stress leads to p53 protein stabilization via phosphorylation that
prevents binding to the p53-induced E3 ligase MDM2 that targets p53 for protea-
some-mediated degradation. However, studies on knock-in mice expressing mutant
p53 proteins with substitutions of residues that are phosphorylated upon stress, e.g.,
Ser18 (corresponding to Ser15 in human p53) and Ser23 (Ser20 in human p53),
have indicated that each of these posttranslational modifications by themselves do
not have any significant effect on either basal p53 levels or induction of p53 upon
cellular stress (Zhang and Chen 2008). This suggests that regulation of p53 is
complex and that multiple mechanisms contribute to modulation of p53 expression
in vivo. Other posttranslational modifications, such as acetylation and sumoylation,
are also important. Interestingly, recent studies have revealed that p53 is regulated
at the RNA level. RNA-binding proteins, such as HuR, L26, RPL26, and nucleolin,
can regulate p53 activity by binding to the 50 or 30 UTR of p53 mRNA and affect
mRNA stability and/or translation (Zhang and Chen 2008). Moreover, the miR-125b
microRNA was shown to target the 30 UTR of p53 and modulate p53-induced
apoptosis during development and stress conditions (Le et al. 2009). Our discovery
of Wrap53 adds another level of complexity as to the role of regulatory RNAs for
p53 function (Mahmoudi et al. 2009).
3 Wrap53
Wrap53 is a natural antisense transcript of p53 that regulates basal and stress-
induced endogenous p53 mRNA and protein expression by targeting the 50 untrans-
lated region of p53 mRNA. Our discovery of Wrap53 is the first demonstration of
antisense-mediated regulation of p53. We baptized this gene Wrap53 for WD40
encoding RNA antisense to p53, a name that has been approved by the HUGO Gene
Nomenclature Committee as the official name of this gene. Wrap53 is located on
chromosome 17p13 and overlaps the p53 gene in a head-to-head fashion (Fig. 1).
The gene has three alternative start exons, exon 1a, 1b, and 1g. Exon 1a overlaps
the first exon of p53 in an antisense fashion. Wrap53 also encodes a protein with
homology to members of the WD40 protein family, thus the name Wrap53. This is
in contrast to many other regulatory RNAs that are noncoding and exert their effect
only at the RNA level. The Wrap53 protein (also denoted WDR79 and TCAB1)
was recently identified as a Cajal body protein that binds and directs small
Cajal body-specific RNAs (scaRNAs), including telomerase RNA, to Cajal bodies
(Venteicher et al. 2009; Tycowski et al. 2009).
p53
Chromosome - Strand
17p13.1 + Strand
Wrap53
Transcription
Binding site for RNA degradation factors
p53 mRNA
degradation Wrap53/p53 RNA interaction
protects p53 mRNA from
degradation
p53
protein
protein
Apoptosis Cell cycle arrest
Fig. 1 Model for Wrap53-mediated regulation of p53. Wrap53a and p53 are coexpressed in cells.
Interaction between the two transcripts via their complementary regions (IIIIII) protects p53
mRNA from degradation. Knockdown of Wrap53a or blockage of Wrap53/p53 hybrid formation
leads to reduced p53 mRNA levels. Conversely, overexpression of Wrap53a increases p53 mRNA
levels, and consequently, p53 protein expression. Thus, Wrap53a will potentiate p53-induced cell
cycle arrest and/or apoptosis in response to cellular stress
Our expression analysis revealed a positive correlation between Wrap53 anti-

sense (Wrap53a) and p53 sense transcripts, and a 100-fold higher level of p53
mRNA over Wrap53a. Other studies have shown that the antisense transcript is
often expressed at much lower levels than the corresponding sense transcript
(Katayama et al. 2005; Fish et al. 2007; Oeder et al. 2007), raising the question as
to how the antisense transcripts might regulate their sense partner despite a consid-
erable discrepancy in expression levels. One possibility is that a transient interac-
tion between the complementary transcripts results in a permanent modification of
the p53 mRNA that protects it from degradation even after detachment of the
Wrap53 mRNA. Upon detachment, Wrap53 is free to move on and target the
next p53 sense transcript in a “hit and run” fashion. Another plausible explanation
for this skewed antisense–sense RNA expression is the formation of endo-siRNA
from sense–antisense transcript pairs followed by strand selective degradation of
one of the transcripts (Borsani et al. 2005; Watanabe et al. 2008; Carlile et al. 2008,
2009). Experimental evidence indicates that the protein coding sense transcript
dictates the strand selection, resulting in accumulation of the sense transcript and
degradation of the antisense transcript (Watanabe et al. 2008; Carlile et al. 2008).
We found that Wrap53a is crucial for p53 expression and function, even though it
is expressed at 100-fold lower levels. Wrap53 knockdown significantly reduces
p53 mRNA and protein expression and this is not due to block of transcription
but instead occurs at the posttranscriptional level. Interestingly, knockdown or
overexpression of p53 had no effect on Wrap53a mRNA levels, indicating that
Wrap53 regulates p53 in a nonreciprocal manner.
Several observations demonstrated that the antisense region of the Wrap53a
transcript (i.e., exon 1a) controls p53 expression by interfering with the sense region
of p53 mRNA (i.e., exon 1). First, overexpression of Wrap53 exon 1a efficiently
induced p53 levels. Second, only siRNAs targeting exon1a but not exon1b or 1g of
Wrap53 downregulate p53 expression. Third, depletion of Wrap53 triggers specific
decay of transcripts containing p53 exon 1 sequences. This was demonstrated using
a reporter construct carrying p53 exon 1 fused to luciferase cDNA. The expression
of this construct was almost entirely shut down in Wrap53 depleted cells whereas
luciferase transcripts lacking p53 exon 1 were unaffected. Fourth, 20 -O-methyl
oligonucleotides that bind to the overlapping region and thus prevent interaction
between p53 and Wrap53 transcripts significantly reduced p53 expression, indicating
that Wrap53/p53 RNA–RNA interaction is required to maintain normal p53 levels
in the cell. It is conceivable that Wrap53/p53 RNA hybridization masks target
sequences in p53 mRNA and thus protects it from degradation. Removal of
Wrap53 RNA or blocking this sequence would expose p53 mRNA to putative
regulatory factors, leading to its degradation. Exon 1 (i.e., 50 UTR) of p53 has
previously been suggested to control its own RNA stability. An element within
this region was shown to be involved in rapid destabilization of p53 mRNA in
chicken and mouse cells, and 50 UTR deletion constructs were significantly more
stabile compared to full length p53 constructs (Kim et al. 2001).
Our findings described above raise the important question as to exactly how
Wrap53 regulates p53 mRNA. Although thousands of natural antisense genes have
been identified, the mechanism of antisense-mediated regulation remains obscure.

It is possible that long double-stranded RNA structures may form in vivo, but they
have very short half-lives due to the presence of destabilizing endogenous factors.
The finding that the pri-miRNA processing protein Drosha cleaves double-stranded
RNA with protruding single stranded ends suggests that antisense/sense double-
stranded RNAs may be substrates for Drosha (Han et al. 2006). Recent studies
have indeed identified endo-siRNAs originating from sense–antisense RNA pairs
produced through the RNA interference pathway (Borsani et al. 2005; Watanabe
et al. 2008). However, other studies argue against RNA interference as a mode of
action of overlapping genes (Faghihi and Wahlestedt 2006; Jen et al. 2005). We
have been unable to detect endo-siRNAs from Wrap53 and p53 transcripts in cells
using various approaches including RNAse protection assay, and RNA duplexes
have only been detected in a few cases of endogenous sense–antisense RNA pairs in
human cells (Katayama et al. 2005; Faghihi and Wahlestedt 2006; Munroe and Zhu
2006; Yu et al. 2008). This indicates that RNA duplexes in higher organisms are too
transient and/or labile to allow detection by currently available methods. The
transient nature of the RNA duplexes may be essential in order to avoid activation
of the cellular interferon-mediated pathway that is triggered by the presence of
viral RNA duplexes, resulting in shutdown of protein synthesis and culminating in
apoptosis.
Wrap53/p53 RNA interaction might also enhance p53 stability by influencing
the folding of p53 mRNA. Specific RNA sequences can have a great impact on the
folding of RNAs, such as in the case of the conserved leader sequences of bacterial
rRNA operons required for the maturation of 16S rRNA (Besancon and Wagner
1999) and the proper formation of 30S ribosomal subunits (Balzer and Wagner
1998). Single-stranded transcripts are thermodynamically highly unstable, and
in particular, transcripts derived from (CpG-rich) bidirectional promoters, such as
the Wrap53/p53 promoters, will fold instantly as transcription progresses. Also, the
high speed of RNA folding (Crothers et al. 1974) should theoretically decrease the
chance of long double-stranded RNA formation even if complementary RNAs are
coexpressed within the same cell. Whether Wrap53 RNA affects the folding of the
p53 RNA remains to be elucidated, but clearly the choice of structures adopted can
have profound effects on the function and stability of an individual transcript. Other
well-characterized RNA quality control pathways, such as nonsense-mediated
decay (NMD), nonstop decay, and AU-rich element (ARE)-directed decay of pre-
and mRNA degradation, may also be involved in antisense-mediated regulation,
such as the observed p53 mRNA degradation following Wrap53 knockdown.
Although the exact mechanism behind Wrap53-mediated p53 regulation
remains to be elucidated, the significance of this regulation was demonstrated in
several ways. We found that Wrap53 appears to have a crucial role in the p53-
dependent DNA damage response since depletion of Wrap53a or blockage of
Wrap53/p53 RNA hybrids prevented p53 protein induction and transactivation of
p53 target genes in cells treated with the DNA damaging agent camptothecin. p53
and Wrap53a transcripts were also induced upon DNA damage, indicating that
Wrap53a not only maintains basal p53 mRNA levels but also plays a role in
stabilizing p53 mRNA in response to DNA damage. In addition, overexpression of

Wrap53a sensitized cells for p53-induced apoptosis (Mahmoudi et al. 2009).
Altogether, these data show that Wrap53 has a considerable impact on p53 function
in response to stress, which provides further evidence for a significant functional
role of antisense transcripts in gene regulation. Moreover, our comparative analysis
revealed that the p53 family member gene p73 also has an overlapping cis-antisense
gene, Wrap73/WDR8, which encodes a protein belonging to the same family as
Wrap53 (Mahmoudi et al. 2009; Koshizuka et al. 2001). Thus, the Wrap73/WDR8
and p73 genes not only overlap in a manner similar to that of Wrap53 and p53 but
WDR8 also encodes a protein with structural and perhaps functional similarities to
Wrap53.
4 Future Perspectives
Loss of p53 function is a key step during tumor development, allowing evasion
of apoptosis and accelerated tumor progression. The control of p53 is complex with
tight regulation at both the posttranscriptional and posttranslational levels. The
identification of Wrap53 as a novel regulator of p53 adds important new insights
into the impact of regulatory RNAs on p53 activity. It also opens the possibility
that dysfunction of Wrap53 itself may contribute to cancer. Could lack of Wrap53a
RNA, or insufficient levels of expression, represent a novel mechanism for p53
inactivation in wild type p53-carrying tumors? Future studies should examine
expression of Wrap53 transcripts in primary tumors of various origin. Moreover,
it will be important to determine if manipulation of Wrap53 expression to either
enhance wild type p53 levels or inhibit expression of mutant p53 and putative
gain-of-function activities may be a useful strategy for cancer treatment.
Acknowledgements We thank the Swedish Cancer Society (Cancerfonden), the Swedish Childhood
Cancer Society (Barncancerfonden), and the King Gustaf V Jubilee Fund for generous financial
support.
References
Balzer M, Wagner R (1998) Mutations in the leader region of ribosomal RNA operons cause
structurally defective 30 S ribosomes as revealed by in vivo structural probing. J Mol Biol
276:547–557
Besancon W, Wagner R (1999) Characterization of transient RNA–RNA interactions important
for the facilitated structure formation of bacterial ribosomal 16S RNA. Nucleic Acids Res
27:4353–4362
Borsani O, Zhu J, Verslues PE et al (2005) Endogenous siRNAs derived from a pair of natural
cis-antisense transcripts regulate salt tolerance in Arabidopsis. Cell 123:1279–1291
Carlile M, Nalbant P, Preston-Fayers K et al (2008) Processing of naturally occurring sense/

antisense transcripts of the vertebrate Slc34a gene into short RNAs. Physiol Genomics
34:95–100
Carlile M, Swan D, Jackson K et al (2009) Strand selective generation of endo-siRNAs from the
Na/phosphate transporter gene Slc34a1 in murine tissues. Nucleic Acids Res 37:2274–2282
Chen J, Sun M, Kent WJ et al (2004) Over 20% of human transcripts might form sense–antisense
pairs. Nucleic Acids Res 32:4812–4820
Crothers DM, Cole PE, Hilbers CW et al (1974) The molecular mechanism of thermal unfolding of
Escherichia coli formylmethionine transfer RNA. J Mol Biol 87:63–88
Engstrom PG, Suzuki H, Ninomiya N et al (2006) Complex loci in human and mouse genomes.
PLoS Genet 2:e47
Faghihi MA, Wahlestedt C (2006) RNA interference is not involved in natural antisense mediated
regulation of gene expression in mammals. Genome Biol 7:R38
Fish JE, Matouk CC, Yeboah E et al (2007) Hypoxia-inducible expression of a natural
cis-antisense transcript inhibits endothelial nitric-oxide synthase. J Biol Chem 282:
15652–15666
the Drosha–DGCR8 complex. Cell 125:887–901
Jen CH, Michalopoulos I, Westhead DR et al (2005) Natural antisense transcripts with coding
capacity in Arabidopsis may have a regulatory role that is not linked to double-stranded RNA
degradation. Genome Biol 6:R51
Katayama S, Tomaru Y, Kasukawa T et al (2005) Antisense transcription in the mammalian
transcriptome. Science 309:1564–1566
Kim H, You S, Foster LK, Farris J, Foster DN (2001) The rapid destabilization of p53 mRNA in
immortal chicken embryo fibroblast cells. Oncogene 20:5118–5123
Koshizuka Y, Ikegawa S, Sano M et al (2001) Isolation, characterization, and mapping of the
mouse and human WDR8 genes, members of a novel WD-repeat gene family. Genomics
72:252–259
Le MT, Teh C, Shyh-Chang N et al (2009) MicroRNA-125b is a novel negative regulator of p53.
Mahmoudi S, Henriksson S, Corcoran M et al (2009) Wrap53, a natural p53 antisense transcript
required for p53 induction upon DNA damage. Mol Cell 33:462–471
Munroe SH, Zhu J (2006) Overlapping transcripts, double-stranded RNA and antisense regulation:
a genomic perspective. Cell Mol Life Sci 63:2102–2118
Oeder S, Mages J, Flicek P et al (2007) Uncovering information on expression of natural antisense
transcripts in Affymetrix MOE430 datasets. BMC Genomics 8:200
Raver-Shapira N, Oren M (2007) Tiny actors, great roles: microRNAs in p53’s service. Cell Cycle
6:2656–2661
Soussi T, Wiman KG (2007) Shaping genetic alterations in human cancer: the p53 mutation
paradigm. Cancer Cell 12:303–312
Tycowski KT, Shu MD, Kukoyi A et al (2009) A conserved WD40 protein binds the cajal body
localization signal of scaRNP particles. Mol Cell 34:47–57
Venteicher AS, Abreu EB, Meng Z et al (2009) A human telomerase holoenzyme protein required
for Cajal body localization and telomere synthesis. Science 323:644–648
Vogelstein B, Lane D, Levine AJ (2000) Surfing the p53 network. Nature 408:307–310
Yelin R, Dahary D, Sorek R et al (2003) Widespread occurrence of antisense transcription in the
human genome. Nat Biotechnol 21:379–386
Yu W, Gius D, Onyango P et al (2008) Epigenetic silencing of tumour suppressor gene p15 by its
antisense RNA. Nature 451:202–206
Zhang J, Chen X (2008) Posttranscriptional regulation of p53 and its targets by RNA-binding
proteins. Curr Mol Med 8:845–849
Antisense Oligonucleotides: Insights
from Preclinical Studies and Clinical Trials
Doreen Kunze, Kai Kraemer, and Susanne Fuessel
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
2 Antisense Oligonucleotides in Cancer Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
2.1 BCL2 (B-cell CLL/lymphoma 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
2.2 XIAP (X-Linked Inhibitor of Apoptosis) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
2.3 Survivin, BIRC5 (Baculoviral IAP Repeat-Containing 5) . . . . . . . . . . . . . . . . . . . . . . . . . . 291
2.4 CLU (Clusterin) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
2.5 TGFB2 (Transforming Growth Factor, b 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
3 Application of Antisense Oligonucleotides in Noncancerous Diseases . . . . . . . . . . . . . . . . . . . 294
3.1 Asthma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
3.2 Cardiovascular Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
3.3 Duchenne Muscular Dystrophy – Exon-Skipping Therapy . . . . . . . . . . . . . . . . . . . . . . . . . 295
3.4 Virus Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
4 Specificity of Antisense-Mediated Gene Silencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
Abstract Since the first pioneering studies using antisense oligonucleotides

(ASOs) in the late 1970s, thousands of publications followed, demonstrating the
remarkableness of antisense action and its enormous application spectrum. In 1998,
Fomivirsen (Vitravene) was the first, and to date the only ASO that gained approval
by the US Food and Drug Administration (FDA) for intravitreous treatment of
cytomegalovirus-induced retinitis in patients with acquired immune deficiency
syndrome (AIDS). Meanwhile, efforts regarding ASO research decreased and
investigations shifted to other molecules, e.g., small interfering RNAs, because
ASO-related problems such as insufficient efficacy and off-target effects are not
yet overcome. However, newer studies using ASOs with improved chemistry or
D. Kunze (*), K. Kraemer, and S. Fuessel

Department of Urology, Medical Faculty, Dresden University of Technology, Fetscherstraße 74,
01307 Dresden, Germany
e-mail: doreen.kunze@uniklinikum-dresden.de
286 D. Kunze et al.
approaches combining ASO treatment with other therapies, such as chemotherapy

or radiation, might bring ASOs back into the spotlight. This chapter will focus on
current in vivo studies and clinical trials of promising ASOs.
Keywords Antisense oligonucleotides Cancer treatment Clinical trials In vivo

studies Asthma Cardiovascular disease Duchenne muscular dystrophy Virus
infection
Abbreviations
ApoB-100 Apolipoprotein B-100

AIDS Acquired immune deficiency syndrome
AML Acute myeloid leukemia
ASO Antisense oligonucleotide
BCL2 B-cell CLL/lymphoma 2
CLL Chronic lymphocytic leukemia
CpG Cytosine–guanine dinucleotide
DMD Duchenne muscular dystrophy
FDA US food and drug administration
FH Familial hypercholesterolemia
HBV Hepatitis B virus
HIV Human immunodeficiency virus
IAP Inhibitor of apoptosis protein
LDH Lactate dehydrogenase
LNA Locked nucleic acid
LDL-c Low density lipoprotein cholesterol
miRNA Micro ribonucleic acid
miR-122 MicroRNA-122
NSCLC Nonsmall cell lung cancer
PCa Prostate cancer
PMO Phosphoroamidate morpholino oligomer
PNA Peptide nucleic acid
PPMO Peptide-conjugated phosphoroamidate morpholino oligomer
PS Phosphorothioate
TGFB2 Transforming growth factor, b 2
XIAP X-linked inhibitor of apoptosis
20 -OMe 20 -O-Methyl
20 -MOE 20 -O-(20 -Methoxy)ethyl
Antisense Oligonucleotides: Insights from Preclinical Studies and Clinical Trials 287
Nucleus Cell membrane Cytoplasma
RNA polymerase
DNA STOP
a DNA
Transcription ASO
RNA Triplex formation
Pre-mRNA
Post-transcriptional
modification Impairment of mRNA
mRNA processing b
RNase H activation
Ribosome
STOP mRNA c
d Steric blockade
Protein
Fig. 1 Mechanism of ASO action: (a) interference with transcription via triplex formation with
complementary DNA; (b) inhibition of splicing or destabilization of mRNA by hindrance of
polyadenylation and 50 -capping after hybridization to pre-mRNA; (c) mRNA degradation after
induction of RNase H activity; (d) translational arrest following binding of the ASO in the region
of the initiation codon of the target mRNA and sterical blocking of the ribosome
1 Introduction
Antisense oligonucleotides (ASOs) are short (15–20 nucleotides), single-stranded

DNA molecules that are designed complementary to the mRNA of a selected target
gene. After introduction into the cell, ASOs hybridize with the target mRNA via
Watson–Crick base pairing and inhibit protein biosynthesis either by induction of
RNase H-mediated mRNA cleavage or by steric blockade of the ribosomes (Fig. 1).
Besides these two main modes of action, other mechanisms are proposed. ASOs can
bind to DNA and inhibit transcription due to triplex formation or affect mRNA
processing after hybridization with pre-mRNA (Crooke 2004a).
The first described ASOs were phosphodiester oligonucleotides, which are
rapidly degraded by nucleases in vivo. Consequently, chemical modifications of
the phosphodiester linkage, the heterocycle, or the sugar were introduced into the
oligonucleotides to increase their stability (Fig. 2). In “first generation” ASOs, a
nonbridging oxygen atom of the phosphodiester bond is replaced by a sulfur atom to
create a phosphorothioate (PS) backbone. These PS–ASOs are currently the most
commonly used ones, even though they can induce nonsequence-specific toxicities
because of their affinity to bind to proteins, e.g., serum albumin (Kurreck 2003;
288 D. Kunze et al.
O B O B
1st
O H O H
– –
O P O O P S
O O
Unmodified 2nd PS
O B O B
3rd O O CH3 O O
– – O CH3
O P O O P O
O O
2'-OMe 2'-MOE
O B
O B
B N
N
O O O
O
– O P N
O P O O NH
O O
LNA PNA PMO
Fig. 2 Examples for the three generations of chemically modified ASOs: phosphorothioate (PS)
backbone; 20 -O-methyl-(20 -OMe)- and 20 -O-methoxyethyl-(20 -MOE)-RNA-substitutions; locked
nucleic acid (LNA), peptide nucleic acid (PNA), and phosphoroamidate morpholino (PMO)
modifications. B – bases (adenine, thymine, guanine, cytosine)
Crooke 2004a). Furthermore, the PS-backbone can lead to reduced target-mRNA

binding. To overcome these problems, “second generation” ASOs with 20 -alkyl
modifications of the ribose were developed, most importantly 20 -O-methyl
(20 -OMe) and 20 -O-(20 methoxy) ethyl (20 -MOE) substitutions. These ASOs show
increased target affinity and enhanced tissue half-life due to improved nuclease
resistance (Zellweger et al. 2001). Furthermore, they have the potential for oral
administration (Tillman et al. 2007). Since 20 -OMe- and 20 -MOE–ASOs are not
capable of inducing RNase H, they exert their activity via inhibition of translation.
To recover RNase H mobilization, gapmer ASOs composed of unmodified or
PS-modified nucleotides in the center and 20 -OMe- or 20 -MOE-elements at the
ends were developed (Monia et al. 1993). Meanwhile, ASOs of the “third generation”
(e.g., phosphoroamidate morpholino oligomers (PMOs), locked nucleic acids
(LNA), or peptide nucleic acids (PNA)) with further improved properties, such as
enhanced target affinity and nuclease resistance as well as reduced nonspecific
activity, are analyzed (Kurreck 2003).
The main field of application for ASOs is the treatment of diseases that involve
the overexpression of a detrimental gene, e.g., tumor growth or viral infections.
Additionally, oligonucleotides containing unmethylated cytosine–guanine dinucle-
otide (CpG) motifs can stimulate the mammalian immune system (Vollmer and Krieg
2009). Hence, they are tested as vaccine adjuvants for cancer, asthma, and allergies.
Since the discovery of ASOs in the late 1970s, numerous in vitro and in vivo studies
have been performed showing the possibilities and limitations of these constructs.
In 1998, fomivirsen, a 21-mer PS–ASO, inhibiting viral proteins from the major
immediate early transcriptional unit, received FDA approval for treatment of
cytomegalovirus-induced retinitis in AIDS patients. Up to now, no other ASO has
reached the clinic although promising candidates are in development.
2 Antisense Oligonucleotides in Cancer Treatment
Numerous targets, particularly tumor-related genes encoding oncoproteins or

signaling molecules, were inhibited in various in vivo studies and subsequently
in clinical trials, e.g., PRKCA, RAF1, and HRAS (Crooke 2004b; Stahel and
Zangemeister-Wittke 2003). Since defective apoptosis (¼ programmed cell death)
mechanisms substantially contribute to tumor development and progression as well
as to the development of resistances to standard anticancer therapies, this chapter
focuses on promising ASOs targeted at important proteins with antiapoptotic
function. Furthermore, the antisense-mediated inhibition of the multifunctional
cytokine “transforming growth factor, b 2” (TGFB2) might have a promising future.
2.1 BCL2 (B-cell CLL/lymphoma 2)
G3139 (Genasense, Oblimersen Sodium), an 18-mer first-generation PS–ASO

containing two CpG motifs, is extensively studied in numerous clinical trials. Its
target, the antiapoptotic protein BCL2, blocks the release of cytochrome C from the
mitochondria and consequently inhibits the activation of the caspase cascade (Call
et al. 2008). BCL2 is overexpressed in many tumors, including chronic leukemia,
melanoma, prostate, and lung cancer and promotes resistance to chemotherapy
(Reed 1995). In vitro and in vivo studies proved that BCL2 protein and mRNA
expression levels can be reduced concentration- and time-dependently by G3139.
Additionally, these studies showed chemo-sensitizing activity of the drug (reviewed
in Klasa et al. 2002). Recently, promising preclinical studies using bispecific ASOs
with binding sites for two different targets, e.g., BCL2/EGFR and BCL2/clusterin,
and combination treatments with two target-specific ASOs plus chemotherapy
(G3139/docetaxel/MYC-ASO) were published showing a possible future of antisense
therapy (Leonetti et al. 2007; Rubenstein et al. 2009).
290 D. Kunze et al.
At present, there are 16 phase I/II and 6 phase III clinical trials with G3139,
either alone or in combination with various anticancer agents, in hematologic
malignancies and solid tumors ongoing (www.clinicaltrials.gov)1. In 2000, the
first phase III study started comparing G3139 plus dacarbazine versus dacarbazine
alone in 771 patients with advanced melanoma (Bedikian et al. 2006). Median
overall survival in the combination group was longer (9.0 vs. 7.8 months) but did
not reach statistical significance. However, the study showed a correlation between
pretreatment serum lactate dehydrogenase (LDH) level and outcome after G3139
therapy. Survival and overall response significantly improved in the G3139 group
in participants with normal LDH, whereas patients with elevated LDH did not
benefit from G3139 treatment. Therefore, AGENDA, a second randomized phase
III trial in chemo-naı̈ve patients with advanced melanoma and low baseline LDH
was initiated comparing dacarbazine G3139 treatment (www.clinicaltrials.gov).
Altogether, 241 patients with relapsed or refractory chronic lymphocytic leukemia
(CLL) were enrolled in a phase III study with fludarabine plus cyclophosphamide
G3139 (O’Brien et al. 2007). Response rate (complete plus nodular partial
response) was significantly higher (17% vs. 7%) and longer in the G3139 combina-
tion arm. In a phase III study enrolling 503 patients with acute myeloid leukemia
(AML), the addition of G3139 to chemotherapy with cytarabine and daunorubicin
did not change patients’ outcome (Marcucci et al. 2007). Likewise, no significant
difference in time to tumor progression was found after dexamethasone G3139
treatment in a phase III trial involving 224 patients with relapsed or refractory
multiple myeloma (Chanan-Khan et al. 2009). Up to now, no results regarding the
phase III trial examining G3139 in combination with docetaxel in patients with
nonsmall cell lung cancer (NSCLC) have been published.
Most important adverse events reported from the phase III trials so far are grade
3 and 4 neutropenia and thrombocytopenia (Bedikian et al. 2006; O’Brien et al.
2007). Furthermore, catheter-related problems, which are due to the continuous
intravenous infusion of the PS–ASO, increased in the G3139 groups. Alternative
administration routes like short intravenous infusions or subcutaneous injections
might produce relief (Lin et al. 2007; www.genta.com). Safety and maximum
tolerated dose of G3139 administered once or twice per week as 2 h intravenous
infusion are currently examined in a phase I clinical trial in patients with solid
tumors (www.clinicaltrials.gov).
SPC2996, an LNA-modified ASO directed at BCL2, is clinically tested in CLL
(www.santaris.com). In the first dose-escalating study, all six patients in the treat-
ment group with maximum drug concentration (4 mg/kg/dose) showed a reduction
in lymphocyte count (Tilly et al. 2007). A second phase I/II clinical trial examining
new dosing regimens is completed (www.santaris.com). So far, however, no results
have been published.
1
All websites mentioned in the text were viewed on July 15th, 2009.
2.2 XIAP (X-Linked Inhibitor of Apoptosis)
AEG35156 (GEM640) is a gapmer ASO without CpG motif targeting XIAP. It

consists of 11 central PS-nucleotides that are flanked by four 20 -OMe-nucleotides
on each side (Lacasse et al. 2005). XIAP, also known as “baculoviral IAP repeat-
containing 4” (BIRC4), is the only member of the inhibitor of apoptosis protein
(IAP) family that directly binds and inhibits both initiator and effector caspases
(Schimmer et al. 2006). The ability to suppress apoptosis triggered by different
stimuli such as the mitochondrial and the death receptor-mediated pathways makes
XIAP a promising target in antisense therapy.
Expression of XIAP is upregulated in different tumor entities, including NSCLC,
prostate cancer (PCa), and AML. XIAP knockdown mediated by AEG35156
increased tumor cell sensitivity towards chemotherapeutics in different cancer xeno-
graft models (reviewed in Lacasse et al. 2005). The first dose-escalation study
including 38 patients with advanced tumors examined different dosing regimes of
AEG35156, i.e., 7 or 3 days of continuous intravenous infusion in a 21-day treatment
cycle (Dean et al. 2009). XIAP knockdown was detected in peripheral-blood
mononuclear cells but not in tumor biopsies obtained from five patients. Maximum-
tolerated AEG35156 doses were 125 mg/m2/day in the seven and 213 mg/m2/day in
the 3-day-cohort, respectively. One patient with non-Hodgkin0 s lymphoma showed a
marked but short-lived reduction in circulating tumor cells. Two participants (breast
cancer, malignant melanoma) experienced unconfirmed partial responses but pro-
gressed after treatment. Only grade 1 or 2 adverse events were reported at drug doses
of 48 and 96 mg/m2/day, respectively. However, at higher doses, grade 3 or 4 adverse
events like elevated alanine transaminase and aspartate transaminase levels, throm-
bocytopenia, lymphopenia, and pulmonary venous thrombus occurred.
AEG35156 is currently examined in eight phase I/II clinical trials, thereof once
as single agent in hematologic malignancies, once together with sorafenib in
patients with advanced hepatocellular carcinoma and seven times in combination
with chemotherapy in patients with AML, NSCLCS, pancreatic, or breast cancer
(www.clinicaltrials.gov).
2.3 Survivin, BIRC5 (Baculoviral IAP Repeat-Containing 5)
LY2181308 (ISIS 23722) is a 20 -MOE–ASO targeting survivin. This IAP family

member is strongly expressed in malignant, embryonic, and fetal tissues while it
is absent in nearly all differentiated tissues (reviewed in Altieri 2003). Besides its
antiapoptotic activity, survivin seems to play a role in angiogenesis. Its overexpres-
sion is associated with resistances to chemotherapy and radiation (reviewed in Ryan
et al. 2009). Recently, antitumoral activity of an ASO targeted at survivin was
shown in a bladder cancer xenograft model in mice. Furthermore, the study
292 D. Kunze et al.
demonstrated feasibility and antiproliferative effects of combined ASO-mediated

inhibition of survivin, hTERT, and VEGF in vitro (Kunze et al. 2008).
In preclinical studies, LY2181308 mediated survivin inhibition, induced apoptosis,
and sensitized PCa cells to paclitaxel (Fisker et al. 2007). Patients with advanced or
metastatic malignancies were treated intravenously with LY2181308 in a phase I
clinical trial. The maximum tolerated dose was 750 mg (Talbot et al. 2008). Tumor
biopsies, obtained from 22 patients before treatment and 48 h after the last injection,
demonstrated reduction of nucleic and cytoplasmic survivin protein expression in
11/17 and 5/14 of the evaluable pairs (Talbot et al. 2009). Adverse events were mild
to moderate, with no grade 3 or 4 toxicities (Talbot et al. 2008). In 2008, two phase
II clinical trials started either analyzing LY2181308 docetaxel in hormone
refractory PCa or LY2181308 in combination with idarubicin and cytarabine in
patients with AML (www.clinicaltrials.gov).
Recently, EZN3042 (SPC3042), a survivin-targeting 16-mer LNA-modified
ASO (gapmer: central PS-nucleotides that are flanked by 7 LNAs) entered clinical
development in a phase I/II trial in patients with solid tumors and lymphomas
(www.santaris.com). Previously, EZN3042 had been demonstrated in vivo to
inhibit prostate and lung cancer growth and to sensitize these cells to paclitaxel
treatment (www.enzon.com; Hansen et al. 2008).
2.4 CLU (Clusterin)
OGX-011 (Custirsen) is a second-generation gapmer ASO composed of 13 central

PS-oligonucleotides that are flanked by four 20 MOE-nucleotides at both the 30 - and
50 -ends (Chi et al. 2008). It contains one CpG motif and is targeted at CLU exon II
mRNA AUG translation initiation site. CLU, also known as “testosterone-repressed
prostate message 2,” “sulfated glycoprotein-2,” or “apolipoprotein J,” is a stress-
induced cytoprotective chaperone protein that is overexpressed in a variety of
cancers including those of the bladder, breast, lung, and prostate (reviewed in Chi
et al. 2008). Its expression is induced by standard anticancer treatments, including
radiation and hormone ablation therapy (Miyake et al. 2000; Zellweger et al. 2002).
In humans, two isoforms exist, a proapoptotic nuclear one and an antiapoptotic
secreted one (Zhang et al. 2006), whereof OGX-011 selectively inhibits the expres-
sion of the antiapoptotic form (Cao et al. 2005).
Treatment with OGX-011 reduced CLU expression and enhanced tumor cell
sensitivity towards chemo- and radiotherapy in cell culture and xenograft models of
various tumor entities (Zellweger et al. 2001; Cao et al. 2005). Preclinical studies
did not reveal significant signs of toxicity when OGX-011 is applied at doses of up
to 50 mg/kg in mice or of up to 10 mg/kg in monkeys (Chi et al. 2008). At the
highest dose, alterations in liver function (elevated transaminase levels) and
immune stimulation (lymphohistiocytic cell infiltrates) in mice and minor evidence
of complement activation in monkeys were the primary toxicities.
Up to now, approximately 300 patients have been treated with OGX-011 in six
phase I and II clinical trials (Chi et al. 2008). In the first phase I dose-escalation
study, 25 patients with high-risk localized PCa were treated with 40–640 mg
OGX-011 given seven times as a 2 h intravenous infusion prior to radical prosta-
tectomy. Besides the safety of this ASO, the biological activity was demonstrated,
too. The study showed that OGX-011 treatment is well tolerated, with adverse
events limited to grade 1 or 2 (e.g., leucopenia, thrombocytopenia, fever, fatigue)
(Chi et al. 2005). Furthermore, a statistically significant dose-dependent reduction
of CLU expression in PCa and lymph node tissue was detected after OGX-011
treatment in comparison to control tissues taken from a tumor bank.
In ongoing clinical studies, OGX-011 is applied in combination with docetaxel,
mitoxantrone, or gemcitabine–cisplatin in patients with prostate, lung, or breast
cancer (www.clinicaltrials.gov). Preliminary results indicate that combined treat-
ment with OGX-011 plus standard chemotherapy is well tolerated (reviewed in Chi
et al. 2008). The mostly single-armed study design hampers the interpretation of
response rates and survival data. However, in a randomized phase II trial, encour-
aging improvement of median survival (median follow-up 32 months) was seen in
82 patients with metastatic castration-resistant PCa who have been treated with
docetaxel þ prednisone þ OGX-011 (27.5 months) compared to docetaxel þ pred-
nisone alone (16.9 months), though the primary endpoint, 50% PSA decline from
baseline, was achieved in both treatment arms (Chi et al. 2009; www.oncogenex.
com). Currently, two randomized phase III trials with OGX-011 plus docetaxel with
overall survival or durable pain palliation as primary end points in patients with
castrate-resistant PCa are in preparation (http://www.isispharm.com).
2.5 TGFB2 (Transforming Growth Factor, b 2)
AP 12009 (Trabedersen) is a PS–ASO complementary to TGFB2, which is over-

expressed in numerous tumor entities. TGFB2 is a multifunctional cytokine that
contributes to tumor development by regulating proliferation, angiogenesis, immu-
nosuppression, as well as invasion and metastasis. In vitro, AP 12009 mediated
inhibition of TGFB2 protein expression and tumor cell migration (reviewed in
Schlingensiepen et al. 2008).
In 24 patients with malignant glioma, intratumoral AP 12009 injection by
convection-enhanced delivery was well tolerated and resulted in complete tumor
remission in two patients (Schlingensiepen et al. 2008). In a phase IIB study, 145
patients with recurrent or refractory high-grade glioma were randomized either to
receive 10 or 80 mM AP 12009 or standard chemotherapy. In a subgroup of 39
patients with anaplastic astrocytoma, treatment with 10 mM AP 12009 was most
effective, with a survival rate of 83.3% after 24 months compared to 41.7% in the
control group (Bogdahn et al. 2009). Therefore, a phase III clinical trial in patients
294 D. Kunze et al.
with recurrent or refractory anaplastic astrocytoma (SAPPHIRE) has been started in

December 2008 (www.clinicaltrials.gov).
Furthermore, in a phase I dose-escalation study in 33 patients with advanced
malignant melanoma, pancreatic, or colorectal carcinoma, effects of intravenous
AP 12009 infusions are examined with two different treatment schedules. Preliminary
results show excellent drug safety and tolerability as well as encouraging survival
data, e.g., complete response in one patient with pancreatic carcinoma (Oettle et al.
2009). Hence, two actively controlled phase II clinical trials in patients with
pancreatic carcinoma and malignant melanoma, respectively, are in planning stages.
3 Application of Antisense Oligonucleotides

in Noncancerous Diseases
In addition to antitumor therapy, ASOs are tested in the treatment of noncancerous

diseases. Particularly, the inhibition of targets overexpressed in liver and kidney,
the organs showing the highest ASO concentrations after systemic drug delivery,
seems to be promising. Furthermore, ASOs have high potential in the treatment of
lung diseases and, due to their ability to affect splicing, in the therapy of severe
muscular dystrophy.
3.1 Asthma
TPI ASM8 is a drug for asthma treatment, which consists of two PS–ASOs targeted
at human chemokine receptor 3 (CCR3) and the common b-chain of IL-3, IL-5, and
GM-CSF receptors, respectively. A study in monkeys showed safety and tolerability
of TPI ASM8 after 14 days of inhalation of up to 2.5 mg/kg/day (Guimond et al.
2008). ASOs were mainly localized in the pulmonary tract with only limited
distribution in plasma, liver, and kidney.
In humans with mild atopic asthma, inhalation of 1.5 mg TPI ASM8 on four
consecutive days reduced influx of sputum inflammatory cells by 46% (Gauvreau
et al. 2008). Furthermore, TPI ASM8 treatment decreased the early asthmatic
response and mRNA levels of allergen-induced targets in sputum-derived cells
without serious adverse events. A new phase II study examining efficacy and safety
of escalating dose regimens of TPI ASM8 in patients with allergic asthma has been
started in April 2009 (www.clinicaltrials.gov).
3.2 Cardiovascular Disease
ISIS 301012 (mipomersen), a 20-mer PS–ASO with five 20 -MOE-modified nucleo-

tides at each end, is targeted at apolipoprotein B-100 (apoB-100), the main
structural protein of low density lipoprotein cholesterol (LDL-c) (Yu et al. 2007).
The inhibition of elevated LDL-c levels in patients is an attractive approach, since
high LDL-c is a risk factor for atherosclerosis. In particular, ASO-mediated apoB-
100 inhibition seems to be promising, because apoB-100 is synthesized in the liver,
the organ where ISIS 301012 predominantly concentrates after intravenous or
subcutaneous application (Yu et al. 2007).
In a phase I study with 36 healthy volunteers, a dose-dependent reduction of
serum apoB-100 and LDL-c was measured after injections of 50–400 mg ISIS
301012 (Kastelein et al. 2006). No serious adverse events were reported. Similar
results were obtained in patients with high levels of cholesterol in the blood
(reviewed in Yu et al. 2009). Mipomersen can be safely administered in combina-
tion with other LDL-c lowering drugs like simvastatin or ezetemibe (Yu et al.
2009). It is currently examined in three phase II and four phase III clinical trials
(www.clinicaltrials.gov). Recently, a placebo-controlled phase III clinical trial in
51 patients with homozygous familial hypercholesterolemia (FH), a genetic dis-
order characterized by high LDL-c level leading to increased risk of cardiovascular
disease, met its primary and secondary endpoints (www.isispharm.com, press
release 05/20/09). FH patients, being on standard lipid-lowering therapy, were
treated weekly with 200 mg ISIS 301012 for 26 weeks. LDL-c, apoB-100, and
total cholesterol were significantly decreased. Reported adverse events were injec-
tion site reactions, flu-like symptoms, and elevations in liver transaminases.
3.3 Duchenne Muscular Dystrophy – Exon-Skipping Therapy
Duchenne muscular dystrophy (DMD) is a common and severe X-linked hereditary

muscle disease caused by mutations in the DMD gene. Due to a shift in the
transcripts’ reading frame, complete dystrophin expression is lost. ASO-mediated
exon-skipping therapy aims at the specific removal of the targeted, defective exon.
Changing the out-of-frame mutation into an in-frame mutation should restore
synthesis of a truncated, semifunctional protein, thereby reducing the severe DMD
phenotype. DMD can be triggered by various mutations. Theoretically, 83% of them
should be treatable with exon-skipping therapy (Aartsma-Rus et al. 2009).
Recent studies in mice showed for the first time restoration of dystrophin expres-
sion in all muscle cells, including cardiac muscle, after intravenous injection of a
peptide-conjugated PMO (PPMO, exhibits further improved stability and cellular
uptake) that mediates exon 23 skipping (Jearawiriyapaisarn et al. 2008). Further
studies aim at simultaneous skipping of multiple exons, since the high number
of different DMD causing mutations makes ASO-mediated single exon skipping
applicable only for a limited number of patients. Feasibility of this approach was
recently shown in DMD dogs (Yokota et al. 2009). A mixture of three PMOs
(120–200 mg/kg in total), which mediates skipping of exons 6–9 was intravenously
296 D. Kunze et al.
injected weekly or biweekly for 5–22 weeks. Two weeks after treatment, increased
dystrophin expression and reduced inflammatory signals were detected.
AVI-4658 is a PMO designed to skip exon 51 of the DMD gene. It is currently
examined in two phase I/II clinical trials (www.clinicaltrials.gov). First results
demonstrated that injection of 0.09 or 0.9 mg AVI-4658 into the exterior digitorum
brevis muscle has been well tolerated and has significantly increased dystrophin
synthesis in DMD patients (www.avibio.com, press release 01/21/09). The second
trial, which was started in January 2009, will evaluate effects of intravenous
drug application. Furthermore, AVI-5038, a PPMO for skipping of exon 50, is in
preclinical development (www.avibio.com).
PRO051 is a 20 OMe-PS–ASO mediating exon 51 skipping. In a first study, four
DMD patients received a single injection of 0.8 mg PRO051 into the tibialis anterior
muscle (van Deutekom et al. 2007). Twenty-eight days later, biopsy samples from
the patients were obtained showing induction of dystrophin expression.
3.4 Virus Infections
Currently, only few antiviral ASOs are in stage of development. Three PMO–ASOs
targeted at VP35, VP24, and RNA polymerase L mRNAs of Ebola Zaire virus,
respectively, showed as single agents and in combination, profound virus inhibition
in mice (Warfield et al. 2006). In rhesus macaques, only the combination was
effective and protected three out of four primates from lethal Ebola virus dose.
ASOs against VP35 and VP24 are now tested in the drug AVI-6002. AVI-6003,
also based on PMO chemistry, is analyzed in the treatment of Marburg virus. It
prevented 100% of treated monkeys from virus related death (www.avibio.com).
The number of treated animals, however, is not mentioned.
A novel approach in antisense therapy is the inhibition of microRNAs (miRNAs),
small noncoding RNAs that posttranscriptionally regulate expression of estimated
30% of the protein-coding genes (reviewed in Wiemer 2007). MicroRNA-122
(miR-122) is a liver-specific miRNA that is implicated in the replication of hepatitis
C virus (reviewed in Niepmann 2009). Mice were treated on three consecutive
days with 2.5–25 mg/kg of a 16-mer LNA-modified ASO targeted at miR-122.
Intravenous ASO injection reduced miR-122 expression in murine livers in a
concentration-dependent manner (Elmén et al. 2007). Effects were strongest
24 h after treatment and disappeared after 2 weeks. No hepatotoxicity was induced.
Further studies in monkeys showed safety, drug uptake in hepatocytes, and
consequent miR-122 decrease after systemic treatment with 3 or 10 mg/kg ASO
(Elmén et al. 2008). In May 2008, a clinical phase I study in healthy men was
started to evaluate the safety of SPC3649, an LNA targeted at miR-122 (www.
clinicaltrials.gov). A phase II study examining SPC3649 in patients with hepatitis
C is in planning stages (www.santaris.com).
4 Specificity of Antisense-Mediated Gene Silencing
In theory, the nature of antisense technique provides an excellent specificity for

the selected target mRNA due to Watson–Crick base-pairing that allows entire
hybridization only between ASO and target mRNA. In practice, the biological
effects sometimes differ from those expected. Besides the specific targeting of a
certain mRNA, effects that are not directly related to antisense hybridization with
the target mRNA have been observed frequently, e.g., toxicities, proinflammatory
reactions, or changes in the expression of untargeted genes. Important sources of
adverse in vitro effects are reagents commonly used to promote cellular uptake
of ASOs, mainly lipid-based transfection agents. In the following paragraph, off-
target effects directly caused by the ASO molecule itself, either by its nucleic acid
backbone chemistry or by its nucleotide sequence, are reviewed in more detail.
In terms of toxicology, sequence-independent off-target effects are associated
with a certain toxicological profile that characterizes ASOs as a particular class of
drugs. To date, several reproducible “class effects” of ASOs have been described
in animal experiments. Primate toxicology studies revealed acute effects like
hypotension, activation of the alternative complement pathway, and inhibition of
the intrinsic coagulation pathway as the most serious toxicities (Jason et al. 2004).
Complement activation as well as inhibited blood coagulation seem to be related to
protein binding of ASOs. Factor H, which normally inhibits complement activation,
is bound by ASOs, which may lead to its inactivation followed by complement
activation (Henry et al. 1997). The procoagulant thrombin can also be bound by
ASOs causing a decreased thrombin activity and an increased coagulation time
in the presence of ASOs (Wasan et al. 2002). In general, the cardiovascular
response seems to be associated with the chemical structure of ASOs, particularly
with PS-backbone modifications. The charged backbone of PS–ASOs promotes
sequence-independent interactions with certain proteins, which may cause cellular
toxicity. Second-generation ASOs can be used to overcome toxicities related to the
backbone (Kurreck 2003).
Aprinocarsen (ISIS 3521), a PS–ASO targeting PRKCA, was tested in a dose-
escalating clinical phase I trial revealing transiently increased complement proteins
C3a and Bb without clinical evidence of complement activation. Moreover, a
transiently inhibited coagulation pathway (prolonged partial thromboplastin time)
in relation to the ASO dose but without clinical significance has been shown
(Advani et al. 2005). In contrast, GEM231, a gapmer ASO targeted at PKA RIa,
also caused a prolonged partial thromboplastin time but with no relation to ASO
dose and without complement activation (Goel et al. 2003).
The aforementioned sequence-independent off-target effects of ASOs were
relatively moderate at intended therapeutic doses as reported in several phase I/II
trials (Chi et al. 2008; Talbot et al. 2008; Dean et al. 2009). A major unexpected
side effect that might be caused by ASOs is intense stimulation of the immune
system. This effect was shown to be dependent on ASO sequence but irrespective of
its antisense function.
298 D. Kunze et al.
Guanosine-rich ASOs designed to target a particular mRNA can bind to DNA

via Hoogsteen hydrogen bonds resulting in the formation of intermolecular quad-
ruplex structures that are stabilized by the presence of G-quartets (Dapić et al.
2003). Quadruplex formation is unfavorable for ASOs, because it impairs their
function of target inhibition by reducing their concentration available for hybridi-
zation. Furthermore, biological effects of ASOs caused by target inhibition can be
masked by quadruplex effects. Various biological effects of quadruplex-forming
oligonucleotides have been described ranging from antiproliferative and antiviral
responses to inhibition of specific enzymatic activities like that of topoisomerase 1
(Marchand et al. 2002; Bates et al. 2009). Although the off-target effects of G-rich
oligonucleotides are generally regarded as undesirable for antisense experiments,
the distinct properties of these molecules can be exploited for cancer therapy, e.g.,
increased nuclease resistance afforded by the quadruplex structure and enhanced
cellular uptake (Bates et al. 2009). However, the mechanism of action is rather the
specific binding of a certain protein (aptamer effect) than antisense hybridization.
Bates et al. discovered a G-rich oligonucleotide (AS1411) with anticancer activity
on several models and found binding of the cancer-associated protein nucleolin as
mechanism of action. AS1411 is now in phase II clinical trials for cancer therapy
(Bates et al. 2009).
Another particular sequence that can cause off-target effects consists of CpG
dinucleotides. Unmethylated CpG motifs are known to stimulate immune responses
in vivo mainly by upregulation of toll-like receptor 9 expression (Agrawal and
Kandimalla 2003). Two 50 purines and two 30 pyrimidines flanking the CpG motif
result in an amplification of immune stimulation. However, the immunostimulatory
effects of CpG oligonucleotides can be utilized for therapeutic approaches.
Hartmann et al. screened a broad panel of CpG oligonucleotides to stimulate
proliferation of B cells from rhesus monkeys and chimpanzees in vitro. The most
active construct was used as an adjuvant in these primates immunized against
hepatitis B virus (HBV). Increased antihepatitis B antibody titres were detected in
animals receiving vaccine and adjuvant, compared to those receiving the vaccine
only (Hartmann et al. 2000). A randomized, double-blind controlled clinical trial in
HBV-susceptible HIV-infected individuals revealed a rapid, higher, and more
sustained HBV seroprotection and increased HBV-specific T helper cell response
to HBV vaccine when the vaccine was administered together with the CpG oligo-
nucleotide CPG 7909 (Cooper et al. 2005).
Regarding tumor therapy, it is difficult to separate true antisense effects of ASOs
containing CpG motifs from CpG-related off-target effects, since off-target effects
can contribute to the anticancer activity by immune stimulation and induction of
several cytokines, e.g., IL12, IL6, IFNG, and TNF (Agrawal and Kandimalla 2001).
G3139, containing two CpG motifs, has shown antitumor activity in vitro and
in vivo. However, several off-target effects have been reported and it is not clear
to which extent its antitumor function is related to BCL2 inhibition. Cytosine C-5
methylation of G3139 resulted in a loss of its ability to activate mouse B cells
in vitro. This was associated with a significant decrease of the antitumor and
chemosensitization activity in mouse xenograft models in comparison to
unmethylated G3139 suggesting a potent contribution of the immune stimulatory

function of G3139 to its antitumor activity (Gekeler et al. 2006). Although unre-
lated to the classical CpG effects in immune stimulation, Lai et al. reported an
association of G3139’s cytostatic activity against PCa cells in vitro with a bis-CpG
motif (CGTGC) but not with BCL2 inhibition (Lai et al. 2003).
The existence of particular oligonucleotide motifs that cause sequence-dependent
off-target effects should be considered for the design of ASOs with special view on
adequate controls. Exclusion of G-rich sequences and CpG motifs or methylation of
cytosine residues are efficient possibilities to circumvent the side-effects connected
to these sequences.
ASOs are shown to regulate the expression of several genes either independent
of their action on the target mRNA or as a consequence of that. Anderson et al.
characterized the changes in gene expression of PCa cells caused by G3139
transfection and identified an ASO-specific signature containing stress-inducible
genes, which might be mainly related to the PS-backbone (Anderson et al. 2006).
Accordingly, Stessl et al. showed an off-target signature of proteins that have been
changed by G3139-treatment but without relation to BCL2-signaling (Stessl et al.
2009). An off-target signature of differentially expressed genes has also been
reported for partially PS-modified ASOs targeted at telomerase in bladder cancer
cells (Kraemer et al. 2006). Although an oligonucleotide sequence of more than 17
nucleotides has a high probability of being unique in the human genome, ASOs can
cause off-target gene regulation by incomplete hybridization with nontargeted
mRNAs. Alternatively, secondary effects can be the reason. However, it is difficult
to differentiate between these kinds of regulations, which makes the effects hard to
predict but not necessarily toxic to cells, tissues, animals, or humans.
5 Conclusion
Despite several limitations regarding stability, delivery, and off-target effects,

specific gene silencing mediated by hybridization of relatively simple molecules
like ASOs with a selected target mRNA still shows high potential for therapeutic
applications. Currently, several promising ASO drug candidates with high potential
to reach clinical application are examined in in vivo studies and clinical trials
for treatment of cancerous and noncancerous diseases. Regarding toxicities, the
majority of adverse effects known from clinical trials are associated with backbone
modifications of ASOs but not with their antisense action on the target mRNA.
A promising way to overcome these limitations might be the formation of complexes
based on nanoparticles, which deliver ASOs as active compounds and comprise
targeting moieties like antibodies for delivery to a certain type of cells (Sp€ankuch
et al. 2008; Wang et al. 2009).
300 D. Kunze et al.
References
Aartsma-Rus A, Fokkema I, Verschuuren J et al (2009) Theoretic applicability of antisense-

mediated exon skipping for Duchenne muscular dystrophy mutations. Hum Mutat 30:293–299
Advani R, Lum BL, Fisher GA et al (2005) A phase I trial of aprinocarsen (ISIS 3521/LY900003),
an antisense inhibitor of protein kinase C-alpha administered as a 24-hour weekly infusion
schedule in patients with advanced cancer. Invest New Drugs 23:467–477
Agrawal S, Kandimalla ER (2001) Antisense and/or immunostimulatory oligonucleotide therapeutics.
Curr Cancer Drug Targets 1:197–209
Agrawal S, Kandimalla ER (2003) Modulation of toll-like receptor 9 responses through synthetic
immunostimulatory motifs of DNA. Ann NY Acad Sci 1002:30–42
Altieri DC (2003) Survivin, versatile modulation of cell division and apoptosis in cancer. Onco-
gene 22:8581–8589
Anderson EM, Miller P, Ilsley D et al (2006) Gene profiling study of G3139- and Bcl-2-targeting
siRNAs identifies a unique G3139 molecular signature. Cancer Gene Ther 13:406-414
Bates PJ, Laber DA, Miller DM et al (2009) Discovery and development of the G-rich
oligonucleotide AS1411 as a novel treatment for cancer. Exp Mol Pathol 86:151–164
Bedikian AY, Millward M, Pehamberger H et al (2006) Bcl-2 antisense (oblimersen sodium) plus
dacarbazine in patients with advanced melanoma: the Oblimersen Melanoma Study Group.
J Clin Oncol 24:4738–4745
Bogdahn U, Schneider T, Oliushine V et al (2009) Randomized, active-controlled phase IIb study
with trabedersen (AP 12009) in recurrent or refractory high-grade glioma patients: Basis for
phase III endpoints. J Clin Oncol 27:2037
Call JA, Eckhardt SG, Camidge DR (2008) Targeted manipulation of apoptosis in cancer treatment.
Lancet Oncol 9:1002–1011
Cao C, Shinohara ET, Li H et al (2005) Clusterin as a therapeutic target for radiation sensitization
in a lung cancer model. Int J Radiat Oncol Biol Phys 63:1228–1236
Chanan-Khan AA, Niesvizky R, Hohl RJ et al (2009) Phase III randomised study of dexamethasone
with or without oblimersen sodium for patients with advanced multiple myeloma. Leuk
Lymphoma 50:559–565
Chi KN, Eisenhauer E, Fazli L et al (2005) A phase I pharmacokinetic and pharmacodynamic
study of OGX-011, a 20 -methoxyethyl antisense oligonucleotide to clusterin, in patients with
localized prostate cancer. J Natl Cancer Inst 97:1287–1296
Chi KN, Zoubeidi A, Gleave ME (2008) Custirsen (OGX-011): a second-generation antisense
inhibitor of clusterin for the treatment of cancer. Expert Opin Investig Drugs 17:1955–1962
Chi KN, Hotte SJ, Yu E et al (2009) Mature results of a randomized phase II study of OGX-011
in combination with docetaxel/prednisone versus docetaxel/prednisone in patients with
metastatic castration-resistant prostate cancer. J Clin Oncol 27:5012
Cooper CL, Davis HL, Angel JB et al (2005) CPG 7909 adjuvant improves hepatitis B virus
vaccine seroprotection in antiretroviral-treated HIV-infected adults. AIDS 19:1473–1479
Crooke ST (2004a) Antisense strategies. Curr Mol Med 4:465–487
Crooke ST (2004b) Progress in antisense technology. Annu Rev Med 55:61–95
Dapić V, Abdomerović V, Marrington R et al (2003) Biophysical and biological properties of
quadruplex oligodeoxyribonucleotides. Nucleic Acids Res 31:2097–2107
Dean E, Jodrell D, Connolly K et al (2009) Phase I trial of AEG35156 administered as a 7-day and
3-day continuous intravenous infusion in patients with advanced refractory cancer. Clin Oncol
27:1660–1666
Elmén J, Lindow M, Silahtaroglu A et al (2007) Antagonism of microRNA-122 in mice by
systemically administered LNA-antimiR leads to up-regulation of a large set of predicted
target mRNAs in the liver. Nucleic Acids Res 36:1153–1162
Elmén J, Lindow M, Sch€ utz S et al (2008) LNA-mediated microRNA silencing in non-human
Fisker N, Westergaard M, Hansen HF et al (2007) Survivin mRNA antagonists using locked

