2020 Osteoporosis and Osteoarthritis.-humana

Methods in
Molecular Biology 2221
Andre J. van Wijnen

Marina S. Ganshina Editors
Osteoporosis
and Osteo-
arthritis
Second Edition
METHODS IN MOLECULAR BIOLOGY
Series Editor
John M. Walker
School of Life and Medical Sciences
University of Hertfordshire
Hatfield, Hertfordshire, UK
For further volumes:

http://www.springer.com/series/7651
For over 35 years, biological scientists have come to rely on the research protocols and
methodologies in the critically acclaimed Methods in Molecular Biology series. The series was
the first to introduce the step-by-step protocols approach that has become the standard in all
biomedical protocol publishing. Each protocol is provided in readily-reproducible step-by-
step fashion, opening with an introductory overview, a list of the materials and reagents
needed to complete the experiment, and followed by a detailed procedure that is supported
with a helpful notes section offering tips and tricks of the trade as well as troubleshooting
advice. These hallmark features were introduced by series editor Dr. John Walker and
constitute the key ingredient in each and every volume of the Methods in Molecular Biology
series. Tested and trusted, comprehensive and reliable, all protocols from the series are
indexed in PubMed.
Osteoporosis and Osteoarthritis
Second Edition
Edited by
Andre J. van Wijnen and Marina S. Ganshina

Stem Cell Therapy and Skeletal Regeneration Lab, Mayo Clinic, Rochester, MN, USA
Editors
Andre J. van Wijnen Marina S. Ganshina
Stem Cell Therapy and Skeletal Stem Cell Therapy and Skeletal
Regeneration Lab Regeneration Lab
Mayo Clinic Mayo Clinic
Rochester, MN, USA Rochester, MN, USA
ISSN 1064-3745 ISSN 1940-6029 (electronic)

Methods in Molecular Biology
ISBN 978-1-0716-0988-0 ISBN 978-1-0716-0989-7 (eBook)
https://doi.org/10.1007/978-1-0716-0989-7
© Springer Science+Business Media, LLC, part of Springer Nature 2021

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, 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.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to
be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been
made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer
Nature.
The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.
Preface
Musculoskeletal regenerative medicine and orthopedic research examine molecular mechan-

isms, cellular pathways, and animal models for skeletal degeneration. The translational
end-goal of these studies is to develop treatments for aging patients either with joint
problems that restrict their mobility, or with increased fracture risk due to reduced bone
mineral density bone loss. This Methods in Molecular Biology volume is intended to assist
investigators concerned with research topics broadly related to osteoporosis, osteoarthritis,
intervertebral disc degeneration, as well as other musculoskeletal disorders.
In this second edition of Osteoporosis and Osteoarthritis, we have enlisted the expertise of
investigators in the field to provide key updates on well-established methods that were
previously incorporated in the earlier edition. In addition, we recruited additional experts
who provided new chapters to cover recently emerging sophisticated techniques and meth-
ods that are essential for an in-depth and state-of-the-art understanding of skeletal develop-
ment and homeostasis, and the pathological mechanisms that cause skeletal degeneration.
The first ten chapters provide detailed methods for cell culture models for examining
skeletal biology at the cellular and molecular levels, as well as methodology required for the
application of powerful and novel recent techniques such as single-cell sequencing, high-
throughput sequencing of novel epigenetic modifications in DNA, fluorescence recovery
after photobleaching (FRAP) microscopy, and in silico modeling of signal networks and
functional output at the cell level. Furthermore, we have incorporated seven chapters on
powerful and informative animal models for understanding skeletal disorders and repair.
Notably, we included one chapter on how to measure pain because it is a key symptom for
patients with bone or joint disorders in a clinical setting.
We are indebted to all authors for their willingness to share detailed protocols that have
been developed, implemented, and/or tested in their research groups for use in the fields of
skeletal biology and musculoskeletal regeneration. We thank John Walker for his encour-
agement to pursue the second edition of this volume.
We trust that the contents will inspire and benefit the next generation of investigators
interested in solving essential questions in skeletal biology that may eventually benefit
patients with bone and joint disorders.
Rochester, MN, USA Andre J. van Wijnen

Marina S. Ganshina
v
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
PART I CELLULAR AND MOLECULAR BIOLOGY OF OSTEOARTHRITIS

AND OSTEOPOROSIS
1 Isolation of Murine and Human Osteocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Matthew Prideaux, Amber Rath Stern, and Lynda F. Bonewald
2 Expansion and Chondrogenic Differentiation of Human Bone
Marrow-Derived Mesenchymal Stromal Cells. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Roberto Narcisi, Wendy J. L. M. Koevoet, and Gergo J. V. M. van Osch
3 A Novel Enzymatic Digestion Approach for Isolation
and Culture of Rodent Bone Marrow Mesenchymal Progenitors. . . . . . . . . . . . . . 29
Leilei Zhong, Lutian Yao, and Ling Qin
4 Isolation of Nucleus Pulposus and Annulus Fibrosus Cells
from the Intervertebral Disc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Guus G. H. van den Akker, Andy Cremers, Donatus A. M. Surtel,
Willem Voncken, and Tim J. M. Welting
5 Engineering Cartilage Tissue by Co-culturing of Chondrocytes
and Mesenchymal Stromal Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Yao Fu, Carlo A. Paggi, Amel Dudakovic, Andre J. van Wijnen,
Janine N. Post, and Marcel Karperien
6 Generation of Induced Pluripotent Stem Cells. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
David R. Deyle
7 Specimen Preparation for Single-Cell Sequencing Analysis
of Skeletal Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Shawon Debnath and Matthew B. Greenblatt
8 Mapping 5-Hydroxymethylcytosine (5hmC) Modifications
in Skeletal Tissues Using High-Throughput Sequencing . . . . . . . . . . . . . . . . . . . . . 101
Fiorella Carla Grandi and Nidhi Bhutani
9 Using FRAP to Quantify Changes in Transcription Factor
Dynamics After Cell Stimulation: Cell Culture, FRAP, Data Analysis,
and Visualization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Kannan Govindaraj and Janine N. Post
10 Quantitative Molecular Models for Biological Processes: Modeling
of Signal Transduction Networks with ANIMO . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Sakshi Khurana, Janet Huisman, Stefano Schivo, and Janine N. Post
vii
viii Contents
PART II IN VIVO MODELS OF SKELETAL TISSUE INJURY, DEGENERATION,

AND REPAIR
11 Generation and Characterization of Mouse Models

for Skeletal Disease. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Gabrielle E. Foxa, Ye Liu, Lisa M. Turner, Alexander G. Robling,
Tao Yang, and Bart O. Williams
12 Drill Hole Models to Investigate Bone Repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Zhijun Li and Jill A. Helms
13 Generation and Experimental Outcomes of Closed Femoral
Fracture in Mice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Joseph L. Roberts, Christopher W. Kinter, and Hicham Drissi
14 Mouse Models of Osteoarthritis: Surgical Model
of Post-traumatic Osteoarthritis Induced by Destabilization
of the Medial Meniscus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Kirsty L. Culley, Purva Singh, Samantha Lessard, Mengying Wang,
Brennan Rourke, Mary B. Goldring, and Miguel Otero
15 Immunostaining of Skeletal Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
Crystal Idleburg, Madelyn R. Lorenz, Elizabeth N. DeLassus,
Erica L. Scheller, and Deborah J. Veis
16 Mapping Regional Cortical Bone Responses to Local
Changes in Loading and Systemic Stimuli. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Sara H. Windahl, Peter J. Delisser, and Gabriel L. Galea
17 Pain and Activity Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
David H. H. Molstad and Elizabeth W. Bradley
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
Contributors
NIDHI BHUTANI • Department of Orthopaedic Surgery, Stanford University, Stanford, CA,

USA
LYNDA F. BONEWALD • Indiana Center for Musculoskeletal Health and Department of
Anatomy, Cell Biology and Physiology, Indiana University, Indianapolis, IN, USA
ELIZABETH W. BRADLEY • Department of Orthopedics, University of Minnesota, Minneapolis,
MN, USA; Stem Cell Institute, University of Minnesota, Minneapolis, MN, USA
ANDY CREMERS • Laboratory for Experimental Orthopedics, Department of Orthopedic
Surgery, Maastricht University, Maastricht, The Netherlands
KIRSTY L. CULLEY • Orthopedic Soft Tissue Research Program, HSS Research Institute, The
Hospital for Special Surgery, New York, NY, USA
SHAWON DEBNATH • Department of Pathology and Laboratory Medicine, Weill Cornell
Medicine, New York, NY, USA
ELIZABETH N. DELASSUS • Musculoskeletal Research Center, Histology and Morphometry
Core, Washington University, St. Louis, MO, USA; Department of Orthopedics,
Washington University, St. Louis, MO, USA
PETER J. DELISSER • Veterinary Specialist Services, Brisbane, Australia
DAVID R. DEYLE • Department of Medical Genetics, Mayo Clinic, Rochester, MN, USA
HICHAM DRISSI • Department of Orthopaedics, Emory University School of Medicine,
Atlanta, GA, USA
AMEL DUDAKOVIC • Department of Orthopedic Surgery, Mayo Clinic, Rochester, MN, USA;
Department of Biochemistry and Molecular Biology, Mayo Clinic, Rochester, MN, USA
GABRIELLE E. FOXA • Program for Skeletal Disease and Tumor Microenvironment, Center for
Cancer and Cell Biology, Van Andel Institute, Grand Rapids, MI, USA
YAO FU • Department of Developmental BioEngineering, TechMed Centre, University of
Twente, Enschede, The Netherlands
GABRIEL L. GALEA • Developmental Biology and Cancer, UCL GOS Institute of Child
Health, London, UK; Comparative Biomedical Sciences, Royal Veterinary College, London,
UK
MARY B. GOLDRING • Orthopedic Soft Tissue Research Program, HSS Research Institute, The
KANNAN GOVINDARAJ • Developmental BioEngineering, TechMed Centre, University of
Twente, Enschede, The Netherlands
FIORELLA CARLA GRANDI • Department of Orthopaedic Surgery, Stanford University,
Stanford, CA, USA
MATTHEW B. GREENBLATT • Department of Pathology and Laboratory Medicine, Weill
Cornell Medicine, New York, NY, USA; Research Division, Hospital for Special Surgery,
New York, NY, USA
JILL A. HELMS • School of Medicine, Stanford University, Palo Alto, CA, USA
JANET HUISMAN • Developmental BioEngineering, TechMed Centre, University of Twente,
Enschede, The Netherlands; Student BioMedical Engineering, University of Twente,
Enschede, The Netherlands
ix
x Contributors
CRYSTAL IDLEBURG • Musculoskeletal Research Center, Histology and Morphometry Core,

Washington University, St. Louis, MO, USA; Department of Orthopedics, Washington
University, St. Louis, MO, USA
MARCEL KARPERIEN • Department of Developmental BioEngineering, TechMed Centre,
University of Twente, Enschede, The Netherlands
SAKSHI KHURANA • Developmental BioEngineering, TechMed Centre, University of Twente,
Enschede, The Netherlands
CHRISTOPHER W. KINTER • Department of Orthopaedics, Emory University School of
Medicine, Atlanta, GA, USA
WENDY J. L. M. KOEVOET • Department of Otorhinolaryngology, Erasmus MC, University
Medical Center, Rotterdam, The Netherlands
SAMANTHA LESSARD • Orthopedic Soft Tissue Research Program, HSS Research Institute, The
YE LIU • Program for Skeletal Disease and Tumor Microenvironment, Center for Cancer
and Cell Biology, Van Andel Institute, Grand Rapids, MI, USA
ZHIJUN LI • School of Medicine, Stanford University, Palo Alto, CA, USA
MADELYN R. LORENZ • Musculoskeletal Research Center, Histology and Morphometry Core,
Washington University, St. Louis, MO, USA; Division of Bone and Mineral Diseases,
Department of Medicine, Washington University, St. Louis, MO, USA
DAVID H. H. MOLSTAD • Department of Orthopedics, University of Minnesota, Minneapolis,
MN, USA
ROBERTO NARCISI • Department of Orthopedics, Erasmus MC, University Medical Center,
Rotterdam, The Netherlands
MIGUEL OTERO • Orthopedic Soft Tissue Research Program, HSS Research Institute, The
CARLO A. PAGGI • Department of Developmental BioEngineering, TechMed Centre,
JANINE N. POST • Department of Developmental BioEngineering, TechMed Centre,
MATTHEW PRIDEAUX • Indiana Center for Musculoskeletal Health and Department of
Anatomy, Cell Biology and Physiology, Indiana University, Indianapolis, IN, USA
LING QIN • Department of Orthopaedic Surgery, Perelman School of Medicine, University of
Pennsylvania, Philadelphia, PA, USA
JOSEPH L. ROBERTS • Department of Orthopaedics, Emory University School of Medicine,
Atlanta, GA, USA
ALEXANDER G. ROBLING • Department of Anatomy and Cell Biology, Indiana University
School of Medicine, Indianapolis, IN, USA; Indiana Center for Musculoskeletal Health,
Indianapolis, IN, USA; Richard L. Roudebush VA Medical Center, Indianapolis, IN,
USA
BRENNAN ROURKE • Orthopedic Soft Tissue Research Program, HSS Research Institute, The
ERICA L. SCHELLER • Musculoskeletal Research Center, Histology and Morphometry Core,
Department of Medicine, Washington University, St. Louis, MO, USA
STEFANO SCHIVO • Department of Computer Science, Open University, Heerlen, The
Netherlands
PURVA SINGH • Orthopedic Soft Tissue Research Program, HSS Research Institute, The
Contributors xi
AMBER RATH STERN • Engineering Systems Inc., Charlotte, NC, USA

DONATUS A. M. SURTEL • Laboratory for Experimental Orthopedics, Department of
Orthopedic Surgery, Maastricht University, Maastricht, The Netherlands
LISA M. TURNER • Pathobiology and Biorepository Team, Center for Cancer and Cell Biology,
Van Andel Institute, Grand Rapids, MI, USA
GUUS G. H. VAN DEN AKKER • Laboratory for Experimental Orthopedics, Department of
Orthopedic Surgery, Maastricht University, Maastricht, The Netherlands
GERGO J. V. M. VAN OSCH • Department of Orthopedics, Erasmus MC, University Medical
Center, Rotterdam, The Netherlands; Department of Otorhinolaryngology, Erasmus MC,
University Medical Center, Rotterdam, The Netherlands
ANDRE J. VAN WIJNEN • Department of Orthopedic Surgery, Mayo Clinic, Rochester, MN,
USA; Department of Biochemistry and Molecular Biology, Mayo Clinic, Rochester, MN,
USA
DEBORAH J. VEIS • Musculoskeletal Research Center, Histology and Morphometry Core,
Department of Medicine, Washington University, St. Louis, MO, USA; Department of
Pathology and Immunology, Washington University School of Medicine, St. Louis, MO,
USA
WILLEM VONCKEN • Department of Molecular Genetics, Maastricht University, Maastricht,
The Netherlands
MENGYING WANG • Orthopedic Soft Tissue Research Program, HSS Research Institute, The
TIM J. M. WELTING • Laboratory for Experimental Orthopedics, Department of Orthopedic
Surgery, Maastricht University, Maastricht, The Netherlands
BART O. WILLIAMS • Program for Skeletal Disease and Tumor Microenvironment, Center for
Cancer and Cell Biology, Van Andel Institute, Grand Rapids, MI, USA
SARA H. WINDAHL • Division of Pathology, Department of Laboratory Medicine, Karolinska
University Hospital, Karolinska Institutet, Huddinge, Sweden
TAO YANG • Program for Skeletal Disease and Tumor Microenvironment, Center for Cancer
and Cell Biology, Van Andel Institute, Grand Rapids, MI, USA
LUTIAN YAO • Department of Orthopaedic Surgery, Perelman School of Medicine, University
of Pennsylvania, Philadelphia, PA, USA
LEILEI ZHONG • Department of Orthopaedic Surgery, Perelman School of Medicine,
University of Pennsylvania, Philadelphia, PA, USA
Part I
Cellular and Molecular Biology of Osteoarthritis

and Osteoporosis
Chapter 1
Isolation of Murine and Human Osteocytes

Matthew Prideaux, Amber Rath Stern, and Lynda F. Bonewald
Abstract
Osteocytes are thought to be the mechanosensors of bone by sensing mechanical loads imposed upon the
bone and transmitting these signals to the other bone cells to initiate bone modeling and remodeling. The
location of osteocytes deep within bone is ideal for their function. However, this location makes the study
of osteocytes in vivo technically difficult. There are several methods for obtaining and culturing primary
osteocytes for in vitro experiments and ex vivo observation. In this chapter, several proven methods are
discussed including the isolation of avian osteocytes from chicks and osteocytes from calvaria and long
bones of young mice. A detailed protocol for the isolation of osteocytes from hypermineralized bone of
mature and aged animals is provided. In addition, a modified version of this protocol that can be used to
isolate osteocytes from human trabecular bone is described.
Key words Osteocyte, Isolation, Age, Culture, Collagenase, Mice
1 Introduction
Osteocytes are the most abundant of the bone cells and recently
found to be multifunctional [1, 2]. They serve as orchestrators of
bone remodeling and regulators of mineral homeostasis. They are
the mechanosensors of bone, sensing imposed bone loads, and
translating these mechanical signals into biological signals of bone
modeling and remodeling. They are housed in cave-like voids
within the bone called lacunae. Their location deep within the
mineralized bone matrix is ideal for their cellular functions, but
makes their observation and study difficult. Methods to isolate
these bone matrix-embedded cells have been developed through-
out the years and vary by the species, state, and extent of minerali-
zation of the bone.
In 1992, the group of Peter Nijweide was the first to describe
the isolation of osteocytes from 18-day-old chick embryos. Their
approach yielded a relatively pure population of osteocytes based on
morphology [3]. In 1995, Kumegawa and colleagues published a
method for isolating primary avian osteocytes from the parietal
Andre J. van Wijnen and Marina S. Ganshina (eds.), Osteoporosis and Osteoarthritis, Methods in Molecular Biology, vol. 2221,
https://doi.org/10.1007/978-1-0716-0989-7_1, © Springer Science+Business Media, LLC, part of Springer Nature 2021
3
4 Matthew Prideaux et al.
bones of 16-day-old chick embryos [4]. Osteocyte morphology

and possible dedifferentiation into osteoblasts was noted in this
study. The method proved reproducible and useful in isolating
primary avian osteocytes for study by other researchers [5–
12]. The bones isolated from the chick embryos are essentially
paper thin and not yet mineralized. The parietal bone is flexible
and easy to digest, making the isolation of avian embryonic osteo-
cytes rather quick and straightforward. However, the drawbacks of
this initial method were that the primary osteocytes were very
young themselves because they were isolated from embryonic
bone, and they were avian, not mammalian. The need to develop
isolation methods for osteocytes from other species was apparent.
A method for isolating primary osteocytes from the calvaria of
neonatal rats was described in 1995 [13]. Mikuni-Takagaki et al.
characterized the osteoblast-osteocyte lineage by describing the
subpopulations of isolated cells. These methods were utilized in
several subsequent publications on the investigation the mechan-
otransduction of osteocytes [14–16]. These studies showed that
the various populations of isolated bone cells responded to
mechanical strain in different manners and at different magnitudes,
providing insight into the highly strain responsive nature of osteo-
cytes. Other researchers have also utilized this isolation method in
their studies of primary neonatal rat calvaria osteocytes [17].
Calvaria from young chicks and neonatal rats are all very thin
and easily processed using sequential collagenase digestions and
calcium chelation with EDTA (ethylenediaminetetraacetic acid).
The rationale for these sequential steps is that removal of mineral
exposes collagen fibers that if digested will release cells embedded
in mineralized tissue. Studies utilizing these primary osteocytes can
provide insight into the behavior of osteocytes during early devel-
opment but are not suitable for the study of osteocytes from
skeletally mature bone, and do not allow the comparison between
primary osteocytes isolated from animals of different ages, species,
and genotypes. The calvaria are also not bones that are typically
mechanically loaded longitudinally during everyday activity and
regularly modeled and remodeled, such as the long bones (femurs,
humeri, and tibiae). Methods for isolating calvarial osteocytes have
also been adapted and applied to the isolation of osteocytes from
neonatal and very young murine long bones with success, and were
even utilized in the creation of several osteocyte-like cell lines from
mice of 2–3 months of age [18–20]. This method was used to
compare osteoblast and osteocyte function and gene expression in
several studies [21–23].
To study the effects of age on osteocytes and osteocytes isolated
from high bone mass mice, a method for isolating primary osteo-
cytes from hypermineralized bone was still needed. When the
methods for isolating primary osteocytes from hypomineralized
bone such as young calvaria and long bones were employed for
Isolation of Murine and Human Osteocytes 5
hypermineralized bone, they produced a very low yield rate mainly

yielding only the surface cells and shallowly embedded osteocytes.
When characterized, the populations of cells were mixed with
considerable variation from isolation to isolation. We recently pub-
lished a method for isolating osteocytes from hypermineralized
bone utilizing nine sequential collagenase and EDTA treatments.
It is similar to previous methods, but key differences are that the
periosteum was removed, and a tissue homogenizer was employed
prior to the final digestion [24]. This method has been utilized by
several laboratories to isolate primary osteocytes from mature
(4–6 months) and aged (22–24 months) murine bone [25, 26].
Although animal models provide a valuable tool to investigate
the pathobiology of human diseases, there are distinct differences
between murine and human bone formation and remodeling
[27]. Therefore, it is essential that discoveries made using mouse
osteocytes are also confirmed in human cells. As for studies using
murine osteocytes, this has proven difficult due to the location of
the osteocytes buried deep within the bone matrix. In vitro studies
can be performed by differentiating human osteoblast cell lines into
osteocytes, but these cell lines often fail to fully recapitulate the
mature osteocyte phenotype [28, 29]. We therefore describe an
additional technique for the isolation of primary human osteocytes
for in vitro culture, which retain similar characteristics to osteocytes
in vivo. This technique will enable researchers to translate findings
in murine osteocyte cell lines and primary cells into human osteo-
cyte biology.
2 Materials
This technique facilitates the isolation of osteocytes from skeletally

mature bone (older than 3–4 months) to aged bone
(22–24 months), and was originally published in Biotechniques
[24]. A modified protocol adapting this technique for isolating
osteocytes from human bone is also described [30]. The early
digestions from both techniques (where noted) can also be used
for obtaining primary osteoblasts. Prior to starting the isolation,
several solutions and media must be prepared. Many of these solu-
tions can be prepared the day before the isolation procedure,
however the collagenase solution should be freshly prepared on
the day of the procedure to ensure optimal activity of the enzyme.
2.1 Isolation 1. Collagenase solution: Dissolve 300 active units/mL of collage-

of Osteocytes from nase type IA (Sigma-Aldrich, St. Louis, MO) in α-minimal
Skeletally Mature essential medium (αMEM). 50 mL is adequate for an isolation
Mouse Bone from one or two mice (see Note 1).
2. EDTA solution: Prepare the 5 mM ethylenediaminetetraacetic

acid tetrasodium salt dehydrate (EDTA) solution in magne-
sium- and calcium-free Dulbecco’s phosphate-buffered solu-
tion (DPBS) with 1% bovine serum albumin. Bring to a neutral
pH of 7.4 by adding HCl. 30 mL is adequate for osteocyte
isolation from one or two mice (see Note 2).
3. Primary bone cell culture medium: On the day before the
isolation, supplement α-minimal essential medium (αMEM)
with 5% heat-inactivated fetal bovine serum (FBS), 5% heat-
inactivated calf serum (CS), and 1% penicillin and streptomycin
(PS). This culture medium is chosen based on the culture of the
MLO-Y4 osteocyte cell line [19]. Store at 4 ˚C.
4. Collagen-coated plates: On the day before the isolation and in a
sterile tissue culture hood, dilute sterile collagen in previously
filtered sterilized 0.02 M acetic acid to a final concentration of
0.15 mg/mL (see Note 3). General use, 8 mL for coating a
100 mm dish. Coat plates for 1 h at room temperature. Tilt to
remove excess collagen and save. This solution can be reused
approximately 6 times and should be kept at 4 C. To use plates
immediately, it is best to rinse the plate with PBS to remove
residual acid; otherwise dry the plates for 1 h (without rinsing
with PBS) with the lids off before storing at 4 C.
5. Surgical instruments to dissect and mince bones: forceps, sur-
gical scissors, and scalpels.
6. 25-, and/or 27-gauge needles and 1 mL syringes.
7. 100% Ethanol.
8. 70% Ethanol.
9. Hank’s balanced salt solution (HBSS) calcium and magnesium-
free.
10. α–Minimal essential medium (αMEM).
11. Heat-inactivated fetal bovine serum (FBS).
12. Penicillin and streptomycin.
13. Gentamicin (optional).
14. 6-Well petri dishes (non-TC treated).
15. 100 mm Petri dishes (non-TC treated).
16. Shaker in incubator.
17. Tissue homogenizer (Medimachine (BD Biosciences, San Jose,
CA)) with a stainless steel mincing screen with a pore size of
50 μm).
2.2 Isolation 1. Collagenase solution: Dissolve 2 mg/mL (approx.

of Osteocytes from 250–300 U) collagenase type II in α-MEM supplemented
Human Trabecular with L-glutamine, ribonucleosides and deoxyribonucleosides,
Bone 1% penicillin/streptomycin, and sterile filter. 85 mL of collage-
nase solution is sufficient to isolate osteocytes from 1.5 g of
trabecular bone.
2. EDTA/BSA solution: Dissolve tetra sodium EDTA in calcium

and magnesium-free Hank’s balanced salt solution (HBSS) to a
concentration of 5 mM and add 0.1% BSA. Adjust the pH to
7.4 with concentrated HCl and sterile filter. 85 mL of EDTA
solution is sufficient for 1.5 g of bone.
3. Osteoblast media preparation: Add 10% heat-inactivated FBS,
1% penicillin and streptomycin to α-MEM.
4. Osteocyte media preparation: Add 2.5% FBS, 1.8 mM potas-
sium phosphate, 1% penicillin and streptomycin.
5. Collagen-coated plates: These are prepared the same as for the
murine protocol.
6. Sterile surgical tools: Bone cutters, forceps, scalpel. These are
sterilized by autoclaving prior to use.
7. 70% Ethanol.
8. Heat-inactivated FBS.
9. Calcium and magnesium-free HBSS.
10. α-MEM with L-glutamine, ribonucleosides and
deoxyribonucleosides.
3 Methods
The osteocyte isolation protocol from mouse bone takes approxi-

mately 10–12 h from the time the mice are sacrificed to the time
that the bone particles are plated. The length of time depends on
the number of mice used and familiarity of the researchers with the
protocol. The protocol for human bone has fewer digests and takes
approximately 4–5 h. We have found that preparing additional
collagenase or EDTA treatments after the sixth digest does not
yield a significant increase in cell number.
3.1 Isolation 1. Aseptically dissect the long bones (femurs, tibiae, and humeri)
of Osteocytes from from the mice using surgical scissors or scalpel. Be sure not to
Skeletally Mature break any of the bones at this point and also try to keep the
Mouse Bone abdomen intact during the dissection to reduce contamination
potential (see Note 4).
2. After dissection of bones and removal of as much soft tissue as
possible, place them in 100 mm petri dishes containing αMEM
with 1% penicillin and streptomycin (and gentamicin (25 μg/
mL)—optional) (see Note 4).
3. Remove any remaining muscle and connective tissue from the
bones and scrape away the periosteum using a scalpel (see
Note 4).
4. Wash the bones in sequential dishes/wells of a six-well plate

filled with αMEM +10% penicillin and streptomycin to remove
fur and other contaminants.
5. Place bones in a 100 mm petri dishes with fresh αMEM with 1%
penicillin and streptomycin (and gentamycin (25 μg/mL)—
optional).
6. Cut off the bone epiphyses and flush the marrow out using a
needle and syringe.
7. Wash the hollowed bone pieces again in αMEM with 1% peni-
cillin and streptomycin (and gentamicin (25 μg/mL)—
optional).
8. Cut the bones in half lengthwise and then cut into 1–2 mm
lengths using a scalpel.
9. As the bone pieces are cut place in HBSS for a brief wash.
10. Collagenase Treatment 1: Incubate the bone pieces in warmed
collagenase solution for 25 min (see Note 5).
11. Aspirate the solution and keep for cell plating (if interested in
Digest 1 cells) (see Notes 6 and 7).
12. Wash the bone pieces with HBSS three times with 5 mL each,
each time adding the HBSS rinse to the aspirated solution for
cell plating.
13. Pellet, resuspend, and plate the cells on collagen-coated plates
using the primary bone cell culture medium.
14. Collagenase Treatment 2: Repeat steps 10–13.
15. Collagenase Treatment 3: Repeat steps 10–13, again. Com-
bine cells with those from step 14 (see Note 8).
16. EDTA Treatment 1: Incubate the bone pieces in warmed
EDTA solution for 25 min (see Note 5).
18. Repeat steps 12 and 13.
21. Repeat steps 12 and 13 (see Note 8).

32. Aspirate the solution and keep for cell plating (see Notes 6 and 7).
34. Mince the bone pieces in αMEM utilizing a tissue
homogenizer.
35. Directly plate the resulting suspension of bone particles in
αMEM on collagen-coated plates adding additional primary
bone cell culture medium if needed (see Note 13).
3.2 Isolation 1. Under sterile conditions, use the bone cutters to remove tra-
of Osteocytes from becular bone pieces from the surgical samples (see Note 14).
Human Trabecular 2. Further dissect the bone pieces into 1–2 mm fragments using
Bone the bone cutters or a scalpel.
3. Place the bone pieces (1–1.5 g, wet weight) into a 50 mL
culture tube.
4. Wash the bone pieces 3 times with 20 mL HBSS containing
Pen/Strep.
5. Digest the bone pieces in 20 mL pre-warmed collagenase type
II (2 mg/mL in α-MEM containing Pen/Strep) for 25 min
with gentle shaking at 37 C.
6. Remove the collagenase from the bone pieces. Wash the bone
pieces 3 times with 10 mL HBSS (see Note 15).
7. Incubate the bone pieces in a second digest of 20 mL collage-
nase for 25 min. Collect the supernatant and save. Wash the
bone pieces 3 times in 10 mL HBSS and add the rinsate to the
saved collagenase solution.
8. Repeat step 7 (third collagenase digest) and combine the
digest solutions and the rinsates. Centrifuge at 500 g for
5 min and resuspend the pelleted cells in 3 mL of osteoblast
media. Divide between 2–3 wells of a collagen-coated 12-well
plate (see Notes 16 and 17).
9. Add 40 mL EDTA solution (5 mM in HBSS with 0.1% BSA,

pH 7.4) to the bone pieces and incubate for 25 min at 37 C
with gentle shaking.
10. Remove and keep the EDTA cell suspension. Rinse the bone
pieces 3 times with 10 mL HBSS and add the rinsate to the
EDTA. Centrifuge at 500 g and resuspend in 2 mL osteocyte
media.
11. Repeat the collagenase digest as in step 7. Centrifuge the
collagenase and rinsate at 500 g and resuspend the cells in
2 mL osteocyte media.
12. Repeat the EDTA treatment as in steps 9 and 10. Combine the
resuspended cells with those from steps 10 and 11 and centri-
fuge at 500 g. Resuspend in 3 mL osteocyte media (see
Note 18).
13. Split the cells between 2–3 wells of a collagen-coated 12-well
plate. Culture for 4–5 days to allow the cells to attach. Replace
half the culture media with fresh osteocyte media 48 h after
plating (see Note 19).
4 Notes
1. The collagenase solution must be prepared fresh the morning

of the isolation.
2. The EDTA solution can be prepared the day before the isola-
tion and stored at 4 ˚C.
3. Use a chilled pipet so the collagen does not stick.
4. Steps 1 and 2 can be conducted on a lab bench. Steps 3–35
should be performed in a sterile laminar flow hood.
5. 8 mL of solution per well in a six-well plate works well for the
long bones from 1–2 mice.
6. The issue of maintaining cell density is quite crucial for the cell
attachment and survival of the later digests. It is recommended
for an isolation using 1–2 mature mice where it is desired to
plate each digest individually, one should use a 6-well plate
format. If similar digests are combined together, digests 7–9
for example, one should use a 100 mm dish format. The bone
particles derived from 1–2 mature mice can be split between
two wells of a 6-well plate.
7. Cells will be immediately visible in digests 1–9 (for cell count-
ing and trypan blue staining) and should attach to the plate
within 24–48 h. These will be primarily surface cells such as
fibroblasts and osteoblasts.
8. These will be primarily osteoblastic cells.
9. These will be a mix of osteoblastic and osteocytic cells.

10. These will be primarily osteoblastic and osteocytic cells. Each
subsequent serial digest will yield a greater percentage of osteo-
cytic cells.
11. These will be primarily osteocytic cells.
12. At this point, the bone pieces can also be used for isolation of
osteocyte mRNA as described previously [31].
13. Do not disturb the bone particle cultures for at least 48 h.
Moving the dish will cause movement of the bone particles and
therefore hinder the attachment of the osteocytes. It is recom-
mended to leave the bone particles for as long as possible,
adding additional primary bone cell medium to the dishes at
72 h, and changing to fresh medium at 4 or 5 days post culture.
It is recommended to use the osteocytic cultures for experi-
mental purposes before day 7 as that is when they were char-
acterized in the BioTechniques manuscript [24]. Prolonged
culture will otherwise lead to dedifferentiation/loss of pheno-
type or an overgrowth of the cultures by any contaminating
fibro- or osteoblasts.
14. Human surgical samples may comprise femoral knee shavings
or trabecular bone removed from the hip during arthroplasty.
Dedicated surgical tools such as bone cutters are preferred to
remove the trabecular bone pieces as scalpel blades are usually
not strong enough for human bone.
15. The first digest will contain mainly fibroblasts and marrow
cells. These cells can be kept and plated or discarded if only
osteoblast and osteocyte fractions are required.
16. If the osteoblast fraction is not required, the collagenase solu-
tion from digests 2 and 3 can be discarded.
17. The bone pieces at this stage contain primarily osteocytes.
These bone pieces can be cultured in osteocyte media without
further digestions, to provide a population of osteocytes still
within their native extracellular matrix.
18. Combining the cells from digests 4–6 and the additional cen-
trifugation/resuspension step removes any remaining traces of
collagenase and helps the cells to adhere to the culture plate.
19. Culturing the cells for more than 5 days may lead to loss of
osteocyte morphology and phenotype. The cells from digests
4–6 may also be suspended in a collagen type I gel, which
results in the cells acquiring a highly dendritic 3D morphology
similar to osteocytes in vivo [30, 32].
References
1. Bonewald LF (2011) The amazing osteocyte. J genes. Biochem Biophys Res Commun 246
Bone Miner Res 26(2):229–238 (2):404–408
2. Dallas SL, Prideaux M, Bonewald LF (2013) 15. Mikuni-Takagaki Y (1999) Mechanical
The osteocyte: an endocrine cell and more. responses and signal transduction pathways in
Endocr Rev 34(5):658–690 stretched osteocytes. J Bone Miner Metab 17
3. van der Plas A, Nijweide PJ (1992) Isolation (1):57–60
and purification of osteocytes. J Bone Miner 16. Mikuni-Takagaki Y et al (1996) Distinct
Res 7(4):389–396 responses of different populations of bone
4. Tanaka K et al (1995) Time-lapse microcine- cells to mechanical stress. Endocrinology 137
matography of osteocytes. Miner Electrolyte (5):2028–2035
Metab 21(1–3):189–192 17. Gu G et al (2005) Estrogen protects primary
5. Aarden EM et al (1996) Adhesive properties of osteocytes against glucocorticoid-induced apo-
isolated chick osteocytes in vitro. Bone 18 ptosis. Apoptosis 10(3):583–595
(4):305–313 18. Kato Y et al (2001) Establishment of an osteoid
6. Ajubi NE et al (1996) Pulsating fluid flow preosteocyte-like cell MLO-A5 that spontane-
increases prostaglandin production by cultured ously mineralizes in culture. J Bone Miner Res
chicken osteocytes—A cytoskeleton- 16(9):1622–1633
dependent process. Biochem Biophys Res 19. Kato Y et al (1997) Establishment of an
Commun 225(1):62–68 osteocyte-like cell line, MLO-Y4. J Bone
7. Kamioka H, Honjo T, Takano-Yamamoto T Miner Res 12(12):2014–2023
(2001) A three-dimensional distribution of 20. Woo SM et al (2011) Cell line IDG-SW3 repli-
osteocyte processes revealed by the combina- cates osteoblast-to-late-osteocyte differentia-
tion of confocal laser scanning microscopy and tion in vitro and accelerates bone formation
differential interference contrast microscopy. in vivo. J Bone Miner Res 26(11):2634–2646
Bone 28(2):145–149 21. Zhao S et al (2002) MLO-Y4 osteocyte-like
8. Kamioka H et al (2007) Primary cultures of cells support osteoclast formation and activa-
chick osteocytes retain functional gap junctions tion. J Bone Miner Res 17(11):2068–2079
between osteocytes and between osteocytes 22. Kramer I et al (2010) Osteocyte Wnt/beta-
and osteoblasts. Microsc Microanal 13 catenin signaling is required for normal bone
(2):108–117 homeostasis. Mol Cell Biol 30(12):3071–3085
9. Kamioka H et al (2006) Fluid shear stress 23. Nakashima T et al (2011) Evidence for osteo-
induces less calcium response in a single pri- cyte regulation of bone homeostasis through
mary osteocyte than in a single osteoblast: RANKL expression. Nat Med 17
implication of different focal adhesion forma- (10):1231–1234
tion. J Bone Miner Res 21(7):1012–1021 24. Stern AR et al (2012) Isolation and culture of
10. Klein-Nulend J et al (1995) Pulsating fluid flow primary osteocytes from the long bones of skel-
increases nitric oxide (NO) synthesis by osteo- etally mature and aged mice. BioTechniques 52
cytes but not periosteal fibroblasts--correlation (6):361–373
with prostaglandin upregulation. Biochem 25. Jahn K et al (2012) Skeletal muscle secreted
Biophys Res Commun 217(2):640–648 factors prevent glucocorticoid-induced osteo-
11. Klein-Nulend J et al (1995) Sensitivity of cyte apoptosis through activation of beta-
osteocytes to biomechanical stress in vitro. catenin. Eur Cell Mater 24:197–209. discus-
FASEB J 9(5):441–445 sion 209-10
12. Westbroek I et al (2000) Differential stimula- 26. Kalajzic I et al (2013) In vitro and in vivo
tion of prostaglandin G/H synthase-2 in approaches to study osteocyte biology. Bone
osteocytes and other osteogenic cells by pulsat- 54(2):296–306
ing fluid flow. Biochem Biophys Res Commun 27. Jilka RL (2013) The relevance of mouse mod-
268(2):414–419 els for investigating age-related bone loss in
13. Mikuni-Takagaki Y et al (1995) Matrix miner- humans. J Gerontol A Biol Sci Med Sci 68
alization and the differentiation of osteocyte- (10):1209–1217
like cells in culture. J Bone Miner Res 10 28. Prideaux M et al (2014) SaOS2 osteosarcoma
(2):231–242 cells as an in vitro model for studying the tran-
14. Kawata A, Mikuni-Takagaki Y (1998) Mechan- sition of human osteoblasts to osteocytes. Cal-
otransduction in stretched osteocytes—Tem- cif Tissue Int 95(2):183–193
poral expression of immediate early and other
29. Bodine PV, Vernon SK, Komm BS (1996) 31. Qing H et al (2012) Demonstration of osteo-
Establishment and hormonal regulation of a cytic perilacunar/canalicular remodeling in
conditionally transformed preosteocytic cell mice during lactation. J Bone Miner Res 27
line from adult human bone. Endocrinology (5):1018–1029
137(11):4592–4604 32. Skottke J, Gelinsky M, Bernhardt A (2019) In
30. Prideaux M et al (2016) Isolation of osteocytes vitro co-culture model of primary human
from human trabecular bone. Bone 88:64–72 osteoblasts and osteocytes in collagen gels. Int
J Mol Sci 20(8):1998
Chapter 2
Expansion and Chondrogenic Differentiation of Human Bone

Marrow-Derived Mesenchymal Stromal Cells
Roberto Narcisi, Wendy J. L. M. Koevoet, and Gergo J. V. M. van Osch
Abstract
Human bone marrow-derived mesenchymal stem/stromal cells (BM-MSC) are adult multipotent progeni-
tor cells that can be isolated from bone marrow. BM-MSCs have the ability to be expanded and differ-
entiated into the chondrogenic lineage in vitro. Here we describe a standardized method to expand and
chondrogenically differentiate human BM-MSCs, highlighting how to overcome technical challenges and
indicating the most common readout parameters to evaluate the chondrogenic differentiation capacity.
Key words Tissue engineering, Regenerative medicine, Cartilage, Chondrogenesis, Mesenchymal

stem cells
1 Introduction
In this manuscript we describe the materials and methods we use to

isolate, expand, and chondrogenically differentiate human bone
marrow-derived mesenchymal stem/stromal cells (BM-MSC). His-
torically, some important milestones substantially contributed to
improve the understanding and the use of BM-MSC as research
tool. Self-renewing bone marrow-derived stem cells have been
originally identified by experiment on serial ectopic transplantation
of bone marrow, thanks to their ability to recapitulate the genera-
tion of complex bone structures including the bone marrow and
stroma [1, 2]. Later, the characterization and the definition of these
cells evolved, and with it, the name by which those cells have been
called [3–6], bringing the scientific community to define, in 2006,
the minimum requirements for in vitro cultured BM-MSC [7],
including surface marker expression profile. However, this is still a
developing area of research and surface markers originally sug-
gested as essential for MSC multilineage differentiation have been
demonstrated to be less crucial for chondrogenesis [8]. An over-
view of MSC (sup)population’s surface profiles was discussed in a
15
16 Roberto Narcisi et al.
previous book chapter (https://doi.org/10.1007/978-3-319-

53316-2_2; Chapter 2).
In 1998, the first robust in vitro chondrogenic induction media
for BM-MSC was developed by Johnstone et al. [9], and the next
year the in vitro multilineage differentiation potential of BM-MSC
was characterized [10]. Like in vivo, the heterogeneity of BM-MSC
is also present in vitro. Important studies revealed the heteroge-
neous nature of BM-MSC in vitro [11, 12] and via in vivo
transplantation-based assays [13, 14]. This heterogeneity is also
linked with the limited proliferation capacity of BM-MSC that
was found to be associated with loss of differentiation capacity
[11, 15]. To improve proliferation and subsequent differentiation
capacity of BM-MSC, the effect of different growth factor supple-
mentations has been tested during expansion, such as WNT, EGF,
PDGF but also platelet lysate or platelet-rich plasma. FGF2, how-
ever, was one of the first successfully tested factors used to stimulate
expansion capacity while maintaining chondrogenic differentiation
capacity of BM-MSC [16].
Based on this knowledge, in the course of the last two decades
we have established in our laboratory robust protocols to efficiently
expand and chondrogenically differentiate BM-MSC. Our current
materials and methods, as well as all the small—yet important—
details to successfully perform the assays, are here described.
2 Materials
All the experiments regarding the manipulation of cells of human

origin described in this manuscript are performed under sterile
conditions using a Biological Safety Cabinet. All materials and
solutions for cell culture should be previously sterilized according
with the manufacture’s instruction for each component.
2.1 Expansion of 1. Complete expansion medium (CEM; see also Note 1):
BM-MSCs (See (a) MEM-α, nucleosides.
Table 1)
(b) 10% Fetal calf serum (FCS; see Note 2).
(c) 1 ng/mL Fibroblast growth factor-2 (FGF2).
(d) 104 M Ascorbic acid-2-phosphate.
(e) 1.5 μg/mL Fungizone (Amphotericin B).
(f) 50 μg/mL Gentamicin.
2. CEM for enhanced expansion and chondrogenic differentia-
tion capacity (CEM+):
(a) CEM.
(b) 250 ng/mL WNT3A or 2.5 μM CHIR99021, both
canonical WNT agonists (see Note 3).
Expansion and Chondrogenic Differentiation of Human Bone Marrow-Derived. . . 17
Table 1
Summary of the materials for the expansion and passaging of BM-MSC
Materials Company Catalogue

®
VACUETTE buis, NH (heparinized tubes) Greiner Bio-One 455051
3% acetic acid with methylene blue StemCell 07060
Technology
Bürker counting chamber VWR International 631-1159
BV
MEM-α, nucleosides Gibco 22571038
Fetal calf serum (FCS; see Note 2) Life technologies Lot:
41Q2047K
Fibroblast growth factor-2 (FGF-2) Serotec PHP105
Ascorbic acid-2-phosphate Sigma-Aldrich A8960
Fungizone Gibco 15290026
Gentamicin Gibco 15750-037
CHIR99021 Stemgent 04-0004
WNT3A In house –
production
Trypsin-EDTA (0.25%) phenol red Gibco T4049
Dulbecco’s phosphate-buffered saline, no calcium, no magnesium Gibco 14190-169
(DPBS)
Culture flasks T175 BD Falcon® 353112
®
Culture flasks T75 BD Falcon 353136
Culture flasks T25 BD Falcon® 353108
Centrifuge tubes, 50 mL, conical bottom Greiner Bio-One T2318-
500EA
Centrifuge tubes, 15 mL, conical bottom Westburg P91015
3. Passaging of BM-MSCs:
(a) CEM.
(b) Trypsin-EDTA (0.25%) phenol red (Gibco).
(c) Dulbecco’s phosphate-buffered saline, no calcium, no
magnesium (DPBS).
2.2 Chondrogenic 1. Complete chondrogenic medium (CCM; see Note 4):

Differentiation of BM- (a) DMEM-high glucose, GlutaMAX™ Supplement,
MSCs (See Table 2) HEPES (DMEM-HG).
(b) Insulin transferring selenic acid (ITS+).
Table 2
Summary of the materials for the chondrogenic differentiation of BM-MSC
Materials Company Catalogue

Bürker counting chamber VWR International BV 631-1159
DMEM-high glucose, GlutaMAX(TM), HEPES Gibco 32430027
Insulin transferring selenic acid (ITS+) B&D Bioscience 354352
L-proline Sigma-Aldrich P5607
Sodium pyruvate Gibco 11360-039
Transforming growth factor-β1 (TGF-β1) R&D Systems 240-B
Ascorbic acid-2-phosphate Sigma-Aldrich A8960
Dexamethasone Sigma-Aldrich D4902
IWP2 (see Note 5) Stemgent 04-0034
Fungizone Gibco 15290026
Gentamicin Gibco 15750-037
Polypropylene centrifuge tubes, screw caps (15 mL) VWR International BV TPPA91019
(c) 40 μg/mL L-Proline.