nucleic acid, potential for molecular cancer therapy. Nucleosides Nucleotides Nucleic Acids
26:1427–1430
Gauvreau GM, Boulet LP, Cockcroft DW et al (2008) Antisense therapy against CCR3 and the
common beta chain attenuates allergen-induced eosinophilic responses. Am J Respir Crit Care
Med 177:952–958
Gekeler V, Gimmnich P, Hofmann HP et al (2006) G3139 and other CpG-containing immunos-
timulatory phosphorothioate oligodeoxynucleotides are potent suppressors of the growth of
human tumor xenografts in nude mice. Oligonucleotides 16:83–93
Goel S, Desai K, Bulgaru A et al (2003) A safety study of a mixed-backbone oligonucleotide
(GEM231) targeting the type I regulatory subunit alpha of protein kinase A using a continuous
infusion schedule in patients with refractory solid tumors. Clin Cancer Res 9:4069–4076
Guimond A, Viau E, Aubé P et al (2008) Advantageous toxicity profile of inhaled antisense
oligonucleotides following chronic dosing in non-human primates. Pulm Pharmacol Ther
21:845–854
Hansen JB, Fisker N, Westergaard M et al (2008) SPC3042: a proapoptotic survivin inhibitor. Mol
Cancer Ther 7:2736–2745
Hartmann G, Weeratna RD, Ballas ZK et al (2000) Delineation of a CpG phosphorothioate
oligodeoxynucleotide for activating primate immune responses in vitro and in vivo. J Immunol
164:1617–1624
Henry SP, Giclas PC, Leeds J et al (1997) Activation of the alternative pathway of complement by
a phosphorothioate oligonucleotide: potential mechanism of action. J Pharmacol Exp Ther
281:810–816
Jason TLH, Koropatnicka J, Berg RW (2004) Toxicology of antisense therapeutics. Toxicol Appl
Pharmacol 201:66–83
Jearawiriyapaisarn N, Moulton HM, Buckley B et al (2008) Sustained dystrophin expression
induced by peptide-conjugated morpholino oligomers in the muscles of mdx mice. Mol Ther
16:1624–1629
Kastelein JJ, Wedel MK, Baker BF et al (2006) Potent reduction of apolipoprotein B and low-
density lipoprotein cholesterol by short-term administration of an antisense inhibitor of
apolipoprotein B. Circulation 114:1729–1735
Klasa RJ, Gillum AM, Klem RE et al (2002) Oblimersen Bcl-2 antisense: facilitating apoptosis in
anticancer treatment. Antisense Nucleic Acid Drug Dev 12:193–213
Kraemer K, Schmidt U, Fuessel S et al (2006) Microarray analyses in bladder cancer cells:
inhibition of hTERT expression down-regulates EGFR. Int J Cancer 119:1276–1284
Kunze D, Wuttig D, Kausch I et al (2008) Antisense-mediated inhibition of survivin, hTERT and
VEGF in bladder cancer cells in vitro and in vivo. Int J Oncol 32:1049–1056
Kurreck J (2003) Antisense technologies Improvement through novel chemical modifications. Eur
J Biochem 270:1628–1644
Lacasse EC, Kandimalla ER, Winocour P et al (2005) Application of XIAP antisense to cancer and
other proliferative disorders: development of AEG35156/ GEM640. Ann NY Acad Sci
1058:215–234
Lai JC, Benimetskaya L, Santella RM et al (2003) G3139 (oblimersen) may inhibit prostate cancer
cell growth in a partially bis-CpG-dependent non-antisense manner. Mol Cancer Ther
2:1031–1043
Leonetti C, Biroccio A, D’Angelo C et al (2007) Therapeutic integration of c-myc and bcl-2
antisense molecules with docetaxel in a preclinical model of hormone-refractory prostate
cancer. Prostate 67:1475–1485
Lin C, Papadopoulos K, Patnaik A et al (2007) Oblimersen can be administered by subcutaneous
(SC) and brief intravenous (IV) infusion: Clinical pharmacokinetics and pharmacodynamics
(PK/PD) in patients with advanced cancer. J Clin Oncol 25:14083
302 D. Kunze et al.
Marchand C, Pourquier P, Laco GS et al (2002) Interaction of human nuclear topoisomerase I with

guanosine quartet-forming and guanosine-rich single-stranded DNA and RNA oligonucleotides.
J Biol Chem 277:8906–8911
Marcucci G, Moser B, Blum W et al (2007) A phase III randomized trial of intensive induction and
consolidation chemotherapy þ/ oblimersen, a pro-apoptotic Bcl-2 antisense oligonucleotide
in untreated acute myeloid leukemia patients >60 years old. J Clin Oncol 25:7012
Miyake H, Chi KN, Gleave ME (2000) Antisense TRPM-2 oligodeoxynucleotides chemosensitize
human androgen-independent PC-3 prostate cancer cells both in vitro and in vivo. Clin Cancer
Res 6:1655–1663
Monia BP, Lesnik EA, Gonzalez C et al (1993) Evaluation of 20 -modified oligonucleotides
containing 20 -deoxy gaps as antisense inhibitors of gene expression. J Biol Chem 268:
14514–14522
Niepmann M (2009) Activation of hepatitis C virus translation by a liver-specific microRNA. Cell
Cycle 8:1473–1477
O’Brien S, Moore JO, Boyd TE et al (2007) Randomized phase III trial of fludarabine plus
cyclophosphamide with or without oblimersen sodium (Bcl-2 antisense) in patients with
relapsed or refractory chronic lymphocytic leukemia. J Clin Oncol 25:1114–1120
Oettle H, Hilbig A, Seufferlein T et al (2009) Interim results of the phase I/II study of trabedersen
(AP 12009) in patients with pancreatic carcinoma, malignant melanoma, or colorectal carcinoma.
J Clin Oncol 27:4619
Reed JC (1995) Regulation of apoptosis by bcl-2 family proteins and its role in cancer and
chemoresistance. Curr Opin Oncol 7:541–546
Rubenstein M, Tsui P, Guinan P (2009) Treatment of prostate and breast tumors employing mono-
and bi-specific antisense oligonucleotides targeting apoptosis inhibitory proteins clusterin and
bcl-2. Med Oncol. doi:10.1007/s12032-009-9254-4
Ryan BM, O’Donovan N, Duffy MJ (2009) Survivin: a new target for anti-cancer therapy. Cancer
Treat Rev. doi:10.1016/j.ctrv.2009.05.003
Schimmer AD, Dalili S, Batey RA et al (2006) Targeting XIAP for the treatment of malignancy.
Cell Death Differ 13:179–188
Schlingensiepen KH, Fischer-Blass B, Schmaus S et al (2008) Antisense therapeutics for tumor
treatment: the TGF-beta2 inhibitor AP 12009 in clinical development against malignant
tumors. Recent Results Cancer Res 177:137–50
Sp€ankuch B, Steinhauser I, Wartlick H et al (2008) Downregulation of Plk1 expression by receptor-
mediated uptake of antisense oligonucleotide-loaded nanoparticles. Neoplasia 10:223–234
Stahel RA, Zangemeister-Wittke U (2003) Antisense oligonucleotides for cancer therapy-an
overview. Lung Cancer 41(Suppl 1):S81–88
Stessl M, Marchetti-Deschmann M, Winkler J et al (2009) A proteomic study reveals unspecific
apoptosis induction and reduction of glycolytic enzymes by the phosphorothioate antisense
oligonucleotide oblimersen in human melanoma cells. J Proteomics. doi:10.1016/j.jprot.
2009.06.001
Talbot D, Davies J, Callies S et al (2008) First human dose study evaluating safety and pharmaco-
kinetics of LY2181308, an antisense oligonucleotide designed to inhibit survivin. J Clin Oncol
26:3518
Talbot D, Davies J, Olsen A et al (2009) Pharmacodynamic (PD) evaluation of LY2181308 in
patients with metastatic malignancies. J Clin Oncol 27:3507
Tillman LG, Geary RS, Hardee GE (2007) Oral delivery of antisense oligonucleotides in man.
J Pharm Sci 97:225–236
Tilly H, Coiffier B, Michallet AS et al (2007) Phase I/II study of SPC2996, an RNA antagonist of
Bcl-2, in patients with advanced chronic lymphocytic leukaemia (CLL). J Clin Oncol 25:7036
van Deutekom JC, Janson AA, Ginjaar IB et al (2007) Local dystrophin restoration with antisense
oligonucleotide PRO051. N Engl J Med 357:2677–2686
Vollmer J, Krieg AM (2009) Immunotherapeutic applications of CpG oligodeoxynucleotide TLR9
agonists. Adv Drug Deliv Rev 61:195–204
Wang Y, Liu X, Chen L et al (2009) Tumor delivery of antisense oligomer using trastuzumab within
a streptavidin nanoparticle. Eur J Nucl Med Mol Imaging. doi:10.1007/s00259-009-1201-2
Warfield KL, Swenson DL, Olinger GG et al (2006) Gene-specific countermeasures against Ebola
virus based on antisense phosphorodiamidate morpholino oligomers. PLoS Pathog 2:e1
Wasan EK, Waterhouse D, Sivak O et al (2002) Plasma protein binding, lipoprotein distribution
and uptake of free and lipid-associated BCL-2 antisense oligodeoxynucleotides (G3139) in
human melanoma cells. Int J Pharm 241:57–64
Wiemer EA (2007) The role of microRNAs in cancer: no small matter. Eur J Cancer
43:1529–1544
Yokota T, Lu QL, Partridge T et al (2009) Efficacy of systemic morpholino exon-skipping in
duchenne dystrophy dogs. Ann Neurol 65:667–676
Yu RZ, Kim TW, Hong A et al (2007) Cross-species pharmacokinetic comparison from mouse
to man of a second-generation antisense oligonucleotide, ISIS 301012, targeting human
apolipoprotein B-100. Drug Metab Dispos 35:460–468
Yu RZ, Geary RS, Flaim JD et al (2009) Lack of pharmacokinetic interaction of mipomersen
sodium (ISIS 301012), a 20 -O-methoxyethyl modified antisense oligonucleotide targeting
apolipoprotein B-100 messenger RNA, with simvastatin and ezetimibe. Clin Pharmacokinet
48:39–50
Zellweger T, Miyake H, Cooper S (2001) Antitumor activity of antisense clusterin oligonucleotides
is improved in vitro and in vivo by incorporation of 20 -O-(2-methoxy)ethyl chemistry.
J Pharmacol Exp Ther 298:934–940
Zellweger T, Chi K, Miyake H et al (2002) Enhanced radiation sensitivity in prostate cancer by
inhibition of the cell survival protein clusterin. Clin Cancer Res 8:3276–3284
Zhang Q, Zhou W, Kundu S et al (2006) The leader sequence triggers and enhances several
functions of clusterin and is instrumental in the progression of human prostate cancer in vivo
and in vitro. BJU Int 98:452–460
What can the New Hammerhead Ribozyme
Structures Teach us About Design?
William G. Scott
Contents
1 Introduction to the Hammerhead Ribozyme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
1.1 The Genomic Ribozymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
1.2 What is a Hammerhead Ribozyme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
1.3 Minimal and Full-Length Hammerhead Ribozymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
1.4 Expanding Biological Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
2 Hammerhead Ribozyme Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
2.1 Three-Dimensional Structure of Minimal Hammerhead Ribozymes . . . . . . . . . . . . . . . 312
2.2 Three-Dimensional Structures of Full-Length Hammerhead Ribozymes . . . . . . . . . . . 313
3 Structure and Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
3.1 Acid–Base Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
3.2 Metal Ions? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
3.3 Substrate Binding and Specificity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
4 Hammerhead Structure, Function, and Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
4.1 Minimal Hammerheads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
4.2 Full-Length Hammerheads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
Abstract The hammerhead ribozyme is a small, self-cleaving genomic ribozyme

whose substrate-targeting properties are quite flexible. It catalyzes a phosphodiester
backbone cleavage reaction that can be exploited for antisense-type applications in
which it is desirable to cleave the target RNA. To better understand the require-
ments for rational hammerhead ribozyme design, the natural history, secondary and
tertiary structures, and reaction mechanism are reviewed in detail. Specifically,
significant advances in our understanding of how the hammerhead ribozyme works
W.G. Scott
Department of Chemistry and Biochemistry and The Center for the Molecular Biology of RNA,
University of California at Santa Cruz, 228 Sinsheimer Laboratories, Santa Cruz, CA 95064, USA
e-mail: wgscott@chemistry.ucsc.edu
306 W.G. Scott
have taken place since 2003, rendering previous assumptions about therapeutic
hammerhead ribozyme design largely obsolete. The requirement for a tertiary
contact between Stems I and II to be present in order to achieve a highly active
ribozyme in vivo is described, and design requirements that enable straightforward
incorporation of the tertiary contact are explicitly described. This analysis is only
possible with crystal structures of two classes of full-length natural hammerhead
ribozymes that became available in 2006 and 2008.
Keywords Ribozyme RNA Ribozyme gene regulation Ribozyme mechanism

Ribozyme structure Ribozyme design
1 Introduction to the Hammerhead Ribozyme
Prior to the 1980s, all enzymes were thought to be proteins. RNA was thought to
play a mostly subservient role in cellular biochemistry. tRNAs were merely
adapter molecules that were employed by the ribosome translational apparatus
to read the genomic message, immortalized in DNA, from an intermediary mRNA
transcript, and translate it into the protein sequence corresponding to the DNA
sequence. The ribosome itself was recognized to be an RNA–protein complex, but
conventional wisdom suggested that the ribosomal rRNA was merely scaffolding
that enabled the required collection of ribosomal proteins to assemble. RNA
viruses were considered a rare exception to the Central Dogma of Molecular
Biology, and genetic regulation via RNA interference mechanisms would remain
unimagined for decades. Every step of each of a myriad of biochemical reactions
that comprised the complex entangled web of metabolic pathways was catalyzed
by an enzyme, as were DNA replication and transcription. These enzymes were
always proteins.
The discovery that RNA, like proteins, also can have catalytic activity was
therefore a complete surprise. The Group I intron was shown to have self-splicing
activity in the absence of protein cofactors (Kruger et al. 1982), and the RNA
subunit of the RNA–protein complex enzyme RNase P was shown to be the
catalytic subunit of this precursor-tRNA processing enzyme (Guerrier-Takada
et al. 1983). Both of these ribozymes were comprised of RNA sequences that
were several hundred nucleotides in length and were believed (correctly) to have
rather complex secondary and tertiary structures and catalytic mechanisms. The
third ribozyme to be discovered was rather more simple and compact; it was the
hammerhead self-cleaving motif found in the genome of the satellite RNA of
tobacco ringspot virus (Prody et al. 1986). Subsequently, several other self-splicing
and self-cleaving RNAs have been discovered (Fedor 2009), and the most profound
discovery, both in terms of cellular biochemistry and evolutionary biology, was the
realization that the peptidyl transferase activity within the ribosome is a ribozyme
(Noller et al. 1992; Steitz and Moore 2003).
What can the New Hammerhead Ribozyme Structures Teach us About Design? 307
1.1 The Genomic Ribozymes
The hammerhead self-cleaving RNA was the first of several self-cleaving RNAs, or
ribozymes, to be found in the context of RNA genomes (Cochrane and Strobel
2008). The hammerhead motif, first discovered in the satellite RNA of tobacco
ringspot virus (Prody et al. 1986), has been found in a number of other plant satellite
virus RNAs, viroid RNAs, and related genomic elements. In every case, the ham-
merhead RNA is involved in the rolling circle replicative mechanism of RNA
genome replication (Fig. 1). In addition, several other self-cleaving ribozymes
with different sequences have subsequently been discovered, and all catalyze the
same chemical reaction in the same sort of biological context. This self-cleavage
reaction is not a hydrolysis reaction but rather a phosphodiester isomerization
reaction, wherein nucleophilic attack of the 20 -OH upon the adjacent phosphate
results in backbone cleavage, leaving 20 ,30 -cyclic phosphate and 50 -OH termini
(Fig. 2).
1.2 What is a Hammerhead Ribozyme
The hammerhead RNA sequence within satellite RNA genomes occurs at the
interface of two monomeric segments of a linear concatamer following rolling-
circle replication (Fig. 1). Although it is, in that context, a single self-cleaving
strand of RNA that is capable of catalyzing only a single, albeit highly specific,
cleavage reaction, the hammerhead RNA can be artificially engineered to create a
true multiple-turnover ribozyme simply by separating the molecule into discrete
enzyme and substrate strands. The latter constructs are typically studied in vitro and
also correspond to hammerhead ribozyme sequences that have been used for
targeting other RNAs. Minimal hammerhead ribozymes have typical Km values
of 10 mm, and turnover rates of about 1 substrate molecule/minute, whereas full-
length hammerhead ribozymes have a similar Km but may be 1,000-fold faster.
1.3 Minimal and Full-Length Hammerhead Ribozymes
Soon after the discovery of the hammerhead self-cleaving motif, the minimal
sequence required for self-cleavage activity was identified (Uhlenbeck 1987; Haseloff
and Gerlach 1989). The minimal hammerhead sequence consists of a central core
region of 13 mostly invariant nucleotides flanked by three A-form Watson–Crick
base-paired helical sequences whose detailed sequence is comparatively less impor-
tant (Fig. 3a). The highly conserved central region for the most part is incapable of
forming canonical Watson–Crick base-pairs and was identified as likely giving rise to
a tertiary structure that enabled the RNA to possess catalytic activity.
308 W.G. Scott
Self-cleavage reactions
N-N
Minimal N-N Phosphodiester bond
Hammerhead N-N isomerization
self-cleaving N-N
RNA sequence C-G
A-U
A
A C
NNNNNN G NNNN
NNNNNN A NNNN
C
G U
U AG
Fig. 1 Rolling circle replication. A single-stranded, covalently-closed circular RNA genome is

replicated by the host cell’s RNA polymerase. The polymerase copies the template (red) proces-
sively, creating a long linear complementary concatomeric copy (blue) that must then cleave itself
into linear monomeric fragments that can then recircularize to form single-stranded templates for
the second half of the replicative process. The cleavage sites are autolytic in the absence of protein,
and correspond to the minimal hammerhead sequence shown. The cleavage reaction is a readily
reversible phosphodiester isomerization reaction, which permits ligation into monomeric circles to
take place subsequent to the self-cleavage processing reaction
Minimal hammerhead ribozymes received an immense amount of attention in

terms of biochemical and biophysical characterization. Every one of the functional
groups on each of the conserved nucleotides has been modified to dissect its
particular contribution to catalysis, often with conflicting results (McKay 1996;
Wedekind and McKay 1998, Blount and Uhlenbeck 2005), and several crystal
structures (Pley et al. 1994; Scott et al. 1995, 1996; Murray et al. 1998a, b, 2000;
O O
G12
O O
N N O G8 N
NH2 N O
NH2
N G H+
N O NH2 N G NH2
H+ N O
H H
–O N O N
O O– N H -O N
H
O O P O O O P O
O
O O O
O -O O
OH OH
N N
O O N N O O
N1.1
H2N
O H2N O
C17
Enzyme-Substrate Complex Transition-State
O
N O
N NH2
N G NH2
HO
N
H
O -O H N
O N
P O
O O
O O
O
OH
N N O O
H2N O
Enzyme-Product Complex
Fig. 2 The chemical mechanism of hammerhead ribozyme self-cleavage. The 20 -H of C17 is

abstracted by a base (a transiently deprotonated G12), and the nucleophilic 20 -O of C17 initiates
attack upon the adjacent phosphate of nucleotide 1.1. A proton is supplied to the 50 -O leaving
group, presumably supplied from the ribose of G8, and the cleavage reaction is completed,
generating 20 ,30 -cyclic phosphate and 50 -OH termini, as shown. The transition-state is required
to be in an in-line conformation, as shown
Martick and Scott 2006; Chi et al. 2008), the first of any ribozyme, have been
determined. The crystal structures were only capable of reconciling a subset of
the biochemical experiments designed to probe the catalytic mechanism, and
considerable discord plagued the hammerhead ribozyme biochemical community
(Blount and Uhlenbeck 2005).
All ribozymes, including the hammerhead ribozyme, were originally believed to
be metalloenzymes (Pyle 1993), requiring an obligate Mg2+ for catalysis (Dahm
and Uhlenbeck 1991; Dahm et al. 1993; Peracchi et al. 1997). It has subsequently
been revealed, however, that the hammerhead, in addition to other small self-
cleaving ribozymes, does not strictly require divalent cations for catalysis. Instead,
if a sufficiently high concentration of even nonmetallic monovalent salt is present,
permitting the RNA to fold correctly, it will remain catalytically active, even in a
310 W.G. Scott
Fig. 3 A schematic secondary structure of (a) the minimal and (b) the full-length hammerhead
ribozyme. The conserved residues in the catalytic core are shown explicitly in each case, and the
cleavage site is indicated with a red arrow. The tertiary contact in (b) is indicated in the grey
portion of the schematic diagram. This figure was kindly supplied by Christian Hammann
high concentration of EDTA (Murray et al. 1998a, b). Hence it appeared that the
RNA itself, rather than functioning as a passive scaffold to bind metal ions, instead
must be an active participant in its own chemical catalysis (Scott 1999). This
renewed focus upon the RNA structure itself. However, the crystal structures of
the minimal hammerhead could not be reconciled with this conclusion; none of the
invariant residues were positioned in a way that made their role in catalysis at all
obvious, and the substrate itself was not bound to the enzyme in a way that would
permit the known required in-line attack geometry to be stabilized, as one would
expect from an enzyme (McKay 1996; Blount and Uhlenbeck 2005).
In 2003, two papers appeared (De la Peña et al. 2003; Khvorova et al. 2003) that
had essentially the same conclusion: the minimal hammerhead construct lacked a
tertiary contact between helices I and II that had a rather profound effect upon
hammerhead ribozyme catalysis, despite being distant from the cleavage site
(Fig. 3b). This contact appeared to have little if any sequence conservation between
the large number of natural hammerhead ribozyme sequences that had been identi-
fied to date and had thus escaped notice. However, when the tertiary contact
sequences were included, these natural, full-length hammerheads were observed
to be up to 1,000-fold more active than their minimal counterparts (Khvorova et al.
2003; Canny et al. 2004). Clearly, the tertiary contact imparted some change at the
active site that stimulated catalytic activity.
1.4 Expanding Biological Context
Although the hammerhead ribozyme was originally discovered in RNA virus-like

genomes, it has since been discovered to occur in a few other contexts (Ferbeyre
Fig. 4 The hammerhead ribozyme embedded within the 30 -UTR of the clec2d mRNA transcript,
immediately downstream from the stop codon. The “enzyme” part of the strand is highlighted in
blue, and the “substrate” is highlighted in orange, with the cleavage site indicated
et al. 2000). Highly repetitive DNA sequences in a Schistosome trematode parasite

(Ferbeyre et al. 1998) and in a newt genome (Forster et al. 1988; Luzi et al. 1997),
when transcribed, give rise to RNA satellites that contain hammerhead ribozyme
sequences. More recently, active hammerhead ribozyme sequences have been
discovered in the 30 -untranslated regions of mature mRNAs in a variety of
mammals (Martick et al. 2008a, b), and these are thought to control export and
translation via a riboswitch mechanism that is currently under investigation (Fig. 4).
Additional reports of hammerhead ribozyme sequences in bacteria and eukaryotes
are now emerging. Hammerhead ribozymes may thus turn out to be far more
ubiquitous than originally thought and may play a significant role in RNA-based
regulation of gene expression.
2 Hammerhead Ribozyme Structures
The first structures of an RNA appeared in 1974 with the publication of the yeast
tRNAPhe crystal structures (Robertus et al. 1974; Kim et al. 1974). These structures
revealed that RNA possesses the propensity to fold into comparatively compact
globular protein-like three-dimensional structures, and it also illustrated the impor-
tance of tertiary contacts in stabilizing complex RNA backbone folds (Klug et al.
1974). Another 20 years passed before another complex RNA structure emerged. In
1994, the first structure of a ribozyme was published by McKay and coworkers
(Pley et al. 1994); it consisted of a minimal hammerhead enzyme strand hybridized
with a DNA substrate analog. A second structure of a minimal hammerhead, this
time with an all-RNA substrate (Scott et al. 1995), appeared shortly thereafter,
corroborating the initial structural work.
312 W.G. Scott
2.1 Three-Dimensional Structure of Minimal Hammerhead

Ribozymes
Both minimal hammerhead ribozyme structures revealed a three-stranded junction

in which the cleavage-site nucleotide, C17, is surrounded by invariant residues that
formed a structure analogous to the uridine turn in tRNA (Pley et al. 1994). The
remaining invariant residues augmented Stem II, permitting it to stack upon Stem
III coaxially, at the junction interface (Fig. 5).
The crystal structures of the minimal hammerhead ribozyme frustratingly
created many more questions than compelling explanations for RNA catalysis.
The 12 years subsequent to the publication of these structures saw only increasing
discord; the crystal structure analyses seemed hopelessly irreconcilable with a grow-
ing corpus of biochemical evidence (Blount and Uhlenbeck 2005). Meanwhile,
crystal structures for many of the other ribozymes, including the Group I intron,
the hairpin, HDV, and RNase P, appeared.
Despite observations of hammerhead ribozyme catalysis in a crystal in which the
lattice packing contacts by necessity confined the global positions of the distal
termini of all three flanking helical stems of the minimal hammerhead (Scott et al.
1996; Murray et al. 2002; Scott 2002), many biochemical experiments designed to
probe transition-state interactions and the chemistry of catalysis appeared to be
irreconcilable with the crystal structures. For example, the invariant core residues
G5, G8, G12, and C3 in the minimal hammerhead ribozyme were each observed to
be so fragile that changing even a single exocyclic functional group on any one of
these nucleotides results in abolition of catalytic activity; yet few of these appeared
to form hydrogen bonds involving the Watson–Crick faces of these nucleotides
(McKay 1996). A particularly striking and only recently observed example
Fig. 5 A backbone diagram

of a minimal hammerhead
ribozyme, in which a shorter
strand is the enzyme strand,
shown in blue, and the longer
strand is the substrate strand,
shown in magenta
consisted of G8 and G12, which were identified as possible participants in acid–-

base catalysis (Han and Burke 2005). After it was demonstrated that the hammer-
head ribozyme does not require divalent metal ions for catalysis (Murray et al.
1998a, b), it gradually became apparent that the RNA itself, rather than passively
bound divalent metal ions, must play a direct chemical role in any acid–base
chemistry within the hammerhead active site. It was, however, completely unclear
how G12 and G8 could accomplish this, given the original structures of the minimal
hammerhead ribozyme. In addition, the attacking nucleophile in the original struc-
tures, i.e., the 20 -OH of C17, was not in a position amenable to in-line attack upon
the adjacent scissile phosphate. Perhaps most worrisome were experiments that
suggested that the A-9 and scissile phosphates must come within about 4 Å of one
another in the transition-state, based upon double phosphorothioate substitution and
soft metal ion rescue experiments; the distance between these phosphates in the
crystal structure was about 18 Å, with no clear mechanism for close approach if the
Stem II and Stem I A-form helices were treated as rigid bodies (Wang et al. 1999).
Taken together, these results appeared to suggest that a fairly large-scale confor-
mational change must have taken place to reach the transition-state within the
minimal hammerhead ribozyme structure. For these reasons, the two sets of experi-
ments (biochemical vs. crystallographic) appeared not only to be at odds but to
be completely and hopelessly irreconcilable, generating a substantial amount of
discord in the field (Blount and Uhlenbeck 2005). No compelling evidence for
dismissing either set of experimental results was ever made successfully (although
some claims to the contrary were made in favor of each).
2.2 Three-Dimensional Structures of Full-Length

Hammerhead Ribozymes
A new crystal structure of the full-length hammerhead ribozyme emerged in 2006,

20 years after the hammerhead’s discovery (Martick and Scott 2006). This structure
includes a set of distal tertiary contacts whose importance was largely unrecognized
until 2003, but whose incorporation increases the catalytic prowess by a factor of
1,000. The new crystal structure reveals that this remarkable rate enhancement is a
direct consequence of localized yet dramatic active site conformational changes
that are stabilized by a comparatively distant set of tertiary interactions. The new
structure appears to reconcile twenty years of discord while offering some new
insights into RNA structure and catalysis (Nelson and Uhlenbeck 2006).
2.2.1 Schistosomal Hammerhead Structure
The resolution of this vexing conundrum finally came in 2006 with a 2.2 Å resolu-
tion crystal structure of the full-length hammerhead ribozyme from Schistosoma
314 W.G. Scott
Fig. 6 A backbone diagram

of a full-length Schistosomal
hammerhead ribozyme, in
which the enzyme strand is
shown in blue, and the
substrate strand is shown in
magenta
mansoni (Fig. 6). C17 is now positioned for in-line attack, and the invariant residues
C3, G5, G8, and G12 all appear involved in vital interactions relevant to catalysis.
Moreover, the A9 and scissile phosphates are observed to be 4.3 Å apart, which is
consistent with the idea that, when modified, these phosphates could bind a single
thiophilic metal ion. The structure also reveals how two invariant residues, G-12 and
G-8, are positioned within the active site – consistent with their previously proposed
role in acid–base catalysis. G12 is within hydrogen bonding distance to the 20 -O of
C17, the nucleophile in the cleavage reaction, and the ribose of G8 hydrogen bonds
to the leaving group 50 -O, while the nucleotide base of G8 forms a Watson–Crick
pair with the invariant C3. This arrangement suggests that G12 is the general base in
the cleavage reaction, and that the G8 ribose may function as the general acid
(Fig. 7). The crystal structure of the full-length hammerhead ribozyme thus clearly
addresses the major concerns that appeared irreconcilable with the previous crystal
structures (Nelson and Uhlenbeck 2006, 2008).
In addition to the rearranged cleavage site, one of the most prominent features of
the full-length hammerhead ribozyme structure is the Stem II loop/Stem I bulge
interaction that appears to induce the structural organization of the catalytic core.
The loop/bulge interaction is composed of an intricate network of interhelical
Fig. 7 The disposition of

active-site nucleotides in the
context of a proposed
transition-state structure
noncanonical base pairs and stacks interdigitating Stem II loop into Stem I, kinking
Stem I in such a way as to coaxially align its distal helix on top of the Stem II–Stem
III coaxial arm. The tertiary contacts between the loop and bulge regions induce
structural changes affecting the catalytic core, specifically via the relative under-
winding of Stem I. This interaction imparts a severe bend in the distal part of the
Stem I helix and a pronounced kink in the backbone of the substrate strand at the
cleavage site. These distortions appear to accommodate G-8 and U-7 in the catalytic
pocket and in turn stabilize the rearrangement of the augmented Stem II helix that
enables G-8 to form the Watson–Crick base pair with C-3 in the catalytic pocket.
Concurrently, an overwinding or right-handed twist of Stem II positions the con-
served G-12, A-13, and A-14 precisely against the catalytic-site C-17, helping to
lock the latter in a catalytically active conformation in which C-17 is oriented for
in-line attack (Martick and Scott 2006).
2.2.2 Satellite Viral Hammerhead Ribozyme Structure
The Satellite RNAs all possess a different type of tertiary contact compared to the
Schistosomal hammerhead. In 2008, crystal structures of an unmodified hammer-
head ribozyme derived from the satellite RNA of tobacco ringspot virus (sTRSV)
was published (Chi et al. 2008), permitting comparison of the two types of ham-
merhead (Scott et al. 2009). Briefly, the active site is nearly identical in both types
of hammerhead ribozyme, but the tertiary contacts, though imparting the same net
structural effect upon the active site, are distinctly different in the two cases.
The two classes of tertiary contacts are shown as secondary structural represen-
tations that reflect the tertiary structures in Fig. 8. The only structural feature the
two types of tertiary contacts have in common is the presence of an AU Hoogsteen
316 W.G. Scott
Fig. 8 The secondary structures of two classes of full-length hammerhead ribozymes, with those
represented by satellite virus hammerheads shown in a, and the Schistosoma-like hammerheads,
shown in b. The tertiary contacts are highlighted in light green. Little sequence homology in the
tertiary contact region in apparent
pair. Remarkably, this apparently conserved interaction escaped detection, even by

Eric Westhof’s well-trained eyes. The 2003 paper by Khvorova, Westhof, and
colleagues (Khvorova et al. 2003) compares 13 natural hammerhead ribozyme
sequences, all of the sTRSV class, in an attempt to deduce conserved tertiary
interactions. All of these sequences possess GNRA-like tetraloops capping
Stem II, where the final A in the sequence is always present. This A makes a
Hoogsteen pair with U1.7 in the substrate strand of Stem I. (It also makes
a Watson–Crick pair with another U in the 30 -region of the loop capping Stem I
in the sTRSV hammerhead, but this U appears not to be conserved even within the
second class of hammerheads.) The U that is involved in the Hoogsteen
pair interaction is present in 10 of the 13 sequences analyzed by Westhof and
colleagues; the other three examples have a C at position 1.7. It is noteworthy that
C will participate in the same Hoogsteen pairing interaction if N3 is protonated. If
this is indeed the case, it is possible that the C1.7 hammerhead sequences will show
decreasing activity at more basic pH values, a property that might be exploitable for
control of activity in designer ribozymes.
The structure of the GNRA tetraloop capping Stem II in the sTRSV hammerhead
is unusual as well. The GNRA tetraloop, where the 50 nucleotide is always a G, the
second can be any nucleotide (N), the third is a purine (R), and the fourth is always
an A, is a canonically stable RNA secondary structural motif. The G pairs with the
A via a single hydrogen bond, and the N, R, and A all stack upon one another on the
30 -side of the loop. However, in the context of the sTRSV hammerhead tertiary
contact, the structure of the GNRA tetraloop rearranges so that the above-noted
interactions disappear and the A rearranges to form the Hoogsteen base pair with
U1.7 from Stemloop I.
3 Structure and Mechanism
The cleavage reaction is a phosphodiester isomerization reaction that is initiated by

abstraction of the 20 -hydroxyl proton from its 20 -oxygen, which then becomes the
attacking nucleophile in an “in-line” or SN2(P)-like reaction, although it is not
known whether this proton is removed before or during the chemical step of the
hammerhead cleavage reaction. (The cleavage reaction is technically not bimolecular,
but behaves in the same way a genuine SN2 reaction does; it undergoes inversion of
configuration subsequent to forming an associative transition-state consisting of a
pentacoordinated oxyphosphorane). The attacking and leaving group oxygens will
both occupy the two axial positions in the trigonal bipyramidal transition-state
structure as is required for an SN2-like reaction mechanism.
The 50 -product, as a result of this cleavage reaction mechanism, possesses a
2 ,3 -cyclic phosphate terminus, and the 30 -product possesses a 50 -OH terminus, as
0 0
with nonenzymatic alkaline cleavage of RNA. The reaction is therefore, in principle,

reversible, as the scissile phosphate remains a phosphodiester, and may thus act as
a substrate for hammerhead RNA-mediated ligation without a requirement for ATP
or a similar exogenous energy source. The hammerhead ribozyme-catalyzed reac-
tion, unlike the formally identical nonenzymatic alkaline cleavage of RNA, is a
highly sequence-specific cleavage reaction with a typical turnover rate of approxi-
mately one molecule of substrate per molecule of enzyme per minute at pH 7.5
in 10 mM Mg2+ (the so-called “standard reaction conditions” for the minimal
hammerhead RNA sequence), depending upon the sequence of the particular
hammerhead ribozyme construct measured. This represents an approximately
10,000-fold rate enhancement over the nonenzymatic cleavage of RNA (Stage-
Zimmermann and Uhlenbeck 1998).
3.1 Acid–Base Catalysis
Based upon the arrangement of invariant nucleotides in the full-length hammerhead

active site, as well as the solvent structure in a combined crystallographic and
molecular dynamics investigation, it appears that a specifically bound water molecule
318 W.G. Scott
Fig. 9 The satellite RNA

of tobacco ringspot virus
hammerhead ribozyme. The
crystal structure is of a single
strand of the RNA, which
cleaves slowly due to a G12A
mutation that lowers the pKa
of the general base
(Fig. 9, light blue) accepts a proton from G12. G12 must ionize to function as the
general base, and the proton is replaced by that from the 20 -OH of C17 (Fig. 9,
black). The original G12 proton can then be relayed directly to the 20 -OH of G8 to
replace a proton that must be donated to the 50 -O leaving group of C1.1 (black) as
the phosphodiester backbone is cleaved. This mechanism conserves the number of
protons during the phosphodiester isomerization. It is testable, in that it predicts that
altering the pKa of either the purine base at position 12 or the 20 -OH at position
8 will alter the cleavage rate without inducing gross structural perturbations. There
are also opportunities for transition-state stabilization of the accumulating negative
charges in the pentacoordinated oxyphosphorane. We suggest that either the
exocyclic amine of A9 or a divalent cation can perform this function.
3.2 Metal Ions?
The minimal hammerhead ribozyme was originally believed to be dependent upon

the presence of divalent metal ions for folding and catalysis (Dahm and Uhlenbeck
1991). However, in the presence of a high concentration of monovalent salt,
including molar concentrations of NH4+, the strict requirement for Mg2+ may be
dispensed with (Murray et al. 1998a, b). Nevertheless, the hammerhead ribozyme
appears to be reliant upon Mg2+ under in vivo conditions, and is typically assayed
in vitro in the presence of 10 mM Mg2+, whereas physiological concentrations of
Mg2+ are closer to 1 mM. The apparent Km for Mg2+ in minimal hammerheads
ranges from 10 to 100 mM (Stage-Zimmermann and Uhlenbeck 1998).
Originally, Mg2+ was believed to play a direct role in acid–base catalysis in the
hammerhead ribozyme (Dahm et al. 1993), and although as early as 1998, it was
known that the hammerhead did not require Mg2+ for catalysis, it was only with
the publication of the full-length hammerhead ribozyme structures, which
revealed RNA functional groups positioned for acid/base chemistry, that the
participation of Mg2+ in acid–base catalysis could be ruled out. An additional
potential role for Mg2+, however, is transition-state charge stabilization (Martick
et al. 2008a, b; Lee et al. 2007, 2008). For this, any high concentration of positive
charge should suffice, so the suggestion that Mg2+ or monovalent cations aid in
folding as well as transition-state stabilization appears to be the most consistent
with all of the data, and accounts for the rather high apparent Km values for
Mg2+. What is apparent is that under low ionic strength in vitro assay conditions,
the hammerhead ribozyme needs at least ten times the total physiological con-
centration of Mg2+ to cleave efficiently. Therefore, it is very unlikely that the
minimal hammerhead sequence will be able to fold and cleave efficiently in vivo,
unless folding is assisted by some compensatory mechanism (such as an associated
RNA-binding protein).
The full-length hammerhead ribozyme, even under low ionic strength in vitro
assay conditions, requires only micromolar concentrations of Mg2+ (Khvorova et al.
2003), which is far more consistent with in vivo requirements for activity. It is
likely that the requirement for 10–100 mM Mg2+ for optimal activity of the minimal
hammerhead is partially compensating for the lack of the tertiary contact that
stabilizes the active site. Hence, the full-length hammerhead, which includes a
naturally occurring tertiary contact, would be by far the most preferable starting
point for the design of an in vivo RNA cleaving reagent.
3.3 Substrate Binding and Specificity
One of the most attractive features of the minimal hammerhead ribozyme to those
hoping to design a specific RNA cleavage reagent is that almost all of the enzyme–
substrate binding specificity can be understood in terms of simple Watson–Crick
base-pairing rules. One does not need to know anything about the hammerhead
ribozyme’s tertiary structure in order to design a hammerhead ribozyme to cleave
an RNA substrate of a given sequence. The only sequence restriction that exists, in
terms of choosing a target substrate, is that a RUH nucleotide triplet be present. H is
the cleavage-site nucleotide. It is typically a C but can be any nucleotide apart from
320 W.G. Scott
G. U is uridine. R is either purine, with a preference for G. Because of the tertiary

contacts in the full-length hammerhead ribozyme, whose sequence requirements are
rather poorly understood, it is less apparent what sequence restrictions might exist for
designing a full-length hammerhead ribozyme. For this reason, use of the full-length
hammerhead ribozyme in designing cleavage reagents is often avoided in order to
circumvent such complications.
Fortunately, comparison of the two full-length hammerhead crystal structures
reveals that the only additional conserved nucleotide in the substrate tertiary contact
region is U1.7. The remainder of the Stem I contributions to the tertiary contact are
all contained within the enzyme strand.
4 Hammerhead Structure, Function, and Design
The above considerations lead quite naturally to explicit requirements for hammer-
head ribozyme design.
4.1 Minimal Hammerheads
Minimal hammerhead ribozymes have been employed for many years as antisense
RNA cleaving agents in an attempt to target pathological mRNAs and RNA viruses
(For a recent review, please see Tedeschi et al. 2009). An advantage of the minimal
hammerhead is that the only sequence restriction imposed on the target substrate is
at the cleavage site, where a sequence of the type RUH is required (R is either G or
A at position 16.2, U must be uracil at position 16.1, H can be anything except G at
position 17, the cleavage site nucleotide). The sequences of Stems I and III in a
minimal hammerhead enzyme strand, apart from these restrictions, can then be
tailored to base-pair with any target sequence. Unfortunately, their practical utility
as therapeutic agents has been limited by the apparent need for nonphysiological
concentrations of Mg2+ and their slow (1/min) turnover rates.
4.2 Full-Length Hammerheads
The much greater activity of the full-length hammerhead ribozymes makes them a
more attractive alternative as in vivo RNA cleavage agents. In addition, the discov-
ery of naturally occurring full-length hammerheads in mammalian mRNAs that
downregulate gene expression (Martick et al. 2008a, b), including some that appear
to work intermolecularly, provide a convincing argument that full-length hammer-
heads should be viable in vivo ribozyme nucleases. But because of the requirement
Fig. 10 A schematic
representation of the
Schistosomal hammerhead
ribozyme, with the enzyme
strand shown in blue, and the
substrate strand shown in
orange. The restrictions
imposed on the sequence are
such that any RNA strand
with a sequence of the form ...
NRUHNNNNNNYN... can
be targeted, although the
optimal sequence will have
C17 and U1.7. The crystal
structures reveal that a U (or
possibly C) at position 1.7 is
the only additional sequence
restriction imposed by the
full-length hammerhead
tertiary contact region
...NRUHNNNNNNYN...
between Stems I and II
GU C
16.2 16.1 17
1.1 U1.7
1.2
1.4
1.3
for the tertiary contact, whose sequence requirements before now have been rather
obscure, full-length hammerhead cleavage agents have not been pursued with vigor.
Comparison of the sTRSV and Schistosomal full-length hammerhead crystal
structures reveals that only one base-pairing interaction between enzyme and
substrate strands in the tertiary contact region is conserved. The AU Hoogsteen
pair requires U1.7 to be present (although C might be able to substitute for U at this
position; cf. Fig. 10). The remainder of the specific tertiary interactions appears in
both cases to lie entirely within the enzyme strands of Stem I and Stem-loop II.
Hence the sequence restrictions imposed by the full-length hammerhead on possi-
ble RNA targets is simply ...NRUHNNNNNNYN... where N is any nucleotide, R
can be G or A at position 16.2, U must be uracil at 16.1, H can be any residue at
position 17, the cleavage site, although C is preferred, and Y at position 1.7, (i.e.,
seven nucleotides downstream of the cleavage site), must be U or possibly C.
Examination of the structures therefore strongly suggests that an mRNA
sequence that possesses the ...NRUHNNNNNNYN... motif can be targeted and
cleaved efficiently by designing a hammerhead enzyme strand complementary to
the target sequence. If cleavage is insufficiently efficient, one can subsequently use
in vitro selection to further optimize the enzyme sequence by selecting for variants
in the tertiary contact region.
322 W.G. Scott
References
Blount KF, Uhlenbeck OC (2005) The structure–function dilemma of the hammerhead ribozyme.
Annu Rev Biophys Biomol Struct 34:415–440
Canny MD, Jucker FM, Kellogg E et al (2004) Fast cleavage kinetics of a natural hammerhead
ribozyme. J Am Chem Soc 126:10848–10849
Chi YI, Martick M, Lares M et al (2008) Capturing hammerhead ribozyme structures in action by
modulating general base catalysis. PLoS Biol 6:e234
Cochrane JC, Strobel SA (2008) Catalytic strategies of self-cleaving ribozymes. Acc Chem Res
41:1027–1035
Dahm SC, Uhlenbeck OC (1991) Role of divalent metal ions in the hammerhead RNA cleavage
reaction. Biochemistry 30:9464–9469
Dahm SC, Derrick WB, Uhlenbeck OC (1993) Evidence for the role of solvated metal hydroxide
in the hammerhead cleavage mechanism. Biochemistry 32:13040–13045
Fedor MJ (2009) Comparative enzymology and structural biology of RNA self-cleavage. Annu
Rev Biophys 38:271–299
Ferbeyre G, Bourdeau V, Pageau M et al (2000) Distribution of hammerhead and hammerhead-
like RNA motifs through the GenBank. Genome Res 10:1011–1019
Ferbeyre G, Smith JM, Cedergren R (1998) Schistosome satellite DNA encodes active hammer-
head ribozymes. Mol Cell Biol 18:3880–3888
Forster AC, Davies C, Sheldon CC et al (1988) Self-cleaving viroid and newt RNAs may only be
active as dimers. Nature 334:265–267
Guerrier-Takada C, Gardiner K, Marsh T et al (1983) The RNA moiety of ribonuclease P is the
catalytic subunit of the enzyme. Cell 35:849–857
Han J, Burke JM (2005) Model for general acid-base catalysis by the hammerhead ribozyme:
pH-activity relationships of G8 and G12 variants at the putative active site. Biochemistry
44:7864–7870
Haseloff J, Gerlach WL (1989) Sequences required for self-catalysed cleavage of the satellite
RNA of tobacco ringspot virus. Gene 82:43–52
Khvorova A, Lescoute A, Westhof E et al (2003) Sequence elements outside the hammerhead
ribozyme catalytic core enable intracellular activity. Nat Struct Biol 10:708–712
Kim SH, Suddath FL, Quigley GJ et al (1974) Three-dimensional tertiary structure of yeast
phenylalanine transfer RNA. Science 185:435–440
Klug A, Ladner J, Robertus JD (1974) The structural geometry of co-ordinated base changes in
transfer RNA. J Mol Biol 89:11–16
Kruger K, Grabowski PJ, Zaug AJ et al (1982) Self-splicing RNA: autoexcision and autocyclization
of the ribosomal RNA intervening sequence of tetrahymena. Cell 31:147–157
De la Peña M, Gago S, Flores R (2003) Peripheral regions of natural hammerhead ribozymes
greatly increase their self-cleavage activity. EMBO J 22:5561–5570
Lee TS, Silva López C, Giambasu GM et al (2008) Role of Mg2þ in hammerhead ribozyme
catalysis from molecular simulation. J Am Chem Soc 130:3053–3064
Lee TS, Silva-López C, Martick M et al (2007) Insight into the role of Mg 2 in hammerhead
ribozyme catalysis from X-ray crystallography and molecular dynamics simulation. J Chem
Theory Comput 3:325–327
Luzi E, Eckstein F, Barsacchi G (1997) The newt ribozyme is part of a riboprotein complex. Proc
Martick M, Scott WG (2006) Tertiary contacts distant from the active site prime a ribozyme for
catalysis. Cell 126:309–320
Martick M, Horan LH, Noller HF et al (2008a) A discontinuous hammerhead ribozyme embedded
in a mammalian messenger RNA. Nature 454:899–902
Martick M, Lee TS, York DM et al (2008b) Solvent structure and hammerhead ribozyme catalysis.
Chem Biol 15:332–342
McKay DB (1996) Structure and function of the hammerhead ribozyme: an unfinished story. RNA
2:395–403
Murray JB, Dunham CM, Scott WG (2002) A pH-dependent conformational change, rather than
the chemical step, appears to be rate-limiting in the hammerhead ribozyme cleavage reaction.
J Mol Biol 315:121–130
Murray JB, Seyhan AA, Walter NG et al (1998a) The hammerhead, hairpin and VS ribozymes are
catalytically proficient in monovalent cations alone. Chem Biol 5:587–595
Murray JB, Szöke H, Szöke A et al (2000) Capture and visualization of a catalytic RNA
enzyme–product complex using crystal lattice trapping and X-ray holographic reconstruction.
Mol Cell 5:279–287
Murray JB, Terwey DP, Maloney L et al (1998b) The structural basis of hammerhead ribozyme
self-cleavage. Cell 92:665–673
Nelson JA, Uhlenbeck OC (2006) When to believe what you see. Mol Cell 23:447–450
Nelson JA, Uhlenbeck OC (2008) Hammerhead redux: does the new structure fit the old biochemical
data? RNA 14:605–615
Noller HF, Hoffarth V, Zimniak L (1992) Unusual resistance of peptidyl transferase to protein
extraction procedures. Science 256:1416–1419
Peracchi A, Beigelman L, Scott EC et al (1997) Involvement of a specific metal ion in the
transition of the hammerhead ribozyme to its catalytic conformation. J Biol Chem
272:26822–26826
Pley HW, Flaherty KM, McKay DB (1994) Three-dimensional structure of a hammerhead
ribozyme. Nature 372:68–74
Prody GA, Bakos JT, Buzayan JM et al (1986) Autolytic processing of dimeric plant virus satellite
RNA. Science 231:1577–580
Pyle AM (1993) Ribozymes: a distinct class of metalloenzymes. Science 261:709–714
Robertus JD, Ladner JE, Finch JT et al (1974) Structure of yeast phenylalanine tRNA at 3
A resolution. Nature 250:546–551
Scott WG (1999) RNA structure, metal ions, and catalysis. Curr Opin Chem Biol 3:705–710
Scott WG (2002) Visualizing the structure and mechanism of a small nucleolytic ribozyme.
Methods 28:302–306
Scott WG, Finch JT, Klug A (1995) The crystal structure of an all-RNA hammerhead ribozyme:
a proposed mechanism for RNA catalytic cleavage. Cell 81:991–1002
Scott WG, Martick M, Chi YI (2009) Structure and function of regulatory RNA elements:
ribozymes that regulate gene expression. Biochim Biophys Acta 1789:634–641
Scott WG, Murray JB, Arnold JR et al (1996) Capturing the structure of a catalytic RNA
intermediate: the hammerhead ribozyme. Science 274:2065–2069
Stage-Zimmermann TK, Uhlenbeck OC (1998) Hammerhead ribozyme kinetics. RNA 4:875–889
Steitz TA, Moore PB (2003) RNA, the first macromolecular catalyst: the ribosome is a ribozyme.
Trends Biochem Sci 28:411–418
Tedeschi L, Lande C, Cecchettini A et al (2009) Hammerhead ribozymes in therapeutic target
discovery and validation. Drug Discov Today 14:776–783
Uhlenbeck OC (1987) A small catalytic oligoribonucleotide. Nature 328:596–600
Wang S, Karbstein K, Peracchi A et al (1999) Identification of the hammerhead ribozyme metal
ion binding site responsible for rescue of the deleterious effect of a cleavage site phosphor-
othioate. Biochemistry 38:14363–14378
Wedekind JE, McKay DB (1998) Crystallographic structures of the hammerhead ribozyme:
relationship to ribozyme folding and catalysis. Annu Rev Biophys Biomol Struct 27:475–502
microRNA Biogenesis and its Impact on RNA
Interference
Stefanie Grund and Sven Diederichs
Contents
1 The microRNA Biogenesis Pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
1.1 microRNA Gene Transcription . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
1.2 microRNA Editing: Small Changes Affect Many Steps . . . . . . . . . . . . . . . . . . . . . . . . . . 327
1.3 pri-miRNA Cleavage by the Microprocessor Complex . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
1.4 Nuclear Export of the microRNA Precursors by Exportin-5 . . . . . . . . . . . . . . . . . . . . . . 332
1.5 The RISC Loading Complex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
1.6 Terminal Loop Removal by Dicer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
1.7 Ago2 Jumps the Queue: Generation of the ac-pre-miRNA . . . . . . . . . . . . . . . . . . . . . . . 335
1.8 miRNA Duplex Unwinding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
1.9 Strand Selection: Who Becomes the Guide? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
1.10 Mediators of RNA Silencing: The Argonaute Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
1.11 Half-Life and Degradation of microRNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
2 Implication for RNAi Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
2.1 Potentials and Challenges of siRNAs as a Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
2.2 siRNA Versus shRNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
2.3 shRNA-miR Library: Transferring microRNA Structures to Synthetic shRNAs . . . 343
2.4 Enhancement of RNAi by microRNA Biogenesis Factors . . . . . . . . . . . . . . . . . . . . . . . . . . 344
2.5 microRNA Biogenesis in Health and Disease: Basis for RNAi Therapy . . . . . . . . . . . 346
2.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
Abstract microRNAs (miRNAs) are small, noncoding, single-stranded RNAs that

control diverse key cellular pathways at the posttranscriptional level. Their mode of
action is translational repression or degradation of target mRNAs containing
S. Grund and S. Diederichs (*)

Helmholtz-University-Group “Molecular RNA Biology & Cancer”, German Cancer Research
Center (DKFZ), Im Neuenheimer Feld 280 (B150), 69120, Heidelberg, Germany
Institute of Pathology, University of Heidelberg, Im Neuenheimer Feld 220 (B150), 69120,
Heidelberg, Germany
e-mail: s.diederichs@dkfz.de
326 S. Grund and S. Diederichs
complementary sequences. As many miRNAs act in crucial cellular pathways, their

dysregulation can result in various diseases including cancer.
Here, we summarize recent insights into the complex processing pathway gen-
erating the mature, functional miRNA. Cleavage of the primary miRNA transcripts
(pri-miRNAs) by the microprocessor complex Drosha-DGCR8/Pasha releases
60–70 nt hairpin structures, the pre-miRNAs. After export to the cytoplasm
mediated by Exportin-5, the RNase III-like Dicer completes processing in conjunc-
tion with its dsRBD partner protein (TRBP in mammals and LOQS in flies). The
small RNA duplexes are unwound, and one strand, the guide strand, is incorporated
together with Argonaute proteins into the RNA-induced silencing complex (RISC).
Multiple studies in recent years have revealed that every step of this processing
pathway can be regulated and that certain miRNAs do not follow this general
processing pathway but use a variety of other processing and regulatory options
for their maturation.
Importantly, the miRNA processing and effector proteins also provide the
essential machinery for RNA interference (RNAi). While ectopically delivered
RNA (like dsRNA, siRNA, or shRNA) functions as specificity component to
knockdown target genes, the processing and effector machinery has to be contrib-
uted by the targeted cell. Also, several miRNA processing factors can be used to
enhance RNAi. Thus, a deeper understanding of miRNA processing, regulation,
and function is an essential prerequisite to optimize experimental RNAi and enable
therapeutic RNAi approaches.
Keywords microRNA miRNA processing miRNA biogenesis Argonaute

RNAi siRNA shRNA
Abbreviations
miRNAs microRNAs
pri-miRNAs primary miRNA transcripts
Pol II RNA Polymerase II
Tudor-SN Tudor staphylococcal nuclease
bp basepairs
nt nucleotide
dsRBD double-stranded RNA binding domain
OB-fold oligonucleotide/oligosaccharide binding fold
shRNAs short hairpin RNAs
ac-pre-miRNA Ago2-cleaved pre-miRNA
microRNA Biogenesis and its Impact on RNA Interference 327
1 The microRNA Biogenesis Pathway
Since RNAi was first described in the nematode Caenorhabditis elegans (Fire et al.
1998), it has been found to be a widespread phenomenon in eukaryotic organisms.
Over the past years, hundreds of miRNA genes have been identified in animals,
plants, and viruses, making them one of the largest gene families. Several different
miRNAs can control one mRNA target cooperatively and each miRNA can bind to
many different mRNAs. Thus, more than 30% of all human genes are predicted to
be subject to miRNA regulation (Lewis et al. 2005). Hence, the control of miRNA
expression is by itself a key step in regulating target mRNAs.
1.1 microRNA Gene Transcription
Analysis of the genomic position of miRNAs revealed that the majority resides in
intergenic regions (Lagos-Quintana et al. 2003), indicating that they are autono-
mous transcription units. The miRNA genes embedded within known transcripts
are primarily found in intronic regions (Lagos-Quintana et al. 2003; Rodriguez et al.
2004; Kim and Kim 2007), although some are found in exonic locations, such as the
untranslated regions of mRNAs (Rodriguez et al. 2004; Kim and Kim 2007).
Animal miRNA genes are often localized in close proximity to each other forming
clusters (Lau et al. 2001; Lagos-Quintana et al. 2003), which are transcribed
polycistronically (Lee et al. 2002). In contrast, this polycistronic arrangement is
quite rare in plants with few exceptions (Guddeti et al. 2005; Zhang et al. 2008).
Transcription of miRNA genes is mediated by RNA Polymerase II (Pol II) as
primary miRNA transcripts (pri-miRNAs) have been shown to contain the hall-
marks of Pol II transcripts, a cap structure and a poly(A) tail (Cai et al. 2004; Lee
et al. 2004a). Further, expression of miRNAs was decreased by the Pol II-specific
inhibitor a-amanitin, and Pol II has been shown to associate with miRNA promoters
(Lee et al. 2004a; Bortolin-Cavaille et al. 2009). Transcription by Pol II enables
tissue-specific or developmental control by regulatory transcription factors. In
several cases, where the miRNA gene is transcribed together with a protein-coding
gene as a single transcription unit, the regulated expression pattern appears to be
coordinated (Baskerville and Bartel 2005).
The largest human miRNA gene cluster, C19MC, is embedded in Alu repeat
sequences. It was proposed that this cluster is unique in being transcribed by RNA
Pol III (Borchert et al. 2006). However, another study claims that C19MC miRNAs
are encoded within introns of a Pol II transcript (Bortolin-Cavaille et al. 2009).
1.2 microRNA Editing: Small Changes Affect Many Steps
During transcription, RNAs undergo various maturation processes, such as 50

capping, splicing, 30 end processing, polyadenylation, and RNA editing. The most
frequent form of RNA editing is the conversion of individual adenosines to inosines

in double-stranded RNAs by the action of ADARs (adenosine deaminases acting on
RNA). In vertebrates, three ADAR family members have been identified (ADAR1
to ADAR3), although ADAR3 has not yet been proven to be catalytically active. In
addition, ADAR1 yields two isoforms due to usage of alternative translation
initiation codons (Patterson and Samuel 1995). While ADAR2 and the short form
of ADAR1 are expressed constitutively and are localized in the nucleoplasm and
the nucleolus, the long isoform of ADAR1 is induced by interferon and located
primarily in the cytoplasm (Patterson and Samuel 1995; Desterro et al. 2003).
Cotranscriptional A ! I base conversion (Ryman et al. 2007) results in a sequence
different from the one encoded by the DNA template as inosines are interpreted as
guanosines by cellular machineries. Since the miRNA precursors form stem-loop
structures, they were also considered to be potential targets of A ! I editing.
Indeed, several studies in the past years have revealed that certain miRNA pre-
cursors are edited by ADAR enzymes. This sequence change has far reaching
consequences regarding processing and target site recognition of miRNAs as
structural and base pairing properties are altered.
Nishikura and colleagues have shown that editing of pri-miR-142 interferes with
the Drosha cleavage step, which leads to a reduction of mature miR-142 (Yang et al.
2006b). The edited, unprocessed pri-miRNA is degraded by the Tudor staphylo-
coccal nuclease (Tudor-SN), which preferentially cleaves dsRNA with multiple
IU wobble pairs (Scadden 2005; Yang et al. 2006b). In contrast, editing of two
other pri-miRNAs aids in Drosha processing and increases pre-miRNA levels
(Kawahara et al. 2008). While an effect of RNA editing on pre-miRNA export
into the cytoplasm has not been reported yet, it has been shown to abolish a further
downstream step, the cleavage by the Dicer-TRBP complex (Kawahara et al.
2007a). Some A ! I conversions, however, have no impact on either of the two
cleavage steps, leading to the expression of edited mature miRNAs if the edited site
resides in the mature miRNA sequence (Pfeffer et al. 2005). MiRNAs with altered
sequences result in targeting of mRNAs different from those targeted by the
unedited miRNAs, especially when the A ! I conversion is located in the seed
sequence (Kawahara et al. 2007b). Since selection of the guide strand, which is
incorporated into the RISC complex, depends highly on the stability of the miRNA
duplex (Khvorova et al. 2003; Schwarz et al. 2003), and A ! I editing is thought
to affect stability properties, RNA editing might also affect selection of the
“functional” guide strand.
In summary, A ! I RNA editing adds another layer of complexity, increasing
the pool of cellular miRNAs.
1.3 pri-miRNA Cleavage by the Microprocessor Complex
A typical primary transcript of miRNA genes comprises an imperfect stem of 33

basepairs (bp) in length with a terminal loop and adjacent single-stranded
sequences (Han et al. 2006) (Fig. 1). The hairpin, containing the future miRNA in
either the 50 or 30 half of its stem, is excised by the endonuclease Drosha, which
liberates the 60–70 nt long pre-miRNA with a two nucleotide (nt) overhang at its
30 -end (Lee et al. 2003; Denli et al. 2004). This process, also referred to as
“cropping”, occurs cotranscriptionally and precedes the splicing reaction (Kim
and Kim 2007; Morlando et al. 2008). Notably, the Drosha cleavage reaction is
no hindrance for the spliceosome to execute its function because continuity of the
intron is not a prerequisite (Dye et al. 2006; Kim and Kim 2007).
Drosha is a member of the highly conserved RNase III enzyme family, contain-
ing two RNase III domains, a double-stranded RNA binding domain (dsRBD), and
a long N-terminal region with unknown function (Filippov et al. 2000; Wu et al.
2000). By intramolecular dimerization, the two RNase III domains form a single
processing center with two catalytic sites that each cut one strand of the stem (Han
et al. 2004a). Drosha is present in two different complexes. In the smaller complex,
referred to as the Microprocessor, Drosha binds to its cofactor DGCR8 (DiGeorge
syndrome critical region 8; also known as Pasha in C. elegans and Drosophila
melanogaster) (Denli et al. 2004; Gregory et al. 2004; Han et al. 2004a). This
interaction is essential for the conversion of pri-miRNAs into pre-miRNAs
(Gregory et al. 2004; Han et al. 2004a). While Drosha is crucial for the catalysis,
DGCR8 establishes specificity for pri-miRNAs and determines the cleavage site
11 bp distant from the junction of the stem base and the flanking single-stranded
RNA (ssRNA) (Gregory et al. 2004; Han et al. 2006). Binding of the dsRBDs of
DGCR8 to the pri-miRNA requires the ssRNA regions, which are therefore indis-
pensible for pri-miRNA processing (Zeng and Cullen 2005; Han et al. 2006).
However, several sequence alterations in pri-miRNAs of human tumors lead to
conformational changes without affecting processing efficiency pointing towards a
certain flexibility of the Microprocessor (Diederichs and Haber 2006). Aside from
the necessity of the basal segments, the stem and a loop at the end of the stem are
important for efficient cleavage (Zeng et al. 2005b; Han et al. 2006). After deter-
mination of the cleavage site by DGCR8, Drosha can interact transiently with this
preformed RNA–protein complex and execute the cut.
Drosha-mediated cleavage represents a critical step in miRNA biogenesis since
this initial processing event defines one end of the mature miRNA. Nonetheless, it
is not compulsory for the generation of pre-miRNAs. Short introns can also form
hairpin structures that resemble pre-miRNAs. These alternative precursors, termed
“mirtrons”, can be spliced and debranched into pre-miRNA-like hairpins. Mirtrons
lack the lower stem of pri-miRNAs and therefore omit processing by the Micropro-
cessor. The debranched hairpins are then exported and further processed by the
canonical miRNA biogenesis pathway (Okamura et al. 2007; Ruby et al. 2007).
1.3.1 Regulation of the Microprocessor
To control Microprocessor activity, Drosha and DGCR8 regulate each other via a
regulatory feedback circuit. Protein–protein interaction of DGCR8 with Drosha
miRNA gene Pol II
ADAR
m7G
transcription & editing
pre-mRNA
m 7G
splicing
DGCR8
Drosha
pri-miRNA
branched pre-mirtron mRNA
AAA
Drosha cleavage debranching
RanGTP
Exp-5
5’ nucleus
pre-miRNA 3’
cytoplasm
Dicer
Dicer Ago2 TRBP

RISC loading complex (RLC) 5’
3’
Ago2 TRBP
Ago2 cleavage
Dicer
5’ ac-pre-miRNA
3’
Ago2 TRBP
Dicer cleavage
Dicer cleavage
Dicer Dicer
5’ 5’
miRNA duplex 3’ 3’
Ago2 TRBP Ago2 TRBP
duplex unwinding
& RISC formation
mature miRNA Ago2 passenger strand

active RISC 5’ degardation
Fig. 1 The microRNA biogenesis pathway in vertebrates. MicroRNA (miRNA) genes are tran-
scribed by RNA Polymerase II (Pol II) generating primary transcripts (pri-miRNAs). Cleavage by
the Microprocessor complex consisting of Drosha and DGCR8 results in a 65 nt precursor-
miRNA (pre-miRNA). Intron-derived miRNAs (mirtrons) are generated by splicing and debranch-
ing. The intron resembles the hairpin structure of the pre-miRNA thereby bypassing the Drosha
processing step. After export to the cytoplasm, mediated by Exportin-5 in a Ran-dependent
stabilizes Drosha. In turn, Drosha cleaves hairpins within the Dgcr8 mRNA result-
ing in destabilization of Dgcr8 mRNA (Han et al. 2009; Triboulet et al. 2009).
Apart from this general mechanism, Drosha-mediated cleavage of specific pri-
miRNAs can be influenced by additional factors. The DEAD-box RNA helicases
p68 (DDX5) and p72 (DDX17), which are present in the large Drosha complex,
seem to be involved in the processing of a subset of pri-miRNAs, as deficiency of
these factors decreases their mature miRNA expression levels (Fukuda et al. 2007).
Some auxiliary processing factors act even on individual miRNAs. Transforming
growth factor-b (TGF-b) and bone morphogenetic factors (BMPs) activate specific
SMAD signal transducers, which then form a complex with the Microprocessor via
the RNA helicase p68. This interaction promotes processing of pri-miR-21 into pre-
miR-21 by Drosha (Davis et al. 2008). Although the terminal loop seems to be of
inferior significance for Microprocessor action per se (Han et al. 2006), it seems to
have a fine-tuning role, providing a binding platform for regulatory proteins. The
heterogeneous nuclear ribonucleoprotein A1 (hnRNP A1) protein, for example,
contributes to the biogenesis of only a single miRNA in the miR-17 cluster,
miR-18a (Guil and Caceres 2007). By binding to the conserved loop of the hairpin,
hnRNP A1 remodels the stem structure and thereby creates a more advantageous
cleavage site for Drosha (Michlewski et al. 2008). In contrast, interaction of a stem
cell-specific regulator, lin-28, with the loop of pri-let-7 prevents Drosha cleavage
(Newman et al. 2008; Viswanathan et al. 2008). The repressive effect of lin-28
seems to be antagonized by the KH-type splicing regulatory protein (KSRP), which
promotes maturation of a let-7 and a subset of miRNA precursors by binding to the
terminal loop (Trabucchi et al. 2009). Thus, it is likely that there are far more
factors to be identified, optimizing recruitment and positioning of the Microproces-
sor thereby controlling processing in a coordinated manner.
1.3.2 Primary miRNA Generation in Plants
Plant miRNAs are also derived from long, primary transcripts – although rather
diverse in structure and with longer hairpins than in animals – in a stepwise process
(Kurihara and Watanabe 2004). Nevertheless, biogenesis in plants holds some
substantial differences compared to metazoans. A key characteristic of plant
miRNA maturation is the lack of a Drosha-like protein, which is highly conserved
Fig. 1 (Continued) manner, the pre-miRNA interacts with the preformed ternary complex of
Dicer, TRBP, and Ago2 forming the RISC loading complex (RLC). Dicer removes the terminal
loop of the pre-miRNA creating a miRNA duplex. Pre-miRNAs with a high degree of comple-
mentarity are cleaved by Ago2 in their passenger strand, producing a nicked hairpin called ac-pre-
miRNA, before Dicer cleavage. After unwinding of the miRNA duplex, the passenger strand is
degraded whereas the functional guide strand is loaded onto Ago2, constituting the RNA-induced
silencing complex (RISC), which silences target mRNAs by translational repression or mRNA
cleavage
in animals. One of four Dicer homologs, Dicer-like1 (DCL1; also known as

CARPEL FACTORY, CAF), seems to execute not only the Dicer-like maturation
step but is also responsible for catalyzing the pri- to pre-miRNA conversion
(Kurihara and Watanabe 2004). Both cleavage events seem to succeed quite fast,
as intermediate products are rarely detected. Due to the predominantly nuclear
localization of DCL1 (Papp et al. 2003), mature miRNAs are produced in the
nucleus and not in the cytoplasm like in animals. The whole processing requires
several additional proteins, such as the RNA-binding protein DAWDLE (DDL)
aiding in stabilizing the precursor (Yu et al. 2008), the zinc-finger protein SER-
RATE (SE) (Lobbes et al. 2006; Yang et al. 2006a), and a double-stranded RNA-
binding protein HYPONASTIC LEAVES1 (HYL1) (Han et al. 2004b; Hiraguri
et al. 2005). The latter two interact with DCL1 in the so-called D-bodies (Fang and
Spector 2007; Song et al. 2007).
An interesting particularity in plants is the protection of mature miRNA
duplexes by the S-adenosyl methionine-dependent methyltransferase Hua Enhancer
1 (HEN1). Modified by methyl groups at the 30 end of each strand, miRNAs are
more stable and can escape degradation by the SMALL DEGRADING NUCLE-
ASE (SDN) class of exonucleases (Li et al. 2005; Yang et al. 2006c).
1.4 Nuclear Export of the microRNA Precursors by Exportin-5
After the initial cropping by Drosha, the precursor has to be further processed to
yield the final miRNA. This second cleavage reaction is mediated by the cytoplas-
mic enzyme Dicer. Owing to the different spatial appearance of the two endonu-
cleases, nucleocytoplasmic transit of the pre-miRNA has a pivotal role in miRNA
biogenesis. Like other noncoding RNAs, pre-miRNAs are exported by a member of
the karyopherin family of nucleocytoplasmic transport receptors. Several studies
demonstrated that Exportin-5 is the main if not only transport factor for nuclear
export of pre-miRNAs and that the transport process – characteristic of karyopherin
mediated export – depends on the Ran cycle (Yi et al. 2003; Bohnsack et al. 2004;
Lund et al. 2004). Binding of Exportin-5 to its cargo requires a stem of at least 16 bp
and a short 30 -overhang, an end structure characteristic of RNase III cleavage.
Improperly processed pre-miRNAs with, e.g., 50 -overhangs prevent binding of
Exportin-5 and thus efficient export, highlighting the importance of precise cleav-
age by Drosha (Lund et al. 2004; Zeng and Cullen 2004). Notably, Exportin-5 is not
only required for nucleocytoplasmic transport of pre-miRNAs but also aids in
stabilizing the relatively unstable precursor (Yi et al. 2003; Zeng and Cullen 2004).
In plants, the role of HASTY, the plant homolog of Exportin-5 is not as well
defined as in animals. Since miRNAs are completely processed in the nucleus,
mature miRNAs are found in both the nucleus and the cytoplasm; however, they are
more abundant in the latter compartment. Although loss-of-function mutants of
HASTY reduce miRNA levels, they do not cause accumulation in the nucleus.
Hence, till date, there is no direct evidence that HASTY is involved in miRNA
export. Also, the exact form as which miRNAs leave the nucleus – either as miRNA
itself, loaded into the RISC complex, or in complex with their target mRNAs –
remains still unclear (Park et al. 2005).
1.5 The RISC Loading Complex
Once in the cytoplasm, the pre-miRNA still has to undergo several additional
processing steps before the single-stranded mature miRNA can guide the effector
complex known as RNA-induced silencing complex (RISC) to its target mRNA.
Maturation of the pre-miRNA and RISC loading is performed by the RISC loading
complex (RLC). This complex comprises in the human system the RNases and
Dicer and the structurally related dsRNA binding proteins TRBP and PACT
(Chendrimada et al. 2005; Haase et al. 2005; Lee et al. 2006; MacRae et al.
2008). Like Drosha, Dicer is an RNase III enzyme and removes the terminal loop
by a staggered cut. Albeit TRBP and PACT are not required for the Dicer-mediated
processing step per se, they have a stimulating effect on the cleavage reaction,
influence the efficiency of RNA silencing, and are involved in the recruitment of
Ago2 (Haase et al. 2005; Lee et al. 2006). Ago2 (also known as eIF2C2) acts in
the effector phase of RNAi and cleaves the mRNA targeted for destruction by
the complementary miRNA or mediates – as well as other human Ago proteins –
translational inhibition. Ago2 stability and thus also efficiency of RNAi is regulated
by hydroxylation (Qi et al. 2008).
The general paradigm that RNase III enzymes cooperate with dsRBD proteins
holds true for Drosophila Dicer enzymes. Dicer-1 and Dicer-2 interact with Loqua-
cious (Loqs) and R2D2, respectively. While the heterodimer Dicer-1/Loqs processes
pre-miRNAs, the counterpart Dicer-2/R2D2cleaves long dsRNAs into siRNAs
(Lee et al. 2004b). Loqs is required for efficient pre-miRNA processing and confers
substrate specificity for pre-miRNAs to Dicer-1 (Forstemann et al. 2005; Saito et al.
2005), whereas R2D2 is necessary for RISC loading but not for the cleavage reaction
(Liu et al. 2003; Tomari et al. 2004b).
The assembly of the RLC and the final RISC differs in flies and humans.
In humans, the RLC is formed prior to pre-miRNA binding and Dicer is released
after miRNA incorporation into Ago2 (Gregory et al. 2005; Maniataki and Mour-
elatos 2005). On the contrary, in Drosophila, Dcr-2/R2D2 binds first the siRNA
duplex. Thereafter, Ago2 joins the complex and the single-stranded siRNA is
loaded onto Ago2. This final complex is called holo-RISC and still retains Dcr-2/
R2D2 (Pham et al. 2004; Tomari et al. 2004b). For the miRNA pathway, however,
Dicer-1/Loqs was recently shown to be excluded from the RISC. After processing of
the pre-miRNA and loading of the duplex onto Ago1, Dicer dissociates from the
RLC. Ago1, loaded with the mature miRNA, interacts then with GW182, a P body
component with a role in miRNA-mediated silencing, forming the RISC (Miyoshi
et al. 2009).
1.6 Terminal Loop Removal by Dicer
Once the pre-miRNA joins the RLC, Dicer cleaves the precursor removing the
terminal loop. The cleavage defines the other end of the mature miRNA and results
in a 22 nt long miRNA duplex with 2 nt 30 -overhangs at each end (Bernstein et al.
2001; Provost et al. 2002). While in invertebrates this step necessitates ATP
(Zamore et al. 2000; Nykanen et al. 2001), dicing is ATP-independent in mamma-
lian cells (Provost et al. 2002; Zhang et al. 2002). Studies with dicer mutants in
different organisms have shown the significance of the Dicer reaction for functional
RNAi and its requirement for embryonic development (Knight and Bass 2001;
Bernstein et al. 2003). Dicer proteins are evolutionary conserved throughout
the eukaryotic kingdom except budding yeast (Bernstein et al. 2001). In some
organisms, even multiple Dicer homologs exist, dedicated to mainly one aspect
of RNA silencing, like generation of siRNAs or miRNAs. In contrast to Arabidopsis
and Drosophila, where two Dicer isoforms can share the duties (Lee et al. 2004b;
Xie et al. 2004), the single Dicer enzymes in nematodes and humans process both
dsRNA and miRNA precursors.
As members of the RNase III family, Dicer enzymes comprise two neighboring
RNase III domains responsible for the catalytic reaction and a dsRBD that interacts
with dsRNA in vitro (Provost et al. 2002). The N-terminus harbors in addition a
DExH/DEAH box RNA helicase domain, a PAZ domain and a domain of
unknown function (DUF283). The RNA-binding PAZ domain, named after the
Piwi, Argonaute, and Zwille proteins, is also found in the Argonaute protein family
and adopts a topology related to the oligonucleotide/oligosaccharide-binding fold
(OB-fold). The binding pocket for dsRNAs with a two nucleotide 30 -overhang
within the PAZ domains enables the anchoring and recognition of pre-miRNAs
processed by Drosha (Song et al. 2003; Lingel et al. 2004; Ma et al. 2004).
An additional loop in the PAZ domain of the Dicer family alters the electrostatic
potential of the surface surrounding the binding pocket compared to the Argonaute
PAZ. Due to this difference, substrate recognition and transfer of the substrate to
other complexes might differ between the two PAZ-domain families (Macrae et al.
2006). Notably, some Dicer proteins, such as Drosophila Dicer-2, Arabidopsis
DCL-4, or Schizosaccharomyces pombe Dicer, do not contain a PAZ domain.
In these species, adaptor molecules might compensate the lack of a PAZ domain
providing an alternative to recognize pre-miRNAs or the organism – like in the case
of S. pombe – does not encode miRNAs.
The crystal structure of the parasite Giardia intestinalis Dicer shed light onto the
question how the Dicer cleavage site is determined. G. intestinalis Dicer is a
minimalist among the Dicer enzymes as it consists only of a PAZ domain and
two consecutive RNase III domains. However, it generates RNA duplexes of
25–27 nt from dsRNA in a similar way to human Dicer (Zhang et al. 2002; Macrae
et al. 2006). Unlike the nuclear RNase III Drosha, which is dependent on the “ruler”
activity of its partner DGCR8/Pasha, Dicer is capable of measuring the distance of
about 25 nt from the 30 -end on its own. This length is defined by the distance
between the PAZ domain that binds the 30 - and the 50 -end and the RNase IIIa
domain, forming the catalytic center (Zhang et al. 2004; Macrae et al. 2006). The
connecting helix between both domains is predicted to form in all Dicer homologs,
and thus also the mechanism of cleavage site determination might be conserved
(Macrae et al. 2006).
1.6.1 Control Mechanisms of the Dicer Cleavage
Dicer activity is also subject to diverse regulatory mechanisms. One striking

example is the tissue-specific Dicer cleavage of pre-miR-138-2. While the precur-
sor is expressed ubiquitously, the mature miR-138 is found only in certain cell
types. As the precursor is exported to the cytoplasm, the lack of mature miR-138
seems due to inhibition of Dicer although the mechanism and involved factors
are unclear (Obernosterer et al. 2006). Another regulatory layer is mediated by
the Dicer interaction partner TRBP. In the absence of TRBP, the N-terminal
DExD/H-box helicase domain of human Dicer displays a low rate of substrate
cleavage. Binding of TRBP stimulates Dicer-mediated catalysis possibly by induc-
ing a conformational change. This mechanism could prevent unintentional activity
of free Dicer before incorporation into the RLC (Ma et al. 2008). Additionally, its
product, the mature let-7 miRNA, controls Dicer activity creating a negative
feedback loop. Let-7 downregulates Dicer by binding to complementary sequences
in the dicer mRNA (Forman et al. 2008; Tokumaru et al. 2008). In vitro, maturation
of pre-let-7 itself is inhibited by the presence of Lin-28 (Rybak et al. 2008), a factor
that has also regulatory function in the Microprocessor step (see above). Further-
more, Lin-28 mediates 30 -terminal uridylation of pre-let-7 in the cytoplasm. The
elongated tail of pre-let-7 impedes Dicer cleavage and targets the precursor for
degradation constituting another regulatory mechanism for Dicer activity for a
specific miRNA (Heo et al. 2008). Lastly, KSRP, also known as transcription factor
FBP2 with regulatory roles in the nucleus, acts as an antagonist of Lin-28 and
promotes Dicer processing of a subset of pre-miRNAs (Trabucchi et al. 2009).
1.7 Ago2 Jumps the Queue: Generation of the ac-pre-miRNA
After joining the preformed RLC, certain pre-miRNAs undergo an additional

endonucleolytic processing step prior to Dicer cleavage (Diederichs and Haber
2007). Ago2 cleaves in highly complementary stems the prospective passenger
strand, which is not designated to become the mature miRNA, 12 nt from its
30 -end. The generated processing intermediate, termed Ago2-cleaved pre-miRNA
(ac-pre-miRNA), joins then the canonical pathway and is processed by Dicer. The
biological function of the ac-pre-miRNA is still unsolved but is speculated to
influence strand selection or to alleviate removal of the passenger strand by analogy
to siRNA processing (Matranga et al. 2005; Rand et al. 2005).
1.8 miRNA Duplex Unwinding
As soon as the miRNA duplex is generated by Dicer, Dicer and its binding partners
TRBP and PACT dissociate from the miRNA. Since only one strand is finally
retained in the active RISC complex, the duplex has to be separated into their
component strands: the guide strand, complementary to the target and mediating
RNAi, and the nonfunctional passenger strand, which is subsequently degraded.
Separation of double-stranded RNA molecules is usually mediated by helicases
using energy derived from ATP hydrolysis. Albeit several helicases have been
shown to associate with proteins that act in RISC formation or activity and
influence RNA silencing, a direct involvement in strand unwinding could till date
not be proven. For example, the DEAD box helicases Gemin3/4, RCK/p54, and the
putative DExD-box helicase MOV10 interact with Argonaute proteins and are
required for posttranscriptional silencing (Mourelatos et al. 2002; Meister et al.
2005; Chu and Rana 2006). MOV10 homologs in flies (Armitage) or plants
(SDE-3) play also a role in RNAi (Cook et al. 2004; Tomari et al. 2004a). RHA/
DHX9, a member of the DEAH-containing family of RNA helicases unwinds
dsRNA (Lee and Hurwitz 1992) and aids in active RISC loading by promoting
the association of siRNA with Ago2 (Robb and Rana 2007). Although this obser-
vation points towards a role of RHA in siRNA duplex unwinding, it is unclear
whether this possible function also applies to miRNAs. For miRNA let-7 unwind-
ing, the ATP-dependent helicase p68/DDX5 is sufficient in vitro. In accordance
with this, the lack of p68/DDX5 inhibits let-7 miRNA function (Salzman et al.
2007). These findings and the multitude of potential factors in the unwinding
process suggest that specific helicases may regulate particular subclasses of miRNAs.
Whether they participate directly in unwinding the duplexes or whether they rather
remodel the RLC to facilitate miRNA loading remains to be investigated. As RISC
assembly and RISC loading can be accomplished in an ATP-independent manner
(Gregory et al. 2005; Maniataki and Mourelatos 2005; MacRae et al. 2008), the
general necessity of helicases in this process is challenged.
Another factor implicated in strand separation is the endonuclease Ago2. In the
siRNA pathway, the effector protein Ago2, which is loaded with the siRNA duplex,
cleaves the passenger strand to reduce the internal stability (Matranga et al. 2005;
Rand et al. 2005). This destabilization is necessary for strand dissociation and
facilitates removal of the passenger strand. Mismatches and unpaired bulges in
some miRNA hairpin stems could render the cleavage-assisted strand separation
mechanism not only feasible but also unnecessary due to their inherent thermody-
namic instability. However, Ago2 is able to cleave the passenger strand of si-like
miRNA precursor stems with a highly complementary sequence like the let-7
miRNA family (Diederichs and Haber 2007) (see above). Thus, Ago2-mediated
passenger strand cleavage and generation of the ac-pre-miRNA could facilitate
strand dissociation and RISC activation also in the miRNA pathway – at least for
miRNA precursors that resemble siRNAs with a high degree of complementarity in
the middle of the hairpin stem. In Drosophila, a novel complex of Translin and
Trax, called C3PO (component 3 promoter of RISC), aids in unwinding of siRNA

duplexes and removes the remnants of the passenger strand caused by the Ago2-
mediated nick. Thereby, C3PO supports active RISC formation and enhances RNAi
(Liu et al. 2009). For miRNA duplexes loaded onto the weak endonuclease Dro-
sophila Ago1, a slicer-dependent strand separation is not possible. In this case,
structural features of miRNAs – mismatches in the seed and the 30 mid region – are
pivotal for strand unwinding (Kawamata et al. 2009). If this mechanism holds true
also for the mammalian nonslicer Ago proteins, Ago1, Ago3, and Ago4, has to be
investigated.
1.9 Strand Selection: Who Becomes the Guide?
In theory, unwinding of the miRNA duplex yields two different mature miRNAs
with the potential to become the effector strand. However, the two strands are not
equally competent in entering the RISC. Therefore, only one strand from each
duplex, the so-called guide strand, is predominantly incorporated into the RISC.
The remaining strand, named passenger strand, is primarily targeted for degrada-
tion, although both strands can be functionally active (Ro et al. 2007; Okamura
et al. 2008). Which strand is chosen to participate in RNAi and which is condemned
to be destroyed is predestined by the thermodynamic properties of the base pairs at
the ends of the duplex. The strand whose 50 -end is less stably paired to its
counterpart is loaded into RISC (Khvorova et al. 2003; Schwarz et al. 2003).
In Drosophila, the binding partner of Dicer-2, R2D2, senses the functional asym-
metry of the two siRNA strands and binds the strand with the greater double-
stranded character. R2D2 orients the Dcr-2/R2D2 heterodimer on the siRNA in
such a way that the correct strand is handed to Ago2, thereby facilitating RISC
loading (Liu et al. 2003; Tomari et al. 2004b).
1.10 Mediators of RNA Silencing: The Argonaute Proteins
RISC is the effector complex that silences target transcripts posttranscriptionally by

degradation or translational inhibition. Aside from the mature miRNA and the
target mRNA, this ribonucleoprotein complex contains a member of the Argonaute
family as core protein, which binds to the single-stranded miRNA, and also
includes accessory factors, such as GW182, aiding in the effector step. The Argo-
naute family is the largest protein family involved in RNAi and can be divided
into the Ago and the Piwi subfamily. While members of the Ago subfamily are
ubiquitously expressed, expression of Piwi proteins is restricted to the germ line
(Sasaki et al. 2003). The whole Argonaute family is defined by three characteristic
domains, the PAZ, the MID, and the PIWI domain. While the MID domain anchors
the guide strand by binding to its 50 -phosphate (Ma et al. 2005; Parker et al. 2005;
Yuan et al. 2005; Wang et al. 2008b), the PAZ domain (as described above)
provides a binding pocket for dsRNAs with 30 overhangs (Song et al. 2003; Lingel
et al. 2004; Ma et al. 2004; Wang et al. 2008b) suggesting that the miRNA after
generation by Dicer could be directly handed over from the Dicer-PAZ to the
Argonaute-PAZ domain. The PIWI domain structurally resembles an RNase H
fold and provides the endonuclease activity for mRNA target cleavage (Song et al.
2004; Rivas et al. 2005; Yuan et al. 2005). Structural studies with a ternary complex
consisting of a Thermus thermophilus argonaute protein, a guide DNA and a target
RNA revealed the processes within this complex that happen upon binding to the
RNA target (Wang et al. 2008a, 2009). The binding of the target RNA starts with
basepairing to the 50 end of the guide DNA, which is anchored in the MID domain.
Progressing duplex formation liberates the 30 end of the guide DNA from its
anchoring site in the PAZ domain. This release leads to a conformational change,
which brings the cleavage site of the target RNA in close proximity to the catalytic
residues in the PIWI domain.
Although most organisms encode several Argonaute proteins (ranging from one
in S. pombe to 27 in C. elegans), which are functionally not redundant, many of
them are endonucleolytically inactive. In humans, only Ago2, also known as
eIF2C2, is equipped with the so-called “slicer” activity (Meister et al. 2004).
However, Ago2 is not the only Argonaute protein associated with miRNAs. The
remaining three members of the Ago subfamily, Ago1, Ago3, and Ago4, interact
with miRNAs as well (Meister et al. 2004). How miRNAs are partitioned between
effector complexes with different Ago proteins and what are the functional
consequences of this sorting in the mammalian system is currently unknown.
More about sorting of small RNAs in different RISC variants is known in flies
and plants. In Drosophila, siRNAs and miRNAs are generated by two different
Dicer enzymes (Lee et al. 2004b). It was assumed that due to their different
biogenesis, the two classes of small RNAs are predetermined to be loaded into a
particular RISC variant, miRNAs into Ago1-RISCs and siRNAs into Ago2-RISCs
(Okamura et al. 2004; Saito et al. 2005). In contrast, Zamore and colleagues could
show that Argonaute loading is not coupled to the biogenesis pathway of miRNAs
or siRNAs but that sorting in one or the other complex depends rather on the
intrinsic structure of the small RNA (Forstemann et al. 2007; Tomari et al. 2007).
The heterodimer Dcr-2/R2D2 enforces incorporation of perfectly matched short
RNAs (e.g., siRNAs) into Ago2-RISC, while bulged or mismatched miRNAs are
excluded. In analogy, a yet unidentified mechanism sorts nonperfectly matched
miRNAs into Ago1-RISCs. For small RNAs with a structure in between that of a
completely complementary siRNA duplex and a typical miRNA duplex with bulges
and mismatches, both RISCs compete (Tomari et al. 2007). Importantly, sorting
into one or the other RISCs determines the fate of the target mRNA as the
two Argonaute proteins are functionally specialized: Ago1-RISCs only silence
imperfectly matched targets whereas Ago2-RISCs, which have the stronger cata-
lytic capacity, silence targets that are fully complementary to the guide strand
(Forstemann et al. 2007). Notably, only those guide strands – irrespective from
which arm of the pre-miRNA they originate – are capable of associating with
Drosophila Ago2, which have a 50 -end derived from an accurate cleavage by Drosha
or Dicer (Seitz et al. 2008). In Arabidopsis, sorting of small RNAs is directed by the
50 -terminal nucleotide; e.g., joining AGO1, the Argonaute protein prevailing in the
miRNA pathway, requires a uridine at the 50 -end (Seitz et al. 2008).
1.10.1 Regulation of Ago Activity and Ago-Mediated Silencing Mechanisms
Once loaded onto an Argonaute protein, the miRNA guides the complex to a fully
or partially complementary mRNA target. Target recognition is considered to
require primarily full complementarity of nucleotides 2–7, termed the miRNA
“seed” sequence, but other nucleotides have been shown to aid in this process
(Doench and Sharp 2004; Brennecke et al. 2005; Grimson et al. 2007). The shortness
of the seed region enables each miRNA to regulate a large number of target genes.
The extent of complementarity between the miRNA and the target is likely the
key determinant for the mechanism of regulation. Highly complementary target
sites – which particularly occur in plants but rarely in animals – cause slicing and
subsequent decay of the target by 50 - and 30 -exonucleases (Zamore et al. 2000). For
siRNAs, the 50 -end of the guide strand defines the position of the target cleavage
between the nucleotides paired to bases 10 and 11 of the siRNA (Elbashir et al.
2001b,c). Most animal miRNAs form imperfect hybrids with their target mRNA,
which precludes endonucleolytic cleavage, due to central mismatches (bases 9–12)
(Elbashir et al. 2001c). Instead, they promote translational attenuation or exonu-
cleolytic mRNA decay. The underlying molecular mechanism of this silencing
pathway is still under debate. Accounting the multitude of steps that have been
proposed to be targeted by RISC action, it is likely that more than one mechanism is
involved. For example, evidences exist that translation inhibition can be mediated
at the level of translation initiation as well as elongation. Additionally, models for
premature termination or cotranslational protein degradation have been suggested
(Eulalio et al. 2008; Wu and Belasco 2008). By triggering the removal of the target
poly(A) tail and subsequent decapping, miRNAs can cause exonucleolytic mRNA
decay (Behm-Ansmant et al. 2006; Wu et al. 2006; Eulalio et al. 2007b). The
enzymes for deadenylation and decapping localize to processing bodies (P bodies),
cytoplasmic sites of mRNA turnover (Eulalio et al. 2007a). Since Argonaute
proteins, miRNAs, and their targets colocalize with these foci, miRNA targets
could be directed to P bodies to be separated from the translation machinery and
be accessible for the decay components (Liu et al. 2005; Pillai et al. 2005; Sen and
Blau 2005). Targeting of Ago2 to P bodies is regulated by its phosphorylation status
(Zeng et al. 2008). The environment of the P bodies per se is not essential for target
silencing as miRNA function is not or only marginally affected by the loss of
P-bodies. Hence, enrichment of the RISC components in P bodies seems to be the
consequence and not the cause of silencing (Eulalio et al. 2007a).
Considering their core function in translation inhibition by various means, it is
not surprising that Argonaute proteins are mainly localized in the cytoplasm and
enriched in processing bodies. Nevertheless, nuclear functions for Ago proteins
have been reported in worms, flies, plants, and S. pombe (Lippman and Martienssen
2004; Matzke and Birchler 2005). Evidence is accumulating that human Ago proteins
also function in the nucleus. They were detected in nuclear fractions (Robb et al.
2005), and Ago1 and Ago2 have also been shown to associate with the promoter
region of the progesterone receptor (Janowski et al. 2006). Additionally, siRNAs
directed against promoter regions resulted in Ago1- or Ago2-dependent transcrip-
tional repression (Morris et al. 2004; Janowski et al. 2006; Kim et al. 2006). Importin-
8 is one factor involved in the transport of Ago2 into the nucleus (Weinmann et al.
2009). In addition to the effector proteins, mature miRNAs have also been detected
in the nucleus. The human miR-29b is even predominantly detected in the nucleus
(Meister et al. 2004; Hwang et al. 2007). Nuclear import of miR-29b is mediated
by a 30 -terminal hexanucleotide motif, which acts as a nuclear localization element
(Hwang et al. 2007). These examples of reimport of a miRNA into the nucleus
and the nuclear localization and action of effector proteins raise the possibility
that RISC-dependent gene silencing could occur also in humans at the transcriptional
level.
1.11 Half-Life and Degradation of microRNA
For steady-state miRNA expression levels, their stability, turnover, and degradation
could be as important as the regulation of miRNA maturation. So far, rather little is
known about this aspect of the miRNA life cycle.
Generally, mature miRNAs are rather stable, as they persist long in the cell after
depletion of miRNA processing factors (Lee et al. 2003; Gregory et al. 2004).
However, miRNA levels can drop rapidly under certain conditions, albeit the
mechanism is yet unidentified (Pedersen et al. 2007). One factor known to play a
role in miRNA homeostasis is Ago2, as miRNA levels depend on the amount of
Ago2 in the cell (Diederichs and Haber 2007). Exoribonucleases responsible for
miRNA degradation have till date only been discovered in two organisms: SDN1 in
Arabidopsis and XRN-2 in C. elegans (Ramachandran and Chen 2008; Chatterjee
and Grosshans 2009). XRN-2 aids in miRNA release from the Argonaute complex
and mediates miRNA turnover. Both separation and degradation are antagonized
by the presence of target RNA implicating a coordination of miRNA and target
RNA levels (Chatterjee and Grosshans 2009). An alternative way to impede miRNA
activity could be the shielding of miRNA-binding sites by interaction with RNA-
binding proteins (Bhattacharyya et al. 2006; Kedde et al. 2007).
2 Implication for RNAi Technology
The discovery of gene silencing by small RNA molecules, termed RNAi, was a
milestone in biology by providing a novel level of gene expression regulation.
Beyond the scientific importance, RNAi opened up the possibility for experimental
and therapeutic applications. However, since the small RNA molecules, miRNAs
or siRNAs, provide only the specificity and depend on the endogenous miRNA
pathway to mediate RNAi, a detailed knowledge of miRNA biogenesis is manda-
tory to use RNAi successfully as tool (Fig. 2).
miRNA gene
shRNA-mir
shRNA
m7G
construct
pri-miRNA
AAA
shRNA
5’ nucleus
pre-miRNA 3’
RanGTP
cytoplasm
Exp-5
5’ Ago2
Dicer 3’
Ago2 cleavage
Dicer cleavage
siRNA 5’ ac-pre-miRNA
3’
Dicer
Dicer cleavage
5’ 5’
miRNA duplex 3’ 3’
duplex unwinding
& RISC formation
mature miRNA Ago2 passenger strand

active RISC 5’ degardation
Fig. 2 RNAi takes advantage of the endogenous miRNA biogenesis machinery and effector
pathway. To use RNAi as a tool, artificial RNAs can be used, which enter the endogenous pathway
at different steps. While shRNAs or shRNA-mirs, which are transcribed from a plasmid by
Polymerase III or II, undergo (almost) the entire processing pathway, synthetic siRNA duplexes
join the pathway only in the cytoplasm. For simplicity, only those factors improving RNAi are
shown. Exp-5; Exportin-5
2.1 Potentials and Challenges of siRNAs as a Tool
In basic research, a fundamental approach to investigate the biological role of a

particular gene is the use of loss-of-function studies. In vertebrates, the lack of
methodologies for a gene knockout was a major drawback to investigate deletion
phenotypes. With the revolutionary discovery of RNAi by Mello and Fire in 1998, a
powerful tool for silencing of specific gene products became available (Fire et al.
1998). An initial obstacle for successful application in mammalian cells was the
existence of an antiviral response triggered by foreign dsRNAs. Accumulating
dsRNAs that are longer than 30 bp activate two enzymes: protein kinase R (PKR)
and 20 ,50 -oligoadenylate synthetase (20 ,50 -AS). Active PKR phosphorylates the
translation initiation factor eIF2a, which leads ultimately to a global inhibition of
translation, whereas 20 ,50 -AS causes activation of RNase L, a ribonuclease that
degrades mRNAs nonspecifically. The nonspecific effects of the immune response
would disguise any sequence-specific regulation generated by RNAi. The solution
turned out to be quite simple: Long dsRNAs are processed to 21–23 nt short
siRNAs, which mediate RNAi (Hammond et al. 2000; Zamore et al. 2000), and
the dsRNA-induced nonspecific responses are dependent on at least 30 bp length of
the dsRNA (Minks et al. 1979; Manche et al. 1992). Hence, attempts to use short
siRNAs in analogy to the successful application of synthetic siRNAs in Drosophila
(Elbashir et al. 2001b) have also been made in the mammalian system (Caplen et al.
2001; Elbashir et al. 2001a). Indeed, the small RNAs were sufficient to mediate
RNAi and due to their shortness without causing unspecific inhibitory effects
caused by PKR activation. Since this breakthrough, RNAi has been established as
a standard methodology for gene silencing in mammalian cells and opened up new
possibilities for gene-targeted therapy approaches.
However, it turned out that RNAi triggered by artificial siRNAs is not as specific
as originally assumed. Off-target effects have been observed in a concentration-
dependent manner likely caused by induction of certain signaling pathways or
knockdown of genes other than the desired target (Sledz et al. 2003; Persengiev
et al. 2004). Another reason for off-target effects might be the possibility that both
strands can be loaded into RISC (Ro et al. 2007; Okamura et al. 2008) resulting not
only in silencing of the target by the guide strand but also in undesired repression of
mRNAs complementary to the passenger strand. Studies on requirements for
correct strand selection revealed that the strand with the less stably paired 50 -end
is chosen for incorporation into RISC (Khvorova et al. 2003; Schwarz et al. 2003).
Hence, off-target effects caused by the complementary siRNA strand can be
minimized by designing an asymmetric siRNA with high internal instability at
the 50 -end of the desired guide strand or by chemically modifying the 50 -end of the
passenger strand to prevent phosphorylation and thus RISC loading (Khvorova et al.
2003; Schwarz et al. 2003). Further, targeting requires primarily complementarity
between the target and the seed sequence comprising nucleotides 2–7 of the siRNA
(Doench and Sharp 2004; Brennecke et al. 2005; Grimson et al. 2007). Therefore,
siRNAs – although absolutely complementary to their desired target region – might
be able to repress also targets, which only partially basepair with the siRNA (Jackson
et al. 2003).
2.2 siRNA Versus shRNA
The subsequent elucidation of the RNAi machinery components and discovery of

the biogenesis pathway of endogenous miRNAs enabled an alternative to siRNAs.
Vectors were developed that produce short hairpin RNAs (shRNAs) mimicking
miRNA precursors, which are processed into siRNAs by the cellular miRNA
biogenesis factors and thus trigger RNAi (Brummelkamp et al. 2002; Paddison
et al. 2002). The question, which system should be used for gene silencing cannot
be answered generally and depends on the particular experiment. Although siRNAs
are easily available with constant quality, one has to consider that the effects
obtained by siRNAs are only transient. Moreover, using siRNAs for genome-
wide applications could be quite cost-intensive. The advantage of shRNAs is the
ability to maintain repression as transfection of shRNA expression vectors leads to
continuous siRNA production. Using advanced viral vector backbones, integration
into the genome is also possible. Another plus is the possibility of using inducible or
regulatable promoters that drive shRNA production under certain conditions or in
specific tissues. However, the big challenge is the design of shRNAs so that they
meet the demands of the miRNA processing factors and that they mediate specific
and efficient target knockdown without off-target effects.
Intensive studies on miRNA biogenesis over the past years broadened our
knowledge about the factors involved in generation of small RNAs and shed light
onto the molecular mechanisms. Detailed mutational analysis, for example,
revealed the structural requirements of the pri-miRNA to be processed efficiently
and accurately by the Microprocessor complex and thus provides a basis for the
rational design of shRNAs (Han et al. 2006). The authors suggest creating a shRNA
with a 33–35 bp stem where the Drosha cleavage site is located 11 bp distant from
the base segments. In case of guide strands derived from the 50 -arm of the hairpin,
position þ1 should contain a mismatch but position þ19 to þ20 should form a
stable basepair. To maximize processing efficiency and accuracy, addition of
flexible ssRNA segments at the basal stem is advantageous (Han et al. 2006).
2.3 shRNA-miR Library: Transferring microRNA Structures

to Synthetic shRNAs
Early approaches with shRNAs – when knowledge about miRNA biogenesis was
still rudimentary – used simple short hairpin RNAs of varying stem length
(19–29 bp) and loops of 4–15 nt transcribed under the control of RNA polymerase
III (Brummelkamp et al. 2002; Paddison et al. 2002). Those vectors have been
successfully used to mediate RNAi. The deeper insights in the action of miRNA
processing factors and the knowledge gained on structural demands for efficient
cleavage and strand incorporation found consideration in the design of shRNA
vectors. The new generation of shRNAs, termed shRNA-miRs, are modeled after
endogenous pri-miRNAs (Dickins et al. 2005; Silva et al. 2005). The starting basis
for such shRNA-miRs is the backbone of the primary miR-30a, which is a well-
studied human miRNA of which the processing sites have been mapped (Zeng and
Cullen 2003). When the miR-30a stem is replaced by heterologous sequences, the
particular miRNA is generated and mediates effective and regulated target gene
inhibition (Zeng et al. 2002, 2005a). Using the backbone of pri-miR-30a, Hannon
and colleagues constructed an improved shRNA-miR library (Silva et al. 2005).
The vector designed for this library contains not only the remodeled pri-miR-30a
but also the 50 - and 30 -flanking regions to optimize ectopic expression (Chen et al.
2004). Mapping the processing sites of the shRNA-miRs allowed the shRNA design
following the thermodynamic asymmetry rule, even when cleavage is inaccurate
and translocated 1 bp (Seitz et al. 2008). One criterion for efficient RNAi is the
amount of siRNA that is available for incorporation into RISC (Siolas et al. 2005).
In comparison with first generation constructs containing simple short hairpins, the
shRNA-miR vectors produced about 12 times more small RNAs (Silva et al. 2005).
The expression of the shRNAs is like in the first generation library driven by a U6
promoter (Pol III dependent), since this gave together with expression from a CMV
promoter (Pol II dependent) the best results regarding efficiency and consistency of
target repression. However, one advantage of the advanced library system is the
possibility to swap the shRNA-miR cassette into other vectors without transferring
the U6 promoter. By recombination, the shRNA-miR can thus be combined with
any other desired promoter (e.g., promoters for inducible or tissue-specific expres-
sion) (Dickins et al. 2005; Stegmeier et al. 2005). With the resulting shRNA-miR
library (Hannon–Elledge library), the majority of human and mouse genes is
covered on average by two shRNA-miRs (Silva et al. 2005), which allows the
functional analysis on a single gene basis or in genome-wide approaches.
2.4 Enhancement of RNAi by microRNA Biogenesis Factors
Mimicking miRNA biogenesis intermediates in order to enable entrance of syn-

thetic molecules into the endogenous miRNA pathway promises efficient repres-
sion of the target of interest. However, utilization of the miRNA biogenesis and
effector pathway also leads to competition with the natural substrates, endogenous
miRNA precursors, for limiting processing factors. Overexpression of artificial
shRNAs thus interferes with endogenous RNAi and causes undesired side effect
(Grimm et al. 2006). Knowledge of factors with saturable activity could on
the other hand also be used to optimize RNAi. The pre-miRNA export factor,
Exportin-5, which is also required for the export of shRNAs (Yi et al. 2003;
Bohnsack et al. 2004; Lund et al. 2004) is expressed at low levels. Indeed, over-
expression of Exportin-5 enhances RNAi triggered by shRNAs (Yi et al. 2005).
However, the effect caused by Exportin-5 overexpression is not absolutely specific.
Reporter constructs are silenced by elevated Exportin-5 levels even in the absence
of a corresponding ectopic miRNA. Further, repression of the transcript also
occurred when the miRNA-binding sites were mutated, making an approach with
Exportin-5 more prone to off-target effects (Diederichs et al. 2008).
Apart from Exportin-5, other factors of the miRNA biogenesis and effector
machinery have been tested for their ability to improve RNAi. While Drosophila
Dicer-2 could enhance RNAi (Dietzl et al. 2007), human Dicer did not affect RNAi
potency in mammalian cells (Diederichs et al. 2008). In contrast, human Ago2
overexpression led to increased let-7a activity towards a reporter constructs con-
taining let-7a binding sites (Diederichs et al. 2008). This effect seems not to be due
to a simultaneous increase of mature let-7a miRNA, which is derived from a
construct encoding the let-7a primary miRNA, as overexpression of the other
Argonaute proteins, Ago1, Ago3, and Ago4, results also in elevated levels of
mature let-7a but not in enhanced RNAi. Importantly, the effect obtained by over-
expression of Ago2 is extremely specific, as it acts only towards perfectly matched
binding sites, and endogenous RNAi activity is not affected. That coexpression of
Ago2 allows reduction of the pri-miRNA construct concentration for transfection
is another asset (Diederichs et al. 2008). Thereby, off-target effects caused by
oversaturation of the endogenous pathway (Grimm et al. 2006) can be reduced.
Especially high-throughput screens using shRNA or siRNA libraries can profit from
these effects. False-positive or false-negative results of such screens could be
reduced by coexpression of Ago2, thereby increasing specificity and efficiency of
RNAi. Moreover, overexpression of Ago2 could be of use for medical applications
based on RNAi technology. The mechanism behind the effect of Ago2 overexpres-
sion on RNAi is likely a combination of several modes of action. First, Ago2
improves the stability of mature miRNAs. However, increased expression levels
of mature miRNAs alone cannot be decisive for a more potent RNAi response.
Overexpression of the other human Argonaute proteins, Ago1, Ago3, and Ago4,
enhances miRNA expression, as well (Diederichs and Haber 2007), but RNAi
efficiency was not altered (Diederichs et al. 2008). Hence, a property unique to
Ago2 seems to account for RNAi enhancement. Ago2 differs from the other argo-
naute proteins by its intrinsic endonuclease activity, which operates at two different
levels: during miRNA processing and in the effector phase. Certain pre-miRNAs
with high complementarity are cleaved by Ago2 in their passenger strand creating an
ac-pre-miRNA intermediate, whereby strand dissociation is thought to be facilitated
(Matranga et al. 2005; Rand et al. 2005; Diederichs and Haber 2007). This function in
processing seems to be important for enhancing RNAi as the effect is more
pronounced when shRNAs, which undergo this cleavage event, are used instead
of siRNAs, which enter the pathway at one of the last steps (Diederichs et al.
2008). In the effector phase, the endonuclease activity of Ago2 mediates cleavage
of the target mRNA paired with the small RNA (Meister et al. 2004). Target mRNA
cleavage and subsequent degradation might result in a more potent RNAi response
than translation inhibition as the target is irrevocably destroyed. In contrast, mRNAs,

whose translation is just attenuated, can be reused for protein synthesis at a later time
point. Overexpression of Argonaute proteins, which prevent translation as well as
overexpression of RNase-deficient Ago2 mutants, diminish the potency of RNAi,
which speaks for a competition between argonaute proteins for the mature miRNA.
The more mature miRNA is assembled into the most potent Argonaute protein, Ago2,
the more efficient is the silencing effect. The necessity of perfect complementarity for
mRNA target cleavage explains why Ago2-mediated enhanced RNAi is highly
specific and therefore reduces the risk of off-target effects.
2.5 microRNA Biogenesis in Health and Disease:

Basis for RNAi Therapy
RNAi has not only become an important scientific tool but holds great potential
for therapeutic applications due to its potency and specificity to inhibit gene
expression. Diseases that could be treated by an RNAi approach are, e.g., cancer
and infections caused by viruses such as the human immunodeficiency virus (HIV)
or the hepatitis C virus (HCV). Although initial successful experiments in cell
culture show promise, several substantial hindrances have to be overcome before
the first RNAi based therapeutics will be available.
The first important task is to find a good target. This is particularly problematic
for viral diseases in case the virus has a high mutation rate. Alternatively to
targeting the viral RNA directly, cellular cofactors required for infection could
be downregulated. However, this approach would also target noninfected cells
potentially resulting in adverse effects. Next to the potency and specificity of the
siRNA/shRNA for therapeutic targets, the intracellular delivery of the RNAi drug is
a major challenge, which will be discussed elsewhere.
In general, successful RNAi-based therapy approaches critically depend on our
comprehensive understanding of the miRNA pathway in health and disease. In
cancer cells, a global reduction of mature miRNA levels is observed (Lu et al. 2005;
Lee et al. 2008). One possible reason for this aberrant miRNA pattern might be
defective miRNA processing factors or components of the effector machinery
(Thomson et al. 2006). Consistently, cellular transformation and tumorigenesis
was promoted by knockdown of the miRNA processing factors Drosha, Dgcr8,
and Dicer (Kumar et al. 2007). In cases where cancer results from impaired miRNA
maturation, drugs based on RNAi are probably not suitable, as they require an
intact processing and effector machinery to function effectively. But not only
knowledge of miRNA regulation in pathological cases is crucial for medical
application. RNAi is such a powerful method because it benefits from an endoge-
nous machinery enhancing its potency and specificity. This inevitably means that a
detailed understanding of the complex regulatory network of expression and
maturation of miRNAs in general is a prerequisite for RNAi drug development.
Thus, the biggest advantage of RNAi is also its major challenge.
2.6 Concluding Remarks
Since the initial successful transfer of the naturally occurring miRNA phenomenon
to the practical application of RNAi, our knowledge on the biogenesis pathway of
small RNAs has immensely increased. Our deeper understanding enabled the
development of more efficient approaches by accounting for the requirements of
the endogenous machinery. Also, future progress in RNAi as genetic tool or in
clinical applications will go hand in hand with advances in our understanding of the
molecular mechanisms and the complex regulatory network of miRNA biogenesis
and action as both depend on the same machinery.
References
Baskerville S, Bartel DP (2005) Microarray profiling of microRNAs reveals frequent coexpression

with neighboring miRNAs and host genes. RNA 11:241–247
Behm-Ansmant I, Rehwinkel J, Doerks T et al (2006) mRNA degradation by miRNAs and GW182
requires both CCR4:NOT deadenylase and DCP1:DCP2 decapping complexes. Genes Dev
20:1885–1898
Bernstein E, Kim SY, Carmell MA et al (2003) Dicer is essential for mouse development. Nat
Genet 35:215–217
Bhattacharyya SN, Habermacher R, Martine U et al (2006) Relief of microRNA-mediated
translational repression in human cells subjected to stress. Cell 125:1111–1124
Bohnsack MT, Czaplinski K, Gorlich D (2004) Exportin 5 is a RanGTP-dependent dsRNA-
binding protein that mediates nuclear export of pre-miRNAs. RNA 10:185–191
Borchert GM, Lanier W, Davidson BL (2006) RNA polymerase III transcribes human microRNAs.
Nat Struct Mol Biol 13:1097–1101
Bortolin-Cavaille ML, Dance M, Weber M et al (2009) C19MC microRNAs are processed from
introns of large Pol-II, non-protein-coding transcripts. Nucleic Acids Res 37:3464–3473
Brennecke J, Stark A, Russell RB et al (2005) Principles of microRNA-target recognition. PLoS
Biol 3:e85
Brummelkamp TR, Bernards R, Agami R (2002) A system for stable expression of short interfering
RNAs in mammalian cells. Science 296:550–553
Cai X, Hagedorn CH, Cullen BR (2004) Human microRNAs are processed from capped, poly-
adenylated transcripts that can also function as mRNAs. RNA 10:1957–1966
Caplen NJ, Parrish S, Imani F et al (2001) Specific inhibition of gene expression by small double-
stranded RNAs in invertebrate and vertebrate systems. Proc Natl Acad Sci USA 98:9742–9747
Chatterjee S, Grosshans H (2009) Active turnover modulates mature microRNA activity in
Caenorhabditis elegans. Nature 461:546–549
Chen CZ, Li L, Lodish HF et al (2004) MicroRNAs modulate hematopoietic lineage differentiation.
Science 303:83–86
Chendrimada TP, Gregory RI, Kumaraswamy E et al (2005) TRBP recruits the Dicer complex to
Chu CY, Rana TM (2006) Translation repression in human cells by microRNA-induced gene
silencing requires RCK/p54. PLoS Biol 4:e210
Cook HA, Koppetsch BS, Wu J et al (2004) The Drosophila SDE3 homolog armitage is required
for oskar mRNA silencing and embryonic axis specification. Cell 116:817–829
Davis BN, Hilyard AC, Lagna G et al (2008) SMAD proteins control DROSHA-mediated
microRNA maturation. Nature 454:56–61
Microprocessor complex. Nature 432:231–235
Desterro JM, Keegan LP, Lafarga M et al (2003) Dynamic association of RNA-editing enzymes
with the nucleolus. J Cell Sci 116:1805–1818
Dickins RA, Hemann MT, Zilfou JT et al (2005) Probing tumor phenotypes using stable and
regulated synthetic microRNA precursors. Nat Genet 37:1289–1295
Diederichs S, Haber DA (2006) Sequence variations of microRNAs in human cancer: alterations in
predicted secondary structure do not affect processing. Cancer Res 66:6097–6104
Diederichs S, Haber DA (2007) Dual role for argonautes in microRNA processing and posttran-
scriptional regulation of microRNA expression. Cell 131:1097–1108
Diederichs S, Jung S, Rothenberg SM et al (2008) Coexpression of Argonaute-2 enhances RNA
interference toward perfect match binding sites. Proc Natl Acad Sci USA 105:9284–9289
Dietzl G, Chen D, Schnorrer F et al (2007) A genome-wide transgenic RNAi library for conditional
gene inactivation in Drosophila. Nature 448:151–156
Dye MJ, Gromak N, Proudfoot NJ (2006) Exon tethering in transcription by RNA polymerase II.
Mol Cell 21:849–859
Elbashir SM, Lendeckel W, Tuschl T (2001b) RNA interference is mediated by 21- and 22-nucleo-
tide RNAs. Genes Dev 15:188–200
Elbashir SM, Martinez J, Patkaniowska A et al (2001c) Functional anatomy of siRNAs for
Eulalio A, Behm-Ansmant I, Izaurralde E (2007a) P bodies: at the crossroads of post-transcriptional
pathways. Nat Rev Mol Cell Biol 8:9–22
Eulalio A, Rehwinkel J, Stricker M et al (2007b) Target-specific requirements for enhancers of
decapping in miRNA-mediated gene silencing. Genes Dev 21:2558–2570
Eulalio A, Huntzinger E, Izaurralde E (2008) Getting to the root of miRNA-mediated gene
silencing. Cell 132:9–14
Fang Y, Spector DL (2007) Identification of nuclear dicing bodies containing proteins for microRNA
biogenesis in living Arabidopsis plants. Curr Biol 17:818–823
Filippov V, Solovyev V, Filippova M et al (2000) A novel type of RNase III family proteins in
eukaryotes. Gene 245:213–221
Forman JJ, Legesse-Miller A, Coller HA (2008) A search for conserved sequences in coding
regions reveals that the let-7 microRNA targets Dicer within its coding sequence. Proc Natl
Acad Sci USA 105:14879–14884
Forstemann K, Tomari Y, Du T et al (2005) Normal microRNA maturation and germ-line stem cell
maintenance requires Loquacious, a double-stranded RNA-binding domain protein. PLoS Biol
3:e236
Forstemann K, Horwich MD, Wee L et al (2007) Drosophila microRNAs are sorted into function-
ally distinct argonaute complexes after production by dicer-1. Cell 130:287–297
Fukuda T, Yamagata K, Fujiyama S et al (2007) DEAD-box RNA helicase subunits of the Drosha
complex are required for processing of rRNA and a subset of microRNAs. Nat Cell Biol
9:604–611
Grimson A, Farh KK, Johnston WK et al (2007) MicroRNA targeting specificity in mammals:
determinants beyond seed pairing. Mol Cell 27:91–105
Guddeti S, Zhang DC, Li AL et al (2005) Molecular evolution of the rice miR395 gene family. Cell
Res 15:631–638
Guil S, Caceres JF (2007) The multifunctional RNA-binding protein hnRNP A1 is required for
processing of miR-18a. Nat Struct Mol Biol 14:591–596
Haase AD, Jaskiewicz L, Zhang H et al (2005) TRBP, a regulator of cellular PKR and HIV-1 virus
expression, interacts with Dicer and functions in RNA silencing. EMBO Rep 6:961–967
Hammond SM, Bernstein E, Beach D et al (2000) An RNA-directed nuclease mediates post-
transcriptional gene silencing in Drosophila cells. Nature 404:293–296
Han J, Lee Y, Yeom KH et al (2004a) The Drosha-DGCR8 complex in primary microRNA
processing. Genes Dev 18:3016–3027
Han MH, Goud S, Song L et al (2004b) The Arabidopsis double-stranded RNA-binding protein
HYL1 plays a role in microRNA-mediated gene regulation. Proc Natl Acad Sci USA
101:1093–1098
the Drosha–DGCR8 complex. Cell 125:887–901
Han J, Pedersen JS, Kwon SC et al (2009) Posttranscriptional crossregulation between Drosha and
DGCR8. Cell 136:75–84
Heo I, Joo C, Cho J et al (2008) Lin28 mediates the terminal uridylation of let-7 precursor
MicroRNA. Mol Cell 32:276–284
Hiraguri A, Itoh R, Kondo N et al (2005) Specific interactions between Dicer-like proteins and
HYL1/DRB-family dsRNA-binding proteins in Arabidopsis thaliana. Plant Mol Biol
57:173–188
Hwang HW, Wentzel EA, Mendell JT (2007) A hexanucleotide element directs microRNA
nuclear import. Science 315:97–100
Janowski BA, Huffman KE, Schwartz JC et al (2006) Involvement of AGO1 and AGO2 in
mammalian transcriptional silencing. Nat Struct Mol Biol 13:787–792
Kawahara Y, Zinshteyn B, Chendrimada TP et al (2007a) RNA editing of the microRNA-151
precursor blocks cleavage by the Dicer–TRBP complex. EMBO Rep 8:763–769
Kawahara Y, Zinshteyn B, Sethupathy P et al (2007b) Redirection of silencing targets by
adenosine-to-inosine editing of miRNAs. Science 315:1137–1140
Kawahara Y, Megraw M, Kreider E et al (2008) Frequency and fate of microRNA editing in
human brain. Nucleic Acids Res 36:5270–5280
Kawamata T, Seitz H, Tomari Y (2009) Structural determinants of miRNAs for RISC loading and
slicer-independent unwinding. Nat Struct Mol Biol 16:953–960
Kedde M, Strasser MJ, Boldajipour B et al (2007) RNA-binding protein Dnd1 inhibits microRNA
access to target mRNA. Cell 131:1273–1286
bias. Cell 115:209–216
Kim YK, Kim VN (2007) Processing of intronic microRNAs. EMBO J 26:775–783
Kim DH, Villeneuve LM, Morris KV et al (2006) Argonaute-1 directs siRNA-mediated transcriptional
gene silencing in human cells. Nat Struct Mol Biol 13:793–797
Knight SW, Bass BL (2001) A role for the RNase III enzyme DCR-1 in RNA interference and
germ line development in Caenorhabditis elegans. Science 293:2269–2271
Kumar MS, Lu J, Mercer KL et al (2007) Impaired microRNA processing enhances cellular
transformation and tumorigenesis. Nat Genet 39:673–677
Kurihara Y, Watanabe Y (2004) Arabidopsis micro-RNA biogenesis through Dicer-like 1 protein
functions. Proc Natl Acad Sci USA 101:12753–12758
Lagos-Quintana M, Rauhut R, Meyer J et al (2003) New microRNAs from mouse and human.
RNA 9:175–179
Lau NC, Lim LP, Weinstein EG et al (2001) An abundant class of tiny RNAs with probable
regulatory roles in Caenorhabditis elegans. Science 294:858–862
Lee CG, Hurwitz J (1992) A new RNA helicase isolated from HeLa cells that catalytically
translocates in the 30 to 50 direction. J Biol Chem 267:4398–4407
Lee Y, Jeon K, Lee JT et al (2002) MicroRNA maturation: stepwise processing and subcellular
localization. EMBO J 21:4663–4670
Lee Y, Ahn C, Han J et al (2003) The nuclear RNase III Drosha initiates microRNA processing.
Nature 425:415–419
Lee Y, Kim M, Han J et al (2004a) MicroRNA genes are transcribed by RNA polymerase II.
EMBO J 23:4051–4060
Lee YS, Nakahara K, Pham JW et al (2004b) Distinct roles for Drosophila Dicer-1 and Dicer-2 in
the siRNA/miRNA silencing pathways. Cell 117:69–81
Lee Y, Hur I, Park SY et al (2006) The role of PACT in the RNA silencing pathway. EMBO
J 25:522–532
Lee EJ, Baek M, Gusev Y et al (2008) Systematic evaluation of microRNA processing patterns in
tissues, cell lines, and tumors. RNA 14:35–42
Lewis BP, Burge CB, Bartel DP (2005) Conserved seed pairing, often flanked by adenosines,
indicates that thousands of human genes are microRNA targets. Cell 120:15–20
Li J, Yang Z, Yu B, Liu J et al (2005) Methylation protects miRNAs and siRNAs from a 30 -end
uridylation activity in Arabidopsis. Curr Biol 15:1501–1507
Lingel A, Simon B, Izaurralde E et al (2004) Nucleic acid 30 -end recognition by the Argonaute2
PAZ domain. Nat Struct Mol Biol 11:576–577
Lippman Z, Martienssen R (2004) The role of RNA interference in heterochromatic silencing.
Nature 431:364–370
Liu Q, Rand TA, Kalidas S et al (2003) R2D2, a bridge between the initiation and effector steps of
the Drosophila RNAi pathway. Science 301:1921–1925
Liu J, Valencia-Sanchez MA, Hannon GJ et al (2005) MicroRNA-dependent localization of
targeted mRNAs to mammalian P-bodies. Nat Cell Biol 7:719–723
Liu Y, Ye X, Jiang F et al (2009) C3PO, an endoribonuclease that promotes RNAi by facilitating
RISC activation. Science 325:750–753
Lobbes D, Rallapalli G, Schmidt DD et al (2006) SERRATE: a new player on the plant microRNA
scene. EMBO Rep 7:1052–1058
Lu J, Getz G, Miska EA et al (2005) MicroRNA expression profiles classify human cancers.
Nature 435:834–838
Lund E, Guttinger S, Calado A et al (2004) Nuclear export of microRNA precursors. Science
303:95–98
Ma JB, Ye K, Patel DJ (2004) Structural basis for overhang-specific small interfering RNA
recognition by the PAZ domain. Nature 429:318–322
Ma JB, Yuan YR, Meister G et al (2005) Structural basis for 50 -end-specific recognition of guide
RNA by the A. fulgidus Piwi protein. Nature 434:666–670
Ma E, MacRae IJ, Kirsch JF et al (2008) Autoinhibition of human dicer by its internal helicase
domain. J Mol Biol 380:237–243
MacRae IJ, Ma E, Zhou M et al (2008) In vitro reconstitution of the human RISC-loading complex.
Manche L, Green SR, Schmedt C, Mathews MB (1992) Interactions between double-stranded
RNA regulators and the protein kinase DAI. Mol Cell Biol 12:5238–5248
Matranga C, Tomari Y, Shin C et al (2005) Passenger-strand cleavage facilitates assembly of