(d) 1 mM Sodium pyruvate.
(e) 10 ng/mL Transforming growth factor-β1 (TGF-β1).
(f) 104 M Ascorbic acid-2-phosphate (Sigma-Aldrich).
(g) 100 nM Dexamethasone.
(h) 1.5 μg/mL Fungizone (Amphotericin B).
(i) 50 μg/mL Gentamicin.
2. A medium to induce chondrogenesis with reduced hypertro-
phy based on the use of the WNT inhibitor IWP2 was recently
developed (see Note 5).
3 Methods
The most commonly used method for the isolation of BM-MSC

from bone marrow aspirates is by plastic adherence. Bone marrow
aspirates need to be collected in heparinized tubes and processed
ideally within 6 h and anyway no later than 18 h from the harvest-
ing time from the patient/donor. During this time the samples
need to be kept at 4 C. Recently, in a comparative analysis, we
showed the effect of different ways of harvesting bone marrow
aspirates from patients on the chondrogenic differentiation capacity
of BM-MSC [17], while maintaining the same isolation procedure.
After determining the total volume of the aspirate, the cells are
counted by mixing 20 μL of bone marrow aspirate with 380 μL of
3% acetic acid with methylene blue (or comparable volumes in a
1:20 ratio). After 2 min the membrane of white and red blood cells
will be lysate and the remaining white blood cell nuclei can be
counted, for example, with a Bürker counting chamber. The cells
are then seeded at a density of approximately 75,000 25,000
cells/cm2 in CEM (see Subheading 2.1) for a total volume of 20 mL
in a T175 flasks or 8.5 mL in a T75 flasks (see Notes 6 and 7). In
this way BM-MSCs are isolated by their ability to adhere to plastic
culture flasks. After 24 h, nonadherent cells are removed by 3
washing with DPBS+2% FCS and adherent cells cultured in stan-
dard conditions (5% CO2 at 37 C) for up to 14 days. During this
time, cell colonies appear and grow at a speed that depend on
individual BM-MSC donor (Fig. 1a). Medium is renewed twice a
week. When BM-MSCs neared confluence (Fig. 1b) or after a
maximum of 14 days, they are detached and passaged for further
expansion (see Subheading 2.1, step 3 and Subheading 3.1). At this
point we have Passage-1 (P1) BM-MSC.
3.1 Expansion and P1 BM-MSC are 2 washed with DPBS by adding around 50% of
Passaging of BM- the volume used for the expansion media, then Trypsin-EDTA is
MSCs added (3 mL for a T175 flask and 1.25 mL for a T75) for 3–4 min
in the culture incubator (37 C and 5%CO2) till all the cells look
detached after a check with the microscope. The cell suspension is
harvested in appropriate sized tubes by adding CEM (containing
FCS) in order to neutralize the effect of the Trypsin-EDTA solu-
tion. For the harvesting, it is suggested to use a Trypsin-EDTA/
CEM ratio of at least a 1:4. The cell suspension is then centrifuged
at 300 g for 6–8 min in 50 or 15 mL tubes, the resulting cellular
pellet resuspended in CEM or CEM+ and the cell counted, for
example, with a Bürker counting chamber. Next, the BM-MSC are
reseeded at the density of 2300 cells/cm for further expansion (see
Note 6). Generally, BM-MSC take 4–6 days to reach
sub-confluence (75–85% confluence; Fig. 1b) after seeding (see
Note 8). In order to keep as much as possible the BM-MSC in
continuous proliferation, we do not grow them till confluence
(Fig. 1c).
Every passage we monitor the BM-MSC for their morphology
(Fig. 2) and expansion speed. These parameters are known to be
influenced over time in vitro and have an impact on BM-MSC
differentiation capacity, and one reason may be cellular senescence
[18–20]. For example, when during expansion BM-MSC take
>50% more time to reach the sub-confluence compared to the
time needed to go from P1 to P2, we stop the culture and do not
proceed with the differentiation assay (see Note 9). By default, we
would then no longer perform surface marker analysis on our
plastic-adherent-selected BM-MSC (see Note 10).
Fig. 1 Representative images of BM-MSC forming a colony 2 days after isolation (a), at 75–85% confluence (b)
and over-confluent (c). Scale bar ¼ 200 μm
Fig. 2 Representative images of BM-BMCs showing their classical morphology during the expansion (left
panel) and a more enlarged morphology (right panel, black arrow) usually associate with aging and low
expansion rate (right panel). Scale bar ¼ 100 μm
3.2 Chondrogenic BM-MSC are trypsinized as indicated in Subheading 3.1. However,

Differentiation of BM- differentially compared to the passaging protocol, after centrifuga-
MSCs tion at 300 g for 6–8 min, the cells are resuspended in CCM (see
Subheading 2.2). After counting, for example, with a Bürker count-
ing chamber, BM-MSC are divided in aliquots of 200,000 cell/
0.5 mL of CCM in 15 mL polypropylene tubes (see Note 11) and
centrifuged at 300 g for 6–8 min to form a pellet (Fig. 3a). The
generated pellet cultures are immediately transferred in an incuba-
tor for cell culture (37 C, 5%CO2) for the next 24 h. At this stage,
the pellets should roundup and look like a sphere-like cellular
cluster (Fig. 3b). Now, it is suggested to loosen the pellet from
the bottom of the tube by tapping it gently. From this moment-on
the pellets are cultured with CCM (medium renew twice per week)
for the following 3–5 weeks (see Subheading 2.2, Notes 5 and 12)
depending on the research question and the history of the
BM-MSC batch (see Note 13). A first positive macroscopic indica-
tion that chondrogenesis is taking place is the increasing size of the
pellet over time (Fig. 3c).
Fig. 3 Representative images of pellet cultures (200,000 BM-MSCs/tube) immediately after centrifugation (a),
after 24 h in the presence of CCM (b), and after 10 days of chondrogenic induction in CCM (c). Arrows indicate
the pellets on the bottom of the tubes
3.3 Evaluation of In this section you will find the general information about our
Chondrogenic standard assays to evaluate chondrogenic differentiation. This sec-
Differentiation tion is not meant to provide all the details necessary to perform the
Capacity of BM-MSC assays, but more to guide you toward the minimum set the analysis
we suggest to perform to verify chondrogenic maturation of the
tissue.
3.3.1 Histological After chondrogenic differentiation we generally fixate the pellet

Analysis cultures in 4% formaldehyde for 24 h (required for our Collagen
type-X antibody), or for a maximum of 60 h. Then the samples are
dehydrated, paraffin-embedded, cut in 6 μm thick sections, col-
lected in microscope glass slides (VWR International BV, cat.: KN
ITVS112751FEA.01), and dried for a minimum of 16 h at 37 C.
We generally perform matrix analysis on the following proteins:
1. Chondrogenesis.
Sulfated glycosaminoglycan by Thionin or Safranin-O,
Immunostaining for Collagen type-II (Developmental Studies
Hybridoma Bank; cat.: II-II6B3)
2. Hypertrophy.
Immunostaining for Collagen type-X (Quartett; cat.:
1_CO097–05; or ThermoFisher, cat,: 14–9771-82).
To verify cartilage stability, we usually perform an in vivo assay
via the subcutaneous implantation of the chondrogenically differ-
entiated pellets in NMRI nu/nu mice for a minimum of 6 weeks,
after which we perform the histological analyses. However, in case
the in vivo samples are expected to have formed calcified tissue
and/or bone, after sample retrieval from the animal we first fix
the samples for up to 1 week (depending on the size of the tissue)
in 4% formaldehyde and then perform a decalcification step either
with 10% EDTA in PBS (pH 7.4) or with formic acid 10% in water
for up to 10 days. Next, we proceed with the histological analysis as
mentioned above.
3.3.2 Transcript Analysis After chondrogenic differentiation we extract the mRNA by first
(See Note 13) mechanically disrupting the pellet with a micro pestle (Eppendorf,
cat.: 0030 120.973) in a 1.5 mL Eppendorf tube containing
350 μL of ice-cold RNA-STAT60 solution (Gentaur, cat:
CS-111-200; see Note 14). At this stage the samples can be stored
at 80 C till the moment of the RNA extraction (see Note 15),
which is done following the manufacturer’s instructions.
We generally perform RT-PCR analysis on the following genes
1. Chondrogenic genes.
COL2A1 and ACAN (optional: COL2A1-IIB, SOX9)
2. Hypertrophic genes.
COL10A1 and ALPL (optional: RUNX2).
We have a panel of several housekeeper/reference genes and
among them the most used are GAPDH, RPS27a, β-ACTIN, 18S,
and HTPR1. We select our reference gene based on its stability
across the conditions and/or time-points, with the intention to
select one stable gene applicable for all the experiments. However,
in case of high variations between the conditions of the same
experiments (generally 1 Ct value) we apply the “best house-
keeper index” (BHI [21]). We run the BHI by using at least
3 housekeeper/reference genes.
4 Notes
If not differentially indicated, all the components need to be stored

at 4 C till expiring date. For all the growth factors that need to be
reconstituted, it is suggested to centrifuge the vial briefly before
opening, to bring the contents to the bottom. Always check data-
sheet of each individual products for any additional
recommendation.
1. Preparation of 500 mL of CEM medium:
(a) 446.5 mL of MEM-α *
(b) 50 mL FCS (10% of the total volume; see Note 2)
(c) 3 mL Fungizone (167; previously aliquoted in 3 mL/
tube and stored in 20 C)
(d) 500 μL Gentamicin (1000)
This preparation can be stored up to 6 weeks at 4 C.
In addition to this, we freshly add to each medium renew
FGF2 and ascorbic acid-2-phosphate, prepared as follows:
(e) FGF2 preparation: we first dissolve the powder in order to
obtain a 5 μg/μL solution in TRIS pH 7.6; we make
aliquots of 5 μL in 1.5 mL Eppendorf tubes and store
them at 20 C. When needed, the 5 μL aliquot are
thawed and 995 μL of CEM (without ascorbic acid-2-

phosphate) is added for a final stock concentration
5 ng/μL. This is to be considered a 5000 stock solution.
These aliquots can be stored at 4 C and used for 14 days.
Never refreeze the dissolved growth factors.
(f) Ascorbic acid-2-phosphate preparation: dissolve the pow-
der in MEM-α (without FCS) to obtain 14.48 mg/mL
solution and store in ready-to-use aliquots at 20 C.
This is to be considered a 500 stock solution. After
thawing, store ascorbic acid-2-phosphate for maximum
24 h at 4 C.
*Between 2010 and 20,111 we moved from DMEM
to MEMα as basic media. This was based on the results of
a comparative study between the two media published in
2009 [22], and on our personal experience
(unpublished data).
2. Preparation of FCS:
(a) Defrost overnight at 4 C.
(b) Heat-inactivate by transferring the FCS at 56 C for
30 min. It is suggested to gently shake/mix the FCS
bottle every 5–8 minduring the incubation at 56 C.
(c) Make aliquot of 50 mL and store at 20 C until use.
It is highly recommended to avoid the use of different
FCS batches for experiments belonging to the same proj-
ect. Different FCS batches may be significantly different in
their composition, resulting in different outcomes in term
of expansion and differentiation capacity of BM-MSC. It
is a standard procedure in our laboratory to screen 3 to
5 different batches of serum from at least 2 different
companies for their capacity to support expansion and
differentiation capacity of BM-MSC. Then, we purchase
the entire selected FCS batch for long-term use (
2 year). The lot number currently used is Lot:
41Q2047K, as indicated in Table 1.
3. WNT3A and CHIR99021 are canonical WNT agonist, the
former acting directly on WNT receptors on the cellular sur-
face, the latter by inhibiting the glycogen synthase kinase
3 (GSK-3) intracellularly.
(a) WNT3a preparation: WNT3A was purified from cell cul-
ture medium conditioned by Drosophila S2 cells modified
with a mouse WNT3A expression vector, using affinity
and gel filtration chromatography as described [23]. A
stock solution of 50 μg/mL in PBS + 1% of 3-[(3-chola-
midopropyl)dimethylammonio]-1-propanesulfonate
(CHAPS) is then prepared and stored at 80 C for up to
6 months. Once thawed, the stock solution can be stared

at 4 C for a maximum of 24 h. Note that WNT3A is heat
sensitive and in our experiment we use to renew the
medium containing WNT3A every day. WNT3A is also
commercially available from different companies, and
although we never performed comparative studies, we
are now successfully using WNT3A from R&D Systems
and Sigma-Aldrich for culturing other cell types
(unpublished data).
(b) CHIR99021 preparation: A stock solution of 5 or 10 mM
is prepared by adding dimethyl sulfoxide (DMSO) at
45 to 55 C under gentle shaking. When dissolved, ali-
quots are prepared according to the experimental needs.
This stock solution can be stored for 1 month at 20 C.
Once thawed, the stock solution can be stared at 4 C for a
maximum of 24 h. We found that the effect of
CHIR99021 is comparable to the effect of WNT3A on
cell proliferation and subsequent chondrogenic differenti-
ation. However, stimulation with CHIR99021 for more
than 2 weeks results in changing in cellular morphology
with negative effect on expansion and differentiation
capacity [24]. We therefore suggest to use CHIR99021
as WNT3A substitute only for short-term experiments
(10 days).
4. Preparation of 500 mL of CCM medium.
(a) 485.5 mL DMEM-HG
(b) 5 mL ITS+ (100)
(c) 1 mL of Proline of 20 mg/mL in DMEM-HG (500;
store at 4 C for 6 months)
(d) 5 mL of Sodium pyruvate (100)
(e) 3 mL Fungizone (167; previously aliquoted in 3 mL/
tube and stored in 20 C)
(f) 500 μL Gentamicin (1000)
This preparation can be stored up to 6 weeks at 4 C.
In addition to this, we freshly add to medium TGF-β1*,
ascorbic acid-2-phosphate and dexamethasone, prepared
as follows:
(g) TGF-β1 preparation: we first dissolve the powder in order
to obtain a 2 μg/mL solution in 4 mM HCL +0.1%
bovine serum albumin (BSA); we make aliquots of
500 μL in 1.5 mL Eppendorf tubes and store them at
20 C or 80 C up to 1 year. This is to be considered a
200 stock solution. Once thawing can be stored at 4 C
and used for 14 days. Never refreeze the dissolved growth
factors.
(h) Ascorbic-acid preparation: dissolve the powder in

DMEM-HG (without FCS) and then follow the same
indication as described in Note 1.
(i) Dexamethasone preparation: we prepare a 1 mM stock
solution by adding 10 mL of 100% ethanol to 3.92 g of
dexamethasone powder. This is a 10.000 solution. We
store in aliquots of around 1 mL at 20 C for up to
1 year.
*We use TGF-β1 for our chondrogenic cultures.
However, TGF-β3 and sometimes TGF-β2 are used
while maintaining the rest of the protocol unchanged.
We and others compared the effect of different TGF-β
isoforms on chondrogenic differentiation and TGF-β1 or
TGF-β3 has been suggested to be the most potent [25–
27], with differences mainly related to the readout para-
meters taken in account for the analysis and, possibly, to
the bioactivity of the recombinant protein purchased from
different companies.
5. N-(6-Methyl-2-benzothiazolyl)-2-[(3, 4, 6, 7-tetrahydro-4-
oxo-3-phenylthieno[3, 2-dpyrimidin-2-yl)thio]-acetamide
(IWP2) is a small molecule that specifically inhibits the matu-
ration (endogenous production) of WNT proteins by blocking
the Porcupine-mediated WNT palmitoylation. IWP2 is
prepared as 2 mM stock solution (1000) by adding
1.072 mL of dimethyl sulfoxide (DMSO) in 1 mg of IWP2
powder. We store aliquots of 50–100 μL at 80 C for up to
2 years. During chondrogenic differentiation experiments,
medium containing IWP2 is renewed every other day, and is
added starting from day 10 to day 14 of chondrogenic induc-
tion [18]. Earlier administration of CCM+ media to the pellets,
in our hands, reduced chondrogenic differentiation
(unpublished data).
6. We use between 0.11 (T175 flasks) and 0.16 (T75) mL/cm2 of
culture medium for our experiments. However, for each read-
out parameter (e.g., expansion, RT-PCR, Western blot), we
design our experiments in order to avoid direct comparison
between cells cultured with different amount of expansion
media per cm2.
7. We suggest to retain a minimum of 50% of the total volume as
CEM. Given that, during the first seeding, all the components
of the CEM should be adjusted in order to have all the com-
ponents of the CEM at the right final concentration, despite
the presence of the bone marrow aspirate.
8. Do not allow the cells to reach confluence during expansion.
This, in our hands, negatively influences their chondrogenic
differentiation capacity (unpublished data).
9. Unless specifically requested by the research question, we gen-

erally use P5 BM-MSC for our experiments.
10. Although a specific set of surface markers have been proposed
in 2006 to identify BM-MSCs [7], this list is continuously
updated and there is no general consensus on the minimum
requirements a BM-MSC needs to have in terms of surface
marker expression. Moreover, we recently showed that some
of the originally selected markers, such as CD105, are not
associated with the chondrogenic differentiation capacity of
BM-MSC [8].
11. The use of polypropylene tubes is necessary to prevent cell
attachment to the side of the tube before and during centrifu-
gation. Always remember to leave the lid partially open to allow
air-gas exchange. It is also possible to generate 100,000 cell/
pellet. In our hands, chondrogenic differentiation progresses
similarly compared to the 200,000 cell/pellets; however, the
reduced amount of cells in the 100,000 cell pellets can be a
limiting factor to perform subsequent analysis (e.g., RT-PCR
or Western Blot).
12. Especially in the first 2 weeks of culturing, the pellets have the
tendency to attach to the bottom of the tube. It is therefore
suggested to gently shake/tap the tube during the medium
renewal, in order to get the pellet detached. This will allow a
more homogeneous diffusion of the media components and
chondrogenic differentiation.
13. Although our standard chondrogenic differentiation protocol
consists in 5 weeks of differentiation [9], when we use
BM-MSC from a previously tested donor, or the research
question we are investigating requires, for example, to investi-
gate early chondrogenic events, we apply a 3- or 4-week differ-
entiation protocol, which has the sole difference of being 2 or
1 week shorter, with no difference in media composition (see,
for example, Cleary et al. [28]).
14. Until the end of 2019 we used RNA-Bee (Gentaur, cat.:
CS-105B) for the extraction of RNA from pellet cultures and
other soft tissues. Due to the fact that RNA-Bee was no longer
available and after a series of comparative validation tests
regarding extraction efficiency and purity among different
extraction protocols, we have moved to RNA-STAT60.
15. By convention, and unless differently indicated, we harvest the
samples for RT-PCR analysis 24 1 h after the last medium
renewal.
Acknowledgements
This work is part of Medical Delta RegMed4 program.
References
1. Tavassoli M, Crosby WH (1968) Transplanta- their use in cell therapy. Exp Hematol
tion of marrow to extramedullary sites. Science 28:707–715
161:54–56 13. Gronthos S, Zannettino AC, Hay SJ, Shi S,
2. Friedenstein AJ, Petrakova KV, Kurolesova AI, Graves SE, Kortesidis A et al (2003) Molecular
Frolova GP (1968) Heterotopic of bone mar- and cellular characterisation of highly purified
row. Analysis of precursor cells for osteogenic stromal stem cells derived from human bone
and hematopoietic tissues. Transplantation marrow. J Cell Sci 116:1827–1835
6:230–247 14. Sacchetti B, Funari A, Michienzi S, Di
3. Pittenger MF, Discher DE, Peault BM, Phin- Cesare S, Piersanti S, Saggio I et al (2007)
ney DG, Hare JM, Caplan AI (2019) Mesen- Self-renewing osteoprogenitors in bone mar-
chymal stem cell perspective: cell biology to row sinusoids can organize a hematopoietic
clinical progress. NPJ Regen Med 4:22 microenvironment. Cell 131:324–336
4. Owen M, Friedenstein AJ (1988) Stromal stem 15. Digirolamo CM, Stokes D, Colter D, Phinney
cells: marrow-derived osteogenic precursors. DG, Class R, Prockop DJ (1999) Propagation
Ciba Found Symp 136:42–60 and senescence of human marrow stromal cells
5. Caplan AI (1991) Mesenchymal stem cells. J in culture: a simple colony-forming assay iden-
Orthop Res 9:641–650 tifies samples with the greatest potential to
6. Bianco P, Gehron RP (2000) Marrow stromal propagate and differentiate. Br J Haematol
stem cells. J Clin Invest 105:1663–1668 107:275–281
7. Dominici M, Le Blanc K, Mueller I, Slaper- 16. Martin I, Muraglia A, Campanile G,
Cortenbach I, Marini F, Krause D et al Cancedda R, Quarto R (1997) Fibroblast
(2006) Minimal criteria for defining multipo- growth factor-2 supports ex vivo expansion
tent mesenchymal stromal cells. The Interna- and maintenance of osteogenic precursors
tional Society for Cellular Therapy position from human bone marrow. Endocrinology
statement. Cytotherapy 8:315–317 138:4456–4462
8. Cleary MA, Narcisi R, Focke K, van der 17. Sivasubramaniyan K, Ilas DC, Harichandan A,
Linden R, Brama PA, van Osch GJ (2016) Bos PK, Santos DL, de Zwart P et al (2018)
Expression of CD105 on expanded mesenchy- Bone marrow-harvesting technique influences
mal stem cells does not predict their chondro- functional heterogeneity of mesenchymal
genic potential. Osteoarthr Cartil 24:868–872 stem/stromal cells and cartilage regeneration.
Am J Sports Med 46:3521–3531
9. Johnstone B, Hering TM, Caplan AI, Gold-
berg VM, Yoo JU (1998) In vitro chondrogen- 18. Narcisi R, Cleary MA, Brama PA, Hoogduijn
esis of bone marrow-derived mesenchymal MJ, Tuysuz N, ten Berge D et al (2015) Long-
progenitor cells. Exp Cell Res 238:265–272 term expansion, enhanced chondrogenic
potential, and suppression of endochondral
10. Pittenger MF, Mackay AM, Beck SC, Jaiswal ossification of adult human MSCs via WNT
RK, Douglas R, Mosca JD et al (1999) Multi- signaling modulation. Stem Cell Rep
lineage potential of adult human mesenchymal 4:459–472
stem cells. Science 284:143–147
19. Gruber HE, Somayaji S, Riley F, Hoelscher
11. Muraglia A, Cancedda R, Quarto R (2000) GL, Norton HJ, Ingram J et al (2012)
Clonal mesenchymal progenitors from human Human adipose-derived mesenchymal stem
bone marrow differentiate in vitro according to cells: serial passaging, doubling time and cell
a hierarchical model. J Cell Sci 113 senescence. Biotech Histochem 87:303–311
(Pt 7):1161–1166
20. Banfi A, Bianchi G, Notaro R, Luzzatto L,
12. Banfi A, Muraglia A, Dozin B, Cancedda R, Quarto R (2002) Replicative
Mastrogiacomo M, Cancedda R, Quarto R aging and gene expression in long-term cul-
(2000) Proliferation kinetics and differentia- tures of human bone marrow stromal cells.
tion potential of ex vivo expanded human Tissue Eng 8:901–910
bone marrow stromal cells: implications for
21. Pfaffl MW, Tichopad A, Prgomet C, Neuvians Hypertrophy in mesenchymal stem cell chon-
TP (2004) Determination of stable drogenesis: effect of TGF-beta isoforms and
housekeeping genes, differentially regulated chondrogenic conditioning. Cells Tissues
target genes and sample integrity: Best- Organs 192:158–166
Keeper—Excel-based tool using pair-wise cor- 26. Cals FL, Hellingman CA, Koevoet W, Baaten-
relations. Biotechnol Lett 26:509–515 burg de Jong RJ, van Osch GJ (2012) Effects
22. Chen HH, Decot V, Ouyang JP, Stoltz JF, of transforming growth factor-beta subtypes
Bensoussan D, de Isla NG (2009) In vitro on in vitro cartilage production and minerali-
initial expansion of mesenchymal stem cells is zation of human bone marrow stromal-derived
influenced by the culture parameters used in mesenchymal stem cells. J Tissue Eng Regen
the isolation process. Biomed Mater Eng Med 6:68–76
19:301–309 27. Barry F, Boynton RE, Liu B, Murphy JM
23. Willert K, Brown JD, Danenberg E, Duncan (2001) Chondrogenic differentiation of mes-
AW, Weissman IL, Reya T et al (2003) Wnt enchymal stem cells from bone marrow:
proteins are lipid-modified and can act as stem differentiation-dependent gene expression of
cell growth factors. Nature 423:448–452 matrix components. Exp Cell Res
24. Narcisi R, Arikan OH, Lehmann J, Ten 268:189–200
Berge D, van Osch GJ (2016) Differential 28. Cleary MA, Narcisi R, Albiero A, Jenner F, de
effects of small molecule WNT agonists on Kroon LMG, Koevoet W et al (2017) Dynamic
the multilineage differentiation capacity of regulation of TWIST1 expression during chon-
human mesenchymal stem cells. Tissue Eng drogenic differentiation of human bone
Part A 22:1264–1273 marrow-derived mesenchymal stem cells.
25. Mueller MB, Fischer M, Zellner J, Berner A, Stem Cells Dev 26:751–761
Dienstknecht T, Prantl L et al (2010)
Chapter 3
A Novel Enzymatic Digestion Approach for Isolation

and Culture of Rodent Bone Marrow Mesenchymal
Progenitors
Leilei Zhong, Lutian Yao, and Ling Qin
Abstract
Bone marrow mesenchymal stem cells (MSCs) are promising therapeutic tools for tissue repair and
treatment of a number of human diseases. As a result, there is substantial interest in characterizing and
expanding these cells to uncover their therapeutic potential. Bone marrow mesenchymal progenitors,
containing both MSCs and their proliferative progeny, are commonly isolated from the central region of
rodent long bones. However, challenges exist in expanding these central mesenchymal progenitors in
culture. We have designed an enzymatic digestion protocol to isolate mesenchymal progenitors within
rodent long bones that resides close to the bone surface, which we termed endosteal mesenchymal
progenitors. These cells are more metabolically active and more responsive to external stimuli compared
to central mesenchymal progenitors. Therefore, they represent a biologically important target for MSC
research. This chapter describes the approach in detail how to isolate and culture endosteal mesenchymal
progenitors as well as their central counterparts from rodent long bones.
Key words Mesenchymal stem cells, Endosteal mesenchymal progenitors, Bone marrow, Enzymatic
digestion, Colony forming unit-fibroblast
1 Introduction
Almost a half century ago, Alexander Friedenstein and colleagues

pioneered a flushing method to isolate bone marrow cells from the
central region of rodent long bones for culturing plastic-adherent
and clonogenic fibroblastoid mesenchymal progenitors
[1, 2]. Since then, the flushing method has become a standard
technique to isolate mesenchymal progenitors from rodents in
laboratories. Mesenchymal progenitor cultures are heterogeneous
and consist of mesenchymal stem cells (MSCs) and their prolifera-
tive and more differentiated offspring [3]. In addition to their
multi-lineage differentiation ability, these cells are immunosuppres-
sive and capable of homing to injured tissues and secreting a
number of bioactive molecules that promote wound repair and
29
30 Leilei Zhong et al.
tissue regeneration. Therefore, they have been intensively investi-

gated as potential therapeutic tools for tissue repair and for treat-
ment of a number of diseases including Crohn’s disease and graft-
versus-host disease [4]. However, these central mesenchymal pro-
genitors are anatomically distant from trabecular and cortical bone
surfaces where constant replenishment of bone forming osteoblasts
by mesenchymal progenitors is required. There are also challenges
associated with the rapid expansion of these cells in culture. For
example, central mesenchymal progenitor cultures, especially those
from mice, are difficult to grow in vitro and have limited prolifera-
tive ability [5–8].
The endosteal bone marrow is the portion of the bone marrow
that is close to the bone surface, including cells that are within the
trabecular bone and close to the endocortical bone surface [9]. We
previously demonstrated that, in rodents, endosteal bone marrow
cells contain a much higher frequency of mesenchymal progenitors
than central bone marrow cells [8]. These endosteal mesenchymal
progenitors have similar cell surface marker expression
(Sca-1+CD105+CD29+CD73+ CD71+CD44+ CD45 CD34 )
and multi-lineage differentiation ability to central progenitors.
However, they form much larger colony forming unit-fibroblast
(CFU-F) colonies due to their higher proliferative ability and can
be passaged more times in culture than their central counterparts.
They also exhibit greater immunosuppressive activity both in vitro
and in a mouse model of inflammatory bowel disease. Moreover,
aging, a major contributing factor for osteoporosis, dramatically
decreases their number, while injection of parathyroid hormone, an
anabolic treatment for osteoporosis, strongly increases their num-
ber. Hence, these endosteal mesenchymal progenitors are more
biologically relevant to skeletal homeostasis and disease than central
mesenchymal progenitors [8]. Particularly, for in vitro culturing
and expansion, endosteal mesenchymal progenitors are more suit-
able than central mesenchymal progenitors.
This chapter describes an enzymatic digestion method to iso-
late endosteal mesenchymal progenitors from rat and mouse long
bones, along with the concomitant isolation of their central coun-
terparts from the same bones. Methods to quantify, culture, and
differentiate these cells are also presented.
2 Materials
2.1 Animals This protocol was developed from experiments performed with
Sprague-Dawley rats and C57Bl/6 mice (see Note 1). It is also
compatible with other strains of mice we have tested, such as 129.
2.2 Instruments 1. Class II biological safety cabinet/cell culture hood and a hori-
zontal laminar flow clean bench: both should be equipped with
a UV light for decontamination.
A Digestion Method for Isolating Bone Marrow Mesenchymal Progenitors 31
2. Tissue culture incubator with temperature and gas composi-

tion controls.
3. Mini shaker that can be placed inside a tissue culture incubator.
4. Inverted microscope with phase-contrast ability.
5. Benchtop centrifuge with a swing-bucket rotor.
6. Sterile surgical scissors, surgical forceps, scalpel handles, and
scalpel blades (#22).
7. Sterile 0.2 μm syringe filter.
8. Sterile 70 μm cell strainer.
9. Sterile 10 mL syringes with 25- and 27-gauge needles.
10. Sterile 15 and 50 mL polypropylene conical tubes.
11. Pipet-aid, sterile serological pipettes (5 and 10 mL) and
gibson-type micropipettes and tips (20, 200, and 1000 μL).
12. 25 cm2 tissue culture flasks with vented seal caps, 100 mm
tissue culture dishes, 100 mm Petri dishes.
13. Hemocytometer.
2.3 Reagents 1. 70% Ethanol.

and Media 2. Dulbecco’s phosphate-buffered saline (PBS).
3. Flushing medium: αMEM supplemented with 1% fetal bovine
serum (FBS), 100 IU/mL penicillin, and 100 μg/mL
streptomycin.
4. Protease solution: 2 mg/mL collagenase A (Roche Diagnos-
tics, Indianapolis, IN) and 2.5 mg/mL trypsin dissolved in
Dulbecco’s PBS and filter sterilized using a syringe filter. This
solution should be freshly prepared just before starting the
isolation process.
5. Growth medium for rat mesenchymal progenitors: αMEM
supplemented with 15% FBS, 100 IU/mL penicillin, and
100 μg/mL streptomycin.
6. Growth medium for mouse mesenchymal progenitors: αMEM
supplemented with 15% FBS, 0.1% β-mercaptoethanol, 20 mM
glutamine, 100 IU/mL penicillin, and 100 μg/mL
streptomycin.
7. Osteogenic medium: αMEM containing 10% FBS, 10 nM
dexamethasone, 10 mM β-glycerophosphate, and 50 μg/mL
L-ascorbic acid (AA, see Note 2), 100 IU/mL penicillin, and
100 μg/mL streptomycin.
8. Adipogenic medium: αMEM with 10% FBS, 0.5 mM isobutyl-
methylxanthine, 10 mM indomethacin, 1 μM dexamethasone,
and 10 μg/mL insulin, 100 IU/mL penicillin, and 100 μg/mL
streptomycin.
9. Chondrogenic medium: High glucose DMEM, 0.1 μM dexa-

methasone, 50 μg/mL AA, 40 μg/mL L-proline, 100 μg/mL
sodium pyruvate, 1 ITS+, and 10 ng/mL TGFβ3, 100 IU/
mL penicillin, and 100 μg/mL streptomycin.
10. 3% Acetic acid with methylene blue.
11. 3% Crystal violet in methanol.
12. 0.05% Trypsin/EDTA solution.
13. 0.4% Trypan blue solution.
3 Methods
3.1 Harvest 1. Euthanize the animal by CO2 inhalation.

of Rodent Hind Long 2. Immediately transfer the dead animal to a clean bench
Bones pre-decontaminated by UV radiation. Place the animal on a
flat surface on its back and wet the pelt thoroughly with 70%
ethanol.
3. Using sterile forceps and scissors, incise and peel back the skin
surrounding the hind long bones. Remove the bilateral hind
long bones by cutting through the hip and ankle joints using
sharp surgical scissors. Cut through the knee joint to separate
the tibia and femur (see Note 3).
4. Place the long bones in a 100 mm Petri dish filled with 10 mL
of flushing medium. Transfer the petri dish with the bones to a
tissue culture hood for bone marrow harvesting.
3.2 Isolation 1. Under a sterile tissue culture hood, use forceps and a scalpel to
of Central remove all of the soft tissue surrounding the bones. After
and Endosteal Bone removal of the soft tissue, place the long bones into a new
marrow Cells 100 mm Petri dish filled with 10 mL of flushing medium.
2. Cut off both ends of the tibia and femur at the growth plate
with a scalpel.
3. Fill a syringe with 5 (mouse) or 10 (rat) mL of flushing medium
per animal (2 tibiae and 2 femurs). Attach a 25- (rat) or
27-gauge needle (mouse) to the syringe.
4. With the prefilled syringe, drill a hole at each end of the bone
and press down the plunger at one end to force medium
through the bone. This will flush the bone marrow out of the
bone through the opposite end of the bone.
5. Reverse the bone and repeat the flushing from the other end of
the bone. Use 1 (mouse) or 2 (rat) mL of flushing medium to
flush out each bone. Bone marrow cells released by flushing
mainly come from the central part of the diaphyseal shaft, and
hence are central bone marrow cells (Fig. 1 step 1, see Note 4).
The bone marrow cells that are located in close proximity to the
Fig. 1 Representative images of a rat femur during the isolation of endosteal bone marrow. Step 1: flush out
central bone marrow cells from a bone that is free of its surrounding soft tissue and with both ends removed at
the growth plates; Step 2: predigest the whole bone to remove the periosteal progenitors and then
longitudinally cut the bones into two halves; Step 3: gently wash the bones to remove loosely attached
bone marrow; Step 4: digest bone fragments to collect endosteal bone marrow cells. BM: bone marrow.
Reprinted from Bone, 53(2), Siclari et al., Mesenchymal progenitors residing close to the bone surface are
functionally distinct from those in the central bone marrow. 575–86, Copyright (2013), with permission from
Elsevier
endosteum remain attached to the bone after the flushing and

are only released by enzymatic digestion (For information
about how to culture central bone marrow mesenchymal pro-
genitors, see Note 5).
6. After removing the central bone marrow cells, scrape the out-
side surface of the bones a few times with a scalpel blade and
then place the bones into a 15 mL tube containing 5 mL of
protease solution (8 mouse bones or 4 rat bones per tube).
7. Place the tubes on a mini shaker in a tissue culture incubator
and shake for 20 minutes at 37 C. This step removes perios-
teum and its associated periosteal progenitors from the long
bones (Fig. 1 step 2, see Note 6).
8. After digestion, wash the bones with flushing medium twice,
and then longitudinally cut the bones into two halves (Fig. 1
step 2, see Note 7).
9. Use a syringe to gently wash the inside of the bones with
flushing medium to remove loosely attached bone marrow
(Fig. 1 step 3, see Note 8).
10. Place the bone fragments into a 15 mL tube containing 5 mL
of protease solution (8 mouse bones or 4 rat bones per tube)
and perform the second digestion step for 60 minutes as
described in Subheading 3.2, step 7 (Fig. 1 step 4).
11. Add 5 mL of growth medium to neutralize the protease solu-
tion, and then collect the supernatant. Wash bone fragments
twice with growth medium to collect all remaining cells.
12. Pass the cells through a 70 μm cell strainer to remove debris (see
Note 9).
13. Perform a cell count by diluting a cell aliquot 1:10 in 3% acetic

acid with methylene blue to lyse the red blood cells.
14. Count the nucleated cells using a hemocytometer under a
microscope. The expected cell recovery of endosteal bone
marrow is 8–12 106 cells per mouse and 12–20 106 cells
per rat.
3.3 CFU-F Assays 1. Seed 1 106 mononuclear endosteal bone marrow cells per
of Endosteal 25 cm2 flask in the growth medium and incubate the culture at
Mesenchymal 37 C in 5% CO2 in a humidified tissue culture incubator (see
Progenitors Note 10).
2. Typically, after 5 (rat) or 7 (mouse) days of incubation, most
colonies should contain more than 50 fibroblastic cells (Fig. 2a,
b, see Note 11). At this point, remove medium and wash flasks
twice with PBS.
3. Stain the colonies with 3% crystal violet in methanol for at least
1 hour at RT.
4. Rinse flasks thoroughly with tap water to remove unbound
stain.
5. Air-dry the flasks completely.
6. Count the number of CFU-F colonies under an inverted micro-
scope with a 4 objective. We recommend drawing lines on the
bottom of the flask to divide the surface into 8 regions in order
to facilitate counting. Only count colonies larger than 50 cells.
The expected CFU-F frequency of endosteal bone marrow cells
is about 80–150 CFU-Fs per 1 106 mononuclear cells from
both mouse and rat (Fig. 2c, d). The size of CFU-F colonies is
normally larger in endosteal bone marrow compared to central
bone marrow (Fig. 2c, d, see Note 12).
3.4 Culture 1. To culture endosteal mesenchymal progenitors, seed 5 106

and Differentiation mononuclear endosteal bone marrow cells per 100 mm tissue
of Endosteal culture dishes in the growth medium and incubate at 37 C in
Mesenchymal 5% CO2 in a humidified tissue culture incubator.
Progenitors 2. Change medium every 2–3 days.
3. When the cells reach 80–90% confluence or when individual
CFU-F colonies have expanded so that they are in close prox-
imity to each other (about 8–10 days after plating), cells should
be passaged for expansion. Aspirate the medium and wash the
cells with PBS.
4. Add 3 mL of 0.05% trypsin/EDTA to the cells and incubate for
2–3 min in the tissue culture incubator. Examine under the
microscope to confirm that about 70–90% of the cells are
detached from the plate. If not, return the plate to the incuba-
tor for another 2 min. Typically, endosteal mesenchymal pro-
genitors require less digestion time compared to central cells.
Fig. 2 CFU-F assays of rodent endosteal bone marrow cells compared to their central counterparts. (a)
Representative images of 25 cm2 flasks with mouse central and endosteal CFU-F colonies after staining. Note
that the initial seeding densities are 3 106 and 1 106 cells per flask for central and endosteal bone
marrow, respectively. (b) Representative images and morphologies of mouse CFU-F colonies at low (top) and
high (bottom) magnification. (c, d) Quantification of CFU-F frequency and diameter of endosteal bone marrow
cells from mouse (c) and rat (d). **: p < 0.01 vs. central
5. Neutralize the trypsin by adding 3 mL of growth medium and

gently pipet up and down with a 10 mL serological pipette to
obtain a single cell suspension.
6. Transfer the cells into a 15 mL tube and centrifuge at 300 g
for 5 min at room temperature.
7. Resuspend the pellet in 3 mL of growth medium and count the
number of cells. To count the number of live mesenchymal
progenitors, dilute an aliquot of the cells in trypan blue solu-
tion to quantify live and dead cells using a hemocytometer and
a light microscope.
8. Plate the cells at a density of 0.5 106/100 mm dish. The cells
that grow up are passage 1 cells.
9. Change medium every 2–3 days. Normally cells reach 80–90%
confluency within 5–6 days. Lift and split cells at a ratio of
1:5–1:3 for expansion (see Note 13). The cells can also be
stored in liquid nitrogen from passage 2–3 for future use.
Fig. 3 Endosteal mesenchymal progenitors are capable of multi-lineage differentiation. (a) In vitro osteogenic
differentiation as detected by von Kossa staining. (b) In vitro adipogenic differentiation as detected by oil red O
staining. (c) In vitro chondrogenic pellet differentiation as detected by alcian blue staining
10. To differentiate into osteoblast or adipocyte lineages (Fig. 3a,

b), endosteal mesenchymal progenitors are first expanded to
confluence in the growth medium and then switched to osteo-
genic or adipogenic medium for 3 and 1 weeks, respectively.
The differentiation media should be changed every 2–3 days.
11. To differentiate into chondrocyte pellet (Fig. 3c), endosteal
mesenchymal progenitors are resuspended in chondrogenic
medium at 1 106 cells/mL and seeded in a V-bottomed
96-well plate at 200 μL/well. Centrifuge the plate at 300 g
for 5 min to form a cell pellet, which appears as a fuzzy cloud
on the first day. Change media every 3 days for 3 weeks.
4 Notes
1. With this protocol, we have successfully isolated and cultured

endosteal mesenchymal progenitors from 1- to 4-month-old
Sprague-Dawley rats and 1- to 16-month-old C57Bl/6 mice.
The number of endosteal mesenchymal progenitors decreases
significantly with aging [8]. Therefore, young animals (1–2-
month-old) are the best source for obtaining these progenitors.
In addition, the bones of young animals are much easier to cut
while those from old animals tend to shatter and require more
force during cutting (see Note 7).
2. Since AA is unstable and rapidly oxidizes in water, it should be
freshly added to the medium from a frozen stock (50 mg/mL)
just before a medium change.
3. Long bones should be harvested under sterile conditions. If a

sterile bench is not available, long bones can be dissected in a
tissue culture hood. Frequently dip the forceps and scissors in
70% ethanol to prevent contamination into the cultures. All
procedures should be performed as quickly as possible to
achieve a high yield of viable mesenchymal progenitors.
4. Be sure to hold the bone tightly with a pair of forceps during
flushing to avoid dropping the bone. It is recommended to use
a syringe with a screw-tip to prevent the needle from detaching
from the syringe during flushing.
5. To culture central mesenchymal progenitors, flush the central
bone marrow cells into a 50 mL conical tube. Gently pipet the
cell suspension up and down to break up the clumps of bone
marrow. Centrifuge the cells at 300 g for 5 min and resus-
pend the pellet in 5 mL of growth medium. Filter the cell
suspension through a cell strainer and count the number of
cells in the same way as described in Subheading 3.2, step 13.
The expected yield of central bone marrow is 30–50 106 cells
per mouse and 120–200 106 cells per rat. Seed 3 106 cells
per 25 cm2 flask for CFU-F assays and 30–50 106 per
100 mm dish for expansion. CFU-F staining and counting
are performed 7 (rat) and 10 (mouse) days later as described
in Subheading 3.3. The expected frequency is about
20–50 CFU-Fs/1 106 mononuclear cells (Fig. 2c, d). Cen-
tral mesenchymal progenitors are cultured in the same growth
medium as endosteal mesenchymal progenitors but they grow
much slower. Split central mesenchymal progenitors at 1:2 or
1:3 when passaging.
6. It is important to remove all soft tissue, especially the perios-
teum, surrounding the long bones to avoid the contamination
of mesenchymal progenitors from undesired sources. Perios-
teum contains periosteal progenitors that have similar charac-
teristics to bone marrow mesenchymal progenitors
[10]. Several previous studies also used collagenase digestion
of flushed, minced, or chopped bone fragments to increase the
yield of bone marrow mesenchymal progenitors [11–
16]. However, they were likely to have contamination of peri-
osteal progenitors because they did not remove the periosteum
from the bone. We have demonstrated by histology that scrap-
ing and predigestion of the bones is sufficient to remove peri-
osteal cells, and therefore prevent contamination of periosteal
progenitors into the endosteal bone marrow [8].
7. Bones from old animals easily shatter during cutting. To mini-
mize the amount of shattering, while holding the bone tightly
with a pair of forceps, use a scalpel blade to first mark a longi-
tudinal line on the outside of the bone surface. Then, slowly
and forcefully cut along this line.
8. You can choose to harvest endosteal bone marrow cells from

diaphyseal and metaphyseal regions separately. To do so, use a
blade to cut the longitudinally halved bones at the junctions
between the metaphysis and diaphysis. Then, enzymatically
digest the diaphyseal and metaphyseal bone fragments sepa-
rately. We have found that the majority of endosteal mesenchy-
mal progenitors reside in the metaphysis [8].
9. Endosteal bone marrow cells have a tendency to clump. Always
mix cell suspension well before counting and seeding. If you
intend to use these cells directly for flow cytometry, pass the
cells through a cell strainer immediately before flow analysis.
10. We prefer using 25 cm2 flasks over 6-well plates for CFU-F
assays. Due to the concaved surface of the wells in a 6-well
plate, we have found that the colonies tend to grow more at the
center of the wells and become difficult to count individually.
We have found the colonies to be more evenly distributed and
easier to count in 25 cm2 flasks.
11. Do not change the medium of cells plated for a CFU-F assay.
Overall, minimize the amount of disturbance to the flasks after
plating. If possible, allow the cells to remain undisturbed in the
tissue culture incubator till counting. This ensures the optimal
accuracy of the CFU-F assay.
12. The adherence and proliferative ability of mesenchymal pro-
genitors varies significantly depending on culture conditions. It
is recommended to test different batches of FBS in the growth
medium to select one that gives the greatest number of
CFU-Fs and optimal colony morphology and to use this one
batch through the entire project.
13. To maintain a healthy cell population, it is advisable to passage
cells at 80–90% confluence and to avoid over-confluence. End-
osteal mesenchymal progenitors grow much better and have a
much shorter doubling time in culture than the commonly
used central progenitors. While mouse central mesenchymal
progenitors normally reach senescence and stop growth at
5–10 passages, we found that mouse endosteal mesenchymal
progenitor cultures keep proliferating beyond 20 passages [8].
Acknowledgments
This work was supported by NIH grants K01DK071988 and