siRNA into Ago2-containing RNAi enzyme complexes. Cell 123:607–620
Matzke MA, Birchler JA (2005) RNAi-mediated pathways in the nucleus. Nat Rev Genet 6:24–35
Meister G, Landthaler M, Patkaniowska A et al (2004) Human Argonaute2 mediates RNA
cleavage targeted by miRNAs and siRNAs. Mol Cell 15:185–197
Meister G, Landthaler M, Peters L et al (2005) Identification of novel argonaute-associated
proteins. Curr Biol 15:2149–2155
Michlewski G, Guil S, Semple CA et al (2008) Posttranscriptional regulation of miRNAs harboring
conserved terminal loops. Mol Cell 32:383–393
Minks MA, West DK, Benvin S et al (1979) Structural requirements of double-stranded RNA for
the activation of 20 , 50 -oligo(A) polymerase and protein kinase of interferon-treated HeLa cells.
J Biol Chem 254:10180–10183
Miyoshi K, Okada TN, Siomi H et al (2009) Characterization of the miRNA-RISC loading
complex and miRNA-RISC formed in the Drosophila miRNA pathway. RNA 15:1282–1291
Morlando M, Ballarino M, Gromak N et al (2008) Primary microRNA transcripts are processed
co-transcriptionally. Nat Struct Mol Biol 15:902–909
Morris KV, Chan SW, Jacobsen SE et al (2004) Small interfering RNA-induced transcriptional
gene silencing in human cells. Science 305:1289–1292
Mourelatos Z, Dostie J, Paushkin S et al (2002) miRNPs: a novel class of ribonucleoproteins
containing numerous microRNAs. Genes Dev 16:720–728
Newman MA, Thomson JM, Hammond SM (2008) Lin-28 interaction with the Let-7 precursor
loop mediates regulated microRNA processing. RNA 14:1539–1549
Nykanen A, Haley B, Zamore PD (2001) ATP requirements and small interfering RNA structure in
the RNA interference pathway. Cell 107:309–321
Obernosterer G, Leuschner PJ, Alenius M et al (2006) Post-transcriptional regulation of microRNA
expression. RNA 12:1161–1167
Okamura K, Ishizuka A, Siomi H et al (2004) Distinct roles for Argonaute proteins in small
RNA-directed RNA cleavage pathways. Genes Dev 18:1655–1666
Okamura K, Hagen JW, Duan H et al (2007) The mirtron pathway generates microRNA-class
regulatory RNAs in Drosophila. Cell 130:89–100
Okamura K, Phillips MD, Tyler DM et al (2008) The regulatory activity of microRNA* species
has substantial influence on microRNA and 30 UTR evolution. Nat Struct Mol Biol 15:354–363
Papp I, Mette MF, Aufsatz W et al (2003) Evidence for nuclear processing of plant micro RNA and
short interfering RNA precursors. Plant Physiol 132:1382–1390
Park MY, Wu G, Gonzalez-Sulser A et al (2005) Nuclear processing and export of microRNAs in
Arabidopsis. Proc Natl Acad Sci USA 102:3691–3696
Parker JS, Roe SM, Barford D (2005) Structural insights into mRNA recognition from a PIWI
domain-siRNA guide complex. Nature 434:663–666
Patterson JB, Samuel CE (1995) Expression and regulation by interferon of a double-stranded-
RNA-specific adenosine deaminase from human cells: evidence for two forms of the deami-
nase. Mol Cell Biol 15:5376–5388
Pedersen IM, Cheng G, Wieland S et al (2007) Interferon modulation of cellular microRNAs as an
antiviral mechanism. Nature 449:919–922
Persengiev SP, Zhu X, Green MR (2004) Nonspecific, concentration-dependent stimulation and
repression of mammalian gene expression by small interfering RNAs (siRNAs). RNA 10:12–18
Pfeffer S, Sewer A, Lagos-Quintana M et al (2005) Identification of microRNAs of the herpesvirus
family. Nat Methods 2:269–276
Pham JW, Pellino JL, Lee YS et al (2004) A Dicer-2-dependent 80s complex cleaves targeted
mRNAs during RNAi in Drosophila. Cell 117:83–94
Pillai RS, Bhattacharyya SN, Artus CG et al (2005) Inhibition of translational initiation by Let-7
MicroRNA in human cells. Science 309:1573–1576
Provost P, Dishart D, Doucet J et al (2002) Ribonuclease activity and RNA binding of recombinant
human Dicer. EMBO J 21:5864–5874
Qi HH, Ongusaha PP, Myllyharju J et al (2008) Prolyl 4-hydroxylation regulates Argonaute
2 stability. Nature 455:421–424
Ramachandran V, Chen X (2008) Degradation of microRNAs by a family of exoribonucleases in
Arabidopsis. Science 321:1490–1492
Rand TA, Petersen S, Du F, Wang X (2005) Argonaute2 cleaves the anti-guide strand of siRNA
during RISC activation. Cell 123:621–629
Rivas FV, Tolia NH, Song JJ et al (2005) Purified Argonaute2 and an siRNA form recombinant
human RISC. Nat Struct Mol Biol 12:340–349
Ro S, Park C, Young D et al (2007) Tissue-dependent paired expression of miRNAs. Nucleic
Acids Res 35:5944–5953
Robb GB, Rana TM (2007) RNA helicase A interacts with RISC in human cells and functions in
RISC loading. Mol Cell 26:523–537
Robb GB, Brown KM, Khurana J et al (2005) Specific and potent RNAi in the nucleus of human
cells. Nat Struct Mol Biol 12:133–137
Rodriguez A, Griffiths-Jones S, Ashurst JL et al (2004) Identification of mammalian microRNA
host genes and transcription units. Genome Res 14:1902–1910
Nature 448:83–86
Rybak A, Fuchs H, Smirnova L et al (2008) A feedback loop comprising lin-28 and let-7 controls
pre-let-7 maturation during neural stem-cell commitment. Nat Cell Biol 10:987–993
Ryman K, Fong N, Bratt E et al (2007) The C-terminal domain of RNA Pol II helps ensure that
editing precedes splicing of the GluR-B transcript. RNA 13:1071–1078
Saito K, Ishizuka A, Siomi H et al (2005) Processing of pre-microRNAs by the Dicer-1-Loqua-
cious complex in Drosophila cells. PLoS Biol 3:e235
Salzman DW, Shubert-Coleman J et al (2007) P68 RNA helicase unwinds the human let-7
microRNA precursor duplex and is required for let-7-directed silencing of gene expression.
J Biol Chem 282:32773–32779
Sasaki T, Shiohama A, Minoshima S et al (2003) Identification of eight members of the Argonaute
family in the human genome small star, filled. Genomics 82:323–330
Scadden AD (2005) The RISC subunit Tudor-SN binds to hyper-edited double-stranded RNA and
promotes its cleavage. Nat Struct Mol Biol 12:489–496
Seitz H, Ghildiyal M, Zamore PD (2008) Argonaute loading improves the 50 precision of both
MicroRNAs and their miRNA strands in flies. Curr Biol 18:147–151
Sen GL, Blau HM (2005) Argonaute 2/RISC resides in sites of mammalian mRNA decay known
as cytoplasmic bodies. Nat Cell Biol 7:633–636
Silva JM, Li MZ, Chang K et al (2005) Second-generation shRNA libraries covering the mouse
and human genomes. Nat Genet 37:1281–1288
Siolas D, Lerner C, Burchard J et al (2005) Synthetic shRNAs as potent RNAi triggers. Nat
Biotechnol 23:227–231
Song JJ, Liu J, Tolia NH et al (2003) The crystal structure of the Argonaute2 PAZ domain reveals
an RNA binding motif in RNAi effector complexes. Nat Struct Biol 10:1026–1032
Song L, Han MH, Lesicka J et al (2007) Arabidopsis primary microRNA processing proteins
HYL1 and DCL1 define a nuclear body distinct from the Cajal body. Proc Natl Acad Sci USA
104:5437–5442
Stegmeier F, Hu G, Rickles RJ et al (2005) A lentiviral microRNA-based system for single-copy

polymerase II-regulated RNA interference in mammalian cells. Proc Natl Acad Sci USA
102:13212–13217
Thomson JM, Newman M, Parker JS et al (2006) Extensive post-transcriptional regulation of
microRNAs and its implications for cancer. Genes Dev 20:2202–2207
Tokumaru S, Suzuki M, Yamada H et al (2008) let-7 regulates Dicer expression and constitutes a
negative feedback loop. Carcinogenesis 29:2073–2077
Tomari Y, Du T, Haley B et al (2004a) RISC assembly defects in the Drosophila RNAi mutant
armitage. Cell 116:831–841
Tomari Y, Matranga C, Haley B et al (2004b) A protein sensor for siRNA asymmetry. Science
306:1377–1380
Tomari Y, Du T, Zamore PD (2007) Sorting of Drosophila small silencing RNAs. Cell
130:299–308
Trabucchi M, Briata P, Garcia-Mayoral M et al (2009) The RNA-binding protein KSRP promotes
the biogenesis of a subset of microRNAs. Nature 459:1010–1014
Triboulet R, Chang HM, Lapierre RJ et al (2009) Post-transcriptional control of DGCR8 expres-
sion by the Microprocessor. RNA 15:1005–1011
Viswanathan SR, Daley GQ, Gregory RI (2008) Selective blockade of microRNA processing by
Lin28. Science 320:97–100
Wang Y, Juranek S, Li H et al (2008a) Structure of an argonaute silencing complex with a seed-
containing guide DNA and target RNA duplex. Nature 456:921–926
Wang Y, Sheng G, Juranek S et al (2008b) Structure of the guide-strand-containing argonaute
silencing complex. Nature 456:209–213
Wang Y, Juranek S, Li H et al (2009) Nucleation, propagation and cleavage of target RNAs in Ago
silencing complexes. Nature 461:754–761
Weinmann L, Hock J, Ivacevic T et al (2009) Importin 8 is a gene silencing factor that targets
argonaute proteins to distinct mRNAs. Cell 136:496–507
Wu L, Belasco JG (2008) Let me count the ways: mechanisms of gene regulation by miRNAs and
siRNAs. Mol Cell 29:1–7
Wu H, Xu H, Miraglia LJ et al (2000) Human RNase III is a 160-kDa protein involved in
preribosomal RNA processing. J Biol Chem 275:36957–36965
Wu L, Fan J, Belasco JG (2006) MicroRNAs direct rapid deadenylation of mRNA. Proc Natl Acad
Sci USA 103:4034–4039
Xie Z, Johansen LK, Gustafson AM et al (2004) Genetic and functional diversification of small
RNA pathways in plants. PLoS Biol 2:E104
Yang L, Liu Z, Lu F et al (2006a) SERRATE is a novel nuclear regulator in primary microRNA
processing in Arabidopsis. Plant J 47:841–850
Yang W, Chendrimada TP, Wang Q et al (2006b) Modulation of microRNA processing and
expression through RNA editing by ADAR deaminases. Nat Struct Mol Biol 13:13–21
Yang Z, Ebright YW, Yu B et al (2006c) HEN1 recognizes 21-24 nt small RNA duplexes and
deposits a methyl group onto the 20 OH of the 30 terminal nucleotide. Nucleic Acids Res
34:667–675
Yi R, Doehle BP, Qin Y et al (2005) Overexpression of exportin 5 enhances RNA interference
mediated by short hairpin RNAs and microRNAs. RNA 11:220–226
Yu B, Bi L, Zheng B et al (2008) The FHA domain proteins DAWDLE in Arabidopsis and SNIP1
in humans act in small RNA biogenesis. Proc Natl Acad Sci USA 105:10073–10078
Yuan YR, Pei Y, Ma JB et al (2005) Crystal structure of A. aeolicus argonaute, a site-specific
DNA-guided endoribonuclease, provides insights into RISC-mediated mRNA cleavage. Mol
Cell 19:405–419
Zamore PD, Tuschl T, Sharp PA et al (2000) RNAi: double-stranded RNA directs the
ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals. Cell 101:25–33
Zeng Y, Cullen BR (2003) Sequence requirements for micro RNA processing and function in
human cells. RNA 9:112–123
Zeng Y, Cullen BR (2004) Structural requirements for pre-microRNA binding and nuclear export
by Exportin 5. Nucleic Acids Res 32:4776–4785
Zeng Y, Cullen BR (2005) Efficient processing of primary microRNA hairpins by Drosha requires
flanking nonstructured RNA sequences. J Biol Chem 280:27595–27603
Zeng Y, Cai X, Cullen BR (2005a) Use of RNA polymerase II to transcribe artificial microRNAs.
Methods Enzymol 392:371–380
Zeng Y, Yi R, Cullen BR (2005b) Recognition and cleavage of primary microRNA precursors by
the nuclear processing enzyme Drosha. EMBO J 24:138–148
Zeng Y, Sankala H, Zhang X et al (2008) Phosphorylation of Argonaute 2 at serine-387 facilitates
its localization to processing bodies. Biochem J 413:429–436
Zhang H, Kolb FA, Brondani V et al (2002) Human Dicer preferentially cleaves dsRNAs at their
termini without a requirement for ATP. EMBO J 21:5875–5885
Zhang H, Kolb FA, Jaskiewicz L et al (2004) Single processing center models for human Dicer and
bacterial RNase III. Cell 118:57–68
Zhang B, Pan X, Stellwag EJ (2008) Identification of soybean microRNAs and their targets. Planta
229:161–182
MicroRNAs in Epithelial Antimicrobial
Immunity
Jun Liu, Guoku Hu, Rui Zhou, Kristen M. Drescher, and Xian-Ming Chen
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
2 Abundant Expression of miRNAs in Epithelial Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
3 Regulation of miRNA Expression in Epithelial Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
4 MicroRNAs in the Regulation of Epithelial Antimicrobial Defense . . . . . . . . . . . . . . . . . . . . . . 360
4.1 MicroRNAs and Maintenance of Epithelial Barrier Integrity . . . . . . . . . . . . . . . . . . . . . . . 360
4.2 MicroRNAs and Regulation of Epithelial Intracellular Signaling Pathways . . . . . . . . 361
4.3 MicroRNAs and Expression of B7-Costimulatory Molecules in Epithelial Cells . . . 363
4.4 MicroRNAs in the Exosomes Released from Epithelial Cells . . . . . . . . . . . . . . . . . . . . . . 363
4.5 MicroRNAs-Mediated Antivirus Response in Epithelial Cells . . . . . . . . . . . . . . . . . . . . . 363
5 Conclusion and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Abstract The importance of microRNAs (miRNA) in regulating immunity is

becoming increasingly apparent as more miRNA targets are discovered and the
molecular mechanisms underlying miRNA gene regulation becomes unraveled.
miRNA are abundant and their expression is finely controlled in human epithelial
cells at skin and mucosal sites. Recent studies indicate that miRNAs appear to
regulate a diverse spectrum of epithelial cell functions including maintaining
epithelial barrier integrity, refining intracellular signaling, and controlling epithelial
immune responses to inflammatory stimuli and pathogens. Expanding our knowl-
edge of the role of miRNA in epithelial immunoregulation as well as identifying
miRNAs of pathogenetic significance will enhance our knowledge of epithelial
immunobiology and immunopathology. This chapter will review recent advances in
the identification and expression of epithelial cell miRNAs and highlight the
J. Liu, G. Hu, R. Zhou, K.M. Drescher, and X.-M. Chen (*)

Department of Medical Microbiology and Immunology, Creighton University Medical Center,
Omaha, NE 68178, USA
e-mail: xianmingchen@creighton.edu
356 J. Liu et al.
functional significance of miRNA expression in immunoregulation of epithelial

antimicrobial defense.
Keywords Epithelium Immunoregulation Infection MicroRNAs Noncoding

RNAs
Abbreviations
TLR Toll-like receptor

NLRs NOD-like receptors
MHC Major histocompatibility complex
miRNAs MicroRNAs
EMT Epithelial mesenchymal transition
PAMPs Pathogen-associated molecular patterns
MyD88 Myeloid differentiation primary response gene 88
NF-kB Nuclear factor-k B
TTP Tristetraprolin
IRAK1 Interleukin-1 receptor-associated kinase 1
IL Interleukin
TRAF6 TNF receptor associated factor 6
CCL5 C–C motif ligand 5
IFN Interferon
CIS Cytokine-inducible Src homology 2-containing protein
SOCS Suppressors of cytokine signaling protein
TIR Toll/IL-1 receptor
LPS Lipopolysaccharide
VSV Vesicular stomatitis virus
CHC Chronic hepatitis C
1 Introduction
Epithelial cells at skin and mucosal sites represent the host’s first line of defense
against microbial infection. Beyond their role in creating a physical barrier to
potential disease agents, epithelial cells are critical in the initiation, regulation,
and resolution of both innate and adaptive immune responses to infection (Sansonetti
2004; Viswanathan and Hecht 2000). Epithelial cells express several pathogen-
sensing receptors, such as Toll-like receptors (TLRs) and NOD-like receptors
(NLRs). Through these receptors, epithelial cells recognize pathogen infection
MicroRNAs in Epithelial Antimicrobial Immunity 357
and evoke diverse responses including antimicrobial peptide and cytokine release
(Akira and Takeda 2004). Epithelial cells also express a wide range of proteins
critical to both the innate and adaptive immune responses, including major histo-
compatibility complex (MHC) class I and II, B7 costimulatory molecules, chemo-
kines, cytokines, and prostaglandins, which work in concert to eliminate microbes
(Viala et al. 2004; Menendez and Brett 2007). The epithelial immune response to
infection is finely controlled and reflects a delicate balance between effector
functions and the potential of the immune response to cause damage to healthy
tissue (Han and Ulevitch 2005; Shibolet and Podolsky 2007; Michael et al. 2006).
The importance of microRNAs (miRNA) in regulating normal cellular functions
is becoming increasingly clear as more miRNA targets are discovered and the
molecular mechanisms underlying miRNA gene regulation becomes unraveled
(Bartel 2009). Recent studies implicate specific miRNAs in controlling regulation
of cellular differentiation, determination of cell fate (cell death and proliferation),
initiation and regulation of antimicrobial immunity, control of inflammatory
responses, and activation of intracellular signaling pathways in epithelial cells
(Moschos et al. 2007; Chen et al. 2007; Pedersen et al. 2007; Gong et al. 2009;
Zhai et al. 2008; Yi et al. 2006; Andl et al. 2006; Harris et al. 2006; Hino et al. 2007;
Lu et al. 2007; Yi et al. 2008; Otsuka et al. 2007; Jopling et al. 2005; Sarasin-
Filipowicz et al. 2009). Because these functions are also involved in the host’s
response to infection, it is likely that miRNAs modify the epithelial immune
response to permit optimal responses to pathogens. This chapter summarizes recent
advances in the identification and expression of epithelial cell miRNAs and highlights
the functional significance of miRNA expression in immunoregulation of epithelial
antimicrobial defense, including maintenance of epithelial barrier integrity, regulation
of intracellular signaling pathways, and epithelial antiviral defense. Major findings in
this emerging field are summarized in Table 1.
2 Abundant Expression of miRNAs in Epithelial Cells
Studies on numerous human cell lines derived from normal epithelial cells demon-
strated the abundant expression of miRNA molecules in epithelial cells (Chen et al.
2007; Mattick and Makunin 2005; Liu et al. 2004; Krichevsky et al. 2003). As an
example, over 70 miRNAs were detected in H69 cells, a line of human biliary
epithelial cells (Chen et al. 2007). As expected, the miRNA expression profiles
are distinct among epithelial cells from different tissues as demonstrated by micro-
array technology and validated by quantitative real-time PCR approach
(Krichevsky et al. 2003). MicroRNA-122, a tissue-specific miRNA in hepatocytes
(the major epithelial cell type in the liver), has been found to play essential roles in
diverse hepatic functions including tumorigenesis, antiviral responses, and choles-
terol biosynthesis (Girard et al. 2008). Unique miRNA expression profiles have also
been described in normal epithelial cells from lung, breast, stomach, prostate,
colon, and pancreas (Calin and Croce 2006; Lu et al. 2005; Volinia et al. 2006).
358 J. Liu et al.
Table 1 Experimental charactering of miRNAs in epithelial antimicrobial immunity

MicroRNAs Potential functions in epithelial cells References
miR-338-3p, miR-451 Control the polarity establishment in (Tsuchiya et al. 2009)
epithelial cells
miR-203 Involved in skin morphogenesis by (Yi et al. 2008)
targeting p65
miR-146a/b Modulation of cytokine production in (Perry et al. 2008)
lung epithelial cells
miR-1, miR-30, miR-128, Inhibition of HCV replication by (Pedersen et al. 2007)
miR-196, miR-296, miR-351, targeting virus genome in
miR-431, miR-448 hepatocytes
Induced by interferon-b
miR-122 Permissive of HCV replication (Jopling et al. 2005;
in vitro Sarasin-Filipowicz
Positively correlated with interferon et al. 2009)
therapy
miR-24,miR-93 Inhibition of VZV replication by (Otsuka et al. 2007)
targeting virus genome
let-7 Promotes of C. parvum infection (Chen et al. 2007)
clearance in biliary epithelial cells
Targeting TLR4
miR-105 Regulation of TLR2 expression (Benakanakere et al.
2009)
miR-98 Modulation NF-kB activity via (Hu et al. 2009)
targeting CIS
miR-513 Regulation of immune response in (Gong et al. 2009)
biliary epithelial cells by targeting
B7-H1 costimulatory molecule
Decrease by interferon-g
Additionally, miRNA signatures in cancer cell lines of epithelial origin are distinct
from those in normal epithelial cells. The unique expression patterns of these
powerful gene regulators in a tissue- and differentiation-specific manner further
solidifies the critical nature of miRNAs in host homeostasis and defense from
pathogens.
3 Regulation of miRNA Expression in Epithelial Cells
Distinct miRNA expression signatures are observed in a tissue-specific manner,

suggesting that expression of these molecules is exquisitely controlled. Similar to
other RNA molecules, most miRNAs are initially transcribed as primary transcripts by
Poly II in the nucleus (Kim et al. 2009), supporting the hypothesis that miRNA genes
are transcriptionally regulated. Recent studies of human intergenic miRNA genes
structure have revealed potential transcriptional regulation of miRNAs by transcrip-
tion factors in different cell types including epithelial cells (Taganov et al. 2006).
Accordingly, it has been postulated that extracellular stimuli, such as cytokines,
chemokines, and pathogen infection, modulate transactivation of miRNA genes
through activation of intracellular signaling pathways (Taganov et al. 2006).
Triggering the downstream signaling pathways of the Toll/interleukin (IL)-1

receptor (TIR) superfamily members is a key step in initiating immune responses
in many cell types (Gottipati et al. 2008). Recent data suggest that TLR-mediated
and tissue type-dependent expression of specific miRNAs may be a common theme
in immune regulation. For example, miR-155 expression is regulated by TLR/
myeloid differentiation primary response gene 88 (MyD88)/nuclear factor-kB
(NF-kB) signaling in macrophages and monocytes (Tili et al. 2007), but not in
lung epithelial cells. Instead, a set of 46 miRNAs have been found to be upregulated
in lung-derived epithelial cells following lipopolysaccharide (LPS) treatment,
accompanied by a reduction in inflammatory cytokine production (Moschos et al.
2007). Alterations in miRNA expression are not solely limited to upregulation of
miRNA expression. Activation of the TLR/MyD88/NF-kB pathway has been
reported to downregulate certain miRNAs in some cell types as well. For example,
the activation of TLR/MyD88/NF-kB decreases let-7 expression in H69 cells
(Chen et al. 2007).
IL-1b has recently been reported to induce miR-146a expression in A549 cells, a
cell line derived from a human epithelial lung carcinoma, via NF-kB activation. IL-
1b-induced miR-146a expression was also confirmed in primary bronchial epithe-
lial cells and in a SV40 transformed bronchial BEAS-2B epithelial cell line.
Because induction of this miRNA was only observed at high IL-1b concentrations,
it is postulated that miRNA-146a participates in severe inflammation. Conversely,
following exposure to high dose IL-1b, at least six miRNAs including miR-26b,
miR-104, miR-195, miR-296, miR-299, and let-7g were downregulated in lung
alveolar epithelial cells.
Besides TIR-mediated signaling, cytokines and chemokines may also regulate
miRNA expression in epithelial cells during immune reactions. Several recent
studies have begun to address the impact of soluble mediators on miRNA expres-
sion. Treatment of human hepatocytes with interferon (IFN)-a induced significant
alterations in expression of 30 miRNAs (Pedersen et al. 2007). Specifically, miR-1,
miR-30, miR-128, miR-196, miR-296, miR-351, miR-431, and miR-448 were
upregulated. Conversely, several miRNAs, including miR-122, a liver-specific
miRNA, were downregulated. Recently, we examined the effect of IFN-g treatment
on miRNA expression in human biliary epithelial (H69) cells and demonstrated that
IFN-g treatment induced a universal downregulation of miRNA expression in these
cells (Gong et al. 2009). MicroRNA-mediated translational silencing is also compro-
mised in human bronchial epithelial BEAS-2B cells following stimulation with IL-4
and TNF-a (Zhai et al. 2008). Together, the current literature demonstrates the
responsiveness of miRNA expression to immune stimuli that are both pathogen-
(i.e., LPS) and host-derived (i.e., IL-4). Obviously, our current knowledge is too
limited to elucidate the roles of specific miRNAs during immune responses, particu-
larly in considering the redundant, synergistic, and antagonistic nature of the soluble
mediator network. Further studies in this area may shed novel insights regarding our
understanding of the subtle alterations required to appropriately regulate the immune
system’s response to challenges.
360 J. Liu et al.
4 MicroRNAs in the Regulation of Epithelial

Antimicrobial Defense
4.1 MicroRNAs and Maintenance of Epithelial Barrier Integrity
Epithelial barrier integrity is a key element in epithelial defenses against infection.

Epithelial polarity is essential to epithelial barrier integrity, and several miRNAs
have recently been described as crucial regulators of epithelial cell polarity and
development of epithelial–mesenchymal transition (EMT). EMT describes the
molecular reprogramming and phenotypic changes involved in the conversion of
polarized immotile epithelial cells to motile mesenchymal cells. This process
allows the remodeling of epithelial tissues during embryonic development and is
implicated in tumor metastasis, and more recently, in epithelial responses to
infection. Several studies have identified the miR-200 family and miR-205 as key
regulators of EMT and enforcers of the epithelial phenotype (Bracken et al. 2008).
The miR-200 family participates in a signaling network with the E-cadherin
transcriptional repressors ZEB1/dEF1 and ZEB2/SIP1, and TGF b2. This miRNA
family is not only postulated to facilitate maintenance of stable epithelial or
mesenchymal states but also allows reversible switching between these states in
response to EMT effectors (such as TGF b).
MicroRNA-203 is highly expressed in mature skin cells but minimally expressed
in skin precursor cells. To test the impact of aberrant (that is, premature) miR-203
expression in skin, transgenic mice were generated using the keratin 14 promoter to
drive miR-203 expression. These transgenic mice had significantly thinner back
skin epidermis than their wild-type littermates. Blocking studies in vitro using
primary cells suggest that p63 may be a target of miR-203. p63, an important
regulator of stem cell maintenance in stratified epithelial tissues, is not repressed in
the absence of either Dicer or miR-203 (Yi et al. 2008). The study also showed that
miR-203 induced cell-cycle exit and impaired proliferative potential of epithelial
stem cells, but had little effect on terminal differentiation marker expression.
Thus, miR-203 appears to function as a switch between proliferation and terminal
differentiation via a mechanism repressing progenitor cell proliferation. Given that
epithelial progenitor cells provide a continuous source of differentiated epithelial
cells, these studies highlight the importance of miRNAs in the maintenance
of homeostasis of various epithelial barriers (Hino et al. 2007; Lu et al. 2007;
Yi et al. 2008).
Utilizing an epithelial differentiation model with T84 cells, Tsuchiya et al.
demonstrated that miR-338-3p and miR-451 contribute to the formation of epi-
thelial basolateral polarity by facilitating translocalization of b1 integrin to the
basolateral membrane. Of the 250 miRNA screened in the study, the expression of
four miRNA (miR-33a, 210, 338-3p, and 451) were significantly elevated in the
differentiated stage of T84 cells, defined by when epithelial cell polarity was
established in the cells (Tsuchiya et al. 2009). Loss-of- and gain-of-function
analyses revealed that blockage of endogenous miR-338-3p or miR-451 using

miRNA-specific antisense oligonucleotides inhibited translocalization of b1
integrin to the basolateral membrane. Conversely, inhibition of miR-210 or
miR-33a had no effect on b1 integrin translocation. Simultaneous transfection
of synthetic miR-338-3p and miR-451 into T84 cells accelerated b1 integrin
translocation to the basolateral membrane. Therefore, both miR-338-3p and
miR-451 are necessary for the development of epithelial cell polarity, further
demonstrating a role for miRNAs in the development and maintenance of epithe-
lial cell polarity. A recent study examined the role of miR-29a, known to target
tristetraprolin (TTP), a protein involved in the degradation of messenger RNAs
with AU-rich 30 -untranslated regions in epithelial cell polarity. In these studies,
overexpression of miR-29a resulted in a loss of epithelial cell polarity in a
tumorigenic mouse mammary cell line (Gebeshuber et al. 2009).
Yi and colleagues provided novel insight into how miRNAs function in concert
in a tissue-specific manner (Yi et al. 2006). By comparing the unique miRNA
expression profiles in epidermis and hair follicles, miRNAs were classified into
groups according to the similarity of their 50 ends. Although miRNAs within a single
group may be independently transcribed from separate genes, their transcription
appeared to be coordinately expressed. For example, five members of the miR-200
family and four members of the miR-19/20 family were predominately expressed in
epidermis, while miR-199 family members were exclusively expressed in hair
follicles. The coexpression of multiple miRNAs with similar seed sequences from
the same lineage suggests that the miRNAs function together to permit effective
suppression of specific target genes within the cells and thus may be important in
maintaining tissue-specific characteristics.
4.2 MicroRNAs and Regulation of Epithelial Intracellular

Signaling Pathways
While the previous section described the impact of immune stimulation via either
pathogen-associated molecular patterns (PAMPs) or host immunomodulators on
miRNA expression, miRNAs also appear to impact signaling pathways directly.
Several recent studies reported that miRNAs control TLRs protein expression in
epithelial cells under certain conditions. MicroRNA-105 modulates TLR2 expres-
sion in human oral keratinocytes and let-7 targets TLR4 expression in human
cholangiocytes (Chen et al. 2007; Benakanakere et al. 2009). Cryptosporidium
parvum infection induced let-7 downregulation via the TLR/MyD88/NF-kB path-
way activation and enhanced TLR4 expression. Experimental manipulation
of let-7i expression caused reciprocal alterations in the infection dynamics of
C. parvum in vitro (Chen et al. 2007). Since TLRs recognize PAMPs and are key
modulators of epithelial cell immune responses to microbial infection, these data
support the critical role of miRNAs to host-cell regulatory responses to microbial
362 J. Liu et al.
infection. Downstream mediators of TIR signaling pathways are also regulated by

specific miRNAs. Potential targets of miR-146 are IL-1 receptor-associated kinase
1 (IRAK1) and TNF receptor-associated factor 6 (TRAF6) (Taganov et al. 2006),
both of which are key components of the TLR/NF-kB signaling pathway (Gottipati
et al. 2008; Chen 2005). Upregulation of miR-146 following LPS stimulation is
postulated to be a negative feedback regulator that inhibits TLR/NF-kB signaling
in macrophages and monocytes (Taganov et al. 2006). MicroRNA-146a was
also induced in human lung alveolar epithelial cells by exposure to high concentra-
tions of IL-1b (Perry et al. 2008). Transfection of the alveolar epithelial cells
with miR-146 precursor prior to cytokine exposure in vitro decreased IL-8 and
C-C motif ligand 5 (CCL5) release, indicating that miR-146 is likely involved in
regulating release of soluble mediators participating in the early phases of an
immune response, either directly or indirectly. Conversely, transfection of the
same cells with anti-miR-146a increased the release of these two cytokines follow-
ing stimulation with IL-1b. The effect of miR-146a on IL-1b-induced IL-8 and
CCL5 release in lung epithelial cells appears to be independent of both IRAK1 and
TRAF6. Transfection of these cells with miR-146a precursor did not decrease
IRAK1 or TRAF6 levels. Furthermore, although three genes involved in secretion
(syntaxin-3, synaptotagmin-1, and sec23 interacting protein) are also the predicted
targets for miR-146a, their protein levels were not affected by miR-146a precursor
transfection.
The cytokine-inducible Src homology 2-containing protein (CIS) and suppres-
sors of cytokine signaling (SOCS) proteins belong to a family of intracellular
proteins, which are key physiological regulators of cytokine responses in many
cell types, including epithelial cells. Each CIS/SOCS protein contains a SH2
domain and a SOCS box and functions in a classical negative-feedback loop that
inhibits cytokine signaling by interacting with the JAK-STAT signaling cascades
(Yoshimura et al. 2007). Epithelial cell CIS/SOCS expression is stimulated follow-
ing pathogen recognition via TLRs (Narayana and Balaji 2008). CIS/SOCS proteins
also provide negative feedback regulation to limit TLR signaling (Baetz et al 2004).
In human biliary epithelial cells, LPS stimulation or C. parvum infection induces
CIS at the posttranscriptional level by activating the TLR signaling pathway.
Further study revealed that miR-98 and let-7 target CIS and lead to translational
suppression of this protein. LPS stimulation and C. parvum infection activate TLR
signaling to downregulate miR-98 and let-7, which alleviates miR-98/let-7-
mediated CIS translational suppression thereby resulting in CIS protein expression.
In addition, induction of CIS expression is associated with positive-feedback
regulation of the NF-kB signaling pathway (Hu et al. 2009). These findings raise
the possibility that through miRNA-mediated posttranscriptional gene regulation,
TLR signaling may regulate expression of genes not activated at the transcriptional
level. These findings also implicate miRNA-mediated gene regulation as partici-
pants in CIS/SOCS expression to ensure finely controlled epithelial immunity
against microbial infection.
4.3 MicroRNAs and Expression of B7-Costimulatory

Molecules in Epithelial Cells
In addition to delivering intracellular signals, miRNAs are implicated in regulating

the expression of membrane proteins important in modulating epithelial immune
reactions. B7-H1, a member of the B7 costimulatory molecule family, is critical in
modulating cell-mediated immune responses, including epithelial–T cell interac-
tions (Gong et al. 2009). Our recent work showed that B7-H1 expression was
suppressed at the translational level in resting human biliary epithelial cells
in vitro and transfection of these cells with anti-miR-513 induced B7-H1 protein
expression. Downregulation of miRNA-513 was required for IFN-g-induced B7-H1
protein expression. Transfection of miR-513 precursor decreased IFN-g-induced
B7-H1 expression, demonstrating that miR-513 downregulation may be key to
IFN-g-induced B7-H1 induction. Moreover, transfection of biliary epithelial cells
with the miR-513 precursor inhibited B7-H1-associated apoptotic cell death in
cocultured human T cells, demonstrating the functional significance of miR-513 in
biliary epithelial cell–T cell interactions during an immune response. Interestingly,
C. parvum infection similarly decreases miR-513 expression to induce B7-H1
translation in H69 cells, suggesting that the release of translational repression on
B7-H1 via downregulation of miR-513 may be a common cellular response to
immune stimulation in biliary epithelial cells.
4.4 MicroRNAs in the Exosomes Released from Epithelial Cells
Exosomes are small membrane vesicles derived from multivesicular bodies or

endocytic-like lipid raft domains of the plasma membrane and are found in many
cell types including epithelial cells. Currently, it is postulated that exosomes
mediate cell–cell communication via exosomal shuttle of molecules (Valadi et al.
2007; Hunter et al. 2008). miRNA have been identified in exosomes released from
cultured mast cells, suggesting that exosome-mediated transport of miRNAs may
provide a novel mechanism of gene regulation between cells (Valadi et al. 2007).
Epithelial cell-secreted exosomes have been shown to express high levels of
MHC–peptide complexes that are capable of modulating immune responses
(B€uning et al. 2008; Kesimer et al. 2009). Given that miRNAs have been shown
to impact immune responses, it would be interesting to determine if exosomes from
epithelial cells also contain miRNAs and thus modulate epithelial–immune cell
interactions via exosomal delivery of miRNAs.
4.5 MicroRNAs-Mediated Antivirus Response in Epithelial Cells
The role of epithelial cell miRNAs in the control of microbial infections has
recently been investigated. Otsuka et al. demonstrated that Dicer knockout mice
364 J. Liu et al.
were highly susceptible to vesicular stomatitis virus (VSV) infection (Otsuka et al.
2007). To study the mechanism of this increase in VSV susceptibility, segments of
the VSV genome were fused to the 30 UTR of a luciferase reporter gene and the
resulting plasmid transfected into RAW 264.7 cells. Three VSV genome sequences
were identified that decreased reporter gene expression in these cells. Further
studies demonstrated that four miRNAs (miR-24, miR-93, miR-146, and miR-378)
were highly expressed by the RAW 264.7 cells with the potential to target VSV.
None of these specific miRNAs were detected in macrophages isolated from Dicer
knockout mice. Transfection of the RAW 264.7 cells with either anti-miR-24 or anti-
miR-93 resulted in 4- to 5-fold increase in virus titer, suggesting that miR-24 and
miR-93 are involved in controlling VSV replication in the host. Although it remains
to be determined if these miRNAs inhibit VSV replication in the virus’s natural hosts,
the abundant expression of these specific miRNAs at the site of VSV replication in
the epithelial layer suggests that miR-24 and miR-93 may participate in defense of the
epithelial barrier.
In hepatocytes, eight IFN-inducible miRNAs (miR-1, miR-30, miR-128, miR-
196, miR-296, miR-351, miR-431, and miR-448) have been shown to have near
perfect complementarities between their seed sequences and the hepatitis C virus
(HCV) RNA genome (Pedersen et al. 2007). Transfection of HCV replicon-containing
hepatocytes with precursors of these eight miRNAs decreased the levels of HCV
RNA accumulation in the cells. Functional inhibition of these particular miRNAs
abrogated the inhibitory effect of IFN on HCV replication in hepatocytes. IFN-b
treatment also decreased miR-122 expression, a liver specific miRNA essential
for HCV replication in hepatocytes (Jopling et al. 2005). The downregulation of
miR-122 in response to IFN-b further enhanced the antiviral effects of this cytokine.
Together, the data suggest a novel mechanism involving miRNA-mediated gene
targeting to fight HCV infection in hepatocytes (Pedersen et al. 2007).
The initial data derived from these in vitro studies are not compatible with some
in vivo data thus far. A very recent research examined miR-122 levels in liver
biopsies from 42 patients with chronic hepatitis C (CHC) undergoing IFN treatment
(Sarasin-Filipowicz et al. 2009). Pretreatment levels of miR-122 in nonresponders
were several times lower than miR-122 levels in responders. Given the results from
in vitro studies suggesting that miR-122 is crucial for efficient replication of HCV
in hepatocytes, this finding in CHC patients is unexpected and suggests that the
impact of miR-122 on HCV replication may be less pronounced in vivo than it is
in vitro, probably a result of the complex in vivo interactions that are difficult to
model in tissue culture.
5 Conclusion and Perspectives
The study of miRNAs is flourishing in the decade after their discovery. It is clear
that miRNAs have the potential to affect every aspect of cellular function, from cell
differentiation and proliferation to apoptotic death. miRNA appear to regulate a
diverse spectrum of epithelial cell functions including epithelial cell development,

refining intracellular signaling, and controlling epithelial immune responses to
inflammatory stimuli and pathogens. MicroRNA expression is a double-edged
sword, and aberrant miRNA expression has been implicated in pathogenesis of
various inflammatory diseases of the skin and mucosa. In the near future, distinct
miRNA signatures involved in fine-control of intracellular signaling and expression
of proteins, including antimicrobial peptides, cytokines, and chemokines, adhesion,
and costimulatory molecules, should further define the role of miRNAs in epithelial
immune responses. In addition, identification of miRNAs of major pathogenic
significance in persistent inflammatory reactions of the skin and at mucosal sites
could provide rationale for the design and implementation of new immunothera-
peutic strategies for treatment of these diseases. Unraveling the regulatory circuits
of miRNAs in epithelial biology is in its infancy but will likely yield new insights
into our understanding of epithelial immunobiology and immunopathology.
References
Akira S, Takeda K (2004) Toll-like receptor signaling. Nat Rev Immunol 4:499–511
Andl T, Murchison EP, Liu F et al (2006) The miRNA-processing enzyme dicer is essential for the
morphogenesis and maintenance of hair follicles. Curr Biol 16:1041–1049
Baetz A, Frey M, Heeg K et al (2004) Suppressor of cytokine signaling (SOCS) proteins indirectly
regulate toll-like receptor signaling in innate immune cells. J Biol Chem 279:54708–54715
Bartel DP (2009) MicroRNAs: target recognition and regulatory functions. Cell 136:215–233
Benakanakere MR, Li Q, Eskan MA et al (2009) Modulation of TLR2 protein expression by a
miR-105 in human oral keratinocytes. J Biol Chem 284:23107–23115
Bracken CP, Gregory PA, Khew-Goodall Y et al (2008) MicroRNAs as regulators of epithelial–
mesenchymal transition. Cell Cycle 7:3112–3118
B€uning J, von Smolinski D, Tafazzoli K et al (2008) Multivesicular bodies in intestinal epithelial
cells: responsible for MHC class II-restricted antigen processing and origin of exosomes.
Immunology 125:510–521
Calin GA, Croce CM (2006) MicroRNA signatures in human cancers. Nat Rev Cancer 6:857–866
Chen ZJ (2005) Ubiquitin signalling in the NF-kappaB pathway. Nat Cell Biol 7:758–765
Chen XM, Splinter PL, O’Hara SP et al (2007) A cellular micro-RNA, let-7i, regulates Toll-like
receptor 4 expression and contributes to cholangiocyte immune responses against Cryptospo-
ridium parvum infection. J Biol Chem 282:28929–28938
Gebeshuber CA, Zatloukal K, Martinez J et al (2009) MiR-29a suppresses tristetraprolin, which is
a regulator of epithelial polarity and metastasis. EMBO Rep 10:400–405
Girard M, Jacquemin E, Munnich A et al (2008) MiR-122, a paradigm for the role of microRNAs
in the liver. J Hepatol 48:648–656
Gong AY, Zhou R, Hu G et al (2009) MicroRNA-513 regulates B7-H1 translation and is involved
in IFN-gamma-induced B7-H1 expression in cholangiocytes. J Immunol 182:1325–1333
Gottipati S, Rao NL, Fung-Leung WP (2008) IRAK1: a critical signaling mediator of innate
immunity. Cell Signal 20:269–276
Han J, Ulevitch RJ (2005) Limiting inflammatory responses during activation of innate immunity.
Nat Immunol 6:1198–1205
Harris KS, Zhang Z, McManus MT et al (2006) Dicer function is essential for lung epithelium
morphogenesis. Proc Natl Acad Sci USA 103:2208–2213
366 J. Liu et al.
Hino K, Fukao T, Watanabe M (2007) Regulatory interaction of HNF1-alpha to microRNA-194

gene during intestinal epithelial cell differentiation. Nucleic Acids Symp Ser (Oxf) 51:415–416
Hu G, Zhou R, Liu J et al (2009) MicroRNA-98 and let-7 confer cholangiocyte expression of
cytokine-inducible Src homology 2-containing protein in response to microbial challenge.
J Immunol 183:1617–1624
Hunter MP, Ismail N, Zhang X et al (2008) Detection of microRNA expression in human
peripheral blood microvesicles. PLoS One 3:e3694
Jopling CL, Yi M, Lancaster AM et al (2005) Modulation of hepatitis C virus RNA abundance by a
liver-specific MicroRNA. Science 309:1577–1581
Kesimer M, Scull M, Brighton B et al (2009) Characterization of exosome-like vesicles released
from human tracheobronchial ciliated epithelium: a possible role in innate defense. FASEB
J 23:1858–1868
Kim VN, Han J, Siomi MC et al (2009) Biogenesis of small RNAs in animals. Nat Rev Mol Cell
Biol 10:126–139
Krichevsky AM, King KS, Donahue CP et al (2003) A microRNA array reveals extensive
regulation of microRNAs during brain development. RNA 9:1274–1281
Liu CG, Calin GA, Meloon B et al (2004) An oligonucleotide microchip for genome-wide
microRNA profiling in human and mouse tissues. Proc Natl Acad Sci USA 101:9740–9744
Nature 435:834–838
Lu Y, Thomson JM, Wong HY et al (2007) Transgenic over-expression of the microRNA
miR-17-92 cluster promotes proliferation and inhibits differentiation of lung epithelial
progenitor cells. Dev Biol 310:442–453
Mattick JS, Makunin IV (2005) Small regulatory RNAs in mammals. Hum Mol Genet 1:R121–132
Menendez A, Brett FB (2007) Defensins in the immunology of bacterial infections. Curr Opin
Immunol 19:385–391
Michael K, Toby L, Victor N (2006) Innate immunity gone awry: linking microbial infections to
chronic inflammation and cancer. Cell 124:823–835
Moschos SA, Williams AE, Perry MM et al (2007) Expression profiling in vivo demonstrates rapid
changes in lung microRNA levels following lipopolysaccharide-induced inflammation but not
in the anti-inflammatory action of glucocorticoids. BMC Genomics 8:240
Narayana Y, Balaji KN (2008) NOTCH1 up-regulation and signaling involved in Mycobacterium
bovis BCG-induced SOCS3 expression in macrophages. J Biol Chem 283:12501–12511
Otsuka M, Jing Q, Georgel P et al (2007) Hypersusceptibility to vesicular stomatitis virus infection
in Dicer1-deficient mice is due to impaired miR24 and miR93 expression. Immunity
27:123–134
Pedersen IM, Cheng G, Wieland S et al (2007) Interferon modulation of cellular microRNAs as an
antiviral mechanism. Nature 449:919–922
Perry MM, Moschos SA, Williams AE et al (2008) Rapid changes in microRNA-146a expression
negatively regulate the IL-1beta-induced inflammatory response in human lung alveolar
epithelial cells. J Immunol 180:5689–5698
Sansonetti PJ (2004) War and peace at mucosal surfaces. Nat Rev Immunol 4:953–964
Sarasin-Filipowicz M, Krol J, Markiewicz I et al (2009) Decreased levels of microRNA miR-122
in individuals with hepatitis C responding poorly to interferon therapy. Nat Med 15:31–33
Shibolet O, Podolsky DK (2007) TLRs in the Gut IV. Negative regulation of Toll-like receptors
and intestinal homeostasis: addition by subtraction. Am J Physiol Gastrointest Liver Physiol
292:G1469–G1473
Taganov KD, Boldin MP, Chang KJ et al (2006) NF-kappaB-dependent induction of microRNA
miR-146, an inhibitor targeted to signaling proteins of innate immune responses. Proc Natl
Acad Sci USA 103:12481–12486
Tili E, Michaille JJ, Cimino A et al (2007) Modulation of miR-155 and miR-125b levels following
lipopolysaccharide/TNF-alpha stimulation and their possible roles in regulating the response to
endotoxin shock. J Immunol 179:5082–5089
Tsuchiya S, Oku M, Imanaka Y et al (2009) MicroRNA-338-3p and microRNA-451 contribute to

the formation of basolateral polarity in epithelial cells. Nucleic Acids Res 37:3821–3827
Valadi H, Ekström K, Bossios A et al (2007) Exosome-mediated transfer of mRNAs and micro-
RNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol 9:654–659
Viala J, Sansonetti P, Philpott DJ (2004) Nods and ‘intracellular’ innate immunity. C R Biol
327:551–555
Viswanathan VK, Hecht G (2000) Innate immunity and the gut. Curr Opin Gastroenterol
16:546–551
Volinia S, Calin GA, Liu CG et al (2006) A microRNA expression signature of human solid
tumors defines cancer gene targets. Proc Natl Acad Sci USA 103:2257–2261
Yi R, O’Carroll D, Pasolli HA et al (2006) Morphogenesis in skin is governed by discrete sets of
differentially expressed microRNAs. Nat Genet 38:356–362
Yi R, Poy MN, Stoffel M et al (2008) A skin microRNA promotes differentiation by repressing
‘stemness’. Nature 452:225–229
Yoshimura A, Naka T, Kubo M (2007) SOCS proteins, cytokine signalling and immune regulation.
Nat Rev Immunol 7:454–465
Zhai Y, Zhong Z, Chen CY et al (2008) Coordinated changes in mRNA turnover, translation, and
RNA processing bodies in bronchial epithelial cells following inflammatory stimulation. Mol
Cell Biol 28:7414–7426
Emerging Roles of Long Noncoding RNAs
in Gene Expression and Intracellular
Organization
Tetsuro Hirose
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
2 Intracellular Behaviors of ncRNAs Distinct from Those of mRNAs . . . . . . . . . . . . . . . . . . . . . 370
3 Unique Pathways for Long ncRNA Biogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
4 ncRNA Functions in the Regulation of Gene Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
4.1 Regulation of Transcription Factor Activity by Long ncRNAs . . . . . . . . . . . . . . . . . . . . . 374
4.2 ncRNA Transcription Affects Adjacent Gene Expression . . . . . . . . . . . . . . . . . . . . . . . . . . 376
4.3 ncRNA Recruits or Modulates Epigenetic Factors on the Chromosome . . . . . . . . . . . . 378
4.4 ncRNAs Regulate Gene Expression at Posttranscriptional Steps . . . . . . . . . . . . . . . . . . . 380
5 Structural Roles of ncRNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
6 ncRNAs in Biomedical Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
7 Future Directions for ncRNA Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
Abstract In the postgenomic era, we have learned that large numbers of RNAs that
do not code for proteins, the so-called noncoding RNAs (ncRNAs), are transcribed
from large portions of the intergenic regions in mammalian genomes. The
biological significance of these ncRNAs remains elusive. Although the research
is still limited, recent progress has revealed that several ncRNAs play important
roles in various steps of gene expression, including epigenetic chromatin regula-
tion, transcription, RNA processing, protein assembly, and transport. Novel ncRNA
functions in the organization of intracellular structures have also been reported.
Major ncRNA subsets are expressed in a tissue-specific manner and some are
induced by external stimuli. The expression of numerous ncRNAs is drastically
changed in some types of cancer cells, suggesting that ncRNAs may be involved in
disease as well as in physiological events. Thus, understanding ncRNA functions
T. Hirose
Biomedicinal Information Research Center, National Institute of Advanced Industrial Science
and Technology (AIST), 2-4-7 Aomi, Koutou, 135-0064, Tokyo, Japan
e-mail: tets-hirose@aist.go.jp
370 T. Hirose
and mechanisms of action will open new opportunities for developing RNA-based
technologies for pharmaceutical application.
Keywords Transcriptome Noncoding RNA Epigenetic regulation Nuclear body
1 Introduction
In the twenty-first century, postgenomic transcriptomic analyses, including full-

length cDNA sequencing and tiling array analyses, have revealed that large num-
bers of transcripts that are unlikely to code for polypeptides are produced from
regions covering large portions of the human and mouse genomes (Bertone et al.
2004; Carninci et al. 2005; Cheng et al. 2005; Katayama et al. 2005; Birney et al.
2007; Kapranov et al. 2007). This discovery led to a rethinking of the general
scheme of gene expression, the so-called central dogma in which the genes were
thought to primarily produce functional proteins. Indeed, the recent report from the
ENCODE project (Birney et al. 2007) estimated that 93% of the human genome is
transcribed into some kind of RNA, whereas only 2% of the whole human genome
codes for proteins (International Human Genome Sequencing Consortium 2004).
Therefore, most of the RNAs transcribed from the human genome must be nonpro-
tein coding transcripts. These transcripts are commonly termed “noncoding RNAs
(ncRNAs)”. The biogenesis pathways, intracellular localization, expression, and
function of most of these long ncRNAs remain largely enigmatic. In this chapter,
I focus on the functions and characteristics of a subset of ncRNAs that have been
recently annotated, although these may be just the tip of the ncRNA iceberg.
2 Intracellular Behaviors of ncRNAs Distinct from Those

of mRNAs
The transcriptome has primarily been analyzed using the sequences of full-length
cDNAs that were constructed from transcripts possessing both a cap structure at the
50 terminus and a poly(A) tail at the 30 terminus; therefore, the ncRNAs that
emerged through these analyses are likely to be products of RNA polymerase II.
The biogenesis pathway of protein-coding mRNAs that are produced by RNA
polymerase II has been studied for a long time and is well characterized (Fig. 1)
(Dreyfuss et al. 2002). It is an amazingly organized multistep process in which
quality control mechanisms eliminate aberrantly synthesized mRNAs that may
eventually produce harmful polypeptides. In contrast, the biogenesis pathways of
RNA polymerase II-produced ncRNAs remain to be investigated. To what extent
are ncRNAs subject to the rules of mRNA biogenesis (Fig. 1)? One form of quality
control in mRNA biogenesis, nonsense-mediated decay (NMD), recognizes aberrant
mRNAs to be eliminated by detecting the presence of a premature termination
Emerging Roles of Long Noncoding RNAs in Gene Expression and Intracellular 371
mRNA ncRNA
AAAAA
AAAAA
Nuc
Cyt
AAAAA
Functional polypeptide
Fig. 1 Comparison of the defined expression pathway for protein-coding genes and a putative
pathway for noncoding genes. The biogenesis pathway and intracellular localization of most
ncRNAs remain to be investigated. The mechanisms of ncRNA biogenesis that are distinct from
those for mRNA will be the focus of studies in the near future
codon in mRNAs (Chang et al. 2007; Isken and Maquat 2007). However, canonical
ncRNAs lack meaningful ORFs and instead have plenty of termination codons,
which is why they are called “noncoding” RNAs. Indeed, two ncRNAs, gas5 and
UHG, which are known to be host genes of small nucleolar RNAs (snoRNAs) are
rapidly degraded by the NMD pathway in mammalian cells (Tycowski et al. 1996;
Smith and Steitz 1998; Ideue et al. 2007). Because mammalian snoRNAs are
processed from excised introns, some of the spliced host genes are thought to be
nonfunctional RNAs that can be rapidly degraded in the cytoplasm. The degrada-
tion is arrested by either cycloheximide treatment or knockdown of a NMD factor,
UPF1, indicating that these ncRNAs are natural NMD targets (Tycowski et al.
1996; Smith and Steitz 1998; Ideue et al. 2007). Recently, a genome-wide search of
NMD targets in Arabidopsis plants revealed that 20% of known ncRNAs are
natural NMD targets, and the majority of these 20% are natural antisense ncRNAs
(Kurihara et al. 2009). The majority of NMD-insusceptible ncRNAs may avoid
entering the NMD pathway because ncRNA genes lack introns. ncRNAs are
retained in the nucleus and are never exported to the cytoplasm, and ncRNAs
would not be recognized by ribosomes even if they were transported to the
cytoplasm. A genome-wide analysis of gene structures for mRNAs and ncRNAs
in mouse revealed an apparent structural difference, in that 72% of the mapped
ncRNA sequences were uninterrupted by an intron, whereas only 18% of the
mRNAs were unspliced (Ravasi et al. 2006). As described below, numerous long
ncRNAs that have been functionally annotated play important biological roles in
nuclear events, suggesting that an ncRNA subpopulation is retained in the nucleus
and therefore would not encounter the NMD machinery in the cytoplasm. Indeed,
our data on 70 long ncRNAs selected from a human full-length cDNA database
(Sasaki et al. 2007) showed that >70% of them are predominantly localized to the
nuclear fraction and are therefore insusceptible to NMD (our unpublished results).
372 T. Hirose
3 Unique Pathways for Long ncRNA Biogenesis
Some long ncRNAs mature via unique RNA processing events that have not been
observed in the well-characterized mRNA biogenesis pathway (Srikantan et al.
2000; Dreyfuss et al. 2002; Kiyosawa et al. 2005; Ganesan and Rao 2008; Wilusz
et al. 2008). All eukaryotic mRNAs possess poly(A) tails at their 30 termini, with the
exception of histone mRNAs, which lack the poly(A) tail at their 30 termini and
instead possess conserved stem–loop structures that contribute to their cell cycle-
dependent regulation. The two long ncRNA genes for Malat-1 (multiple endocrine
neoplasia a [MENa]) and MENb adjoin each other at a chromosome 11 locus in
human and on the syntenic locus on chromosome 19 in mouse (Guru et al. 1997;
Hutchinson et al. 2007). The 30 terminus of the major Malat-1 transcript lacks a poly
(A) tail (Fig. 2a). The evolutionarily conserved tRNA-like structure located down-
stream of the 30 terminus is recognized and processed by the tRNA processing
enzyme RNase P in vitro, resulting in the creation of a nonpolyadenylated 30
terminus in the Malat-1 ncRNA (Fig. 2a) (Wilusz et al. 2008). Subsequently, a
similar tRNA-like structure was found in the downstream region of the 30 terminus
of the MENb ncRNA, which is transcribed from the chromosome locus adjacent to
the Malat-1 gene. The tRNA-like structure is also recognized and processed by
RNase P, creating a nonpolyadenylated 30 terminus in the MENb ncRNA (Sunwoo
et al. 2009). The cleaved downstream portion of the Malat-1 precursor containing
the tRNA-like structure is further processed by RNaseZ and CCA-adding enzyme
to produce a tRNA-like small RNA (mascRNA) that is subsequently transported to
the cytoplasm. The tRNA-like portion of the MENb precursor accumulates at low
levels, probably due to its unstable structure. The mechanism protecting the non-
polyadenylated 30 termini from RNA degradation has remained elusive. A U-rich
region located upstream of the 30 terminus of Malat-1 is critical for Malat-1’s stable
accumulation, raising the possibility that the U-rich region interacts with the
genomically encoded A-rich tract located at the 30 terminus (Wilusz et al. 2008).
This model is consistent with that proposed for the stable nuclear accumulation of
Kaposi’s sarcoma-associated herpesvirus polyadenylated nuclear RNA (Conrad
et al. 2006, 2007). Although the significance of these nonpolyadenylated 30 termini
of Malat-1 and MENb ncRNAs, as well as the function of mascRNA, remains
explainable, their presence is noteworthy as it represents a novel mechanism in
which the 30 termini of long ncRNAs that are RNA polymerase II products are
processed by a tRNA-processing enzyme to produce small RNAs. Because the long
ncRNAs have been mainly identified by large-scale transcriptome analyses in
which full-length cDNA clones derived from polyadenylated RNAs were sequenced,
long ncRNAs with nonpolyadenylated 30 termini may be missing from transcriptome
lists. A genome-wide analysis of natural antisense transcripts revealed that vast
amounts of antisense noncoding transcripts of various sizes are expressed in
mice, including large numbers of nonpolyadenylated and nuclear-localized ncRNAs
(Kiyosawa et al. 2005). Although we cannot rule out the possibility that the
detected nonpolyadenylated ncRNAs are transcriptionally nascent RNAs, the further
a
AAAAAAAAA
RNase P
Malat-1 (nuclear speckle)