R01DK095803 (to LQ). This study was supported by the Penn
Center for Musculoskeletal Disorders Histology Core
(P30-AR069619).
References
1. Friedenstein AJ, Chailakhjan RK, Lalykina KS 9. Ellis SL, Grassinger J, Jones A et al (2011) The
(1970) The development of fibroblast colonies relationship between bone, hemopoietic stem
in monolayer cultures of Guinea-pig bone mar- cells, and vasculature. Blood 118:1516–1524
row and spleen cells. Cell Tissue Kinet 10. van Gastel N, Torrekens S, Roberts SJ et al
3:393–403 (2012) Engineering vascularized bone: osteo-
2. Friedenstein AJ, Gorskaja JF, Kulagina NN genic and proangiogenic potential of murine
(1976) Fibroblast precursors in normal and periosteal cells. Stem Cells 30:2460–2471
irradiated mouse hematopoietic organs. Exp 11. Morikawa S, Mabuchi Y, Kubota Y et al (2009)
Hematol 4:267–274 Prospective identification, isolation, and sys-
3. Lindner U, Kramer J, Rohwedel J et al (2010) temic transplantation of multipotent mesen-
Mesenchymal stem or stromal cells: toward a chymal stem cells in murine bone marrow. J
better understanding of their biology? Transfus Exp Med 206:2483–2496
Med Hemother 37:75–83 12. Xu S, De Becker A, Van Camp B et al (2010)
4. Wang S, Qu X, Zhao RC (2012) Clinical appli- An improved harvest and in vitro expansion
cations of mesenchymal stem cells. J Hematol protocol for murine bone marrow-derived
Oncol 5:19.5 mesenchymal stem cells. J Biomed Biotechnol
5. Short B, Brouard N, Occhiodoro-Scott T et al 2010:105940
(2003) Mesenchymal stem cells. Arch Med Res 13. Nakamura Y, Arai F, Iwasaki H et al (2010)
34:565–571 Isolation and characterization of endosteal
6. Krishnappa V, Boregowda SV, Phinney DG niche cell populations that regulate hemato-
(2013) The peculiar biology of mouse mesen- poietic stem cells. Blood 116:1422–1432
chymal stromal cells--oxygen is the key. 14. Ohishi M, Ono W, Ono N et al (2012) A novel
Cytotherapy 15:536–541 population of cells expressing both hemato-
7. Banfi A, Muraglia A, Dozin B et al (2000) poietic and mesenchymal markers is present in
Proliferation kinetics and differentiation poten- the normal adult bone marrow and is aug-
tial of ex vivo expanded human bone marrow mented in a murine model of marrow fibrosis.
stromal cells: implications for their use in cell Am J Pathol 180:811–818
therapy. Exp Hematol 28:707–715 15. Short BJ, Brouard N, Simmons PJ (2009) Pro-
8. Siclari VA, Zhu J, Akiyama K et al (2013) Mes- spective isolation of mesenchymal stem cells
enchymal progenitors residing close to the from mouse compact bone. Methods Mol
bone surface are functionally distinct from Biol 482:259–268
those in the central bone marrow. Bone 16. Zhu H, Guo ZK, Jiang XX et al (2010) A
53:575–586 protocol for isolation and culture of mesenchy-
mal stem cells from mouse compact bone. Nat
Protoc 5:550–560
Chapter 4
Isolation of Nucleus Pulposus and Annulus Fibrosus Cells

from the Intervertebral Disc
Guus G. H. van den Akker, Andy Cremers, Donatus A. M. Surtel,
Willem Voncken, and Tim J. M. Welting
Abstract
Cells isolated from the intervertebral disc are often used for in vitro experimentation. Correctly separating
the intervertebral disc tissue in annulus fibrosus and nucleus pulposus is particularly challenging when
working with surplus material from surgery or specimens from donors with an advanced age. Moreover,
lineage controls are only sparsely reported to verify tissue of origin. Here we describe an approach to
intervertebral disc cell isolation from human and bovine origin.
Key words Intervertebral disc, Cell isolation, Nucleus pulposus, Annulus fibrosus, Collagenase, Cell
characterization
1 Introduction
The intervertebral disc (IVD), located between vertebra, is an

avascular, non-innervated tissue of musculoskeletal origin that
allows flexion and rotation of the spinal column. The IVD com-
prises a central cartilaginous nucleus pulposus (NP) surrounded by
the fibrous annulus fibrosus (AF) with highly oriented collagen
fibers. At the rostral and caudal sides, the IVD is enclosed by the
end plates that connect to the vertebral body. End plate cartilage is
thought to be similar to articular cartilage [1]. The end plates are
important for fluid flow and nutrient supply from the vertebral
body to the NP and AF. NP cells originate from the embryonic
notochord in mice, and presumably this holds for other mamma-
lians as well [2]. By contrast, the AF originates from the embryonic
sclerotome [3]. NP cells constitute a chondrocyte phenotype that
in part resembles articular chondrocytes [4]. Importantly, cells of
the inner AF, directly bordering the NP, also acquire a chondrocyte
phenotype similar to the NP [5]. The outer AF consists predomi-
nantly of fibroblastic cells.
41
42 Guus G. H. van den Akker et al.
IVD cells are isolated from tissues for cellular studies. Multiple
types of (sequential) enzymatic digestions at different concentra-
tions and durations have been reported. Digestion time, type of
enzyme, and its concentration are known to affect cell viability
[6]. In addition, the time post-mortem at which cells are isolated
may affect cell viability and/or phenotype. For human post-mortem
IVD tissue, maximum sampling times of 24–72 h have
been reported [1, 7]. Surgical surplus material and animal tissues
are typically obtained within hours of dissection or death of the
animal. Surgical surplus material from scoliosis correction surgery
often consists of smaller fragments; this demands careful selection,
labeling, orientation, and dissection to obtain NP and AF cell
populations. In case of surplus material from IVD herniation sur-
gery it is recommended to verify tissue of origin by histological
analyses (NP/AF). Histological scoring of IVD degeneration in
human and certain animal samples can be done by various systems,
e.g., Mankin [8], Thompson [9], or modified Mankin [10]. Based
on our earlier work, we recommend a pre-isolation comparison
and/or histological analyses of NP and AF tissue of the same
donor to ensure correct isolation of the cell type of interest [11].
Following enzymatic digestion of the tissues, cells are strained,
washed, and plated at a high cell density in monolayer cultures or
brought into a 3D culture environment. A plethora of (pseudo) 3D
culture options are available, i.a. pellets, micromasses, alginate,
Matrigel, etc. This choice is mainly guided by the prevailing
research question, e.g., to expand cells, limit chondrocyte dediffer-
entiation, or mimic the microenvironment [3, 12–14]. Importantly,
different cell culture media were shown to affect marker gene
expression of IVD cells [15]. DMEM or mixtures of DMEM with
alpha-MEM or Ham’s F-12 are recommended for expansion of
IVD cells. Finally, it is recommended to perform post-isolation
measurements of IVD marker expression to confirm the successful
isolation of NP or (outer) AF cell populations [11, 16, 17].
The ORS/Spine Research Interest group released a consensus
definition of young healthy human NP cells [18]. The recom-
mended markers are: stabilized expression of HIF-1, GLUT-1,
aggrecan/collagen II ratio >20, Shh, Brachyury, KRT18/19,
CA12, and CD24 to assess variation between NP cells. However,
it may not be practical to perform all recommended measurements
for each individual donor. Moreover, NP cells from aged or degen-
erated tissue are known to have different expression profiles
[19]. Nevertheless, we advocate the systematic measurement of a
limited marker subset of NP and AF phenotypic marker genes as
standard practice to acquire a reference for and record on common
characteristics of isolated types. This is in particular valuable when
both NP and AF cells are isolated from one donor. These markers
provide an internal reference for the donor and the experimenter. A
subset of marker genes that we have successfully used in situ and
Isolation of Cells from the Intervertebral Disc 43
in vitro to assess correct isolation of NP and outer AF cells from

bovine and human donors consists of Brachyury/T, Keratin
19 (KRT19), Secreted frizzled-related protein 2 (SFRP2), and
Collagen type 12 alpha 1 (COL12A1) [11, 16, 17]. In our experi-
ence this constitutes a more sensitive and reliable method com-
pared to Collagen type 2/Collagen type 1 mRNA ratio to
characterize NP cells [11, 17]. The systematic measurement of
these markers will aid in a more uniform description of IVD cell
populations used in literature.
2 Materials
All solution are prepared in a sterile cell culture environment.

Prepared reagents are stored at 4 C unless indicated otherwise.
Cell culture reagents are prewarmed at 37 C prior to use. The
required waste disposal regulations should be followed.
2.1 Medical Ethical Prior to performing the procedures described below, it is obligatory
Approval for Human to obtain approval from a local medical ethical committee for the
Tissues use of tissues from deceased individuals or surgical leftover material
for research purposes prior to performing the procedures described
below.
2.2 Cell Culture 1. Fetal calf serum (FCS): Stock is thawed overnight at 4 C.
Decomplement FCS for 30 min at 56 C in a water bath. Fill
out 25 mL FCS per sterile polypropylene tube. Store at 20 C
in upright position.
2. Anti-Anti, Antibiotic Antimycotic (A/A), Gibco 15240-062:
Stock (100) is thawed at 37 C in a water bath. Aliquot
5.5 mL per tube, store at 20 C in upright position.
3. Non-essential amino acids (NEAA), Gibco 11140-035: Stock
(100) is thawed at 37 C in a water bath. Stock is stored at
4 C.
4. Sodium chloride (0.9%): Dissolve 9 g sodium chloride (suitable
for cell culture, >99% purity, Mw ¼ 58.44 g/mol) in 1 L
MQ-water. Sterilize the solution by autoclaving (121 C,
15 min). Store at room temperature.
5. DMEM/F12 (1:1) Glutamax, Gibco 31331-028.
6. DMEM/F12 (1:1), HEPES buffered, Gibco 31330-038.
7. Collagenase II, Invitrogen 17101-015: 10 concentrated
stock ¼ 300 U/mL or 1.1%. Dissolve in HEPES-buffered
DMEM/F12 with 1% A/A. Aliquot in 5 mL portions in
50 mL tubes and store at 20 C until use. Thaw and dilute
to 1 (NP) or 2 (AF) in HEPES-buffered DMEM/F12 with
1% A/A.
8. IVD expansion medium: DMEM/F12, Glutamax (440 mL),

10% decomplemented FCS (50 mL), 1 A/A (5 mL of 100
stock), 1 NEAA (5 mL of 100 stock), Limited shelf life,
store at 4 C.
9. 0.5% Trypsin-EDTA (TE), Gibco 15400-054: Stock (100) is
thawed for 30 min at 37 C in a water bath. Dilute 10 with
0.9% sodium chloride and aliquot 5 mL per 50 mL tube (10
stock), dilute again 10 in Sodium Chloride (0.9%) to obtain
1 stock. Store at 20 C in upright position. Keep 1 stock
at 4 C after thawing.
10. Cell strainer 70 μm, Greiner, 542070.
11. Freeze-down medium (10 mL, 2 concentrated stock):
DMEM/F12, Glutamax (5 mL), FCS (3 mL ¼ 30%), DMSO
(2 mL ¼ 20%). This solution can be aliquoted and stored at
20 C. Avoid repeated freeze/thaw cycles.
12. Mr. Frosty, Nalgene, VWR, 479-3200. Fill with isopropanol
until the mark on the side and refrigerate to 4 C.
13. Isopropanol, p.a Merck 1.09634.1011.
14. DMSO, cell culture grade, Santa Cruz, SC-358801.
15. Cryovials (2 mL), Greiner, 126277.
2.3 Cell Counting 1. Bürker Turk cell counter: Depth 0.1 mm, surface per square
0.04 mm2.
2. Trypan blue (0.4%), Sigma-Aldrich Chemie T8154.
3. Phosphate-buffered saline (PBS, 20 concentrated stock):
87.52 g Sodium chloride (suitable for cell culture, >99%
purity, Mw ¼ 58.44 g/mol), 14.16 g disodium hydrogen
phosphate dihydrate (>99.5% purity, Mw ¼ 177.99 g/mol),
2.15 g potassium dihydrogen phosphate (>99.5% purity,
Mw ¼ 136.09 g/mol), fill up to 500 mL. pH is between the
required range of 7.2–7.4, no further pH-correction is needed.
Stock is diluted 20 in MQ water to obtain 1 stock. Sterilize
by autoclaving (121 C, 15 min).
2.4 IVD Marker Gene RNA was isolated according to standard protocols and cDNA
Analyses by RT-qPCR synthesis was done using random hexameres. RT-qPCR reactions
were set up in Takyon™ No Rox SYBR Master Mix dTTP blue
(Eurogentec) in a CFX96 Real-Time PCR Detection System
(Biorad) according to the MIQE guideline [20]. Primer pair
sequences for human and bovine were described before and are
provided in Tables 1 and 2 [11, 16, 17].
Table 1
RT-qPCR primer sequences for human NP (T & KRT19) and AF (SFRP2 & COL12A1) marker genes
Human
Gene FW primer (50 -30 ) RV primer (50 -30 )

Brachyury/T CCACCTGCAAATCCTCATCCT TTGGAGAATTGTTCCGATGAGC
KRT19 GCAGTCACAGCTGAGCATGAA TCCGTTTCTGCCAGTGTGTCT
SFRP2 TGTGCCACGGCATCGA TCGTGGCCCAGCAGGTT
COL12A1 TGACAACCCTTTCCGACACA CTCCTCACGGTTCTAAAATTTGC
ACTB CCTGGCACCCAGCACAAT GCCGATCCACACGGAGTACT
PPIB CCTGCTTCCACCGGATCAT CGTTGTGGCGCGTAAAGTC
Beta Actin (ACTB) and Peptidyl-Prolyl Isomerase B (PPIB) are recommended as stable reference genes
Table 2
RT-qPCR primer sequences for bovine NP (T & Krt19) and AF (Sfrp2 & Col12a1) marker genes
Bovine
Gene FW primer (50 -30 ) RV primer (50 -30 )

bBrachyury/T CACACGGCTGCGAAAGGTA TGAACTGTCGGAATAGGTTGGA
bKrt19 GACCTGCGGGACCAGATTCTC GTCAGCCTCCACACTCATGC
bSfrp2 CAGGACAACGACCTTTGCAT TCACATACCTTTGGAGCTTCCT
bCol12a1 ACCGGCTACACTGTGACCTA TCCAGGCGCATCTCTTTGG
bGapdh CACCCACGGCAAGTTCAAC TCTCGCTCCTGGAAGATGGT
bRps14 CATCACTGCCCTCCACATCA TTCCAATCCGCCCAATCTTCA
Bovine glyceraldehyde-3-phosphate dehydrogenase (bGapdh) and bovine ribosomal protein S14 (bRps14) are recom-
mended as stable reference genes
3 Methods
Procedures are carried out at room temperature in a laminar flow

cabinet unless indicated otherwise.
3.1 Tissue 1. Tissues are obtained as fresh as possible from human or animal
Dissection and Cell donors (<24 h post-mortem).
Isolation It is recommended to transfer tissues as quickly as possible
to a sterile solution of PBS with 1 A/A (4 C).
2. A sterile metal plate or large sterile petri dish is used for sorting
tissue pieces and dissection with a sterile scalpel blade (size 20)
in a laminar flow cabinet (see Notes 1, 2 and Fig. 1 for tips on
recognizing the boundaries of tissues).
Fig. 1 Dissection of an intact bovine tail IVD. (a) The bovine tail IVD was separated from the vertebral body. The
border between NP and inner AF (iAF) is difficult to recognize by eye. A transversal cut from outer AF (oAF) and
the gradual loss of lamellae become obvious. (b) Pushing the AF toward an open position (indicated by arrows)
with a pincer reveals the demarcation of the inter zone and allows excision of the central NP. (c) Correct
isolation of amorphous proteoglycan-rich NP tissue was confirmed by formalin fixed, paraffin-embedded
histology (Safranin O staining)
3. Mince sorted NP and AF tissue pieces to a size of roughly

1 mm3 using a scalpel blade (size 10). See Notes 3 and 4 for
tips on handling and rapid processing.
4. Incubate human NP tissue pieces in collagenase type II (30 U/
mL ¼ 0.11%) in a sterile and covered Erlenmeyer while shaking
in a water bath at 37 C for 2 h or until completely digested (see
Notes 5 and 6 for AF tissue). Use 2.5–5 mL Collagenase II per
gram tissue, with a minimum of 10 mL. Replace scalpel blades
when excising a new donor or tissue.
5. After digestion of the tissue, pipet the solution on top of a
70 μm cell strainer in a sterile 50 mL tube (see Note 7 on
caution for notochordal cell isolation and Note 8 for incom-
plete digestion). Replace the cell strainer when flow through is
substantially decreased. Flush the cell strainer twice with 12 mL
0.9% sodium chloride.
6. Spin cells down in a 50 mL tube at 316 RCF for 8 min. Handle
pellet with care (see Note 9).
7. Remove the collagenase solution and floating debris by aspira-

tion and add 25 mL 0.9% sodium chloride. Resuspend cells
briefly by gentle pipetting and spin down for 8 min at 316 RCF.
8. Repeat the washing step above with 25 mL 0.9% sodium chlo-
ride, resuspend briefly, and spin down for 8 min at 316 RCF.
3.2 Cell Counting 1. Resuspend cells in an appropriate volume of culture medium,

approximately 5 mL for human tissue donors or 10 mL for
animal tissue donors, depending on the amount of starting
material. Gently pipet up and down 10 after macroscopically
visible clumps have disappeared to acquire single-cell
suspensions.
2. Take 15 μL cell suspension and mix 1:1 with 15 μL trypan blue
in an Eppendorf tube. Apply 10 μL of this mixture to an
appropriate counting vessel (perform in duplicate).
3. Count 25 fields of trypan blue negative cells (living cells) on a
Bürker-Turk. Count two sides of each square (see Fig. 2).
4. Calculate the total number of living cells in the cell suspension.
Multiply the counted cell number (25 fields) by the dilution
factor 2*104 to obtain the cell concentration in mL. The total
cell number is obtained by multiplying with the total volume
in mL.
3.3 Expansion 1. Plate freshly isolated cells at a density of 30.000 cells/cm2 in an

of Cells in Monolayer appropriate polystyrene culture vessel in IVD expansion
medium. Plating efficiency can differ between donors and
Fig. 2 Use of the Bürker Turk counting chamber. The cells of a single field are
counted within the square including the cells overlapping with the top and left
sides of the small square and not those overlapping with the right and bottom
sides. Only cells enclosed in the large square are counted
depends on cell type. Moreover, NP chondrocytes attach

slower (5–7 days) than AF fibroblasts (3–5 days).
2. Refresh media after cell attachment to the culture surface
(5–7 days for NP cells, see Note 10).
3. Culture cells up to 80% confluency (P0) and passage (Px +1,
etc.). It is recommended to count cells at every subsequent
passage and calculate population doublings, as these are more
informative than passage numbers.
3.4 Passaging IVD Passaging of IVD cells is usually done 1:4 and it may take 4–5 days
Cells in Monolayer for human NP cells to reach the same confluency after each passage.
Adjust this ratio accordingly; if less confluent, consider splitting 1:2
or 1:3 and if more confluent consider splitting 1:5 or 1:6. AF cells
usually reach confluency faster. An example of typical morphology
of human NP and AF cells is provided below (see Fig. 3).
1. Remove culture medium and wash with an appropriate amount
of 0.9% sodium chloride
Optional: wash twice to reduce the required trypsinization
time. This can be particularly useful at P0 when the cells have
generated a large amount of extracellular matrix.
2. Remove all sodium chloride without disturbing the cells and
add X mL Trypsin-EDTA according to the amounts indicated
below.
Fig. 3 Typical morphology of NP and AF cells in vitro. Cells were isolated from three human scoliotic IVD donors
and photographed using a Nikon Eclipse TE200 microscope. NP cells tend to prefer cell-cell contact and can be
more spindle shaped when compared to AF cells. D donor, pX passage number. Bar represents 20 μm
T25 ¼ 25 cm2 ¼ 1.0 mL TE.

T75 ¼ 75 cm2 ¼ 2.0 mL TE.
T175 ¼ 175 cm2 ¼ 3.0 mL TE.
3. Incubate 3–5 min in the CO2 incubator at 37 C.
4. Tap the flask to loosen the cells and check under an inverted
microscope if the cells are detached. Put back in the CO2
incubator if not sufficient cells have detached (should be
>90%).
5. Add 5–10 mL cell culture medium to the trypsinized cells to
neutralize the Trypsin-EDTA.
6. Transfer the solution to a sterile 50 mL tube and centrifuge for
5 min at 316 RCF at RT.
7. Remove the supernatant and resuspend the pellet in an appro-
priate amount of IVD expansion medium (4–8 mL is conve-
nient for passaging 1:4). Carefully pipet up and down 10 after
the last clump has disappeared to make a single-cell solution.
8. Count cells and replate for further expansion, an experiment or
for freezing down. Write down the number of seeded cells at
the start and end of each passage to calculate population dou-
bling level (see Note 11).
3.5 Freezing Down 1. Write down the donor number, cell type, passage number, and
IVD Cells date on the cryotubes.
2. Trypsinize and count the cells according to “Subheading 3.4.”
3. Centrifuge for 5 min at 316 RCF at room temperature (RT).
4. Remove supernatant and resuspend the IVD cells at a density of
2. 106 cells/mL.
5. Add cold 2 freeze down medium 1:1 drop-by-drop and mix
at the same time.
Note: be quick, DMSO at this concentration is toxic to the
cells.
6. Aliquot 1 mL cell suspension (1. 106 cells) per cryotube and
close the cap. Put cryotubes on ice immediately when freezing
multiple donors.
7. Put the vials in a refrigerated Mr. Frosty and store overnight at
80 C to freeze the cells.
Note: A maximum of 18 cryotubes fits in one Mr. Frosty.
8. The next day, transfer the cryotubes to a liquid nitrogen con-
tainer for long-term storage.
3.6 Thawing IVD 1. Transport tubes from the liquid nitrogen on ice to the cell
Cells from the Liquid culture laboratory.
Nitrogen 2. Release any liquid nitrogen that might have gotten into the
tube by slowly unscrewing the cap in the flow cabinet.
3. Rapidly thaw the cells in a 37 C water bath. Important: stop

when a small clump of ice is still visible.
4. Transfer the cells to a 50 mL tube and add 9 mL RT IVD
expansion medium dropwise and mix at the same time.
5. Centrifuge for 5 min at 316 RCF at RT.
6. Remove supernatant and plate the cells in a T75 culture flask
with 12 mL IVD expansion medium.
7. The next day, refresh medium to remove DMSO that might
have been released from the cells overnight. The far majority
(>80%) should attach and survive the freeze/thawing proce-
dure. Cells should be allowed to recover for a day or two prior
to use in experiments.
3.7 RT-qPCR In brief, RT-qPCR reactions are set up with 15 ng cDNA, 300 nM
Analyses of NP and AF forward and 300 nM reverse primers. The program was: 50 C for
Marker Genes 2 min, denaturation at 95 C for 10 min, 40 cycles of amplification
(15 s 95 C and 1 min 60 C) and followed by a melting curve
(95–50 C, 1 C/min). The standard curve method was used to
calculate gene expression and this was normalized to stable refer-
ence genes indicated in Tables 1 and 2.
4 Notes
1. It can be challenging to distinguish the NP from AF in human

aged/herniation surgery or bovine tail specimens. Cut trans-
versally into the AF toward the NP using a scalpel blade (see
Fig. 1a). It should become clear from the separation between
the lamellar AF tissue when the AF transitions into the inter-
zone and NP (see Fig. 1b). Store representative samples for
histological confirmation and grading (standard formalin fixa-
tion, paraffin embedding, and histology, Fig. 1c).
2. For preadolescence tissue samples, the separation between NP
and AF should be easily discernable by eye. Store representative
samples for histological grading.
3. Do not leave tissue pieces for too long outside the sodium
chloride container in which it was collected to avoid drying.
4. Mincing can be challenging for the AF. It is sometimes better
to limit the dissection time and digest slightly larger AF tissue
pieces.
5. Human outer AF tissue pieces tend to be tougher and it is
advisable to use 60 U/mL (instead of 30 U/mL) collagenase
(0.22%) solution and digestion usually takes 4–6 h.
6. For certain animal tissues, such as bovine tail AF tissue, it is
recommended to perform an overnight (16 h) digestion.
7. The 70 μm cell strainer is thought to prevent isolation of

notochordal cells that can have a diameter of up to 100 μm.
8. If a minority of small AF tissue pieces have not digested
completely within the indicated time periods, proceed with
cell isolation.
9. Pellet will be very unstable and can detach easily.
10. Not all cells will attach within this time frame, but the majority
should attach if the initial cell isolate had a good viability
(>90%).
11. The formula for PDL calculation ¼ 3.32(log (total viable cells
at harvest/total viable cells at seed)) [21]. The PDL of each
passage should be plotted in a cumulative figure.
Acknowledgements
This work was supported by grants from ReumaNederland and

Stichting de Weijerhorst.
References
1. Power KA, Grad S, Rutges JPHJ, Creemers 6. Molinos M, Almeida CR, Goncalves RM, Bar-
LB, van Rijen MHP, O’Gaora P, Wall JG, bosa MA (2015) Improvement of bovine
Alini M, Pandit A, Gallagher WM (2011) Iden- nucleus pulposus cells isolation leads to identi-
tification of cell surface–specific markers to tar- fication of three phenotypically distinct cell
get human nucleus pulposus cells: expression subpopulations. Tissue Eng Part A 21
of carbonic anhydrase XII varies with age and (15–16):2216–2227. https://doi.org/10.
degeneration. Arthritis Rheum 63 1089/ten.tea.2014.0461
(12):3876–3886. https://doi.org/10.1002/ 7. Thorpe AA, Binch ALA, Creemers LB,
art.30607 Sammon C, Le Maitre CL (2016) Nucleus
2. McCann MR, Séguin CA (2016) Notochord pulposus phenotypic markers to determine
cells in intervertebral disc development and stem cell differentiation: fact or fiction? Onco-
degeneration. J Dev Biol 4(1):3. https://doi. target 7(3):2189–2200. https://doi.org/10.
org/10.3390/jdb4010003 18632/oncotarget.6782
3. Pattappa G, Li Z, Peroglio M, Wismer N, 8. Mankin HJ, Dorfman H, Lippiello L, Zarins A
Alini M, Grad S (2012) Diversity of interverte- (1971) Biochemical and metabolic abnormal-
bral disc cells: phenotype and function. J Anat ities in articular cartilage from osteo-arthritic
221(6):480–496. https://doi.org/10.1111/j. human hips. II Correlation of morphology
1469-7580.2012.01521.x with biochemical and metabolic data. J Bone
4. Minogue BM, Richardson SM, Zeef LA, Free- Joint Surg Am 53(3):523–537
mont AJ, Hoyland JA (2010) Transcriptional 9. Thompson JP, Pearce RH, Schechter MT,
profiling of bovine intervertebral disc cells: Adams ME, Tsang IK, Bishop PB (1990) Pre-
implications for identification of normal and liminary evaluation of a scheme for grading the
degenerate human intervertebral disc cell phe- gross morphology of the human intervertebral
notypes. Arthritis Res Ther 12(1):R22–R22. disc. Spine (Phila Pa 1976) 15(5):411–415.
https://doi.org/10.1186/ar2929 https://doi.org/10.1097/00007632-
5. Hayes AJ, Isaacs MD, Hughes C, Caterson B, 199005000-00012
Ralphs JR (2011) Collagen fibrillogenesis in 10. Thomas CM, Fuller CJ, Whittles CE, Sharif M
the development of the annulus fibrosus of (2007) Chondrocyte death by apoptosis is
the intervertebral disc. Eur Cell Mater associated with cartilage matrix degradation.
22:226–241 Osteoarthr Cartil 15(1):27–34. https://doi.
org/10.1016/j.joca.2006.06.012
11. van den Akker GGH, Koenders MI, van de Loo 11(1):e0144497–e0144497. https://doi.org/
FAJ, van Lent PLEM, Blaney Davidson E, van der 10.1371/journal.pone.0144497
Kraan PM (2017) Transcriptional profiling dis- 17. van den Akker GGH, Surtel DAM, Cremers A,
tinguishes inner and outer annulus fibrosus from Rodrigues-Pinto R, Richardson SM, Hoyland
nucleus pulposus in the bovine intervertebral JA, van Rhijn LW, Welting TJM, Voncken JW
disc. Eur Spine J 26(8):2053–2062. https://doi. (2014) Novel immortal human cell lines reveal
org/10.1007/s00586-017-5150-3 subpopulations in the nucleus pulposus.
12. Tang X, Jing L, Richardson WJ, Isaacs RE, Arthritis Res Ther 16(3):R135. https://doi.
Fitch RD, Brown CR, Erickson MM, Setton org/10.1186/ar4597
LA, Chen J (2016) Identifying molecular phe- 18. Risbud MV, Schoepflin ZR, Mwale F, Kandel
notype of nucleus pulposus cells in human RA, Grad S, Iatridis JC, Sakai D, Hoyland JA
intervertebral disc with aging and degenera- (2015) Defining the phenotype of young
tion. J Orthop Res 34(8):1316–1326. healthy nucleus pulposus cells: recommenda-
https://doi.org/10.1002/jor.23244 tions of the spine research interest group at
13. Huang Y-C, Hu Y, Li Z, Luk KDK (2018) the 2014 annual ORS meeting. J Orthop Res
Biomaterials for intervertebral disc regenera- 33(3):283–293. https://doi.org/10.1002/
tion: current status and looming challenges. J jor.22789
Tissue Eng Regen Med 12(11):2188–2202. 19. Rutges J, Creemers LB, Dhert W, Milz S,
https://doi.org/10.1002/term.2750 Sakai D, Mochida J, Alini M, Grad S (2010)
14. Johnson WEB, Stephan S, Roberts S (2008) Variations in gene and protein expression in
The influence of serum, glucose and oxygen on human nucleus pulposus in comparison with
intervertebral disc cell growth in vitro: implica- annulus fibrosus and cartilage cells: potential
tions for degenerative disc disease. Arthritis Res associations with aging and degeneration.
Ther 10(2):R46–R46. https://doi.org/10. Osteoarthr Cartil 18(3):416–423. https://
1186/ar2405 doi.org/10.1016/j.joca.2009.09.009
15. Schubert AK, Smink JJ, Pumberger M, 20. Bustin SA, Benes V, Garson JA, Hellemans J,
Putzier M, Sittinger M, Ringe J (2018) Standar- Huggett J, Kubista M, Mueller R, Nolan T,
disation of basal medium for reproducible culture Pfaffl MW, Shipley GL, Vandesompele J,
of human annulus fibrosus and nucleus pulposus Wittwer CT (2009) The MIQE guidelines:
cells. J Orthop Surg Res 13:14. https://doi.org/ minimum information for publication of quan-
10.1186/s13018-018-0914-y titative real-time PCR experiments. Clin Chem
16. van den Akker GGH, Surtel DAM, Cremers A, 55(4):611. https://doi.org/10.1373/
Richardson SM, Hoyland JA, van Rhijn LW, clinchem.2008.112797
Voncken JW, Welting TJM (2016) Novel 21. Roth S (1974) Tissue Culture. Methods and
immortal cell lines support cellular heterogene- Applications. Paul F. Kruse, Jr. , M. K. Patter-
ity in the human annulus Fibrosus. PLoS One son, Jr. Q Rev Biol 49(2):150–151. https://
doi.org/10.1086/408024
Chapter 5
Engineering Cartilage Tissue by Co-culturing

of Chondrocytes and Mesenchymal Stromal Cells
Yao Fu, Carlo A. Paggi, Amel Dudakovic, Andre J. van Wijnen,
Janine N. Post, and Marcel Karperien
Abstract
Co-culture of chondrocytes and mesenchymal stromal cells (MSCs) has been shown to be beneficial in
engineering cartilage tissue in vitro. In these co-cultures, MSCs increase the proliferation and matrix
deposition of chondrocytes. The MSCs accomplish this beneficial effect by so-called trophic actions.
Thus, large cartilage constructs can be made with a relatively small number of chondrocytes. In this chapter,
we describe different methods for making co-cultures of MSCs and chondrocytes. We also provide detailed
protocols for analyzing MSC-chondrocyte co-cultures with cell tracking, proliferation assays, species-
specific polymerase chain reactions (PCR), rheological analysis, compression analysis, RNA-sequencing
analysis, short tandem repeats analysis, and biochemical examination.
Key words Chondrocytes, Mesenchymal stromal cells, Co-culture, Trophic effects, Cartilage engi-
neering, Matrix deposition
1 Introduction
Partial replacement of chondrocytes by alternative cell sources can

reduce the number of chondrocytes needed to engineering carti-
lage constructs in vitro [1–3]. Hendriks et al., co-cultured bovine
primary chondrocytes with human expanded chondrocytes, human
dermal fibroblasts, mouse embryonic stem cells, mouse-3T3 feeder
cells, or human mesenchymal stromal cells (MSCs) in cell pellets
[4]. Their data indicated that cartilage matrix deposition increased
in co-culture pellets. Replacement of 80% of the chondrocytes with
other cell types resulted in similar amounts of GAG production
when compared to pure chondrocyte pellets. This beneficial effect
on cartilage formation is most prominent in co-cultures of chon-
drocytes with mesenchymal stromal cells [5]. In a more recent
study, we used a xenogeneic co-culture model of human MSCs
and bovine chondrocytes to study the contribution of each cell
type to cartilage matrix formation [6, 7]. Our data showed a
53
54 Yao Fu et al.
significant decrease in MSCs in co-culture pellets over time, result-

ing in an almost homogeneous cartilage tissue predominantly
derived from the initially seeded chondrocytes. Our data showed
that the beneficial effect of co-culture is largely due to increased
chondrocyte proliferation and matrix formation, while chondro-
genic differentiation of the MSCs only marginally contributed to
cartilage formation. We also demonstrated that these observations
present in co-culture pellets of chondrocytes and MSCs are inde-
pendent of donor variation and culture conditions [8]. Subsequent
experiments indicated that increased secretion of fibroblast growth
factor 1 (FGF1) in co-culture of MSCs and chondrocytes is respon-
sible for increased chondrocyte proliferation in pellet co-cultures
[9]. Thrombospondin-2 has also been reported to be secreted by
MSCs to promote chondrogenic differentiation both in vitro and
in vivo [10].
To facilitate the engineering of large cartilage constructs it is
beneficial to combine the co-culture with a support matrix. Many
different scaffolds have been described in literature. Of these,
hydrogels are particularly promising for cartilage tissue engineer-
ing. We showed that the beneficial co-culture effect on cartilage
matrix formation is preserved in a hydrogel matrix consisting of
agarose or tyramine-conjugated hyaluronic acid and tyramine-
conjugated dextran [11]. The tyramine-conjugated polymers
cross-link in situ by a biocompatible enzymatic cross-linking reac-
tion. The method is compatible with incorporation of chondro-
cytes and mesenchymal stem cells which can be mixed with the
polymer conjugates prior hydrogel gelation. The cross-linked
hydrogel supports long-term cell survival and matrix deposition
in vitro [11]. This matrix production can even be further acceler-
ated by incorporation of cellular micro-aggregates of 50–250 cells
that can be uniformly produced using a high-throughput microwell
platform [12]. Our data showed that micro-aggregation of the
expanded monoculture cells prior to incorporation in the hydrogel
boosted the chondrogenic potential and cartilage matrix formation
both in vitro and in vivo [12, 13]. Different co-culture systems
were subsequently performed in the injectable hydrogels or as
micro-aggregates. The beneficial effects of MSCs in co-cultures
were also observed in these studies.
These reports revealed that the beneficial effects in co-cultures
were largely due to the stimulation of proliferation and the matrix
formation of chondrocytes induced by a trophic effect of the MSCs
[14]. There is a dynamic regulation and interplay of stem cell-
derived cytokines that influence tissue survival, repair, and regener-
ation, including the activation of resident and circulating stem cells
[15]. These studies point to a dominant role of the MSCs in
stimulating resident or co-implanted chondrocytes to initiate a
regenerative response. These findings starting from observations
in vitro have been confirmed in animal studies and recently as well
as a clinical trial [16, 17].
Methods for Co-culturing of Chondrocytes and MSCs 55
Moreover, while chondrogenic differentiation of MSCs can be

triggered by chondrocytes, it also induces death of the MSCs [6]. It
was postulated that the underlying mechanism is likely related to
one of the deliberate “suicide” programs present within cells
[14]. The programmed suicide death has usually two main forms,
apoptosis and autophagy [18–20]. These suicide programs are
usually induced intrinsically or extrinsically by external stimuli
such as mechanical stress, oxidative processes, and drug treatments.
Further studies need to be performed to investigate the mechanism
behind the trophic effects of MSCs in co-cultures.
2 Materials
2.1 Cell Sources 1. Bovine primary chondrocytes (bPCs) are isolated from full-
thickness cartilage knee biopsies of female calves that are
approximately 6 months old. Cartilage is separated and
digested to extract primary chondrocytes (see Subheading 3.1).
2. Human primary chondrocytes (hPCs) are obtained from full-
thickness cartilage dissected from knee biopsies of a patient
undergoing total knee replacement (see Subheading 3.2).
3. Human MSCs (hMSCs) are isolated from bone marrow aspi-
rates of healthy donors (see Note 1).
2.2 Media, Solutions, 1. Chondrocyte proliferation medium contains DMEM supple-

Chemicals, and Kits mented with 10% FBS, 1 nonessential amino acids, 0.2 mM
ascorbic acid 2-phosphate (AsAP), 0.4 mM proline, 100 U
penicillin /mL, and 100 μg/mL streptomycin.
2. Chondrogenic differentiation medium contains DMEM sup-
plemented with 40 μg/mL of proline, 50 μg/mL ITS-premix,
50 μg/mL of AsAP, 100 ug/mL of sodium pyruvate, 10 ng/
mL of transforming growth factor beta 3 (TGFβ3, R&D sys-
tem), 10–7 M of dexamethasone, 500 ng/mL of bone mor-
phogenetic protein 6 (BMP6, R&D system), 100 U penicillin/
mL, and 100 μg/mL streptomycin.
3. MSC proliferation medium contains α-MEM plus 10% fetal
bovine serum, 1% L-glutamine, 0.2 mM ascorbic acid,
100 U/mL penicillin, 10 μg/mL streptomycin, and 1 ng/
mL basic fibroblast growth factor (bFGF, R&D system).
4. Proteinase K digestion buffer: 1 mg/mL proteinase K (Sigma) in
Tris/EDTA buffer (pH 7.6), 18.5 μg/mL iodoacetamide, and
1 μg/mL pepstatin A. The proteinase K solution can be stored in
aliquots at 20 C for several weeks. After one thaw, do not
freeze again. Tris/EDTA buffer: Dissolve 6.055 g Tris and
0.372 g EDTA · 2 H2O in 1000 mL of H2O. Adjust pH to 7.6.
5. PBE buffer: 14.2 g/L Na2HPO4 and 3.72 g/L Na2EDTA,
pH 6.5.
56 Yao Fu et al.
6. GAG stock solution: 50 mg/mL, 17.5 mg of cysteine-HCl was

dissolved in 10 mL of PBE buffer; GAG stock solution can be
aliquoted and stored in 20 C freezer. A GAG working
solution is 250 times dilution of GAG stock solution in PBE
buffer, which contains 200 μg/mL of GAG.
7. DMMB solution: add 9.5 mL of 0.1 M HCl solution to
90.5 mL of d2H2O plus 0.304 g of glycine and 0.237 g of
NaCl; adjust to pH 3 before adding 1.6 mg of DMMB to the
buffer. When stored in the dark at RT, the solution is stable for
3 months; filter to get rid of precipitates before use.
8. Tyramine-conjugated polymers: dextran/hyaluronic acid-
tyramine conjugates dissolved in appropriate medium
(Hy2Care). To obtain stable hydrogel constructs the end con-
centration of the polymers should be between 5 and 10 wt/v %.
9. Organic fluorescent dye (CM-DiI), Click-iT® EdU Imaging
Kit, and the CyQuant DNA Kit (Molecular Probes).
10. QIAamp DNA Mini Kit and RNeasy Mini Kit (Qiagen).
11. iScript cDNA Synthesis kit and iQ SYBR Green Supermix
(Bio-Rad).
12. Direct-zolTM RNA kit (Zymo Research).
13. TruSeq RNA library preparation kit (Illumina).
14. PowerPlex 16 System (Promega).
15. Collagenase type II (Worthington).
16. Click-iT® EdU Imaging Kit (Invitrogen).
17. A round-bottom ultra-low attachment 96 wells plate
(Corning).
18. Cryomatrix (Shandon).
19. DMMB (1, 9-Dimethyl-Methylene Blue, Sigma).
20. DMTMM (4-(4,6-Dimethoxy-1,3,5-triazin-2-yl)-4-methyl-
morpholinium Chloride, Fluorochem Ltd. UK).
21. Milli-Q water (Milli-Q Advantage A10 system with 0.22 μm
Millipak®-40 Express filter).
22. Horseradish peroxidase (HRP, Sigma-Aldrich).
23. Hydrogen peroxide (H2O2, 30%, Sigma-Aldrich).
24. Ultrapure™ agarose (Invitrogen).
2.3 Equipment 1. BD pathway 435 confocal microscope (BD Biosciences).

2. ELISA reader (TECAN).
3. MyiQ2 Two-Color Real-Time PCR Detection System
(Bio-Rad).
4. PTEE mold for hydrogel formation (see Fig. 1a, b).
Fig. 1 Schematic illustration of PTEE mold for hydrogel formation (a and b) and PDMS convex mold for micro-
aggregation (c)
5. PDMS convex mold for micro-aggregation (see Fig. 1c).

6. Rheometer (MCR 302, Anton Paar).
7. Texture Analyser TA-HD plus (StableMicro Systems Ltd.).
3 Methods
3.1 Isolation of 1. Human cartilage tissue was obtained from total knee or hip
Human Articular joint replacement.
Chondrocytes 2. Cartilage tissue is cut from underlying bone and connective
tissue with scalpels and chopped into pieces of approximately
2 2 mm.
3. Digest cartilage pieces for 20–22 h in collagenase type II
(0.15%) in DMEM supplemented with penicillin (100 U/
mL) and streptomycin (100 mg/mL).
3.2 Isolation of 1. Collect bone marrow aspirates in sterile heparin tubes.