AAAAAAAAA
mascRNA (cytoplasm)
AAAAA AAAAA
Fig. 2 New pathways for ncRNA biogenesis. (a) 30 end processing of Malat1 and MENb ncRNAs by
cleavage with RNase P. The downstream tRNA-like structure acts as a processing signal that is
recognized by RNase P. The tRNA-like region cleaved from the Malat-1 precursor is stably
accumulated as mascRNA and transported to the cytoplasm, whereas the mature Malat1 is exclu-
sively localized in the nucleus. (b) Posttranscriptional processing and capping of mature mRNAs and
ncRNAs. Extensive analysis of CAGE tags revealed the existence of small RNAs produced from
mature mRNAs and long ncRNAs, some of which have been spliced. The small RNAs may be
endonucleolytically processed, followed by the addition of a cap structure at their 50 termini
exploration of ncRNA termini may reveal unexpected structures in transcript termini

and associated processing mechanisms.
Hundreds of miRNAs are primarily transcribed as long RNA polymerase II
transcripts and then sequentially cleaved by RNase III-like Drosha and Dicer in
mammalian cells. The recently identified Piwi-interacting RNAs are likely to be
processed from the longer precursor RNAs. Therefore, some ncRNAs act as pre-
cursors of various small RNA species. The recently identified long ncRNA, meiotic
recombination hotspot locus (mrhl) RNA, is 2.4 kb in length and exclusively
localized to the nucleus. mrhlRNA is processed with Drosha to produce an 80-nt
nuclear-localized small RNA that is never processed further into smaller RNAs
(Ganesan and Rao 2008). This raises the possibility that the canonical premiRNA
processing pathway may be utilized for processing other ncRNAs as well.
Large numbers of small RNAs identified using next-generation sequencing
technology were found to significantly overlap with cap analysis of gene expression
(CAGE) tags, which are thought to mark the 50 ends of long RNA transcripts with a
374 T. Hirose
cap structure (Fig. 2b) (Fejes-Toth et al. 2009). Although many CAGE tags mark
transcription start sites, significant numbers were found in exonic regions and, in
some cases, even crossed splice junctions, indicating that they must have arisen
from at least partially processed mRNAs. Therefore, it has been proposed that
mature long transcripts, including both mRNAs and long ncRNAs, can be pro-
cessed posttranscriptionally to yield small RNAs, which are subsequently modified
by the addition of a cap structure (Fejes-Toth et al. 2009) (Fig. 2b).
4 ncRNA Functions in the Regulation of Gene Expression
Numerous ncRNAs are involved in various aspects of transcriptional regulation

through several distinct processes: the regulation of transcription factor activity, the
regulation of adjacent gene transcription by transcriptional interference, and the
recruitment of epigenetic regulatory factors onto chromosomes.
4.1 Regulation of Transcription Factor Activity by Long ncRNAs
The activity of transcription factors is controlled by various mechanisms, including

posttranslational modifications, assembly into active complexes, and subcellular
relocation. Several ncRNAs participate in these processes in mammalian cells. The
best-characterized long ncRNA is steroid receptor RNA coactivator (SRA), which
was first discovered in a cDNA expression screen for transcriptional coactivators of
the estrogen receptor (Lanz et al. 1999). SRA is 0.9 kb in length and forms stable
stem–loop structures that act to assemble a large transcriptional activation complex
containing the nuclear receptor and additional proteins (Lanz et al. 1999, 2002;
Hatchell et al. 2006; Colley and Leedman 2009). There is striking evidence from
biochemical studies that mPus1p, an RNA pseudouridine synthetase, is a coactivator
of the retinoic acid receptor that functions cooperatively with SRA (Zhao et al. 2004).
These biochemical studies have suggested that mPus1 pseudouridylates SRA
ncRNA, a posttranscriptional modification that may be important for transcriptional
activation by mRARg nuclear receptor. Although no other posttranscriptional
modifications of long ncRNAs have been identified, regulation of SRA by pseu-
douridylation raises the intriguing possibility that long ncRNAs are the targets of
various posttranscriptional modifications such as those observed in tRNAs,
snRNAs, and rRNAs.
Vertebrate Dlx genes are members of the homeodomain protein family that play
critical roles in differentiation and migration of neurons, as well as in craniofacial
and limb patterning, during development (Feng et al. 2006). The Dlx genes are
expressed in gene clusters, and ultraconserved intergenic enhancers have been
identified for the Dlx-5/6 and Dlx-1/2 loci (Ghanem et al. 2003; Yu et al. 2008).
Two ncRNA splicing isoforms, Evf-1 and Evf-2, are transcribed from the Dlx-5/6
region (Feng et al. 2006; Kohtz and Fishell 2004). Evf-1 is a 2.7-kb polyadenylated
RNA, and its expression is developmentally regulated (Kohtz and Fishell 2004).
The Evf-2 ncRNA (3.8 kb) specifically cooperates with the homeodomain protein
Dlx-2 to increase the transcriptional activity of the Dlx-5/6 enhancer region.
Interestingly, a stable complex containing the Evf-2 ncRNA/Dlx-2 homeodomain
protein forms in the nucleus (Feng et al. 2006). These data suggest that the Evf-2/
Dlx-2 complex stabilizes the interaction between Dlx-2 and its target Dlx-5/6
enhancer sequences to increase transcriptional activity. The possible role of Evf-2
produced from the genomic ultraconserved region suggests that other ultracon-
served regions may function to produce ncRNA regulators for key developmental
processes.
The NRON (noncoding repressor of NFAT) ncRNA was identified in an exten-
sive RNAi screening of about 500 long ncRNAs that were selected as evolutionarily
conserved ncRNAs from the FANTOM mouse full length cDNA database
(Willingham et al. 2005). The NRON ncRNA is alternatively spliced to produce
0.8–3.7 kb isoforms. Knockdown of NRON results in a dramatic activation of Ca++-
dependent NFAT (nuclear factor of activated T-cell) activity, suggesting that
NRON represses NFAT function. Characterization of NRON-binding proteins by
RNA affinity purification was used to identify the mechanism of NRON action and
members of the importin-b family, which likely function in NFAT nuclear trafficking,
were identified (Fig. 3a). Taken together, these findings show that NRON represses
NFAT function by capturing nuclear transporter proteins to prevent the nuclear
import of NFAT (Fig. 3a) (Willingham et al. 2005).
The heat-shock RNA-1 (HSR1) ncRNA was identified as a necessary factor for
activation of the heat-shock transcription factor 1 (HSF1) upon heat shock (Fig. 3b)
(Shamovsky et al. 2006). HSF1 has an important role in the heat-shock response in
vertebrates by inducing the expression of heat-shock proteins (HSPs) and other
cytoprotective proteins. HSF1 is present in unstressed cells in an inactive mono-
meric form and becomes activated by heat and other stress stimuli. HSF1 activation
involves trimerization to acquire site-specific DNA-binding activity, which is
negatively regulated by interaction with certain HSPs. HSF1 activation by heat
shock is mediated by a ribonucleoprotein complex containing the translation
elongation factor eEF1A and HSR1 ncRNA. Both HSR1 and eEF1A are required
for HSF1 activation in vitro, and the specific knockdown of HSR1 impairs the heat-
shock response, rendering cells thermosensitive (Fig. 3b).
In addition to specific transcription factors, RNA polymerase II is directly
targeted by ncRNAs (Allen et al. 2004; Espinoza et al. 2004; Mariner et al.
2008). Several SINE-encoded RNAs such as B2 in mouse and Alu in human
directly bind to RNA polymerase II during heat shock and globally inhibit
mRNA transcription. Thus, ncRNAs are involved in the various aspects of tran-
scriptional regulation by modulating the activities of transcription factors. The
ncRNAs described above are effectors for specific transcription factors, suggesting
that many more ncRNAs will be discovered that regulate the activity of hundreds of
transcription factors in mammalian cells.
376 T. Hirose
a b p23
Immunophilin
HSP90
HSF
Heat shock
p23
HSP90 Immunophilin
Ca++ signal
HSF
eEF1A
NFAT
P
NFAT
HSR1
HSF
calcineurin
NRON NFAT
importinb
HSF Active
c PRC2 complex
d
Long ncRNAs
IRES
Recruitment and
histone methylation
spliceosome
K27K27K27 IRES
3Me3Me3Me
ribosome
Fig. 3 Diverse functions of long ncRNAs involved in various processes of gene expression. (a)
NRON ncRNA captures importin-b and negatively controls the nuclear transport of the NFAT
transcription factor upon Ca++ signaling. (b) NSR1 ncRNA and the associated eEF1A facilitate the
trimerization of the HSF1 transcription factor upon heat shock. (c) An ncRNA subset including
Hotair, RepA, and Kcnqot1 act to recruit the histone modification complex to the specific target
region on the chromosome. (d) The antisense ncRNA regulates pre-mRNA splicing to produce the
translatable mRNA isoform possessing an IRES
4.2 ncRNA Transcription Affects Adjacent Gene Expression
In contrast to transcription factors, which act on promoters, transcription of an

ncRNA across the promoter region of a downstream protein-coding gene can
directly interfere with transcription factor binding, and thus prevent the protein-
coding gene from being expressed. This phenomenon, termed transcriptional inter-
ference, was first reported in the Saccharomyces cerevisiae SER3 gene (Martens
et al. 2004, 2005). The transcription of SER3 is tightly repressed during growth in
rich medium, whereas the upstream promoter region of this gene is highly tran-
scribed under these conditions and produces a nonprotein-coding RNA (SRG1).
Expression of the SRG1 RNA is required for the repression of SER3, and
the repression occurs by a transcription-interference mechanism in which SRG1
transcription across the SER3 promoter interferes with the binding of activators
(Fig. 4a).
a b
Basic transcription
factors
Transcription factors
SRG1 RNA DHFR

Triple helix
SER3
DHFR
Triple helix
c d
TLS/FUS I. Retention and function in cis
p300 CBP
CCND
II. Transcription event
III. Diffuse and act in trans
Fig. 4 Promoter-associated ncRNAs regulate transcription through distinct mechanisms. (a) Tran-
scriptional interference. The transcription of SRG ncRNA from the flanking region inhibits SER3
gene transcription in cis. (b). The transcript from the flanking promoter region of a DHFR gene
forms a triple helix with promoter DNA that leads to interference with the transcriptional initiation
of the DHFR gene. The ncRNAs act both in cis and in trans. (c). The ncRNAs that are transcribed
from the CCND promoter region bind to an RNA-binding protein, TLS/FUS, which inhibits CBP/
p300 histone acetyl transferase activity and is allosterically regulated by the bound ncRNAs.
(d). Three basic mechanisms of action of promoter-associated ncRNA. First, the produced ncRNA
is retained at the transcription site and functions only in cis. Second, transcription interference
inhibits transcription downstream. The transcription event itself is enough to inhibit the down-
stream gene in cis. Third, the produced ncRNA may diffuse and act on the promoter region both in
cis and in trans
Transcriptional interference mechanisms regulate key developmental decisions,

including the expression of homeotic ultrabithorax genes in Drosophila (Petruk
et al. 2006) and the switch to enter meiosis in S. cerevisiae (Hongay et al. 2006). In
mammalian cells, a few examples of transcriptional interference with varied
mechanisms of action have been found. ncRNA transcription induces heterochro-
matin formation at the human p15 tumor suppressor gene locus (Yu et al. 2008). In
this case, the heterochromatin status is maintained even after ncRNA synthesis is
turned off, suggesting that the transient expression of ncRNAs may be required to
set the chromosome status but are dispensable for its maintenance (Yu et al. 2008).
A long ncRNA transcribed from a region upstream of the major DHFR promoter
represses expression of the downstream DHFR protein-coding gene (Fig. 4b)
(Blume et al. 2003; Martianov et al. 2007). However, the ncRNA inhibits DHFR
expression both in cis and in trans by forming an RNA–DNA triple helix with the
378 T. Hirose
DHFR promoter and directly interacting with TFIIB, which results in the disruption
of the preinitiation complex at the DHFR promoter (Fig. 4a) (Martianov et al. 2007).
In contrast, ncRNA transcription positively regulates the expression of adjacent
genes. Recently, it was reported that transcription of long ncRNAs upstream of the
Schizosaccharomyces pombe fbp1+ locus induces chromatin remodeling that is
critical for transcriptional activation of the downstream protein-coding gene (Hirota
et al. 2008). In this case, ncRNA transcription is initiated in a stepwise manner from
multiple sites in the fbp1+ promoter, causing chromatin opening to proceed pro-
gressively toward the mRNA transcription start site. The insertion of a transcrip-
tional terminator within these ncRNA regions prevents downstream chromatin
remodeling, resulting in inefficient transcriptional induction because of the reduced
recruitment of transcription factors to the fbp1+ promoter. Noncoding transcription
also plays a role in activation of the S. cerevisiae PHO5 gene. A 2.4-kb antisense
ncRNA transcribed from near the 30 end of the yeast PHO5 gene acts to evict
histones from the PHO5 gene promoter during the repression under high phosphate
conditions, making it possible to respond rapidly to the signal for gene activation
(Uhler et al. 2007). In Drosophila, intergenic transcription through cis-acting
negative regulatory elements in the promoter regions prevents the silencing of
certain Hox genes by Polycomb group (PcG) proteins (Bender and Fitzgerald
2002; Hogga and Karch 2002; Schmitt et al. 2005; Preker et al. 2008). This seems
to be the opposite of transcriptional interference, where the transcription of ncRNAs
prevents binding of the positive transcription factors to the promoter regions.
An important question is whether chromatin remodeling is induced by the
transcription events that produce ncRNAs or by the produced ncRNAs themselves.
To answer this question, researchers examined whether the ncRNA can act in trans,
or whether knockdown of the accumulated ncRNAs disturbs the possible ncRNA’s
regulatory function. At least at the yeast Ser3, IME4, and PHO5 loci, and Drosophila
Ubx locus, it appears to be the act of noncoding transcription rather than the ncRNA
itself that contributes to the regulation of the expression of protein-coding genes
because these cognate ncRNAs cannot act in trans (Hongay et al. 2006; Uhler et al.
2007). On the other hand, at least at the human DHFR locus, the ncRNA does work
in trans (Martianov et al. 2007). Therefore, depending on the gene locus, ncRNA
transcription can act either negatively or positively. In some cases, the act of tran-
scription is sufficient to have functional consequences, but it is likely that many of the
ncRNAs play diverse regulatory roles with currently unknown mechanisms.
4.3 ncRNA Recruits or Modulates Epigenetic Factors

on the Chromosome
Increasing evidence indicates that epigenetic regulation is a major role of long

ncRNAs. For example, it was suggested that long ncRNAs recruit PcG proteins,
which bind to and silence the expression of more than a thousand mammalian genes
in specific genomic locations (Silva et al. 2003; Petruk et al. 2006). There are
precedents for ncRNA involvement in X-chromosome inactivation (XCI) in female
mammalian cells. The XCI center harbors two major long ncRNA genes, the 17-kb
Xist and its antisense repressor, the 40-kb Tsix (Masui and Heard 2006; Peters and
Robson 2008). Tsix blocks Xist expression and prevents the recruitment of silenc-
ing factors in cis on the future active X chromosome. In contrast, on the future
inactive X chromosome, Tsix is downregulated, leading to Xist expression and the
spread of Xist RNA along the chromosome (Masui and Heard 2006; Peters and
Robson 2008). The accumulation of Xist transcripts correlates with chromatin
alteration (Masui and Heard 2006; Payer and Lee 2008), but how Xist directs
these alterations is unknown. Recently, the catalytic subunit of Polycomb-repressive
complex 2 (PRC2) was found to directly bind an independent, shorter ncRNA derived
from Xist RNA called RepA (Zhao et al. 2008). The RepA ncRNA is transcribed
from the Repeat A region of the Xist gene and has been proposed to play a key role in
the early stages of mammalian XCI (Zhao et al. 2008).
Hotair is the only long ncRNA that recruits the PcG complex to another
chromosome locus in trans (Rinn et al. 2007). Hotair, a 2.2-kb ncRNA that is
transcribed from the HOXC locus, regulates the HOXD locus in trans by recruiting
the PRC2 complex (Rinn et al. 2007). Recently, Guttman et al. performed a
genome-wide search for histone H4K36 trimethylation marks in intergenic regions
and found that large numbers of large intergenic ncRNAs (lincRNAs) (Guttman
et al. 2009, Khalil et al. 2009) further showed that numerous lincRNAs are
associated with PRC2 and negatively regulate thousands of protein-coding genes.
Therefore, the interaction of a long ncRNA with the PRC2 complex is a general
epigenetic regulatory mechanism for recruiting chromosomal silencing factors to
specific chromosomal locations (Fig. 3c).
Trithorax group (TrxG) proteins counteract the silencing functions of PcG
proteins to maintain active transcription states. Certain ncRNAs from the Hox
loci were shown to interact directly with the histone methyltransferase Ash1
in vitro and were proposed to target TrxG proteins to chromatin (Sanchez-Elsner
et al. 2006). In fact, ectopic expression of these ncRNAs in trans was found to
activate Ubx gene expression (Sanchez-Elsner et al. 2006). On the other hand,
another report failed to observe a similar association between ectopic expression of
these ncRNAs and transcriptional activation, and instead suggested that Ubx gene
expression is inhibited by transcriptional interference from ncRNAs from the
flanking promoter region (Petruk et al. 2006; Payer and Lee 2008). Although
there are still some discrepancies in the literature, it appears that certain ncRNAs
play key roles in maintaining the active or inactive state of the chromosome by
modulating the recruitment of PcG and TrxG proteins.
Numerous long ncRNAs have been suggested to function as key players in
uniparental expression due to genomic imprinting (Petruk et al. 2006; Royo and
Cavaillé 2008). In the mouse placenta, the 108-kb nuclear-retained Airn ncRNA
is required for the paternal-specific silencing in cis of a 400-kb region that includes
the Slc22a2, Slc22a3, and Igf2r genes (Sleutels et al. 2002; Seidl et al. 2006).
Although Airn covers the imprinted locus on the paternal chromosome, Airn
380 T. Hirose
preferentially accumulates at the Slc22a3 promoter rather than localizing uniformly

at the entire imprinted domain (Nagano et al. 2008). Airn interacts with G9a, a
histone H3K9 methyltransferase, to silence the paternal Slc22a3 promoter. Airn is
also required for silencing Igf2r on the paternal chromosome, although by a
different mechanism. This G9a depletion results in the loss of Slc22a3 imprinting
but has no effect on Igf2r, which remains monoallelically expressed. Another
imprinted ncRNA, Kcnq1ot1, accumulates nonuniformly along the Kcnq1 genomic
locus and interacts with the G9a and PcG silencing factors (Pandey et al. 2008), as
well as acting through transcriptional interference (Kanduri et al. 2006; Mohammad
et al. 2008). The interaction of Kcnq1ot1 ncRNA with the silencing factors inacti-
vates the Kcnq1 region of the chromosome only in the placenta. Therefore, genes in
the Kcnq1 region are imprinted only in the placenta (Pandey et al. 2008).
Recent large-scale cDNA sequencing with the next-generation sequencers and
tiling array analyses revealed that large numbers of ncRNAs are transcribed from
the promoter regions of numerous protein-coding genes (Core et al. 2008; Preker
et al. 2008; Seila et al. 2008). In mammalian cells, the majority of these transcripts
are rapidly degraded by exosomes (Preker et al. 2008). The precise functions of
the promoter-associated ncRNAs remain largely elusive. Recently, the ncRNAs
produced from the cyclin D1 (CCND1) promoter region were shown to function as
allosteric effectors of an RNA-binding protein, TLS/FUS (Wang et al. 2008). These
ncRNAs are normally in low abundance (less than two copies per cell), but are
induced in response to DNA damage and remain associated with the chromatin in
the CCND1 promoter region. Upon association with these ncRNAs, the conforma-
tion of the TLS protein alters in order to bind to and inhibit the enzymatic activities
of the histone acetyltransferases CBP and p300, which consequently induces
silencing of CCND1 transcription (Fig. 4c). This complex mechanism may solely
be one example of the regulatory functions of promoter-associated ncRNAs.
What is notable is that the ncRNA utilizes its functional sequence in a way that
is obviously distinct from the function of the promoter-associated ncRNAs in
transcriptional interference.
4.4 ncRNAs Regulate Gene Expression at Posttranscriptional

Steps
Our knowledge of the posttranscriptional regulation of gene expression by long

ncRNAs remains limited. There are several examples of natural antisense tran-
scripts modulating the alternative splicing patterns of their overlapping genes
(Krystal et al. 1990; Munroe and Lazar 1991; Hastings et al. 2000; Yan et al.
2005). Recently, an additional example of a natural antisense RNA regulating
alternative splicing, and the biological significance of this regulation, was reported
for the Zeb2/Sip1 gene locus (Beltran et al. 2008). The translation of the Zeb/Sib1
mRNA requires an internal ribosome entry site (IRES). In epithelial cells, the
50 untranslated region (UTR) containing the IRES is spliced out of the mature
mRNA; however, upon the signal for the epithelial–mesenchymal transition, the
antisense RNA that is complementary to the 50 splice site of this intron is induced
and blocks the splicing of this intron (Fig. 3d). The resultant mature mRNA is able
to be translated into Zeb2/Sib1 protein through the IRES in the 50 UTR (Fig. 3d)
(Beltran et al. 2008).
Pseudogene transcripts act to regulate the levels of their homologous coding
mRNAs. The transcriptional reduction caused by transgene integration into the
vicinity of the expressed pseudogene of Makorin1 destabilizes the Makorin1
mRNA in trans via an RNA sequence element within the 50 region of Makorin1
that is homologous between Makorin1 and its pseudogene (Hirotsune et al. 2003).
These findings demonstrate a novel and specific regulatory role for an expressed
pseudogene as well as demonstrating additional functional significance for
ncRNAs.
Currently, it is recognized that miRNAs and the related small RNAs broadly
impact gene expression in various developmental and physiological conditions.
Long ncRNAs have been reported to modulate the function or biogenesis of such
small RNAs. Caenorhabditis elegans rncs-1 is an 800-nt, starvation-induced
ncRNA that is highly base-paired (Hellwig and Bass 2008). Rncs-1 ncRNA is
not a substrate for Dicer because of branched structures at its termini. Rncs-1
RNA inhibits Dicer cleavage of a second dsRNA in vitro, and the expression of
rncs-1 leads to varying mRNA levels of several Dicer-regulated genes in vivo
(Hellwig and Bass 2008). Arabidopsis thaliana IPS1 (Induced by Phosphate
Starvation 1) contains a motif with sequence complementarity to the phosphate
(Pi) starvation-induced miRNA miR-399, but the pairing is interrupted by a
mismatched loop at the expected miRNA cleavage site (Franco-Zorrilla et al.
2007). The IPS1 ncRNA is not cleaved, but instead sequesters miR-399. IPS1
overexpression results in increased accumulation of the miR-399 target, PHO2
mRNA, and concomitantly leads to reduced shoot Pi content. That is, IPS1 acts by
‘target mimicry’ to inhibit miRNA activity (Franco-Zorrilla et al. 2007). Thus,
rncs-1 and IPS1 ncRNAs both have the potential to act as molecular decoys by
competing with the actual RNA substrates. In the transcriptome data, large
numbers of ncRNA-like transcripts that overlap with mRNA 30 -UTRs have been
registered; these ncRNAs may be decoys for specific regulatory factors that
interact with the 30 UTR.
5 Structural Roles of ncRNAs
The mammalian cell nucleus contains more than ten membrane less suborganelles
that serve specialized functions (Lamond and Spector 2003; Misteli 2005). Recent
works suggest that long ncRNAs serve as key structural components in some of
these suborganelles. Paraspeckles are a relatively newly discovered subnuclear
structure with unknown function (Fox et al. 2002). Paraspeckles are observed
382 T. Hirose
as 10–20 granular foci in the interphase cell nucleus and contain numerous RNA-
binding proteins, including PSP1, PSP2/CoAA, p54/nrb, PSF, and the 68-kDa
subunit of cleavage factor I m (Fig. 5a) (Fox et al. 2002, 2005 Dettwiler et al.
2004). Interestingly, RNase A treatment disrupts the structural integrity of para-
speckles (Fox et al. 2005), suggesting that RNA is a critical component of these
nuclear structures. Recently, three groups independently identified MENe and -b as
the critical RNA components of paraspeckles (Clemson et al. 2009; Sasaki et al.
2009; Sunwoo et al. 2009). Two MEN isoforms, MENe/NEAT1 (3.7 kb) and
MENb (23 kb), are transcribed from a single RNA polymerase II promoter, but
differ in the location of their 30 end. The MENe/b depletion phenotype was
examined in human and mouse cells by knockdown with chimeric antisense
oligonucleotides (Sasaki et al. 2009; Sunwoo et al. 2009) or siRNA (Clemson
et al. 2009). MENe/b knockdown resulted in the disruption of the paraspeckles
but not other nuclear bodies, indicating that these long ncRNAs are required for
paraspeckle establishment and maintenance (Fig. 5a) (Clemson et al. 2009; Sasaki
et al. 2009; Sunwoo et al. 2009). Experiments examining the interaction between
MENe/b ncRNAs and the known paraspeckle proteins, and RNAi knockdown of
paraspeckle proteins, revealed the significance of the interaction of MENe/b
ncRNAs with at least two paraspeckle RNA-binding proteins, p54/nrb and PSF,
in the organization of paraspeckle structure (Fig. 5b) (Clemson et al. 2009;
a
knockdown control
PSF MENε/β merged
b
PSF
assembly
MENb
PSP1
p54
MENe
Fig. 5 Structural roles of ncRNAs. (a) MENe/b ncRNA (magenta) is specifically localized to the
nuclear paraspeckles where the RNA-binding protein PSF (green) is colocalized (merged). Upon
knockdown of MENe/b with ASO, the paraspeckle structures disintegrate (panels labeled “knock-
down”). (b) A model of paraspeckle organization achieved by cooperative association of ncRNAs
and RNA-binding proteins. MENb, the longer isoform, interacts with p54/nrb and PSF proteins to
form the core structure, then MENe, the shorter isoform, and PSP1 protein may bind to construct
the intact paraspeckle structure
Sasaki et al. 2009; Sunwoo et al. 2009). Physiological roles of paraspeckles have
not been well established, although it has been reported that the paraspeckles serve
as the nuclear-retention sites of mRNA subsets. The CTN-RNA is specifically
localized to paraspeckles and in response to external stimuli, it is endonucleolyti-
cally cleaved at its long 30 -UTR to produce a shorter mCat2 mRNA that is
transported to the cytoplasm, where it is then translated (Prasanth et al. 2005).
The CTN-RNA contains a long inverted repeat sequence in its 30 UTR that is A-to-I
edited. Artificial reporter mRNAs containing inverted repeated Alu elements in
their 30 UTRs are bound by a p54/nrb-containing complex, which prevents their
export to the cytoplasm and tends to localize them to the paraspeckles (Chen et al.
2008). The detailed mechanism of nuclear retention of mRNAs through the para-
speckles remains elusive, but a recent report showed that the MENe/b ncRNA-
dependent formation of intact paraspeckle structure is required for the extents of
nuclear retention of mRNA subsets that are usually retained in the nucleus (Chen and
Carmichael 2009). It would be most intriguing to pursue the roles of MENe/b
ncRNAs in mRNA nuclear retention.
Thermal and chemical stresses induce the formation of transient nuclear struc-
tures called nuclear stress bodies (nSBs) (Biamonti 2004). These contain HSF1 and
a specific subset of pre-mRNA splicing factors. nSBs are assembled on specific
pericentromeric heterochromatic domains containing satellite III (SatIII) DNA. In
response to stress, these domains change their epigenetic status from heterochro-
matin to euchromatin and are transcribed into polyadenylated SatIII RNAs that
remain associated with nSBs. Downregulation of SatIII RNAs significantly affects
the recruitment of RNA splicing factors to nSBs without altering the association of
HSF-1 with these structures. Thus, SatIII RNAs have a major role in the formation
of nSBs (Valgardsdottir et al. 2005).
RNA also has a structural role in the organization of the cytoskeleton and the
mitotic spindle. In Xenopus oocytes, the Xlsirts ncRNA and the VegT mRNA are
integrated within the cytoskeleton and are required for proper organization of the
cytokeratin cytoskeleton (Kloc et al. 2005, 2007). Downregulation of either tran-
script disrupts the cytokeratin network, but not the actin cytoskeleton. VegT mRNA
may act as an RNA itself, because blocking its translation had no effect on the
cytokeratin network (Heasman et al. 2001; Kloc et al. 2005). Meanwhile, the
mitotic spindles were found to associate with various RNAs, including ribosomal
RNAs and a number of uncharacterized transcripts (Blower et al. 2005). RNase A
treatment, but not translation inhibitors, disrupts spindle assembly and causes the
spindle to collapse, indicating that these RNA species play a role in spindle
assembly in M phase (Blower et al. 2005). A number of subcellular structures are
involved in RNA biogenesis and metabolism, and contain RNA components.
Malat-1 is localized exclusively in the splicing speckle, along with numerous
splicing factors (Hutchinson et al. 2007), and Gomafu, a neuron-specific long
ncRNA, is localized in a novel subnuclear structure that remains to be characterized
(Sone et al. 2007). Therefore, further examples will appear that long ncRNAs play
architectural roles in organizing intracellular structures.
384 T. Hirose
6 ncRNAs in Biomedical Research
Long ncRNAs have great potential to be sensitive molecular markers, because

many long ncRNAs are misregulated in various diseases, especially cancer (Prasanth
and Spector 2007). For example, the DD3/PCA3 and PCGEM1 ncRNAs are signifi-
cantly overexpressed in prostate tumors and therefore can be used for the diagnosis of
prostate cancer (Bussemakers et al. 1999; Srikantan et al. 2000; de Kok et al. 2002).
In nonsmall-cell lung cancer (NSCLC), metastasis is associated with increased
expression of the nuclear-localized Malat-1/MENa ncRNA (Ji et al. 2003). Indeed,
Malat-1 overexpression is a prognostic parameter for poor NSCLC patient survival
and can be used to identify early-stage NSCLC patients who are at risk of developing
metastases. In addition to NSCLC, recent studies have also reported the overexpres-
sion of Malat-1 in uterine endometrial stromal sarcoma and hepatocellular carcinoma
(Yamada et al. 2006; Lin et al. 2007). The expression of numerous other long
ncRNAs (e.g., H19, BC1, BC200, His1, NCRMS, and OCC1) has been reported to
be elevated in specific cancer cells. Because the exact physiological roles of these
ncRNA transcripts remain elusive, it is not known if they affect tumor initiation
and/or progression. Long ncRNAs remain an unexplored area in disease research;
hopefully, knowing more about them will allow us to identify new therapeutic targets.
7 Future Directions for ncRNA Research
Among the thousands of long ncRNAs whose sequences have been deposited in
databases, limited numbers have been functionally annotated. Future ncRNA
research should include the categorization of these RNA species based on common
features in their structures, functions, binding partners, or other characteristics.
Various bioinformatic approaches have been used to identify RNA sequence motifs
that may characterize ncRNA subsets. However, no sequence motifs that could be
used to categorize ncRNAs have been identified so far. Identification of the binding
partners of specific ncRNAs would also be a useful approach to categorize the
ncRNAs; however, there have been technical difficulties in the analysis of RNA–
protein interactions because of nonspecific associations in vitro that are distinct
from what is seen in DNA–protein interactions. Even in standard immunoprecipita-
tion experiments, the reassociation of RNA-binding proteins with RNA after cell
lysis can lead to misidentification of binding partners (Mili and Steitz 2004). To
circumvent these technical difficulties, a new method called CLIP has been devel-
oped. In this method, the in vivo RNA–protein (RNP) complexes are covalently
crosslinked by UV irradiation, followed by immunoprecipitation of the crosslinked
RNP complex (Ule et al. 2005). The CLIP method has been combined with high-
throughput DNA sequencing (HITS–CLIP) to globally identify the target RNA
sequences of certain RNA-binding proteins (Licatalosi et al. 2008; Chi et al. 2009).
For categorization of long ncRNAs, HITS–CLIP analysis will provide useful
information about ncRNA species that bind to a common RNA-binding protein and
about the sequences of the target motifs.
Another goal of ncRNA research is to identify physiological phenomena asso-
ciated with the action of a particular ncRNA. RNA interference (RNAi), currently
the most popular method for investigating RNA function, is used to knock down a
specific RNA, after which any phenotypic alterations are examined. Because
numerous long ncRNAs seem to function in nuclear processes, the knockdown
should be targeted to nuclear-localized ncRNA molecules. However, in mammalian
cells, the RNAi machinery is believed to exclusively localize in the cytoplasm,
making RNAi a poor choice to knock down nuclear ncRNAs, with a few exceptions
(Fig. 6a). Instead, antisense deoxyoligonucleotides (ASOs), which are usually
modified to increase their stability, are recognized as the most effective method
for nuclear RNA knockdown. The introduced ASOs form DNA–RNA hybrids that
are specifically recognized by endogenous RNase H, which degrades the RNA
strand of the hybrid (Fig. 6a and b). In the past, the introduction of ASO into the
nucleus was inefficient, resulting in ASO effectiveness that was often low, but
recently it was reported that ASO can be efficiently delivered into the nucleus
by using the nucleofection method, resulting in efficient depletion of various
nuclear-localized ncRNAs (Fig. 6c) (Ideue et al. 2009). As mentioned above, it is
a b
RNAi ASO
RNA
RNaseH
mRNA
Cytoplasmic ncRNA
decay
Nuclear ncRNA c
ASO ASO siRNA
FP 84 FP 4
G Δ G Δ8
6 24 6 24 6 24 6 24
New method to
Transcriptional interference
arrest transcription
U84
Fig. 6 The experimental system for the functional analysis of nuclear ncRNAs. (a) Canonical
RNAi is thought to occur exclusively in the cytoplasm of mammalian cells; therefore, an antisense
oligonucleotide (ASO) is utilized to knock down nuclear ncRNAs. Knockdown of the posttran-
scriptionally accumulated ncRNA may not inhibit transcriptional interference. A new method to
arrest the transcription of specific ncRNAs would be required to explore their roles in transcrip-
tional interference. (b). The principle of ASO action in the nucleus. The administered ASO forms
an RNA–DNA hybrid with the target RNA, in which the RNA strand is specifically cleaved by
endogenous RNase H. (c) An example of the knockdown of a nuclear ncRNA. U84 snoRNA is
efficiently knocked down with ASO but is not susceptible to siRNA
386 T. Hirose
noteworthy that nuclear ncRNA subsets are implicated in transcriptional inter-

ference, in which the RNA itself may be nonfunctional. To prevent the effect of
transcriptional interference, the expression of the regulatory ncRNAs needs to be
impeded at the transcription step. The technology for the specific and efficient arrest
of transcription of a specific gene, which will be necessary to explore the diverse
functions of nuclear ncRNAs, remains to be developed.
The number of papers published about long ncRNAs has increased in the past
few years, and there is no doubt that many researchers have begun to be interested
in the roles of these enigmatic molecules that are produced from mostly whole
genomic regions. Because there is a correlation between biological complexity and
the extent of the noncoding regions in the whole genome that produce ncRNAs,
one attractive model hypothesizes that ncRNAs have roles in the acquisition of
the complex biological phenomena seen in mammalian cells (Taft et al. 2007).
Consistent with this hypothesis, a long ncRNA termed HAR1F is transcribed from a
human-accelerated region of the human genome (Pollard et al. 2006). HAR1F is
specifically expressed in the Cajal–Retzius neurons of developing human neocortex
(Pollard et al. 2006), and it possesses a specific secondary structure that, unlike that
of its chimpanzee counterpart, resembles a cloverleaf structure (Beniaminov et al.
2008). These results suggest that HAR1F may play a role that is specific in human
brain.
Research on long ncRNAs is just now taking off and has begun to unveil the
diverse and important functions of long ncRNAs, from fundamental roles in gene
expression to higher-order physiological phenomena such as human-specific brain
functions and various diseases such as cancers. It is likely that additional diverse
ncRNA functions will be elucidated in the near future and can be expected to
provide new viewpoints for pharmaceutical applications, such as drugs targeted to
specific ncRNA motifs or to the interaction surface between ncRNAs and bound
proteins.
References
Allen TA, Von Kaenel S, Goodrich JA et al (2004) The SINE-encoded mouse B2 RNA represses
mRNA transcription in response to heat shock. Nat Struct Mol Biol 11:816–821
Beltran M, Puig I, Peña C et al (2008) A natural antisense transcript regulates Zeb2/Sip1
gene expression during Snail1-induced epithelial–mesenchymal transition. Genes Dev 22:
756–769
Bender W, Fitzgerald DP (2002) Transcription activates repressed domains in the Drosophila
bithorax complex. Development 129:4923–4930
Beniaminov A, Westhof E, Krol A (2008) Distinctive structures between chimpanzee and human
in a brain noncoding RNA. RNA 14:1270–1275
Bertone P, Stolc V, Royce TE et al (2004) Global identification of human transcribed sequences
with genome tiling arrays. Science 306:2242–2246
Biamonti G (2004) Nuclear stress bodies: a heterochromatin affair? Nat Rev Mol Cell Biol
5:493–498
Birney E, Stamatoyannopoulos JA, Dutta A et al (2007) Identification and analysis of functional
elements in 1% of the human genome by the ENCODE pilot project. Nature 447:799–816
Blower MD, Nachury M, Heald R et al (2005) A Rae1-containing ribonucleoprotein complex is

required for mitotic spindle assembly. Cell 121:223–234
Blume SW, Meng Z, Shrestha K et al (2003) The 50 -untranslated RNA of the human dhfr minor
transcript alters transcription pre-initiation complex assembly at the major (core) promoter.
J Cell Biochem 88:165–180
Bussemakers MJ, van Bokhoven A, Verhaegh GW et al (1999) DD3: a new prostate-specific gene,
highly overexpressed in prostate cancer. Cancer Res 59:5975–5979
Carninci P, Kasukawa T, Katayama S et al (2005) The transcriptional landscape of the mammalian
genome. Science 309:1559–1563
Chang YF, Imam JS, Wilkinson MF (2007) The nonsense-mediated decay RNA surveillance
pathway. Annu Rev Biochem 76:51–74
Chen LL, DeCerbo JN, Carmichael GG (2008) Alu element-mediated gene silencing. EMBO
J 27:1694–1705
Chen LL, Carmichael GG (2009) Altered nuclear retention of mRNAs containing inverted repeats
in human embryonic stem cells: functional role of a nuclear noncoding RNA. Mol Cell
35:467–478
Cheng J, Kapranov P, Drenkow J et al (2005) Transcriptional maps of 10 human chromosomes at
5-nucleotide resolution. Science 308:1149–1154
Chi SW, Zang JB, Mele A et al (2009) Argonaute HITS–CLIP decodes microRNA–mRNA
interaction maps. Nature 460:479–486
Clemson CM, Hutchinson JN, Sara SA et al (2009) An architectural role for a nuclear noncoding
RNA: NEAT1 RNA is essential for the structure of paraspeckles. Mol Cell 33:717–726
Colley SM, Leedman PJ (2009) SRA and its binding partners: an expanding role for RNA-binding
coregulators in nuclear receptor-mediated gene regulation. Crit Rev Biochem Mol Biol
44:25–33
Conrad NK, Mili S, Marshall EL et al (2006) Identification of a rapid mammalian deadenylation-
dependent decay pathway and its inhibition by a viral RNA element. Mol Cell 24:943–953
Conrad NK, Shu MD, Uyhazi KE et al (2007) Mutational analysis of a viral RNA element that
counteracts rapid RNA decay by interaction with the polyadenylate tail. Proc Natl Acad Sci
USA 104:10412–10417
Core LJ, Waterfall JJ, Lis JT (2008) Nascent RNA sequencing reveals widespread pausing and
divergent initiation at human promoters. Science 322:1845–1848
de Kok JB, Verhaegh GW, Roelofs RW et al (2002) DD3(PCA3), a very sensitive and specific
marker to detect prostate tumors. Cancer Res 62:2695–2698
Dettwiler S, Aringhieri C, Cardinale S et al (2004) Distinct sequence motifs within the 68-kDa
subunit of cleavage factor Im mediate RNA binding, protein–protein interactions, and subcel-
lular localization. J Biol Chem 279:35788–35797
Dreyfuss G, Kim VN, Kataoka N (2002) Messenger-RNA-binding proteins and the messages they
carry. Nat Rev Mol Cell Biol 3:195–205
Espinoza CA, Allen TA, Hieb AR et al (2004) B2 RNA binds directly to RNA polymerase II to
repress transcript synthesis. Nat Struct Mol Biol 11:822–829
Fejes-Toth K, Sotirova V, Sachidanandam R et al (2009) Post-transcriptional processing generates
a diversity of 50 -modified long and short RNAs. Nature 457:1028–1032
Feng J, Bi C, Clark BS et al (2006) The Evf-2 noncoding RNA is transcribed from the Dlx-5/6
ultraconserved region and functions as a Dlx-2 transcriptional coactivator. Genes Dev
20:1470–1484
Fox AH, Lam YW, Leung AK et al (2002) Paraspeckles: a novel nuclear domain. Curr Biol 12:13–25
Fox AH, Bond CS, Lamond AI (2005) P54nrb forms a heterodimer with PSP1 that localizes to
paraspeckles in an RNA-dependent manner. Mol Biol Cell 16:5304–5315
Franco-Zorrilla JM, Valli A, Todesco M et al (2007) Target mimicry provides a new mechanism
for regulation of microRNA activity. Nat Genet 39:1033–1037
Ganesan G, Rao SM (2008) A novel noncoding RNA processed by Drosha is restricted to nucleus
in mouse. RNA 14:1399–1410
388 T. Hirose
Ghanem N, Jarinova O, Amores A et al (2003) Regulatory roles of conserved intergenic domains

in vertebrate Dlx bigene clusters. Genome Res 13:533–543
Guru SC, Agarwal SK, Manickam P et al (1997) A transcript map for the 2.8-Mb region containing
the multiple endocrine neoplasia type 1 locus. Genome Res 7:725–735
Guttman M, Amit I, Garber M et al (2009) Chromatin signature reveals over a thousand highly
conserved large non-coding RNAs in mammals. Nature 458:223–227
Hastings ML, Ingle HA, Lazar MA et al (2000) Post-transcriptional regulation of thyroid hormone
receptor expression by cis-acting sequences and a naturally occurring antisense RNA. J Biol
Chem 275:11507–11513
Hatchell EC, Colley SM, Beveridge DJ et al (2006) SLIRP, a small SRA binding protein, is a
nuclear receptor corepressor. Mol Cell 22:657–668
Heasman J, Wessely O, Langland R et al (2001) Vegetal localization of maternal mRNAs is
disrupted by VegT depletion. Dev Biol 240:377–386
Hellwig S, Bass BL (2008) A starvation-induced noncoding RNA modulates expression of Dicer-
regulated genes. Proc Natl Acad Sci USA 105:12897–12902
Hirota K, Miyoshi T, Kugou K et al (2008) Stepwise chromatin remodelling by a cascade of
transcription initiation of non-coding RNAs. Nature 456:130–134
Hirotsune S, Yoshida N, Chen A et al (2003) An expressed pseudogene regulates the messenger-
RNA stability of its homologous coding gene. Nature 423:91–96
Hogga I, Karch F (2002) Transcription through the iab-7 cis-regulatory domain of the bithorax
complex interferes with maintenance of Polycomb-mediated silencing. Development
129:4915–4922
Hongay CF, Grisafi PL, Galitski T et al (2006) Antisense transcription controls cell fate in
Saccharomyces cerevisiae. Cell 127:735–745
Hutchinson JN, Ensminger AW, Clemson CM et al (2007) A screen for nuclear transcripts
identifies two linked noncoding RNAs associated with SC35 splicing domains. BMC Genomics
8:39
Ideue T, Sasaki YT, Hagiwara M et al (2007) Introns play an essential role in splicing-dependent
formation of the exon junction complex. Genes Dev 21:1993–1998
Ideue T, Hino K, Kitao S et al (2009) Efficient oligonucleotide-mediated degradation of nuclear
noncoding RNAs in mammalian cultured cells. RNA 15:1578–1587
International Human Genome Sequencing Consortium (2004) Finishing the euchromatic sequence
of the human genome. Nature 431:931–945
Isken O, Maquat LE (2007) Quality control of eukaryotic mRNA: safeguarding cells from
abnormal mRNA function. Genes Dev 21:1833–1856
Ji P, Diederichs S, Wang W et al (2003) MALAT-1, a novel noncoding RNA, and thymosin 4
predict metastasis and survival in early-stage nonsmall cell lung cancer. Oncogene
22:8031–8041
Kanduri C, Thakur N, Pandey RR (2006) The length of the transcript encoded from the Kcnq1ot1
antisense promoter determines the degree of silencing. EMBO J 25:2096–2106
Kapranov P, Cheng J, Dike S et al (2007) RNA maps reveal new RNA classes and a possible
function for pervasive transcription. Science 316:1484–1488
Khalil AM, Guttman M, Huarte M et al (2009) Many human large intergenic noncoding RNAs
associate with chromatin-modifying complexes and affect gene expression. Proc Natl Acad Sci
USA 106:11667–11672
Kiyosawa H, Mise N, Iwase S et al (2005) Disclosing hidden transcripts: mouse natural sense-
antisense transcripts tend to be poly(A) negative and nuclear localized. Genome Res
15:463–474
Kloc M, Wilk K, Vargas D et al (2005) Potential structural role of non-coding and coding RNAs in
the organization of the cytoskeleton at the vegetal cortex of Xenopus oocytes. Development
132:3445–3457
Kloc M, Bilinski S, Dougherty MT (2007) Organization of cytokeratin cytoskeleton and germ

plasma in the vegetal cortex of Xenopus laevis oocytes depends on coding and non-coding
RNAs: three-dimensional and ultrastructural analysis. Exp Cell Res 313:1639–1651
Kohtz JD, Fishell G (2004) Developmental regulation of EVF-1, a novel non-coding RNA
transcribed upstream of the mouse Dlx6 gene. Gene Expr Patterns 4:407–412
Krystal GW, Armstrong BC, Battey JF (1990) N-myc mRNA forms an RNA–RNA duplex with
endogenous antisense transcripts. Mol Cell Biol 10:4180–4191
Kurihara Y, Matsui A, Hanada K et al (2009) Genome-wide suppression of aberrant mRNA-like
noncoding RNAs by NMD in Arabidopsis. Proc Natl Acad Sci USA 106:2453–2458
Lamond AI, Spector DL (2003) Nuclear speckles: a model for nuclear organelles. Nat Rev Mol
Cell Biol 4:605–612
Lanz RB, McKenna NJ, Onate SA et al (1999) A steroid receptor coactivator, SRA, functions as an
RNA and is present in an SRC-1 complex. Cell 97:17–27
Lanz RB, Razani B, Goldberg AD et al (2002) Distinct RNA motifs are important for coactivation
of steroid hormone receptors by steroid receptor RNA activator (SRA). Proc Natl Acad Sci
USA 99:16081–16086
Licatalosi DD, Mele A, Fak JJ et al (2008) HITS–CLIP yields genome-wide insights into brain
alternative RNA processing. Nature 456:464–469
Lin R, Maeda S, Liu C et al (2007) A large noncoding RNA is a marker for murine hepatocellular
carcinomas and a spectrum of human carcinomas. Oncogene 26:851–858
Mariner PD, Walters RD, Espinoza CA et al (2008) Human Alu RNA is a modular transacting
repressor of mRNA transcription during heat shock. Mol Cell 29:499–509
Martens JA, Laprade L, Winston F (2004) Intergenic transcription is required to repress the
Saccharomyces cerevisiae SER3 gene. Nature 429:571–574
Martens JA, Wu PY, Winston F (2005) Regulation of an intergenic transcript controls adjacent
gene transcription in Saccharomyces cerevisiae. Genes Dev 19:2695–2704
Martianov I, Ramadass A, Serra Barros A et al (2007) Repression of the human dihydrofolate
reductase gene by a non-coding interfering transcript. Nature 445:666–670
Masui O, Heard E (2006) RNA and protein actors in X-chromosome inactivation. Cold Spring
Harb Symp Quant Biol 71:419–428
Mili S, Steitz JA (2004) Evidence for reassociation of RNA-binding proteins after cell lysis:
implications for the interpretation of immunoprecipitation analyses. RNA 10:1692–1694
Misteli T (2005) Concepts in nuclear architecture. Bioessays 27:477–487
Mohammad F, Pandey RR, Nagano T et al (2008) Kcnq1ot1/Lit1 noncoding RNA mediates
transcriptional silencing by targeting to the perinucleolar region. Mol Cell Biol 28:3713–3728
Munroe SH, Lazar MA (1991) Inhibition of c-erbA mRNA splicing by a naturally occurring
antisense RNA. J Biol Chem 266:22083–22086
Nagano T, Mitchell JA, Sanz LA et al (2008) The air noncoding RNA epigenetically silences
transcription by targeting G9a to chromatin. Science 322:1717–1720
Pandey RR, Mondal T, Mohammad F et al (2008) Kcnq1ot1 antisense noncoding RNA mediates
lineage-specific transcriptional silencing through chromatin-level regulation. Mol Cell
32:232–246
Peters J, Robson JE (2008) Imprinted noncoding RNAs. Mamm Genome 19:493–502
Petruk S, Sedkov Y, Riley KM et al (2006) Transcription of bxd noncoding RNAs promoted by
trithorax represses Ubx in cis by transcriptional interference. Cell 127:1209–1221
Payer B, Lee JT (2008) X chromosome dosage compensation: how mammals keep the balance.
Annu Rev Genet 42:733–772
Petruk S, Sedkov Y, Brock HW et al (2007) A model for initiation of mosaic HOX gene expression
patterns by non-coding RNAs in early embryos. RNA Biol 4:1–6
Pollard KS, Salama SR, Lambert N et al (2006) An RNA gene expressed during cortical develop-
ment evolved rapidly in humans. Nature 443:167–172
Prasanth KV, Prasanth SG, Xuan Z et al (2005) Regulating gene expression through RNA nuclear
retention. Cell 123:249–263
390 T. Hirose
Prasanth KV, Spector DL (2007) Eukaryotic regulatory RNAs: an answer to the ‘genome com-
plexity’ conundrum. Genes Dev 21:11–42
Preker P, Nielsen J, Kammler S et al (2008) RNA exosome depletion reveals transcription
upstream of active human promoters. Science 322:1851–1854
Ravasi T, Suzuki H, Pang KC et al (2006) Experimental validation of the regulated expression of
large numbers of non-coding RNAs from the mouse genome. Genome Res 16:11–19
Rinn JL, Kertesz M, Wang JK et al (2007) Functional demarcation of active and silent chromatin
domains in human HOX loci by noncoding RNAs. Cell 129:1311–1323
Royo H, Cavaillé J (2008) Non-coding RNAs in imprinted gene clusters. Biol Cell 100:149–166
Sanchez-Elsner T, Gou D, Kremmer E et al (2006) Noncoding RNAs of trithorax response
elements recruit Drosophila Ash1 to Ultrabithorax. Science 311:1118–1123
Sasaki YT, Sano M, Ideue T et al (2007) Identification and characterization of human non-coding
RNAs with tissue-specific expression. Biochem Biophys Res Commun 357:991–996
Sasaki YT, Ideue T, Sano M et al (2009) MENe/b noncoding RNAs are essential for structural
integrity of nuclear paraspeckles. Proc Natl Acad Sci USA 106:2525–2530
Schmitt S, Prestel M, Paro R (2005) Intergenic transcription through a polycomb group response
element counteracts silencing. Genes Dev 19:697–708
Seidl CI, Stricker SH, Barlow DP (2006) The imprinted Air ncRNA is an atypical RNAPII
transcript that evades splicing and escapes nuclear export. EMBO J 25:3565–3575
Seila AC, Calabrese JM, Levine SS et al (2008) Divergent transcription from active promoters.
Science 322:1849–1851
Shamovsky I, Ivannikov M, Kandel ES et al (2006) RNA-mediated response to heat shock in
mammalian cells. Nature 440:556–560
Silva J, Mak W, Zvetkova I et al (2003) Establishment of histone h3 methylation on the inactive X
chromosome requires transient recruitment of Eed-Enx1 polycomb group complexes. Dev Cell
4:481–495
Sleutels F, Zwart R, Barlow DP (2002) The non-coding Air RNA is required for silencing
autosomal imprinted genes. Nature 415:810–813
Smith CM, Steitz JA (1998) Classification of gas5 as a multi-small-nucleolar-RNA (snoRNA) host
gene and a member of the 50 -terminal oligopyrimidine gene family reveals common features of
snoRNA host genes. Mol Cell Biol 18:6897–6909
Sone M, Hayashi T, Tarui H et al (2007) The mRNA-like noncoding RNA Gomafu constitutes a
novel nuclear domain in a subset of neurons. J Cell Sci 120:2498–2506
Srikantan V, Zou Z, Petrovics G et al (2000) PCGEM1, a prostate-specific gene, is overexpressed
in prostate cancer. Proc Natl Acad Sci USA 97:12216–12221
Sunwoo H, Dinger ME, Wilusz JE et al (2009) MENe/b nuclear-retained non-coding RNAs are
up-regulated upon muscle differentiation and are essential components of paraspeckles.
Genome Res 19:347–359
Taft RJ, Pheasant M, Mattick JS (2007) The relationship between non-protein-coding DNA and
eukaryotic complexity. Bioessays 29:288–299
Tycowski KT, Shu MD, Steitz JA (1996) A mammalian gene with introns instead of exons
generating stable RNA products. Nature 379:464–466
Uhler JP, Hertel C, Svejstrup JQ (2007) A role for noncoding transcription in activation of the
yeast PHO5 gene. Proc Natl Acad Sci USA 104:8011–8016
Ule J, Jensen K, Mele A et al (2005) CLIP: a method for identifying protein-RNA interaction sites
in living cells. Methods 37:376–386
Valgardsdottir R, Chiodi I, Giordano M et al (2005) Structural and functional characterization of
noncoding repetitive RNAs transcribed in stressed human cells. Mol Biol Cell 16:2597–2604
Wang X, Arai S, Song X et al (2008) Induced ncRNAs allosterically modify RNA-binding proteins
in cis to inhibit transcription. Nature 454:126–130
Willingham AT, Orth AP, Batalov S et al (2005) A strategy for probing the function of noncoding
RNAs finds a repressor of NFAT. Science 309:1570–1573
Wilusz JE, Freier SM, Spector DL (2008) 3’ end processing of a long nuclear-retained noncoding
RNA yields a tRNA-like cytoplasmic RNA. Cell 135:919–932
Yamada K, Kano J, Tsunoda H et al (2006) Phenotypic characterization of endometrial stromal
sarcoma of the uterus. Cancer Sci 97:106–112
Yan MD, Hong CC, Lai GM et al (2005) Identification and characterization of a novel gene Saf
transcribed from the opposite strand of Fas. Hum Mol Genet 14:1465–1474
Zhao J, Sun BK, Erwin JA et al (2008) Polycomb proteins targeted by a short repeat RNA to the
mouse X chromosome. Science 322:750–756
Zhao X, Patton JR, Davis SL et al (2004) Regulation of nuclear receptor activity by a pseudouridine
synthase through post transcriptional modification of steroid receptor RNA activator. Mol Cell
15:549–558
Noncoding RNAs as Therapeutic Targets
Maciej Szymański and Jan Barciszewski
Contents
1 RNA-Dependent Regulation of Gene Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
1.1 Gene Regulation Through Epigenetic Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
1.2 Controlling Transcription Machinery Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
1.3 Posttranscriptional Regulatory Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
2 The Medical Perspective: Noncoding RNAs
in Human Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
2.1 MicroRNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
2.2 mRNA-Like ncRNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
2.3 Other Transcripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
3 NcRNA-Based Therapeutic Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
Abstract Noncoding RNAs are key players in the regulation of complex cellular
processes. Over the years, it has been demonstrated that their abnormal expression
is associated with many human pathologies including developmental and neurobe-
havioral disorders, diabetes, obesity, and cancer. A wide spectrum of activities and
a large number of genes that are regulated by RNA-dependent mechanisms makes
the ncRNA molecules attractive targets for developing next generation therapeutic
agents. This chapter outlines the principles of RNA regulation, the involvement
of various RNAs in human diseases, and the strategies of application of
ncRNA-targeted therapeutic approaches.
Keywords Cancer Noncoding RNAs Therapy
M. Szymański (*) and J. Barciszewski

Institute of Bioorganic Chemistry of the Polish Academy of Sciences, Noskowskiego 12, 61-704,
Poznan, Poland
e-mail: mszyman@ibch.poznan.pl; Jan.Barciszewski@ibch.poznan.pl
394 M. Szymański and J. Barciszewski
1 RNA-Dependent Regulation of Gene Expression
A strict control of expression of genetic information contained within the genome is

crucial for correct development and proper functioning of every organism. For many
years, it was assumed that the organization, growth, development, and functioning
of highly complex organisms require a large number of different proteins that
were recognized as the only biomolecules responsible for catalytic, structural, and
regulatory functions. The progress in genomic studies demonstrated that the differ-
ences in complexity and organization of organisms cannot be accounted for by the
differences in the repertoire of proteins encoded by their genomes. In fact, the
number of protein coding genes identified in mammalian genomes was significantly
lower than expected. The mammalian genes count of 20,000–25,000 (Lander et al.
2001; Waterston et al. 2002) when compared with 19,000 in Caenorhabditis elegans
(C. elegans Sequencing Consortium 1998) or 25,000–40,000 genes in plants did
not seem to confirm earlier supposition that the complexity of organisms is directly
correlated with the number of protein-coding genes. This finding was one of the
greatest surprises coming from the analysis of the human genome sequence. Yet
another unexpected finding was the size of protein-coding part. In human genome,
there are less than 2% of nucleotides that are actually encoding proteins or open
reading frames of the corresponding mRNAs. The problem of protein diversity is
partially solved in higher organisms by the alternative splicing found in the majority
of human genes (Clark et al. 2007) and resulting in production of multiple proteins
from a single primary transcript. However, the answer to the low gene number –
high complexity problem seems to be associated with the transcripts originating
from the apparently noncoding DNA. There is, in fact, a positive correlation
between the biological complexity of organisms and the size of nonprotein-coding
fractions of their genomes (Taft et al. 2007).
In the pregenomic era, it was already known that there is a large portion of
nuclear DNA that does not code for proteins or known classes of infrastructural
RNAs. Apart from the gene control elements providing binding sites for protein
factors regulating transcription, it was assumed that most of this noncoding DNA,
often regarded as “junk,” does not perform any particular function. In recent years,
these views on the organization and functions of genomic DNA were challenged by
the discovery of a widespread presence of diverse regulatory mechanisms involving
a variety of untranslated RNAs referred to as noncoding RNAs (ncRNA) (Mattick
and Makunin 2006; Szymanski and Barciszewski 2006).
Although some of the ncRNAs have been isolated nearly 30 years ago, their
function remained elusive. The real magnitude of the problem was only appreciated
after the data from large-scale cDNA sequencing and genome tiling arrays
became available (Carninci et al. 2005; Cheng et al. 2005). It is now obvious
that, in addition to the protein-coding mRNAs and infrastructural RNAs, there are
thousands of other RNAs transcribed from the genomic DNA. Currently, it is
estimated that almost all of the mammalian genomic DNA is transcribed from
both strands and the number of RNA-coding genes is at least equal to that of
Noncoding RNAs as Therapeutic Targets 395
protein-coding ones (Beiter et al. 2009). It has been proposed that the ncRNAs,
including transcripts from the independent transcriptional units as well as those
processed from intronic sequences constitute elements of the regulatory networks
responsible for controlling gene expression. The increase in complexity of regu-
latory networks dependent on ncRNA accompanied by a decrease in the size of
protein-coding portion of the genome would contribute to increased complexity of
the living systems (Mattick 1994, 2001, 2003, 2009a). Thus, the subtle changes in
expression or function of regulatory RNAs may be essential for the evolution and
development of eukaryotes.
The whole output of transcription from the genomic DNA (the transcriptome)
consists of the protein-coding part (open reading frames) serving as template for
protein biosynthesis and noncoding part including untranslated regions of mRNAs,
ncRNA, and introns removed from primary transcripts of both protein-coding and
noncoding genes.
The ncRNA fall into two broad categories: housekeeping or infrastructural
RNAs essential for protein biosynthesis (tRNA, rRNA), RNA processing
(snRNA), and modifications (snoRNA) and other elementary cellular functions
(e.g., telomerase RNA, RNase P RNA). Expression of these RNAs is constitutive
and their levels in the cells do not change.
In the last decade, it was realized that there also exists a large group of ncRNA
molecules performing regulatory functions affecting expression of numerous genes.
Unlike the housekeeping RNAs that are essential for the fundamental life functions
and thus expressed in every cell, the regulatory ncRNAs are expressed in a very
specific strictly regulated manner, and their repertoire varies depending on tissue or
cell type, developmental stage, or the changes of biotic or abiotic environmental
conditions (Mercer et al. 2009; Szymanski and Barciszewski 2006).
Although in recent years a significant progress was made in identifying new
ncRNA species, there are still many unsolved questions concerning the mechanisms
underlying their role in regulatory pathways. Various ncRNAs act on different
stages of gene expression. They are involved in processes affecting chromatin
structure and its transcriptional competence, activity of transcription factors, and
the fate of RNAs on the posttranscriptional level.
1.1 Gene Regulation Through Epigenetic Mechanisms
ncRNA play a key role in epigenetic processes associated with silencing of genes
within imprinted clusters (Mohammad et al. 2009), X chromosome inactivation
(Payer and Lee 2008), DNA methylation (Imamura et al. 2004a), and heterochromatin
formation at the centromeres (Iida et al. 2008). Genomic imprinting is unique to
mammals and represents epigenetic modification directing expression of imprinted
alleles according to parent of origin (Mohammad et al. 2009). A common feature of
clusters of imprinted genes is expression of ncRNAs that are required for transcrip-
tional inactivation of protein-coding genes (Morison et al. 2005). The precise role of
most of these imprinted ncRNAs in silencing is unknown, but their expression was
shown to be essential for this process. Aberrant expression of imprinted genes on
human chromosome 11p15 is associated with Beckwith–Wiedemann syndrome
(BWS) and several human cancers (Enklaar et al. 2006). In most human tissues,
the paternal allele produces a 91 kb long nuclear antisense ncRNA (LIT1/Kcnq1ot1)
originating from an unmethylated CpG island (KvDMR1) within intron 10 of the
maternally expressed KCNQ1 (KvLQT1) gene (Niemitz et al. 2004). Methylation
status of KvDMR1 and expression of LIT1 RNA constitute the key elements
controlling imprinted expression of the BWS associated genes. Demethylation of
the maternal KvDMR1 allele or its deletions are among the most frequent causes of
the BWS (DeBaun et al. 2002). It has been demonstrated that LIT1 RNA expression is
necessary for transcriptional inactivation of several maternally expressed genes on the
paternal chromosome (Horike et al. 2000). The silencing effect spreads bidirectionally
(Thakur et al. 2004), affecting genes located downstream and upstream from
KvDMR1, and its extent depends on the length of LIT1 RNA (Mancini-Dinardo
et al. 2006; Kanduri et al. 2006). In the mouse embryo, the region affected by Lit1
RNA spans 400 kb and is extended to over 780 kb in the placenta (Terranova et al.
2009). Transcriptional repression depends on Polycomb group complex Eed-Ezh2
and repressive methylations of histone H3 (Umlauf et al. 2004; Terranova et al. 2009).
Moreover, it has been shown that Lit1 RNA participates in establishing a nuclear
domain comprising exclusively silenced genes (Terranova et al. 2009).
Similar bidirectional silencing induced by ncRNA occurs at the imprinted
cluster on mouse chromosome 17. Differential methylation of the imprinting
control region within intron 2 of insulin-like growth factor type-2 receptor (Igf2r)
regulates expression of maternally expressed genes (Wutz et al. 1997). An anti-
sense, unspliced 108 kb long Air RNA (antisense Igf2r) overlapping 30 kb of Igf2r
gene is transcribed from the unmethylated paternal allele. Transcriptional silencing
affects the sense Igf2r gene and two other genes, Slc22a2 and Slc22a3, 110 and
155 kb downstream from Igf2r, respectively (Sleutels et al. 2003).
The Air RNA was shown to directly interact with the chromatin at Slc22a3
promoter, which correlated with the presence of repressive modifications of histone
H3. The silencing depended on the activity of histone methyltransferase G9a and a
whole length Air transcript (Nagano et al. 2008).
Epigenetic gene silencing induced by noncoding transcripts is not restricted to
imprinted genes clusters. An antisense transcript was found to downregulate
expression of the cyclin-dependent kinase inhibitor p15 that plays a role of tumor
suppressor gene. In leukemia, it was demonstrated that increased transcription of
the p15 antisense RNA (p15AS) results in reduced expression of the sense tran-
script. It has been found that the effect was due to epigenetic mechanism-inducing
heterochromatin formation. The repression of p15 was maintained even in the
absence of antisense transcript (Yu et al. 2008). Although there are limited data
available, it has been suggested that the RNA-induced gene silencing may represent
a more general mechanism and could account for regulatory potential of other
natural antisense transcripts (Pauler et al. 2007) that may be associated with
approximately 70% of loci in the mammalian genome (Katayama et al. 2005).
A special instance of epigenetic silencing extending to the whole chromosome is

X chromosome inactivation that results in formation of a transcriptionally silent
nuclear compartment (Chaumeil et al. 2006; Jonkers et al. 2008). Transcriptional
silencing of all but one X-chromosomes is critical for equalization of gene dosage
from the X chromosome between male (XY) and female (XX) cells (Chow and
Heard 2009). An essential element required for the initiation of this process is a
long mRNA-like ncRNA XIST (X-inactive specific transcript) expressed from the
XIC (X-inactivation center) locus (Newall et al. 2001). Upon transcription, XIST
RNA binds to the chromatin of the X chromosome undergoing inactivation and
recruits chromatin silencing factors including Polycomb group (PcG) repressive
complexes 1 (PRC1) and 2 (PRC2) (Plath et al. 2004; de Napoles et al. 2004).
NcRNAs were also involved in the regulation of DNA methylation. Expression
of tissue-specific isoforms of rat sphingosine kinase 1 (Sphk1) is regulated by a
tissue-specific, differentially methylated region (T-DMR) embedded in a 3.7 kb
long CpG island (Imamura et al. 2001, 2004a). Apart from the protein-coding
transcripts, Sphk1 locus produces an array of noncoding antisense (Khps1) RNAs.
One of these antisense transcripts overlaps the T-DMR and induces methylation of
nonCpG and demethylation of CpG sites (Imamura et al. 2004a). The conservation
of the SPHK1 gene structure, the presence of the T-DMR region, tissue specific
methylation, and multiple tissue-specific mRNA isoforms between rodents and
primates suggested that the antisense RNA-dependent methylation may also regu-
late SPHK1 expression in humans (Imamura et al. 2004b). It is also possible, that
similar mechanism employing ncRNAs overlapping T-DMRs may be responsible
for regulation of other genes producing tissue-specific mRNA isoforms (Shiota
2004; Turner et al. 2006).
1.2 Controlling Transcription Machinery Activity
ncRNA were also identified as factors modulating activity of proteins directly

involved in transcription. An ncRNA Evf-2 expressed from the ultraconserved
region (UCR) in the intergenic region between Dlx-5 and Dlx-6 genes encoding
homeobox-containing proteins. Evf-2 RNA forms a stable complex with another
homeodomain protein Dlx-2 and cooperatively enhances transcription of the tran-
scription Dlx-5 and Dlx-6 genes (Feng et al. 2006). Transcriptional cis- and trans-
activation dependent on expression of Evf-2 RNA is required for the proper
development of brain (Bond et al. 2009).
Another ncRNA involved in the regulation of transcription is 7SK RNA that
acts as an inhibitor of positive transcription elongation factor b (P-TEFb) (Nguyen
et al. 2001; Yang et al. 2001). 7SK RNA together with hexamethylene bisaceta-
mide-induced protein 1 (HEXIM1) and/or HEXIM2 protein binds P-TEFb and
inhibits its cyclin-dependent kinase 9 (Cdk9) activity required for phosphorylation
of the C-terminal domain of the RNA polymerase II largest subunit (Yik et al.
2004). Another protein involved in 7SK RNA activity is the La-related protein 7
(LARP7) binding a conserved 30 -terminal U-rich region and constituting an integral

part of the 7SK RNP (Markert et al. 2008). Suppression of the P-TEFb activity was
also found to affect alternative splicing (Barboric et al. 2009).
NcRNA also plays a part in the modulation of activity of nuclear receptors of
steroid hormones. Steroid receptor activator RNA (SRA RNA) was shown to act as
a coactivator of nuclear receptors of progestin, estrogen, androgen, glucocorticoid,
retinoic acid, thyroid hormone, and vitamin D (Lanz et al. 1999; Kawashima et al.
2003; Zhao et al. 2004; Hatchell et al. 2006). Alternative splicing generates tissue-
specific isoforms 0.7–1.5 kb long, differing in 5- and 30 -terminal regions and posses-
sing a common central core. The activity of various SRA mutants demonstrated the
importance of RNA secondary structure elements but not the integrity of putative
open reading frame (Lanz et al. 2002). Transcription initiation from an alternative
start site yields an mRNA encoding two SRAP proteins that are conserved among
vertebrates (Emberley et al. 2003; Chooniedass-Kothari et al. 2004; Leygue 2007).
Interestingly, the SRAP proteins are also involved in the regulation of steroid
hormones responsive genes (Kawashima et al. 2003; Chooniedass-Kothari et al.
2006). The interactions between SRA RNA and hormone receptors are mediated
by hormone receptors through a steroid receptor coactivator 1 (SRC-1) (Lanz et al.
1999) or transcriptional repressors: SHARP, (SMRT/HDAC1 associated repressor
protein) (Shi et al. 2001) and SLIRP (SRA stem–loop interacting RNA-binding
protein) (Hatchell et al. 2006). The interaction of SRA RNA with SRC-1 depends
on binding DEAD-box RNA-binding proteins p68 and p72 (Watanabe et al. 2001)
that also mediate activation of MyoD transcription factor essential for muscle
development (Caretti et al. 2006).
Another ncRNA involved in the regulation of transcription factor’s activity is
20 bp long double-stranded small modulatory RNA (NRSE smRNA) expressed in
neural cells and responsible for switching on neuron-specific genes associated with
a promoter element NRSE/RE1 (neuron-restrictive silencer element/repressor
element 1). NRSE/RE1 constitutes a binding site for a NRSF/REST (neuron-restrictive
silencing factor/repressor element 1 silencing transcription factor) playing a role of
transcriptional repressor blocking expression of neuron-specific genes in nonneural
tissues (Abrajano et al. 2009). NRSE smRNA binds the REST protein preventing its
interactions with a corepressor, which in turn leads to transcription of NRSE/RE1
controlled genes (Kuwabara et al. 2004).
NcRNAs also play an important role in transcriptional regulation during heat
shock. Expression of heat shock proteins depends on activation of heat shock
transcription factor 1 (HSF-1), a constitutively expressed protein present in an
inactive monomeric form in unstressed cells. Upon heat shock, the HSF-1 protein
trimerizes and binds promoters of heat shock protein genes increasing up to
200-fold their expression (Anckar and Sistonen 2007). Recently, it has been
demonstrated that the process of HSF-1 activation requires presence of a ribonu-
cleoprotein complex of a translation elongation factor eEF1A and a large HSR1
ncRNA (heat-shock RNA 1) (Shamovsky et al. 2006). Both eEF1A and HSR1 RNA
are required for activation and have been shown to form a complex with HSF1
(Shamovsky and Nudler 2009).
Apart from the induction of heat shock proteins, there is a general inhibition of
transcription by RNA polymerase II. This effect is brought about by short inter-
spersed elements (SINEs) transcripts, binding the core PolII and inhibiting tran-
scription at the initiation step. In mouse, the regulatory RNAs are transcribed from
B2 SINEs (Allen et al. 2004), while in humans, the same role is played by Alu
transcripts (Mariner et al. 2008). At the molecular level, inhibition is achieved
by ncRNAs preventing contacts between polymerase and the promoter DNA
(Yakovchuk et al. 2009). Reactivation of transcription occurs either through disso-
ciation of ncRNA or its removal by as yet unknown factors (Espinoza et al. 2004).
1.3 Posttranscriptional Regulatory Mechanisms
The potential of forming complementary duplexes by two RNA molecules sug-

gested that such interactions may provide a basis for regulatory mechanisms. In
fact, it seems that most of the regulatory RNAs operate at the posttranscriptional
steps of gene expression on the basis of recognition between sense–antisense
sequences influencing RNA splicing, translation, and affecting target molecule’s
stability. The most obvious candidates to perform such tasks are antisense RNAs.
An analysis of mouse transcriptomes revealed that in over 70% of loci transcripts
are generated from both strands resulting in overlapping pairs of sense:antisense
RNAs (Katayama et al. 2005). However, the role of most of these antisense
transcripts is not known, and a regulatory role was demonstrated only in a few
cases. It has been shown that, at least in some cases, antisense transcripts can
regulate sense genes via RNAi pathway by forming dsRNA, cleaved by Dicer to
produce endogenous siRNAs (Watanabe et al. 2008).
Antisense transcripts were demonstrated to downregulate the expression of the
HOXA11 gene, but the observed effect was not likely to be due to sense–antisense
RNA duplex formation (Chau et al. 2002). The complementary interactions were,
however, observed between NPPA (natriuretic peptide precursor A) mRNA and its
antisense partner NPPA-AS. One of the alternatively spliced NPPA-AS isoforms
downregulated expression of intron-retained NPPA mRNA variant (Annilo et al.
2009). In several cases including Saf (Yan et al. 2005), c-erbA (Munroe and Lazar
1991), and N-myc (Krystal et al. 1990), expression of antisense transcripts was
linked to alternative splicing. Recently, an antisense RNA was reported for the
Zeb2/Sip1 gene encoding a transcriptional repressor of E-cadherin. Expression of
Zeb2/Sip1 depends on an internal ribosome entry site (IRES) that is normally
spliced out of the mRNA. The retention of intron containing the IRES and conse-
quently translation of Zeb2/Sip1 protein occurs during epithelial to mesenchymal
transition and depends on expression of an antisense RNA that by hybridization
with the 50 -splice site interferes with spliceosome binding which, in turn, prevents
removing the IRES from the mRNA (Beltran et al. 2008).
Although most of the sense–antisense pairs of transcripts originate from over-
lapping regions (cis-antisense), there are also trans-antisense RNAs transcribed
from separate transcriptional units. It has been reported that long antisense tran-
scripts originating from pseudogenes can form duplexes with their corresponding
mRNAs. The double-stranded RNAs are further processed by Dicer to generate
endogenous siRNAs that participate in RISC-mediated cleavage of protein-coding
transcripts and downregulation of the gene (Tam et al. 2008). This observation
demonstrated that although pseudogenes were mostly regarded as nonfunctional
remnants of past events, they may also play an important role in regulatory
pathways.
The largest and best understood class of posttranscriptional regulators is micro-
RNAs that in recent years attracted most attention from researchers. This is
primarily due to their wide spectrum of activities and their involvement in the
regulation of many genes playing critical role in development and differentiation
(Stefani and Slack 2008). The ability to posttranscriptionally modulate expression
of many genes makes microRNAs, together with transcription factors, the primary
determinants of unique gene expression patterns in eukaryotic cells (Chen and
Rajewsky 2007; Hobert 2004).
Most of the animal microRNAs are involved in translational regulation depend-
ing on binding partially complementary sequences within target mRNAs’ 30 -UTRs
(Bartel 2004). Mechanism of translational repression depends on the presence of
7-methyl guanosine cap (Pillai et al. 2005) and one of the Argonaute family proteins
(Ago2) that interferes with cap binding by the translation initiation factor eIF4E on
microRNA-bound mRNA (Kiriakidou et al. 2007). Another possibility is binding of
initiation factor eIF6 by the microRNA–mRNA complex that prevents translation
by blocking association of the ribosomal subunits (Chendrimada et al. 2007). In
addition to translational inhibition, microRNAs can induce deadenylation of
mRNA, thus decreasing its stability (Fabian et al. 2009; Beilharz et al. 2009).
Not all animal microRNAs act as translational repressors. Some are involved
in posttranscriptional control by directing cleavage of target mRNAs by Dicer
nuclease as in the case of mir-196 targeting Hox-B8 mRNA (Yekta et al. 2004).
However, microRNA-induced hydrolysis requires full complementarity between
target mRNA and microRNA.
Apart from regulating translation and stability of protein coding genes, micro-
RNAs can also be involved in posttranscriptional regulation of ncRNA. Comple-
mentary sites for miR-155 and miR-24-1 were found within UCRs differentially
expressed in chronic lymphocytic leukemias (CLLs) suggesting that there may exist
links between microRNAs and expression of other ncRNAs (Calin et al. 2007).
Localized translation of mRNA in the neurons of rodents and primates is
regulated by PolIII transcripts BC1 and BC200 (Martignetti and Brosius 1995)
identified in ribonucleoprotein particles in cell bodies and in dendrites (Tiedge et al.
1991, 1993). BC200 was also found in complexes with SYNCRIP (synaptotagmin-
binding cytoplasmic RNA interacting protein), a component of mRNA transport
granules involved in localized protein synthesis at postsynaptic sites. (Duning et al.
2008). It has been shown that BC1 interferes with the formation of a stable 48S
preinitiation complex. BC1 specifically inhibits activity of eIF4A helicase and forms
stable complexes with the poly(A)-binding protein (PABP) (Wang et al. 2002;
Lin et al. 2008). At synapses, BC1 and BC200 RNAs were also found in complexes
with FMRP (fragile X mental retardation associated protein) participating in trans-
lational repression of a subset of FMRP-regulated mRNAs (Zalfa et al. 2003, 2005).
NcRNAs can also regulate alternative splicing as demonstrated in the case
o serotonin receptor 5-HT(2C)R. Splice site selection in exon Vb of 5-HT(2C)R
is controlled by HBII-52 snoRNA that binds a silencing element. In Prader–Willi
syndrome patients, lacking expression of HBII-52 the 5-HT(2C)R mRNA isoforms
differ from those found in healthy individuals (Kishore and Stamm 2006). Recently,
the involvement of MBII-52, the mouse homolog of HBII-52, in alternative splicing
has been confirmed for five other pre-mRNAs containing alternative exons. It has
also been shown that there is an alternative, shorter isoform associate with hnRNPs,
differing in structure from the canonical C/D box snoRNA that is responsible for
this process (Kishore et al. 2010).
Posttranscriptional regulation by ncRNAs may also involve subcellular locali-
zation of proteins. A calcium-responsive transcription factor NFAT (nuclear factor
of activated T cells) controls expression of genes involved T-cell-mediated immune
response and in the development of vascular and nervous system and muscles. It has
been shown that the nuclear localization of NFAT is regulated by an NRON ncRNA
(noncoding repressor of NFAT), which by interactions with nuclear import factors
specifically inhibits nucleocytoplasmic trafficking of NFAT, but not of other
transcription factors (Willingham et al. 2005).
2 The Medical Perspective: Noncoding RNAs

in Human Diseases
The diversity of ncRNA functions and their involvement in processes linked to cell
growth, differentiation, and development makes them very susceptible targets for
mutations resulting in various pathologies. In fact, the number of ncRNAs, belong-
ing to various classes of transcripts, implicated in human diseases is constantly
growing. Although, there is not always a clear-cut distinction between a cause and
an outcome, aberrant expression of many ncRNAs has been observed in human
cancers and neurobehavioral and developmental disorders.
2.1 MicroRNAs
An extraordinary potential of microRNAs as regulatory molecules is a consequence

of their mode of action that allows any single microRNA to control expression of
multiple targets thus affecting many cellular processes. On the other hand, single
mRNAs can be regulated by several different microRNAs with different expression
patterns. It has been estimated that about 25% of human genes can be targeted by
multiple microRNAs. It has been noted that mRNAs encoding nuclear proteins
controlling expression of genes involved in development and differentiation often

possess target sites for 15 and more microRNAs. Among cancer-related genes,
there are many that can potentially be regulated by over 30 microRNAs (Hon and
Zhang 2007). The distribution of microRNA-coding regions in human and mouse
genomes correlates with regions of genetic aberrations linked to various forms of
cancer, which suggested that microRNAs could play a role of either oncogenes or
tumor suppressors (Calin et al. 2004; Sevignani et al. 2007).
The analysis of microRNAs expression profiles revealed that each cell line
possesses a unique microRNA signature and that malignant growth is generally
accompanied by an overall reduction of microRNAs consistent with the role of
microRNAs in maintaining differentiated state of the cells (Lu et al. 2005).
Decreased expression of microRNAs can result from deficiency of proteins
involved in processing and maturation of microRNA precursors (Karube et al.
2005; Kumar et al. 2007) as demonstrated in the case of downregulation of mature
miR-143 and miR-145 in colorectal cancers (Michael et al. 2003).
It has been demonstrated that there are very specific patterns of microRNA
expression in normal tissues significantly different from those in malignant cells.
In colorectal cancer, malignant transformation results in altered expression of 13
microRNAs. Moreover, the stage of the tumor correlated with the levels of miR-31
(Bandrés et al. 2006). MicroRNA profiling can also differentiate various subtypes
of cancer of the same origin (Blenkiron et al. 2007). Thus, the microRNA expres-
sion profiles provide very sensitive molecular markers of oncogenesis that in
many respects is superior when compared with profiling using expression data
for protein-coding genes (Ciafre et al. 2005; Lu et al. 2005; Murakami et al.
2006; Roldo et al. 2006).
In recent years, there was a significant progress in deciphering the molecular
mechanisms underlying the involvement of particular microRNAs in processes
associated with oncogenesis. Although generally microRNA expression is signifi-
cantly reduced in malignant cells, particular microRNAs can show either up- or
downregulation consistent with their role as either oncogenes or tumor suppressors
(Kent and Mendell 2006), though it is not always possible to draw a line between
the two.
MicroRNAs have been shown to play a role in regulating apoptosis-related
genes. MiR-21 is an oncogenic microRNA showing elevated expression in many
types of cancer (Krichevsky and Gabriely 2009). It has been shown that MiR-21-
suppresses apoptosis (Chan et al. 2005; Si et al. 2007) by targeting an mRNA
encoding tumor suppressor protein Programmed Cell Death 4 (PDCD4) (Frankel
et al. 2008). Moreover, miR-21 downregulates expression of other tumor suppres-
sors involved in proliferation, cell migration, and invasion: PTEN (phosphatase and
tensin homolog) (Meng et al. 2007) and TPM1 (tropomyosin 1) (Zhu et al. 2007).
Another ubiquitous microRNA-controlling cell proliferation is let-7, which is signifi-
cantly downregulated in many cancers (Peter 2009). Let-7 suppresses expression of
human RAS genes (Johnson et al. 2005) and other proto-oncogenes involved in the
regulation of a cell cycle (Johnson et al. 2007).
Regulation of cell growth and apoptosis has been proposed for miR-16 family
members that target genes responsible for cell cycle progression from G0/G1 to S
phase, including regulators of G1 phase CDK6, CDC27, an activator of NF-kB
signaling CARD10 (Linsley et al. 2007; Cloonan et al. 2008), and BCL2 that plays
a role of antiapoptotic factor (Cimmino et al. 2005).
MicroRNAs also play an important role in viral infections. Human cytomegalo-
virus (HCMV) encodes a microRNA hcmv-mirR-UL112 targeting major histocom-
patibility complex class I-related chain B (MICB) involved in the recognition and
killing of virus-infected cells by natural killer cells (Stern-Ginossar et al. 2007).
MicroRNAs involved in reprogramming endothelial cells by targeting the cellular
transcription factor MAF (musculoaponeurotic fibrosarcoma oncogene homolog)
were identified in Kaposi sarcoma herpesvirus (KSHV) (Hansen et al. 2010).
MicroRNAs have been implicated in the origins of certain neurological disorders
including Alzheimer’s disease and Parkinson’s disease. One of the mechanisms
controlling expression of amyloid precursor responsible for Alzheimer’s depends
on several miR-20a family microRNAs (miR-20a, miR-17-5p and miR-106b),
and decreased levels of miR-29a/b-1 and miR-106b have been reported in brains
of Alzheimer’s disease patients (Hebert et al. 2008a; Hebert et al. 2008b). In
Parkinson’s disease patients, there are reduced levels of miR-133b essential for
development and functions of midbrain dopaminergic neurons (Kim et al. 2007).
MicroRNA miR-181b overexpressed in the cortex of schizophrenia patients may
contribute to etiology of this disorder. The putative miR-181b targets include genes
essential for neuronal functions (e.g., receptors involved in synaptic transmission),
development of nervous system and cell differentiation, and the experimental
verification indicated two mRNAs encoding ionotropic glutamate receptor (GRIA2)
and calcium sensor gene visinin-like 1 (VSNL1). GRIA2 and VSNL1 are crucial for
neuronal functions involved in signal transduction and synaptic plasticity and have
been earlier linked to schizophrenia (Beveridge et al. 2008).
Pathologies linked to microRNAs may also result from mutations changing
microRNA-binding sites within 30 -UTRs of target genes. One of the genes regulated
by let-7 microRNA is HMGA2 (High Mobility Group A2) protein showing elevated
expression in several human tumors and involved in remodeling of chromatin.
Chromosomal rearrangements or mutations disrupting let-7 binding result in onco-
genic transformation (Mayr et al. 2007; Lee and Dutta 2007; Klemke et al. 2010). In
Parkinson’s disease, an overexpression of the fibroblast growth factor 20 (FGF20) is
due to a single nucleotide substitution within the 30 -UTR of FGF20 mRNA that
affects miR-433 target site resulting in increased expression of FGF20 protein
in vitro and in vivo (Wang et al. 2008). A polymorphism at the microRNA target
site within SLITRK1 (SLIT and Trk-like 1) gene responsible for growth of neurons is
responsible for the origin of the Touret’s syndrome (Abelson et al. 2005).
The list of diseases and molecular pathways that have been linked to changes in
microRNA expression is growing. MicroRNAs have been implicated in the origin
of such conditions as diabetes (Tang et al. 2008), obesity (Xie et al. 2009),
cardiovascular diseases (Mishra et al. 2009), autoimmunological disorders (Pauley
et al. 2009), and many others.
2.2 mRNA-Like ncRNAs
Although currently there are limited data concerning precise functions of most of
ncRNAs, there is a growing body of evidence that most of these transcripts are in
fact functional and participate in the regulation of a wide array of cellular processes
(Mattick 2009b). It is, therefore, not surprising that the aberrant expression of many
of them was found to be associated with pathological states.
Deregulated expression of imprinted genes including those encoding ncRNAs is
linked to severe congenital, developmental disorders. At least in some cases, the
RNA is responsible for establishing and maintaining monoallelic expression of
imprinted genes by inducing epigenetic changes as demonstrated for LIT1 or Air
RNAs (for details see Sect. 1.1).
Abnormal expression of imprinted ncRNAs has been observed in various forms
of human cancer. H19 RNA is a product of a maternally expressed gene on
chromosome 11p15.5 located downstream from the paternally expressed IGF2
(insulin like growth factor 2) gene. H19 RNA is expressed primarily during
embryonic development in most fetal tissues (Gabory et al. 2006, 2009). Biallelic
expression of either gene due to epigenetic changes in the differentially methylated
region (DMR) upstream of the H19 can cause malignant cell growth (Manoharan
et al. 2004). Because the effect of H19 RNA differs in various forms of cancer, it
was described either as a tumor suppressor reducing tumorigenicity and growth
(Hao et al. 1993) or as an oncogene promoting cancer progression (Fellig et al.
2005; Berteaux et al. 2005; Lottin et al. 2005). It has been proposed that H19 RNA
may have different effect depending on its alternative splicing or interactions with
RNA-binding proteins (Matouk et al. 2004; Ioannidis et al. 2004). The biological
effect may also depend on microRNA miR-675 for which H19 plays a role of a host
gene (Cai and Cullen 2007). The same imprinted genes cluster produces a paternal-
expressed PEG8/IGF2AS RNA (paternally expressed gene 8, IGF2 antisense) that
shows elevated expression in several fetal cancers (Okutsu et al. 2000). In colorectal
and esophageal cancers, the epigenetic changes altering the imprinting status
and the expression of the LIT1 RNA were observed (Nakano et al. 2006; Soejima
et al. 2004).
An involvement of imprinted ncRNA in tumor suppression was described for the
transcript from the MEG3 (maternally expressed gene 3), highly expressed in
human pituitary whose expression is lost in pituitary adenomas and most human
cancer cell lines. Expression of MEG3 RNA in cancer cells stimulates expression of
p53 followed by activation of p53 downstream targets. It has been demonstrated
that the effect depends on proper folding of MEG3 RNA that is required for p53
activation and growth suppression (Zhang et al. 2009).
Long mRNA-like ncRNAs were often identified as transcripts showing elevated
expression in malignant cells. One of the molecular markers of metastasis in lung
adenocarcinoma is MALAT-1 RNA (metastasis associated in lung adenocarcinoma
transcript 1) (Ji et al. 2003). MALAT-1 is expressed in normal human and mouse
tissues (Ji et al. 2003; Hutchinson et al. 2007) and its overexpression is observed in
human breast, pancreas, lung, colon, prostate, and liver carcinomas (Lin et al. 2006).
Several cancer-associated translocations were shown to disrupt the MALAT-1 locus
(Davis et al. 2003; Kuiper et al. 2003; Rajaram et al. 2007). MALAT-1 was also
identified as a nuclear-enriched abundant transcripts (NEAT2) present in SC35
nuclear speckles (Hutchinson et al. 2007).
In over 90% of prostate cancers, there is a significant overexpression of several
alternatively spliced isoforms of DD3/PCA3 RNA (prostate cancer antigen 3)
(Bussemakers et al. 1999; de Kok et al. 2002). Another prostate-specific transcript
with markedly elevated levels in malignant cells is PCGEM1 RNA (Srikantan et al.
2000) expression, of which correlates with increased cell proliferation and colony
formation (Petrovics et al. 2004). Another effect of PCGEM1 observed in LNCaP
cell line was a resistance to doxorubicin-induced apoptosis and delayed induction
of p53 and p21 (Fu et al. 2006).
An ncRNA, CUDR (cancer up-regulated drug resistant), was identified in sev-
eral human cancer cell lines including hepatocellular, breast, colon, and lung
carcinomas as well as HeLa cells. CUDR expression correlates with resistance to
apoptosis-inducing drugs doxorubicin and etoposide. Although the molecular basis
of drug-resistance is not known, it has been proposed that CUDR RNA is involved
in the downregulation of caspase 3 that shows reduced levels in CUDR-expressing
cells (Tsang et al. 2007). As a very strong and specific ncRNA marker of cancer-
ogenesis, CUDR may represent a good target for the development of anticancer
drugs.
Several other mRNA-like ncRNAs have also been identified as cancer-associated
transcripts. The NCRMS (noncoding RNA in rhabdomyosarcoma (RMS)) shows
elevated expression in alveolar RMS, neuroblastoma, and synovial sarcoma (Chan
et al. 2002). OCC-1 RNA (overexpressed in colon carcinoma 1) is a specific marker
of colon carcinoma (Pibouin et al. 2002).
The number of novel ncRNAs with altered expression in cancers is growing
each year. A thorough analysis of expressed sequence tags (ESTs) from head, neck,
and thyroid allowed identification of new intronic, antisense RNAs (Reis et al.
2005). Another source of novel ncRNAs are the UCRs discovered by comparing
mammalian genomes (Bejerano et al. 2004) that were shown to be highly tran-
scribed and produce ncRNAs. An analysis of expression of 481 UCRs in CLL
demonstrated that they are differentially expressed in normal and malignant human
cells (Calin et al. 2007). Moreover, specific expression profiles of UCRs were
consistent with poor and good prognosis.
Human endogenous retroviruses (HERVs) were also implicated in cancerogenesis
(Galli et al. 2005; Mangeney et al. 2005; de Parseval et al. 1999). In patients with
bladder transitional cell carcinoma (TCC), an ncRNA UCA1 (urothelial carcinoma
associated 1) belonging to a HERV-H family was identified as a specific, highly
expressed molecular marker (Wang et al. 2006).
Abnormal expression of long ncRNAs is not limited to cancer cells alone.
A specific ncRNA PRINS RNA (psoriasis susceptibility-related RNA gene induced
by stress) is overexpressed in response to UV-B irradiation or viral infection in
patients with psoriasis (Sonkoly et al. 2005). MIAT RNA is associated with
increased risk of myocardial infarction (Ishii et al. 2006). Silencing of a-globin

gene (HBA2) through methylation induced by an antisense transcript was found to
be the cause of a-thalassemia. Due to chromosomal deletion, a truncated transcript
from the LUC7L gene is produced that induces methylation of the CpG island
controlling expression of the HBA2 at early stages of the development (Tufarelli
et al. 2003).
Transcriptional suppression of the FXN gene responsible for Friedreich ataxia
(FRDA) is caused by expanded GAA triplet-repeat sequence within FXN intron 1.
An antisense RNA FAST-1 (FXN Antisense Transcript – 1) was shown to be highly
upregulated in patients with FRDA. FAST-1 overlaps chromatin insulator protein
CTCF (CCCTC-binding factor)-binding site in the FXN 50 -UTR. Depletion of
CTCF results in enrichment of H3 histone with trimethylated lysine at position 9
and recruitment of heterochromatin protein 1. The consequence of these epigenetic
changes is heterochromatin formation and transcriptional silencing of the FXN
gene (De Biase et al. 2009).
In DiGeorge syndrome (DGS) critical region (DGCR), a DGCR5 gene produces
several alternatively spliced ncRNAs expressed during human and mouse embryo-
genesis (Sutherland et al. 1996). Another ncRNA implicated in DGS is antisense
22k48 transcribed from the first intron of HIRA, a gene deleted in DGS. 22k48 is
expressed in neurons as several alternatively spliced RNAs (Pizzuti et al. 1999). In
case of an autistic patient, a translocation t(7;13)(q31.2;q21) disrupts a complex
RAY1/ST7 locus encoding two proteins ST7 and RAY1 and a series of sense and
antisense ncRNAs (Vincent et al. 2000). Translocations disrupting ncRNA coding
genes DISC2 (disrupted in schizophrenia 2) and PSZA11q14 (putative schizophrenia
associated gene from 11q14) have been identified in schizophrenia (Millar et al.
2000; Polesskaya et al. 2003). Both DISC2 and PSZA11q14 overlap protein-coding
genes and have been proposed to act as cis-acting antisense regulators, but so far
there are no data that would confirm such hypotheses. Expression of antisense
transcript for the b-secretase-1 (BACE1) has been also proposed to contribute to
pathogenesis of Alzheimer’s disease (Faghihi et al. 2008).
2.3 Other Transcripts
Certain human pathologies are also associated with abnormal expressions of other
classes of ncRNAs including snoRNA and RNA polymerase III transcribed BC200
RNA. Imprinted expression abnormalities of the genes within the human chromo-
some 15q11–q13 are associated with Angelman (AS) and Prade–Willi syndromes
(PWS) that are caused by maternal and paternal deficiencies, respectively
(Horsthemke and Wagstaff 2008). Duplications of this region have been also
found in about 1% of autistic patients (Koochek et al. 2006). The paternal allele
specifies a large neuron-restricted transcript several hundred kilobases long encoding
splicesomal protein (SNRPN), several small nucleolar RNAs and the antisense
transcript to an ubiquitin ligase (UBE3A-AS) expressed from the maternal allele
(Horsthemke and Wagstaff 2008). Although the deficiency of one of the paternally
expressed snoRNAs (HBII-85) is regarded as a primary cause of PWS, there is
still little known about the mechanisms underlying it (Sahoo et al. 2008; de Smith
et al. 2009).
The deletion involving a region in 6q14–q16 frequently found in breast and
prostate cancer patients was found to contain a gene encoding U50 snoRNA
(Verhagen et al. 2002; Dong et al. 2009). Transcriptional downregulation and
genomic deletions of the U50 was confirmed in a number of breast cancer cell
lines. U50 snoRNA can be considered a tumor suppressor as its expression resulted
in the inhibition of colony formation in breast cancer cell lines (Dong et al. 2009).
Another snoRNA-encoding gene that might be involved in regulation of tumor
growth is GAS5 RNA (growth arrest-specific transcript 5), a snoRNA host gene that
induces growth arrest and apoptosis (Mourtada-Maarabouni et al. 2009).
3 NcRNA-Based Therapeutic Strategies
The involvement of ncRNAs, and microRNAs in particular, in many human dis-

eases prompted research on the development of therapeutics that would target
microRNAs or microRNA-regulated pathways. An obvious advantage of targeting
microRNAs is their involvement in a wide spectrum of complex pathways and their
influence on expression of many different genes (Esau et al. 2006). This approach
contrasts with most of current strategies focused on specific targets, either on
mRNA or protein levels.
Depending on the role of a given microRNA in the regulation of cancer-related
genes, the strategies may involve either knocking down overexpression of onco-
genic microRNAs or restoring expression of tumor suppressor microRNA. Anti-
microRNA therapies can affect expression of multiple genes contributing to
malignant growth as was demonstrated for the miR-15a/miR-16-1 cluster that targets
two antiapoptotic oncogenes BCL2 and MCL1 as well as other genes involved in
carcinogenesis including Jun, MSH2, or WT1 (Wilms tumor 1) (Calin et al. 2008).
Currently, most of the strategies aimed at the downregulation of microRNA
levels are based on the application of modified antisense oligonucleotides – antag-
omirs – which are able to inhibit functions of specific microRNAs by preventing
their binding to target mRNAs (Davis et al. 2006). The modifications of ribose
20 -hydroxyl groups by 20 -O-methyl, 20 -O-methoxyethyl or locked nucleic acid
(LNA) and backbone modification with phosphorothioate provide higher affinity
for target RNAs and resistance to nucleases, which is a crucial factor that has to be
considered in the development of RNA-based therapeutics. However, it has been
shown that the modifications, while generally increasing the affinity and nuclease
resistance, show differences in effectiveness depending on their position in the
antisense oligonucleotide (Davis et al. 2006).
An oncogenic microRNA miR-21 is overexpressed in many forms of cancer and
shows proproliferative and antiapoptotic activity by targeting p53- and transforming
growth factor-b (TGFb)-dependent tumor suppression pathways as well as apoptosis-

related genes (Papagiannakopoulos et al. 2008; Chen et al. 2008). Inactivation
of miR-21 with antagomirs in glioblastoma resulted in activation of caspase and
apoptosis (Chan et al. 2005). It has been also shown that overexpression of miR-23a-
27a-24-2 cluster in human embryonic kidney HEK293T cells resulted in induction of
apoptosis by both caspase-dependent and independent pathways and sensitization to
TNF-a cytotoxicity (Chhabra et al. 2009).
Therapeutic potential of microRNA overexpression was demonstrated in glio-
blastoma, which shows downregulation of miR-34a. The result of transfection of
miR-34a into human glioma and medulloblastoma cells was a significant reduction
of c-Met protein and inhibition of tumor growth (Li et al. 2009). Downregulation or
the absence of human miR-34a resulting from deletion within 1p36 is found in
cultured cell lines and primary tumors. Transfection of these cell lines with
precursors of miR-34a leads to inhibition of cell growth and induction of apoptosis
probably through downregulation of BCL2 and MYCN oncogenes (Cole et al.
2008). On the other hand, MYCN trans-activate expression of the microRNA
17-5p-92 cluster that has oncogenic properties increasing proliferation and colony
formation through downregulation of p21 and BIM tumor suppressors implicated in
negative regulation of cell cycle progression and apoptosis. Application of antag-
omir against miR17-5p rescued expression of p21 and BIM proteins resulting in
apoptosis and inhibited proliferation (Fontana et al. 2008).
Although most of the proof of principle work was done in cultured cells, the
feasibility of antimicroRNA therapeutics have also been demonstrated in vivo.
LNA-antimiR oligonucleotides targeting miR-122 injected into mouse were
efficiently delivered into liver and repressed miR-122-regulated genes (Elmén
et al. 2008). Similar results were obtained with inhibition of miR-219, which
modulated behavioral responses associated with disrupted NMDA receptor trans-
mission (Kocerha et al. 2009).
Because the aberrant expression of many microRNAs is encountered in a wide
spectrum of human diseases, antagomirs can be envisioned as new generation
therapeutics for as diverse disorders as cancer and Down syndrome (Kuhn et al.
2008).
Another approach to regulation of microRNA function is application of decoy
molecules mimicking mRNAs’ 30 -UTRs with multiple target sites for particular
microRNAs.
Such microRNA sponges act as competitive inhibitors, binding microRNAs and
preventing their binding to mRNAs 30 -UTRs. The efficiency of such approach was
demonstrated to be as high as antisense antagomirs (Ebert et al. 2007). Moreover,
the utilization of microRNA-binding sites allows regulation of the activity of
related microRNA species belonging to the same family and recognizing the
same targets (Ebert et al. 2007; Care et al. 2007).
Yet another strategy is based on application of antisense oligonucleotides
that would mask microRNA-binding sites within 30 -UTR of target mRNA. LNA-
modified oligonucleotides complementary to the miRNA binding sites in the 30 UTR
of cardiac pacemaker channel genes HCN2 and HCN4 significantly upregulated

HCN2/HCN4 expression (Xiao et al. 2007).
The efficiency of microRNA-based regulation also prompted research on the
possibility of custom-made, artificial microRNAs that would display selective
activity. In human cells, an artificial microRNA targeting chemokine receptor
CXCR4 was shown to effectively inhibit protein expression (Liang et al. 2007).
This approach, however, requires better understanding of all aspects of microRNA
regulation, which we are only beginning to comprehend.
Acknowledgments This work was supported by grants from the Polish Ministry of Science and
Higher Education.
References
Abelson JF, Kwan KY, O’Roak BJ et al (2005) Sequence variants in SLITRK1 are associated with
Tourette’s syndrome. Science 310:317–320
Abrajano JJ, Qureshi IA, Gokhan S et al (2009) REST and CoREST modulate neuronal subtype
specification, maturation and maintenance. PLoS One 4:e7936
Allen TA, Von Kaenel S, Goodrich JA et al (2004) The SINE-encoded mouse B2 RNA represses
mRNA transcription in response to heat shock. Nat Struct Mol Biol 11:816–821
Anckar J, Sistonen L (2007) Heat shock factor 1 as a coordinator of stress and developmental
pathways. Adv Exp Med Biol 594:78–88
Annilo T, Kepp K, Laan M (2009) Natural antisense transcript of natriuretic peptide precursor A
(NPPA): structural organization and modulation of NPPA expression. BMC Mol Biol 10:81
Bandrés E, Cubedo E, Agirre X et al (2006) Identification by Real-time PCR of 13 mature
microRNAs differentially expressed in colorectal cancer and non-tumoral tissues. Mol Cancer
5:29
Barboric M, Lenasi T, Chen H et al (2009) 7SK snRNP/P-TEFb couples transcription elongation
with alternative splicing and is essential for vertebrate development. Proc Natl Acad Sci USA
106:7798–7803
Beilharz TH, Humphreys DT, Clancy JL et al (2009) microRNA-mediated messenger RNA
deadenylation contributes to translational repression in mammalian cells. PLoS One 4:e6783
Beiter T, Reich E, Williams RW et al (2009) Antisense transcription: a critical look in both
directions. Cell Mol Life Sci 66:99–112
Bejerano G, Pheasant M, Makunin I et al (2004) Ultraconserved elements in the human genome.
Science 304:1321–1325
Beltran M, Puig I, Pena C et al (2008) A natural antisense transcript regulates Zeb2/Sip1 gene
expression during Snail1-induced epithelial–mesenchymal transition. Genes Dev 22:756–769
Berteaux N, Lottin S, Monté D et al (2005) H19 mRNA-like noncoding RNA promotes breast
cancer cell proliferation through positive control by E2F1. J Biol Chem 280:29625–29636
Beveridge NJ, Tooney PA, Carroll AP et al (2008) Dysregulation of miRNA 181b in the temporal
cortex in schizophrenia. Hum Mol Genet 17:1156–1168
Blenkiron C, Goldstein LD, Thorne NP et al (2007) MicroRNA expression profiling of human
breast cancer identifies new markers of tumour subtype. Genome Biol 8:R214
Bond AM, Vangompel MJ, Sametsky EA et al (2009) Balanced gene regulation by an embryonic
brain ncRNA is critical for adult hippocampal GABA circuitry. Nat Neurosci 12:1020–1027
Bussemakers MJ, van Bokhoven A, Verhaegh GW et al (1999) DD3: a new prostate-specific gene,
highly overexpressed in prostate cancer. Cancer Res 59:5975–5979
C. elegans Sequencing Consortium (1998) Genome sequence of the nematode C. elegans: a

platform for investigating biology. Science 282:2012–2018
Cai X, Cullen BR (2007) The imprinted H19 noncoding RNA is a primary microRNA precursor.
RNA 13:313–316
Calin GA, Cimmino A, Fabbri M et al (2008) MiR-15a and miR-16-1 cluster functions in human
leukemia. Proc Natl Acad Sci USA 105:5166–5171
Calin GA, Liu CG, Ferracin M et al (2007) Ultraconserved regions encoding ncRNAs are altered
in human leukemias and carcinomas. Cancer Cell 12:215–229
Calin GA, Sevignani C, Dumitru CD et al (2004) Human microRNA genes are frequently located at
fragile sites and genomic regions involved in cancers. Proc Natl Acad Sci USA 101:2999–3004
Care A, Catalucci D, Felicetti F et al (2007) MicroRNA-133 controls cardiac hypertrophy. Nat
Med 13:613–618
Caretti G, Schiltz RL, Dilworth FJ et al (2006) The RNA helicases p68/p72 and the noncoding
RNA SRA are coregulators of MyoD and skeletal muscle differentiation. Dev Cell 11:547–560
Carninci P, Kasukawa T, Katayama S et al (2005) The transcriptional landscape of the mammalian
genome. Science 309:1559–1563
Chhabra R, Adlakha YK, Hariharan M, Scaria V, Saini N (2009) Upregulation of miR-23a-27a-
24-2 cluster induces caspase-dependent and -independent apoptosis in human embryonic
kidney cells. PLoS One 4:e5848
Chan AS, Thorner PS, Squire JA et al (2002) Identification of a novel gene NCRMS on
chromosome 12q21 with differential expression between Rhabdomyosarcoma types. Onco-
gene 21:3029–3037
Chan JA, Krichevsky AM, Kosik KS (2005) MicroRNA-21 is an antiapoptotic factor in human
glioblastoma cells. Cancer Res 65:6029–6033
Chau YM, Pando S, Taylor HS (2002) HOXA11 silencing and endogenous HOXA11 antisense
ribonucleic acid in the uterine endometrium. J Clin Endocrinol Metab 87:2674–2680
Chaumeil J, Le Baccon P, Wutz A et al (2006) A novel role for Xist RNA in the formation of a
repressive nuclear compartment into which genes are recruited when silenced. Genes Dev
20:2223–2231
Chen Y, Liu W, Chao T et al (2008) MicroRNA-21 down-regulates the expression of tumor
suppressor PDCD4 in human glioblastoma cell T98G. Cancer Lett 272:197–205
Chen K, Rajewsky N (2007) The evolution of gene regulation by transcription factors and
microRNAs. Nat Rev Genet 8:93–103
Chendrimada TP, Finn KJ, Ji X et al (2007) MicroRNA silencing through RISC recruitment of
eIF6. Nature 447:823–828
Cheng J, Kapranov P, Drenkow J et al (2005) Transcriptional maps of 10 human chromosomes at
5-nucleotide resolution. Science 308:1149–1154
Chooniedass-Kothari S, Hamedani MK, Troup S et al (2006) The steroid receptor RNA activator
protein is expressed in breast tumor tissues. Int J Cancer 118:1054–1059
Chooniedass-Kothari S, Emberley E, Hamedani MK et al (2004) The steroid receptor RNA
activator is the first functional RNA encoding a protein. FEBS Lett 566:43–47
Chow J, Heard E (2009) X inactivation and the complexities of silencing a sex chromosome. Curr
Opin Cell Biol 21:359–366
Ciafre SA, Galardi S, Mangiola A et al (2005) Extensive modulation of a set of microRNAs in
primary glioblastoma. Biochem Biophys Res Commun 334:1351–1358
Cimmino A, Calin GA, Fabbri M et al (2005) miR-15 and miR-16 induce apoptosis by targeting
BCL2. Proc Natl Acad Sci USA 102:13944–13949
Clark TA, Schweitzer AC, Chen TX et al (2007) Discovery of tissue-specific exons using
comprehensive human exon microarrays. Genome Biol 8:R64
Cloonan N, Brown MK, Steptoe AL et al (2008) The miR-17-5p microRNA is a key regulator of
the G1/S phase cell cycle transition. Genome Biol 9:R127
Cole KA, Attiyeh EF, Mosse YP et al (2008) A functional screen identifies miR-34a as a candidate
neuroblastoma tumor suppressor gene. Mol Cancer Res 6:735–742
Davis IJ, His B, Arroyo JD et al (2003) Cloning of an alpha-TFEB fusion in renal tumors
harboring the t(6;11)(p21;q13) chromosome translocation. Proc Natl Acad Sci USA 100:
6051–6056
Davis S, Lollo B, Freier S et al (2006) Improved targeting of miRNA with antisense oligonucleo-
tides. Nucleic Acids Res 34:2294–2304
De Biase I, Chutake YK, Rindler PM et al (2009) Epigenetic silencing in Friedreich ataxia is
associated with depletion of CTCF (CCCTC-binding factor) and antisense transcription. PLoS
One 4:e7914
de Kok JB, Verhaegh GW, Roelofs RW et al (2002) DD3(PCA3), a very sensitive and specific
marker to detect prostate tumors. Cancer Res 62:2695–2698
de Napoles M, Mermoud JE, Wakao R et al (2004) Polycomb group proteins Ring1A/B link
ubiquitylation of histone H2A to heritable gene silencing and X inactivation. Dev Cell
7:663–676
de Parseval N, Alkabbani H, Heidmann T (1999) The long terminal repeats of the HERV-H human
endogenous retrovirus contain binding sites for transcriptional regulation by the Myb protein.
J Gen Virol 80:841–845
de Smith AJ, Purmann C, Walters RG et al (2009) A deletion of the HBII-85 class of small
nucleolar RNAs (snoRNAs) is associated with hyperphagia, obesity and hypogonadism. Hum
Mol Genet 18:3257–3265
DeBaun MR, Niemitz EL, McNeil DE et al (2002) Epigenetic alterations of H19 and LIT1
distinguish patients with Beckwith–Wiedemann syndrome with cancer and birth defects. Am
J Hum Genet 70:604–611
Dong XY, Guo P, Boyd J et al (2009) Implication of snoRNA U50 in human breast cancer. J Genet
Genomics 36:447–454
Duning K, Buck F, Barnekow A et al (2008) SYNCRIP, a component of dendritically localized
mRNPs, binds to the translation regulator BC200 RNA. J Neurochem 105:351–359
Ebert MS, Neilson JR, Sharp PA (2007) MicroRNA sponges: competitive inhibitors of small
RNAs in mammalian cells. Nat Methods 4:721–726
Elmén J, Lindow M, Silahtaroglu A et al (2008) Antagonism of microRNA-122 in mice by
systemically administered LNA-antimiR leads to up-regulation of a large set of predicted
target mRNAs in the liver. Nucleic Acids Res 36:1153–1162
Emberley E, Huang GJ, Hamedani MK et al (2003) Identification of new human coding steroid
receptor RNA activator isoforms. Biochem Biophys Res Commun 301:509–515
Enklaar T, Zabel BU, Prawitt D (2006) Beckwith–Wiedemann syndrome: multiple molecular
mechanisms. Expert Rev Mol Med 8:1–19
Esau C, Davis S, Murray SF et al (2006) miR-122 regulation of lipid metabolism revealed by
in vivo antisense targeting. Cell Metab 3:87–98
Espinoza CA, Allen TA, Hieb AR et al (2004) B2 RNA binds directly to RNA polymerase to
repress transcript synthesis. Nat Struct Mol Biol 11:822–829
Fabian MR, Mathonnet G, Sundermeier T et al (2009) Mammalian miRNA RISC recruits CAF1
and PABP to affect PABP-dependent deadenylation. Mol Cell 35:868–880
Faghihi MA, Modarresi F, Khalil AM et al (2008) Expression of a noncoding RNA is elevated in
Alzheimer’s disease and drives rapid feed-forward regulation of beta-secretase. Nat Med
14:723–730
Fellig Y, Ariel I, Ohana P et al (2005) H19 expression in hepatic metastases from a range of human
carcinomas. J Clin Pathol 58:1064–1068
Feng J, Bi C, Clark BS et al (2006) The Evf-2 noncoding RNA is transcribed from the Dlx-5/6
ultraconserved region and functions as a Dlx-2 transcriptional coactivator. Genes Dev
20:1470–1484
Fontana L, Fiori ME, Albini S et al (2008) Antagomir-17-5p abolishes the growth of therapy-
resistant neuroblastoma through p21 and BIM. PLoS One 3:e2236
Frankel LB, Christoffersen NR, Jacobsen A et al (2008) Programmed cell death 4 (PDCD4) is an
important functional target of the microRNA miR-21 in breast cancer cells. J Biol Chem
283:1026–1033
Fu X, Ravindranath L, Tran N et al (2006) Regulation of apoptosis by a prostate-specific and
prostate cancer-associated noncoding gene, PCGEM1. DNA Cell Biol 25:135–141
Gabory A, Ripoche MA, Le Digarcher A (2009) H19 acts as a trans regulator of the imprinted gene
network controlling growth in mice. Development 136:3413–3421
Gabory A, Ripoche MA, Yoshimizu T et al (2006) The H19 gene: regulation and function of a non-
coding RNA. Cytogenet Genome Res 113:188–193
Galli UM, Sauter M, Lecher B et al (2005) Human endogenous retrovirus rec interferes with germ
cell development in mice and may cause carcinoma in situ, the predecessor lesion of germ cell
tumors. Oncogene 24:3223–3228
Hansen A, Henderson S, Lagos D et al (2010) KSHV-encoded miRNAs target MAF to induce
endothelial cell reprogramming. Genes Dev 24:195–205
Hao Y, Crenshaw T, Moulton T et al (1993) Tumour-suppressor activity of H19 RNA. Nature
365:764–767
Hatchell EC, Colley SM, Beveridge DJ et al (2006) SLIRP, a small SRA binding protein, is a
nuclear receptor corepressor. Mol Cell 22:657–668
Hebert SS, Horre K, Nicolai L et al (2008a) MicroRNA regulation of Alzheimer’s amyloid
precursor protein expression. Neurobiol Dis. doi:10.1016/j.nbd.2008.11.009
Hebert SS, Horre K, Nicolai L et al (2008b) Loss of microRNA cluster miR-29a/b-1 in sporadic
Alzheimer’s disease correlates with increased BACE1/beta-secretase expression. Proc Natl
Acad Sci USA 105:6415–6420
Hobert O (2004) Common logic of transcription factor and microRNA action. Trends Biochem Sci
9:462–468
Hon LS, Zhang Z (2007) The roles of binding site arrangement and combinatorial targeting in
microRNA repression of gene expression. Genome Biol 8:R166
Horike S, Mitsuya K, Meguro M et al (2000) Targeted disruption of the human LIT1 locus defines
a putative imprinting control element playing an essential role in Beckwith–Wiedemann
syndrome. Hum Mol Genet 9:2075–2083
Horsthemke B, Wagstaff J (2008) Mechanisms of imprinting of the Prader–Willi/Angelman
region. Am J Med Genet 146A:2041–2052
Hutchinson JN, Ensminger AW, Clemson CM et al (2007) A screen for nuclear transcripts
identifies two linked noncoding RNAs associated with SC35 splicing domains. BMC Geno-
mics 8:39
Iida T, Nakayama J, Moazed D (2008) siRNA-mediated heterochromatin establishment requires
HP1 and is associated with antisense transcription. Mol Cell 31:178–189
Imamura T, Yamamoto S, Ohgane J et al (2004a) Non-coding RNA directed DNA demethylation
of Sphk1 CpG island. Biochem Biophys Res Commun 322:593–600
Imamura T, Miyauchi-Senda N, Tanaka S et al (2004b) Identification of genetic and epigenetic
similarities of SPHK1/Sphk1 in mammals. J Vet Med Sci 66:1387–1393
Imamura T, Ohgane J, Ito S et al (2001) CpG island of rat sphingosine kinase-1 gene: tissue-
dependent DNA methylation status and multiple alternative first exons. Genomics 76:113–125
Ioannidis P, Kottaridi C, Dimitriadis E et al (2004) Expression of the RNA-binding protein CRD-
BP in brain and non-small cell lung tumors. Cancer Lett 209:245–250
Ishii N, Ozaki K, Sato H et al (2006) Identification of a novel non-coding RNA, MIAT, that
confers risk of myocardial infarction. J Hum Genet 51:1087–1099
Ji P, Diederichs S, Wang W et al (2003) MALAT-1, a novel noncoding RNA, and thymosin beta4
predict metastasis and survival in early-stage non-small cell lung cancer. Oncogene 22:8031–8041
Johnson SM, Grosshans H, Shingara J et al (2005) RAS is regulated by the let-7 microRNA family.
Cell 120:635–647
Johnson CD, Esquela-Kerscher A, Stefani G et al (2007) The let-7 MicroRNA represses cell
proliferation pathways in human cells. Cancer Res 67:7713–7722
Jonkers I, Monkhorst K, Rentmeester E et al (2008) Xist RNA is confined to the nuclear territory of
the silenced X chromosome throughout the cell cycle. Mol Cell Biol 28:5583–5594
Kanduri C, Thakur N, Pandey RR (2006) The length of the transcript encoded from the Kcnq1ot1
antisense promoter determines the degree of silencing. EMBO J 25:2096–2102
Karube Y, Tanaka H, Osada H et al (2005) Reduced expression of Dicer associated with poor
prognosis in lung cancer patients. Cancer Sci 96:111–155
Kawashima H, Takano H, Sugita S et al (2003) A novel steroid receptor co-activator protein
(SRAP) as an alternative form of steroid receptor RNA-activator gene: expression in prostate
cancer cells and enhancement of androgen receptor activity. Biochem J 369:163–171
Kent OA, Mendell JT (2006) A small piece in the cancer puzzle: microRNAs as tumor suppressors
and oncogenes. Oncogene 25:6188–6196
Kim J, Inoue K, Ishii J et al (2007) A MicroRNA feedback circuit in midbrain dopamine neurons.
Science 317:1220–1224
Kiriakidou M, Tan GS, Lamprinaki S et al (2007) An mRNA m7G cap binding-like motif within
human Ago2 represses translation. Cell 129:1141–1151
Kishore S, Khanna A, Zhang Z et al (2010) The snoRNA MBII-52 (SNORD 115) is processed into
smaller RNAs and regulates alternative splicing. Hum Mol Genet 19(7):1153–1164
Kishore S, Stamm S (2006) The snoRNA HBII-52 regulates alternative splicing of the serotonin
receptor 2C. Science 311:230–232
Klemke M, Meyer A, Hashemi Nezhad M et al (2010) Loss of let-7 binding sites resulting from
truncations of the 30 untranslated region of HMGA2 mRNA in uterine leiomyomas. Cancer
Genet Cytogenet 196:119–123
Kocerha J, Faghihi MA, Lopez-Toledano MA et al (2009) MicroRNA-219 modulates NMDA
receptor-mediated neurobehavioral dysfunction. Proc Natl Acad Sci USA 106:3507–3512
Koochek M, Harvard C, Hildebrand MJ et al (2006) 15q duplication associated with autism in a
multiplex family with a familial cryptic translocation t(14;15)(q11.2;q13.3) detected using
array-CGH. Clin Genet 69:124–134
Krichevsky AM, Gabriely G (2009) miR-21: a small multi-faceted RNA. J Cell Mol Med
13:39–53
Krystal GW, Armstrong BC, Battey JF (1990) N-myc mRNA forms an RNA–RNA duplex with
endogenous antisense transcripts. Mol Cell Biol 10:4180–4191
Kuhn DE, Nuovo GJ, Martin MM et al (2008) Human chromosome 21-derived miRNAs are
overexpressed in down syndrome brains and hearts. Biochem Biophys Res Commun
370:473–477
Kuiper RP, Schepens M, Thijssen J et al (2003) Upregulation of the transcription factor TFEB in t
(6;11)(p21;q13)-positive renal cell carcinomas due to promoter substitution. Hum Mol Genet
12:1661–1669
Kumar MS, Lu J, Mercer KL et al (2007) Impaired microRNA processing enhances cellular
transformation and tumorigenesis. Nat Genet 39:673–677
Kuwabara T, Hsieh J, Nakashima K et al (2004) A small modulatory dsRNA specifies the fate of
adult neural stem cells. Cell 116:779–793
Lander ES, Linton LM, Birren B et al (2001) Initial sequencing and analysis of the human genome.
Nature 409:860–921
Lanz R, Razani B, Goldberg AD et al (2002) Distinct RNA motifs are important for coactivation of
steroid hormone receptors by steroid receptor RNA activator (SRA). Proc Natl Acad Sci USA
99:16081–16086
Lanz RB, McKenna NJ, Onate SA et al (1999) A steroid receptor coactivator, SRA, functions as an
RNA and is present in an SRC-1 complex. Cell 97:17–27
Lee YS, Dutta A (2007) The tumor suppressor microRNA let-7 represses the HMGA2 oncogene.
Genes Dev 21:1025–1030
Leygue E (2007) Steroid receptor RNA activator (SRA1): unusual bifaceted gene products with
suspected relevance to breast cancer. Nucl Recept Signal 5:e006
Li Y, Guessous F, Zhang Y et al (2009) MicroRNA-34a inhibits glioblastoma growth by targeting