Human Bone marrow 2. Pour aspirate into 50 mL Falcon tubes.
Mesenchymal Stromal
3. Remove red blood cells by incubating 100 μL aliquots of
Cells
aspirate with 900 μL red blood cell lysing buffer for
5–10 min on ice or until transparent.
4. Count cell numbers with trypan blue staining. Plate cells at
50,000/cm2 in T75 in MSC proliferation medium plus 1%
heparin.
3.3 Cell Tracking of 1. Trypsinize bovine or human chondrocytes and resuspend in

Cell Populations in PBS at a concentration of 2 106 cells/mL.
Pellet Co-cultures with 2. Incubate the cells with the fluorescent dye CM-DiI (final con-
Organic Fluorescent centration of 4 μM) at 37 C for 5 min followed by incubation
Dyes CM-DiI at 4 C for 15 min.
3. Wash cells two times by suspending cells in PBS followed by
collecting cells by centrifuging at 300 g for 3 min.
58 Yao Fu et al.
3.4 Co-culture of 1. Trypsinize hMSCs and suspend in chondrocyte proliferation

bPCs and hMSCs in medium at a concentration of 1 106 cells/mL. Resuspend
Pellets labeled bPCs or hPCs from Subheading 3.1 at the same con-
centration as hMSCs in chondrocyte proliferation medium.
2. Mix hMSCs with bPCs or hPCs at ratios of 80/20% and
50/50%. Seed a total of 200,000 cells in one well of a round-
bottom ultra-low attachment 96 wells plate in chondrocyte
proliferation medium.
3. Use monoculture of hMSCs only or bPCs only or hPCs only as
controls. Cell numbers per well are the same as in co-culture
pellets.
4. Make pellets by centrifugation of the plate at 500 g for 5 min.
5. Xenogeneic co-cultures (bPCs and hMSCs), including
corresponding controls, are cultured in chondrocyte prolifera-
tion medium at all times.
6. For allogenic co-cultures (hPCs and hMSCs), including
corresponding controls, medium is changed to chondrogenic
differentiation medium (see Subheading 2.2) on the second day
after seeding.
3.5 Co-culture of 1. Dissolve the tyramine-conjugated polymers in sterile PBS and

bPCs and hMSCs in incubate the polymer solution with HRP overnight at 4 C.
Injectable Hydrogels 2. Trypsinize hMSCs and bPCs and resuspend in chondrocyte
proliferation medium. Mix hMSCs with bPCs at ratios of
80/20%.
3. Combine the cell mixture with polymers solution mentioned
above at concentration of ten million cells / mL. Pure polymers
groups without cells are also prepared as control groups.
4. Add the freshly prepared 0.3% hydrogen peroxide (H2O2) to
the mixture and immediately transfer to the mold (see Subhead-
ing 2.3) using a 1 mL pipette after brief vortex (see Note 2).
5. After gelation, transfer the gels to six-well plates and incubate
in chondrogenic differentiation medium at all times.
3.6 Co-culture of 1. Agarose microwell chips containing ~5000 microwells with a

bPCs and hMSCs in diameter of 200 μm were fabricated by pouring a 3% Ultrapur-
Micro-aggregates eTM agarose solution on convex molds of PDMS (see Subhead-
ing 2.3).
2. After agarose solidification, separate the agarose chips from the
PDMS molds and transfer into 12-well culture plates covering
with PBS (see Note 3).
3. Trypsinize hMSCs and bPCs and resuspend in chondrocyte
proliferation medium. Mix hMSCs with bPCs at ratios of
80/20%.
4. Seed the cell mixture onto the agarose chips resulting in aggre-
gates consisting of 100 cells. Immediately after seeding, centri-
fuge the agarose chips at 300 g for 3 min to allow the cells to
settle down in the microwells and then incubate at 37 C for
24 h to form micro-aggregates.
5. The micro-aggregates can be obtained by flushing the well
plates with medium and centrifugation.
3.7 Examination of 1. 2 or 3 days after making pellets, add EdU (5-ethynyl-2-

Cell Proliferation in 0
-deoxyuridine, provided in Click-iT® EdU Imaging Kit) to
Pellets by EdU the culture medium of pellets at a concentration of 10 μM.
Labeling and Staining 2. Harvest samples for analysis, 24 h later by transferring pellets to
Eppendorf tubes.
3. Wash cell pellets with PBS and fix with 10% formalin for
15 min.
4. Embed samples in cryomatrix and cut 10 μm sections with a
cryotome.
5. Permeabilize sections and stain for EdU with Click-iT® EdU
Imaging Kit according to the manufacturer’s protocol. In this
kit, nuclei are counterstained with LP435 (Hoechst 33342,
provided in Click-iT® EdU Imaging Kit).
3.8 Image 1. Make fluorescent images with a BD pathway 435 confocal

Acquisition and microscope (see Note 4).
Analysis by 2. Capture three separate images for each pellet section, using
Fluorescent BP536/40 (Alexa 488), BP593/40 (DiI), and LP435
Microscopy (Hoechst 33342) and pseudo color green, red, and blue,
respectively.
3. Open blue image of one pellet section with ImageJ
software [21].
4. Set threshold by clicking drop-down menu via Image –
> Adjust –> Threshold (see Note 5).
5. Open particle analyzer via Analyze –> analyze particles.
6. Set area restrictions: 100-infinite; choose Display results,
Exclude on edges, Include holes; click OK to count NUMBER
of total cell (see Note 6).
7. Open red image of the same pellet section; set threshold as

described above (see Note 7).
8. Open image calculator via Process- > Image calculator.
9. Select blue image in the box of Image 1; select red image in the
box of Image 2; select “AND” in the box of Operation; then
click OK to generate a new image named “result of blue.”
10. Run “Analyze particles” on new image “result of blue” with
the same setting as above to count NUMBER of red cell.
60 Yao Fu et al.
11. Open green image of the same pellet; set threshold and area
restriction (see step 6) to count NUMBER of green cell.
12. Run “Image calculator” by selecting green image in Image
1 box and red image in Image 2 box, with AND in Operation
box to generate new image named “result of green.”
13. Run “Analyze particles” on new image “result of green” with
same setting as above to count NUMBER of green plus red cell.
14. Input all NUMBERs into an Excel spreadsheet and perform
the following calculations: Rate of EdU positive
Chondrocyte¼ NUMBER of green plus red cell NUMBER of
red cell 100%; Rate of EdU positive MSCs¼(NUMBER of green
cell – NUMBER of green plus red cell)(NUMBER of total cell – N
UMBER of red cell) 100%; Labeling efficiency ¼ NUMBER of
red cell NUMBER of total cell100% (see Note 8).
3.9 Quantitative GAG 1. Perform glycosaminoglycan (GAG) and DNA assay at the end
and DNA Assay of co-culture (i.e., 4 weeks).
2. Wash cell pellets (n ¼ 6) with PBS and freeze pellets overnight
at 80 C.
3. Digest pellets in 500 μL of proteinase K digestion buffer (see
Note 9) for more than 16 h at 56 C.
4. To prepare a standard curve, make dilution series of cysteine-
HCl, according to Table 1.
5. Add 5 μL of a 2.3 M NaCl solution and 25 μL of the samples or
the standard in one well of a 96-well non-tissue culture treated
plate.
6. Add 150 μL of the DMMB (1, 9-Dimethyl-Methylene Blue)
solution (see Subheading 2.2) and read the absorbance at
520 nm on an ELISA reader. Figure 2 gives an example of a
standard curve (see Subheading 2.2).
7. Determine cell number by quantification of total DNA using a
CyQuant DNA Kit, according to the manufacturer’s
instructions.
Table 1
Series dilution of GAG standards
GAG amount Blank 0.5 μg 1 μg 1.5 μg 2 μg 2.5 μg

GAG working solution (see Note 9) 0 μL 10 μL 20 μL 30 μL 40 μL 50 μL
PBE buffer (see Note 10) 100 μL 90 μL 80 μL 70 μL 60 μL 50 μL
Fig. 2 An example of standard curve for GAG quantification. The blank (25 μL
PBE, 5 μL 2.3 M NaCl and 150 μL DMMB solution) has an O.D. value of
0.18 0.03
3.10 Cell Tracking 1. Perform species-specific PCR to determine the ratio of MSCs
with Species-Specific and chondrocytes in xenogeneic co-culture (hMSCs and bPCs)
PCR pellets at the end of culture (i.e., 4 weeks).
2. Isolate DNA samples of pellets with a QIAamp DNA Mini Kit
according to the manufacturer’s protocols.
3. Extract RNA samples of pellets with an RNeasy Mini Kit (see
Note 10).
4. Reverse-transcribe one microgram of total RNA into cDNA
using the iScript cDNA Synthesis kit.
5. Perform species-specific quantitative PCR (qPCR) on genomic
DNA or cDNA samples by using the iQ SYBR Green
Supermix.
6. Carry out PCR Reactions on MyiQ2 Two-Color Real-Time
PCR Detection System under the following conditions: cDNA
is denatured for 5 min at 95 C, followed by 45 cycles consist-
ing of 15 s at 95 C, 15 s at 60 C, and 30s at 72 C.
7. Generate a melting curve for each reaction to test primer dimer
formation and nonspecific priming.
8. The primers for real-time PCR, either species specific or cross-
species specific, are listed in Tables 2 and 3.
9. For each gene, standard curves are obtained by serial dilutions
of cDNA (see Note 11). Figure 3 gives an example of standard
curve for qPCR.
62 Yao Fu et al.
Table 2
Forward (F) and Reverse (R) primers used for quantitative PCR on genomic DNA
Product
Gene name Primer sequence size Gene bank no.
0 0
Cross-species F: 5 GCATTGCCCTCAACGACCA 3 179 OR NC_000012 &
GAPDH R: 50 CACCACCCTGTTGCTGTAGCC 30 171a NC_007303
Human-specific F: 50 TTCCACCCATGGCAAATTCC 30 131 NC_000012
GAPDH R: 50 TTGCCTCCCCAAAGCACATT 30
Bovine-specific F: 50 AGCCGCATCCCTGAGACAAG 30 132 NC_007303
GAPDH R: 50 CAGAGACCCGCTAGCGCAAT 30
a
Product size of human genomic GAPDH is 179, of bovine genomic GAPDH is 171
Table 3
Forward (F) and Reverse (R) primers used for quantitative RT-PCR
Product
Gene name Primer sequence size Gene bank no.
Cross-species F: 50 GCGCAAGTACTCCGTGTGGA 30 123 NM_001101 &
β-actin R: 50 AAGCATTTGCGGTGGACGAT 30 NM_173979
Cross-species F: 50 AGCTCACTGGCATGGCCTTC 30 116 NM_002046&
GAPDH R: 50 CGCCTGCTTCACCACCTTCT 30 NM_001034034
Human-specific F: 50 CGCTCTCTGCTCCTCCTGTT 30 82 NM_002046
GAPDH R: 50 CCATGGTGTCTGAGCGATGT 30
Bovine-specific F: 50 GCCAT CACTG CCACC CAGAA 30 207 NM_001034034
GAPDH R: 50 GCGGCAGGTCAGATCCACAA 30
Human-specific F: 50 TTCCCATCGTGCCTTTCCA 30 121 NM_013227
aggrecan R: 50 AACCAACGATTGCACTGCTCTT 30
Bovine-specific F: 50 CCAAGCTCTGGGGAGGTGTC 30 98 NM_173981
aggrecan R: 50 GAGGGCTGCCCACTGAAGTC 30
Human-specific F: 50 GGCGGGGAGAAGACGCAGAG 30 129 NM_001844
collagen II R: 50 CGCAGCGAAACGGCAGGA 30
Bovine-specific F: 50 AGGTCTGACTGGCCCCATTG 30 101 NM_001001135
collagen II R: 50 CTCGAGCACCAGCAGTTCCA 30
Human-specific F: 50 GGCAGAAATGGCCGAGACG 30 150 NM_001851
collagen IX R: 50 CCCTTTGTTAAATGCTCGCTGA 30
Bovine-specific F: 50 GGACTCAACACGGGTCCACA 30 102 XM_601325
collagen IX R: 50 ACAGGTCCAGCAGGGCTTTG 30
Standard Unknown
35
30
Threshold Cycle
25
20
15
10
-0,5 0 0,5 1 1,5 2 2,5
Log Starting Quantity,copy number
SYBR E=118, 1% R^2=0, 996 slope=-2, 953 y-int=22, 150

SYBR1 E=110, 3% R^2=0, 998 slope=-3, 097 y-int=27, 560
SYBR2 E=110, 4% R^2=0, 993 slope=-3, 095 y-int=30, 151
Fig. 3 An example of standard curve for qPCR. SYBR, SYBR1, and SYBR2 stand for three different primer sets
10. Use Bio-Rad iQ5 optical system software (version 2.0) to

calculate copy numbers for each condition using the standard
curve as reference.
11. Ratio of bovine or human cells in the xenogenic co-culture
pellets are defined as the proportion of human or bovine
GAPDH copy numbers as percentage of the total copy num-
bers of both human and bovine genes determined by species-
specific PCR using genomic DNA as a template.
12. The relative mRNA expression level of bovine or human genes
in xenogenic co-cultures is determined by normalizing the
values using cross-species-specific GAPDH and β-actin
primers.
3.11 Rheological 1. Perform Rheological measurements to determine the storage

Analysis modulus of hydrogels incorporated with MSC-Chondrocytes
co-cultures pellets at the end of culture (i.e., 4 weeks) (see
Note 12).
2. Wash the hydrogels with PBS (see Note 13).
3. Start the RheoCompass™ software. Load the related parallel
plate measuring system.
4. Adjust the measuring chamber temperature at 20 C.
5. Load one sample on the chamber and select the measuring
protocol. This protocol set parameters with a 0.05 N normal
force in the oscillatory mode with 0.5% strain and 1 Hz, which
was in the LVE range according to measured frequency and
strain sweeps. Figure 4a gives an example of a standard curve.
64 Yao Fu et al.
Fig. 4 Example figures of mechanical properties analysis. (a) Storage modulus curve for rheological analysis.
(b) Stress–strain curve for compressive test
3.12 Compression 1. Perform texture analysis to determine the compressive modu-

Test lus of hydrogels prepared and equilibrated for the rheological
testing.
2. Load the sample on the loadcell at room temperature. The
sample undergo three compression cycles with a maximum
strain of 50% using a compression speed of 0.05 mm/s. Fig-
ure 4b gives an example of a stress strain curve.
3. The initial compressive modulus was calculated from the stress–
strain curves using a linear slope at a strain ranging from 0% to
2.5%. The high strain compressive modulus was calculated
from the stress–strain curves using a linear slope at a strain
ranging from 40% to 49.5%. The percentage of energy
dissipated during a compression cycle was calculated by divid-
ing the surface of the hysteresis loop by the surface under the
compression trace.
3.13 RNA- 1. Perform RNA-sequencing analysis to determine the gene

Sequencing Data expression of sex-mismatched MSCs and human chondrocytes
Analysis in pellet co-cultures (see Note 14).
2. Isolate total RNA samples from pellets using the Direct-zolTM
RNA kit. Perform the RNA-sequencing procedure with the
RNA samples. The values obtained are normalized and
expressed as reads per kilobase pair per million mapped reads
(RPKM).
3. To remove the values related to noise or highly influenced by
noise, consider an RPKM average value above 0.3. For average
value is considered the average obtained by combining the
value of a gene for all the samples.
1X
n
RPKMaverage ¼ RPKMn > 0:3
n
i¼1
where n is the number of samples considered in the analysis (see

Note 15).
4. The sex-mismatch of the MSC and chondrocyte donors can be
used to attribute gene expression to either one of the donors.
To determine the RPKM level of the two donors in the co-cul-
ture, the ratio between the monoculture (male donor) pellets
and the co-culture (male + female donors) is calculated.
RPKMmonoculture
%RPKMmale ¼
RPKMcoculture
%RPKMfemale ¼ 100 %RPKMmale
The percentage obtained is the expected value of the gene

expression of the male and female donor in the co-culture.
Moreover, the same value can be used to determine the
expected value of the co-culture with the assumption of cells
not interacting with each other.
RPKM Cocultureexpected ¼ %RPKMmale ∗RPKM donor1mono

þ %RPKMfemale ∗RPKM donor2mono
where RPKM donor1mono and RPKM donor2mono are the real

value of the individual donors cultured in monoculture and the
%RPKMmale/female is the % obtained in step 3.
66 Yao Fu et al.
5. To determine the fold change, compare the real values with the
expected values:
RPKM Coculturereal
Fc ¼
RPKM Cocultureexpected
6. Values above 2 are considered highly upregulated and values
below 0.5 downregulated (see Note 16).
Fc < 0:5&Fc > 2 ðhigh sensitivity and specificityÞ

7. Add the upregulated gene names into STRING software to
determine gene-interconnections and GO-pathways. Perform
the same procedure for the downregulated (as shown in
Fig. 5a).
8. Add the entire data set of raw RPKM data to MORPEUS
software. The system gives as output the heatmap of all the
genome and allow samples comparison (as shown in Fig. 5b).
9. Up-load the upregulated or downregulated genes in PAN
THER. This software determines the function of the individual
genes and creates clusters of genes involved in the same path-
way (as shown in Fig. 5c).
Fig. 5 Example figures of RNA-sequencing data analysis. (a) Heatmap of collagen-related genes comparison in
different conditions. (b) Inter-connecting cluster showing collagen-related genes
3.14 Short Tandem 1. Perform STR analysis to determine the ratio of MSCs and
Repeats (STR) Analysis chondrocytes in allogeneic co-cultures (hMSCs and hPCs)
pellets.
2. Extracts genomic DNA samples from pellets (n ¼ 6) with the
QIAamp DNA Mini Kit.
3. Amplify the sixteen loci of the kit PowerPlex 16 System, type
“sequence,” and analyze all loci according to manufacturer’s
protocol.
4. Compare monocultures of hMSCs or hPCs to find informative
alleles only present in either the hMSCs or the hPCs donor (see
Note 17).
5. Make electropherograms of the informative loci.
6. In the resulting electropherogram, calculate the area under the
peaks, which reflects the abundance of the alleles.
7. The sum of the area under the peak for the two donor-specific
alleles represents a relative amount of DNA for this donor.
8. Calculate the relative DNA amount for both the hMSC and the
hPC donor.
9. Calculate the ratio of hMSCs and hPCs in the pellet by
dividing through the total amount of relative DNA present
in the pellet.
4 Notes
1. We define the “primary” cells (bPCs, hPCs and hMSCs) in this

manuscript as cells with low passage number (<2) without
immortalization.
2. The vortex speed should be carefully adjusted. The speed
should be high enough to properly mix the solutions, while
not negatively affecting the incorporated cells.
3. Before cell seeding, the chips were preincubated in culture
medium overnight at 37 C.
4. Using montage capture, images of high resolutions were
obtained covering the entire section of a pellet. Choose the
20 objective. Use standard setting for the microscope and
software.
5. Thresholds can usually be set by clicking “Dark background”
option on the “Threshold window.” If large artifacts appear, set
threshold manually by adjusting the threshold bars so that the
objects are red; click set and then ok.
6. By setting 100-infinite, any artifacts smaller than 100 pixel2
(10 pixel 10 pixel) will be excluded. In images made with
68 Yao Fu et al.
20 objective, cell nuclei (either bovine or human) are larger

than 100 pixel2.
7. Setting the threshold for red image is tricky. Labeling efficiency
is calculated to estimate the accuracy of threshold setting.
Labeling efficiency should be similar to the ratio of chondro-
cytes used to establish the co-cultures particularly in early time
points (up to a few days maximum) after establishing the
culture.
8. It is possible to automatically analyze all images by running
customized plugins, which are written specifically for counting
cells in different colors, using macro language of ImageJ. Basic
knowledge about computer programing is required. Our plu-
gin is available upon request.
9. Reading of absorbance at 520 nm gives variations. Always do
triplicates for standards and samples.
10. Co-culture pellets usually contain a lot of extracellular matrix,
which makes it very difficult to extract RNA. After washing
with PBS, pellets must be snap frozen with liquid nitrogen and
smashed with pestle and mortar. Add lysis buffer to mortar to
collet total RNA. To get 1 μg of RNA, at least 3 pellets
(200,000 cells per pellet) are needed.
11. Take equal amount of cDNA from all samples in the same
experiment to make a stock solution of cDNA templates.
From the stock solution, make a series dilution: 1, 4,
16, 64, and 256 times. Run standards on the same plate
as Unknown (samples to be tested), then make standards
curves with Ct values in Bio-Rad iQ5 optical system software
(version 2.0).
12. The cylindrical hydrogels were prepared in 8 mm diameter
times 1.5 mm high molds.
13. At least three specimens were tested for each sample.
14. To distinguish the two individual human cells, male and female
donors were utilized in this pellet co-cultures. The male donors
present phenotypic markers, such as UTY or RPS4Y1, which
are gender-chromosome specific and can provide insight on the
RPKM level expressed from the individual donor in the co-cul-
ture. These markers are remarkably stably expressed in a male
cell and thus can be used to provide an estimate of the % of male
cells in a male-female co-culture.
15. A second polish step, depending on what is the final objective
of the study, is to remove the MIR and SNOR RNAs from the
data. MIR refers to the micro-RNA and SNOR refers to small
nuclear-RNA.
16. If no values are obtained with this threshold, a lower value

(1.5) is considered for the upregulation and a higher value
(0.75–0.8) for the downregulation. However, the sensitivity
and specificity are reduced which could lead to misleading
results/assumptions.
17. Theoretically, a random pair of human individuals has at least
one locus (within the 16 loci tested in the kit), which is infor-
mative, except for identical twins. Normally, 2–3 loci are infor-
mative to distinguish the hMSC and the hPC donor at the
DNA level.
References
1. Wu L, Cai X, Zhang S, Karperien M, Lin Y 9. Wu LG, Leijten J, van Blitterswijk C, Karperien

(2013) Regeneration of articular cartilage by M (2013) Fibroblast growth factor 1 is a
adipose tissue derived mesenchymal stem cells: mesenchymal stromal cell secreted factor sti-
perspectives from stem cell biology and molec- mulating proliferation of osteoarthritic chon-
ular medicine. J Cell Physiol 228:938–944 drocytes in co-culture. Stem Cells Dev
2. Leijten JCH, Georgi N, Wu L, van Blitterswijk 22:2356
CA, Karperien M (2013) Cell sources for artic- 10. Jeong SY, Kim DH, Ha J, Jin HJ, Kwon SJ,
ular cartilage repair strategies: shifting from Chang JW et al (2013) Thrombospondin-2-
monocultures to cocultures. Tissue Eng Part secreted by human umbilical cord blood-
B-Re 19:31–40 derived mesenchymal stem cells promotes
3. Moreira-Teixeira LS, Georgi N, Leijten J, Chondrogenic differentiation. Stem Cells
Wu L, Karperien M (2011) Cartilage tissue 31:2136
engineering. Endocr Dev 21:102–115 11. Jin R, Teixeira LM, Dijkstra PJ, Van Blitters-
4. Hendriks JAA, Miclea RL, Schotel R, de wijk CA, Karperien M, Feijen J (2010)
Bruijn E, Moroni L, Karperien M et al (2010) Enzymatically-crosslinked injectable hydrogels
Primary chondrocytes enhance cartilage tissue based on biomimetic dextran–hyaluronic acid
formation upon co-culture with a range of cell conjugates for cartilage tissue engineering.
types. Soft Matter 6:5080–5088 Biomaterials 31:3103–3113
5. Acharya C, Adesida A, Zajac P, Mumme M, 12. Teixeira LM, Leijten JC, Sobral J, Jin R, van
Riesle J, Martin I et al (2012) Enhanced chon- Apeldoorn AA, Feijen J et al (2012) High
drocyte proliferation and mesenchymal stromal throughput generated micro-aggregates of
cells chondrogenesis in coculture pellets medi- chondrocytes stimulate cartilage formation
ate improved cartilage formation. J Cell Physiol in vitro and in vivo. Eur Cell Mater
227:88–97 23:387–399
6. Wu L, Leijten JC, Georgi N, Post JN, van 13. Leijten J, Teixeira LM, Bolander J, Ji W,
Blitterswijk CA, Karperien M (2011) Trophic Vanspauwen B, Lammertyn J et al (2016)
effects of mesenchymal stem cells increase Bioinspired seeding of biomaterials using
chondrocyte proliferation and matrix forma- three dimensional microtissues induces chon-
tion. Tissue Eng Part A 17:1425–1436 drogenic stem cell differentiation and cartilage
7. Hildner F, Concaro S, Peterbauer A, formation under growth factor free conditions.
Wolbank S, Danzer M, Lindahl A et al (2009) Sci Rep 6:36011
Human adipose-derived stem cells contribute 14. Fu Y, Karbaat L, Wu L, Leijten J, Both SK,
to chondrogenesis in coculture with human Karperien M (2017) Trophic effects of mesen-
articular chondrocytes. Tissue Eng Part A chymal stem cells in tissue regeneration. Tissue
15:3961–3969 Eng Part B Rev 23:515–528
8. Wu L, Prins HJ, Helder MN, van Blitterswijk 15. Gnecchi M, Zhang Z, Ni A, Dzau VJ (2008)
CA, Karperien M (2012) Trophic effects of Paracrine mechanisms in adult stem cell signal-
mesenchymal stem cells in chondrocyte ing and therapy. Circ Res 103:1204–1219
co-cultures are independent of culture condi- 16. Dahlin RL, Kinard LA, Lam J, Needham CJ,
tions and cell sources. Tissue Eng Part A Lu S, Kasper FK et al (2014) Articular chon-
18:1542–1551 drocytes and mesenchymal stem cells seeded on
70 Yao Fu et al.
biodegradable scaffolds for the repair of carti- 19. Levine B, Kroemer G (2008) Autophagy in the
lage in a rat osteochondral defect model. Bio- pathogenesis of disease. Cell 132:27–42
materials 35:7460–7469 20. Hara T, Nakamura K, Matsui M, Yamamoto A,
17. de Windt TS, Vonk LA, Slaper-Cortenbach IC, Nakahara Y, Suzuki-Migishima R et al (2006)
van den Broek MP, Nizak R, van Rijen MH Suppression of basal autophagy in neural cells
et al (2017) Allogeneic mesenchymal stem causes neurodegenerative disease in mice.
cells stimulate cartilage regeneration and are Nature 441:885
safe for single-stage cartilage repair in humans 21. Abramoff M, Magelhaes P, Ram S (2004)
upon mixture with recycled autologous chon- Image processing with ImageJ. Biophoton Int
drons. Stem Cells 35:256–264 11:8
18. Elmore S (2007) Apoptosis: a review of pro-
grammed cell death. Toxicol Pathol
35:495–516
Chapter 6
Generation of Induced Pluripotent Stem Cells

David R. Deyle
Abstract
Induced pluripotent stem cells (iPSCs) are generated from somatic cells that have been reprogrammed by
the ectopic expression of defined embryonic transcription factors. This technology has provided investiga-
tors with a powerful tool for modeling disease and developing treatments for human disorders. This chapter
will provide the researcher with some background on iPSCs and details on how to produce
MEF-conditioned medium, prepare mitotically arrested mouse embryonic fibroblasts (MEFs), create
iPSCs using viral vectors, passage iPSCs, and cryopreserve iPSCs. The methods offered here have been
used in many laboratories around the world and the reader can initially follow these methods. However, not
all cell types are easily transduced using viral vectors and other methods of delivering the reprogramming
transcription factors may need to be tested.
Key words Induced pluripotent stem cells, Pluripotency, Reprogramming, iPSC isolation, Mouse
embryonic fibroblasts, Cryopreservation, iPSC passaging
1 Introduction
In 2006, a major advance in stem cell biology was reported by

Takahashi and Yamanaka [1]. They showed that after the introduc-
tion of a combination of different transcription factors by retroviral
transduction into mouse embryonic fibroblasts could be repro-
grammed into embryonic-like cells, designated induced pluripotent
stem cells. Within the next year, Takahashi et al. [2] and Yu et al. [3]
had adapted this technology for the successful reprogramming of
human somatic cells (skin fibroblasts). To be reprogrammed,
somatic cells must ectopically express four transcription factors
OCT4, SOX2, KLF4 or MYC, and NANOG or LIN28. Fully
reprogrammed human iPSCs express the embryonic antigens
SSEA3, TRA-1-60, TRA-1-81, DNMT3β, and REX1 [4] and
have the capacity to differentiate into the three germ layers, meso-
derm, ectoderm, and endoderm. Like embryonic stem cells, human
iPSCs can divide infinitely and be differentiated into all somatic cell
types.
71
72 David R. Deyle
Since the first iPSC experiment, many alternative methods of

generating iPSCs have been developed, including ones that use
integrating, excisable, and non-integrating viral vectors, as well as
nonviral systems. Each of these methods has its own advantages and
disadvantages, which are discussed in a review by Robinton and
Daley [5]. Fibroblasts have been the most common cell type to
reprogram, likely because these cells are readily available or can be
safely obtained by skin biopsy and are easy to culture in the labora-
tory. A number of different cell types, such as neural stem cells, liver
cells, keratinocytes, amniotic cells, adipose cells, bone marrow stro-
mal cells, and blood have also been reprogrammed into iPSCs using
a variety of different reprogramming methods. This chapter
describes a method of generating iPSC from fibroblasts or mesen-
chymal stem cells using two different forms of integrating viruses.
2 Materials
2.1 Reagents 1. Dispase.

2. Phosphate-buffered saline (PBS), without Ca++ or Mg++,
pH 7.4.
3. Dulbecco’s modified Eagle’s media/Nutrient Mixture F-12
(DMEM/F12) with GlutaMAX and sodium pyruvate.
4. Knockout Serum Replacement (Knockout SR) (Invitrogen).
5. Penicillin/Streptomycin (Pen/Strep).
6. Non-essential amino acids (NEAA).
7. β-Mercaptoethanol.
8. Recombinant human fibroblast growth factor basic (bFGF).
9. DMEM high glucose (Cellgro).
10. Heat-inactivated fetal bovine serum (FBS).
11. Trypsin.
12. Dimethyl sulfoxide.
13. Gelatin.
14. Rho-associated protein kinase (ROCK) inhibitor (Millipore)
(see Note 1).
15. Defined, serum-free cryopreservation medium (mFreSR)
(Stem Cell Technologies).
16. Trypan blue, 0.4%.
17. Lentiviral (LV) reprogramming vectors for transduction (see
Note 2):
(a) LV-OCT4.
(b) LV-SOX2.
(c) ©LV-LIN28.
(d) LV-NANOG.
Generation and Culture of Human iPSCs 73
18. Polybrene.
19. Foamy viral reprogramming vector (ΔΦ53MOSKMETNW)
(see Note 3).
20. CytoTune™-iPS Reprogramming System (see Note 4).
(a) CytoTune® 2.0 KOS (human KLF4, OCT4, and SOX2).
(b) CytoTune® 2.0 hc-MYC (human cMYC).
(c) ©CytoTune® 2.0 hKLF4 (human KLF4).
21. Passage 2 (P2), DR4 mouse embryonic fibroblasts (MEFs) (see
Note 5).
22. DMEM low glucose.
23. L-Glutamine.
2.2 Equipment 1. Laminar flow Biological Safety Cabinet.

2. Incubator, water jacketed and humidified with 5% CO2, main-
tained at 37 C.
3. Centrifuge with swinging bucket rotor, capable of holding
various tube sizes.
4. Inverted microscope with interference phase optics.
5. Dissecting microscope.
6. Glass hemocytometer.
7. Vacuum aspiration source with tubing and waste container.
8. 37 C Water bath.
9. Electric or manual pipet filler/dispenser.
10. 80 C Freezer.
11. Liquid nitrogen dewar storage unit.
12. Bunsen burner.
13. P20, P200, P1000 pipettes.
14. Irradiator.
15. Isopropanol freezing containers.
2.3 Special 1. Bio-Cool III controlled-rate freezer.

Equipment/Supplies 2. Sterile plugged Cassou straws (Veterinary concepts, #04170).
for Freezing iPSCs in
3. Plastic goblet cylindrical holder for loaded straws (Veterinary
Straws
concepts, #04910).
4. Aluminum cryocanes.
5. Methanol.
6. Very long forceps.
2.4 Supplies 1. Cotton plugged, sterile Pasteur pipettes 5 ¾ in and a bulb.

2. 5-, 10-, and 25-mL sterile serological pipets.
74 David R. Deyle
3. 50-mL Sterile conical centrifuge tubes.

4. 15-mL Sterile conical centrifuge tubes.
5. P20, P200, P1000 filtered, sterilized tips.
6. Nunc cell culture cryogenic tubes (cryovials).
7. 6-, 12-, and 24-well tissue culture plates.
8. 35-, 60-, and 100 mm tissue culture dishes.
9. Sterile 100-, 250-, 1000-mL bottles.
10. 1 mL Syringe.
11. Straw Adapter, 1/4 cc (Veterinary concepts).
2.5 Working 1. bFGF solution: 10 μg bFGF (see Note 6), 1 mL PBS, and
Solutions 20 μL Knockout SR. Dispense into 8–125 μL aliquots and
store FGF solution and store at -20 C.
2. Human embryonic stem cell medium (HESCM): 500 mL
DMEM/F12 with GlutaMAX and sodium pyruvate, 100 mL
Knockout SR (final conc. ~17%), 6 mL Pen/Strep (final conc.
~1%), 6 mL NEAA (final conc. ~1%), bFGF (final conc. 2 ng/
mL) (see Note 6), 0.6 mL 0.1 M β-mercaptoethanol (0.35 mL
β-mercaptoethanol in 50 mL H2O). Pass medium through
0.22 μm filter into sterile bottle. The medium can be stored
at 4 C for 14 days. Before an experiment, warm the medium to
37 C.
3. Dispase: 100 mL PBS, 100 mg of dispase (final conc. 1 mg/
mL). Pass dispase through 0.22 μm filter into sterile bottle.
Dispase can be stored at 4 C for 7 days. Before an experiment,
warm dispase to room temperature.
4. MEF culture medium: 500 mL DMEM high glucose, 50 mL
FBS (final conc. ~10%), and 5 mL Pen/Strep (final conc. ~1%).
The medium can be stored at 4 C for 14 days. Before an
experiment, warm the medium to 37 C.
5. Mesenchymal Stem Cell Medium (MSCM): 500 mL DMEM
low glucose, 50 mL FBS (final conc. ~10%), 5 mL L-glutamine,
and 5 mL Pen/Strep (final conc. ~1%). The medium can be
stored at 4 C for 14 days. Before an experiment, warm the
medium to 37 C.
6. Gelatin stock: 100 gm of gelatin in 1 liter of double-distilled
water and autoclave. 0.5% gelatin stock solution can be stored
at room temperature.
7. Gelatin working solution: 100 mL of 0.5% gelatin stock solu-
tion plus 400 mL sterile water. 0.1% Gelatin working solution
can be stored at room temperature.
8. MEF freeze medium (FM): 50% MEF culture medium, 40%
FBS, and 10% DMSO.
9. Straw freezing medium (FMs): 70% HESCM, 20% Knockout

SR, and 10% DMSO.
10. Cryovial freezing medium (FMc): mFreSR plus 10 μM ROCK
inhibitor.
3 Methods
3.1 Expansion of This protocol is based on C. Xu et al. [6]. We use passage

Passage 2 DR4 MEFs 3 (P3) MEFs for the production of MEF-conditioned medium. It
is best to expand P2 DR4 MEFs (see Note 7) and freeze cells for
later use.
1. Thaw vial of P2 DR4 MEFs by placing vial of cells in 37 C
water bath without submersing the cap. Swirl gently.
2. When no crystals remain, wipe vial with 70% ethanol and place
in cell culture hood.
3. Use a sterile pipet to transfer cells to 50 mL conical tube
containing 5 mL MEF culture medium.
4. Centrifuge at 800 g for 5 min.
5. Aspirate medium and resuspend cells in 10 mL of MEF culture
medium.
6. Add 1 mL of cell suspension to 100 mm dish for a total of
10 dishes.
7. Add 9 mL of MEF culture medium and gently rock dish to
disperse cells evenly.
8. Place cells in 37 C incubator.
9. Aspirate medium and add 10 mL new MEF culture medium
every 3 days.
10. When cells are confluent, aspirate medium and wash cells with
4 mL PBS.
11. Add 2 mL 0.05% trypsin to each dish and incubate at 37 C for
3–5 min.
12. Collect using MEF culture medium and pool into a 50 mL
conical tube.
14. As cells are spinning, make fresh FM.
15. Aspirate medium off cells and resuspend cells in 10 mL of FM.
16. Place 1 mL of cell suspension in cryovial and put in room
temperature isopropanol freezing containers and place con-
tainers in -80 C overnight.
76 David R. Deyle
17. The next day store cryogenic tubes in liquid nitrogen. Alter-
nately, MEF-conditioned medium can be produced immedi-
ately following the expansion of MEFs.
18. When cells are confluent, aspirate medium and wash 8 dishes
with 4 mL PBS. Save remaining dishes for generation of
MEF-conditioned medium.
3–5 min.
20. Collect using MEF culture medium and pool the 8 dishes into
a 50 mL conical tube.
22. As cells are spinning, make FM.
24. Add 1 mL of cells to cryogenic tubes, place in Styrofoam
container, and freeze at 80 C.
25. The next day store cryogenic tubes in liquid nitrogen.
3.2 Production of This protocol should yield approximately 3.36 L of

MEF-Conditioned MEF-conditioned medium. Collected medium is stored at
Medium -20 C. When MEF-conditioned medium is thawed for use, add
fresh bFGF to 2 ng/mL and pass through a 22 μm filter.
1. Wash 2 confluent 100 mm dishes containing P3 MEFs with
4 mL PBS each.
3–5 min.
3. Collect using MEF culture medium and pool the 2 dishes into a
50 mL conical tube.
medium.
6. Plate 1 mL of cells in a 100 mm dish 12 dishes.
7. Add 9 mL MEF culture medium and gently rock dish to
8. Change medium on all dishes after 2 days.
9. When cells are confluent (usually 3 or 4 days), wash each dish
with 4 mL PBS.
10. Add 0.05% trypsin and incubate at 37 C for 3–5 min.
11. Collect all dishes with 12 mL MEF culture medium and place
in 50 mL conical tube.
12. Add an additional 24 mL MEF culture medium to the 50 mL
conical tube and centrifuge at 800 g for 5 min.
13. Aspirate medium and resuspend in 24 mL MEF culture

medium.
14. Irradiate cells with 40 Gy (see Note 8).
15. Count cells with hemocytometer and trypan blue.
16. Plate 4.0–4.5 106 cells per 100 mm dish (typically
12 dishes).
17. Add MEF culture medium to final volume of 10 mL.
18. The next day cells should be confluent, rinse each dish with
4 mL PBS.
19. Add 40 mL of HESCM to each dish.
20. Next day harvest medium from all dishes into 2 250 mL
bottles.
21. Add fresh 40 mL HESCM to each dish.
22. Repeat steps 20 and 21 for total of 7 days.
3.3 Preparation of 1. Thaw vial of P2 DR4 MEFs by placing vial of cells in 37 C

Irradiated MEF Stocks water bath without submersing the cap. Swirl gently.
in Cryovials 2. When no crystals remain, wipe vial with 70% ethanol and place
in cell culture hood.
3. Use a sterile pipet to transfer cells to 50 mL conical tube
containing 5 mL MEF culture medium.
medium.
6. Add 1 mL of cell suspension to 100 mm dish for a total of
10 dishes.
7. Add 9 mL of MEF culture medium and gently rock dish to
8. Place cells in 37 C incubator.
9. Aspirate medium and add 10 mL new MEF culture medium
every 3 days.
10. When cells are confluent, aspirate medium and wash 8 dishes
with 4 mL PBS each.
11. Add 2 mL 0.05% trypsin to each washed dish and incubate at
37 C for 3–5 min.
conical tube.
14. Resuspend cells in 42 mL MEF culture medium.
15. Plate 1 mL of cells into 100 mm dish 42 dishes.
78 David R. Deyle
16. Add 9 mL MEF culture medium to each dish and gently rock
dish to disperse cells evenly.
17. Remaining 2 100 mm dishes aspirate medium and wash cells
with 4 mL PBS.
3–5 min.
conical tube.
21. As cells are spinning, make fresh FM.
tainers in 80 C overnight.
25. Aspirate medium on cultured dishes and add 10 mL new MEF
culture medium every 3 days.
26. When cells are confluent, aspirate medium and wash cells with
4 mL PBS. (Collect 6 dishes at a time).
3–5 min.
28. Use 8 mL of MEF culture medium to collect each dish and
pool 3 dishes into a 50 mL conical tube.
29. Use 10 mL MEF culture medium to rinse 3 100 mm dishes
and add to 50 mL conical tube.
30. Centrifuge both tubes at 800 g for 5 min.
31. Aspirate medium off 50 mL conical tubes and collect next
6 dishes.
32. Repeat steps 19 and 24 until all dishes have been harvested.
33. Resuspend both cell pellets in 10 mL MEF culture medium
and combine into one tube.
34. Count cells with hemocytometer and trypan blue.
35. Irradiate cells with 40 Gy (see Note 8).
36. Centrifuge at 800 g for 5 min and aspirate medium.
37. Prepare fresh FM.
38. Resuspend cells in freezing medium at 4 106 cells per mL (see
Notes 9 and 10).
tainers in 80 C overnight.
Table 1
Gelatin per culture dish size
Culture dish/plate size 0.1% gelatin added

100 mm 7 mL
60 mm 3.5 mL
35 mm 2 mL
6 well 1 mL per well
12 well 0.5 mL per well
24 well 0.5 mL per well
3.4 Preparing 1. In sterile hood, add autoclaved 0.1% gelatin to tissue culture
Irradiated-MEF Plates dishes as indicated in Table 1.
for iPSCs 2. Place dishes containing gelatin in incubator at 37 C for a
minimum of 5 min or maximum of several hours.
3. Thaw MEFs (see Note 11).
4. Transfer cells from cryovial to 15 mL conical tube with 5 mL
pipet.
5. Add 4 mL MEF culture medium.
6. Centrifuge at 800 g for 5 min and aspirate medium.
7. Cells are resuspended as desired and plated (see Note 12).
8. Allow MEFs to attach for minimum 8 h, best to allow MEFs to
attach overnight.
9. Plated MEFs can be used for up to 7 days.
3.5 Generating iPSCs This protocol is based on two previously published papers for
from Fibroblasts/ lentiviral and foamy viral vector-mediated reprogramming [3, 7].
Mesenchymal Stem
1. Seed human fibroblasts/MSCs at 2 105 cells per well in
Cells Using Lenti and 6-well plate in MEF culture medium/MSCM.
Foamy Viral
2. Incubate at 37 C overnight.
Reprogramming
Vectors 3. Aspirate off medium and add 2 mL fresh medium to each well.
4. For lentiviral infections, add 4 μg/mL polybrene to each well
and infect each well with LV vectors at corresponding multi-
plicity of infection (MOI) (see Table 2) (see Note 13).
10. For excisable FV reprogramming vector transductions, infect
each well with foamy viral reprogramming vector (ΔΦ53M
OSKMETNW) at MOI 1.
5. Prepare 2 MEF-coated dishes (see Subheading 3.4) for each
infected well.
6. The next day, wash transduced wells with 2 mL PBS.
80 David R. Deyle
Table 2
Lentiviral reprogramming vectors
Reprogramming vector MOI

LV-OCT4 1
LV-SOX2 0.015
LV-LIN28 0.5
LV-NANOG 1
7. Add 0.5 mL 0.05% trypsin to each well and incubate at 37 C

for 3–5 min.
8. Collect each with 4.5 mL MEF culture medium/MSCM and
place in 50 mL conical tube.
10. Aspirate medium off conical tubes and MEF dishes.
11. Resuspend transduced cells in 20 mL of the appropriate
growth medium.
12. Plate 10 mL of viral transduced cells onto MEFs, two dishes of
transduced cells per each well containing MEFs.
13. The next day wash each dish with 4 mL PBS twice.
14. Change medium on all dishes to HESCM.
15. Change medium daily with HESCM for the next 8 days.
16. After 8 days, change medium to MEF-conditioned medium
and replace daily until clones have been picked.
17. From days 14–20, scan dishes and monitor changing cell mor-
phology using inverted microscope.
18. Between 21 and 30 days after transduction, colonies should be
ready to be isolated.
19. One day prior to picking colonies, plate MEFs into 24-well
plates. Prepare one well of MEFs for each iPS colony to be
picked.
20. Prepare sterile Pasteur pipette hooked cutting tool using Bun-
sen burner by:
(a) Heat pipet 1/3 distance from end of tip and make a
90 degree bend in pipette (Fig. 1).
(b) Allow it to cool and apply more heat near bend. Then pull
glass apart to form pointed pipette end (see Note 14).
(c) ©Remove excess glass by tapping with another pipet to
generate cutting tool.
21. Wipe dissecting microscope with 70% ethanol and place into
tissue culture hood (see Note 15).
Fig. 1 Creating the cutting tool from a Pasteur Pipet. To create the cutting tool to pick iPSC colonies, heat
Pasteur pipette over Bunsen burner and make a 90 degree bend approximately 1/3 distance from the tip (a, b)
and allow to cool. Carefully holding the end, place pipette into flame again distal to the 90 degree bend, and as
the glass softens, quickly pull bend and end away from each other (c, d). Remove excess glass by tapping with
another pipet to generate cutting tool (e, f)
22. While visualizing colony under dissecting microscope, use

pipette to gently cut around colony to mechanically dissociate
it from MEFs. It is best to cut colony into pieces (see Note 16).
23. Once colony has been freed, use P1000 pipettor to collect
colony pieces and transfer to a well of a well plate.
24. Repeat steps 22 and 23 until all desired colonies have been
picked (see Note 17).
25. Allow outgrowth of the picked colony (usually 3–5 days) and
passage colony into new well of a 24-well plate using dispase
(see Subheading 3.7).
26. Expand clones by serial passage cells into larger culture dishes
to meet laboratory needs.
3.6 Generating iPSCs 1. Seed human fibroblasts/MSCs at 2 105 cells per well in
from Fibroblasts/ 6-well plate in MEF culture medium/MSCM.
Mesenchymal Stem 2. Incubate at 37 C overnight.
Cells Using
3. Aspirate off medium and add 2 mL fresh medium to each well.
Nonintegrating Sendai
Viral Reprogramming
4. Transduce cells using the CytoTune® 2.0 Sendai reprogram-
ming vectors corresponding multiplicity of infection (MOI)
Vectors
(see Table 3).
5. The next day, aspirate off medium and add 2 mL fresh MEF
culture medium/MSCM to each well.
6. Two days later, aspirate off medium and add 2 mL fresh MEF
82 David R. Deyle
Table 3
Sendai reprogramming vectors
Reprogramming vector MOI

®
CytoTune 2.0 KOS 5
®
CytoTune 2.0 hc-MYC 5
CytoTune® 2.0 hKLF4 3
7. Two days later, aspirate off medium and add 2 mL fresh MEF
8. Prepare 2 MEF-coated dishes (see Subheading 3.4) for each
infected well.
9. The next day, wash transduced wells with 2 mL PBS.
10. Add 0.5 mL 0.05% trypsin to each well and incubate at 37 C
for 3–5 min.
11. Collect each with 4.5 mL MEF culture medium/MSCM and
place in 50 mL conical tube.
13. Aspirate medium off conical tubes and MEF dishes blue.
14. Resuspend transduced cells in 20 mL of fresh MEF culture
medium/MSCM.
15. Count cells with hemocytometer and trypan.
16. Plate 50,000 viral transduced cells onto one 10 cm MEF dish
and 100,000 viral transduced cells on the other per each well
containing MEFs.
17. The next day change medium on all dishes to HESCM.
18. Change medium daily with HESCM for the next 8 days.
19. From days 18 to 21, scan dishes and monitor changing cell
morphology using inverted microscope.
20. Pick colonies as described in Subheading 3.5 (Steps 19–26).
3.7 Passaging iPSCs 1. Aspirate medium off iPSCs.

with Dispase 2. Add dispase (0.5 mL for wells in 24- and 12-well plates, 1 mL
for wells in a 6-well plate, 2 mL for a 6 cm dish, 4 mL for a
10 cm dish).
3. Incubate at 37 C until colony edges start to curl (2–5 min).
4. Aspirate dispase and gently add appropriate amount of wash
medium to rinse cells (see Note 18). Be careful not to dislodge
colonies (see Note 19).
5. Aspirate wash and add 1–4 mL of wash medium.
6. Detach colonies from dish using Pasteur pipet or P1000 for

small wells (12- and 24-well plates) and a 5 mL pipet can be
used for larger wells and dishes. Try to keep the colonies as
large as possible.
7. Spin 200 g for 5 min at room temperature.
8. Aspirate gently (pellets can be loose, so you can use a 5 mL
pipette here).
9. Resuspend in desired volume of HESM. Aliquot cells to MEF
dishes containing the rest of the necessary HESM medium
volume. Place in incubator.
3.8 Freezing iPSCs in This protocol is based on the published report by Ware et al. [8].
Straws
1. Prepare Bio-Cool III controlled-rate freezer by filling tank with
methanol, turning it On, setting the start temperature to –
10 C, setting the speed to 1 degree/minute, setting the end
temperature to –33 C, setting the hold minutes to 600 or
greater, and setting the alarm to 0 C. Press RUN to start
cooling to –10 C (see green light under “start”).
2. Put canes with buckets in freezer to cool.
3. Prepare fresh FMs (5 mL).
4. Collect 100 mm dish containing 70–80% confluent, healthy
appearing iPSCs (1 100 mm dish ¼ 5 straws).
5. Aspirate medium off iPSCs.
6. Add 4 mL of dispase to each dish.
8. Aspirate dispase and gently add 4 mL of wash medium to rinse
cells (see Note 18). Be careful not to dislodge colonies (see
Note 19).
9. Aspirate wash and add 4 mL of wash medium.
10. Detach colonies from dish using a 5 mL pipet. Try to keep the
colonies as large as possible.
12. Aspirate gently. Pellets may be loose, so you can use a 5 mL
pipette here.
13. Resuspend cells in 1.1 mL FMs.
14. Attach straw adaptor to syringe.
15. Attach sterile, plugged straw to the adaptor.
16. Draw up about 1/2 inch FMs, being careful to avoid cells.
17. Draw up about 1/8–1/4 inch of an air bubble.
18. Draw up most of straw with cells, about 0.2 mL (mix cells first),
but leave 1/2 inch to go.
84 David R. Deyle
19. Pull straw out of medium, then draw up last 1/2 inch to plug.
Leave 1/2 inch air at bottom.
20. Place back in sterile package.
21. Repeat steps 17–21 with the other 4 straws.
22. Briefly flame bottom of straw and use gloved hand to press
melted straw together to seal.
23. Repeat step 23 on top of straw.
24. Return to sterile package and bring to Biocool III freezer.
25. The time between cell resuspension to placing straws contain-
ing iPSCs into Biocool III freezer should be at least 15 min.
26. Bring liquid N2 tank near Bio-Cool III controlled-rate freezer.
Make sure canes are ready with buckets.
27. Place straws in buckets at –10 C for about 1 min.
28. Cool long forceps in N2, then grab straw(s) with forceps above
air bubble, and allow ice crystal to form above the bubble.
Then place straws back in bucket in Biocool III freezer.
29. Once all straws set, press RUN again on Biocool III freezer to
start controlled-rate freezing.
30. When frozen, load straws into liquid N2 tanks.
3.9 Freezing iPSCs in 1. Thaw the required amount of mFreSR, store on ice.
Cryovials 2. Prepare FMc, store on ice.
3. Label cryovials.
4. Prechill isopropanol freezing containers at 4 C.
5. Collect 100 dish of iPSCs (1 100 mm dish ¼ 5–6 cryovials).
6. Aspirate medium from iPSCs.
7. Add 4 mL of dispase to each dish.
9. Aspirate dispase and gently add 4 mL of wash medium to rinse
cells (see Note 18). Be careful not to dislodge colonies (see
Note 19).
10. Aspirate and add 4 mL of wash medium.
11. Detach colonies from dish using a 5 mL pipet. Try to keep the
colonies as large as possible.
13. Aspirate gently (pellets may be loose, so you can use a 5 mL
pipette here).
14. Gently resuspend the cells in 6 mL cold FMc. Take care to leave
the cell aggregates larger than would normally be done for
passaging.
15. Transfer 1 mL of cells to each labeled cryovial.
16. Place vials into an isopropanol freezing container and place the
container in -80 C freezer overnight.
3.10 Thawing iPSCs 1. Fill 15 mL conical tube with 5 mL HESM or wash solution.
from Straw 2. Take straw from liquid N2 and place in a beaker of room
temperature water (dH2O, tap).
3. Wipe straw with 70% ethanol.
4. Cut bottom of straw with sterile scissors over waste bin.
5. Cut top of straw over 15 mL tube, while holding little bit
with plug.
6. Let cells drain into tube.
7. Spin 200 g for 5 min at room temperature, resuspend as
desired and plate onto MEF-coated dish.
3.11 Thawing iPSCs 1. Rapidly thaw cells in a 37 C water bath by gently shaking the
from Cryovial cryovial continuously until only a small frozen pellet remains.
2. Remove cryovial from the water bath and wipe with 70%
ethanol.
3. Transfer the cells to a 15 mL conical tube.
4. Add 1.5 mL of warm HESCM dropwise to the tube over
5 min, gently mixing as the medium is added.
5. Add an additional 1.5 mL of warm HESCM dropwise to the
tube over 1–2 min.
6. Spin 200 g for 5 min at room temperature, resuspend as
desired, and plate onto MEF-coated dish.
4 Notes
1. ROCK inhibitor has been shown to block apoptosis of disso-

ciated cultured human embryonic stem cells and thereby
increasing survival [9].
2. Lentiviral vectors can be purchased from Sigma. The plasmids
are also available from Addgene and each lentiviral vector can
be produced as described by Yu et al. [3].
3. The reprogramming plasmid pΔΦ53MOSKMETNW is avail-
able from Dr. David Russell’s laboratory and the foamy viral
vectorΔΦ53MOSKMETNW can be produced as described in
Gharwan et al. [10].
4. The CTS™ CytoTune™-iPS 2.1 Sendai Reprogramming Kit
can be purchased from ThermoFisher [11, 12].
86 David R. Deyle
5. DR4 MEFs can be purchased from ATCC and Capital Bios-

ciences or they can be isolated from E13 DR4 embryos as
described in Manipulating the Mouse Embryo [13].
6. bFGF cannot be refrozen. Once thawed it is best to use the
entire aliquot and discard remainder.
7. DR4 MEFs are resistant to neomycin, hygromycin, puromycin,
and 6-thioguanine.
8. MEFs can also be mitotically arrested with Mitomycin C by
aspirating off the medium on a confluent MEF 100 mm dish.
Add 6 mL of MEF culture medium with 10 μg/mL mitomycin
C and incubate at 37 C for 3 h. Remove medium and wash
3 times with 5 mL PBS. MEFs can be harvested with trypsin
and cryopreserved as described in Subheading 3.3.
9. MEFs cannot be stored once thawed. To minimize waste,
MEFs can be frozen at higher or lower concentrations for
different needs; for example, lower concentration vials can be
used when only a small number of plates or wells are needed for
experiments.
10. Alternatively, MEFs can also be frozen in straws. Resuspend
cells in FM at 2 107 cells per mL. Attach straw adaptor to
1 mL syringe and then fix sterile straw into straw adaptor with
plug near adaptor. Draw up a small amount of FM into straw
with no cells (~1/2 inch) and then draw in 1/8 inch air to
make a bubble spacer. Draw up 0.2 mL of the cell suspension
into straw by pulling on syringe plunger leaving a small amount
of air at the bottom of the straw. Briefly flame bottom of straw
and use gloved hand to press melted straw together to seal and
then repeat on top of straw. Place straws in Styrofoam container
and place in freeze at 80 C overnight and then store in liquid
nitrogen using Plastic goblet cylindrical holder and aluminum
cryocanes.
11. Thaw cryovial in 37 C water bath or straw in 15 mL conical
tube containing room temperature water. For straws, add 5 mL
MEF culture medium into 15 mL tube, wipe straw with 70%
ethanol, and then cut bottom of straw with sterile scissors.
While holding straw in 15 mL tube, use scissors to cut the
top of straw to release cells and then flick straw to drain all fluid.
12. One cryovial (4 106 cells) seeds 2–3100 mm dishes, 4–6
60 mm dishes, or 2 plates (6-well, 12-well, etc.).
13. For an MOI of 1 add 100,000 viral particles to 100,000 cells. If
titer is 1 108 particles/mL, add 1 μL of virus to the cells.
14. Each cutting pipet can be flamed and reused for isolating other
iPSC clones. However, after 3 to 4 uses the tip comes dull, so it
is best to make multiple cutting pipets prior to picking
colonies.
15. A dedicated hood with microscope inside is best, but a 70%

ethanol clean microscope can be used. Not all laboratories will
be equipped with glass that will accommodate the microscope
eyepieces. The dissecting microscope can be placed inside a
standard tissue culture hood with the front guard open, but
be careful not to contaminate cells while picking colonies.
16. Cut a circle around colony and then make an X through the
colony. The colony will be further broken to chunks when
being picked up with P1000 pipet.
17. Not all clones will be fully reprogrammed, which is necessary
for expansion. It is best to pick as many colonies as possible to
maximize the number of isolated clones.
18. DMEM/F12 can be used as wash medium, but HESCM is
better. HESCM medium that is greater than 14 days old and
cannot be used as culture medium can be used as a wash
solution.
19. Dispase is not inactivated by medium, so dilution is required.
By carefully rinsing dishes, there is one less spin step.
References
1. Takahashi K, Yamanaka S (2006) Induction of 8. Ware CB, Nelson AM, Blau CA (2005)
pluripotent stem cells from mouse embryonic Controlled-rate freezing of human ES cells.
and adult fibroblast cultures by defined factors. BioTechniques 38(6):879–880. 882-3
Cell 126(4):663–676 9. Watanabe K et al (2007) A ROCK inhibitor
2. Takahashi K et al (2007) Induction of pluripo- permits survival of dissociated human embry-
tent stem cells from adult human fibroblasts by onic stem cells. Nat Biotechnol 25(6):681–686
defined factors. Cell 131(5):861–872 10. Gharwan H et al (2007) Transduction of
3. Yu J et al (2007) Induced pluripotent stem cell human embryonic stem cells by foamy virus
lines derived from human somatic cells. Science vectors. Mol Ther 15(10):1827–1833
318(5858):1917–1920 11. Fusaki N et al (2009) Efficient induction of
4. Chan EM et al (2009) Live cell imaging distin- transgene-free human pluripotent stem cells
guishes bona fide human iPS cells from par- using a vector based on Sendai virus, an RNA
tially reprogrammed cells. Nat Biotechnol 27 virus that does not integrate into the host
(11):1033–1037 genome. Proc Jpn Acad Ser B Phys Biol Sci
5. Robinton DA, Daley GQ (2012) The promise 85(8):348–362
of induced pluripotent stem cells in research 12. Seki T et al (2010) Generation of induced plu-
and therapy. Nature 481(7381):295–305 ripotent stem cells from human terminally dif-
6. Xu C et al (2001) Feeder-free growth of undif- ferentiated circulating T cells. Cell Stem Cell 7
ferentiated human embryonic stem cells. Nat (1):11–14
Biotechnol 19(10):971–974 13. Hogan B (1994) Manipulating the mouse
7. Deyle DR et al (2012) Normal collagen and embryo: a laboratory manual. Cold Spring
bone production by gene-targeted human Harbor Laboratory Press, Plainview, NY
osteogenesis imperfecta iPSCs. Mol Ther 20
(1):204–213
Chapter 7
Specimen Preparation for Single-Cell Sequencing Analysis

of Skeletal Cells
Shawon Debnath and Matthew B. Greenblatt
Abstract
Recent work emphasizes that bone comprises numerous mesenchymal cell types each with different
biologic functions, and deconvoluting the functions of these cells requires technical approaches with
single-cell resolution, such as single-cell RNA sequencing (scRNA-seq). A critical step in conducting a
successful single-cell sequencing study of bone is generation of a single-cell suspension of skeletal cells while
preserving cell viability. Here we describe a method to prepare single-cell suspensions from skeletal tissue in
preparation for single-cell sequencing studies. Also included are optional steps for fluorescence-activated
cell sorting (FACS) of skeletal cells, which allows subsequent scRNA-seq studies to be focused on specific
populations of interest.
Key words Single-cell sequencing, Bone, Osteoblasts, FACS
1 Introduction
Recent work increasingly indicates that skeletal mesenchyme con-

tains a rich set of several distinct cell types, including LEPR+
stromal cells/Cxcl12-abundant reticular cells, distinct growth
plate-resident stem cells, Acta2-positive cells, CD146+ cells, and
periosteal and calvarial stem cells marked by CTSK-cre. As multiple
members of each of these and other molecularly defined subsets
appear to contribute to the pools of osteoblasts and other tradi-
tionally defined skeletal cell types, methods allowing for classifica-
tion of cells into these and other definitions at the single-cell level,
namely single-cell RNA sequencing and fluoresce activated cell
sorting (FACS)/flow cytometry, are needed to understand bone
metabolism in terms of these molecularly defined cell subsets. A key
step in applying most of these single-cell technologies is efficient
generation of a single-cell suspension while minimizing the impact
of the isolation procedure on the cell viability and transcriptional
profile. With practice and optimization, generation of these single-
cell suspensions of primary skeletal mesenchymal cells can be
89
90 Shawon Debnath and Matthew B. Greenblatt
robust, as evidenced by a recent and growing slate of studies

successfully generating and corroborating insights from scRNA-
seq studies in bone [1–6]. Here we provide a protocol suitable for
isolation of most types of skeletal mesenchymal cells from a variety
of human or mouse tissues, which will produce a single-cell suspen-
sion suitable for downstream analysis by flow cytometry/FACS
and/or single-cell sequencing.
The importance of optimizing these methods for cell isolation
has been highlighted by work on muscle suggesting that the enzy-
matic digestion step is a key point where the in vivo transcriptional
profile of cells is disrupted [7]. Thus, optimization of cell isolation
procedures must aim to minimize the transcriptional impact of the
isolation procedure at the same time that cell yield is maximized.
While direct assessment of the transcriptional impact of the isola-
tion procedure itself is difficult, requiring specialized techniques,
maximization of cell viability (>90%) and minimization of the
procedure duration, particularly enzymatic digestion steps, are eas-
ily accessible basic proxies for the overall impact of the procedure
on cells.
We here focus solely on specimen preparation for scRNA-seq
studies; however, equally important is the selection of scRNA-seq
methodology and subsequent methods for data analysis. Discussion
of considerations in selection of these methods in the context of
skeletal biology have been recently published [8]. Each scRNA-seq
methodology has different requirements and limitations in terms of
numbers of input cells that must be considered when planning
specimen preparation. In particular, methods such as CEL-Seq2
and SMART-seq will require that specimens undergo index sorting
prior to analysis, aiming to place single cells in as many wells as
possible in a microtiter plate. In contrast, most other methods
utilize cells in suspension as the input, and for these methods
FACS is optional, though it can have utility in helping to remove
cell doublets, acellular debris from the specimen in addition to
allowing the scRNA-seq study to focus on a specific population of
interest through the use of cell surface or genetically encoded
markers. Thus, optional methods for both index sorting and
FACS are included here.
Therefore, the following sections focuses on preparation of the
bone samples for scRNA seq studies. However, this protocol has
also been successfully used to sort different mesenchymal cell
populations by flow cytometry to determine their cellular hierarchy
and physiologic contribution in bone using various types of trans-
plantation studies.
Isolation of Skeletal Cells for scRNA-seq 91
2 Materials
The digestion buffer and DNase I solution must be prepared fresh

each time prior to use. Tissue suspension media, FACS buffer, and
wash buffer should be made regularly and can be stored up to
2 months at 4 C. The FACS staining and blocking buffers and
primary and secondary antibody dilutions must all be prepared
fresh each time prior to use. Any unused reagents should be dis-
carded according to disposal regulations of the individual’s
institution.
2.1 Reagents 1. Tissue digestion buffer: Collagenase P (Roche, cat.

11213857001) and Dispase II (Roche, cat. 04942078001)
should always be stored at 4 C when not in use. To prepare
the digestion buffer, weigh required amount of Collagenase P
and Dispase II using a scale capable of detecting 0.1 mg. Dis-
solve both reagents in α-MEM media containing 2% fetal
bovine serum (FBS) using 15 or 30 mL falcon tubes. It is
important to prepare the digestion mix in media with 2%
FBS, since presence of serum greatly improves the viability of
mesenchymal cells including the rare cell types which otherwise
can be lost in media without serum. Gently swirl the reagents in
the falcon tube until dissolved. Do not dissolve reagents by
vortexing. The final digestion buffer should contain Collage-
nase P at 1 mg/mL and Dispase II at 2 mg/mL. Store the
digestion buffer on ice before adding it to the tissue sample.
2. DNase I solution: Add DNase I at 2 units/mL in α-MEM basal
media without serum. Presence of serum will interfere with
DNase activity. Mix gently and do not vortex. Keep the
DNase I solution on ice before use.
3. Tissue suspension media: Prepare tissue suspension media in
bulk using α-MEM basal media and 2% FBS and store it at 4 C.
4. FACS buffer: Prepare FACS buffer in bulk using distilled
phosphate-buffered saline (DPBS), 2% FBS, 1 mM ethylene
diamine tetra acetic acid (EDTA). Store at 4 C.
5. Blocking buffer: Prepare fresh blocking buffer by adding rat
anti-mouse CD16/CD32 (clone 2.4G2, BD bioscience) anti-
body to FACS buffer at a dilution of 1:50. Keep on ice.
6. Primary antibody solution: Prepare fresh primary antibody
solution by adding the fluorophore-conjugated antibodies at
a dilution of 1:100 in brilliant stain buffer (BD bioscience).
Keep on ice and protect it from light (see Note 1).
7. Secondary antibody solution: If an unconjugated primary anti-
body is used, a fluorophore-conjugated secondary antibody
directed against the primary antibody needs to be used. Prepare
fresh secondary antibody solution at a dilution of 1:500 in

brilliant stain buffer. Keep on ice and protect it from light (see
Note 1).
8. DAPI solution: Prepare DAPI (40 , 6-diamidino-2-phenylin-
dole) stock solution in PBS at a concentration of 1 μg/mL.
Keep it cold and protect it from light.
3 Methods
Carry out all procedures at room temperature until specified.
3.1 Tissue Sample 1. Euthanize the animals according to institutional guidelines.

Preparation Make a cut on the back of the mouse with fine scissor and
pull out the skin exposing the mouse femur and tibia. Carefully
remove the skin exposing the mouse knee joint and hip joint/
pelvic girdle. Carefully pull the mouse femur from the pelvic
girdle separating the femur head. It is easy to separate mouse
femur head above 3 weeks of age. However, it may be difficult
to do so in young mice up to 2 weeks of age. In that case, we
first recommend to make a cut with fine scissors taking a
portion of the pelvic joint and the femur head and then care-
fully separating out the femur head with fine forceps. After
isolating the entire femur, strip away as much muscle as possi-
ble. Keep the mouse femur in tissue suspension media in a
12-well plate. Please refer to Fig. 1 (see Notes 2–5).
2. Pick up the femur with forceps and place it on a glass slab such
as the one used in casting SDS-PAGE gels. Use two razor
blades to chop the femur into small pieces until a finely minced
tissue is generated. Thorough mincing of femur section is very
critical to ensure good yield of mesenchymal cells. While
mincing, make sure the chopped tissue is never dry, since it
will affect cell viability. It should always be moist and
Fig. 1 Digestion of mouse femurs. Isolated 15-day-old mouse femur suspended in media in a 12-well plate (a).
Mouse femurs placed on a glass slab to be chopped by razor blades (b). Texture of a thoroughly minced mouse
femur sample (c). Slurry texture of an appropriately digested mouse femur that can be easily pulled up by
pipetting (d)
Table 1
Suggested digestion conditions based on mouse age and specimen site
Digestion DNasel Falcon Shaking

buffer solution tube Temerature speed Time
Source volume (mL) volume (mL) size (mL) ( C) (rpm) (min)
Mouse femur 1 1 15 37 200 15–20
(7-day-old)
Mouse femur 1 1 15 37 200 30–40
(15-day-old)
Mouse femur 1.5 1.5 15 37 200 50–60
(1-month-old)
Mouse femur 2 2 30 37 225 60
(2-month-old)
Mouse femur 3 3 30 37 225 60
(6-month-old)
Bone organoid in 1 1 15 37 200 15
mammary fat pad
surrounded by tiny amount of suspension medium. Please refer

to Fig. 1 (see Notes 6 and 7).
3. Transfer the minced tissue homogenate into a 15 mL falcon
tube. Add 1 mL of tissue digestion buffer to the tube. We
recommend to add 1 mL of digestion buffer to digest two
femurs within 2 weeks of age. Quickly spin the tube for a
minute at 524 g to pull down any minced tissue that is
attached to the surface. Incubate the tube at 37 C, gently
shaking at 200 rpm in a shaker. The digestion process will
approximately take 30–40 min for a 2-week-old mouse femur.
Resuspend the tissue digest several times with 1 mL pipet to
ensure that the homogenate is slurry. At this point, the tissue
digestion step is complete (see Notes 8, 9 and Table 1).
4. Add 10 mL of tissue suspension media to the falcon tube.
Centrifuge the tube at 524 g for 10 min at 4 C.
5. Discard the media by gently turning the falcon tube upside
down without disturbing the pellet. We recommend not to use
vacuum to remove excess liquid, since it can accidentally dis-
lodge the pellet. The media can also be discarded using manual
pipetting method.
6. Resuspend the tissue pellet thoroughly in 1 mL of DNase I
solution. Incubate the tube at 37 C, gently shaking at 200 rpm
inside the shaker for 10 min. Please refer to Table 1.
7. Add 8–9 mL of tissue suspension media and thoroughly sus-

pend the tissue homogenate by vigorously pipetting the sample
for several times. This step is highly critical. Thorough suspen-
sion will cause physical dissociation of tissue, and thereby min-
imize cell loss eventually leading to better single-cell
suspension.
8. Once the tissue is thoroughly suspended, filter it through a
70 μm nylon mesh into a 30 mL falcon tube. Make sure, you
filter the tissue suspension using a 5 mL or 10 mL pipet.
During the filtration, the pipetting force should be strong
enough to push the tissue filtrate through the nylon mesh
without causing any splash. This will cause physical dissociation
of tissue giving rise to single-cell suspension. It will need some
practice to obtain the suitable pipetting force. To increase cell
recovery, collect the remaining filtrate attached to the other
side of the filter using a 1 mL pipet.
9. Transfer the filtered cell suspension from 30 mL falcon tube to
a 15 mL falcon tube. We have observed that using 15 mL
falcon tube for the later stages of sample preparation for
FACS staining minimizes cell loss during the washing steps.
10. Centrifuge the tubes at 524 g for 10 min at 4 C. Discard
the liquid by gently turning the falcon tube upside down
without disturbing the pellet.
11. Gently add 5–10 mL of ice-cold FACS buffer to the pellet. Do
not suspend the cell pellet. This will minimize cell loss.
12. Centrifuge the tubes at 524 g for 5 min at 4 C. Discard the
liquid as much as possible without disturbing the pellet. Keep
the tube on ice. The samples are now ready for FACS staining.
3.2 FACS Staining Skeletal mesenchymal cells can be subjected to single-cell sequenc-
ing with or without sorting. However, sorting provides efficient
elimination of contaminating cell population such as hematopoietic
and endothelial cells and additionally facilitates focus on specific
mesenchymal populations if desired. The following steps need to be
carried out on ice (see Note 10).
1. Add 200 μL of freshly prepared blocking buffer to the samples.
Suspend the pellet thoroughly by using 1 mL pipet. Do not
vortex. Incubate the tubes on ice for 30 min (see Note 11).
2. Add 200 μL of freshly prepared primary antibody solution to
the tubes. Suspend the pellet thoroughly by using 1 mL pipet.
Do not vortex. Incubate the tubes on ice for 45 min to 1 h.
Protect the tubes from light. After adding equal volumes of
blocking buffer and primary antibody solution, the final work-
ing solution of the blocking buffer is 1:100 and primary anti-
body solution is 1:200 (see Note 12).
Fig. 2 Schematic representation of isolation of mesenchymal cells by FACS. DAPI-negative live cells are
analyzed for cell size and singlets are isolated from the doublets. Mesenchymal cells (Ter119, CD45,
CD31) can be sorted by exclusion of mature red blood cells (Ter119+), hematopoietic cells (CD45+), and
endothelial cells (CD31+)
3. For washing, add 10 mL of ice-cold FACS buffer to the tubes.

Centrifuge the tubes at 524 g for 10 min at 4 C.
4. Gently decant the supernatant without resuspending or other-
wise disturbing the cell pellet. This will minimize cell loss.
5. Repeat steps 3 and 4 twice. The centrifugation time can be cut
to 5 min during the second and third wash.
6. Add 200 μL of freshly prepared secondary antibody solution to
the tubes. Thoroughly suspend the pellet by pipetting. Do not
vortex. Incubate the tubes on ice for 20–30 min. Protect the
tubes from light.
7. Wash the samples twice following steps 3 and 4. Do not
resuspend the cell pellet.
8. Add 500 μL of ice-cold FACS buffer to the tubes. Suspend the
pellet thoroughly and transfer it to appropriate FACS tubes.
Filter the cell suspension through a 40 μm nylon mesh during
the transfer. This will remove any undesirable cell aggregates
that may clog the nozzle of the sorter.
9. Add DAPI to the cell suspension at a dilution of 1:100 and
gently mix by tapping the tubes. Do not vortex. The samples
are ready for sorting (see Note 13).
10. A template strategy for sorting mesenchymal cells is shown in
Fig. 2. Generally, DAPI negative live cells are chosen. Forward
versus side scatter gating (FSC-A versus SSC-A) is used to
choose the cell population of interest based on size and cyto-
plasmic complexity. FSC-H versus FSC-W and SSC-H versus
SSC-W gates allow exclusion of doublets. Mesenchymal cells
can be sorted by eliminating Ter119+ erythroid cells, CD45+
hematopoietic cells, and CD31+ endothelial cells. After gating
on mesenchymal cells, additional cell surface or genetic mar-
kers can then be applied to specify subpopulations of mesen-
chymal cells if desired.
3.3 Sample During FACS or after bone tissue digestion without further FACS
Collection isolation, samples must be collected or suspended using a method
for Single-Cell that both preserves cell viability and is compatible with the down-
Sequencing stream single-cell sequencing methodology. The approach will vary
depending on whether a droplet based (e.g., Drop-seq, 10 Geno-
mics) or index sorting based (e.g., SMART-seq, MARS-seq,
CEL-seq) methodology is subsequently employed [8].
1. For droplet based assays using Drop Seq or 10 Genomics
platforms, mesenchymal cells are sorted in bulk to conduct
single-cell sequencing. Sorted mesenchymal cells should be
collected in collection media using 1.5 mL Eppendorf tubes.
Collection media should be free of serum, and basal media
containing 0.5% bovine serum albumin is suitable for many
applications. Sorting cells into PBS should be avoided, since
it will cause significant loss of sensitive cell types such as mes-
enchymal stem cells. We recommended to follow the guidelines
for the specific sequencing platform utilized.
2. Index-based techniques (such as CEL-seq, Smart-seq) require
sorting single cells into individual wells of 96-well or 384-well
plates. In many cases, the collection plate will be provided by
the sequencing facility and are pre-loaded with unique bar-
codes, primers, and lysis buffer. Following sorting, the plates
are sealed and snap frozen on dry ice before submitting the
samples for sequencing. Validation experiments must be con-
ducted to determine that sorting produces as few empty wells
as possible while avoiding doublets. This validation can take the
form of sorting a test cell line with direct microscopic visuali-
zation of cell capture after colonies are grown from the sorted
cells.
4 Notes
1. We recommend that the primary antibody solution should be

made at a 1:100 dilution. However, the appropriate dilution
may need to be adjusted depending on the antibody concen-
tration and the cell number. This also holds true for secondary
antibodies.
2. The procedure described here is suitable for generation of
single-cell suspensions suitable for FACS or single-cell
sequencing for a broad range of skeletal cells, including LEPR
+ stromal cells/Cxcl12-abudnant reticular (CAR) cells, osteo-
calcin+ osteoblasts, articular and other growth plate resident
chondrocytes. Additionally, several endothelial populations are
also present and small numbers of CTSK+ osteoclast-lineage
cells can be isolated. However, this protocol is notably not
suitable for isolation of osteocytes. It is possible that other

populations may not be well represented or may require addi-
tional optimization for efficient isolation.
3. The endpoint of this procedure will be a single-cell suspension
containing large amounts of hematopoietic cells alongside skel-
etal mesenchymal cells and small amounts of populations,
including endothelial cells. For many applications, removal of
hematopoietic cells is desirable, and this can be achieved
through CD45-based gating strategies when FACS is
employed after isolation. Early hematopoietic progenitors, par-
ticularly of the erythroid lineage, may express weak CD45, so
inclusion of an additional negative selection marker such as
glycophorin A is suggested for efficient depletion. In cases
where it is not feasible or undesirable to include a FACS step
after isolation, magnetic bead-based depletion offers an alter-
native approach. In some cases, pre-depletion of hematopoietic
cells may be employed prior to FACS to increase the frequency
of cell types of interest and thereby decrease the instrument
time required. However, magnetic bead based depletion stra-
tegies must be carefully validated for nonspecific depletion of
populations of interest sticking to the depletion column. For
this reason, we typically avoid magnetic bead depletion. Alter-
natively, marrow flushing or centrifugation to remove marrow
cells can be employed, but similarly runs the risk of depleting
mesenchymal populations of interest or introducing an addi-
tional source of variation in population frequencies. A strength
of single-cell sequencing studies is that they are relatively
robust in the face of contamination with unwanted cell types,
so purification need only be robust enough that the popula-
tions of interest are adequately represented and to avoid incur-
ring unnecessary expense in sequencing irrelevant cell types.
4. When dealing with vertebral bone, it is important to thor-
oughly separate any neural tissue from the specimen, as neural
support cells can be challenging to distinguish from mesenchy-
mal cells in some forms of analysis due to their sharing many
markers with mesenchymal cells. For instance, both Schwann
cells and subsets of skeletal mesenchymal cells express
CD200 [9].
5. Microdissection or physical fractionation of the specimen prior
to cell isolation can offer a simple but powerful method to later
annotate populations identified based on anatomic location.
6. For extremely substantial specimens, such as large samples of
human or large animal bone, we have found success by using a
chisel to perform the initial physical disassociation of the speci-
men before progressing to chopping specimens with razor
blades. Use eye protection with a chisel or other forceful
method of physical disassociation. Avoid specimen drying dur-

ing this step by frequent irrigation with medium.
7. The two most important parameters in terms of optimizing cell
yield and viability are the physical dissociation of tissue prior to
digestion and the conditions of the enzymatic digestion itself.
Thus, it is critical that physical dissociation be extremely thor-
ough, ideally almost the point that the specimen takes on the
appearance of coarse sand, before proceeding to digestion. For
the physical disassociation, we most commonly use chopping
with a razor blade performed on a glass plate, such as used for
casting SDS-PAGE gels. Mincing with very fine surgical scis-
sors while the specimen is in medium in a 6-well plate is an
acceptable alternative. We generally find that mouse femurs
until 1 month of age can be thoroughly minced using razor
blades. However, efficient mincing of femurs from older mice
using razor blades can be difficult to achieve. In that case, the
mouse femurs can be crushed using a mortar and pestle. Make
sure that the crushed tissue sample is always kept moist. Do not
add digestion buffer to the samples while mincing or crushing.
Addition of digestion buffer after the tissue is physically dis-
sociated will improve cell recovery. We emphasize that mincing
tissue samples is greatly preferred over crushing wherever feasi-
ble, as mincing increases cell yield and viability, ultimately
helping to preserve rare cell types such as endothelial cells and
mesenchymal stem cells.
8. To increase cell recovery, certain parameters may need optimi-
zation. This includes (1) the volume of digestion buffer and
DNase I solution, (2) the diameter of the falcon tubes, and
(3) the conditions and duration of tissue digestion. It is critical
to adjust the volume of digestion buffer and DNase I solution
according to mouse age. We recommend adding 1 mL of
digestion buffer and, subsequently, 1 mL of DNase I solution
per mouse femur sample between 0 and 15 days of age. The
volume of digestion buffer and DNase I solution needs to be
increased for samples from older mice. Please refer to Table 1
for additional information. To ensure optimum digestion, we
do not recommend combining samples from multiple mice.
Each mouse sample should be kept separate and individually
subjected to digestion. Further, the diameter of the falcon tube
is important to optimize the tissue digestion process. For
example, to digest 8-week-old mouse femurs, 30 mL falcon
tubes should be used. Increase in surface area will ensure better
tissue digestion, resulting in a better yield of mesenchymal
cells. Please refer to Table 1.
9. Additionally, it is imperative to optimize the tissue digestion
process. Make sure that the tissue homogenate is not over- or
under-digested. We recommend to digest the tissue at 37 C in
an enclosed shaker shaking at 200 rpm. Digesting tissue sam-

ples at room temperature instead of using a temperature-
stabilized system will lead to poor cell recovery. Shaking the
tubes at a modest speed of 200 rpm will increase the enzymatic
digestion process by maximizing the contact between minced
tissue and the digestion buffer. Speeds should not exceed
225 rpm, as this can lead to enzymatic degradation or nega-
tively impact cell viability. Duration of tissue digestion is
another important parameter to consider. Generally, the overall
digestion process can range between 30 and 60 min. Digestion
of 2-week-old mouse femurs takes about 30 min, whereas
2-month-old mouse femurs will take an hour to digest. How-
ever, there can be exceptions. For example, digestion of a bone
ossicle/organoid grown in the mouse mammary fat pad only
needs 15 min, and 7-day-old mouse femurs can be digested in
15–20 min. We emphasize that there is no fixed time of tissue
digestion, and therefore this duration should be optimized for
every type of tissue preparation. We recommend monitoring
the digestion process at regular intervals, looking to see if the
lumps of tissue have started to dissolve in the digestion buffer.
As tissue starts to dissolve, the color of the tissue changes from
reddish brown to a more whitish hue. The final step of tissue
digestion can be confirmed by gently pipetting the homoge-
nate, which can now be easily pipetted up and down and will
resuspend in a more homogenous slurry texture. For conve-
nience, please refer to Fig. 1. Also, please refer to the guidelines
in Table 1 that shows several examples of optimized tissue
digestion conditions.
10. It is challenging to successfully generate a FACS panel contain-
ing conjugated antibodies for more than 6 colors. Optimiza-
tion can be time consuming and requires thorough
understanding of the excitation and emission spectra of each
fluorophore and the FACS instrument. First, one will need to
choose appropriate cell surface markers to identify the popula-
tion of interest. While designing the FACS panel, the fluoro-
phore conjugates used with each antibody require
optimization. The conjugation chemistry varies with different
dyes against a specific antibody and vice versa. For example, the
BV421 conjugated mouse CD200 antibody do not display the
same spectral plot when compared with APC R700 conjugated
mouse CD200 antibody in the presence of other colors. There-
fore, it is easy to miss an important cell type that is rarely
present in the cell population. We therefore, recommend to
initially generate single color stains for each antibody using the
brightest color wherever possible to confirm the presence or
absence of the cells. After initial analysis, the conjugated anti-
body should be tested in the presence of other conjugated
antibodies to check for any discrepancies versus the single
stained sample. Antibodies should also be validated by using

proper isotype controls. To set up initial compensation beads
should be used. Fluorescence minus one (FMO) controls
should be used to set up for additional compensation and to
assess background stain for each color. For each antibody, gates
should generally be drawn as determined by FMO controls to
separate positive and negative populations.
11. Based on size, the cell pellet can be suspended in blocking
buffer and, subsequently, primary antibody solution in a vol-
ume ranging from 200 μL to 500 μL. For a medium-sized cell
pellet (5 106 cells), we recommend a volume of 200 μL. If
time is not a constraint, we encourage initially incubating the
cells with blocking buffer prior to addition of the primary
antibody solution. However, blocking buffer and primary anti-
body solution can be added simultaneously to speed up the
procedure when required.
12. Primary and secondary antibody solutions should be freshly
prepared at the time of use. The emission spectra of the fluor-
ophores can change when kept in dilution buffer for more than
10 h. Tandem fluorophores such as PE-Cy7 and BUV737 are
the ones most impacted by excessive time in the dilution buffer.
All fluorophore conjugated antibodies must be initially diluted
in brilliant stain buffer before adding to cells. Conjugated
antibodies should never be diluted in FACS buffer.
13. The nozzle size plays an important role to determine the
efficiency of sorting and the viability of sorted cells. For exam-
ple, freshly isolated mesenchymal cells should be sorted with
70 μm nozzle, whereas when cultured, cells may enlarge and
need to be sorted with a 100 μm nozzle. Validation experi-
ments must be conducted to optimize sorting conditions. This
includes checking cells for viability and function after sorting.
References
1. Chan CKF, Gulati GS, Sinha R et al (2018) 6. Debnath S, Yallowitz AR, McCormick J et al
Identification of the human skeletal stem cell. (2018) Discovery of a periosteal stem cell med-
Cell 175:43–56.e21 iating intramembranous bone formation.
2. Mizuhashi K, Nagata M, Matsushita Y et al Nature 562:133–139
(2019) Growth plate borderline chondrocytes 7. Machado L, Esteves de Lima J, Fabre O, et al
behave as transient mesenchymal precursor (2017) In situ fixation redefines quiescence and
cells. J Bone Miner Res 34(8):1387–1392 early activation of skeletal muscle stem cells. Cell
3. Chan CKF, Seo EY, Chen JY et al (2015) Iden- Rep 21(7):1982–1993
tification and specification of the mouse skeletal 8. Greenblatt MB, Ono N, Ayturk UM et al (2019)
stem cell. Cell 160:285–298 The unmixing problem: a guide to applying
4. Tikhonova AN, Dolgalev I, Hu H et al (2019) single-cell rna sequencing to bone. J Bone
The bone marrow microenvironment at single- Miner Res 34:1207–1219
cell resolution. Nature 569:222–228 9. Chang C-Y, Lee Y-H, Jiang-Shieh Y-F et al
5. Baryawno N, Przybylski D, Kowalczyk MS et al (2011) Novel distribution of cluster of differen-
(2019) A cellular taxonomy of the bone marrow tiation 200 adhesion molecule in glial cells of the
stroma in homeostasis and leukemia. Cell 177 peripheral nervous system of rats and its modu-
(7):1915–1932.e16 lation after nerve injury. Neuroscience
183:32–46
Chapter 8
Mapping 5-Hydroxymethylcytosine (5hmC) Modifications

in Skeletal Tissues Using High-Throughput Sequencing
Fiorella Carla Grandi and Nidhi Bhutani
Abstract
Cytosine modifications can alter the epigenetic landscape of a cell, affecting the binding of transcription
factors, chromatin organizing complexes, and ultimately affecting gene expression and cell fate.
5-Hydroxymethylcytosine (5hmC) modifications are generated by the Ten-eleven-translocation (TET)
family of enzymes, TET 1, 2, and 3, through the oxidation of methylated cytosines (5mC). The TET
family is capable of further oxidizing 5hmC to 5fC and 5caC, leading to eventual DNA demethylation.
However, 5hmC marks can also exist stably in DNA. Stable 5hmC is enriched in the gene bodies of
activated genes in multiple tissues, as well as associated with regulatory regions such as enhancers. Altera-
tions to 5hmC patterns have now been found in multiple diseases including osteoarthritis. Here, we
describe a method to map 5hmC modifications by next-generation sequencing using a technique based
on the selective modification and enrichment of the 5hmC mark. We additionally provide a bioinformatic
analysis pipeline to interpret the resulting data.
Key words Hydroxymethylcytosine, 5hmC, Pull-down, Next-generation sequencing, Bioinformat-

ics, Gene expression regulation
1 Introduction
DNA methylation has long been an epigenetic mark associated with

gene silencing. The family of cytosine modifications has now been
expanded to include 5hmC, 5fC, and 5caC [1]. Cytosine bases in
the CpG context can be modified to 5-methylcytosine (5mC) by
the DNMT family of proteins and subsequently oxidized by the
TET family of enzymes into 5-hydroxymethylcytosine (5hmC),
and two further oxidized productions 5-formylcytosine (5fC) and
5-carboxylcytosine (5caC) [1]. Alterations in DNA modifications
are associated with cell fate changes during development and tissue
regeneration, while changes in these modifications often mark
disease states. Multiple skeletal diseases including osteoarthritis,
osteoporosis, and rheumatoid arthritis have demonstrated changes
in several DNA modifications [2–4]. We have previously shown that
101
102 Fiorella Carla Grandi and Nidhi Bhutani
5hmC accumulates during chondrogenesis and is altered during

osteoarthritis [5–7]. In addition, 5hmC is increasingly appreciated
as a mark of transcriptionally active genes and can be useful in
identifying a particular cell type and the epigenetic landscape that
defines it [8]. However, the challenge with profiling 5hmC is that it
typically represents only 0.05–0.6% of the total DNA in the cell [9]
and is highly labile.
Below, we detail a protocol for mapping 5hmC by next-gener-
ation sequencing. The protocol is based on the selective targeting
of the 5hmC mark via treatment with the T4 Phage
β-glucosyltransferase that selectively glucosylates the hydroxyl
group [10]. The glucose group, which contains an azide, is then
labeled using biotin, and this biotinylated-glucose modification can
then be selectively pulled-down using commercially available strep-
tavidin beads. The enriched DNA can be used for downstream
genome-wide or targeted 5hmC assessment. The protocol consists
of four major sections: (1) shearing of the genomic DNA and
β-glucosyltransferase treatment for glucosylation of 5hmC moi-
eties, (2) enrichment of DNA fragments, (3) library preparation
and sequencing, and (4) data analyses.
2 Materials
2.1 DNA Shearing 1. Qubit dsDNA BR Assay Kit (Invitrogen) and Qubit
and Glucosylation Fluorometer.
Reaction 2. Covaris S220 Ultrafocused sonicator.
3. Covaris milliTUBE 1 mL AFA Fiber.
4. DNA cleanup columns.
5. PCR Tubes.
6. Active Motif HydroxymethylCollector Kit (see Note 1) or:
(a) T4 Phage β-glucosyltransferase.
(b) UDP-Azide-Glucose.
(c) Biotin solution.
2.2 DNA Fragment 1. Active Motif HydroxymethylCollector Kit or:

Enrichment (a) Dynabeads M-280 Streptavidin.
(b) Bead washing solution: 10 mM Tris–HCl (pH 7.5), 1 mM
EDTA, 2 M NaCl.
2. DNA LoBind Tube 1.5 mL (Eppendorf).
3. TE Buffer (10 mM Tris–HCl, 1 mM EDTA, pH 8).
Mapping 5hmC Modifications using High-Throughput Sequencing 103
2.3 DNA Library 1. NEB Ultra II DNA Library Prep with Sample Purification
Preparation Beads (NEB #E7103).
and Next-Generation 2. NEBNext Multiplex Oligos for Illumina.
Sequencing
3 Methods
3.1 DNA Shearing 1. Extract DNA using standard methods for your skeletal tissue of
and Glucosylation choice. Quantify DNA (see Note 2).
Reaction 2. Shear DNA: Place 2 μg of DNA in a Covaris tube and shear
DNA to 200–300 base pair fragments. Conditions should be
optimized for each sample type based on concentration and
purity. Ideally, you want 2 μg of DNA in a volume of less than
35.5 μL (see Notes 3–5).
3. Set aside 10% of your fragmented input DNA as your input
control for sequencing. This will be used to calculate the back-
ground for your assay. Freeze this sample at 20 C until
Subheading 3.3.
4. Glycosylation reaction: set up the reactions in a 200 μL PCR
tube as in Table 1. Control reactions: A negative control should
be set up without the addition of the β-glucosyltransferase. In
the skeletal system, ATDC5 cells at day 15 of differentiation
can serve as a positive control, as these cells are known to have
high levels of 5hmC [5].
5. Incubate the reaction at 37 C for 1 h in a PCR thermocycler.
6. Spin down the PCR tube to collect all the volume.
7. Add 20 μL of biotin solution (Active Motif) to each reaction.
8. Incubate the reaction for 1 h at 37 C.
9. Purify biotinylated DNA using standard DNA cleanup column.
Make sure that you are using a column that is compatible with
your input DNA amount.
3.2 DNA Fragment 1. Resuspend the streptavidin beads by vortexing and allow to
Enrichment come to room temperature.
2. Set up each enrichment reaction in a 1.5 mL DNA LoBind tube
(Eppendorf) a 200 μL according to Table 2. Incubate the
reaction for 2–3 h at room temperature with end-to-end rota-
tion (see Note 6).
3. After incubation, spin down the tube to collect the volume at
the bottom and place the tube in a magnetic stand. Allow the
beads to pellet on the side of the tube. This may take 1–3 min.
Remove the supernatant (see Note 7).
Table 1
β-glucosyltransferase
Reagent Volume
Buffer 5 μL
10 mM DTT 5 μL
UDP-Azide-glucose 2.5 μL
B-Gluctranserase 2 μL
Sheared DNA ______μL
dH2O To volume
Total volume 50 μL
Table 2
DNA fragment enrichment
Reagents Volume
Streptavidin beads 25 μL
Binding buffer 25 μL
Purified biotinylation reaction 50 μL
Total volume 100 μL
4. Wash beads gently 5 times with 200 μL binding buffer (Active

Motif). For each wash, remove the tube from the stand, pipet
up and down gently 2–3 times making sure all beads are fully
washed from the sides of the tube, and place the tube back on
the magnetic holder. Allow beads to pellet again and
repeat wash.
5. After the final wash step, remove the buffer and resuspend the
washed beads with 30–50 μL TE buffer. Incubate for 30 min at
room temperature with end-to-end rotation. After incubation,
remove supernatant. This supernatant contains the 5hmC
enriched DNA.
6. Purify the DNA using commercially available DNA cleanup kit
for small fragments.
7. Quantify the amount of DNA using Qubit and Bioanalyzer (see
Note 8). Store samples at 20 C until you are ready to
perform sequencing.
8. Optional: locus-specific qPCR to verify pull-down (see Note 9).
(a) Design primers spanning a 90–150 base-pair region on
DNA on your target gene of interest.
(b) Perform a standard qPCR reaction using the Sybr Green

system, using 1–5 ng of DNA in each reaction. Delta Ct
calculations should be performed using the input as the
reference control.
(c) Use the negative control to check for the specificity of
enrichment.
3.3 DNA Library 1. Using the quantification of your immunoprecipitated DNA,

Preparation follow the manufacturer’s instructions for library preparation
and Next-Generation (see Note 10). If you are planning on pooling several samples
Sequencing into one lane, make sure to use barcodes. We find that it is
possible to pool 6–8 samples per Illumina HiSeq 4000 lane.
2. Sequence DNA fragments.
3.4 Data Analysis 1. The quality of your fastq files can be analyzed using
(See Note 11) FASTQC [12].
2. Trim the low quality ends identified using FASTQC using
Trimmomatic [13]. After trimming, run the fastq files back
through FASTQC to ensure that all low quality reads have
been removed.
3. Map reads to the correct species genome using HISAT2
[14]. For mouse and human, we suggest using mm10 and
hg19, respectively, as they have the most updated additional
epigenetic information from the ENCODE and ROADMAP
projects [15–17].
4. 5hmC peaks can be called using MACS [18], with the input
sample as a control for the background to the biotin pull-
down.
5. Differentially hydroxymethylated regions (dHMRs) can be
called using diffREPS [19]. In order to do this, you must first
convert your bam files to bed files using BEDOPS [20]
bamtobed tool.
6. Annotation of peaks or dHMRs can be done using BEDOPS
closest-feature tool.
7. Motif analysis of the DNA nearby your 5hmC can be useful to
determine if specific transcription factors are either being tar-
geted or being used to guide the TET enzymes. This analysis
can be done using HOMER [21], which will predict both
known and de novo motifs.
8. Visualize your 5hmC reads on gene bodies or other genes of
interest using ngsplot [22].
9. Additional downstream analysis options (see Note 12):
(a) Intersect genes with dHMRs or peaks with genes differ-
entially expressed in RNA-sequencing studies.
(b) Perform network or pathway analysis of genes with 5hmC

peaks or dHMRs using IPA (QIAGEN), Enrichr [23, 24],
or STRING [25].
(c) Intersect peaks with ATAC-seq [26] or other ChIP-seq
data sets. Peak intersection can be done easily using bed
files. We recommend Galaxy [27] as a good online tool for
performing these types of analyses quickly.
(d) Intersect peaks or dHMRs with enhancers or superenhan-
cers, which can both be defined by a combination of
histone modifications [28].
4 Notes
1. We recommend using the standardized Active Motif™ kit, as

we find that these reagents give the most reproducible results.
It is possible to buy all the component parts separately, but we
find this has lower efficiency in the overall enrichment of
5hmC. In this protocol, we present our alterations to the
manufacturer’s recommended guidelines, as well as the analysis
pipeline we use after sequencing.
2. We recommend using the Qubit or Bioanalyzer assays as they
are more accurate than Nanodrop based estimations, especially
for DNA that is extracted from tissue samples, which may be
more fragmented or contaminated with trace phenol.
3. There are multiple ways to create DNA fragments. We recom-
mend using sonication, as this results in minimal loss of DNA.
However, DNA can also be sheared using restriction enzymes.
Several platforms exist for DNA sonication, including the
Diagenode and Covaris. We have had the best performance
with systems that perform sonication in chilled water baths, as
this prevents degradation of the fragile 5hmC groups.
4. The input amount of DNA depends on the abundance of
5hmC in your particular skeletal tissue of interest, and the
integrity of the sample. In our experience, depending on the
degradation of DNA, you may require up to 5 μg of DNA for
this assay. In the case of late-stage DMM models of OA, we
found that the limited amount of DNA we could extract from
the joint was more suitable for targeted analysis of 5hmC by
using the Reduced Representation Hydroxymethylation
Profiling [11].
5. To find the ideal shearing conditions, we recommend setting
up a reaction with your target DNA amount and doing a time-
course of shearing time. You can check the shearing either by
running a gel or by Bioanalyzer. We recommend the Bioanaly-
zer trace as it will give you a better indication of the fragment
profile and concentration. A good starting point for the Cov-

aris machine is: duty cycle: 10%, intensity 5, cycles per burst
200, and duration 180 s.
6. The reaction can also be allowed to incubate overnight at 4 C.
7. This supernatant can be analyzed to see what the percent
capture is if desired. To do so, the optional qPCR step with
the supernatant after DNA extraction as one of the samples.
You should see enrichment of your target regions in the pull-
down samples compared to the supernatant.
8. Nanodrop will not give an accurate quantification of the
fragmented DNA.
9. When performing this enrichment on unknown samples, it can
often be useful to check a known gene for 5hmC enrichment to
check that enrichment went well. For example, if you are using
Day 15 differentiated ATDC5 cells as a positive control,
Col2a1 or Acan will be enriched for 5hmC [5].
10. A variety of different DNA library preparation kits can be used.
We recommend NEB’s Ultra II DNA Library Prep with Sam-
ple Purification Beads.
11. A variety of different programs can be used to analyze sequenc-
ing data. Here we detail the pipeline we use for our sequencing
analysis.
12. Make sure that you are selecting only peaks or dHMRS with an
adjusted p-value of 0.05 or lower for downstream analysis.
Acknowledgments
F.C.G. is supported by the NSF Graduate Research Fellowship.