multiple oncogenes. Cancer Res 69:7569–7576
Liang Z, Wu H, Reddy S et al (2007) Blockade of invasion and metastasis of breast cancer cells
via targeting CXCR4 with an artificial microRNA. Biochem Biophys Res Commun 363:
542–546
Lin D, Pestova TV, Hellen CU et al (2008) Translational control by a small RNA: dendritic BC1
RNA targets the eukaryotic initiation factor 4A helicase mechanism. Mol Cell Biol
28:3008–3019
Lin R, Maeda S, Liu C et al (2006) A large noncoding RNA is a marker for murine hepatocellular
carcinomas and a spectrum of human carcinomas. Oncogene 26:851–858
Linsley PS, Schelter J, Burchard J et al (2007) Transcripts targeted by the microRNA-16 family
cooperatively regulate cell cycle progression. Mol Cell Biol 27:2240–2252
Lottin S, Adriaenssens E, Berteaux N et al (2005) The human H19 gene is frequently over-
expressed in myometrium and stroma during pathological endometrial proliferative events. Eur
J Cancer 41:168–177
Nature 435:834–838
Mancini-Dinardo D, Steele SJ, Levorse JM et al (2006) Elongation of the Kcnq1ot1 transcript is
required for genomic imprinting of neighboring genes. Genes Dev 20:1268–1282
Mangeney M, Pothlichet J, Renard M et al (2005) Endogenous retrovirus expression is required for
murine melanoma tumor growth in vivo. Cancer Res 65:2588–2591
Manoharan H, Babcock K, Pitot HC (2004) Changes in the DNA methylation profile of the rat H19
gene upstream region during development and transgenic hepatocarcinogenesis and its role in
the imprinted transcriptional regulation of the H19 gene. Mol Carcinog 41:1–16
Mariner PD, Walters RD, Espinoza CA et al (2008) Human Alu RNA is a modular transacting
repressor of mRNA transcription during heat shock. Mol Cell 29:499–509
Markert A, Grimm M, Martinez J et al (2008) The La-related protein LARP7 is a component of the
7SK ribonucleoprotein and affects transcription of cellular and viral polymerase II genes.
EMBO Rep 9:569–575
Martignetti JA, Brosius J (1995) BC1 RNA: transcriptional analysis of a neural cell-specific RNA
polymerase III transcript. Mol Cell Biol 15:1642–1650
Matouk I, Ayesh B, Schneider T et al (2004) Oncofetal splice-pattern of the human H19 gene.
Biochem Biophys Res Commun 318:916–919
Mattick JS (1994) Introns: evolution and function. Curr Opin Genet Dev 4:823–831
Mattick JS (2001) Non-coding RNAs: the architects of eukaryotic complexity. EMBO Rep
2:986–991
Mattick JS (2003) Introns and noncoding RNAs the hidden layer of eukaryotic complexity. In:
Barciszewski J, Erdmann VA (eds) Noncoding RNAs: molecular biology and molecular
medicine. Eurekah, Austin, pp 12–33
Mattick JS (2009a) Deconstructing the dogma. A new view of the evolution and genetic program-
ming of complex organisms. Ann NY Acad Sci 1178:29–46
Mattick JS (2009b) The genetic signatures of noncoding RNAs. PLoS Genet 5:e1000459
Mattick JS, Makunin IV (2006) Non-coding RNA. Hum Mol Genet 15:R17–29
Mayr C, Hemann MT, Bartel DP (2007) Disrupting the pairing between let-7 and Hmga2 enhances
oncogenic transformation. Science 315:1576–1579
Meng F, Henson R, Wehbe-Janek H et al (2007) MicroRNA-21 regulates expression of the PTEN
tumor suppressor gene in human hepatocellular cancer. Gastroenterology 133:647–658
Mercer TR, Dinger ME, Mattick JS (2009) Long non-coding RNAs: insights into functions. Nat
Rev Genet 10:155–159
Michael MZ, O’ Connor SM, van Holst Pellekaan NG et al (2003) Reduced accumulation of
specific microRNAs in colorectal neoplasia. Mol Cancer Res 1:882–891
Millar JK, Wilson-Annan JC, Anderson S et al (2000) Disruption of two novel genes by a
translocation co-segregating with schizophrenia. Hum Mol Genet 9:1415–1423
Mishra PK, Tyagi N, Kumar M et al (2009) MicroRNAs as a therapeutic target for cardiovascular
diseases. J Cell Mol Med 13:778–789
Mohammad F, Mondal T, Kanduri C (2009) Epigenetics of imprinted long noncoding RNAs.
Epigenetics 4:277–286
Morison IM, Ramsay JP, Spencer HG (2005) A census of mammalian imprinting. Trends Genet
21:457–463
Mourtada-Maarabouni M, Pickard MR, Hedge VL et al (2009) GAS5, a non-protein-coding RNA,
controls apoptosis and is downregulated in breast cancer. Oncogene 28:195–208
Munroe SH, Lazar MA (1991) Inhibition of c-erbA mRNA splicing by a naturally occurring
antisense RNA. J Biol Chem 266:22083–22086
Murakami Y, Yasuda T, Saigo K et al (2006) Comprehensive analysis of microRNA expression
patterns in hepatocellular carcinoma and non-tumorous tissues. Oncogene 25:2537–2545
Nagano T, Mitchell JA, Sanz LA et al (2008) The air noncoding RNA epigenetically silences
transcription by targeting G9a to chromatin. Science 322:1717–1720
Nakano S, Murakami K, Meguro M et al (2006) Expression profile of LIT1/KCNQ1OT1 and
epigenetic status at the KvDMR1 in colorectal cancers. Cancer Sci 97:1147–1154
Newall AE, Duthie S, Formstone E et al (2001) Primary non-random X inactivation associated
with disruption of Xist promoter regulation. Hum Mol Genet 10:581–589
Nguyen VT, Kiss T, Michels AA et al (2001) 7SK small nuclear RNA binds to and inhibits the
activity of CDK9/cyclin T complexes. Nature 414:322–325
Niemitz EL, DeBaun MR, Fallon J et al (2004) Microdeletion of LIT1 in familial Beckwith-
Wiedemann syndrome. Am J Hum Genet 75:844–849
Okutsu T, Kuroiwa Y, Kagitani F et al (2000) Expression and imprinting status of human PEG8/
IGF2AS, a paternally expressed antisense transcript from the IGF2 locus, in Wilms’ tumors.
J Biochem 127:475–483
Papagiannakopoulos T, Shapiro A, Kosik KS (2008) MicroRNA-21 targets a network of key
tumor-suppressive pathways in glioblastoma cells. Cancer Res 68:8164–8172
Pauler FM, Koerner MV, Barlow DP (2007) Silencing by imprinted noncoding RNAs: is tran-
scription the answer? Trends Genet 23:284–292
Pauley KM, Cha S, Chan EK (2009) MicroRNA in autoimmunity and autoimmune diseases.
J Autoimmun 32:189–194
Payer B, Lee JT (2008) X chromosome dosage compensation: how mammals keep the balance.
Annu Rev Genet 42:733–772
Peter ME (2009) Let-7 and miR-200 microRNAs: guardians against pluripotency and cancer
progression. Cell Cycle 8:843–852
Petrovics G, Zhang W, Makarem M et al (2004) Elevated expression of PCGEM1, a prostate-
specific gene with cell growth-promoting function, is associated with high-risk prostate cancer
patients. Oncogene 23:605–611
Pibouin L, Villaudy J, Ferbus D et al (2002) Cloning of the mRNA of overexpression in colon
carcinoma-1: a sequence overexpressed in a subset of colon carcinomas. Cancer Genet
Cytogenet 133:55–60
Pillai RS, Bhattacharyya SN, Artus CG et al (2005) Inhibition of translational initiation by Let-7
microRNA in human cells. Science 309:1573–1576
Pizzuti A, Novelli G, Ratti A et al (1999) Isolation and characterization of a novel transcript
embedded within HIRA, a gene deleted in DiGeorge syndrome. Mol Genet Metabol 67:227–235
Plath K, Talbot D, Hamer KM et al (2004) Developmentally regulated alterations in Polycomb
repressive complex 1 proteins on the inactive X chromosome. J Cell Biol 167:1025–1035
Polesskaya OO, Haroutunian V, Davis KL et al (2003) Novel putative nonprotein-coding RNA
gene from 11q14 displays decreased expression in brains of patients with schizophrenia.
J Neurosci Res 74:111–122
Rajaram V, Knezevich S, Bove KE et al (2007) DNA sequence of the translocation breakpoints in
undifferentiated embryonal sarcoma arising in mesenchymal hamartoma of the liver harboring
the t(11;19)(q11;q13.4) translocation. Genes Chromosomes Cancer 46:508–513
Reis EM, Ojopi EP, Alberto FL et al (2005) Large-scale transcriptome analyses reveal new
genetic marker candidates of head, neck, and thyroid cancer. Cancer Res 65:1693–1699
Roldo C, Missiaglia E, Hagan JP et al (2006) MicroRNA expression abnormalities in pancreatic
endocrine and acinar tumors are associated with distinctive pathological features and clinical
behavior. J Clin Oncol 24:4677–4684
Sahoo T, del Gaudio D, German JR et al (2008) Prader–Willi phenotype caused by paternal
deficiency for the HBII-85 C/D box small nucleolar RNA cluster. Nat Genet 40:719–721
Sevignani C, Calin GA, Nnadi SC et al (2007) MicroRNA genes are frequently located near mouse
cancer susceptibility loci. Proc Natl Acad Sci USA 104:8017–8022
Shamovsky I, Ivannikov M, Kandel ES et al (2006) RNA-mediated response to heat shock in
mammalian cells. Nature 440:556–560
Shamovsky I, Nudler E (2009) Isolation and characterization of the heat shock RNA 1. Methods
Mol Biol 540:265–279
Shi Y, Downes M, Xie W et al (2001) SHARP, an inducible cofactor that integrates nuclear
receptor repression and activation. Genes Dev 15:1140–1151
Shiota K (2004) DNA methylation profiles of CpG islands for cellular differentiation and devel-
opment in mammals. Cytogenet Genome Res 105:325–334
Si ML, Zhu S, Wu H et al (2007) miR-21-mediated tumor growth. Oncogene 26:2799–2803
Sleutels F, Tjon G, Ludwig T et al (2003) Imprinted silencing of Slc22a2 and Slc22a3 does not
need transcriptional overlap between Igf2r and Air. EMBO J 22:3696–3704
Soejima H, Nakagawachi T, Zhao W et al (2004) Silencing of imprinted CDKN1C gene expres-
sion is associated with loss of CpG and histone H3 lysine 9 methylation at DMR-LIT1 in
esophageal cancer. Oncogene 23:4380–4388
Sonkoly E, Bata-Csorgo Z, Pivarcsi A et al (2005) Identification and characterization of a
novel, psoriasis susceptibility-related noncoding RNA gene, PRINS. J Biol Chem 280:
24159–24167
Srikantan V, Zou Z, Petrovics G et al (2000) PCGEM1, a prostate-specific gene, is overexpressed
in prostate cancer. Proc Natl Acad Sci USA 97:12216–12221
Stefani G, Slack FJ (2008) Small non-coding RNAs in animal development. Nat Rev Mol Cell Biol
9:219–230
Stern-Ginossar N, Elefant N, Zimmermann A et al (2007) Host immune system gene targeting by a
viral miRNA. Science 317:376–381
Sutherland HF, Wadey R, McKie JM et al (1996) Identification of a novel transcript disrupted by a
balanced translocation associated with DiGeorge syndrome. Am J Hum Genet 59:23–31
Szymanski M, Barciszewski J (2006) RNA regulation in mammals. Ann NY Acad Sci
1067:461–468
Taft RJ, Pheasant M, Mattick JS (2007) The relationship between non-protein-coding DNA and
eukaryotic complexity. Bioessays 29:288–299
Tang X, Tang G, Ozcan S (2008) Role of microRNAs in diabetes. Biochim Biophys Acta
1779:697–701
Terranova R, Yokobayashi S, Stadler MB et al (2009) Polycomb group proteins Ezh2 and Rnf2
direct genomic contraction and imprinted repression in early mouse embryos. Dev Cell
15:668–679
Thakur N, Tiwari VK, Thomassin H et al (2004) An antisense RNA regulates the bidirectional
silencing property of the Kcnq1 imprinting control region. Mol Cell Biol 24:7855–7862
Tiedge H, Chen W, Brosius J (1993) Primary structure, neural-specific expression, and dendritic
location of human BC200 RNA. J Neurosci 13:2382–2390
Tiedge H, Fremeau RT Jr, Weinstock PH et al (1991) Dendritic location of neural BC1 RNA. Proc
Tsang WP, Wong TWL, Cheung AHH et al (2007) Induction of drug resistance and transformation
in human cancer cells by the noncoding RNA CUDR. RNA 13:890–898
Tufarelli C, Stanley JA, Garrick D et al (2003) Transcription of antisense RNA leading to gene
silencing and methylation as a novel cause of human genetic disease. Nat Genet 34:157–165
Turner JD, Schote AB, Macedo JA et al (2006) Tissue specific glucocorticoid receptor expression,
a role for alternative first exon usage? Biochem Pharmacol 72:1529–1536
Umlauf D, Goto Y, Cao R et al (2004) Imprinting along the Kcnq1 domain on mouse chromosome
7 involves repressive histone methylation and recruitment of Polycomb group complexes. Nat
Genet 36:1296–1300
Verhagen PC, Hermans KG, Brok MO et al (2002) Deletion of chromosomal region 6q14-16 in
prostate cancer. Int J Cancer 102:142–147
Vincent JB, Herbrick JA, Gurling HM et al (2000) Identification of a novel gene on chromosome
7q31 that is interrupted by a translocation breakpoint in an autistic individual. Am J Hum
Genet 67:510–514
Wang G, van der Walt JM, Mayhew G et al (2008) Variation in the miRNA-433 binding site of
FGF20 confers risk for Parkinson disease by overexpression of alpha-synuclein. Am J Hum
Genet 82:283–289
Wang H, Iacoangeli A, Popp S et al (2002) Dendritic BC1 RNA: functional role in regulation of
translation initiation. J Neurosci 22:10232–10241
Wang XS, Zhang Z, Wang HC et al (2006) Rapid identification of UCA1 as a very sensitive and
specific unique marker for human bladder carcinoma. Clin Cancer Res 12:4851–4858
Watanabe M, Yanagisawa J, Kitagawa H et al (2001) A subfamily of RNA-binding DEAD-box
proteins acts as an estrogen receptor a coactivator through the N-terminal activation domain
(AF-1) with an RNA coactivator, SRA. EMBO J 20:1341–1352
Waterston RH, Lindblad-Toh K, Birney E et al (2002) Initial sequencing and comparative analysis
of the mouse genome. Nature 420:520–562
Willingham AT, Orth AP, Batalov S et al (2005) A strategy for probing the function of noncoding
RNAs finds a repressor of NFAT. Science 309:1570–1573
Wutz A, Smrzka OW, Schweifer N et al (1997) Imprinted expression of the Igf2r gene depends on
an intronic CpG island. Nature 389:745–749
Xiao J, Yang B, Lin H et al (2007) Novel approaches for gene-specific interference via manipulat-
ing actions of microRNAs: examination on the pacemaker channel genes HCN2 and HCN4.
J Cell Physiol 212:285–292
Xie H, Sun L, Lodish HF (2009) Targeting microRNAs in obesity. Expert Opin Ther Targets
13:1227–1238
Yakovchuk P, Goodrich JA, Kugel JF (2009) B2 RNA and Alu RNA repress transcription by
disrupting contacts between RNA polymerase II and promoter DNA within assembled com-
plexes. Proc Natl Acad Sci USA 106:5569–5574
Yan MD, Hong CC, Lai GM et al (2005) Identification and characterization of a novel gene Saf
transcribed from the opposite strand of Fas. Hum Mol Genet 14:1465–1474
Yang Z, Zhu Q, Luo K et al (2001) The 7SK small nuclear RNA inhibits the CDK9/cyclin T1
kinase to control transcription. Nature 414:317–322
304:594–596
Yik JH, Chen R, Pezda AC et al (2004) A human immunodeficiency virus type 1 Tat-like arginine-
rich RNA-binding domain is essential for HEXIM1 to inhibit RNA polymerase II transcription
through 7SK snRNA-mediated inactivation of P-TEFb. Mol Cell Biol 24:5094–5105
Zalfa F, Adinolfi S, Napoli I et al (2005) Fragile X mental retardation protein (FMRP) binds
specifically to the brain cytoplasmic RNAs BC1/BC200 via a novel RNA-binding motif. J Biol
Chem 280:33403–33410
Zalfa F, Giorgi M, Primerano B et al (2003) The fragile X syndrome protein FMRP associates
with BC1 RNA and regulates the translation of specific mRNAs at synapses. Cell 112:317–327
Zhang X, Rice K, Wang Y et al (2009) Maternally expressed gene 3 (MEG3) noncoding
ribonucleic acid: isoform structure, expression, and functions. Endocrinology 151(3):939–947
Zhao X, Patton JR, Davis SL et al (2004) Regulation of nuclear receptor activity by a pseudour-
idine synthase through posttranscriptional modification of steroid receptor RNA activator. Mol
Cell 15:549–558
Zhu S, Si ML, Wu H et al (2007) MicroRNA-21 targets the tumor suppressor gene tropomyosin1
(TPM1). J Biol Chem 282:14328–14336
Noncoding RNAs at H19/IGF2 Locus:
Role in Imprinting, Gene Expression,
and Associated Pathologies
Nahalie Berteaux, Nathalie Spruyt, and Eric Adriaenssens
Contents
1 The H19/IGF2 Locus and the Parental Imprinting Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
1.1 Overview and Description of the 11p15.5 Locus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
1.2 The Insulator Model of Imprinting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
2 The mRNA-Like Noncoding RNA H19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428
2.1 Properties and Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
2.2 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
2.3 Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
3 The Noncoding Antisense RNA 91H . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
3.1 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
3.2 Hypothesis About 91H Mechanism of Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
4 H19/IGF2 Locus-Associated Pathologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
4.1 Hormone-Dependent Cancers (Breast, Uterus) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
4.2 Children Syndromes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
Abstract Genomic imprinting is a form of epigenetic regulation whereby some

genes are silenced according to their parental origin. The H19/IGF2 locus located in
the chromosome 11 in p15.5 is the best characterized imprinted cluster. The locus
generates two types of noncoding RNAs: the mRNA-like noncoding RNA H19 and
the antisense RNA 91H. The regulation of H19 and its closely linked and recipro-
cally imprinted neighbor, IGF2, has been studied intensively both as a model for
understanding imprinting control mechanisms and because of its role in human
diseases. As with all imprinted clusters, H19 imprinted expression is regulated by
an Imprinting Control Region (ICR), which controls interactions between promo-
ters and shared enhancers. The locus functions like an “insulator model” in which
N. Berteaux, N. Spruyt, and E. Adriaenssens (*)

Institut de Biology de Lille, CNRS UMR 8161, 1 rue Pr Calmette, BP 447 59021 Lille Cedex,
France
e-mail: eric.adriaenssens@univ-lille1.fr
420 N. Berteaux et al.
trans-factors and epigenetic modifications are also required for full expression of
genes. It is well assumed that H19 RNA functions as a riboregulator of which,
expression is developmentally regulated. Elsewhere, the antisense RNA 91H has
been recently discovered as a large and maternally imprinted noncoding RNA. It
plays a role in the paternal IGF2 regulation and is overexpressed in breast cancer.
This original trans-effect may be due to 91H participation in the three-dimensional
organization of the locus, essential for the appropriate expression of genes. In this
chapter, we summarize our current understanding of the molecular and biological
roles of the ncRNAs expressed at the H19/IGF2 domain, both in terms of their
contribution to genomic imprinting control, as well as in terms of cellular targets
they might interact with. We also review knowledge of the locus-associated patho-
logies such as cancers and children syndromes.
Keywords Non coding RNA Imprinting H19/IGF2 91H Cancer Develop-

mental pathologies
1 The H19/IGF2 Locus and the Parental Imprinting Model
1.1 Overview and Description of the 11p15.5 Locus
The mammalian genome contains a small but growing number of genes that are
subject to genomic imprinting (Edwards and Ferguson-Smith 2007; Verona et al.
2003). Genomic imprinting is a form of epigenetic gene regulation that results in
expression from a single allele in a parent-of-origin-dependent manner. This form of
monoallelic expression is essential to normal mammalian development. While the
precise nature of the initial epigenetic imprint remains an intensively investigated
topic, it is assumed that the parental imprint is set in the germline, when genomes are
in distinct compartments and can be differentially modified. After fertilization, the
parental imprints must survive the reprogramming that takes place in the preimplan-
tation embryo, including DNA demethylation and changes in histones modifications
(Reik et al. 2001). Imprinting is maintained throughout development and then erased
before being reestablished in the next generation’s germline. About 90 genes have
been reported to be imprinted even if some of them probably remain to be discovered
(for a complete list, see http://igc.otago.ac.nz/home.html and http://www.har.mrc.
ac.uk/research/genomicimprinting/maps.html). Despite extensive studies and some
major advancement regarding this intriguing phenomenon, we have not yet fully
characterized the underlying molecular mechanisms of genomic imprinting. Never-
theless, some principal hallmarks of imprinted genes can be listed:
l Gene expression is allele-specific.
l Gene expression is often tissue or stage-specific.
l Many of imprinted genes are found in clusters throughout the genome.
l The clusters contain two or more imprinted genes over a region that can span
1 Mb or more.
Noncoding RNAs at H19/IGF2 Locus: Role in Imprinting, Gene Expression 421
)
i t1
IP AS) )
5- T1
,L
ssc P
as
KC 1C S (T , IM
1-
QT
5
2)
KN -A ssc
(K )
T1 T1
vL
)
AS )
H2
CD A1L (T
M C6
1)
1 D 7K
SL C22 a2)
1 O LQ
C L (T )
AS X A1
R
2 1L
2 ( SS
NQ (p5
SC MT
SL Phld
NQ ( K v
EM P
N
C2 A
F- S
(
3
IG 2-A
TS 4
P H (T
3(
KC Q1
L2
L
CD 4
PM
P1
SC
81
2
F-
RP
N
H
9
S
NA
KC
TH
TR
IG
H1
TS
IN
91
M
P
IC2 IC1
Centromere Telomere
Silent allele P Paternal allele

Expressed allele M Maternal allele
Sense of transcription IC Imprinting Center
Fig. 1 Representation of the human chromosomic region 11p15.5. This 1 Mb-locus includes nine
imprinted genes and two imprinting centers. It is delimited by the two maternally expressed genes
TSSC3 et H19 flanked by two nonimprinted genes NAP1L4 et MRPL23 (Tsang et al. 1995; Paulsen
et al. 1998). This region is homologous with the distal region of chromosome 7 in mouse. In spite
of the phylogenetic maintenance if the region over species, it exhibits some structural and
functional discrepancies. Indeed, the TSSC4 gene is imprinted in mouse but not in human (Paulsen
et al. 2000). The TRPM5 gene is imprinted in human but not in mouse (Prawitt et al. 2000).
A TSSC5 antisense transcript has been revealed in human but not in mouse and its imprinted status
has not yet been defined (Crider-Miller et al. 1997; Cooper et al. 1998). The CD81 gene is
imprinted in human and the imprinted status of the INS (Insulin gene) gene is not determined
(Maher and Reik 2000). The ASCL2 gene is imprinted with maternal expression in mouse whereas
its expression is biallelic in human (Miyamoto et al. 2002; Westerman et al. 2001). Usual names
are indicated and followed by secondary names in brackets. Name of antisense transcripts are
underlined and arrows indicating the sense of transcription are depicted below the chromosome
l Within each cluster, a common regulating region, which is called “imprinting

control region” (ICR, also called IC for imprinting Center or ICE for imprinting
control element), controls the imprinting of all genes in the cluster and can act
over hundreds of kilobases. ICRs are designed as differentially methylated
regions with parental-specific modifications that determine their activity. Deletions
of this region lead to the loss of imprinting of multiple genes of the cluster
(Leighton et al. 1995; Ripoche et al. 1997).
l More recently, it has been reported that noncoding RNA were associated with
imprinted clusters and have an essential role in regulating gene expression.
The H19/IGF2 cluster is located on the human chromosome 11 in p15.5 (homolog
to the murine distal chromosome 7). This 1 Mbp-long domain described in Fig. 1
includes nine imprinted genes and two independent imprinting centers.
MKRN3 NDN MAGEL2 SNRPN UBE3A-as UBE3A ATP10C
AS-SRO PWS-SRO
Mas1 Air Igf2r Slc22a1 Slc22a2 Slc22a3
ICE
Kνlqt1-as, Lit1
Tssc3 Tssc5 Cdkn1c Kνlqt1 Ltrpc5 Tssc4 Cd81 Tssc6 Ascl2
KnDMR1
Fig. 2 Antisense RNA associated to imprinted clusters. Igf-2r/Air, Ube3A/Ube3A-as, and Kcnq1/
Kcnq1ot1 are the three best described examples of long antisense RNA, which take part in
epigenetic modifications and gene silencing within imprinted loci. (a) The region of chromosome
15q11–q13 responsible for the Angelman and Prader–Willi syndromes contains a number of
imprinted genes that are coordinately regulated by an imprinting center that contains two func-
tional elements, the PWS-SRO and the AS-SRO. The 460 kb-long Ube3A-as RNA is initiated in
the imprinting center from the paternal allele. It acts as a host gene for the transcription of several
snoRNA (small nucleolar RNA) and represses the UBE3A gene on the paternal allele (Rougeulle
et al. 1998; Runte et al. 2001; Landers et al. 2004). (b) The IGF2R locus. The imprinting center
(ICE) produces a paternally expressed 108-kb long transcript called Air that is necessary for the
silencing in cis of the genes IGF2R, Slc22a1, Slc22a2, and Slc22a3. Air could be implicated in the
methylation spreading thought the locus. (Rougeulle and Heard 2002; Sleutels et al. 2002).
(c) Within the 11p15.5 region, the Kcnq1 gene contains within its intron 10 the imprinting center
KvDMR1, which harbors bidirectional silencing property. This is linked to a paternal antisense
RNA, Kcnq1ot1 (also called Lit1) initiated from KvDMR1. The Kcnq1ot1 promoter shows a
maternal specific methylation. Kcnq1ot1 transcript has a key role in silencing of genes contained in
the Kcnq1 gene imprinted region and it participates to both silencing activity and methylation
spreading (Pandey et al. 2004; Thakur et al. 2004)
Numerous imprinted genes are associated to long antisense RNAs that overlap
several genes. The best described examples are Igf-2r/Air, Ube3A/Ube3A-as, and
Kcnq1/Kcnq1ot1 (Fig. 2). The first demonstration of a direct implication of an
antisense RNA concerns the Air transcript. It consists of a 108 kb-long transcript
localized in the imprinting center within the IGF2r gene and is necessary for the
paternal repression of the gene of the locus (Rougeulle and Heard 2002; Sleutels
et al. 2002). In spite of its well-established role in imprinting process, the molecular
mechanism remains unclear and authors propose hypothesis of methylation propa-
gation from the IGF2R gene or of repressive ARN/protein complexes formation.
The 460 kb-long Ube3A-as RNA is initiated in the imprinting center of the
Prader–Willi syndrome. It acts as a host gene for the transcription of several
snoRNA (small nucleolar RNA) and represses the UBE3A gene on the paternal
allele (Rougeulle et al. 1998; Runte et al. 2001; Landers et al. 2004).
Within the 11p15.5 region, the ICR2 is located within the intron 10 of the
Kcnq1 gene and harbors bidirectional silencing property. This feature is linked to
an antisense RNA, Kcnq1ot1 (also called Lit1), of which the promoter is
contained in ICR2. Expression of this transcript is exclusively paternal. Indeed,
the Kcnq1ot1 promoter shows a maternal-specific methylation. This differential
epigenetic mark is lost in patients affected by Beckwith–Wiedemann syndrome
(BWS) with RNA biallelic expression (Mitsuya et al. 1999; Lee et al. 1999;
Du et al. 2004). More recently, Pandey et al. (2004) have documented that the
Kcnq1ot transcript has a key role in silencing of genes contained in the Kcnq1
gene imprinted region and that it participates directly or indirectly to the methyl-
ation but without RNA interference mechanisms. Furthermore, interruption of
Kcnq1ot1 RNA production by the insertion of a polyadenylation sequence down-
stream of the promoter also caused a loss of both silencing activity and methyla-
tion spreading. Thus, the antisense RNA plays a key role in the silencing function
of the ICR (Thakur et al. 2004).
Elsewhere, a noncoding antisense RNA has also been described in the mouse
and human IGF2 genes (Moore et al. 1997; Okutsu et al. 2000). Like IGF2, this
2.2 kb mRNA transcript is maternally imprinted and overexpressed in Wilms’
tumors. IGF2-AS was expressed at levels comparable with IGF2 sense expression
derived from promoters P1 and P2 in normal tissue and in breast, ovarian, and
Wilms’ tumor tissues. It is composed of three exons, which overlap the exons 3 and
4 of the IGF2 gene (Vu et al. 2003). Its function remains unknown and its
involvement in imprinting has not yet been demonstrated, but findings indicate
that it is a good marker of Wilms’ tumor (Okutsu et al. 2000).
Finally, discovery of microRNAs (miRNA) and RNA interference could likely
provide new insights on imprinting mechanism. Then, the group of Cavaille has
identified in mouse a short cluster of maternally expressed miRNA genes (miR-431,
miR-433, miR-127, miR-434, and miR-136) transcribed and processed from a gene
antisense to the paternally expressed Rtl1 gene (Retrotransposon-like gene1) (Seitz
et al. 2003). Rtl1, also called Peg11 in sheep, displays homology with the Ty3/gypsy
retrotransposon family and its function is currently unknown (Charlier et al. 2001;
Youngson et al. 2005). Due to this peculiar sense–antisense organization, the
encoded miRNAs are obviously perfectly complementary to Rtl1 mRNA and thus
were predicted to cleave Rtl1 mRNA via RNAi-like mechanisms (Seitz et al. 2003).
Indeed, the predicted RNAi-mediated cleavage sites in the middle of the RNA
duplexes have been experimentally mapped by 50 -RACE (rapid amplification of

cDNA ends) experiments (Davis et al. 2005). These imprinted miRNAs are there-
fore among the very rare miRNAs in animals that act as siRNAs (small interfering
RNAs) (Yekta et al. 2004). From an evolutionary point of view, this trans-allelic
RNAi between imprinted genes (for example, maternally expressed miRNAs
silence a paternally expressed gene) can be viewed as a good illustration of the
parent-conflict theory (Wilkins and Haig 2003; for further discussion, see also
Davis et al. 2005; Lewis and Redrup 2005). In addition, this regulation might
also account for the complex gene regulatory network occurring at the ovine
Dlk1–Gtl2 locus, the so-called polar overdominance phenomena (for further infor-
mation, see Davis et al. 2005; also reviewed in Royo and Cavaillé 2008).
Interestingly, a 23-nucleotide microRNA miR-675 was shown recently to be
processed from the H19 gene and this may in turn regulate mRNAs in development
and/or in oncogenesis. It is endogenously expressed in human keratinocytes and
neonatal mice and overexpressed in cells transfected with human or mouse H19
expression plasmids. These data demonstrate that H19 can function as a primary
microRNA precursor (Cai and Cullen 2007).
1.2 The Insulator Model of Imprinting
The H19 gene is one of the first genes proven to be imprinted. In human, it lies
within 200 kbp downstream of the IGF-2 gene (Zemel et al. 1992). The regulation
of H19 and its closely linked and reciprocally imprinted neighbor, IGF2, has been
studied intensively both as a model for understanding imprinting control mechanisms
and because of its role in human diseases. The two genes are imprinted in an opposite
manner, with the paternal IGF-2 and the maternal H19 alleles being reciprocally
expressed (Giannoukakis et al. 1993; Zhang and Tycko 1992).
1.2.1 DNA Methylation of H19 and IGF2 Genes
The H19 silent paternal allele exhibits several characteristics associated to its
transcriptional repression: it is hypermethylated in the promoter region and in the
50 region in embryonic tissues, the promoter shows a compact chromatin structure
(Bartolomei et al. 1993; Ferguson-Smith et al. 1993) and its histone acetylation rate
is lower than the one of the maternal allele (Grandjean et al. 2001).
Surprisingly, the IGF2 promoter region is not methylated and its chromatin
structure is favorable to a biallelic transcription (Sasaki et al. 1992). However,
two other regions preferentially methylated on the expressed paternal allele have
been identified within the gene: the DMR1 located 3 kbp upstream the P1 promoter
acts as a silencer on the maternal allele when it is unmethylated, and the DMR2,
located within exons 5 and 6 is an activator on the paternal allele when it is
methylated (Feil et al. 1994; Murrell et al. 2001; Constancia et al. 2000). It is
interesting to notice that both regions acquire this differential methylation after
fecundation (it is then a secondary methylation by opposition to primary methylation
that it is established in gamete and allows to distinguish the two parental alleles), and
this implies that it is not responsible for imprinting establishment.
1.2.2 Histone Modifications at the H19/IGF2 Locus
Analysis of histone acetylation state at the murine locus shows that H4 histone
acetylation discrepancies take place within H19 and IGF2 genes in a parental-
specific manner with expressed alleles being more acetylated than silent alleles.
However, the link between DNA methylation, histone hypoacetylation, and gene
expression is established only for the H19 promoter region (Grandjean et al. 2001;
Pedone et al. 1999). Moreover, several studies show that the inhibition of histone
deacetylases deregulates gene expression at the locus with a repression of the H19
active maternal allele, changes in acetylation patterns of the ICR region (Grandjean
et al. 2001), and an IGF2 biallelic expression (Hu et al. 1998; Yang et al. 2003).
Finally, the HDAC recruitment is directly involved in the repression effect of the
insulator protein CTCF described in the next paragraph (Lutz et al. 2000).
1.2.3 The ICR or Imprinting Control Region
The methylation of promoter regions is not sufficient to explain the reciprocal

expression of the two genes, other elements exhibiting primary methylation hall-
marks have been searched within the locus. Discovery of the ICR, located between
the H19 and IGF2 genes, allows understanding the mechanism of imprinting setting.
Existence of a common regulatory region has been initially suggested by dele-
tions assays of the H19 gene region (Leighton et al. 1995; Ripoche et al. 1997). This
sequence, named ICR, is located within the region comprised between 2 and 4 kb
upstream of the transcription start site of the H19 gene and carries primary
methylation marks (Tremblay et al. 1997). ICR has a long-range action to establish
the H19 and IGF2 imprinting during the embryonic development. It is exclusively
methylated on the paternal allele and shows nuclease hypersensibility sites on the
maternal allele (Hark and Tilghman 1998; Kanduri et al. 2000a). These epigenetic
hallmarks are a sign of a protein binding to the sequences. ICR deletion leads to loss
of H19 and IGF2 (Thorvaldsen et al. 1998, 2002). ICR is also necessary to the
imprinting of a H19 transgene in mouse (Elson and Bartolomei 1997). These studies
show the pivotal role of ICR in imprinting.
To understand the long-range and allele specific effect of ICR, deletion and
relocalization assays have been conducted. It revealed the chromatin insulator
activity of the ICR region (Webber et al. 1998; Kaffer et al. 2000; Kanduri et al.
2000b). Indeed, it is able to insulate communication between a promoter and
an enhancer when it is located between the two regions. This activity required a
zinc-finger protein named CTCF (CCTC-binding factor) (Bell et al. 1999). In vitro
and in vivo experiments have shown in mouse that CTCF binds unmethylated ICR
via four consensus sites and the binding is abolished by DNA methylation (Bell and
Felsenfeld 2000; Hark et al. 2000; Kanduri et al. 2000b; Szabo et al. 2000;
Holmgren et al. 2001; Ulaner et al. 2003). In human, there are seven binding sites
but the sixth is only the one to possess a differential methylation (Takai et al. 2001).
The H19/IGF2 insulation model is represented in the Fig. 3. Both genes share a
common set of enhancers located downstream from the H19 gene. In the maternal
allele, the CTCF recruitment on the unmethylated ICR acts as a chromatin bound-
ary and blocks the enhancer access to the IGF2 promoter to prevent its activation.
The H19 gene is then activated (Reik and Murrell 2000; Wolffe 2000).
–137 –12 –2 +10.5 kpb

INSULATOR INSULATOR
+ +
× CTCF CTCF M
IGF-2 H19
DMR1 DMR2 Huc ICR Enh
+ +
×
CTCF P
IGF-2 H19
Fig. 3 Reciprocal imprinting mechanism of H19 and IGF2 genes. Activation of gene expression
is indicated by (þ), repression by (), and inhibition of the enhancer function is represented by a
vertical bar. Relative positions are expressed in kilobase pairs relatively to the H19 transcription
start site. Three principal mechanisms intervene in the regulation: the methylation (represented by
vertical bars), the enhancer activity, and the insulator activity (Hark et al. 2000). Three DNA
regions are differentially methylated according to the allele, the DMR1 and 2 of the IGF2 gene
(violet diamond) and the ICR (blue oval) at 2 kb upstream of the H19 gene. The enhancer
sequences Enh (Enhancer downstream of the H19 gene, green circles) and Huc (enhancers
upstream of the H19 gene, blue circles) represent, respectively, the endodermic and mesodermic
enhancers (Ishihara et al. 2000; Drewell et al. 2002). On the maternal allele, the nonmethylated
ICR contains four consensus CTCF binding sites (Hark et al. 2000). The CTCF DNA binding
produces then a chromatin boundary, which prohibits enhancers to access to the IGF2 gene. The
Huc enhancers may also activate the H19 gene (Drewell et al. 2002). The nonmethylated IGF2
DMR1 (violet diamond) acts as a silencer (Constancia et al. 2000). On the paternal chromosome,
the methylated ICR does not bind any protein but acts as a H19 expression repressor. The Enh
enhancers can then activate the IGF2 promoter and the methylated IGF2 DMR2 also activates
gene expression (Murrell et al. 2001). ICR has a role of transcriptional repressor for the H19 gene
(Srivastava et al. 2000). In 30 of H19, a secondary chromatin boundary independent of the
methylation delimits the imprinted domain (Ishihara and Sasaki 2002)
On the paternal allele, the ICR methylation does not allow CTCF binding and
leads to activation of the distal IGF2 promoter and gene expression (for a review,
see Lewis and Murrel 2004). Finally, the maintenance of the unmethylated state of
maternal ICR is due to the CTCF protein, which prevents the de novo methylation
in this region (Fedoriw et al. 2004; Lewis and Murrel 2004; Szabo et al. 2004). The
mesodermic enhancer activity recently discovered (Drewell et al. 2002), which
intervene in the regulation can be added to this model. Finally, the in vivo CTCF
binding upstream of the Mrpl23 gene could be a chromatin boundary delimiting the
imprinted domain (Ishihara and Sasaki 2002).
1.2.4 Imprinting and Parental Specific Chromatin Loops
The cis elements described previously have a long-range action. They have to
physically interact with each other or with their target to exert their effects.
Chromosome conformation capture (3C) analysis in mice, which assay for physical
interactions between chromosomal regions, have suggested that CTCF has a critical
role in the epigenetic regulation of high-order chromatin structure and gene silenc-
ing over considerable distances in the genome, but the precise nature and function
of the looping is debated (Engel et al. 2008; Kurukuti et al. 2006; Lopes et al. 2003;
Murrell et al. 2004; Yoon et al. 2007). Kurukuti et al. 2006 reported that on the
paternal allele, enhancers interact with the IGF2 promoters whereas on the mater-
nal, this is prevented by CTCF binding within the H19 ICR. They demonstrated that
the maternal-specific silencing of IGF2 results when the ICR interacts with a matrix
attachment region (MAR3) and a differentially methylated region (DMR1) at the
IGF2 locus to generate a tight loop around the IGF2 gene, thereby physically
impeding Igf2 expression. Moreover, CTCF interacts with the three clustered
IGF2 promoters and recruits polycomb repressive complexes that lead to the
allele-specific methylation at lysine 27 of histone H3 (H3-K27) and to the suppres-
sion of the maternal IGF2 promoters (Li et al. 2008). Elsewhere, Murrell et al.
(2004) reported that on the maternal allele, the unmethylated ICR binds to the
DMR1 of IGF2 resulting in an inactive domain where IGF2 is away from the
enhancers. On the paternal allele, the methylated ICR associates with the methy-
lated IGF2 DMR2 moving IGF2 into the active chromatin domain (Dekker et al.
2002). More recently, it has been shown that on the maternal allele, the enhancers
make contacts throughout the H19 coding unit and promoter (Engel et al. 2008;
Kato and Sasaki 2005).
Figure 4 provides a simplified overview of available data about chromatin loop
structures at the H19/IGF2 locus (Kurukuti et al. 2006; Murrell et al. 2004; Weber
et al. 2003). Additional mechanisms exist for an imprint mark, such as chromatin
composition, organization, and histone acetylation or methylation state (Fuks 2003;
Grandjean et al. 2001), even if DNA methylation is by far the best candidate
(Bestor 2000).
enhancers
ICR H19
CTCF
R1
M
AR
DM
3
DMR2
Maternal allele
IGF2
H19
enhancers
ICR
ICR
DMR1 DMR2
DMR2 MAR3
Paternal allele
Fig. 4 Chromatin loop structures at the H19/IGF2 locus. Chromosome conformation capture (3C)
assays revealed long-range physical interactions within the H19/IGF2 region. These chromatin
structures are orchestrated by CTCF and regulate epigenetic hallmarks and gene silencing within
the locus. On the maternal allele, CTCF binding to the ICR prevents interaction between enhancers
and IGF2 promoter. ICR interacts with a matrix attachment region MAR3 and with the differen-
tially methylated region DMR1 within the IGF2 region. This leads to the formation of a silencing
loop, which impedes IGF2 expression. Enhancers can interact with the H19 promoter and activate
its expression (Engel et al. 2008; Kato and Sasaki 2005). Moreover, CTCF interacts with the three
clustered IGF2 promoters and recruits polycomb repressive complexes that lead to the allele-
specific histone methylation and to suppression of the maternal IGF2 promoters (Li et al. 2008).
On the paternal allele, enhancers interact with the IGF2 promoters (Kurukuti et al. 2006). The
methylated ICR associates with the methylated IGF2 DMR2 moving IGF2 into the active
chromatin domain (Dekker et al. 2002). ICR methylation spreads to the H19 promoter that impairs
H19 paternal expression
2 The mRNA-Like Noncoding RNA H19
H19 encodes a spliced and polyadenylated RNA that lacks conserved open reading
frames (ORFs) but does have a conserved secondary RNA structure (Juan et al.
2000). Even if extensive deletions and/or point mutations of an ectopic human H19
RNA generate a protein, (Joubel et al. 1996) no endogenous translation product has
so far been identified (Pachnis et al. 1984, 1988). Therefore, it was quickly proposed
that H19 RNA functions as a riboregulator (Brannan et al. 1990) of which
expression is developmentally regulated. It is abundantly expressed in both extra-

embryonic and fetal tissues and is repressed after birth except in a few adult organs,
particularly in the mammary gland (Douc-Rasy et al. 1993; Dugimont et al. 1995). It
should be emphasized that the role of the H19 gene in cancer is still a matter of
debate. It has been proposed that H19 functions as a tumor suppressor in some
Wilms’ tumors, embryonic rhabdomyosarcoma, and the Beckwith–Wiedmann cancer
predisposing syndrome (Okamoto et al. 1997; Steenman et al. 1994). Consistently,
some studies conclude that it downregulates the IGF2 factor (Wilkin et al. 2000). By
contrast, other studies including ours argue in favor of an oncogenic role of H19 with
a positive correlation with cell aggressiveness (Lottin et al. 2002a; Rachmilewitz
et al. 1995). H19 activation has also been reported in various cancer tissues:breast
(Douc-Rasy et al. 1993; Dugimont et al. 1995; Adriaenssens et al. 1998), bladder
(Ariel et al. 1995; Elkin et al. 1995), lung (Kondo et al. 1995), and esophageal
cancers (Hibi et al. 1996). Its oncogenic role has been well documented in the
bladder, since it is considered as an oncodevelopmental marker (Cooper et al. 1996)
and regulates genes involved in metastasis and blood vessel development (Ayesh
et al. 2002). These observations support a H19 role in tumor invasion and angiogen-
esis. In spite of polemic on the H19 role in cancer, we clearly established the
oncogenic role of H19 (Lottin et al. 2002a) in breast cancer, and we demonstrated
that H19 overexpression in breast cancer cells promotes the cell cycle progression,
by increasing S phase entry (Berteaux et al. 2005).
2.1 Properties and Expression
The H19 gene was discovered in the mouse as a gene under coordinate regulation
with (-feto-protein in the liver (Pachnis et al. 1984). More recently, Juan et al.
(2000) brought evidence for evolutionarily conserved secondary structure in the
H19 RNA from several mammalian species. The H19 gene is an unusual gene, in
that it is transcribed by RNA polymerase II, processed by capping, splicing, and
polyadenylation; but it does not appear to encode a protein. Actually, the particu-
larity of the H19 transcript is its inability to be translated when the 50 untranslated-
region (50 UTR) is not experimentally altered (deletions and/or point mutations)
(Pachnis et al. 1988; Joubel et al. 1996). Furthermore, hypothetical translation of
established sequences from a range of mammalian species shows an absence
of conserved ORFs of any size. Consequently, given the evolutionary conservation
of structure at the RNA level and the absence of conservation at the protein level, it
has been proposed that the mature transcript is the functional product of the H19
gene and that its function requires the ability to fold into a specific secondary
structure. As early as 1990, Brannan et al. (1990) proposed that this gene could act
as a “riboregulator.”
The H19 gene encodes one of the most abundant RNAs in the developing mouse
and human embryo (Pachnis et al. 1984; Brannan et al. 1990). It is expressed at the
blastocyst stage of development and accumulates to high levels in tissues of
endodermal and mesodermal origins (Poirier et al. 1991; Lustig et al. 1994) as well
as ectodermal origin (Ohlsson et al. 1994, Hemberger et al. 1998). H19 is expressed
in the choroid plexus and leptomeninges of the developing mouse fetus (Svensson
et al. 1995) but not in these tissues during human development (Ohlsson et al.
1994). After birth, the gene is repressed in almost all tissues except skeletal muscle
(Pachnis et al. 1984; Leibovitch et al. 1995; Douc-Rasy et al. 1993; Milligan et al.
2000). In other respects, a basal but significant H19 expression is detected at
adulthood in lung, heart, and thymus (Poirier et al. 1991), mammary gland
(Douc-Rasy et al. 1993; Dugimont et al. 1995; Adriaenssens et al. 1999), adrenal
gland (Liu et al. 1995), and uterus (Adriaenssens et al. 1999; Ariel et al. 1997).
2.2 Functions
Since the first mention of H19 in 1984 by Pachnis et al., its functions have only
begun to emerge. It has been reported that H19 RNA was involved in the repression
of the IGF2 oncogene by affecting its transcription (Wilkin et al. 2000) or its
translation (Li et al. 1998). In addition, we brought evidence that the H19 gene
posttranscriptionally upregulates the thioredoxin level, a key protein of the cellular
redox metabolism (Lottin et al. 2002b). Deletion of the H19 gene in mouse
(KO mice) leads to a size and weight increase of about 10% (Ripoche et al.
1997). But the consecutive biallelic IGF2 expression in these animals does not
allow concluding to an H19 direct effect. However, the group of Surani demon-
strated that induction of a targeted H19 biallelic expression via specific silencer
deletion in transgenic mice leads to smaller animals whereas IGF2 expression is not
affected (Drewell et al. 2000). So, the H19 expression pattern suggests that the
transcript assures a major function during the development.
2.3 Regulation
Beyond epigenetic regulations, the H19 gene is submitted to local regulations (i.e.,
on the level of the promoter), by hormones, growth factors, or other cytokines.
Indeed, in mammary cells, H19 is activated by HGF-SF (hepatocyte growth factor/
scatter factor), which has been identified as one of the main paracrin mediators of
morphogenetic epithelial/mesenchymal interactions. This factor has potent moto-
genic, mitogenic, and morphogenic effects on epithelial cells in culture, and H19
RNA synthesis is related to the migratory phenotype of cultured cells. EGF and
FGF-2 also activates H19 but less, whereas IGF-2, TGFb1, and TNFa have no
effect on H19 expression in these cells (Adriaenssens et al. 2002). In uterus, during
estrus (proliferative phase) and metestrus (early secretory phase) phases, and in
breast, during puberty and pregnancy, morphologic changes are associated to peak
of H19 expression. Moreover, its expression is regulated by steroid hormones
estrogen and progesterone, which respectively up- and downregulate the gene
(Adriaenssens et al. 1999). Elsewhere, we also demonstrated a negative regulation
of H19 by the tumor suppressor protein P53 (Dugimont et al. 1998).
H19 and AFP (a-fetoprotein) are abundantly transcripted in mammalian fetal
liver but are rapidly repressed after birth. This repression is partly controlled by the
Afr1 locus (a Fetoprotein Regulator 1) (Pachnis et al. 1984). Recently, Perincheri
et al. (2005) have mapped the locus and identified the murine Zhx gene (Zinc
Fingers and Homeoboxes gene), ortholog to the human ZHX2 gene. This factor is
directly responsible for the H19 postnatal repression in liver but also probably in
other organs since it is expressed in ubiquist manner.
In breast cancer, preferential accumulation of H19 transcripts has been observed
at stroma/epithelium interface (Dugimont et al. 1995; Adriaenssens et al. 1998).
Moreover, scattering and morphogenesis of epithelial cells by a conditioned
medium from fibroblasts induces activation of the H19 gene (Adriaenssens et al.
2002). The H19 gene is then finely regulated by environmental factors and it is
involved in the epithelial/mesenchymal crosstalk, essential during tumorigenesis.
Posttranscriptional regulations have been also reported such as RNA stabilization
by not yet identified proteins (Milligan et al. 2000; Jouvenot et al. 1999) Elsewhere,
a direct association of the human and mouse H19 RNA with the IMP protein
family have been demonstrated (In mouse: CRD-BP (cMyc mRNA coding Region
instability Determinant Binding Protein) and in human: IMP1, 2, and 3 (IGF-II
mRNA-binding protein)). These proteins are able to regulate H19 RNA localization
and stabilization and are also able to bind the IGF2 RNA (Tessier et al. 2004; Runge
et al. 2000; Nielsen et al. 2001, 2004; Liao et al. 2005).
3 The Noncoding Antisense RNA 91H
3.1 Characterization
A more recently identified characteristic of imprinted genes is their association, in

some cases, with noncoding antisense transcripts (ncRNAs), which have been
suggested to constitute a new epigenetic regulatory system (see Fig. 2). These
ncRNAs are not yet clearly classified, but categories with known gene regulatory
functions are emerging. It includes (1) intergenic transcripts that regulate local
chromatin activity, (2) cis-acting long ncRNAs such as Xist involved in chromo-
some inactivation, and (3) ncRNA expressed within imprinting loci such as Air,
KCNQ1OT1, and UBE3A-as involved in domain silencing (Rougeulle and Heard
2002; Rougeulle et al. 1998; Sleutels et al. 2002; Thakur et al. 2004).
These latter transcripts share some characteristics:
l It is always very long RNA (several hundreds of base pairs).
l Its expression is allele-specific.
l The promoter is often located near an imprinting center and/or in an intron of a

coding gene.
l The transcript is implicated in gene silencing and epigenetic modifications.
It regulates expression of overlapped (or not) genes in cis.
Apart from the identification of some IGF2 antisenses transcripts, few data are
available on that topic at the H19/IGF2 locus. We recently identified and character-
ized a new transcript within this locus. It consists of a large intergenic 120 kb-long
RNA that we named 91H since it is transcribed antisense to H19. This nuclear and
short-lived RNA is expressed predominantly from the maternal allele in both mouse
and human within the H19 gene region. Moreover, the transcript is stabilized in
breast cancer cells and overexpressed in human breast tumors. Knockdown experi-
ments showed that 91H inhibition downregulates IGF2 expression in trans. Thus,
91H shares the same characteristics as the other similar ncRNAs do, which are
described above, apart from its trans effect on IGF2 expression. Indeed, we
demonstrated that the maternal 91H transcript is involved in the maintenance of
the paternal IGF2 gene expression (Berteaux et al. 2008).
3.2 Hypothesis About 91H Mechanism of Action
To unravel this possible trans-effect of the 91H RNA, we envisaged a physical

proximity between homologous chromosomes. Recent works using chromosome
conformation capture technology strongly support the notion of epigenetic chro-
mosomal networks. Several different chromosomes converged on the H19 ICR
simultaneously (Ling et al. 2006; Zhao et al. 2006) probably through the CTCF
protein, and these long-range allele-specific chromosomal associations were linked
with epigenetic regulation of transcription in trans. These data support the possi-
bility that the 91H RNA collaborates in vivo with the establishment of chromo-
somal complexes and that this collaboration leads to activation of the IGF2 paternal
allele. The fact that 91H knockdown only slightly affects gene expression from the
maternal allele seems to indicate that 91H effect is not mediated through the ICR/
CTCF complex. However, other candidates such as enhancer DNA regions may be
considered for interacting with the antisense RNA. Indeed, 91H encompasses a
region including the biallelically transcribed HUC sequences. It consists of regu-
latory elements located just upstream of the ICR, which act as strong mesodermal
enhancers and are supposed to activate IGF2 expression on the paternal allele
(Drewell et al. 2002). Accordingly, we proposed a model to explain the 91H
trans-effect on IGF2 expression in which cooperation between mesodermic
(HUC) and endodermic (ENH) enhancer sequences would be required for full
expression of the gene (Fig. 5). Moreover, 91H may attract repressive chromatin
modifications to the maternal allele by trapping factors responsible for DNA or
histone modifications, which become available for the paternal allele when 91H is
disrupted, resulting in IGF2 expression decrease (Rinn et al. 2007; Yu et al. 2008).
–137 –12 –2 +10.5 kbp
´
CTCF
IGF-2 H19 MrpL23
HUC ICR ENH
91H RNA
Limiting
regulatory
factors
?
´
IGF-2 H19 MrpL23

HUC ICR ENH
Fig. 5 Model for 91H trans-effect on IGF2 expression. Two sets of enhancers (HUC and ENH
sequences that correspond, respectively, to mesodermic and endodermic enhancers) regulate IGF2
expression. Both would be required for full expression of the gene. In addition, the two IGF2
alleles would be competing for a common limited stock of regulatory elements (methylation/
acetylation/transcription?). On the maternal allele, 91H would block the locus and prevent the
HUC sequences from interaction with any regulatory factors. These latter would be then directed
on the paternal allele, in the HUC region and/or in the IGF2 promoter region, and would cooperate
with the cis endodermic enhancers resulting in IGF2 enhanced expression. (Arrows indicates
positive regulations whereas lines with bars correspond to inhibitions; ICR: Imprinting Control
Region, CTCF: CCTC-binding factor)
4 H19/IGF2 Locus-Associated Pathologies
4.1 Hormone-Dependent Cancers (Breast, Uterus)
It should be emphasized that the role of the H19 gene in cancer is still a matter
of debate. Investigations in tumor development are delicate on account of the H19
noncoding state, the lack of knowledge of its mechanism of action, and its
imprinted status. Gene expression has been studied in numerous cancer types, but
the gene status appears to be contradictory since it can be either tumor suppressor or
oncogene depending on the studied model. Table 1 provides a nonexhaustive
general survey of bibliographic data available and illustrates their heterogeneity.
Indeed, it has been proposed that H19 functions as a tumor suppressor in some
Wilms’ tumors, embryonic rhabdomyosarcoma, and the Beckwith–Wiedmann can-
cer predisposing syndrome (Okamoto et al. 1997; Steenman et al. 1994). Consis-
tently, some studies conclude that it downregulates the IGF2 factor (Wilkin et al.
2000). By contrast, other studies including ours argue in favor of an oncogenic role
of H19 with a positive correlation with cell aggressiveness (Lottin et al. 2002a;
Rachmilewitz et al. 1995). H19 activation has also been reported in various cancer
Table 1 Overview of bibliographic data about oncogene or tumor suppressor status of the H19
gene
Oncogene Tumor suppressor
Organ, model or Bibliographic references Organ, model or type Bibliographic references
type of cancer of cancer
Transgenic mice Leighton et al. (1995), Transgenic mice Drewell et al. (2000)
Ripoche et al. (1997)
Bladder Ariel et al. (1995, 1997, SHE cells Wiseman et al. (1991),
2000b), Elkin et al. Isfort et al. (1997)
(1995), Cooper et al.
(1996), Ayesh et al.
(2002), Ohana et al.
(2002)
Chorion (JEG-3 Rachmilewitz et al. (1995) G401 cells (WT) Hao et al. (1993)
cells)
Breast Douc-Rasy et al. (1993), JEG-3 cells (Chorion) Li et al. (1998)
Dugimont et al.
(1995), Adriaenssens
et al. (1998), Lottin
et al. (2002a, b)
Lung Kondo et al. (1995) Children liver, Li et al. (1998), Wilkin
Hepatoblastomas et al. (2000)
Uterus Tanos et al. (2004), Lottin Wilms’ tumors Steenman et al. (1994),
et al. (2005) (kidney) Casola et al. (1997)
Esophagus/colon Hibi et al. (1996), Cui Children muscle, Casola et al. (1997)
et al. (2002) rhabdomyosarcoma
Liver Manoharan et al. (2004) Beckwith–Wiedemann Reik et al. (1995)
syndrome
Ovary Tanos et al. (1999), Adrenals Liu et al. (1995),
Chen et al. (2000) Wilkin et al. (2000)
Head/neck Rainho et al. (2001)
Pharynx Ng et al. (2003)
Testicles Verkerk et al. (1997),
Ariel et al. (2000a)
Demonstration of the oncogene/tumor suppressor status in mentioned works is not always clearly
established. Numerous results converge to one or the other of these statuses and from these data,
authors have proposed their hypothesis
tissues: breast (Douc-Rasy et al. 1993; Dugimont et al. 1995; Adriaenssens et al.
1998), bladder (Ariel et al. 1995; Elkin et al. 1995), lung (Kondo et al. 1995), and
esophageal cancers (Hibi et al. 1996). Its oncogenic role has been well documented
in the bladder, since it is considered as an oncodevelopmental marker (Cooper et al.
1996) and regulates genes involved in metastasis and blood vessel development
(Ayesh et al. 2002). These observations support a H19 role in tumor invasion and
angiogenesis. In spite of polemic on the H19 role in cancer, the oncogenic role of
H19 has been clearly established in some models. In bladder and uterus cancers,
H19 is directly associated to tumor progression and is considered as tumor marker
(Ariel et al. 2000b; Lottin et al. 2005). Some authors consider H19 as an oncofetal
RNA (Ariel et al. 1997, 2000a). Ohana et al. (2002) proposed an approach of gene
therapy based on the use of H19 regulatory sequences driving the expression of the
diphtheria toxin. Injection of these constructs in mice-induced tumors derived from

cancerous colon and bladder cells leads to a decrease of the tumor growth (Ohana
et al. 2005). In breast cancer, H19 is overexpressed in 70% of the adenocarcinomas
and its expression is correlated with the “Tumor values,” the presence of hormones
receptors, and tumor invasion (Adriaenssens et al. 1998). Overexpression of an
ectopic H19 gene enhances the tumorigenic properties of breast cancer cells (Lottin
et al. 2002a) and promotes the cell cycle progression, by increasing S phase entry
(Berteaux et al. 2005).
In all cases, deregulation of H19 expression results in genetic alterations leading
to a loss of heterozygote (LOH, for example, in the case of uniparental disomy) or
epigenetic alterations leading to a loss of imprinting (LOI by hypermethylation or
hypomethylation of the H19 promoter or the ICR). It is then often difficult to assign
directly the observed phenotypes to the H19 gene only, since it is tightly connected
to the IGF2 gene, and in most of the cases, the expression of both the genes is
modified. To summarize, the H19 contribution in tumorigenesis depends on the
physiological conditions, the considered tissue, and the nature of deregulations.
4.2 Children Syndromes
The demonstration that the H19 transcript is stabilized in breast cancer cells and
overexpressed in human breast tumors led us to propose a link between antisense
transcription and cancer. Furthermore, we know that the deregulation of the H19/
IGF2 locus is associated to several human fetal syndromes such as BWS and
Silver–Russell syndrome (SRS).
BWS is associated with fetal and postnatal overgrowth and is associated with
embryonic tumors such as children kidney tumors named Wilm’s tumors. BWS can
be caused by a range of different defects. Several distinct genetics or epigenetics
errors involving 11p15 have been identified in different BWS patients. Some
patients have maternal chromosomal rearrangements of 11p15, meaning that
there is a disruption of the chromosome in this region. Other patients have paternal
uniparental disomy of 11p15, meaning that the maternal copy of this region is
replaced with an extra paternal copy. Many other patients have abnormal DNA
methylation in different DMR regions of 11p15, meaning that normal epigenetic
marks that regulate imprinted genes in this region are altered. In some cases, the
expression of IGF2 gene is doubled and expression of H19 gene is silenced. BWS is
often associated with H19 epigenetic inactivation due to ICR hypermethylation.
This overexpression of IGF2 is responsible for symptoms linked to the pathology.
Nevertheless, in some cases, the specific defect causing BWS in an affected patient
may remain unknown. In about one-third of BWS patients, the genetic or epigenetic
mutation is unknown.
SRS is an intrauterine growth delay associated to an altered postnatal growth with
facial dysmorphy and corporal asymmetry. Reasons of SRS are varied, chromosome
mosaic, equilibrated translocation (1; 17 or 17; 20), deletion (8q11–q13). In some
cases, defect in 11p15-5 region has been observed. For numerous patients, this
syndrome is associated with ICR hypomethylation, IGF2 silencing, and H19 biallelic
expression. As BWS, some cases of SRS remain unexplained. The discovery of 91H
gene and its function on gene regulation in 11p15.5 locus can give research lanes to
explore theses unexplained cases of BWS or SRS.
5 Conclusion
There is growing evidence that noncoding transcripts are functional and take part
in most, if not all, complex genetic phenomena in eukaryotes, including RNA
interference-related processes such as transcriptional and posttranscriptional gene
silencing as well as parental imprinting and allelic exclusion. Noncoding RNAs
intervene in general biological processes and the close connection between non-
coding RNAs and epigenetic processes suggests that they compose a hitherto
hidden layer of genomic programming in humans. The next frontier is now the
functional characterization of these molecules to understand the interactions
between the regulatory RNAs and their targets. This is a way to better understand
how an RNA molecule can regulate cell cycle, gene expression, methylation
spreading, or even chromosomal network. The particular 91H antisense transcript
that is produced at the H19/IGF2 locus adds further complexity to the cluster of
imprinted genes in the human imprinted 11p15.5 region. 91H RNA is involved in
the control of paternal IGF2 expression suggesting that a distinct mechanism of
imprinting arises in the locus. This innovating type of regulation in “trans” between
two homologous chromosomes need to be elucidated to understand, at least in part,
the complex regulation that takes place in the imprinted locus (chromosomal
network, methylation. . .). Functional studies would allow using these RNA mole-
cules as tumor markers or therapeutic targets. Particularly, 91H could be a good
candidate to improve the diagnostic and the therapeutic tools for severe children’s
syndromes or adult cancers associated to the 11p15 chromosomal region.
References
Adriaenssens E, Dumont L, Lottin S et al (1998) H19 overexpression in breast adenocarcinoma

stromal cells is associated with tumor values and steroid receptor status but independent of p53
and Ki-67 expression. Am J Pathol 153:1597–1607
Adriaenssens E, Lottin S, Dugimont T et al (1999) Steroid hormones modulate H19 gene
expression in both mammary gland and uterus. Oncogene 18:4460–4473
Adriaenssens E, Lottin S, Berteaux N et al (2002) Cross-talk between mesenchyme and epithelium
increases H19 gene expression during scattering and morphogenesis of epithelial cells. Exp
Cell Res 275:215–229
Ariel I, Lustig O, Schneider T et al (1995) The imprinted H19 gene as a tumor marker in bladder
carcinoma. Urology 45:335–338
Ariel I, Ayesh S, Perlman EJ et al (1997) The product of the imprinted H19 gene is an oncofetal
RNA. Mol Pathol 50:34–44
Ariel I, de Groot N, Hochberg A (2000a) Imprinted H19 gene expression in embryogenesis and
human cancer: the oncofetal connection. Am J Med Genet 91:46–50
Ariel I, Sughayer M, Fellig Y et al (2000b) The imprinted H19 gene is a marker of early recurrence
in human bladder carcinoma. Mol Pathol 53:320–323
Ayesh S, Matouk I, Schneider T et al (2002) Possible physiological role of H19 RNA. Mol
Carcinog 35:63–74
Bartolomei MS, Webber AL, Brunkow ME et al (1993) Epigenetic mechanisms underlying the
imprinting of the mouse H19 gene. Genes Dev 7:1663–1673
Bell AC, Felsenfeld G (2000) Methylation of a CTCF-dependent boundary controls imprinted
expression of the IGF2 gene. Nature 405:482–485
Bell AC, West AG, Felsenfeld G (1999) The protein CTCF is required for the enhancer blocking
activity of vertebrate insulators. Cell 98:387–396
Berteaux N, Lottin S, Monté D et al (2005) H19 mRNA-like non-coding RNA promotes breast
cancer cell proliferation through its positive control by E2F1. J Biol Chem 280:29625–29636
Berteaux N, Aptel N, Cathala G et al (2008) A novel H19 antisense RNA overexpressed in breast
cancer contributes to paternal IGF2 expression. Mol Cell Biol 28:6731–6745
Bestor TH (2000) The DNA methyltransferases of mammals. Hum Mol Genet 9:2395–2402
Brannan CI, Dees EC, Ingram RS et al (1990) The product of the H19 gene may function as an
RNA. Mol Cell Biol 10:28–36
Cai X, Cullen BR (2007) The imprinted H19 noncoding RNA is a primary microRNA precursor.
RNA 13:313–316
Casola S, Pedone PV, Cavazzana AO et al (1997) Expression and parental imprinting of the H19
gene in human rhabdomyosarcoma. Oncogene 14:1503–1510
Charlier C, Segers K, Wagenaar D et al (2001) Human-ovine comparative sequencing of a 250-kb
imprinted domain encompassing the callipyge (clpg) locus and identification of six imprinted
transcripts: DLK1, DAT, GTL2, PEG11, antiPEG11, and MEG8. Genome Res 11:850–862
Chen CL, Ip SM, Cheng D et al (2000) Loss of imprinting of the IGF-II and H19 genes in epithelial
ovarian cancer. Clin Cancer Res 6:474–479
Constancia M, Dean W, Lopes S et al (2000) Deletion of a silencer element in IGF2 results in loss
of imprinting independent of H19. Nat Genet 26:203–206
Cooper MJ, Fischer M, Komitowski D et al (1996) Developmentally imprinted genes as markers
for bladder tumor progression. J Urol 155:2120–2127
Cooper PR, Smilinich NJ, Day CD et al (1998) Divergently transcribed overlapping genes
expressed in liver and kidney and located in the 11p15.5 imprinted domain. Genomics
49:38–51
Crider-Miller SJ, Reid LH, Higgins MJ et al (1997) Novel transcribed sequences within the BWS/
WT2 region in 11p15.5: tissue-specific expression correlates with cancer type. Genomics
46:355–363
Cui H, Onyango P, Brandenburg S et al (2002) Loss of imprinting in colorectal cancer linked to
hypomethylation of H19 and IGF2. Cancer Res 62:6442–6446
Davis E, Caiment F, Tordoir X et al (2005) RNAi-mediated allelic trans-interaction at the
imprinted Rtl1/Peg11 locus. Curr Biol 15:743–749
Dekker J, Rippe K, Dekker M et al (2002) Capturing chromosome conformation. Science
295:1306–1311
Douc-Rasy S, Coll J, Barrois M et al (1993) Expression of the human fetal BAC/H19 gene in
invasive cancer. Int J Oncol 2:753–758
Drewell RA, Brenton JD, Ainscough JF et al (2000) Deletion of a silencer element disrupts H19
imprinting independently of a DNA methylation epigenetic switch. Development 127:3419–3428
Drewell RA, Arney KL, Arima T et al (2002) Novel conserved elements upstream of the H19 gene
are transcribed and act as mesodermal enhancers. Development 129:1205–1213
Du M, Zhou W, Beatty LG et al (2004) The KCNQ1OT1 promoter, a key regulator of genomic

imprinting in human chromosome 11p15.5. Genomics 84:288–300
Dugimont T, Curgy JJ, Wernert N et al (1995) The H19 gene is expressed within both epithelial
and stromal components of human invasive adenocarcinomas. Biol Cell 85:117–124
Dugimont T, Montpellier C, Adriaenssens E et al (1998) The H19 TATA-less promoter is
efficiently repressed by wild-type tumor suppressor gene product p53. Oncogene 16:
2395–2401
Edwards CA, Ferguson-Smith AC (2007) Mechanisms regulating imprinted genes in clusters. Curr
Opin Cell Biol 19:281–289
Elkin M, Shevelev A, Schulze E et al (1995) The expression of the imprinted H19 and IGF-2 genes
in human bladder carcinoma. FEBS Lett 374:57–61
Elson DA, Bartolomei MS (1997) A 50 differentially methylated sequence and the 30 -flanking
region are necessary for H19 transgene imprinting. Mol Cell Biol 17:309–317
Engel N, Raval AK, Thorvaldsen JL et al (2008) Three-dimensional conformation at the H19/
IGF2 locus supports a model of enhancer tracking. Hum Mol Genet 17:3021–3029
Fedoriw AM, Stein P, Svoboda P et al (2004) Transgenic RNAi reveals essential function for
CTCF in H19 gene imprinting. Science 303:238–240
Feil R, Walter J, Allen ND et al (1994) Developmental control of allelic methylation in the
imprinted mouse IGF2 and H19 genes. Development 120:2933–2943
Ferguson-Smith AC, Sasaki H, Cattanach BM et al (1993) Parental-origin-specific epigenetic
modification of the mouse H19 gene. Nature 362:751–755
Fuks F (2003) DNA methyltransferases: from chromatin remodeling to cancer. Med Sci
19:477–480
Giannoukakis N, Deal C, Paquette J et al (1993) Parental genomic imprinting of the human IGF2
gene. Nat Genet 4:98–101
Grandjean V, O’Neill L, Sado T et al (2001) Relationship between DNA methylation, histone H4
acetylation and gene expression in the mouse imprinted IGF2-H19 domain. FEBS Lett
488:165–169
Hao Y, Crenshaw T, Moulton T et al (1993) Tumour-suppressor activity of H19 RNA. Nature
365:764–767
Hark AT, Tilghman SM (1998) Chromatin conformation of the H19 epigenetic mark. Hum Mol
Genet 7:1979–1985
Hark AT, Schoenherr CJ, Katz DJ et al (2000) CTCF mediates methylation-sensitive enhancer-
blocking activity at the H19/IGF2 locus. Nature 405:486–489
Hemberger M, Redies C, Krause R et al (1998) H19 and IGF2 are expressed and differentially
imprinted in neuroectoderm-derived cells in the mouse brain. Dev Genes Evol 208:393–402
Hibi K, Nakamura H, Hirai A et al (1996) Loss of H19 imprinting in esophageal cancer. Cancer
Res 56:480–482
Holmgren C, Kanduri C, Dell G et al (2001) CpG methylation regulates the IGF2/H19 insulator.
Curr Biol 11:1128–1130
Hu JF, Oruganti H, Vu TH et al (1998) The role of histone acetylation in the allelic expression of
the imprinted human insulin-like growth factor II gene. Biochem Biophys Res Commun
251:403–408
Isfort RJ, Cody DB, Kerckaert GA et al (1997) Role of the H19 gene in Syrian hamster embryo cell
tumorigenicity. Mol Carcinog 20:189–193
Ishihara K, Sasaki H (2002) An evolutionarily conserved putative insulator element near the
3’ boundary of the imprinted IGF2/H19 domain. Hum Mol Genet 11:1627–1636
Joubel A, Curgy JJ, Pelczar H et al (1996) The 50 part of the human H19 RNA contains cis-acting
elements hampering its translatability. Cell Mol Biol (Noisy-le-grand) 42:1159–1172
Jouvenot Y, Poirier F, Jami J et al (1999) Biallelic transcription of IGF2 and H19 in individual cells
suggests a post-transcriptional contribution to genomic imprinting. Curr Biol 9:1199–1202
Juan V, Crain C, Wilson C (2000) Evidence for evolutionarily conserved secondary structure in
the H19 tumor suppressor RNA. Nucleic Acids Res 28:1221–1227
Kaffer CR, Srivastava M, Park KY et al (2000) A transcriptional insulator at the imprinted H19/
IGF2 locus. Genes Dev 14:1908–1919
Kanduri C, Holmgren C, Pilartz M et al (2000a) The 5’ flank of mouse H19 in an unusual
chromatin conformation unidirectionally blocks enhancer-promoter communication. Curr
Biol 10:449–457
Kanduri C, Pant V, Loukinov D et al (2000b) Functional association of CTCF with the insulator
upstream of the H19 gene is parent of origin-specific and methylation-sensitive. Curr Biol
10:853–856
Kato Y, Sasaki H (2005) Imprinting and looping: epigenetic marks control interactions between
regulatory elements. Bioessays 27:1–4
Kondo M, Suzuki H, Ueda R et al (1995) Frequent loss of imprinting of the H19 gene is often
associated with its overexpression in human lung cancers. Oncogene 10:1193–1198
Kurukuti S, Tiwari VK, Tavoosidana G et al (2006) CTCF binding at the H19 imprinting control
region mediates maternally inherited higher-order chromatin conformation to restrict enhancer
access to IGF2. Proc Natl Acad Sci USA 103:10684–10689
Landers M, Bancescu DL, Le Meur E et al (2004) Regulation of the large (approximately 1000 kb)
imprinted murine Ube3a antisense transcript by alternative exons upstream of Snurf/Snrpn.
Lee MP, DeBaun MR, Mitsuya K et al (1999) Loss of imprinting of a paternally expressed
transcript, with antisense orientation to KVLQT1, occurs frequently in Beckwith–Wiedemann
syndrome and is independent of insulin-like growth factor II imprinting. Proc Natl Acad Sci
USA 96:5203–5208
Leibovitch MP, Solhonne B, Guillier M et al (1995) Direct relationship between the expression of
tumor suppressor H19 mRNA and c-mos proto-oncogene during myogenesis. Oncogene
10:251–260
Leighton PA, Ingram RS, Eggenschwiler J et al (1995) Disruption of imprinting caused by deletion
of the H19 gene region in mice. Nature 375:34–39
Lewis A, Redrup L (2005) Genetic imprinting: conflict at the Callipyge locus. Curr Biol 15:
R291–R294
Lewis A, Murrel A (2004) Genomic imprinting: CTCF protects the boundaries. Curr Biol
14:284–286
Li T, Hu JF, Qiu X et al (2008) CTCF regulates allelic expression of IGF2 by orchestrating a
promoter-polycomb repressive complex 2 intrachromosomal loop. Mol Cell Biol 28:6473–6482
Li YM, Franklin G, Cui HM et al (1998) The H19 transcript is associated with polysomes and may
regulate IGF2 expression in trans. J Biol Chem 273:28247–28252
Liao B, Hu Y, Herrick DJ et al (2005) The RNA-binding protein IMP-3 is a translational activator
of insulin-like growth factor II leader-3 mRNA during proliferation of human K562 leukemia
Ling JQ, Li T, Hu JF et al (2006) CTCF mediates interchromosomal colocalization between Igf2/
H19 and Wsb1/Nf1. Science 312:269–272
Liu J, Kahri AI, Heikkila P et al (1995) H19 and insulin-like growth factor-II gene expression in
adrenal tumors and cultured adrenal cells. J Clin Endocrinol Metab 80:492–496
Lopes S, Lewis A, Hajkova P et al (2003) Epigenetic modifications in an imprinting cluster are
controlled by a hierarchy of DMRs suggesting longrange chromatin interactions. Hum Mol
Genet 12:295–305
Lottin S, Adriaenssens E, Dupressoir T et al (2002a) Overexpression of an ectopic H19 gene
enhances the tumorigenic properties of breast cancer cells. Carcinogenesis 23:1885–1895
Lottin S, Vercoutter-Edouart AS, Adriaenssens E et al (2002b) Thioredoxin post-transcriptional
regulation by H19 provides a new function to mRNA-like non-coding RNA. Oncogene
21:1625–1631
Lottin S, Adriaenssens E, Berteaux N et al (2005) The human H19 gene is frequently over-
expressed in myometrium and stroma during pathological endometrial proliferative events. Eur
J Cancer 41:168–177
Lustig O, Ariel I, Ilan J et al (1994) Expression of the imprinted gene H19 in the human fetus. Mol
Reprod Dev 38:239–246
Lutz M, Burke LJ, Barreto G et al (2000) Transcriptional repression by the insulator protein CTCF
involves histone deacetylases. Nucleic Acids Res 28:1707–1713
Maher ER and Reik W (2000) Beckwith-Wiedemann syndrome: imprinting in clusters revisited.
J Clin Invest 105:247–252
Manoharan H, Babcock K, Pitot HC (2004) Changes in the DNA methylation profile of the rat H19
gene upstream region during development and transgenic hepatocarcinogenesis and its role in
the imprinted transcriptional regulation of the H19 gene. Mol Carcinog 41:1–16
Milligan L, Antoine E, Bisbal C et al (2000) H19 gene expression is up-regulated exclusively
by stabilization of the RNA during muscle cell differentiation. Oncogene 19:5810–5816
Mitsuya K, Meguro M, Lee MP et al (1999) LIT1, an imprinted antisense RNA in the human
KvLQT1 locus identified by screening for differentially expressed transcripts using mono-
chromosomal hybrids. Hum Mol Genet 8:1209–1217
Miyamoto T, Hasuike S, Jinno Y et al (2002) The human ASCL2 gene escaping genomic
imprinting and its expression pattern. J Assist Reprod Genet 19:240–244
Moore T, Constancia M, Zubair M et al (1997) Multiple imprinted sense and antisense transcripts,
differential methylation and tandem repeats in a putative imprinting control region upstream of
mouse IGF2. Proc Natl Acad Sci USA 94:12509–12514
Murrell A, Heeson S, Reik W (2004) Interaction between differentially methylated regions partitions
the imprinted genes IGF2 and H19 into parent-specific chromatin loops. Nat Genet 36:889–893
Murrell A, Heeson S, Bowden L et al (2001) An intragenic methylated region in the imprinted
IGF2 gene augments transcription. EMBO Rep 2:1101–1106
Nielsen J, Kristensen MA, Willemoës M et al (2004) Sequential dimerization of human zipcode-
binding protein IMP1 on RNA: a cooperative mechanism providing RNP stability. Nucleic
Acids Res 32:4368–4376
Nielsen FC, Nielsen J, Christiansen J (2001) A family of IGF-II mRNA binding proteins (IMP)
involved in RNA trafficking. Scand J Clin Lab Invest Suppl 234:93–99
Ng A, Tang JP, Goh CH et al (2003) Regulation of the H19 imprinting gene expression in human
nasopharyngeal carcinoma by methylation. Int J Cancer 104:179–187
Ohana P, Bibi O, Matouk I et al (2002) Use of H19 regulatory sequences for targeted gene therapy
in cancer. Int J Cancer 98:645–650
Ohana P, Schachter P, Ayesh B et al (2005) Regulatory sequences of H19 and IGF2 genes in
DNA-based therapy of colorectal rat liver metastases. J Gene Med 7:366–374
Ohlsson R, Hedborg F, Holmgren L et al (1994) Overlapping patterns of IGF2 and H19 expression
during human development: biallelic IGF2 expression correlates with a lack of H19 expres-
sion. Development 120:361–368
Okamoto K, Morison IM, Taniguchi T et al (1997) Epigenetic changes at the insulin-like growth
factor II/H19 locus in developing kidney is an early event in Wilms tumorigenesis. Proc Natl
Acad Sci USA 94:5367–5371
Okutsu T, Kuroiwa Y, Kagitani F et al (2000) Expression and imprinting status of human PEG8/
IGF2AS, a paternally expressed antisense transcript from the IGF2 locus, in Wilms’ tumors.
J Biochem (Tokyo) 127:475–483
Pachnis V, Belayew A, Tilghman SM (1984) Locus unlinked to alpha-fetoprotein under the
control of the murine raf and Rif genes. Proc Natl Acad Sci USA 81:5523–5527
Pachnis V, Brannan CI, Tilghman SM (1988) The structure and expression of a novel gene
activated in early mouse embryogenesis. EMBO J 7:673–681
Pandey RR, Ceribelli M, Singh PB et al (2004) NF-Y regulates the antisense promoter, bidirec-
tional silencing, and differential epigenetic marks of the Kcnq1 imprinting control region.
J Biol Chem 279:52685–52693
Paulsen M, Davies KR, Bowden LM et al (1998) Syntenic organization of the mouse distal
chromosome 7 imprinting cluster and the Beckwith-Wiedemann syndrome region in chromo-
some 11p15.5. Hum Mol Genet 7:1149–1159
Paulsen M, El-Maarri O, Engemann S, Strodicke M et al (2000) Sequence conservation and

variability of imprinting in the Beckwith-Wiedemann syndrome gene cluster in human and
mouse. Hum Mol Genet 9:1829–1841
Pedone PV, Pikaart MJ, Cerrato F et al (1999) Role of histone acetylation and DNA methylation in the
maintenance of the imprinted expression of the H19 and IGF2 genes. FEBS Lett 458:45–50
Perincheri S, Dingle RW, Peterson ML et al (2005) Hereditary persistence of alpha-fetoprotein and
H19 expression in liver of BALB/cJ mice is due to a retrovirus insertion in the Zhx2 gene. Proc
Poirier F, Chan CT, Timmons PM et al (1991) The murine H19 gene is activated during embryonic
stem cell differentiation in vitro and at the time of implantation in the developing embryo.
Development 113:1105–1114
Prawitt D, Enklaar T, Klemm G et al (2000) Identification and characterization of MTR1, a novel
gene with homology to melastatin (MLSN1) and the trp gene family located in the BWS-WT2
critical region on chromosome 11p15.5 and showing allele-specific expression. Hum Mol
Genet 9:203–216
Rachmilewitz J, Elkin M, Rosensaft J et al (1995) H19 expression and tumorigenicity of chorio-
carcinoma derived cell lines. Oncogene 11:863–870
Rainho CA, Kowalski LP, Rogatto SR (2001) Loss of imprinting and loss of heterozygosity on
11p15.5 in head and neck squamous cell carcinomas. Head Neck 23:851–859
Reik W, Brown KW, Schneid H et al (1995) Imprinting mutations in the Beckwith–Wiedemann
syndrome suggested by altered imprinting pattern in the IGF2-H19 domain. Hum Mol Genet
4:2379–2385
Reik W, Dean W, Walter J (2001) Epigenetic reprogramming in mammalian development.
Science 293:1089–1093
Reik W, Murrell A (2000) Genomic imprinting silence across the border. Nature 405:408–409
Rinn JL, Kertesz M, Wang JK et al (2007) Functional demarcation of active and silent chromatin
domains in human HOX loci by noncoding RNAs. Cell 129:1311–1323
Ripoche MA, Kress C, Poirier F et al (1997) Deletion of the H19 transcription unit reveals the
existence of a putative imprinting control element. Genes Dev 11:1596–1604
Rougeulle C, Heard E (2002) Antisense RNA in imprinting: spreading silence through air. Trends
Genet 18:434–437
Rougeulle C, Cardoso C, Fontes M et al (1998) An imprinted antisense RNA overlaps UBE3A and
a second maternally expressed transcript. Nat Genet 19:15–16
Royo H, Cavaillé J (2008) Non coding RNAs in imprinted gene clusters. J Biol Cell 100:149–166
Runge S, Nielsen FC, Nielsen J et al (2000) H19 RNA binds four molecules of insulin-like growth
factor II mRNA-binding protein. J Biol Chem 275:29562–29569
Runte M, Huttenhofer A, Gross S et al (2001) The IC-SNURF-SNRPN transcript serves as a host
for multiple small nucleolar RNA species and as an antisense RNA for UBE3A. Hum Mol
Genet 10:2687–2700
Sasaki H, Jones PA, Chaillet JR et al (1992) Parental imprinting: potentially active chromatin of
the repressed maternal allele of the mouse insulin-like growth factor II (IGF2) gene. Genes
Dev 6:1843–1856
Seitz H, Youngson N, Lin SP et al (2003) Imprinted microRNA genes transcribed antisense to a
reciprocally imprinted retrotransposon-like gene. Nat Genet 34:261–262
Sleutels F, Zwart R, Barlow DP (2002) The non-coding Air RNA is required for silencing
autosomal imprinted genes. Nature 415:810–813
Steenman MJ, Rainier S, Dobry CJ et al (1994) Loss of imprinting of IGF2 is linked to reduced
expression and abnormal methylation of H19 in Wilms’ tumour. Nat Genet 7:433–439
Svensson K, Walsh C, Fundele R et al (1995) H19 is imprinted in the choroids plexus and
leptomeninges of the mouse fetus. Mech Dev 51:31–37
Szabo P, Tang SH, Rentsendorj A et al (2000) Maternal-specific footprints at putative CTCF sites
in the H19 imprinting control region give evidence for insulator function. Curr Biol
10:607–610
Szabo PE, Tang SH, Silva FJ et al (2004) Role of CTCF binding sites in the IGF2/H19 imprinting
control region. Mol Cell Biol 24:4791–4800
Takai D, Gonzales FA, Tsai YC et al (2001) Large scale mapping of methylcytosines in CTCF-
binding sites in the human H19 promoter and aberrant hypomethylation in human bladder
cancer. Hum Mol Genet 10:2619–2626
Tanos V, Prus D, Ayesh S et al (1999) Expression of the imprinted H19 oncofetal RNA in
epithelial ovarian cancer. Eur J Obstet Gynecol Reprod Biol 85:7–11
Tanos V, Ariel I, Prus D et al (2004) H19 and IGF2 gene expression in human normal, hyper-
plastic, and malignant endometrium. Int J Gynecol Cancer 14:521–525
Tessier CR, Doyle GA, Clark BA et al (2004) Mammary tumor induction in transgenic mice
expressing an RNA-binding protein. Cancer Res 64:209–214
Thakur N, Tiwari VK, Thomassin H et al (2004) An antisense RNA regulates the bidirectional
silencing property of the Kcnq1 imprinting control region. Mol Cell Biol 24:7855–7862
Thorvaldsen JL, Duran KL, Bartolomei MS (1998) Deletion of the H19 differentially methylated
domain results in loss of imprinted expression of H19 and IGF2. Genes Dev 12:3693–3702
Thorvaldsen JL, Mann MR, Nwoko O et al (2002) Analysis of sequence upstream of the endoge-
nous H19 gene reveals elements both essential and dispensable for imprinting. Mol Cell Biol
22:2450–2462
Tremblay KD, Duran KL, Bartolomei MS (1997) A 50 2-kilobase-pair region of the imprinted
mouse H19 gene exhibits exclusive paternal methylation throughout development. Mol Cell
Biol 17:4322–4329
Tsang P, Gilles F, Yuan L et al (1995) A novel L23-related gene 40 kb downstream of the
imprinted H19 gene is biallelically expressed in mid-fetal and adult human tissues. Hum
Mol Genet 4:1499–1507
Ulaner GA, Yang Y, Hu JF et al (2003) CTCF binding at the insulin-like growth factor-II (IGF2)/
H19 imprinting control region is insufficient to regulate IGF2/H19 expression in human
tissues. Endocrinology 144:4420–4426
Verkerk AJ, Ariel I, Dekker MC et al (1997) Unique expression patterns of H19 in human
testicular cancers of different etiology. Oncogene 14:95–107
Verona RI, Mann MR, Bartolomei MS (2003) Genomic imprinting: intricacies of epigenetic
regulation in clusters. Annu Rev Cell Dev Biol 19:237–259
Vu TH, Chuyen NV, Li T et al (2003) Loss of imprinting of IGF2 sense and antisense transcripts in
Wilms’ tumor. Cancer Res 63:1900–1905
Webber AL, Ingram RS, Levorse JM et al (1998) Location of enhancers is essential for the
imprinting of H19 and IGF2 genes. Nature 391:711–715
Weber M, Hagège H, Murrell A et al (2003) Genomic imprinting controls matrix attachment
regions in the Igf2 gene. Mol Cell Biol 23:8953–8959
Wilkin F, Paquette J, Ledru E et al (2000) H19 sense and antisense transgenes modify insulin-like
growth factor-II mRNA levels. Eur J Biochem 267:4020–4027
Wilkins JF, Haig D (2003) What good is genomic imprinting: the function of parent-specific gene
expression. Nat Rev Genet 4:359–368
Wiseman RW, Montgomery JC, Hosoi J et al (1991) Identification of genes associated with tumor
suppression in Syrian hamster embryo cells. Environ Health Perspect 93:105–109
Westerman BA, Poutsma A, Looijenga LH et al (2001) The Human Achaete Scute Homolog
2 gene contains two promotors, generating overlapping transcripts and encoding two proteins
with different nuclear localization. Placenta 22:511–518
Wolffe AP (2000) Transcriptional control: imprinting insulation. Curr Biol 10:R463–465
Yoon YS, Jeong S, Rong Q et al (2007) Analysis of the H19ICR Insulator. Mol Cell Biol
27:3499–3510
Yang Y, Hu JF, Ulaner GA et al (2003) Epigenetic regulation of IGF2/H19 imprinting at CTCF
insulator binding sites. J Cell Biochem 90:1038–1055
304:594–596
Youngson NA, Kocialkowski S, Peel N et al (2005) A small family of sushi-class retrotransposon-

derived genes in mammals and their relation to genomic imprinting. J Mol Evol 61:481–490
Zemel S, Bartolomei MS, Tilghman SM (1992) Physical linkage of two mammalian imprinted
genes, H19 and insulin-like growth factor 2. Nat Genet 2:61–65
Zhang Y, Tycko B (1992) Monoallelic expression of the human H19 gene. Nat Genet 1:40–44
Zhao Z, Tavoosidana G, Sjolinder M et al (2006) Circular chromosome conformation capture
(4C) uncovers extensive networks of epigenetically regulated intra- and interchromosomal
interactions. Nat Genet 38:1341–1347
Index
A Antiviral therapy, 192

aa. See Amino acids Arabidopsis, 75–78, 81, 82, 84
AAV. See Adeno-associated virus Argonaut (AGO), 63
ABRs. See Auditory brainstem responses Argonaute, 3–6, 10–22, 334, 336–340,
AC4, 76, 86 345, 346
Acid-base catalysis, 313, 314, 317–319 Ago2, 336, 338–340, 345, 346
Active caspase-3, 133 Argonaute 2 (Ago2), 32
Adeno-associated virus (AAV), 136–138, ASOs. See Antisense deoxyoligonucleotides;
140–144, 150, 151, 155 Antisense oligonucleotides
Adenovirus, 136, 138, 139 Auditory brainstem responses (ABRs),
Adipogenesis, 221 215–217
AFB2, 81
AFB3, 81 B
AGO. See Argonaut B2, 66, 72–74, 86
AGO1, 77, 78 2b, 66, 67, 69, 71, 72, 78, 80, 85
Ago2. See Argonaute 2 6b5, 70, 71
Agrobacterium, 69–70, 81, 83 Barrier, 356, 357, 360–361, 364
Agrobacterium tumefaciens, 80 Base pair (bp), 63, 64, 73, 74, 76, 78, 79
Allele-specific silencing of mutant huntingtin, BBB. See Blood–brain barrier
144–148 Beckwith–Wiedemann syndrome
Allele-specific SNPs, 146, 147 (BWS), 396
Alzheimer’s disease, 403, 406 Beet western yellows virus (BWYV),
Amino acids (aa), 73, 82 77, 85
Aminoglycoside antibiotics, 217 Bladder, 211, 219–220
Amplicon, 83, 84 Bladder hyperactivity, 219–220
Anandamide, 211, 220 Blood–brain barrier (BBB), 116–118
Antagomirs, 407, 408 Body mass, 222
Antioxidants, 214, 217 Bone cancer pain, 220, 221
Antisense, 256, 272 bp. See Base pair
Antisense deoxyoligonucleotides (ASOs), BWYV. See Beet western yellows virus
382, 385
Antisense oligonucleotides (ASOs) C
field of application, 289 Caenorhabditis elegans, 61
first generation, 287, 289 CAGE. See Cap analysis of gene expression
second generation, 288, 292, 297 Cancer, 346, 396, 401, 402, 404, 405, 407, 408,
third generation, 288 429, 431–436
445
446 Index
Cancer treatment Dicer, 3–17, 20, 21, 62–65, 68, 73, 74,
BCL2 78, 79, 84, 87, 328, 331–339,
bispecific ASOs, 289 345, 346
G3139, 289, 290, 299 Dicer-substrate siRNA, 161–187
SPC2996, 290 DiGeorge syndrome (DGS), 406
clusterin (CLU) Double-stranded RNA (dsRNAs), 61–64, 68,
OGX-011, 292, 293 73–76, 78, 82, 83, 85–87
survivin, BIRC5 Drosha, 328–334, 339, 343, 346
EZN3042 (SPC3042), 292 dsRNAs. See Double-stranded RNA
LY2181308 (ISIS 23722), 291, 292 DUF283, 63
transforming growth factor, b2 (TGFB2)
AP 12009, 293, 294 E
XIAP E3, 77
AEG35156 (GEM640), 291 EGFR. See Epidermal growth factor receptor
Cap analysis of gene expression (CAGE), E3L, 68, 87
373, 374 Endocytosis, 35, 40, 44, 47
Caveosomal endocytosis pathway, 35 Endogenous siRNA (endo-siRNA), 11–12,
Cell-penetrating peptides (CPPs), 31, 64, 78, 82, 85
39–42, 45 Endoplasmic reticulum (ER), 35–37, 134
Cellular uptake, 31–35, 40, 44–47, 50 endo-siRNA. See Endogenous siRNA
Chalcone synthase (CHS), 62 Endosomal escape, 45
Chronic pain, 162, 163, 165, 167, 184, 186 Enhanced RNAi-1 (eri-1), 82
CHS. See Chalcone synthase Epidermal growth factor receptor (EGFR),
Citrus tristeza virus (CTV), 68, 69, 71, 80, 111, 114, 122
84, 85 Epigenetic regulation, 374, 378, 379
Clinical trials, 49, 50 Ash1, 379
CLIP chromatin remodeling, 378
HITS–CLIP, 384 G9a, 380
immunoprecipitation, 384 genomic imprinting, 379
CMV. See Cucumber mosaic virus histone methytransferase, 379
CNV. See Cucumber necrosis virus polycomb group (PcG), 378–380
Coat protein (CP), 65, 69, 71, 77, 80, 84, 85 PRC2, 379
Combinatorial RNAi, 196–198 trithorax group (TrxG), 379
Conditional RNAi, 142 X-chromosome inactivation (XCI), 379
Conformational change, 313 ER. See Endoplasmic reticulum
CPPs. See Cell-penetrating peptides eri-1. See Enhanced RNAi-1
CRV. See Cymbidium ringspot virus Exosomes, 363
CTV. See Citrus tristeza virus Exportin-5, 330, 332–333, 341, 344, 345
CTV CP, 69, 80, 84 Extracellular RNA (exRNA), 34
Cucumber mosaic virus (CMV), 66, 67, 69,
71, 72, 78, 80, 85 F
Cucumber necrosis virus (CNV), 75 Fatty acid oleoylethanolamide, 221
20 ,30 -Cyclic phosphate, 307, 309, 317 Flock house virus (FHV), 66, 67, 72, 73, 86
Cyclophosphamide, 220 FNY2b, 78
Cymbidium ringspot virus (CRV), 75, 86 Functional genomics, 227–249
G
D General acid, 314
Defense, 356–358, 360–364 General base, 314, 318
Delivery, 29–50 Gene silencing, 225–273
DGCR8, 329–331, 334, 346 Gene therapy, 194–196, 198–201
Diabetic peripheral neuropathy, Genomic imprinting, 395
213–214, 217 GFP. See Green fluorescent protein
Index 447
Glioblastoma, 108, 109, 120–124 serotonin, 218

GNRA tetraloop, 316, 317 Inflammatory pain, 211, 218, 222
Golgi apparatus, 36, 37 In-line attack, 310, 313–315
GPCR knockdown, 163, 184 Interferon (IFN), 96–101
Green fluorescent protein (GFP), 67, 69, 71, Interferon-stimulated genes (ISG), 96,
74, 75, 77, 81–83 99–102
Group I intron, 306, 312 Intracellular release, 38–39
GUS, 70, 71, 84 In vitro selection, 321
Ion channel knockdown, 163
H
91H, 431–433, 435, 436 K
Hairpin, 312 126-kDa, 77
Hammerhead, 305–321 Km, 307, 319
full-length, 307–310, 313–317, 319–321 Knockout mice, 212–214, 218, 219, 221, 222
gene expression, 311, 320
(in) mammals, 311, 320 L
minimal, 307–313, 317–320 Lateral ventricle, 136, 137, 140
natural, 310, 316 Lentiviral vector, 194, 195, 197–201
Schistosomal, 313–315, 321 Lentivirus, 135, 136, 139, 142, 143, 150, 151
HC-Pro. See Helper component-proteinase Lund, E., 332, 345
HD. See Huntington’s disease
HDV, 312 M
Heat shock protein 70, 133 Medium spiny neurons, 142
Heat-shock RNA-1 (HSR1), 375 Meiotic recombination hotspot locus (mrhl)
Helper component-proteinase (HC-Pro), RNA, 373
66, 67, 76, 77, 82, 84, 86 Melanin, 228–234, 237, 238, 240, 242, 244,
HEN1, 76, 77 246, 248–249
Heterozygous, 132, 140, 147 MENa, 372
H19/IGF2, 419–436 MENb, 372, 373, 382
High-throughput screen, 134 MENe and -b, 382
HIV-1, 191–202 Meristem, 80
Hoogsteen, 315–317, 321 Metalloenzymes, 309
HSR1. See Heat-shock RNA-1 Microarray, 133, 135, 149
Human genome, 370, 386 Microinjection, 39, 44
Huntington’s disease (HD), 131–155 Microprocessor, 7, 20
Hyperalgesia, 211–213, 218, 220 MicroRNAs (miRNAs), 2, 6–9, 11–16, 18–23,
Hyperthermia, 211, 212, 222 64, 68, 72, 76–79, 81, 82, 84–87,
Hypoalgesia, 214 115, 119–124, 325–347, 357–365
ac-pre-miRNA, 331, 335, 336, 345
I biogenesis, 102, 325–347
IFN. See Interferon degradation, 332, 340
Ilimaquinone, 36, 37 editing, 327–328
Immunity, 355–365 export, 332–333
Immunoregulation, 357 half-life, 340
Implanted pumps, 154 pre-miRNA, 328–335, 338, 344, 345
Imprinting, 419–436 pre-miRNA processing, 332–335
Inflammation, 211, 217–220, 223 pri-miRNA, 327–332, 343–345
Inflammatory cytokines, 218, 219 pri-miRNA processing, 328, 329, 331, 343
Inflammatory mediators processing factors, 331, 340, 343, 344, 346
adenosine, 218 strand selection, 335, 337, 342
ATP, 218 miR393, 81
bradykinin, 218 miRNA-30 backbone, 101
carrageenan, 218 Mutant huntingtin, 133, 134, 136, 141, 142,
prostaglandins, 218 144–148, 150, 154
448 Index
N meiotic recombination hotspot locus (mrhl)

Nanoparticles, 31, 39, 44, 46, 47 RNA, 373
Natural antisense gene, 281 MENb, 372, 373, 382
Natural antisense RNA, 380 MENe/b, 382, 383
Natural antisense transcripts, 372, 380 MENe/NEAT1, 382
ncRNAs. See Noncoding RNAs microRNAs (miRNAs), 373, 381,
Neurogenic inflammation, 218 400–404, 407–409
Neuropathic pain, 211–214, 222 NRON (noncoding repressor of NFAT),
Nicotiana benthamiana, 65, 67 375, 376
Nicotiana clevelandii, 65 NRSE smRNA, 398
NMD. See Nonsense-mediated decay PCGEM1, 384
NMR, 73 Piwi-interacting RNAs, 373
Nodamura virus (NoV), 73, 86 RepA, 376, 379
Nonallele-specific silencing of huntingtin, RNA coactivator (SRA), 374
135–144 Rncs-1, 381
Noncancerous disease satellite III (SatIII), 383
asthma 7SK RNA, 397, 398
TPI ASM8, 294 small nucleolar RNAs (snoRNAs), 371,
cardiovascular disease 385, 406
ISIS 301012, 294, 295 SRG1, 376, 377
duchenne muscular dystrophy (DMD) Tsix, 379
AVI-4658, 298 UHG, 371
AVI-5038, 296 XIST, 397
exon-skipping, 295–296 Xist, 379
PRO051, 296 Xlsirts, 383
virus infections Non-covalent complexation, 41, 42, 44, 45
AVI-6002, 296 Nonopioid drugs, 221
AVI-6003, 296 Nonsense-mediated decay (NMD),
MicroRNA-122, 296 370, 371
SPC3649, 296 premature termination codon, 370–371
Noncoding repressor of NFAT (NRON), UPF1, 371
375, 376 Non-TLR sensors, 97
Noncoding RNAs (ncRNAs), 369–386, Nonviral delivery systems, 31
419–436 NoV. See Nodamura virus
Airn, 379–380 NOX3, 214, 217
air RNA, 396, 404 NS1, 68, 86
antisense transcripts, 396, 397, 399, Nuclear stress bodies (nSBs), 383
400, 406 Nucleophile, 313, 314, 317
CTN-RNA, 383 Nucleotide (nt), 64, 75, 77–79, 82, 84,
DD3/PCA3, 384 85, 87
Evf-1, 374, 375
Evf-2, 374, 375 O
gas5, 371 Off-target effects, 135, 143, 144, 148, 149,
Gomafu, 383 151–153
HAR1F, 386 Oilseed rape mosaic tobamovirus
heat shock RNA-1 (HSR1), 375 (ORMV), 77
Hotair, 376, 379 Oligodeoxynucleotide, 256, 272
IPS1, 381 Oligonucleotide-based drugs, 31, 38
Kcnq1ot1, 380 Omi/HtrA2, 133, 134
lincRNAs, 379 Opioid drugs, 221
LIT1, 396, 404 Organ of Corti, 214, 215
Malat-1, 372, 373, 383, 384 ORMV. See Oilseed rape mosaic tobamovirus
mascRNA, 372, 373 Ototoxicity, 214–217, 222
Index 449
P R
P0, 77, 78, 85 RDRC. See RNA-dependent RNA
P19, 74–76, 83, 86 polymerase complex
P21, 76 RdRP. See RNA-dependent RNA polymerase
P25, 79, 80, 86 RDVI. See RNAi-directed viral immunity
p53, 277–283 Reactive oxygen species (ROS), 214, 217, 219
Paraspeckles, 381–383 Red chili peppers, 221
Parkinson’s disease, 403 Regulator of gene silencing-calmodulin-like
Pathogen-associated molecular pattern protein (rgs-CaM), 82
(PAMP), 81 rgs-CaM. See Regulator of gene silencing-
Pathologies, 433–436 calmodulin-like protein
PAZ, 63, 64, 77, 78 Rheumatoid arthritis, 218, 219
PAZ domain, 5, 7, 9, 17 RISC. See RNA-induced silencing complex
P-body, 18 RITS. See RNA-induced transcriptional
PcG. See Polycomb group silencing complex
PCR, 215 RNA
PDS. See Phytoene desaturase cleavage, 257, 258, 261, 263–265, 269
PFV. See Primate foamy virus regulation, 277–283
Phosphodiester isomerization, 307, 308, silencing, 1–23
317, 318 structure, 256
Phosphorothioate, 313 RNA-binding protein, 377, 380, 382,
Phosphorothioate-stimulated uptake, 34–35 384, 385
Phytoene desaturase (PDS), 65 68-kDa subunit of cleavage factor
Pigmentation, 229–236, 240–242, 246–249 I m, 382
Ping-Pong amplification loop, 9 p54/nrb, 382
piRNA. See Piwi-interacting RNA PSF, 382
PIWI, 63–65 PSP1, 382
PIWI domain, 5 PSP2/CoAA, 382
Piwi-interacting RNA (piRNA), 4, 6, TLS/FUS, 377, 380
8–11, 17, 64 RNA-dependent RNA polymerase (RdRP), 3,
Piwi subfamily, 4, 8 4, 6, 10, 11, 15–18, 63, 64, 71, 75
Polarity, 358, 360, 361 RNA-dependent RNA polymerase complex
Polycations, 46, 47 (RDRC), 16
Polycomb group (PcG), 378–380 RNA-DNA triple helix, 377
Polycomb-repressive complex 2 (PRC2) RNAi-based therapy for HD, 133, 142, 149
PostGolgi transport, 134 RNAi-directed viral immunity (RDVI), 65–68,
Posttranscriptional gene silencing (PTGS), 72, 76
62, 75 RNA-induced silencing complex (RISC), 3, 5,
Potato virus X (PVX), 66, 67, 75, 79, 80, 12–21, 63, 64, 66, 73, 75, 76, 78,
84, 86 328, 331, 333, 336–340, 342, 344
Potato virus Y (PVY), 65, 67, 86 RNA-induced transcriptional silencing
Preclinical testing, 135 complex (RITS), 16
Primate foamy virus (PFV), 68, 87 RNA interference (RNAi), 2–6, 10, 11, 13, 17,
Protein transduction domains (PTDs), 39, 42, 43 18, 21, 22, 31, 32, 59–87, 107–124,
Pseudogene, Makorin1, 381 163–167, 172, 186, 191–202,
Pseudomonas syringae, 81 212–223, 327, 333, 334, 336, 337,
PTDs. See Protein transduction domains 340–347, 375, 382, 385
PTGS. See Posttranscriptional gene silencing RNAi therapy, 346–347
PVX. See Potato virus X shRNA, 94–96, 99, 100, 103
PVY. See Potato virus Y siRNA, 94–96, 99, 100, 103
RNA–protein (RNP) complex, 384
Q RNA protein interactions, 384
Q2b, 78, 80 RNase3, 76, 85
450 Index
RNase H, 64 Guanosine-rich ASOs, 298

RNase III, 3–7, 17 off-target effects, 297–299
RNase P, 306, 312, 372, 373 sequence-dependent, 299
rncs-1, 78, 79 sequence-independent, 297
RNP complex, 384 SPFMV. See Sweet potato feathery mottle
ROS. See Reactive oxygen species virus
S-phase kinase-related protein 1
S (SKP1), 77
Satellite III (SatIII), 383 SRA. See Steroid receptor RNA coactivator
Satellite RNAs, tobacco ringspot virus, 306, sRNAs. See Small RNAs
307, 315, 318 ssRNA. See single-stranded RNA
SCF. See Skp1-Cul1/Cdc53,-F-box protein Steroid receptor RNA coactivator, 374
Schistosomal, 313–315, 321 Striatum, 132, 135–144, 150, 151, 153, 154
Schistosome, 311 Sugarcane yellow leaf virus (SYLV), 78
Schizophrenia, 403, 406 Sweet potato chlorotic stunt virus
Schizosaccharomyces pombe, 63 (SPCSV), 76, 85
Scissile phosphates, 313, 314, 317 Sweet potato feathery mottle virus
SDN1. See Small RNA degrading nuclease 1 (SPFMV), 76
Seed, 6, 8, 16, 22 SYLV. See Sugarcane yellow leaf virus
Seed sequence, 64 Synthetic siRNA, 133, 136, 140, 141,
Sequence motifs, 101 144, 150, 151
Sequence requirements, 307, 320, 321
SGS3, 75 T
Short hairpin RNA (shRNA), 78, 112, 113, T19, 70, 71
115–118, 135–145, 148, Tas, 68, 86
150–153, 155 ta-siRNA, 15
Short interfering RNA (siRNA), 213–217, 219, Tat, 68, 86
221, 223, 228, 233, 234, 237, TBSV. See Tomato bushy stunt virus
239–246, 248, 249, 256–273, TCV. See Turnip crinkle virus
333–346 TCV CP, 77
Signaling, 357–363, 365 T-DNA. See Transfer DNA
Signal peptides, 36, 37 Tenascin-C, 113
Silencing mechanisms, 339–340 Tertiary contacts, 310, 311, 313, 315–317,
Silencing methodology, 177 319–321
Single-stranded RNA (ssRNA), 76 TEV. See Tobacco etch virus
siRNA-peptide conjugate, 35–38 Thermal sensitivity, 214
siRNAs. See Short interfering RNAs; Small Thermogenesis, 222
interfering RNAs TIR1, 81
SKP1. See S-phase kinase-related protein 1 TLR receptors, 96, 97
Skp1-Cul1/Cdc53,-F-box protein (SCF), 77 TNF-a. See Tumor necrosis factor-a
Slicer, 3–5, 8, 9, 13, 17 Tobacco etch virus (TEV), 65, 66, 86
Small interfering RNA (siRNA), 2–8, 10–19, Tobacco mosaic virus (TMV), 77, 86
21, 32, 41, 42, 44, 48–50, 62–66, 68, Tomato bushy stunt virus (TBSV), 75,
71–78, 81, 82, 84, 85, 112, 113, 83, 86
115–118 Tomato yellow leaf curl geminivirus
Small nucleolar RNAs (snoRNAs), 371, 385 (TYLCV), 73
Small RNA degrading nuclease 1 (SDN1), 82 Transcriptional interference, 374, 376–380,
Small RNAs (sRNAs), 64, 65, 82, 85 385, 386
SPCSV. See Sweet potato chlorotic Transcriptomic analyses, 370, 372
stunt virus CAGE, 373, 374
Specificity cap structure, 370, 374
CpG dinucleotides, 298 ENCODE project, 370
G3139, 298, 299 full-length cDNA, 370, 372
Index 451
next-generation sequencer, 380 TYLCV. See Tomato yellow leaf curl

next-generation sequencing, 373 geminivirus
poly (A) tail, 370, 372 Type III secretion system, 81
RNA polymerase II, 370, 372, 373,
375, 382 U
tiling array, 370, 380 30 -Untranslated regions, 311
Transfer DNA (T-DNA), 67, 69, 80, 81, 83 U6 promoter, 96, 100
Transgenic animals, 136, 140, 144, 148 Uridine turn, 312
Transient receptor potential vanilloid 1 21U-RNAs, 4, 10
(TRPV1), 209–223
Transposons, 64, 85 V
D2642 Triplet deletion, 145 V2, 73–75, 86
Trithorax group (TrxG), 379 VA1, 78, 79, 84, 87
tRNA-like structure, 372, 373 Viral resistance, 195, 198
TRPV1. See Transient receptor potential Viral suppressor of RNAi (VSR), 66, 68–71,
vanilloid 1 73, 85, 86
TRPV1 antagonists, 211, 213, 214, 217, Virus, 358, 364
220–222 Virus-derived siRNAs (viRNA), 71–77
TRPV1 knockout mice, 212–214, 218, 219, VP35, 68, 86
221, 222
Tumor necrosis factor-a (TNF-a), 218, 219 W
Tumor suppression, 404, 408 Wild-type huntingtin, 133–135, 143, 148, 154
Turnip crinkle virus (TCV), 61, 77, 85 Wrap53, 279–283

2010 Book RNATechnologiesAndTheirApplica

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2010 Book RNATechnologiesAndTheirApplica

Uploaded by

Copyright:

Available Formats

RNA Technologies

For further volumes:

RNA Technologies and

ISBN 978-3-642-12167-8 e-ISBN 978-3-642-12168-5

# Springer-Verlag Berlin Heidelberg 2010

Cover design: deblik, Berlin

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

category of therapeutic targets. As miRNA studies continue to be developed, novel

Berlin, Germany Volker A. Erdmann

The Key Features of RNA Silencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Selected Strategies for the Delivery of siRNA In Vitro

RNAi Suppression and Its Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Strategies to Prevent siRNA-Triggered Cellular Toxicity . . . . . . . . . . . . . . . . . 93

RNAi in Malignant Brain Tumors: Relevance to Molecular

Silencing Huntington’s Disease Gene with RNAi . . . . . . . . . . . . . . . . . . . . . . . . . . 131

Application of Dicer-Substrate siRNA in Pain Research . . . . . . . . . . . . . . . . . 161

RNAi Treatment of HIV-1 Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

Application of RNA Interference to Treat Conditions Associated

Harnessing RNAi-Based Functional Genomics to Unravel

mRNA Structure and its Effects on Posttranscriptional

Antisense RNA-Mediated Regulation of the p53 Tumor

Antisense Oligonucleotides: Insights from Preclinical Studies

What can the New Hammerhead Ribozyme Structures

microRNA Biogenesis and its Impact on RNA Interference . . . . . . . . . . . . . . 325

MicroRNAs in Epithelial Antimicrobial Immunity . . . . . . . . . . . . . . . . . . . . . . . 355

Emerging Roles of Long Noncoding RNAs in Gene Expression

Noncoding RNAs as Therapeutic Targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393

Noncoding RNAs at H19/IGF2 Locus: Role in Imprinting,

Eric Adriaenssens Institut de Biology de Lille, CNRS UMR 8161, 1 rue Pr

Safoora Ahmed Department of Dermatology and Biological Chemistry, University

Veenu Aishwarya Division of Hematology/Oncology, Department of Medicine,

Jayavani Aruri Department of Dermatology and Biological Chemistry, University

Jan Barciszewski Institute of Bioorganic Chemistry of the Polish Academy of

Matthias Bauer Department of Neurology, Hertie Institute for Clinical Brain

Nicolas Beaudet Department of Physiology and Biophysics, Faculty of Medicine

Valérie Bégin-Lavallée Department of Physiology and Biophysics, Faculty of

Ben Berkhout Center for Infection and Immunity Amsterdam (CINIMA),

Nahalie Berteaux Institut de Biology de Lille, CNRS UMR 8161, 1 rue Pr

Xian-Ming Chen Department of Medical Microbiology and Immunology,

Sven Diederichs Helmholtz-University-Group “Molecular RNA Biology &

Louis Doré-Savard Department of Physiology and Biophysics, Faculty of

Kristen M. Drescher Department of Medical Microbiology and Immunology,

Marianne Farnebo Department of Oncology-Pathology, Cancer Center Karolinska

Robert M. Friedlander Neuroapoptosis Laboratory, Department of Neurosurgery,

Susanne Fuessel Department of Urology, Medical Faculty, Dresden University of

Anand K. Ganesan Department of Dermatology and Biological Chemistry,

Alan M. Gewirtz Division of Hematology/Oncology, Department of Medicine,

Stefanie Grund Helmholtz-University-Group “Molecular RNA Biology &

Jun-ichiro Hamada Department of Neurosurgery, Graduate School of Medical

Yutaka Hayashi Department of Neurosurgery, Graduate School of Medical

Tetsuro Hirose Biomedicinal Information Research Center, National Institute

Hsiang Ho Department of Dermatology and Biological Chemistry, University of

Guoku Hu Department of Medical Microbiology and Immunology, Creighton

Sarvesh Jajoo Department of Pharmacology, Southern Illinois University School

Tejbeer Kaur Department of Pharmacology, Southern Illinois University School

Kazuyuki Kawakami Divisions of Translational and Clinical Oncology, Kana-

Daisuke Kita Department of Neurosurgery, Graduate School of Medical Science,

Kai Kraemer Department of Urology, Medical Faculty, Dresden University of

Doreen Kunze Department of Urology, Medical Faculty, Dresden University

Sandra D. Laufer Institut für Molekulare Medizin, ZMSZ, Universität zu Lübeck

Jun Liu Department of Medical Microbiology and Immunology, Creighton

Rui Lu Department of Biological Sciences, Louisiana State University, Baton

Toshinari Minamoto Divisions of Translational and Clinical Oncology, Kanazawa