Funds from the Department of Orthopedic Surgery, NIH/NIAMS
R03 (R03AR066356) and R01 (R01AR070865) to NB supported
this work.
References
1. Kohli RM, Zhang Y (2013) TET enzymes, 4. Ramos YFM, Meulenbelt I (2017) The role of
TDG and the dynamics of DNA demethyla- epigenetics in osteoarthritis: current perspec-
tion. Nature 502:472–479. https://doi.org/ tive. Curr Opin Rheumatol 29:119–129.
10.1038/nature12750 https://doi.org/10.1097/BOR.
2. Klein K, Gay S (2015) Epigenetics in rheuma- 0000000000000355
toid arthritis. Curr Opin Rheumatol 27:76–82. 5. Taylor SE, Li YH, Smeriglio P et al (2015)
https://doi.org/10.1097/BOR. Stable 5-hydroxymethylcytosine (5hmC)
0000000000000128 acquisition marks gene activation during chon-
3. Letarouilly J-G, Broux O, Clabaut A (2019) drogenic differentiation. J Bone Miner Res Off
New insights into the epigenetics of osteopo- J Am Soc Bone Miner Res 31(3):524–534.
rosis. Genomics 111:793–798. https://doi. https://doi.org/10.1002/jbmr.2711
org/10.1016/j.ygeno.2018.05.001 6. Taylor SEB, Li YH, Wong WH, Bhutani N
(2015) Genome-wide mapping of DNA
hydroxymethylation in osteoarthritic chondro- Biol 9:R137. https://doi.org/10.1186/gb-

cytes. Arthritis Rheumatol 67:2129–2140. 2008-9-9-r137
https://doi.org/10.1002/art.39179 19. Shen L, Shao N-Y, Liu X et al (2013) diffReps:
7. Taylor SEB, Smeriglio P, Dhulipala L et al detecting differential chromatin modification
(2014) A global increase in sites from ChIP-seq data with biological repli-
5-hydroxymethylcytosine levels marks osteoar- cates. PLoS One 8:e65598. https://doi.org/
thritic chondrocytes. Arthritis Rheumatol 10.1371/journal.pone.0065598
Hoboken NJ 66:90–100. https://doi.org/10. 20. Neph S, Kuehn MS, Reynolds AP et al (2012)
1002/art.38200 BEDOPS: high-performance genomic feature
8. Ecsedi S, Rodrı́guez-Aguilera JR, Hernandez- operations. Bioinformatics 28:1919–1920.
Vargas H (2018) 5-Hydroxymethylcytosine https://doi.org/10.1093/bioinformatics/
(5hmC), or how to identify your favorite cell. bts277
Epigenomes 2:3. https://doi.org/10.3390/ 21. Heinz S, Benner C, Spann N et al (2010) Sim-
epigenomes2010003 ple combinations of lineage-determining tran-
9. Skvortsova K, Zotenko E, Luu P-L et al (2017) scription factors prime cis-regulatory elements
Comprehensive evaluation of genome-wide required for macrophage and B cell identities.
5-hydroxymethylcytosine profiling approaches Mol Cell 38:576–589. https://doi.org/10.
in human DNA. Epigenetics Chromatin 10:16. 1016/j.molcel.2010.05.004
https://doi.org/10.1186/s13072-017-0123-7 22. Shen L, Shao N, Liu X, Nestler E (2014) ngs.
10. Song C-X, Szulwach KE, Fu Y et al (2011) plot: Quick mining and visualization of next-
Selective chemical labeling reveals the genome- generation sequencing data by integrating
wide distribution of 5-hydroxymethylcytosine. genomic databases. BMC Genomics 15:284.
Nat Biotechnol 29:68–72. 101038/nbt.1732 https://doi.org/10.1186/1471-2164-15-
11. Petterson A, Chung TH, Tan D et al (2014) 284
RRHP: a tag-based approach for 23. Chen EY, Tan CM, Kou Y et al (2013) Enrichr:
5-hydroxymethylcytosine mapping at single- interactive and collaborative HTML5 gene list
site resolution. Genome Biol 15:456. https:// enrichment analysis tool. BMC Bioinformatics
doi.org/10.1186/s13059-014-0456-5 14:128. https://doi.org/10.1186/1471-
12. Babraham Bioinformatics - FastQC A Quality 2105-14-128
Control tool for High Throughput Sequence 24. Kuleshov MV, Jones MR, Rouillard AD et al
Data. https://www.bioinformatics.babraham. (2016) Enrichr: a comprehensive gene set
ac.uk/projects/fastqc/. Accessed 29 Sep 2019 enrichment analysis web server 2016 update.
13. Bolger AM, Lohse M, Usadel B (2014) Trim- Nucleic Acids Res 44:W90–W97. https://doi.
momatic: a flexible trimmer for Illumina org/10.1093/nar/gkw377
sequence data. Bioinforma Oxf Engl 25. Szklarczyk D, Morris JH, Cook H et al (2017)
30:2114–2120. https://doi.org/10.1093/bio The STRING database in 2017: quality-
informatics/btu170 controlled protein-protein association net-
14. Kim D, Langmead B, Salzberg SL (2015) works, made broadly accessible. Nucleic Acids
HISAT: a fast spliced aligner with low memory Res 45:D362–D368. https://doi.org/10.
requirements. Nat Methods 12:357–360. 1093/nar/gkw937
https://doi.org/10.1038/nmeth.3317 26. Buenrostro JD, Giresi PG, Zaba LC et al
15. The ENCODE Project Consortium (2012) An (2013) Transposition of native chromatin for
integrated encyclopedia of DNA elements in fast and sensitive epigenomic profiling of open
the human genome. Nature 489:57–74. chromatin, DNA-binding proteins and nucleo-
https://doi.org/10.1038/nature11247 some position. Nat Methods 10:1213–1218.
16. Yue F, Cheng Y, Breschi A et al (2014) A https://doi.org/10.1038/nmeth.2688
comparative encyclopedia of DNA elements in 27. Afgan E, Baker D, van den Beek M et al (2016)
the mouse genome. Nature 515:355–364. The Galaxy platform for accessible, reproduc-
https://doi.org/10.1038/nature13992 ible and collaborative biomedical analyses:
17. Roadmap Epigenomics Consortium, 2016 update. Nucleic Acids Res 44:W3–W10.
Kundaje A, Meuleman W et al (2015) Integra- https://doi.org/10.1093/nar/gkw343
tive analysis of 111 reference human epigen- 28. Hnisz D, Abraham BJ, Lee TI et al (2013)
omes. Nature 518:317–330. https://doi.org/ Super-enhancers in the control of cell identity
10.1038/nature14248 and disease. Cell 155:934–947. https://doi.
18. Zhang Y, Liu T, Meyer CA et al (2008) Model- org/10.1016/j.cell.2013.09.053
based analysis of ChIP-Seq (MACS). Genome
Chapter 9
Using FRAP to Quantify Changes in Transcription Factor

Dynamics After Cell Stimulation: Cell Culture, FRAP, Data
Analysis, and Visualization
Kannan Govindaraj and Janine N. Post
Abstract
Here we show how to measure the mobility of transcription factors using fluorescence recovery after
photobleaching (FRAP). Transcription factors are DNA-binding proteins that, upon binding to specific
DNA motifs, regulate transcription of their target genes. FRAP is a simple, fast, and cost-effective method,
and is a widely used quantitative method to measure the dynamics of fluorescently labeled molecules in
solution, membranes, and inside living cells. Dynamics, specified by the immobile fraction, recovery half-
time, diffusion constant, and ratio of molecules contributing to different phases of FRAP recovery, can be
quantified by FRAP. This can be useful to understand their function in gene regulation. This tutorial is
intended to familiarize the reader with the FRAP procedure to quantify transcription factor dynamics using
a standard confocal microscope and analysis using MATLAB (MathWorks®). This article will guide the
reader through the preconditions of FRAP, and a detailed and step-by-step procedure of preparing cells,
bleaching protocol, data analysis in MATLAB, and visualization of the FRAP data.
Key words Protein dynamics, Transcription factor activity, Fluorescence recovery, FRAP , SOX9,
mGFP, CLSM
1 Introduction
Fluorescence recovery after photobleaching (FRAP) is a biophysical

technique, developed in the late 1970s by Axelrod et al. [1]. FRAP
has been successfully applied to study the mobility of fluorescent
molecules in solution. The discovery of green fluorescent protein
(GFP [2]) and subsequent advances in imaging technologies fur-
ther extended the scope of FRAP to study protein and lipid mobil-
ity in live cells. FRAP can be used to study the mobility of
fluorescently labeled proteins, lipids, and molecules in 2D struc-
tures (e.g., the plasma membrane) and in 3D structures (e.g., nuclei
and cytoplasm) [3]. However, appropriate FRAP models resem-
bling actual reaction kinetics within the system—2D or 3D—under
investigation should be used to interpret the data [4].
109
110 Kannan Govindaraj and Janine N. Post
Transcription factors play a key role in the regulation of gene

expression, and their binding to DNA precedes its activity. There
are numerous theories describing the mechanism of transcription
factor activity, reviewed in [5]. Despite many conflicting and con-
senting theories on their functional mechanism [6], the quantity of
transcription factors bound to DNA, affinity, and duration of bind-
ing seem to be the major contributing factors in exerting their
activity, which could be either transcriptional activation or repres-
sion [5, 7]. Thus, quantifying transcription factor dynamics will
yield useful information on their transcriptional activity.
In this tutorial, we will explain when and how FRAP can be
used to measure transcription factor dynamics. We describe the
FRAP protocol from preconditions of FRAP, such as required
materials and preparatory work, tips and tricks, to data analysis
and visualization. We will discuss how to stimulate cells to study
transcription factor dynamics in response to external factors [8], as
we have done in our studies. To guide the reader through the FRAP
procedure, we explain FRAP terminologies, models for fitting
FRAP data, data analysis using ImageJ (FIJI), MATLAB (Math-
Works®), and data visualization using OriginPro® (OriginLab®)
software. Additionally, we added a step-by-step protocol to per-
form a FRAP experiment using a NIKON confocal microscope as
an online supplement.
1.1 Different FRAP is a widely used method to study the protein dynamics in live
Methods for cells due to its simple setup in a laser confocal microscope. Other
Measuring Protein methods, such as FLAP (fluorescence loss after photoactivation),
Dynamics FLIP (fluorescence loss in photobleaching), FCS (fluorescence
correlation spectroscopy), and SMM (single molecule microscopy)
are also used to study protein kinetics [3]. These methods have
their own advantages and disadvantages, as described in Table 1.
1.2 When to Use FRAP can be used to study the signaling mechanism or to map the
FRAP factors that regulate a cellular process, such as the activity of tran-
scription factors. For this, cells must be transfected with a fluores-
cent fusion protein, thus the protein of interest is tagged with a
fluorescent protein.
The choice of transfection method depends on the cell type and
the protein of interest. It is recommended to make a stably trans-
fected cell line using, for example, CRISPR/Cas9 [9], if a cell line
(which properties are not affected by passage number) is used in the
study, or if overexpression of the protein of interest is toxic to the
cells. This helps to study the protein of interest at the native
expression levels and helps to avoid the transfection step every
time before FRAP. If using primary cells or cells which tend to
(de)differentiate over a number of passages, transient transfection
would be the best option. If the cells are easy to transfect, lipid
mediated transfection may be used. Retroviral or other viral-
Protocol for Quantifying Transcriptional Activity by FRAP 111
Table 1
Advantages and disadvantages of various methods used to study protein dynamics
Methods Advantages Disadvantages Refs.

FRAP l Simple instrumentation setup l Complex computational modeling is [3, 4, 29]
l Many FRAP models reflecting actual required in case of diffusion coupled
processes in the cell are already recovery
available l If fast diffusion dominates the FRAP
l If diffusion and reaction kinetics recovery, slow diffusion may not be

occur in distinct timescales, protein resolvable
dynamics can be readily calculated
FLAP/ l Simple instrumentation setup l Sensitive to acquisition [30, 31]
FLIP l Procedure is similar to FRAP, a small photobleaching
modification is required for l Reaction kinetics (slow diffusion)
calculations dominate fluorescence decay

FCS l More precise protein dynamics l Requires complex instrumentation [32–34]
diffusion rates can be calculated l Application to study the proteins
l Diffusion and reaction kinetics can inside the cell is challenging because
be distinguished even in the diffusion of complex diffusion analytical
coupled mobility models
l FCS requires a low level of
fluorescent molecules, so it is easier

to use on proteins that are expressed
at a low concentration
SMM l Individual fluorescent molecules can l Requires ultrasensitive cameras and [3, 35–37]
be tracked accurate speckle-tracking programs
l More precise protein dynamics l Transfection or microinjection
diffusion rates can be calculated. should be optimized to obtain very

l Diffusion and reaction kinetics can low levels of protein expression
be distinguished even in the diffusion l Long exposure time is necessary to
coupled mobility get good signal-to-noise ratio,

leading to acquisition
photobleaching
mediated transfection can be used for hard-to-transfect cells.

N.B. check local safety regulations, since in most countries the
use of viruses is only allowed in labs with the correct biosafety
levels, which is then also true for the location of the microscope
(BSL-2 in the USA, MLII in the Netherlands). We do not have a
confocal microscope in our MLII lab, so we cannot use viral induc-
tion for our FRAP experiments.
Since the cells with which we work, human primary chondro-
cytes (hPCs) and human mesenchymal stem cells (hMSCs), tend to
dedifferentiate over a number of passages, we used lipid mediated
transient transfection. In these primary cells, we were able to
achieve at least a transfection efficiency of 20% (see Subheading 3).
For training purposes, we use the immortalized C20/A4 cell line.
1.3 FRAP Principle In FRAP a small region of interest (ROI, can be circular, square, or
rectangular) is photobleached with a high-intensity laser, and the
fluorescent recovery is monitored in that ROI in a given timescale
(from seconds to minutes). Post-bleach imaging duration and
frame-rate is determined based on the mobility of the molecule of
interest. The higher the mobility, the lower the imaging time, the
higher the frame rate. For example, if the mobility of a transcription
is relatively high, with a stable plateau reached within a minute,
fluorescence recovery can be recorded for 60 s, at a speed of
4 frames per second (fps). Before photobleaching, a number of
pre-bleach images (usually ~10) are recorded for normalization of
the fluorescence recovery. This process can be translated into a
graph.
As an example, we measure the fluorescence recovery of the
transcription factor SOX9, the master regulator for cartilage for-
mation [10], see Fig. 1. In this graph, one can observe that the
fluorescence in the bleach spot does not reach the level of the
original fluorescence. The plateau level is lower than the
pre-bleach intensity because some of the FRAP-bleached molecules
are immobile within the region of interest that is bleached. Because
of their immobility, they do not contribute to the recovery, while at
the same time occupying binding sites for incoming unbleached
proteins. We therefore name the fraction of molecules that contri-
butes to the recovery the “mobile fraction” and the one that does
not contribute the “immobile fraction.”
Fig. 1 Fluorescence recovery curve of SOX9-mGFP. The fluorescence intensity is

normalized to 100%. The immobile fraction and recovery half-time are indicated
(red dotted lines)
1.4 Mapping Signal For studies of immediate response of transcription factors to extra-
Transduction cellular signals, cells can be stimulated with cytokines (before/
Pathways Regulating during FRAP experiments) in an imaging buffer (see below). For
Transcription Factor example, we transfected the cells with a fluorescently tagged tran-
Mobility scription factor, SOX9-mGFP and treated cells with predicted ago-
nists/antagonists to study their transcriptional regulation
[8]. Stimulation time depends on the mechanism of action of
cytokine and needs to be determined empirically. If the cytokine
exerts fast cellular responses, as we have seen for the SOX9 cata-
bolic factors, such as WNT3a and IL1β, changes were observed
within 20 mins after stimulation (Fig. 2, [8]). However, a slightly
longer stimulation time (60 min) was required for the SOX9 ana-
bolic factor BMP7 (see Fig. 2).
1.5 Quantitation of Two major factors affect the mobility of a molecule in a biological
Protein Mobility system: diffusion and chemical interactions. The rate of diffusion is
determined by the diffusion constant (D) of the molecule, which is
dependent on the size of the molecule, the viscosity of the sur-
rounding medium, including physical structures that hinder diffu-
sion, and the temperature. In the cell, proteins interact with other
molecules, including other proteins and DNA/RNA. The binding
constants of the molecular interactions, binding (kon) and dissocia-
tion (koff) with other molecules, will affect mobility of the protein
of interest.
Fig. 2 FRAP curves showing SOX9-mGFP mobility changes in C20/A4 cells after
treatment with cytokines. Treatment of C20/A4 cells with WNT3a at 10 ng/mL
increased SOX9-mGFP mobility at 20 mins (yellow curve) as compared to the
control (blue curve). In contrast, BMP7 at 10 ng/mL did not change SOX9-mGFP
mobility significantly at 20 min (orange curve) as compared to the control.
However, 100 ng/mL of BMP7 with 1 h treatment decreased the SOX9-mGFP
mobility significantly (gray curve). This indicates that concentration of cytokines
and stimulation time varies among cytokines and should be determined
empirically
Protein kinetic data can be calculated from the FRAP measure-

ments. This includes the fraction of immobile molecules (immobile
fraction, IF), the recovery half-time (τ½, time it takes for the
fluorescence to reach 50% of the final fluorescence), diffusion con-
stant (D), and association (kon) and dissociation (koff) rates
[11]. So, using FRAP, one can calculate the diffusion and reaction
kinetics of any fluorescently labeled protein/lipid in a cell. If the
molecular processes that contribute to the fluorescence recovery
occur in distinct timescales, FRAP is a robust method to study the
protein kinetics.
To calculate the protein kinetic data, different equations/mod-
els can be applied: (i) chemical interaction model, (ii) diffusion
model, and (iii) reaction-diffusion model. In the chemical interac-
tion model, FRAP curves are fitted with an exponential equation
with a single exponent:

F ðt Þ ¼ y0 þ A 1 e t=τ ð1Þ
where y0 is the value of the fluorescent intensity at the first post-

bleach frame, A is the amplitude of mobile population, and τ is the
time constant.
To calculate the half-time recovery of A,
Half time to recover : t § ¼ ln ð2Þ∗τ ð2Þ
where t½ is the time-point at which 50% of the fluorescence is
recovered, τ is the time constant.
Immobile fraction : IF ¼ F I F E ð3Þ
where FI is the initial intensity and FE is the end value of the
recovered intensity.
The effective diffusion constant (D) can be determined from
the FRAP curves based on the method from [12], which describes
how the effective diffusion constants can be calculated from FRAP
curves in combination with its laser bleaching profile. To correct for
diffusion during bleaching, the effective radius for the bleaching
spot is calculated from the user-defined nominal radius. In combi-
nation, a FRAP equation was derived that can be readily fitted to
the normalized FRAP data to extract diffusion constants:

K ∗re 2
F ðt Þ ¼ 1 ∗M f þ ð1 M f Þ∗F0 ð4Þ
8∗D∗t þ re 2 þ rn 2
The laser bleaching profile is incorporated in Eq. 4 with K as
bleach depth parameter, rn as the user-defined nominal radius, and
re as the effective radius. Mf is the mobile fraction, F0 is the post-
bleach intensity, and D is the effective diffusion constant when
considering 3-dimensional diffusion.
If the fluorescence recovery is diffusion-coupled (fast and slow
diffusion occur in similar timescales) or when many processes occur
in overlapping timescales, complex computational modeling is

required for FRAP data analysis. In this case, caution should be
applied when calculating diffusion constants and association and
dissociation rates from the FRAP data [4]. On the other hand,
fluorescence correlation spectroscopy and single molecule micros-
copy can resolve the different molecular processes occurring at
similar timescales and provide precise diffusion constants and asso-
ciation and dissociation rates [3]. For example, in the transcription
factor dynamics, diffusion by unbound protein occurs instantly,
even before capturing the first post-bleach image. The FRAP
curve will be a result of two diffusion processes with distinct time-
scales, fast (A1) and slow (A2) diffusion. Nonspecific binding of the
transcription factor to DNA will be fast moving (A1) and the
fluorescence recovery will be slower at the specific binding (A2).
The ratio (A1/A2) indicates the amount of contribution of these
populations present in the measured cell. In this instant, FRAP
curves can be fitted using a two-component fit:

F ðt Þ ¼ y0 þ A 1 1 e t=τ1 þ A 2 1 e t=τ2 ð5Þ
where A2 is the amplitude of slow diffusing population, τ1 and

τ2 are the time constants of A1 and A2, respectively.
1.6 Analyze FRAP Proper fitting and modeling of the FRAP curve, which reflects the
Data actual biological process, is important to extract useful dynamics
information from the FRAP data. Fitting and modeling should
consider the number of reactions contributing to the FRAP recov-
ery, the timescales, and whether these reactions occur in similar or
distinct timescales. For example, untagged mGFP (OriGene) pro-
tein does not bind to any intracellular target and recovers
completely and within seconds after photobleaching (see Fig. 3a).
FRAP data of mGFP can be easily fit with a single exponential fit (see
Fig. 3b, Eq. 1).
In contrast, SOX9-mGFP binds to DNA and the FRAP recov-
ery curve is the result of two reactions, namely fast (weakly bound
to DNA, nonspecific binding) and slow diffusion (strongly bound
to DNA, specific binding). FRAP data of a transcription factor like
SOX9 should be fit with two (or more) component fits (see Fig. 3c,
Eq. 5). The presence of shoulders in the curve can be an indicator
of the number of processes contributing to the FRAP recovery.
After bleaching, mGFP quickly recovers and attains a plateau, while
SOX9-mGFP slowly recovers and continues to recover (due to
exchange at the binding sites) throughout the recovery timescale,
with a visible shoulder in the recovery curve (see Fig. 3a, red curve).
In our model, we use a two-component fit (see Eq. 5) for
transcription factor dynamics, as two processes at distinct timescales
Fig. 3 Fitting FRAP curves. (a) FRAP curves of mGFP (gray) and SOX9-mGFP (red). (b) The mGFP FRAP curve is
fit with a one-component fit (blue circles are data-points, red line is the fit) and (c) the SOX9-mGFP curve is fit
with a two-component fit (gray and green lines are first- and second-component fit, respectively)
contribute to the FRAP recovery. If more than two processes

contribute to FRAP recovery or if these processes occur in similar
timescales, this FRAP model cannot be applied. In that case a new
model should be built based on initial control experiments for the
analysis of FRAP data.
1.7 Explanation of Immobile fraction (IF, Eq. 3) refers to the fraction fluorescent
FRAP Parameters protein bound to its target, e.g., transcription factor bound to
DNA. For transcription factors there are two populations, namely
fast (A1) and slow (A2) diffusing populations that contribute to
the fluorescence recovery (see Fig. 4). The fraction of protein that
are weakly bound to DNA constitute the fast diffusion population,
whereas the population bound to DNA constitutes the slow diffus-
ing population. The ratio A1/A2 refers to the increase or the
decrease of fluorescent proteins in the A1 compared to A2 fraction
(i.e., the ratio of the fast diffusing population to the slow diffusing
population).
If the fast and the slow diffusing populations are equally present
inside the nucleus, the value of the A1/A2 ratio will be 1. If the
Fig. 4 Schematic representation of a FRAP recovery curve and explanation of parameters. If fast and slow
diffusion occur at two different timescales, the FRAP curve can be split into two phases, as shown. FI: initial
intensity, F0: intensity at time point t0 (first post-bleach intensity), FE: end value of the recovered intensity, t½:
halftime of recovery, immobile fraction (IF) is the population of SOX9-mGFP bound to DNA. A1 is the amplitude
of fast diffusing population of SOX9-mGFP, which is not bound to DNA and contributes to quick recovery. A2 is
the amplitude of slow diffusing population of SOX9-mGFP, which interacts on the various binding sites in
the DNA
value is less than 1, this indicates that the number of fast diffusing
proteins is less than that of the slow diffusing proteins. The diffu-
sion constant (D) is the rate at which a molecule diffuses in a
specific area. In a cellular milieu, the term “effective diffusion
(Deff)” indicates the recovery that mimics diffusion, but at a rate
that is slowed by binding interactions. This is calculated using
Eq. 4. The recovery time is the length of time after bleaching,
required for the fluorescence recovery to reach a constant value.
Recovery half-time (t½, Eq. 3) refers to half of recovery time.
Recovery half-time of A1 and A2 indicates half of the time
required for the recovery of corresponding phase of the FRAP
curve. For detailed information, the reader can refer to [8].
2 Materials
2.1 Materials for Cell 1. Silicone gasket with 1.5 mm thickness (cat#: 70465-2R2,
Culture and EMS, USA).
Transfection 2. Fibreless tissue paper (such as Kimberly-Clark® Professional).
3. 24-Well plates.
4. Microscopic cover glass (12 or 13 mm, Ø).

5. Tweezers to lift the coverslip.
6. Suitable transfection reagent.
(a) Lipofectamine LTX with Plus reagent (cat. # 15338030,
ThermoFisher scientific) (for C20/A4 cells and human
primary chondrocytes). We used 1:1 (w/v) ratio of DNA
and Plus reagent and 2:1 (v/v) ratio of Lipofectamine
LTX to Plus reagent.
(b) Lipofectamine 3000 with P3000 reagent (cat. #
L3000015, ThermoFisher scientific) (for hMSCs). We
used 1:1.5 (w/v) ratio of DNA and P3000 reagent and
2:1 (v/w) ratio of Lipofectamine 3000 to DNA.
7. Plasmid DNA encoding GFP-tagged protein of interest.
8. Confocal microscope.
9. Cells of interest.
10. Proteins/cytokines for cell stimulation.
11. Microscopic glass slide.
12. Imaging buffer.
2.2 Materials for 1. 135 mM NaCl.

FRAP 2. 10 mM KCl.
2.2.1 Imaging Buffer 3. 0.4 mM MgCl2.
(Tyrode’s Solution, See 4. 1 mM CaCl2.
Note 1)
5. 10 mM HEPES, pH adjusted to 7.2.
6. Imaging buffer with above components was filter sterilized and
stored at 20 C.
7. On the day of use, add 20 mM glucose and 0.1% BSA (final
concentration after adding to the buffer).
Alternatively, medium without phenol red can also be used as
imaging media (see Note 2).
2.2.2 Confocal Laser 1. Laser scanning confocal microscope (We used a Nikon A1
Scanning Microscope confocal microscope) with suitable laser lines for excitation of
the fluorescent protein.
2. Option to maintain the physiological temperature, CO2
control.
2.3 Materials for 1. MATLAB with script (available on request).

FRAP Analysis 2. ImageJ (for drift correction).
3. Alternatively, easyFRAP software can be used.
2.4 Materials for 1. Data analysis and graphing software, we used Origin Pro
Data Visualization and (OriginLab®).
Statistical Analysis
3 Methods
3.1 Cell Culture and To study the cellular physiology, cells ideally should be maintained
Transfection in their native state. For example, primary chondrocytes should
maintain their chondrogenic potential during the experiments. To
3.1.1 Cell Culture
prevent dedifferentiation, expand the primary chondrocytes in phy-
sioxical (2.5–5% O2) conditions. Use low passage numbers (<4)
and low population doubling levels (PDL < 5). Culture the cells
without antibiotics especially during and after transfection. We
usually do not serum-starve the cells before FRAP experiments,
unless we stimulate with growth factors.
Culturing with serum or serum-starvation: Depending on the
nature of the study, cells can be cultured with or without serum. If
the dynamics of protein under study is influenced by cell cycle, it
would be appropriate to synchronize the cell cycle by serum starva-
tion. Also, if the study is intended to investigate the long-time
effect of an external growth factor/cytokine cells may need to be
cultured without or with inactivated serum to avoid interference
with the growth factors present in the serum.
3.1.2 Choice of the The fluorophore of choice should be irreversibly photobleachable

Fluorophore at high intensity light and photo-stable at lower light intensities.
The chosen fluorescent protein should not have dimerization, tet-
ramerization, or oligomerization properties. These properties will
lead to complex formation and may impair the function of the
protein of interest [13]. The fluorescent protein should also have
a fast maturation time and be resistant to environmental changes,
such as temperature, O2, and pH. We refer the readers to the
following references to know more about selection of fluorescent
proteins [14–16]. Fluorescent proteins such as mGFP, eGFP,
mRFP, and mCherry have been successfully used in FRAP experi-
ments [17–19].
3.1.3 Transfection FRAP can only be performed on fluorescently labeled molecules.

For the study of transcription factors in primary cells, transient
transfection is the easiest method of introducing a fluorescent
fusion protein. At least 10–15% of transfection efficiency is required
to image sufficient cells for FRAP. Optimization of seeding density,
transfection reagent, plasmid DNA purification, and time of trans-
fection may be necessary. We transfected immortalized juvenile
costal chondrocytes (C20/A4), human primary chondrocytes
(hPCs), and human mesenchymal stem cells (hMSCs) at 70–80%
confluent level and are able to achieve a minimum transfection

efficiency of ~20%. To get a high transfection efficiency, use Opti-
MEM (Invitrogen) medium (without serum and antibiotics) dur-
ing transfection. After 2:00–4:00 h (depending on the cell type),
replace this medium with normal culture medium (without
antibiotics).
3.1.4 Cell Stimulation Cell stimulation time and concentration of cytokines/growth fac-
tors are other important factors to be considered. Stimulation time
depends on the speed of cellular response, the concentration
depends on the efficacy of the cytokine/growth factor, and both
needed to be determined empirically. We stimulated C20/A4 cells
[20] with different concentrations (1 ng/mL, 10 ng/mL and
200 ng/mL) of WNT3a and found that 10 ng/mL was the mini-
mal concentration needed to destabilize SOX9-mGFP from DNA
[8]. In contrast, BMP7 promoted DNA binding of SOX9-mGFP at
100 ng/mL concentration and with a longer incubation time (see
Fig. 2).
3.1.5 Preparing Cells for Day 1: Seed the cells on a sterile, microscopic glass coverslip (12 or
FRAP 13 mm Ø) placed in a 24-well plate (see Note 3).
Day 2: Transfect the cells with the plasmid encoding the fluo-
rescent protein of interest.
Day 3: Perform FRAP using the protocol below. For FRAP
experiments, transfected cells grown on coverslip (on day 2) need
to be prepared in the following way (see Fig. 5).
1. Mount a silicone well divider (Electron Microscopy Sciences,
USA, with right well size) on to a microscopic glass slide (see
Fig. 5a, Note 4).
2. Fill the silicone well with imaging buffer (see Fig. 5b).
3. Place the coverslip in the middle of silicone well filled with
imaging buffer, with the transfected cells facing the imaging
buffer (see Fig. 5c, Note 5).
4. Excess water will be squeezed out and the coverslip will nicely
mount on to the silicone divider. Make sure that no air bubble
is introduced in the buffer. If there is an air bubble, carefully
remove the coverslip with the tweezer and start from step 1.
Once the coverslip with transfected cells is successfully
mounted on to the silicone well, it is ready for FRAP. For this
follow the FRAP protocol in Subheading 3.2.
3.2 FRAP There is no standardized universal protocol for FRAP since the
experimental design depends on bleaching and recovery character-
istics of the fluorescent molecule under study. However, all FRAP
experiments contain three phases, i.e., (1) pre-bleach image acqui-
sition, (2) photobleaching, and (3) post-bleach image acquisition.
Fig. 5 Preparation of the cells grown on coverslip for FRAP. (a) The silicone divider is mounted on to a
microscopic glass slide. (b) Imaging buffer is filled in the silicone well. (c) Transfected cells grown on a
coverslip is placed on the imaging buffer (cells are facing the imaging buffer)
For the FRAP dataset to be comparable across various conditions/

treatments, keep the same parameters for all FRAP measurements.
1. Pre-bleach image acquisition: Pre-bleach images are used to
normalize the fluorescence recovery data during analysis. Usu-
ally, 10 pre-bleach images are sufficient to get averaged
pre-bleach intensity. However, if your fluorophore (such as
mGFP) has a triplet state, in order to exclude the artifacts
introduced by this triplet state of mGFP, more pre-bleach
images (>25) are required. During analysis, use only the aver-
age intensity of last 10 pre-bleach images for normalization.
2. Photobleaching: Photobleaching should be instantaneous
after the pre-bleach. A small region of interest (ROI) is photo-
bleached with a high laser power. The laser used for the image
acquisition can be used for photobleaching as well and the laser
power needs to be determined empirically. Duration of bleach-
ing and number of iterations of high-intensity laser pulse
should be as low as possible to calculate accurate protein
dynamics. Optimal bleaching parameters can be easily deter-
mined using a fixed cell (using 4% paraformaldehyde in PBS)
Fig. 6 Optimizing bleaching laser intensity. Bleaching efficiency of SOX9-mGFP in a fixed cell at different
percentages and μW (at the objective) of laser power as indicated in the image. Bleaching is inefficient at (a)
10% (6.8 μW) and (b) 25% (17.1 μW) laser power, whereas (c) 50% (34.3 μW) and (d) 100% (69.0 μW)
efficiently bleach the SOX9-mGFP and the bleach profile is shown in (e)
expressing the same fluorescent protein (see Fig. 6). The fol-
lowing guidelines can be helpful:
(a) Laser power: Usually, 25–100% laser power is used for
photobleaching. Higher laser power enables faster bleach-
ing, but is toxic to the cells [21]. It should be optimized to
the minimum laser power needed to achieve complete
photobleaching of the ROI (see Fig. 6).
(b) Size of the ROI: The size of the ROI is usually less than
10% of the total fluorescence distribution area. It also
depends on the kinetics of the fluorescent molecule. A
size of 1–2 μm (Ø) ROIs with higher magnification objec-
tives (60/100) give a good signal-to-noise ratio
(SNR). Smaller ROIs will increase the noise.
(c) Shape of the ROI: The ROI can be circular or a rectangle
or a square. Again, this depends on the shape/distribution
of the fluorescent protein. We use a circular ROI for
nuclear proteins, such as transcription factors. However,
the calculation formulae may need to be adjusted accord-
ing to the shape of the ROI [22].
(d) Duration of bleaching: Ideally it should be as short as

possible.
(e) Number of iterations: It is recommended to achieve
complete photobleaching in a single iteration. This is
especially important when studying fast kinetics.
(f) Scan speed: The higher the scan speed, the faster the
bleaching process. However, the bleaching can be
inefficient.
3. Post-bleach image acquisition: Acquire sequence of images
after photobleaching to monitor the dynamics of fluorescence
recovery. The following acquisition guidelines can be used to
resolve the dynamic range and good temporal resolution.
(a) Acquisition frequency: At least 20 images during the
time required for the half of the recovery are needed.
(b) Acquisition duration: Needs to be 10 to 50 times longer
than the halftime of recovery [1] or until the fluorescence
recovery attains a plateau and should be determined
empirically.
(c) Acquisition photobleaching: To resolve precise protein
dynamics, it is necessary to reduce the photobleaching
during image acquisition and the guidelines below can
be helpful.
The general rule is to reduce the photo-toxicity during FRAP,

so acquiring sharp and nice images are not the priority. To minimize
photobleaching and photo-toxicity during FRAP, the following
parameters can be adjusted:
1. Laser power: Laser power should be as low as possible during
pre- and post-bleach image acquisition. For example, to image
mGFP or mGFP-tagged protein, laser powers of ~0.35–0.5%
are ideal on the Nikon confocal microscope. Laser power also
depends on the age of the laser. If ~0.35–0.5% laser power does
not sufficiently illuminate the fluorophore, increase the laser
power.
2. Frame size: Decreasing the frame size will enable faster a scan
rate and less light exposure. (For example, a 125 125 or
256 256 frame size can be used instead of 512 512).
3. Frame rate: The higher the frame rate, the lesser the light
exposure of the cells. Higher frame rates are recommended
especially during the study of rapid kinetics.
4. Averaging: Frame and line averaging should be avoided to
achieve faster frame rates and to reduce photo-toxicity and
photobleaching. Again, obtaining the best image is not a
priority.
5. Pinhole: Opening the pinhole enables to capture more signal

and helps to keep the laser power at the minimal level. But, it
also enlarges the confocal volume and background signal.
However, this will not pose much problem during data analysis
as the SNR will be higher for our FRAP settings.
6. Zoom: Use the zoom option to enlarge the imaging area.
7. Photo-stable fluorophores: Use fluorophores which are
photo-stable at lower light intensities (such as eGFP/mRFP)
to reduce acquisition photobleaching.
3.2.1 Optimal FRAP The optimal FRAP parameters for our study of SOX9-mGFP
Parameters dynamics in the nucleus are given below as an example. Although
we used a Nikon A1, the settings should be easily transferable to a
confocal microscope of a different manufacturer.
1. Pre-bleach images: 25 images.
2. Bleaching: 1 iteration of high-intensity laser pulse (50%,
488 nm laser).
3. Post-bleach images: 240 images (60 s, imaging time).
4. Objective: 60 (water immersion), 1.4 NA. Alternatively, a
silicone oil immersion objective can be used, as long as the
refractive index of the objective is close to that of the cell to
avoid aberrations.
5. Scan mode: Unidirectional.
6. Frame rate: 4 frames/s(fps, for both pre- and post-bleach
imaging). The frame rate can also be set based on the pixel
dwell time.
7. Frame size: 256 256 pixels.
8. Averaging: Normal (No averaging!).
9. Pinhole: 1.2 AU (Z-step size: 0.25 μm and Optical sectioning:
0.77 μm).
10. HV (Gain of the detector): ~80–90% (depends on the intensity
of the fluorescent protein).
11. Offset (of the detector): 1.
12. Laser power: 0.35% (0.12 μW at the objective).
13. Zoom: 7.09.
14. Pixel size: 0.12 μm.
15. Laser: 488 nm.
16. Bleaching: 50% (34.3 μW, at the objective) laser power, 16 fps
(the highest speed possible in the Nikon A1).
17. Export the fluorescent intensity values to a Microsoft Excel
document using the “Export” option and save it in a folder.
This Excel document will be used for data analysis in
MATLAB.
18. Repeat the FRAP experiment for at least 50 cells per condition
and collect FRAP image files and Excel documents containing
FRAP data.
3.2.2 Performing FRAP As a detailed example of a FRAP experiment, we describe

Using a Nikon A1 Confocal performing FRAP on a Nikon A1 confocal microscope. This may
Laser Scanning help the reader to convert the protocol to their own CLSM.
Microscope Nikon’s “NIS elements” software provides convenient user inter-
face and FRAP settings. This session will guide the user through
FRAP setup in “NIS elements” software. Options and values shown
in the snippets below are based on the optimal FRAP parameters we
used. In this protocol, we use a nucleus expressing SOX9-mGFP as
an intracellular organelle to explain the FRAP settings in the NIS
elements software. If your target organelle or region is other than
the nucleus, you can replace the term “nucleus” with your target.
1. Turn on the laser, temperature controller, and the microscope
system.
2. Fasten a stage inlet suitable to mount microscopic cover glass.
3. Start the “NIS elements” software.
4. Following acquisition and analysis controls need to be opened
in the NIS elements software (see Fig. 7).
(a) A1plus Compact GUI (View ! Acquisition controls !
A1plus Compact GUI).
(b) A1plus Stimulation (View ! Acquisition controls !
A1plus Stimulation).
(c) A1plus Scan Area (View ! Acquisition controls !
A1plus Scan Area).
(d) ND Stimulation (View ! Acquisition controls ! ND
Stimulation).
(e) Ti Pad (View ! Acquisition controls ! Ti Pad).
(f) Time Measurement (View ! Analysis controls ! Time
Measurement).
5. “Ti Pad” contains basic microscopic control options. Set the
options as shown below to view the cells in the binocular.
(a) Select “60x objective”.
(b) Select the light path “E100”.
(c) Select “EPI” mode and turn on the shutterns (in the
separate physical device).
(d) Select “green filter” (for GFP).
(e) Zoom option in the “Ti Pad” is the optical zoom. Keep it
at 1.00 in the “Ti Pad” and in the microscope as well.
Fig. 7 “A1plus Compact GUI” provides options to control image acquisition. (a) While “Eye port” is selected
(red box 2), interlock is automatically activated (red box 3) and the rest of the options are automatically
disabled. Samples can be viewed through the binocular. (b) While “Scan” mode is on (red box 1), “Eye port” is
not selected (red box 2), “Interlock” is removed (red box 3), “Unidirectional” scan mode is selected (red box 4),
“Scan speed” is set to 4 fps, “image size” is set to 256 256 pixels (red box 5), “Averaging” is set to normal
(red box 6). “Pinhole” is set to 1.2 AU (red box 7). Needed laser line can be activated by the “laser settings”
option (red box 8), “HV” is set to 88, “Offset” is set to 1, “488 nm” laser line is selected and “laser power” is
set to 0.35% (red box 9)
6. Set the parameters in the “A1plus Scan Area” window, if nec-

essary. Once the parameters are set, drag and position the scan
area (red square) to the middle of the window, and to confirm
the position, right click on it (the red square will turn into
green).
7. Place a drop of milliQ water on the 60 water objective and
place the microscopic glass slide mounted with transfected
cells. The glass slide should be placed in the inverted position,

so that the microscopic coverslip is in contact with the
objective.
8. Look through the “binocular” and focus the cells using white
light. Once the cells are focused, turn on the epi(fluorescent)
light (white light can be turned off) and search for the trans-
fected cells. Click the “EPI” button to turn on/off the epi light
in the “TiPad” window. Following guidelines can be helpful to
select the right cells for FRAP.
(a) The transfected cell morphology should be normal.
(b) Cells having extra weird fluorescent bodies should be
avoided.
(c) Avoid cells expressing very low or very high levels of
fluorescent proteins.
(d) For cells expressing optimal level of fluorescent proteins,
“HV” values would be ~85–90% (for above mentioned
optimal FRAP parameters).
9. Position a nucleus (or other cellular region of interest) expres-
sing fluorescent proteins in the center of the view area (when
looking through the binocular).
10. Change the light path to “L100” and turn on the “PFS” in the
“Ti Pad” window.
11. Set the image acquisition parameters in “A1plus Compact
GUI” window, as shown in Fig. 10. If “Remove Interlock”
button appears in “RED,” click on it, to remove interlock
and turn on the “PFS” in the “Ti Pad” window.
12. Make sure that the correct laser is selected (for example, the
488 laser in case of GFP tagged protein); click on “Scan”
button (top left corner) in the “A1plus Compact GUI” win-
dow. Once the first FRAP measurement is finished after start-
ing NIS elements software, switching between “Eye” and
“Scan” mode can be done in one click by “Eye Port” button
(see Fig. 7, red box 2).
13. A new window showing the centered nucleus will appear.
Focus may need to be slightly adjusted to bring the nucleus
into focus.
14. Turn on “pixel saturation indicator” and see if there are any
saturated pixels in the nucleus (saturated pixels are shown as
red dots by default). If saturated pixels appear, decrease HV
values to reduce the number of saturated pixels (a few, <5
saturated pixels are allowed) and stop “Scan” in the “A1plus
Compact GUI” window. A frozen window showing the desired
nucleus will stay in the screen. (If the nucleus appears in the
window, skip steps 15 and 16 and proceed to step 17).
15. If the nucleus does not appear in the window, the view area
(step 6) or the scan area (step 9) might not be centered
properly.
16. Another way to center the desired nucleus (expressing SOX9-
mGFP) is to right click on the scan area (green square) and
select the option “Scan full field of view” in the “A1plus Scan
Area” window. The image acquisition parameters, such as
frame rate, frame size, and zoom, will change. Again, set the
zoom to 7.09 and scan area (red square) will appear again.
Drag and position it around the nucleus of interest and right
click on the red square to confirm the position (the red square
will turn green). Adjust the image acquisition parameters to
optimal conditions and click “Scan” button in the “A1plus
Compact GUI” window and return to step 13.
17. Draw three ROIs in the frozen window. To draw ROIs, right
click and drag inside the frozen window. Options to choose the
ROI shape will appear on the frozen window and choose
“Circular ROI” option. The mouse pointer will turn into a
“+” sign; click and drag on the frozen window, a red circular
ROI will be drawn.
18. Right click on the “red ROI” and click “ROI properties”
option. “ROI properties” window will appear and set the ROI
dimensions. After setting the values, close the “ROI properties”
window.
19. To draw two more ROIs, either repeat steps 17 and 18 or
simply right click on the current “red ROI” and select “Dupli-
cate Selected ROIs” option.
20. Assign a function to each ROI as mentioned in the order
below. Click on a ROI to select it, to assign a function, right
click on it, a window will appear, and select the function to be
assigned. Keep the below order throughout the FRAP mea-
surements, to get uniform datasets.
(a) “Stimulation ROI” (red, for FRAP measurements).
(b) “Reference ROI” (green, to measure the acquisition
photobleaching).
(c) “Background ROI” (blue, to measure the background
signal).
21. Once functions are assigned, ROIs in the frozen image will be
labeled accordingly and position the ROIs as per following
guidelines. Position the “Stimulation ROI S1:1” (red) in a
representative region close to the center of the nucleus. Posi-
tion the “Reference ROI R:2” (green) as much as away from
the stimulation ROI (but the fluorescent intensity levels should
be identical to stimulation ROI). Position the “Background
ROI B:3” at the outside of the cell.
22. Set the bleaching parameters in “A1plus stimulation” window.

Select the appropriate laser (for example, 488 nm laser for
GFP) for bleaching and set the intensity and the “Scan speed”
as optimized.
23. Set the FRAP duration and file saving options in “ND Stimu-
lation” window (see Fig. 8). Set the parameters as mentioned in
the FRAP protocol. Follow the instructions in the figure
Fig. 8 ND Stimulation settings. (a) FRAP duration and file saving options are set in the “ND stimulation”
window. Check the box next to “Save to File” (red box 1). Select the file saving path and write the file name
(red box 2and 3). A FRAP experiment has three phases: (1) pre-bleach, (2) photobleaching (Stimulation), and
(3) post-bleach image acquisition. Add those phases using “Add” button (red box 4). Three phases are shown
(red box 5). Set the parameters for each phase as follows: Phase #1: Select Acquisition, Set the number of
“Loops” to 25 (25 pre-bleach images), Phase #2: Select Stimulation, Select “S1” ROI, Set the number of
“Loops” to 1 (1 iteration of high-intensity laser), Phase #3: Select Acquisition, Set the number of “Loops” to
240 (240 post-bleach images). Set the interval to “No delay” for all three phases to acquire images
continuously. Once number of loops were set, the duration will be set automatically. Check the box next to
“Perform Time Measurement” (red box 6). Select “Apply Stimulation Settings” (red box 7) and start the FRAP
experiment by clicking ‘Run now’ button (red box 8)
legend to set the parameters and run the FRAP experiment. A

new live window (with the “set file name”) acquiring series of
images as set in “ND Stimulation” will appear.
24. Once FRAP measurements are finished, fluorescent intensity
values for all the three ROIs will be plotted against the function
of time and will be displayed automatically in the “Time Mea-
surement” window.
25. Check the FRAP data shown in the “Time Measurement”
window for the following artifacts.
(a) Acquisition photobleaching (green ROI) should be less
than 10%. If it is more than 10% or any weird fluctuation is
present, move the “Reference ROI” to another location
within the nucleus (but far from the “Stimulation ROI”)
in the “FRAP image stack.”
(b) If the ROI is locked (fixed), right click on the ROI and
select “Unlock Selected ROIs” option. Once ROI is moved
to another location, click the “Measure” button in the
“Time Measurement” window. Repeat this step until the
reference line is free of artifacts as mentioned in Subhead-
ing 3.2.2, step 25(a).
(c) There should be a proper FRAP recovery (red ROI) in the
“Stimulation ROI” intensity values. If there is no recovery
and the intensity values appear flat after photobleaching,
discard the data and check if the cells are healthy.
(d) Background intensity values ideally should be less than
50. If it is more than 50, move the “Background ROI” to
the outside of the cell and repeat the step (b).
(e) XY drift: if there is XY drift in the FRAP image stack, data
can be used after drift correction (image registration, see
Note 6).
(f) Z drift: If the FRAP images have Z drift, discard that
FRAP measurement (see Note 7).
26. Export the fluorescent intensity values to a Microsoft Excel
document using the “Export” option and save it in a folder.
This Excel document will be used for data analysis in
MATLAB.
27. Repeat the FRAP experiment for at least 50 cells per condition
and collect FRAP image files and Excel documents containing
FRAP data.
3.3 Data Analysis It is necessary to check the FRAP data for “XY drift” before data
analysis. Even a small drift can significantly alter the FRAP dynam-
3.3.1 FRAP Data
ics data.
Validation Before Analysis
1. “XY drift” can be corrected by image registration using Fiji or
ImageJ. We used Fiji as it contains all plugins needed to read
Nikon’s “.nd2” file format. However, image registration plu-
gins need to be installed manually in Fiji. We used the “Tem-
plate Matching” plugin for image registration. This can be
downloaded from this link [23, 24]. Follow the instructions
given in the site to install the plugin in Fiji and for image
registration.
2. After image registration, play the FRAP image stack in Fiji to
check for the proper alignment. Still, if the alignment is not
proper, discard this FRAP data.
3. If the alignment is proper, go to the first post-bleach image
(26th image in the stack), draw a “circular ROI” on the bleach
area using Fiji “Circle” tool.
4. Select “Plot Z-axis Profile” option in Fiji (Image ! Stacks !
Plot Z-axis Profile). A new window showing a fluorescent
intensity graph will appear. Extract fluorescent intensity values
from this window and replace the respective column in the
Excel file containing FRAP data.
5. Draw another “Circular ROI” to collect reference and back-
ground fluorescent intensities as well and follow the previous
step (see Note 8).
3.3.2 FRAP Data There are many MATLAB scripts and free software, such as easy-
Analysis in MATLAB FRAP [25] and ImageJ plugins [26], available for FRAP data
analysis. However, we used our own MATLAB script for FRAP
analysis, and this session will guide the user through our script. The
MATLAB script and sample data sets are available on request. Only
the “Microsoft Excel” documents containing FRAP data are
needed for data analysis.
1. Download the “Data Analysis” folder from the above link and
the folder contains necessary MATLAB scripts for FRAP data
analysis and sample data sets.
(a) It has two folders: “Data Files” and “Effective Radius.”
(b) It has two MATLAB scripts FRAPAnalysis and findHT.
2. Place both the MATLAB scripts in the MATLAB home direc-
tory folder.
3. Open the MATLAB and “FRAPAnalysis” script in the
MATLAB.
4. The effective radius of the bleach spot should be determined
before the FRAP analysis. Due to the point spread function of
Fig. 9 Using the bleach profile to calculate the effective radius. (a) “Straight line” tool is selected in Fiji (red
box 1). A straight line is drawn across the bleach ROI of the first post-bleach image (red box 2). (b) Profile of
the first post-bleach image. (c) Profile of the last pre-bleach image. Click on the “More>>” (red box 3) button
and select “Copy All Data” option. Paste it in a separate “.txt” file. (d) Averaged bleach profile and the
calculation of effective radius. Effective radius can be calculated from the center axis (red box 4) and using the
“μm” scale of the “X”-axis. Effective radius of this bleach profile is 2.2 μm
the light, in practice, the size of the bleached ROI will be always
higher than the actual size of the ROI. So, finding the actual
bleaching size and effective radius is necessary to calculate
precise protein dynamics. To know more about effective radius
and nominal radius, refer to [27]. The effective radius of the
bleach spot can be calculated as described below.
(a) Import the “FRAP image stack” to Fiji.
(b) Go to the first post-bleach image (26th image) and draw a
straight line across the bleach spot as shown in the Fig. 9a.
Center of the straight line should be at the center of the
bleach spot.
(c) Click on “Plot Profile” option (Analyze ! Plot Profile) to
get the fluorescent intensity profile along the line.
(d) A new window as shown in Fig. 9 (b or c) will appear.
Copy the fluorescent intensity profile data and paste it in a
new “.txt” file.
(e) Scroll back to the previous image (25th image or the last
image of the pre-bleach series). The “straight line” will
stay on the image stack. Repeat the steps (iii and iv).
(f) MATLAB script recognizes pre-bleach and post-bleach
line intensity profiles by the “file name” of the text files
and they should be named accordingly. The text file con-
taining post-bleach and pre-bleach fluorescent intensity
profile should be suffixed with “ps” and “pe,” respectively.
Remaining characters, especially the file numbers in the

file name, should be same for both the files. Refer to the
sample datasets in the “Effective Radius” folder.
(g) Repeat steps (i – vi) for at least 10 FRAP image stacks and
save those intensity profile files in a folder. (Note: the
length of the line drawn across the bleach profile at step
(ii) should be the same for all FRAP image stacks).
(h) There are two MATLAB scripts (“FindRe” and
“hdrload”) in the “Effective Radius” folder. Copy the
scripts and the intensity files (saved in step vii) to the M
ATLAB home directory folder.
(i) Run “FindRe” script in the MATLAB; an image with fit
with average bleach intensity profile will appear and the
effective radius can be calculated from the bleach profile as
shown in Fig. 9d.
5. Input the following details in the “FRAPAnalysis” script as
shown in Fig. 10.
(a) FRAP data files folder location in the “datapath” (line 11).
(b) The frame number of first photobleach image (line 20).
Fig. 10 Variables to be entered in the “FRAPAnalysis” Matlab script. FRAP data files folder location is
mentioned the “datapath” (red box 1). The frame number of the first post-bleach image and the frame
number at which photobleach correction should be started are mentioned in line 20 and 21, respectively (red
box 2). The diameter of the bleach ROI is 25 pixels and the pixel size is 0.12 (red box 3). The effective radius is
2.2 as calculated in step 5 (red box 4)
(c) The frame number at which photobleach correction to be

started (line 21). (Note: We used mGFP and some of the
mGFP proteins underwent triplet state at the start of the
FRAP experiment. We acquired 25 pre-bleach images and
omitted first 15 images to exclude the triplet state arte-
facts. So, the photobleach correction starts at 16. If there
is no such problem of triplet state artefacts, 10 pre-bleach
images are sufficient. In that case, photobleach correction
will start from 1.)
(d) Diameter of the “Bleaching ROI” and “pixel size” (line 26).
(e) Effective radius of the “Bleaching ROI” as calculated in
the previous step (line 29).
(f) Specify the range of pre-bleach images (line 48): If you
have acquired 25 pre-bleach images and want to normal-
ize your data using the last 10 pre-bleach images (i.e.,
16–25 frames), then it should be “Dimen ¼ blcor(16:25)”.
Adjust the numbers to the range of per-bleach frames
needed for data normalization.
6. Once these variables are entered in the MATLAB script, click
the “Run” button in MATLAB. MATLAB will automatically
extract FRAP data from the excel files and analyze them.
7. Once analysis is complete, a new folder named “Analysis” will
be present in the FRAP data files folder.
8. The “Analysis” folder will contain two folders, “SingleFit” and
“DoubleFit,” and two Microsoft Excel files, “Consolidated”
and “FitResults.”
9. The “DoubleFit” folder will contain all the two-component fit
images for individual FRAP measurements. Open the “Dou-
bleFit” folder and check for the proper fitting of the FRAP
curves (compare to Fig. 3). If there are any improperly fit
images, note the file name of the FRAP experiment and exclude
it from further FRAP analysis.
10. The Microsoft Excel document named, “Consolidated” will
contain processed FRAP intensity values (corrected for acqui-
sition photobleach and background signals) for individual
FRAP measurements. Average those values to get the final
FRAP recovery curve.
11. Another Microsoft Excel document named “FitResults” will
contain the dynamics data, such as immobile fraction, half-time
to recover, diffusion constant values for both single and double
component fits.
12. Continue FRAP analysis in a Microsoft Excel sheet containing
“Two component fit” values for FRAP studies involve two
components for fluorescence recovery, such as transcription
Table 2
Abbreviations in the “FitResults” Microsoft Excel sheet are explained (scripts available upon request)
Abbreviation Data type

y0 Fluorescence intensity value at the “Bleaching ROI” of the first photobleach image
A1 Amplitude of the fast diffusing population
A2 Amplitude of the slow diffusing population
I0 Fluorescence intensity value of the first image of the FRAP image stack at the “Bleaching
ROI”
tau1 Time constant of fast diffusing population (A1)
tau2 Time constant of slow diffusing population (A2)
thalf1 Half-time to recover for fast diffusing population calculated from “tau1”
thalf2 Half-time to recover for slow diffusing population calculated from “tau2”
IF Immobile fraction
D11 Effective diffusion constant of fast diffusing population (A1)
D21 Effective diffusion constant of slow diffusing population (A2)
Ratio A1/A2 ratio
factors. Data name abbreviation in the “FitResults” Microsoft

Excel sheet is described in Table 2.
13. Some of the data in the “FitResults” Microsoft Excel sheet are
for reference purpose only. Useful dynamics data are thalf1,
thalf2, IF, diffusion constants, and ratio of A1/A2.
14. Calculate mean and standard deviation for the useful dynamics
data in the Microsoft Excel sheet itself. These data can be used
for tabular visualization of the FRAP data.
3.4 Data FRAP measures a variety of dynamics data at the single cell level
Visualization and resolution. Proper graphical representation of the FRAP data
Statistics would effectively communicate underlying biological information,
such as different aspects of protein dynamics and presence of differ-
3.4.1 Data Visualization ent cell populations, etc. We use Origin® (93E) software (Origi-
nLab, Northampton, Massachusetts, USA) for making boxplots,
interval plots, scatterplots, and statistical analysis. Fluorescence
intensity values are plotted against the function of time in the
X-axis and shown in the scatterplot (see Fig. 11a). Interval plots
or boxplot with individual data points is a good choice to visualize
protein dynamics data (see Fig. 11b, c).
3.4.2 Statistics Our data does not show a normal distribution because of the
cellular heterogeneity. We therefore used a Mann-Whitney U-test
to calculate statistical significance.
Fig. 11 Visualizing protein dynamics data as measured by FRAP. (a) Scatter plot showing fluorescence
recovery curves. (b) Boxplot and (c) interval plot showing immobile fraction data
4 Notes
1. Maintaining physiological pH and osmolarity during imaging

is important to keep the cells alive and normal. We used Tyr-
ode’s solution [28] buffered with HEPES during imaging.
2. In case of using media without phenol red as imaging media,
either CO2 supply or HEPES buffering is required during
imaging. Avoid incubating cells (in well plates) in the imaging
buffer in the CO2 incubator.
3. Wash the coverslips with “90% EtOH + 10% glacial acetic acid”
solution and rinse with 70% EtOH. Autoclave and keep sterile
until use (see Subheading 2.2.1 day 1).
4. We used a silicone gasket with 1.5 mm thickness (cat#: 70465-
2R2, EMS, USA). We did cut the small circular well and made
the circle slightly bigger to allow for a larger imaging area (see
Subheading 2.2.1, item 1).
5. Coverslips in the 24-well plates can be easily lifted using a 10 μl
tip and a tweezer. Practice this mounting step before doing the
procedure with cells. This step needs prior practice! Gently
press the coverslip with a fibreless tissue paper (such as Kim-
berly-Clark® Professional, Subheading 2.2.1, item 3).
6. Check if the XY drift occurs with every FRAP measurement; if
it occurs frequently, check whether the glass side with coverslip
is properly mounted in the stage inlet. Also, if there is a XY
drift, mark it in the corresponding Microsoft Excel file contain-
ing FRAP data, so that during data validation, corresponding
FRAP image stack can be readily picked for image registration
(see Subheading 3.2.2).
7. If the “PFS” is turned on, usually the Z drift will not occur (see
Subheading 3.2.2).
8. You can create shortcuts for “Template Matching” and “Plot

Z-axis profile” options in the Fiji (Plugins ! Shortcuts ! Add
Shortcuts ! Select preferred shortcut key and command). This
will help to save time during data validation (see Subheading
3.2, step 5).
Acknowledgements
We thank Anne K. Kenworthy for kindly providing the MATLAB

script to calculate the effective diffusion. Their MATLAB script is
partly incorporated in our script. We thank Samantha L. Schwartz
and Diane S. Lidke for their FRAP script and useful discussion. We
thank Irene Siemerink-Konings for reading this article for the
clarity check and suggestions for improvement.
References
1. Axelrod D, Koppel DE, Schlessinger J, factor activity. Biochimica et Biophysica Acta

Elson E, Webb AW (1976) Mobility measure- Gene Regul Mech 1862(1):107–117. https://
ment by analysis of fluorescence photobleach- doi.org/10.1016/j.bbagrm.2018.11.001
ing recovery kinetics. Biophys J 9. Jinek M, Chylinski K, Fonfara I, Hauer M,
16:1055–1069. https://doi.org/10.1016/ Doudna JA, Charpentier E (2012) A program-
S0006-3495(76)85755-4 mable dual-RNA–guided DNA endonuclease
2. Tsien RY (1998) The green fluorescent pro- in adaptive bacterial immunity. Science 337
tein. Annu Rev Biochem 67:509–544. (6096):816. https://doi.org/10.1126/sci
https://doi.org/10.1146/annurev.biochem. ence.1225829
67.1.509 10. Bi W, Deng JM, Zhang Z, Behringer RR, de
3. Fritzsche M, Charras G (2015) Dissecting pro- Crombrugghe B (1999) Sox9 is required for
tein reaction dynamics in living cells by fluores- cartilage formation. Nat Genet 22:85. https://
cence recovery after photobleaching. Nat doi.org/10.1038/8792
Protoc 10(5):660–680. https://doi.org/10. 11. Rayan G, Guet J-E, Taulier N, Pincet F,
1038/nprot.2015.042 Urbach W (2010) Recent applications of fluo-
4. Sprague BL, McNally JG (2005) FRAP analysis rescence recovery after photobleaching
of binding: proper and fitting. Trends Cell Biol (FRAP) to membrane bio-macromolecules.
15(2):84–91. https://doi.org/10.1016/j.tcb. Sensors (Basel, Switzerland) 10
2004.12.001 (6):5927–5948. https://doi.org/10.3390/
5. Pan Y, Tsai C-J, Ma B, Nussinov R (2010) s100605927
Mechanisms of transcription factor selectivity. 12. Blaney Davidson EN, Vitters EL, van Lent PL,
Trends Genet 26(2):75–83. https://doi.org/ van de Loo FA, van den Berg WB, van der
10.1016/j.tig.2009.12.003 Kraan PM (2007) Elevated extracellular matrix
6. Spivakov M (2014) Spurious transcription fac- production and degradation upon bone mor-
tor binding: non-functional or genetically phogenetic protein-2 (BMP-2) stimulation
redundant? BioEssays 36(8):798–806. point toward a role for BMP-2 in cartilage
https://doi.org/10.1002/bies.201400036 repair and remodeling. Arthritis Res Ther 9
7. Orphanides G, Reinberg D (2002) A unified (5):R102. https://doi.org/10.1186/ar2305
theory of gene expression. Cell 108 13. Shaner NC, Patterson GH, Davidson MW
(4):439–451. https://doi.org/10.1016/ (2007) Advances in fluorescent protein tech-
S0092-8674(02)00655-4 nology. J Cell Sci 120(24):4247. https://doi.
8. Govindaraj K, Hendriks J, Lidke DS, org/10.1242/jcs.005801
Karperien M, Post JN (2019) Changes in fluo- 14. Nagy P, Vereb G, Post JN, Friedl€ander E, Szöl-
rescence recovery after photobleaching loősi J (2005) Novel single cell fluorescence
(FRAP) as an indicator of SOX9 transcription approaches in the investigation of signaling at
the cellular level. In: Damjanovich S
(ed) Biophysical aspects of transmembrane sig- reduce cell contraction levels. Lab Chip 11
naling. Springer, Berlin, Heidelberg, pp 33–70. (13):2231–2240. https://doi.org/10.1039/
https://doi.org/10.1007/3-540-26511-2_2 C0LC00641F
15. Zacharias DA, Violin JD, Newton AC, Tsien 25. Koulouras G, Panagopoulos A, Rapsomaniki
RY (2002) Partitioning of lipid-modified MA, Giakoumakis NN, Taraviras S, Lygerou
monomeric gfps into membrane microdomains Z (2018) EasyFRAP-web: a web-based tool
of live cells. Science 296(5569):913. https:// for the analysis of fluorescence recovery after
doi.org/10.1126/science.1068539 photobleaching data. Nucleic Acids Res 46
16. Shaner NC, Steinbach PA, Tsien RY (2005) A (W1):W467–W472. https://doi.org/10.
guide to choosing fluorescent proteins. Nat 1093/nar/gky508
Methods 2:905. https://doi.org/10.1038/ 26. Blumenthal D, Goldstien L, Edidin M, Gheber
nmeth819 LA (2015) Universal approach to FRAP analy-
17. Cardarelli F, Tosti L, Serresi M, Beltram F, sis of arbitrary bleaching patterns. Sci Rep
Bizzarri R (2012) Fluorescent recovery after 5:11655. https://doi.org/10.1038/
photobleaching (FRAP) analysis of nuclear srep11655
export rates identifies intrinsic features of 27. Day CA, Kraft LJ, Kang M, Kenworthy AK
nucleocytoplasmic transport. J Biol Chem 287 (2012) Analysis of protein and lipid dynamics
(8):5554–5561. https://doi.org/10.1074/ using confocal fluorescence recovery after
jbc.M111.304899 photobleaching (FRAP). Curr Prot Cytom C
18. Wu B, Piatkevich KD, Lionnet T, Singer RH, HAPTER:Unit2.19-Unit12.19. https://doi.
Verkhusha VV (2011) Modern fluorescent pro- org/10.1002/0471142956.cy0219s62
teins and imaging technologies to study gene 28. Lidke DS, Nagy P, Heintzmann R, Arndt-Jovin
expression, nuclear localization, and dynamics. DJ, Post JN, Grecco HE, Jares-Erijman EA,
Curr Opin Cell Biol 23(3):310–317. https:// Jovin TM (2004) Quantum dot ligands pro-
doi.org/10.1016/j.ceb.2010.12.004 vide new insights into erbB/HER receptor–-
19. Nissim-Rafinia M, Meshorer E (2011) Photo- mediated signal transduction. Nat Biotechnol
bleaching assays (FRAP & FLIP) to measure 22:198. https://doi.org/10.1038/nbt929
chromatin protein dynamics in living embry- 29. Mueller F, Mazza D, Stasevich TJ, McNally JG
onic stem cells. J Vis Exp 52:2696. https:// (2010) FRAP and kinetic modeling in the anal-
doi.org/10.3791/2696 ysis of nuclear protein dynamics: what do we
20. Goldring MB, Birkhead JR, Suen LF, Yamin R, really know? Curr Opin Cell Biol 22
Mizuno S, Glowacki J, Arbiser JL, Apperley JF (3):403–411. https://doi.org/10.1016/j.
(1994) Interleukin-1 beta-modulated gene ceb.2010.03.002
expression in immortalized human chondro- 30. Beaudouin J, Mora-Bermúdez F, Klee T,
cytes. J Clin Invest 94(6):2307–2316. Daigle N, Ellenberg J (2006) Dissecting the
https://doi.org/10.1172/JCI117595 contribution of diffusion and interactions to
21. Post JN, Lidke KA, Rieger B, Arndt-Jovin DJ the mobility of nuclear proteins. Biophys J 90
(2005) One- and two-photon photoactivation (6):1878–1879. https://doi.org/10.1529/
of a paGFP-fusion protein in live Drosophila biophysj.105.071241
embryos. FEBS Lett 579(2):325–330. 31. Fritzsche M, Lewalle A, Duke T, Kruse K,
https://doi.org/10.1016/j.febslet.2004.11. Charras G (2013) Analysis of turnover dynam-
092 ics of the submembranous actin cortex. Mol
22. Siggia ED, Lippincott-Schwartz J, Bekiranov S Biol Cell 24(6):757–767. https://doi.org/10.
(2000) Diffusion in inhomogeneous media: 1091/mbc.E12-06-0485
theory and simulations applied to whole cell 32. Oh D, Zidovska A, Xu Y, Needleman Daniel J
photobleach recovery. Biophys J 79 (2011) Development of time-integrated multi-
(4):1761–1770. https://doi.org/10.1016/ point moment analysis for spatially resolved
S0006-3495(00)76428-9 fluctuation spectroscopy with high time resolu-
23. Tseng Q (2011) Template matching and slice tion. Biophys J 101(6):1546–1554. https://
alignment - ImageJ plugins. https:// doi.org/10.1016/j.bpj.2011.08.013
sitesgooglecom/site/qingzongtseng/tem 33. Kim SA, Heinze KG, Schwille P (2007) Fluo-
plate-matching-ij-plugin. Accessed Jan 2020 rescence correlation spectroscopy in living
24. Tseng Q, Wang I, Duchemin-Pelletier E, cells. Nat Methods 4:963. https://doi.org/
Azioune A, Carpi N, Gao J, Filhol O, Piel M, 10.1038/nmeth1104
Théry M, Balland M (2011) A new micropat- 34. Herrick-Davis K, Grinde E, Cowan A, Mazur-
terning method of soft substrates reveals that kiewicz JE (2013) Fluorescence correlation
different tumorigenic signals can promote or spectroscopy analysis of serotonin, adrenergic,
muscarinic, and dopamine receptor dimeriza- cycle and G-actin regulation of polymer resto-
tion: the oligomer number puzzle. Mol Phar- ration. Proc Jpn Acad Ser B Phys Biol Sci 86
macol 84(4):630–642. https://doi.org/10. (1):62–83. https://doi.org/10.2183/pjab.
1124/mol.113.087072 86.62
35. Mendoza MC, Besson S, Danuser G (2012) 37. Watanabe N, Mitchison TJ (2002) Single-
Quantitative fluorescent speckle microscopy molecule speckle analysis of actin filament turn-
(QFSM) to measure actin dynamics. Curr Pro- over in lamellipodia. Science 295
toc Cytom Chapter 2:Unit2.18. https://doi. (5557):1083–1086. https://doi.org/10.
org/10.1002/0471142956.cy0218s62 1126/science.1067470
36. Watanabe N (2010) Inside view of cell locomo-
tion through single-molecule: fast F/G-actin
Chapter 10
Quantitative Molecular Models for Biological Processes:

Modeling of Signal Transduction Networks with ANIMO
Sakshi Khurana, Janet Huisman, Stefano Schivo, and Janine N. Post
Abstract
Computational modeling of biological networks is increasing in popularity due to the increased demand for
understanding biological processes. This understanding requires integration of a variety of experimental
data that allows understanding of complex mechanisms regulating cell and tissue function. However, the
mathematical complexity of many modeling tools have thusfar prevented broad adaptation and effective use
by molecular biologists. In this chapter, we show by example how one can start building a model in
ANIMO and how to adapt the model to experimental data. We show how this model can be used for
simulating network activities, testing hypotheses, and how to improve the model using wet-lab data.
Key words Modeling tool, Computational model, Signal transduction network, Signaling pathways,
Signaling cross talk, WNT, NFkB, TGF beta
1 Introduction
1.1 The Need of Over the past decades, advancement in high-throughput technol-
Modeling for Biological ogies have led to a rapid accumulation of genomic and proteomic
Networks data. It has provided a great wealth of information by studying the
gene regulation and gene-protein interactions in health, as well as
in pathological states. However, not all data are utilized in an
efficient manner, and changes in cellular signaling are often incom-
pletely explored. This necessitates a shift from the central dogma of
“hypothesis to experiment to models” toward “big data to models
to hypothesis to new experiment to additional data and more
inclusive models” [1]. One effective way of utilizing information
obtained from genomic and proteomic data is by building compu-
tational models. These models can be an option for storing large
and complex data as well as for hypothesis generation.
Models can be used as a tool to combine data from existing
literature. For example, data obtained on a protein studied in
context of one disease can be used to build a model. This model
can then be used to predict the mechanism of this protein in the
141
142 Sakshi Khurana et al.
context of another disease by adding tissue-specific proteins and

genes. Later, these predictions can be confirmed by wet lab experi-
ments, thereby validating the model predictions and obtaining
important information about the molecular mechanism involved
in the latter tissue. If similar experiments were performed by wet lab
experiments only, then one would have to test many concentrations
and time points for each molecule of interest, leading to longer
time investment to understand the mechanism.
Recently, our group presented a model that was built to under-
stand the signaling cross talk between the wingless related integra-
tion site (WNT) and Interleukin 1 beta (IL1β) signaling pathways
in osteoarthritis. The model was based on a study conducted on
cancer cells, showing that IL1β induced nitric oxide
(NO) production, which upregulated WNT signaling [2]. Our
model predicted that these two pathways are linked in chondrocytes
via induced nitric oxide synthase (iNOS), which was confirmed by
wet lab experiments in human chondrocytes [3]. We later used this
knowledge to incorporate the IL1β into a larger network of 7 signal
transduction pathways that regulate cartilage formation and
homeostasis. In that large network, which we called ECHO (for
Executable CHOndrocyte), we show that healthy cartilage is par-
tially protected against the effects of IL1β exposure due to the
presence of the WNT antagonists DKK1 and FRZB [4]. This way
computational models can be used to find novel cross talk between
pathways in a relatively short time span.
In addition to storage of data or as predictive tool, models can
be used as an alternative for the testing of drugs in animals. Prelim-
inary attempts are already made to limit the animal testing by
developing software that can predict the outcome of various drug
screening assays. In this direction, researchers have developed soft-
ware to test drug compounds and found 87% accuracy with the
animal testing data [5]. Environmental protection agencies have
announced that any funding to mammalian testing will be stopped
by 2035 [6, 7], hence necessitating a step toward more software
developments as an alternative for animal testing.
1.2 Building a Model Computational models can graphically represent biological net-
works based on information obtained from literature, in combina-
tion with a formalization of the network in terms of activity. In
contrast to small networks, for large and complex networks it is
impossible to add all the components of the network. For complex
networks, it would be ideal to include at least the most important
molecules. However, it is not always easy to figure out which
molecules play key roles in the functioning of the network. Many
biological events are, in effect, changes in activity, comparable to
switches / dimmers. For example, changes in concentration, post-
translational modifications (i.e., phosphorylation, methylation) or
localization of a protein, or changes in gene expression are causal
Modeling with ANIMO 143
factors for downstream effects. Therefore, the state or concentra-

tion of molecules can be described in terms of activity. For many
molecules the actual biological activity is not measured, especially
not in the context of the cell. Therefore, the more active a mole-
cule, the more it will affect regulation of downstream factors. To
gain insight into important events led by the molecular interac-
tions, the following guidelines should be considered:
1. Choose to measure a molecule that shows differential down-
stream effects for the modeling network activity, i.e., that can
be used as an output for validating and / or optimization of the
network model.
2. Include overlapping treatments to normalize experimental data
between days or assay batches.
3. Include a positive control for each of the measurements to get
an indication of the potential maximum activity of the mole-
cules in the network. This allows scaling of network activities
between 0 and 100% to construct a nondimensional model.
This way, one can omit the need for precise intracellular
concentrations.
4. Include a negative control (t ¼ 0) for background activity
levels.
5. To study the dynamics of network in an efficient way, it is
important to include multiple time-points. Single time-point
measurements provide poor insight into the dynamic behavior
of a system. Consider the following factors to decide how many
time-points should be measured. Ideally, measurements are
taken from time 0 until the system reaches a steady state.
When peak dynamics are expected, 3 time points are the abso-
lute minimum: one before the peak, one as close as possible to
the peak, and one after the peak. Time range for a biological
network depends on the direct effects or indirect effects of
upstream molecules. In case of signal transduction, direct or
primary effects take 240/480 min whereas indirect effects
might take 24–48 h. High order indirect effects might even
take weeks or months [8]. These higher order effects may be,
for example, effects regulating gene expression of genes that
encode growth factors and cytokines, whose expression will be
new input for the network.
1.3 Precision of Generally, the aim of models is to replicate the behavior of

Models biological systems based on the known properties of individual
components [9]. The precision of a model, i.e., the detail in
which the biological network is represented in terms of number
of molecules and interactions, depends on the availability of experi-
mental data, the need for detailed information, and available
computational tools. Models suffer less from experimental issues
like noise and resolution, thus providing more precise insights at

the cellular and molecular level. Specifically, models based on ordi-
nary differential equations, which are more parameter intensive,
represent biological interactions more precisely [10]. Whereas
modeling tools, such as models based on logic descriptions (Bool-
ean, Fuzzy) as well as the modeling tool we will use in this chapter,
ANIMO, suffer more with precision errors as they are based on
single parameter k-values [11].
1.4 Comparison of Various computational tools are available for building models that
some Modeling define the dynamics of biological networks. In this part, we describe
Formalisms and Tools the formalism that is underlying some of the most-used modeling
tools. Tools based on formalisms described by Boolean and fuzzy
logic models are often used. Tools that belong to this category,
such as CytoCopteR [12], GINsim [13], and BooleSim [14], are
based mainly on discrete transitions and can be used to model
interactions among a large variety of proteins. These tools make it
easier to perform model building and model validation, and can be
used for predictions. On the other side of the tool-spectrum,
ordinary differential equations (ODEs) make a purely continuous
model and can efficiently represent biochemical reactions for large
and complex networks where mass-action approximations are
appropriate [10, 11, 15]. Examples of tools using ODEs are
Odefy [16], COPASI [17], and CellDesigner [18].
An actively growing field for computing graphical representa-
tions of biological networks through literature or high-throughput
data has been gaining interest in the last few years [10]. In this field,
biological networks are typically represented with the help of
graphs, where each component, e.g., gene or protein, is repre-
sented by nodes and potential interactions by edges. Edges can
typically represent a wide range of stimulatory or inhibitory inter-
action modes from direct physical binding to correlated gene
expression or phosphorylation by kinase, etc. For the purpose of
making computational modeling of dynamic biological networks
available to researchers without extensive mathematical training, we
previously developed ANIMO: Analysis of Networks with Interac-
tive Modeling [8, 11, 19].
ANIMO is a tool that helps a biological researcher to build and
analyze a protein interaction network, providing a visual represen-
tation of the activity profile over time. In order to formalize and
correctly analyze the models, ANIMO networks are based on a
Timed Automata model, that can be described as piecewise linear
approximation of ODE models. More information about ANIMO
and the comparison with other tools, as well as example signal
transduction models, can be found in Schivo et al. [11].
In this chapter, we want to give a working example of how to
build a computational model. In this model one can add data and
test hypotheses. Wet-lab experiments can be used to validate and
improve the model, as also described in this chapter.
2 Materials
2.1 Computational To run ANIMO, a desktop or laptop computer is needed with the
Materials following software installed:
1. Java (java.com).
2. UPPAAL (www.uppaal.org) (see Note 1).
3. Cytoscape (www.cytoscape.org) (see Note 2).
An extensive user manual for installation and for creating a
network can be found here: http://fmt.cs.utwente.nl/tools/
animo/
The software to run Java-based programs is provided for free by
Oracle. More information is available on the java.com website.
Cytoscape is an open source project released under the terms of
the GNU Lesser General Public License. UPPAAL is developed by
a collaboration by the universities of Uppsala (Sweden) and Aal-
borg (Denmark), and is free for noncommercial applications in
academia only. All software works under Windows, Mac, and the
most common GNU/Linux OS (see Note 3).
Models used in this chapter, as well as raw data files, can be
found online via:
https://www.utwente.nl/en/tnw/dbe/research/additional-
materials/
2.2 Wet-Lab 1. 96-Well plates (BD biosciences, with white walls and a clear
Materials bottom such that they are suitable for cell culture and allow for
luciferase activity measurements).
2. Cells that are being studied (C2C12, mouse myoblasts, ATCC
®
CRL-1772™).
3. Cell culture media (DMEM), Thermo Fisher.
4. Fetal bovine serum (FBS), Thermo Fisher.
5. Penicillin-Streptomycin (Pen/Strep), Thermo Fisher.
6. Phosphate-buffered saline (PBS), Thermo Fisher.
7. Trypsin, Thermo Fisher.
8. Lipofectamine™ 2000 Transfection Reagent, Thermo Fisher
Scientific.
9. Reporter DNA (16xTOPFlash (TCF Reporter Plasmid
21-170, Sigma Aldrich).
10. Renilla DNA (pRL-CMV, Promega,for normalization of trans-
fection efficiency).
11. Opti-MEM reduced serum medium (ThermoFisher).
12. Lipofectamine LTX (ThermoFisher).
13. Stimulation factors TGF-β3 (R&D systems, 8420-B3-005),

BIO (Sigma, B1686-5MG), and IL-1β (BioLegend, 575102).
14. Dual-Glo® kit (Promega, E2940).
15. Varioskan™ LUX multimode microplate reader
(ThermoFisher).
16. Nuclease-free 1.5 mL Eppendorf tubes.
3 Methods
3.1 Preliminary 1. Determine the pathways that need to be studied. In this exam-
Model and Hypotheses ple, the cross talk between the canonical WNT, NFκB, and
TGF-β pathways will be determined. These pathways play an
important role in various cellular processes, such as cell survival,
differentiation, and growth, but their cross talk is not yet fully
understood. The most important molecules and interactions in
these pathways are shown in Fig. 1 [20–22].
2. In this example, the model building was initiated by placing
nodes in ANIMO for the compounds that can be used to
stimulate each of these pathways (see Fig. 2a). The shape of
the nodes corresponds to the type of molecule they represent
(cytokine, receptor, transcription factor, gene, etc.) while the
Fig. 1 Cross talk between the canonical WNT, NFκB, and TGF-β pathways as found in literature
[20–22]. These pathways play an important role in various cellular processes, such as cell survival,
differentiation, and growth
color indicates the activity of each node. A red color represents

a low activity and a green node suggests high activity of the
molecule. Literature research is necessary to determine the
most important ligands for the network as seen in Fig. 1. In
general, one would start modeling the canonical pathways,
unless it is already described that noncanonical pathways are
important for cross talk between pathways. WNT is the general
ligand to indicate all WNT ligands that stimulate the canonical
β-catenin pathway, e.g., WNT3A, WNT5b, and WNT10a
[23]. BIO [24], 6-bromoindirubin-30 -oxime, is a potent,
reversible, and ATP-competitive GSK-3α/β inhibitor, and is
used to stimulate the Wnt pathway (see Note 4). Interleukin
1 beta (IL-1β) activates the NFκB pathway [25] and TGF-β3
stimulates the TGF-β pathway [26] (see Note 5).
3. Next, a literature search will help to determine which down-
stream transcription factors and genes are up- or downregu-
lated when each of these compounds is added to the cell. This
will describe each general pathway (Wnt, NFκB, and TGF-β)
without its cross talk. For example, it was found that for the
Wnt pathway, the transcription cofactor β-catenin binds to the
transcription factors TCF-LEF [27] to initiate transcription of
metalloproteinases (MMPs 2, 7 and 9 [28]) and that TGF-β3
binds to a receptor of the TGF-β superfamily (Activin receptor-
like kinases 1–7, ALK [29]), which promotes the formation of
the Smad2/3/4 complex, which in turn increases transcription
of the genes with a SMAD Binding Element (SBE) in the
promotor [30–32]. Examples of these genes are
TGF-β-induced protein (TGFBIP), transmembrane prostate
androgen induced RNA (TMEPA), and matrix metalloprotease
2 (MMP2) [33].
4. Add the transcription factors and genes to the model generated
with the ANIMO software (see Fig. 2b). Add edges to describe
the direction of the interaction, e.g., activation or inhibition,
and include dynamics as found in literature (see Notes 6–8) [8].
5. More extensive literature research is then used to investigate the
cross talk between each of the pathways. Additional nodes can
be added if necessary and nodes that overcomplicate the model,
for example, because they only interact with one other protein
in the same pathway, can be removed if this does not interfere
with the dynamics of the downstream nodes. Interactions can
be added in the ANIMO model as edges (see Fig. 2c). In our
example, it was found that NFκB inhibits β-catenin [34, 35] and
that Smad7 inhibits NFκB, β-catenin and the TGF-β receptors
1 (TGFβR1, ALK5) and 2 (TGFβR2) [36–39].
Fig. 2 Flow of generating an ANIMO model to simulate pathway interactions. (a) Start with compounds that can
be used to stimulate each of these pathways. (b) Add transcription factors and genes that are up- or
downregulated when each of these compounds is added to the cell. (c) Add interactions that describe the
cross talk between the studied pathways
6. After completing the model with all nodes and edges, a simula-
tion can be run to generate hypotheses (see Fig. 3). In our
example, we are interested in the influence of the stimuli on
the activation of TCF/LEF and subsequent target gene expres-
sion. Depending on the chosen scenario and reaction con-
stants, different hypotheses will be generated (see Note 9).
The rate constants can, at this stage, be manipulated until the
results roughly correspond to what is described in literature.
However, it is recommended to do the fine-tuning of the
model after the wet-lab experiments.
Fig. 3 Hypotheses of the activation of the TCF/LEF luciferase reporter with different stimulation factors
generated using the ANIMO model as shown in Fig. 2c
7. In Fig. 3, the hypotheses of the activation of a TCF/LEF

luciferase reporter are shown with different stimulation factors.
It was chosen to focus on TCF/LEF since this transcription
factor is found downstream of the canonical WNT pathway and
its expression is dependent on the cross talk between all three
modeled pathways, as is shown in Fig. 2c. The activity of
TCF/LEF is depicted with colors, where red indicates a low
activity and green suggests high activity of the transcription
factor. As can be seen in Fig. 3, the model predicts that TGF-β3
by itself will not activate TCF/LEF, while WNT will strongly
activate TCF/LEF. IL-1β will only partially activate TCF/LEF
due to its direct activation of TCF/LEF while at the same time
inhibiting the β-catenin production. A combination of TGF-β3
and WNT is predicted to fully activate TCF/LEF, indicating
that TGF-β3 does not inhibit WNT in activating TCF/LEF.
Stimulating with both TGF-β3 and IL-1β is not sufficient to
activate TCF/LEF and adding both WNT and IL-1β only
partially activates TCF/LEF. Lastly, it is expected that a stimu-
lation with all three stimuli (TGF-β3, WNT, and IL-1β) will
lead to activation of TCF/LEF.
3.2 Wet-Lab Model predictions can be validated either by looking at existing

Experiments literature that was not used for model-building, or by performing
wet-lab experiments. We chose the latter, since we want to validate
both the model topology, i.e., the cross talk between the pathways
of interest, and the model parameters.
1. Many different techniques can be used to determine the up- or
downregulation of gene transcription and protein translation.
To be able to get an indication of the TCF/LEF expression and
validate the hypotheses as shown in Fig. 3, a reporter assay can
be used. This assay measures the activity of a specific gene
promoter by transfection of a vector containing the promotor
of our gene of interest upstream of a reporter gene, such as
Table 1
TCF/LEF-luciferase expression in C2C12 mouse myoblasts for different stimulation factors (10 ng/mL
TGF-β3, 1 μM BIO, and 10 ng/mL IL-1β) at different time points (17 h, 22 h, 41 h, and 48 h after
stimulation). N ¼ 4
17 hours 22 hours 41 hours 48 hours
Condition Mean Std Mean Std Mean Std Mean Std

1 Control 1.000 1.186 1.000 0.956 1.000 0.952 1.000 0.922
2 TGF-β3 1.620 1.584 1.538 1.455 0.846 0.760 1.309 1.032
3 BIO 3.320 3.367 0.529 0.445 0.380 0.360 0.557 0.515
4 IL-1β 1.122 1.174 0.875 0.749 0.532 0.556 0.909 0.768
5 TGF-β3 + BIO 5.007 4.623 1.108 0.789 0.865 1.027 1.345 1.136
6 TGF-β3 + IL-1β 1.898 2.337 1.322 0.995 1.154 1.079 1.749 1.456
7 BIO + IL-1β 4.766 5.161 0.871 0.711 0.414 0.393 0.871 0.943
Fig. 4 Luciferase expression under control of the TCF/LEF promoter in C2C12 mouse myoblasts for different
stimulation factors (TGF-β3, BIO, and IL-1β) at different time points (17 h, 22 h, 41 h, and 48 h after
stimulation). N ¼ 4
luciferase, green fluorescent protein, or β-galactosidase. In our

example, the expression of TCF-LEF is studied by a reporter
assay using bioluminescent luciferase (see Note 10).
2. For the reporter assay, 96-well plates are seeded with C2C12
mouse myoblasts at a density of 20,000 cells per well in a
volume of 100 μL cell culture medium (DMEM supplemented

with 1% v/v FBS). C2C12 cells are a cell line that can differen-
tiate into myoblasts and myocytes (muscle cells) (see Note 11).
For each time point a separate plate was used and the outer
wells are filled with sterile PBS to reduce evaporation of
medium.
3. For the transfection, lipofectamine (Lipofectamine™ LTX
Transfection Reagent, Thermo Fisher Scientific) is used with
some modifications, see Note 4.
4. For each well, 0.1 μg of reporter DNA (16xTOPflash and
0.005 μg of renilla DNA (pRL-CMV) is combined with
0.1 μL PLUS reagent and diluted into 5 μL of Opti-MEM in
a 1.5 mL tube. In a separate tube, for each well, 0.305 μL of
Lipofectamine is diluted into 5 μL of Opti-MEM.
5. The diluted DNA solution is transferred to the diluted Lipo-
fectamine solution, mixed well by pipetting, and incubated for
5 min at room temperature to form DNA-Lipofectamine
complexes.
6. 10 μL of DNA-Lipofectamine complexes is added to each well.
The cells are then incubated overnight at 37 C and 5% CO2. Is
this without adding DMEM+FBS after 3–6 h?
7. The next day, medium is replaced by fresh culture medium
(DMEM +1 v/v% FBS (see Note 12) + 1 v/v% Pen/Strep)
containing the stimulation factors (10 ng/mL TGF-β3, 1 μM
BIO, 10 ng/mL IL-1β or a combination of these, 75 μL per
well and in quintuplicate) and incubated at 37 C and 5% CO2
(see Note 13).
8. At four different time points (17, 22, 41, and 48 h after
stimulation), the reporter assay is performed. At each time
point, Dual-Glo® Luciferase Reagent (75 μL per well) is
added to the culture medium and incubated for 10 min.
9. After incubation, TCF/LEF-luciferase luminescence is
measured using a VarioSkan LUX. Emission firefly luciferase
measured at 550–570 nm.
10. Next, Dual-Glo® Stop & Glo® Reagent (75 μL per well) is
added to the culture medium and incubated for 10 min.
11. After incubation, renilla-luciferase luminescence is measured at
480 nm using the Varioskan LUX (ThermoFisher).
12. Repeat steps 7–11 for each different time point.
13. Analyze the obtained data (see Note 14). Use the renilla-
luciferase data to normalize for differences in transfection effi-
ciency. Next, average the values for identical conditions and
calculate the standard deviation (see Table 1 and Fig. 4).
14. In Fig. 4, the normalized luciferase expression under the con-

trol of the TCF/LEF promoter in C2C12 mouse myoblasts
can be seen. It is clearly visible that for all conditions, the
expression is highest at the first time point (17 h after stimula-
tion) and decreases at subsequent time points. It is expected
that this observed peak-behavior is due to the fairly short half-
life of the luciferase reporter. If BIO was used to stimulate the
cells, either by itself or in combination with other stimuli, an
increase in expression was observed compared to the control.
For TGF-β3 and IL-1β, no significant increase was seen in the
luciferase expression. It is unknown whether the expression
17 h after stimulation is maximum or if it had been even higher
at earlier time points. Additional wet-lab experiments at earlier
time points will have to be conducted to obtain this data.
3.3 Validation and 1. Identify the differences between the wet-lab results and the
Adjustment of Model hypotheses generated with the preliminary model. These can
be differences in expression (e.g., the wet-lab results show an
increase in expression for a specific gene while this gene is
downregulated in the preliminary model) or in time scale (see
Note 15). In our example, the expression levels correspond
fairly well with the hypotheses, except that peak behavior is
seen in the wet-lab results (see Fig. 4, expression is downregu-
lated after 17 h for all conditions), which was not yet included
in the preliminary model.
2. Add missing nodes and/or connections to the model. For
example, adding a specific node might explain certain behavior
as observed in the wet-lab experiments or a specific inhibition
dynamic can solve problems in the simulated expression. This
completely depends on the specific model and might not always
be necessary. An additional literature search might help identi-
fying missing nodes or connections.
3. Incorporate peak behavior in the model if this is not already
done. In our example, peak behavior was observed in the
experimental results, but was not yet incorporated in the
model. This peak behavior can be explained by the fact that
luciferase fluorescent intensity is measured, which has a fairly
short half-life. By adding auto-inactivation/ inhibitory loops to
each node, this peak behavior could be added to the model to
better fit the experimental data, as described before [8]. In
Fig. 5, the optimized model is shown for our example, which
is based on Fig. 2c and improved using wet-lab data.
4. Optimize the rate parameters k to fit the wet-lab data. This can
be done manually by changing each and every rate parameter
until the simulations match the data from the wet-lab experi-
ment. However, for larger models this can get very time con-
suming and complicated. To make this process a little easier,
Fig. 5 Optimized ANIMO model based on the model as shown in Fig. 2c. Wet-lab results are used to improve
the model. Auto-inactivation/inhibitory loops were added to each node to incorporate peak behavior in the
model
the function “optimize k-value(s)” in ANIMO can be used (see

Note 16). For our example, results after optimization can be
found in Fig. 6. It is also important to remember that the
network topology (i.e., the nodes currently included in the
model and their interactions) should have priority on parame-
ter values: if a network model fits a data set only with a very
specific parameter choice, it is unlikely that the current topol-
ogy of the model is a useful representation of the biological
process. This is because ANIMO interactions are a relatively
rougher abstraction of biochemistry than the model network’s
topology is.
5. Use more data to optimize the model even further. This data
can be obtained by performing more wet-lab experiments (for
example, by measuring more time points or using another
technique such as Western Blotting or qPCR to obtain data
on other nodes in the model) or by using data from other
research groups or papers.
6. In Fig. 6, a comparison can be seen between the wet-lab
experimental data (in blue) and the optimized ANIMO
model (in red). The general behavior of the model corresponds
with the wet-lab data for all tested combinations of stimuli.
However, to be able to improve the model even further, more
wet-lab data will be necessary. Furthermore, these results only
focus on the cross talk of TGF-β3 and IL-1β on the WNT
signaling pathway. To investigate the other cross talks that are
Fig. 6 Results of optimization of rate parameters k to obtain a model behavior (red) that corresponds with the
obtained wet-lab data (blue) for various stimuli
incorporated in the model, such as the effect on WNT and

IL-1β on the TGF-β pathway, a new wet-lab experiment will
need to be setup where, for example, an SBE-luciferase
reporter is used instead of a TCF/LEF-luciferase reporter.
With the current model, it is already possible to generate
hypotheses for this experiment. The cross talk between WNT
and the TGF-β pathway is especially interesting, since TGF-β
induces cartilage differentiation while WNT induces bone for-
mation. By better understanding the cross talk between these
pathways, as can be modeled with ANIMO, it is possible to
better control and direct the differentiation into either bone or
cartilage, which is of great importance in the field of tissue
engineering.
4 Notes
1. For UPPAAL installation: Unzip the downloaded file to a

known location: UPPAAL will be installed there. Complete U
PPAAL installation.
For macOS:
– Open the UPPAAL installation location in Finder.
– Drop the UPPAAL.App icon in your Applications folder.
– Copy the verifyta executable file to a known location. The
installation of UPPAAL is complete.
For windowsOS / LinuxOS:
– Open a console, type \cd PATH_TO_THE_UPPAAL_DI
RECTORY” and press Return; PATH_TO_THE_UPPAA
L_DIRECTORY is the path to the directory where you
installed UPPAAL.
– Type \java -jar uppaal.jar “and press Return.
– The license for UPPAAL will be automatically acquired, and
the main window of UPPAAL user interface will appear: you
may now close that window.
2. Installing ANIMO: ANIMO is free only for academic use. For
commercial licenses, please contact us at Stefano.schivo@ou.nl.
– Run Cytoscape.
– Click the menu command Apps - > App Manager: the App
Manager window will open.
– Click on ANIMO in the list of available apps and then on the
Install button. ANIMO will be automatically downloaded
and installed. You can close the App Manager window.
– ANIMO will automatically try to locate the verifyta execut-
able included in UPPAAL, which is needed to verify Timed
Automata models. If you know the location where you
installed UPPAAL, you can just stop the process (press on
the small X on the right), and indicate the location of the
verifyta executable. You can find it where it was copied in the
bin (bin-Linux, bin-Win32,. .. depending on your operating
system) directory inside the UPPAAL installation directory,
where it was unzipped. ANIMO is correctly installed and
ready to be used.
3. In case of problems accessing a web site with Microsoft Inter-
net Explorer, we advise to try with a different web browser
(such as Mozilla Firefox or Google Chrome) or to update
Internet Explorer.
4. BIO is used as a stimulus for the activation of the WNT

pathway. It acts by inhibiting glycogen synthase kinase 3β
(GSK3β) such that GSK3β cannot phosphorylate β-catenin.
Unphosphorylated β-catenin is not degraded and will therefore
accumulate in the cell, where it can interact with TCF/LEF
transcription factors to initiate transcription of target
genes [40].
5. For each node that is added in ANIMO, the following para-
meters need to be selected: a name to describe the compound,
a type (e.g., kinase, receptor, transcription factor, gene), the
maximum activity (usually kept at 100) and the initial activity
(usually 0 except for the stimulating compounds, which are
either 100 (added to the cell) or 0 (not added). The option
“plotted” needs to be selected to be able to run a simulation for
that node.
6. Try to make the model as simple as possible in this stage. Only
include the most important transcription factor(s) and focus on
one or two genes that are of interest to study.
7. For each edge that is added in ANIMO, the following para-
meters need to be selected:
the effect of the edge (activation or inhibition), a scenario
for the reaction kinetics and a reaction constant k. The descrip-
tion field can be used to include literature that supports each
specific interaction.
8. In this example, the following k-values were chosen for the
initial model (see Fig. 2c)
k-
From node To node Influence Scenario value
TGF-β3 Type 1/2 receptor Activation 1 0.004
Type 1/2 receptor Smad2/3 Activation 1 0.004
Smad2/3 Smad2/3/4 Activation 1 0.004
Smad2/3/4 SBE transcription Activation 1 0.002
factor
SBE transcription SBE luciferase Activation 1 0.001
factor
WNT β-catenin Activation 1 0.004
β-catenin TCF/LEF Activation 1 0.002
transcription factor
TCF/LEF TCF/LEF luciferase Activation 1 0.001
TCF/LEF luciferase CCND1 gene Activation 1 0.001
(continued)
k-
From node To node Influence Scenario value
IL-1β NFκB Activation 1 0.004
NFκB NFκBia gene Activation 1 0.001
Smad2/3 Smad7 Activation 1 0.004
Smad2/3/4 β-catenin Activation 1 0.004
Smad7 NFκB Inhibition 1 0.004
Smad7 β-catenin Inhibition 1 0.004
Smad7 Type 1/2 receptor Inhibition 2 0.004
β-catenin Type 1/2 receptor Inhibition 1 0.004
NFκB TCF/LEF Activation 1 0.002
NFκB Smad7 Activation 1 0.004
NFκB CCND1 gene Activation 1 0.001
NFκB β-catenin Inhibition 1 0.004
9. The choice for the scenario and reaction constant k have a large
effect on the results of the simulation. For the initial model,
most edges are set to scenario 1 (unless described differently in
literature). For the reaction constant it generally holds true that
complex steps such as transcription of a gene are slow and thus
have a low reaction constant k, while other processes such as the
binding of a stimulation factor to a receptor or a transcription
factor to a gene are generally fast and have a high reaction
constant k. For each edge, an initial approximation of the
reaction constant k needs to be made, if possible based on
literature, which can be optimized using results from wet-lab
experiments [8].
10. It is important to note that measuring a reporter is not identical
to measuring the protein directly. In a reporter assay, the
measured signal is a result of a transcription factor (in this
case TCF/LEF) binding to a promotor region in a gene,
resulting in transcription of a protein bound to a reporter
such as luciferase. The luciferase then produces a fluorescent
signal after addition of a substrate, of which the intensity can
then be measured. Since this process includes many additional
steps that sometimes have a limited life-time, the measured
response can differ from the result that would be obtained
when directly measuring the transcription factor TCF/LEF.
11. C2C12 mouse myoblasts is an immortalized mouse myoblast
cell line that undergoes rapid proliferation and can be differ-
entiated into myoblasts and osteoblasts. They are a popular cell
line for studying biochemical pathways. They can be easily

transfected and can be stimulated with a variety of cytokines
and growth factors.
12. 1% Serum is used to starve the cells of signals and reduce
background signal of pathway activity, so that the effect of
the addition of stimuli is clearly visible.
13. Each well plate should contain the same conditions, such that
they can be compared at different time points. Each plate
should also contain 5 replicates of each condition to be able
to calculate the average signal and thereby reduce errors. For
the stimuli used in this example concentrations of 10 ng/mL
TGF-β3, 1 μM BIO, and 10 ng/mL IL-1β were added to the
cells.
14. TCF/LEF-luciferase luminescence and Renilla-luciferase lumi-
nescence data was obtained for each well at four different time
points. Per well, the TCF/LEF-luciferase data is normalized by
dividing the TCF/LEF-luciferase luminescence data by the
Renilla-luciferase luminescence data. Next, the mean lumines-
cence and standard deviation is calculated for each condition at
each time point. A subsequent normalization is done by divid-
ing the mean value for each condition by the mean value of the
control, in which no stimuli were added to the cells. Finally, the
data is plotted for each condition at each time point, as can be
seen in Fig. 4.
15. This is mainly a problem when peak behavior is observed,
where a certain protein or gene is initially upregulated after
addition of a stimulus, but is later inhibited by another protein
that is formed further down the pathway.
16. Before the function “optimize k-value(s)” in ANIMO can be
used, it is essential to already have obtained a graph shape that
is similar to the graph shape found in the wet-lab experiments.
The function “optimize k-value(s)” is only able to change the
k-values, meaning that the graph can be stretched horizontally
and vertically. However, the shape of the graph will never
change, e.g., if no peak is visible in the model before optimiz-
ing the k-values, a peak will never appear after optimization and
nodes and/or connections need to be changed, added or
removed to the model before optimization. In this example,
peak behavior was added to the model by adding an additional
inhibitory node to each node that was already present. The
node that represents a specific protein or gene activates this
inhibitory node with a specific k-value, while the inhibitory
node inhibits the same protein or gene with a slightly higher
k-value. In this way, an inhibitory loop is included in the model
and peak behavior can be included to obtain a graph shape that
is similar to the one found in the wet-lab experiments.
References
1. Liang Y, Kelemen A (2017) Dynamic model- for the analysis of cell signaling networks. Bio-
ing and network approaches for omics time chemistry 49(15):3216–3224. https://doi.
course data: overview of computational org/10.1021/bi902202q
approaches and applications. Brief Bioinform 11. Schivo S, Scholma J, van der Vet PE,
19(5):1051–1068. https://doi.org/10.1093/ Karperien M, Post JN, van de Pol J, Langerak
bib/bbx036 R (2016) Modelling with ANIMO: between
2. Du Q, Zhang X, Liu Q, Zhang X, Bartels CE, fuzzy logic and differential equations. BMC
Geller DA (2013) Nitric oxide production Syst Biol 10(1):56. https://doi.org/10.
upregulates Wnt/β-catenin signaling by inhi- 1186/s12918-016-0286-z
biting Dickkopf-1. Cancer Res 73 12. Terfve C, Cokelaer T, Henriques D,
(21):6526–6537. https://doi.org/10.1158/ MacNamara A, Goncalves E, Morris MK,
0008-5472.CAN-13-1620 Mv I, Lauffenburger DA, Saez-Rodriguez J
3. Zhong L, Schivo S, Huang X, Leijten J, (2012) CellNOptR: a flexible toolkit to train
Karperien M, Post JN (2017) Nitric oxide protein signaling networks to data using multi-
mediates crosstalk between interleukin 1β and ple logic formalisms. BMC Syst Biol 6(1):133.
WNT signaling in primary human chondro- https://doi.org/10.1186/1752-0509-6-133
cytes by reducing DKK1 and FRZB expression. 13. Chaouiya C, Remy E, Mossé B, Thieffry D
Int J Mol Sci 18(11):2491. https://doi.org/ (2004) Qualitative analysis of regulatory
10.3390/ijms18112491 graphs: a computational tool based on a dis-
4. Schivo S, Khurana S, Govindaraj K, Scholma J, crete formal framework. In: Benvenuti L, De
Kerkhofs J, Zhong L, Huang X, van de Pol J, Santis A, Farina L (eds) Positive systems. Lec-
Langerak R, van Wijnen AJ, Geris L, ture Notes in Control and Information Sci-
Karperien M, Post JN (2019) ECHO, the exe- ence, vol 294. Springer, Berlin. https://doi.
cutable CHOndrocyte: a computational model org/10.1007/978-3-540-44928-7_17
to study articular chondrocytes in health and 14. Bock M, Scharp T, Talnikar C, Klipp E (2013)
disease. Cell Signal 68:109471. https://doi. BooleSim: an interactive Boolean network sim-
org/10.1016/j.cellsig.2019.109471 ulator. Bioinformatics 30(1):131–132.
5. Luechtefeld T, Marsh D, Rowlands C, Har- https://doi.org/10.1093/bioinformatics/
tung T (2018) Machine learning of toxicologi- btt568
cal big data enables read-across structure 15. Schivo S, Scholma J, Karperien M, Post JN,
activity relationships (RASAR) outperforming Van De Pol J, Langerak R (2014) Setting para-
animal test reproducibility. Toxicol Sci 165 meters for biological modelsWith ANIMO. In:
(1):198–212. https://doi.org/10.1093/ Electronic Proceedings in Theoretical Com-
toxsci/kfy152 puter Science, EPTCS, pp 35–47. doi:https://
6. Grimm D (2019) U.S. EPA to eliminate all doi.org/10.4204/EPTCS.145.5
mammal testing by 2035. https://www. 16. Krumsiek J, Pölsterl S, Wittmann DM, Theis
sciencemag.org/news/2019/09/us-epa-elimi FJ (2010) Odefy - from discrete to continuous
nate-all-mammal-testing-2035 models. BMC Bioinformatics 11(1):233.
7. Grimm D (2019) EPA plan to end animal test- https://doi.org/10.1186/1471-2105-11-
ing splits scientists. Science 365(6459):1231. 233
https://doi.org/10.1126/science.365.6459. 17. Mendes P, Hoops S, Sahle S, Gauges R, Dada J,
1231 Kummer U (2009) Computational modeling
8. Scholma J, Schivo S, Urquidi RA, Pol JVD, of biochemical networks using COPASI. In:
Karperien M, Post JN (2014) Biological net- Maly IV (ed) Methods in molecular biology
works 101 : Computational modeling for (methods and protocols)-systems biology, vol
molecular biologists. Gene 533(1):379–384. 500. Vol systems biology. Humana Press,
https://doi.org/10.1016/j.gene.2013.10. New York, NY. https://doi.org/10.1007/
010 978-1-59745-525-1_2
9. Brodland GW (2015) How computational 18. Matsuoka Y, Funahashi A, Ghosh S, Kitano H
models can help unlock biological systems. In: (2014) Modeling and simulation using cellde-
Seminars in cell & developmental biology. signer. In: Miyamoto-Sato E, Ohashi H,
Elsevier, pp 62–73. https://doi.org/10. Sasaki H, Nishikawa J-I, Yanagawa H (eds)
1016/j.semcdb.2015.07.001 Methods in molecular biology (methods and
10. Morris MK, Saez-Rodriguez J, Sorger PK, protocols)- Transcription factor regulatory
Lauffenburger DA (2010) Logic-based models networks., vol 1164. Humana Press,
New York, NY. https://doi.org/10.1007/ 30. Frederick JP, Liberati NT, Waddell DS, Shi Y,
978-1-4939-0805-9_11 Wang X-F (2004) Transforming growth factor
19. Schivo S, Scholma J, Wanders B, Camacho β-mediated transcriptional repression of c-myc
RAU, van der Vet PE, Karperien M, is dependent on direct binding of Smad3 to a
Langerak R, van de Pol J, Post JN (2014) novel repressive Smad binding element. Mol
Modeling biological pathway dynamics with Cell Biol 24(6):2546. https://doi.org/10.
timed automata. Ieee J Biomed Health 18 1128/MCB.24.6.2546-2559.2004
(3):832–839. https://doi.org/10.1109/Jbhi. 31. Higuera GA, Hendriks JA, van Dalum J, Wu L,
2013.2292880 Schotel R, Moreira-Teixeira L, van den
20. Eisenmann DM (2005) Wnt signaling. Doel M, Leijten JC, Riesle J, Karperien M,
https://doi.org/10.1895/wormbook.1.7.1. van Blitterswijk CA, Moroni L (2013) In vivo
Accessed 19 Dec 2019 screening of extracellular matrix components
21. Murphy AM, Wong AL, Bezuhly M (2015) produced under multiple experimental condi-
Modulation of angiotensin II signaling in the tions implanted in one animal. Integr Biol 5
prevention of fibrosis. Fibrogenesis Tissue (6):889–898. https://doi.org/10.1039/
Repair 8:7–7. https://doi.org/10.1186/ c3ib40023a
s13069-015-0023-z 32. Jonk LJC, Itoh S, Heldin C-H, ten Dijke P,
22. Morlon A, Munnich A, Smahi A (2005) TAB2, Kruijer W (1998) Identification and functional
TRAF6 and TAK1 are involved in NF-κB acti- characterization of a Smad binding element
vation induced by the TNF-receptor, Edar and (SBE) in the JunB promoter that acts as a
its adaptator Edaradd. Hum Mol Genet 14 transforming growth factor-β, activin, and
(23):3751–3757. https://doi.org/10.1093/ bone morphogenetic protein-inducible
hmg/ddi405 enhancer. J Biol Chem 273
(33):21145–21152. https://doi.org/10.
23. MacDonald BT, Tamai K, He X (2009) Wnt/ 1074/jbc.273.33.21145
beta-catenin signaling: components, mechan-
isms, and diseases. Dev Cell 17(1):9–26. 33. Ranganathan P, Agrawal A, Bhushan R, Cha-
https://doi.org/10.1016/j.devcel.2009.06. valmane AK, Kalathur RKR, Takahashi T, Kon-
016 daiah P (2007) Expression profiling of genes
regulated by TGF-beta: differential regulation
24. Tran FH, Zheng JJ (2017) Modulating the wnt in normal and tumour cells. BMC Genomics
signaling pathway with small molecules. Pro- 8:98–98. https://doi.org/10.1186/1471-
tein Sci 26(4):650–661. https://doi.org/10. 2164-8-98
1002/pro.3122
34. Hyun Hwa C, Hye Joon J, Ji Sun S, Yong
25. Verstrepen L, Bekaert T, Chau TL, Tavernier J, Chan B, Jin Sup J (2008) Crossregulation of
Chariot A, Beyaert R (2008) TLR-4, IL-1R β-catenin/Tcf pathway by NF-κB is mediated
and TNF-R signaling to NF-κB: variations on by lzts2 in human adipose tissue-derived mes-
a common theme. Cell Mol Life Sci 65 enchymal stem cells. Biochim Biophys Acta
(19):2964–2978. https://doi.org/10.1007/ 1783(3):419–428. https://doi.org/10.1016/
s00018-008-8064-8 j.bbamcr.2007.08.005
26. Massagué J (1998) TGF-β signal transduction. 35. Du Q, Geller DA (2010) Cross-regulation
Annu Rev Biochem 67(1):753–791. https:// between Wnt and NF-κB signaling pathways.
doi.org/10.1146/annurev.biochem.67.1.753 For Immunopathol Dis Therap 1(3):155–181.
27. Cadigan KM, Waterman ML (2012) https://doi.org/10.1615/For
TCF/LEFs and Wnt signaling in the nucleus. umImmunDisTher.v1.i3
Cold Spring Harb Perspect Biol 4(11): 36. Wang J, Zhao J, Chu ESH, Mok MTS, Go
a007906. https://doi.org/10.1101/ MYY, Man K, Heuchel R, Lan HY, Chang Z,
cshperspect.a007906 Sung JJY, Yu J (2013) Inhibitory role of Smad7
28. Wu B, Crampton SP, Hughes CCW (2007) in hepatocarcinogenesis in mice and in vitro. J
Wnt signaling induces matrix metalloprotei- Pathol 230(4):441–452. https://doi.org/10.
nase expression and regulates T cell transmigra- 1002/path.4206
tion. Immunity 26(2):227–239. https://doi. 37. Guo X, Wang X-F (2009) Signaling cross-talk
org/10.1016/j.immuni.2006.12.007 between TGF-beta/BMP and other pathways.
29. Heather G, Chun P (2006) Activin receptor- Cell Res 19(1):71–88. https://doi.org/10.
like kinases: structure, function and clinical 1038/cr.2008.302
implications. Endocr Metab Immune Disord 38. Huang N, Li W, Wang X, Qi S (2018) Micro-
Drug Targets 6(1):45–58. https://doi.org/ RNA-17-5p aggravates lipopolysaccharide-
10.2174/187153006776056585 induced injury in nasal epithelial cells by
targeting Smad7. BMC Cell Biol 19(1):1–1. https://doi.org/10.1016/S0092-8674(00)

https://doi.org/10.1186/s12860-018-0152- 80303-7
5 40. Huang X, Zhong L, Hendriks J, Post JN, Kar-
39. Hayashi H, Abdollah S, Qiu Y, Cai J, Xu Y-Y, perien M (2018) The effects of the
Grinnell BW, Richardson MA, Topper JN, WNT-signaling modulators BIO and
Gimbrone MA, Wrana JL, Falb D (1997) The PKF118-310 on the chondrogenic differentia-
MAD-related protein Smad7 associates with tion of human mesenchymal stem cells. Int J
the TGFβ receptor and functions as an antago- Mol Sci 19(2):561. https://doi.org/10.3390/
nist of TGFβ signaling. Cell 89(7):1165–1173. ijms19020561
Part II
In Vivo Models of Skeletal Tissue Injury, Degeneration,

and Repair
Chapter 11
Generation and Characterization of Mouse Models

for Skeletal Disease
Gabrielle E. Foxa, Ye Liu, Lisa M. Turner, Alexander G. Robling, Tao Yang,
and Bart O. Williams
Abstract
Our laboratories have used genetically engineered mouse models (GEMMs) to assess genetic contributions
to skeletal diseases such as osteoporosis and osteoarthritis. Studies on the genetic contributions to OA are
often done by assessing how GEMMs respond to surgical methods that induce symptoms modeling
OA. Here, we will describe protocols outlining the induction of experimental OA in mice as well as detailed
descriptions of methods for analyzing skeletal phenotypes using micro-computerized tomography and
skeletal histomorphometry.
Key words microCT, Histomorphometry, Osteoarthritis
1 Introduction
The analysis of skeletal phenotypes of genetically modified mice has

contributed to advancements of our knowledge about skeletal
development and disease and has provided the foundation for the
development of several FDA-approved treatments for bone-related
diseases. Our laboratories have used genetically engineered mouse
models (GEMMs) to assess genetic contributions to skeletal dis-
eases such as osteoporosis and osteoarthritis. Studies on the genetic
contributions to OA are often done by assessing how GEMMs
respond to surgical methods that induce symptoms modeling
OA. The generation of GEMMs is applicable to all tissue and
disease types and discussion of methodologies for their generation
is available in many other contexts [1, 2]. Here, we focus on several
methods specific for generating and analyzing skeletal phenotypes
in mice. We first describe the generation of an OA-like model via
surgical destabilization of the medial meniscus. We will then
describe the evaluation of skeletal phenotypes via micro-
computerized tomography (microCT) and histomorphometry.
165
166 Gabrielle E. Foxa et al.
2 Materials
2.1 Surgical 1. Experimental mice of the same strain divided into DMM sur-
Destabilization gery group (n 5) and sham surgery control group (n 5).
of the Medial Age of mice should be decided according to the goal of experi-
Meniscus (DMM) ments, but generally mice at or older than 3 months of age are
optimal as their skeletal maturity is reached. All mice should be
housed in a gnotobiotic facility, according to the National
Research Council’s Guide and Institutional Animal Care and
Use Committee-approved protocols for the care and use of
experimental animals (see Note 1).
2. Isoflurane machine: VetEquip Tabletop laboratory animal
anesthesia system (VetEquip # 901820).
3. Isoflurane (Piramal # NDC 66794–017-25).
4. Surgical microscope (Leica # M651).
5. Veterinary operating table heated mat (Peco #
Mediheat V500).
6. 9 mm Wound forceps and clips (Reflex # RS-9260, RS-9262).
7. Wound clip remover (Reflex # RS-9263).
8. Hair removal lotion.
9. Alcohol antiseptic wipes.
10. Micro-iris scissors.
11. Micro-surgical knife (#3 handle and #10/ #11 blade).
12. Sharppoint 15 5 mm blade micro-surgical knife.
13. Epinephrine (1:1000 dilution).
14. Amoxicillin (20 mg/kg).
2.2 Osteoarthritis 1. Histological tissue cassette case.

(OA) Histology 2. Surgical scissors.
3. Tweezer.
4. Deionized water.
5. Ethanol 70%, 80%, 90%, 95%, 100%.
6. Neutral buffered formalin (NBF) 10%.
7. Formic acid 5%.
8. Xylene.
9. Paraffin (Paraplast X-tra).
10. Oven.
11. Stock solution A: 10 g hematoxylin, 500 mL 80% ethanol.
12. Stock solution B: 20 g ferric chloride, 475 mL distilled water,
5 mL hydrochloric acid (36.5–38%).
Generation and Characterization of Mouse Models for Skeletal Disease 167
13. Modified Weigert’s Iron Hematoxylin: equal volume stock

solutions A and B (see Note 2).
14. Acid-alcohol 1.0%: 5 mL hydrochloric acid (36.5–38%),
495.0 mL 70% ethanol.
15. Acetic acid 1.0%: 5.0 mL glacial acetic acid, 495 mL distilled
water.
16. Fast Green 0.08%: 0.20 g Fast Green, 250.0 mL distilled water.
17. Safranin O 1%: 2.5 g Safranin O, 250 mL distilled water.
2.3 Microcomputed All standards, equipment, and software for scanning and analysis are
Tomography produced by Bruker microCT, Kontich, Belgium and Micro Pho-
tonics Inc., Allentown, PA.
1. Neutral buffered formalin (NBF) 10%.
2. Deionized (DI) water.
3. Ethanol (EtOH) 70%.
4. Gauze cut to length of the sample.
5. Plastic conical.
6. Bruker Skyscan 1172.
7. Skyscan 1172 (Scanning software) Version 1.5.26.0.
8. Software NRecon Version 1.7.4.6.
9. GPUReconServer Version 1.7.4.2.
10. Data Viewer Version 1.5.6.3 software.
2.4 Bone Mineral 1. Bone mineral density standards 0.25 and 0.75.
Density Calibration 2. CT Analyzer Version 1.18.8.0 software.
2.5 Trabecular 1. CT Analyzer Version 1.18.8.0 software.

Analysis
of Microcomputed
Tomography Images
2.6 Cortical Analysis 1. CT Analyzer Version 1.18.8.0 software.

of Microcomputed
Tomography Images
2.7 Bone Modeling 1. CTvol Version 2.3.2.0 software.
2.8 Fluorochrome 1. Calcein Green 1.0% stock solution: In a 50 mL conical tube,

Labeling of Bone dissolve 400 mg NaHCO3 in 20 mL of saline (0.9% NaCl).
Slowly add 200 mg calcein (Sigma-Aldrich # C0875). Calcein
will foam if added quickly. pH to 7.4 using 1 N NaOH or 1 N
HCl and filter sterilize into evacuated 20 mL glass vial. Wrap in
foil and store at 4 C (see Note 3).
2. Calcein Green 0.10% working solution: Dilute the Calcein

Green 1.0% stock solution in saline. Pipette 2.0 mL of 1%
calcein stock and 18.0 mL of saline. Wrap in foil and store at
4 C (see Note 4).
3. Alizarin Complexone 2.5% stock solution: In a 50 mL conical
tube, dissolve 400 mg NaHCO3 in 20 mL of saline (0.9%
NaCl). Slowly add 500 mg alizarin complexone (Sigma-Aldrich
# A3882). Alizarin will foam if added quickly. pH to 7.4 using
1 N NaOH or 1 N HCl and filter sterilize into evacuated 20 mL
glass vile. Wrap in foil and store at 4 C.
4. Alizarin Complexone 0.20% working solution: Dilute 2.5%
stock solution with saline to 0.2%. Pipette 0.8 mL of 2.5%
alizarin stock and 9.2 mL of saline.
5. Demeclocycline HCL 0.75% working solution: In a sterile
200 mL bottle, dissolve 1 g demeclocycline hydrochloride
(Sigma-Aldrich # D6140) in 133 mL of sterile saline (0.75%
solution). Place on a rocker plate and rock gently for 20 min.
Aliquot into light-proof 1.5 mL tubes and store at 20 C.
2.9 Fixation 1. DI water.

2. Ethanol 70%.
3. Neutral Buffered Formalin (NBF) 10%.
2.10 Infiltration 1. Dremel hand tool.

2. Automated tissue processor.
3. Ethanol 70%, 95%, 100%.
4. Fisherbrand Histocassette Tissue Cassette.
5. Infiltration solution: 85% Methyl methacrylate (MMA) and
15% dibutyl phthalate (DBP).
6. Methyl methacrylate (MMA) 100%.
7. NBF 10%.
8. Rapid Infiltration Processor (RIP).
9. Xylene.
10. Embedding molds.
11. Glass culture tubes.
2.11 Embedding 1. 600-grit Sandpaper.

and Cross-Sectioning 2. Allen wrench.
3. Binder clips.
4. DI water.
5. Embedding solution: 85% MMA, 15% DBP, and 0.5%
Perdadox 16.
6. Eukitt Mounting Medium Electron Microscopy Sciences #

15320.
7. Extech Labcut 150.
8. 3”.006” ½ Diamond Wafering Blade.
9. Glass culture tubes with rubber stoppers.
10. Glass slides.
11. Lab Armor Bead Bath.
12. Micrometer.
13. Paper towel.
14. Sonicator.
15. Small weight.
2.12 MMA 1. Automatic rotary microtome.

Embedding 2. Bead bath.
and Coronal
3. Clamps.
Sectioning
4. Coating solution: 85 mL DI water, 1 g gelatin powder, and
15 mL glycerol incubate the solution at 42 C in a water bath
until dissolved or visible particles are no longer visible (see
Note 5).
5. Cutting solution: 10 mL 70% ethanol, 90 mL distilled water,
and 2 drops of soap (see Note 6).
6. Embedding solution: 85% MMA, 15% DBP, and 0.5%
Perdadox 16.
7. Ethanol 70%, 75%.
8. Forceps.
9. Fine tip paint brush.
10. Glass slides.
11. Kimwipes.
12. MetaServ 250 Single Grinder/Polisher.
13. Paperclips.
14. Paper towels.
15. Petri dish.
16. Plastic film.
17. Rubber roller.
18. Silicone stopper.
19. Oven.
20. Prefabricated mold for embedding.
21. Acetone.
22. Cellusolve.
23. DI water.
2.13 Golder’s 1. Wiegert’s Iron Hematoxylin: Prepare fresh by combining

Trichrome Stain 50 mL Wiegert’s Solution A and 50 mL Wiegert’s Solution B.
2. Ponceau Fuchsin solution A: 1 g Ponceau, 100 mL DI water.
3. Ponceau Fuchsin solution B: 1 g Acid fuchsin, 100 mL DI
water.
4. Ponceau Fuchsin stock solution: 30 mL Ponceau fuchsin solu-
tion A, 10 mL ponceau fuchsin solution B.
5. Acetic acid 1%.
6. Ponceau Fuchsin working solution: Dilute 10 mL ponceau
fuchsin stock solution in 150 mL 0.2% acetic acid.
7. Phosphotungstic Acid (PTA) Orange G: Combine and filter
2 g Orange G, 4 g phosphotungstic acid, and 100 mL DI
water.
8. Light Green: combine and filter 0.2 g Light Green, 0.2 mL
acetic acid, and 100 mL DI water.
9. DI water.
10. Ventana Symphony tissue processor.
2.14 Paraffin 1. DI water.

Embedding 2. Embedding center (Leica EG1150H).
and Sectioning
3. Ethylenediaminetetraacetic acid (EDTA) 10%: pH 7.4.
4. Ethanol 70%, 80%, 95%, 100%.
5. Fisherbrand Histocassette Tissue Cassette.
6. Forceps.
7. Glass slides.
8. Kimwipes.
9. NBF 10%.
10. Oven.
11. Paraffin/tissue infiltration media.
12. Steel microtome knife.
13. Xylene.
14. Tissue processor.
15. Water bath.
2.15 TRAP Stain 1. Sigma-Aldrich product 387-A: Acid Phosphatase, Leukocyte

(TRAP) kit.
2. Aqueous mounting media.
3. Glass slides.
4. Glass coverslips.
2.16 Slide Imaging 1. ET-DAPI/FITC/Texas Red filter set (Chroma product #

69002).
2. Leica Aperio AT2 High Volume, Digital Whole Slide Scanning.
3. Lumen Dynamics X-Cite Series 120 Q.
4. QImaging QIClick Digital CCD Camera.
5. Zeiss Axio Imager.A2.
6. BIOQUANT OSTEO 19.2.6. software.
2.17 Histo- 1. BIOQUANT OSTEO 19.2.6 software.

morphometry
3 Methods
3.1 DMM Surgery The methods described below that we utilize in the laboratory have
been developed and optimized based on several previous studies
[3–6]. Operate the anesthesia machine according to the VetEquip
user guide and operating manual (http://www.vetequip.com/
pdfs/LAAS%20
Manual.pdf). DMM surgery should be performed on both
knees of each individual mice of the DMM group, and sham
surgery will be performed to another group of mice for controls.
Surgery should be carried out with standard sterile techniques, i.e.,
preoperative instrument sterilization and disinfection, and cleans-
ing of incision area.
1. Anesthetize the mice with 4% isoflurane and 1 L/min oxygen
in the VetEquip tabletop laboratory animal anesthesia chamber.
2. Once the mice are anesthetized, transfer them to a mask and
maintain with 3% isoflurane and 1 L/min oxygen.
3. Place the mice on a veterinary operating table heated mat to
maintain body temperature during surgery.
4. To bare the incision area, apply hair removal lotion on the
medial side of knee for 3 min. Then wipe with alcohol antisep-
tic wipes.
5. In order to expose the medial side of stifle joint capsule, use a
micro-iris scissor or a micro-surgical knife to open a 3 mm
longitudinal incision through the skin and subcutaneous tissue
over the distal patella to proximal tibial plateau.
6. Expose medial meniscotibial ligament (MMTL), which
anchors medial meniscus (MM). During the process, avoid
destabilizing patellar tendon, medial collateral ligament, and
articular cartilage.
7. To prevent bleeding that may block the vision for performing

transection, one drop of 0.1% epinephrine can be applied at the
incision site (optional).
8. Under a surgical microscope can assist accurate transection,
transect the MMTL with a micro-surgical knife (sharppoint
15 5 mm blade) to release the MM. Locating the patellar
ligament first, this can assist the identification of the incision
region.
9. The sham group mice will be operated through the same
approach without MMTL transection.
10. Close the incision together with a 9 mm wound forceps and
clips.
11. To prevent joint infection, Amoxicillin (20 mg/kg) can be
injected after surgery (optional).
12. Remove the clips with a wound clip remover after wound has
healed.
3.2 Osteoarthritis At 4 to 6 weeks (or at specific time point according to the experi-
(OA) Histological mental design) post DMM surgery, OA phenotypes can be graded
Grading or assessed by histological analysis, gene expression assay, microCT,
pain behaviors, or gait analysis. Here we focus on the histological
grading.
1. Collect knees from mice and arrange in a histological tissue
cassette case (see Note 7).
2. To fix the tissue, place the tissue cassettes with mouse knees in
10% NBF for 48 h at 4 C. Rinse in DI water 3 for 10 min
each time and store in 70% ethanol at 4 C.
3. To decalcify the samples, place the fixed tissues with cassettes in
a beaker containing at least 15 volumes of 5% formic acid on a
swirling/rocking platform for 7 days (see Note 8).
4. Rinse the tissues in DI water 4 for 10 min each time.
5. Dehydrate the tissues in 70%, 95%, and 100% ethanol for
multiple rounds 1–2 h each (see Note 9).
6. Clear tissues in xylene for 1 h 2.
7. Infiltrate tissues with paraffin (Paraplast X-tra) at 58 C for 1 h 2.
8. Embed tissues into wax blocks at 62 C.
9. Rehydrate tissues for histology using xylene, absolute alcohol,
95% ethanol, 80% ethanol, then rinse in distilled water (see
Note 10).
10. Cut 5 mm sections of the tissues embedded in the blocks (see
Note 11).
11. Bake the slides at 60 C overnight and cool at room tempera-

ture for 20–30 min.
12. Deparaffinize and hydrate in 70% ethanol and air dry.
13. Incubate the slides with Weigert’s Iron Hematoxylin for
30 min.
14. Wash the slides gently in running tap water for 10 min.
15. Differentiate by single dip in 1% acid-alcohol.
16. Rinse the slides in DI water 3.
17. Incubate the slides with 0.08% Fast Green for 5 min.
18. Rinse the slides in 1.0% acetic acid for 10 s (blot, no rinse).
19. Incubate the slides in 1% Safranin O for 30 min (blot, no rinse).
20. Dehydrate the sections with 1 change of 95% ethanol and two
changes of 100% ethanol by 1 dip each and air dry.
21. Clear the slides in xylene for 1 min, 2 and coverslip.
22. Score histologically to evaluate OA progression in a blinded
manner following the semiquantitative scoring system estab-
lished by Sonya S. Glasson et al. The OA phenotype of all four
quadrants of joint, including the medial femoral condyle,
medial tibial plateau, lateral femoral condyle, and lateral tibial
plateau should be scored (see Note 12).
3.3 Imaging by These methods have been adapted and optimized for use in our
Microcomputed laboratories from several sources [7–9]. Methodology is specific to
Tomography femurs from mice at 3–6 months of age. All microCT methods were
performed under the recommended guidelines of the SkyScan
software Standard Operating Procedures by Bruker MicroCT
(Kontich, Belgium).
1. Fix limbs in 10% NBF for 48 h. Rinse limbs in deionized
(DI) water and store in 70% ethanol.
2. Prepare the limbs by wrapping them in gauze saturated in 70%
ethanol. Load the limbs into a conical submerged in 70%
ethanol while preventing air bubbles from entering. Samples
should not be overlapping and should be snug in the container.
3. Turn the Skyscan 1172 on by turning key to “START” posi-
tion. The key will automatically snap back into the “ON”
position.
4. Open Skyscan 1172 software and start the X-ray tube by push-
ing the X-ray icon. A pop-up window will open with a progress
bar as the tube ages. Once the X-ray is on, an indicator window
will open.
5. Press the “grab image” icon to obtain a continuous preview
from the X-ray camera and establish scan settings (see
Note 13).
6. Perform a flat field correction (see Note 14).

7. Open the chamber on the scanner by pressing the “specimen
door” icon, fasten the sample to a stage, and secure the stage on
the rotation axis by turning the brass ring to tighten it.
8. Close the chamber by pressing the “specimen door” icon, turn
the X-ray back on, and restore the settings established in step 5.
9. Set an oversize scan (see Note 15).
10. Select the “Start scanning” button in the Scout Scan window
to open the settings dialog window and complete scan settings
(see Note 16).
11. Click “OK” to start the scan.
12. After scanning is complete, open NRecon and load a dataset by
choosing “Open Dataset” in the “Actions” tab and choosing
one of the images from a scan. Choosing one image will open
the entire dataset associated with the sample.
13. Perform a thermal correction (see Note 17).
14. Choose a dynamic range based on the histogram of the sample
(see Note 18).
15. Correct beam-hardening (see Note 19).
16. Correct misaligned segments (see Note 20).
17. Reduce ring artifacts (see Note 21).
18. Smooth the images (see Note 22).
19. Choose a region of interest (ROI) to save (see Note 23).
20. Choose a volume of interest (VOI) to save (see Note 24).
21. Reconstruct and save the images (see Note 25).
22. After the images have been reconstructed, open Data Viewer
and load a reconstructed dataset from a “Rec” folder of a
sample.
23. After the sample is open, click the “load for 3D viewing” icon.
24. Use the three views to straighten the sample. Orient all samples
in a project the same way.
25. Under the “View” tab, choose “Single Volume of Interest
(VOI),” and fit the sample within each of the boxes to crop
the dataset and define a volume of interest (VOI).
26. Save the volume of interest (see Note 26).
3.4 Bone Mineral 1. Scan, reconstruct, and orient the bone mineral density stan-
Density Calibration dards following the steps under “Imaging by Microcomputed
Tomography 3.3.”
2. Open the CTan software and load one of the density standards
by choosing the “open image or dataset” icon.
3. Define the upper bound of the density standard by scrolling to

the top and choosing a slice that includes the entire density
standard. Right click the slice in the dataset window and choose
“Set the Top of Selection.” Repeat for the lower bound but
choose “Set the Bottom of the Selection.”
4. Define the region of interest (ROI) to include only the density
standard (see Note 27).
5. Save the ROI (see Note 28).
6. Open the new ROI file for the density standard by selecting the
“open image or dataset icon” and choosing one of the images
in the ROI folder.
7. Apply the ROI to the open file (see Note 29).
8. Identify and record the attenuation coefficient (see Note 30).
9. Repeat steps 2–8 for the other density standard.
10. Under “File,” choose “Preferences,” and under the “Histo-
gram” tab, choose “Calibrate.”
11. Enter the known min and max BMD density values of the
density standards (Ex: 0.25 and 0.75) followed by the
corresponding attenuation coefficients recorded in step 8.
12. Click “OK” to calibrate the density range.
3.5 Trabecular 1. In the CTan software, select the “open image or dataset” icon.
Analysis Choose a dataset within a newly created VOI folder.
of Microcomputed 2. Identify the reference point for the ROI, use a 0.25 mm offset,
Tomography Images and create an ROI 2.5 mm in height (see Note 31).
3. Create a region of interest to include trabecular bone only, by
manually drawing an ROI (see Note 32) or use an automated
program and save the regions. If manually drawing ROIs,
continue to step 6 after this step. To make an automated
program to create an ROI, continue to step 4.
4. Begin the automated trabecular ROI process by, saving the
ROI defined in step 2 and determining the thresholding scale
(see Note 33).
5. Open the ROI saved in step 4 by choosing the “open an
existing image or dataset” icon and choosing an image within
the ROI folder for the sample. Build a task list to create,
analyze, and save an automated trabecular ROI (see Note 34).
After building the task list, skip to step 9.
6. Open a sample’s manual ROI by choosing the “open an exist-
ing image or dataset” icon and choosing an image within the
trabecular ROI folder for the sample.
7. Apply the manual ROI to the open file and determine the
thresholding scale (see Note 35).
8. Choose the “custom processing preview” icon and select the

“Internal” tab to build a task list to analyze the trabecular bone
(see Note 36).
9. Run the task list on multiple samples using the batch manager
(see Note 37).
3.6 Cortical Analysis 1. In the CTan software, select the “open image or dataset” icon.
of Microcomputed 2. Identify the reference point for the ROI and set the height to
Tomography Images 0.6 mm (see Note 38).
3. Create a region of interest to include the entire area of the bone
(see Note 39).
4. Save the cortical ROI (see Note 40).
5. Open a sample ROI by choosing the “open an existing image
or dataset” icon and choosing an image within the cortical ROI
folder for the sample.
6. Apply the ROI to the open file and determine the thresholding
scale (see Note 41).
7. Build a 2D analysis task list using a combination of functions
until the cortical bone that remains in the “binary selection
preview” closely matches the raw images (see Note 42).
8. Build a histogram Task List by constructing a task list the same
as the “2D analysis Task List” except for the 2D analysis and
3D model functions. Add ROI shrink-wrap, reload, histogram
functions to the end of the list (see Note 43).
9. Run the task list on multiple samples using the batch manager.
Repeat for each task list (see Note 44).
3.7 Modeling Several modifications can be made when generating visual repre-
Skeletal Phenotypes sentations of scanned samples. The guidelines below are for a basic,
gray-scale model.
1. Using CTvol, open a 3D file by selecting the “open 3D object”
icon and choosing the p3g file within the ROI folder of the
desired sample.
2. Use the wheel on the computer mouse to zoom in and out on
the sample.
3. Select “Options” and choose “Objects” to open up the
“Objects” window. Adjust color, emission, diffusion, reflec-
tion, and opacity as desired within this window.
4. Select the “move camera” or “move objects” icon and use your
mouse to adjust the sample to the desired position and angle.
5. Save the Image by selecting “Save Image” from the “File”
dropdown menu.
3.8 Mouse Labeling schedules will vary depending on anticipated phenotype

Fluorochrome and project design. The following schedule is for mice 6 months of
Injections age with fluorochrome labeled surfaces at four different time
points.
1. At 4 weeks of age, perform intraperitoneal (IP) injections on
mice with 0.10% working calcein solution. Inject 0.2 mL per
20 g of body weight for a 10 mg/kg dose (see Note 45).
2. At 10 weeks of age, perform IP injections on mice with 0.20%
working alizarin complexone solution. Inject 0.2 mL per 20 g
of body weight for a 20 mg/kg dose (see Note 45).
working demeclocycline hydrochloride solution. Inject
0.250 mL per 20 g mouse 150 IP/100 subcutaneous for a
75 mg/kg dose (see Note 45).
working calcein solution. Inject 0.2 mL per 20 g of body
weight for a 10 mg/kg dose.
5. Sacrifice animals at 26 weeks of age.
3.9 Infiltration 1. Fix the femurs in 10% NBF for 48 h, followed by storage in 70%
ethanol.
2. Remove excess tissue from femurs and move to individual
histocassette tissue cassettes.
3. Using the Dremel hand tool, notch each bone ¼ to ½ of the
way through in the midshaft region. Return each bone to its
cassette.
4. Dehydrate and clear the samples for 2–4 h in 70% ethanol, 95%
ethanol, 100% ethanol, and xylene (see Note 46).
5. Transfer the specimens to xylene until infiltration.
6. Within the RIP unit infiltrate the samples under 15 psi using
100% MMA for 24 h (infiltration 1), the infiltration solution
for 24 h (infiltration 2), followed by the infiltration solution for
at least 2 days (infiltration 3).
7. Transfer bone tissue to individual, prefabricated embedding
molds for coronal sectioning or glass culture tubes for cross-
sectioning.
3.10 Femoral When processing tissues that are light sensitive due to fluoro-
Embedding chrome labeling, keep in the dark in between processing steps.
and Cross-Sectioning
1. Add a few drops of glass culture tube embedding solution into
a labeled glass culture tube and insert the femur, ensuring the
head of the femur is closest to the open end of the tube (see
Note 47).
2. Cap the glass tube with a rubber stopper and transfer the
beaker into the bead bath at 25 C for optimal polymerization.
3. Check the glass tubes once daily until the polymerization pro-
cess is complete (see Note 48).
4. Remove the polymerized rod containing the femur from the
glass culture tube by gently shattering the surrounding glass.
5. Load the resulting plastic rod into the specimen holder of the
Extech Labcut 150. Secure the rod into the specimen holder by
tightening down the adjustment bolts with an Allen wrench.
6. Line up the plastic rod with the diamond wafering blade to the
midshaft of the femur and move the specimen arm so it sits at
the back of the Labcut 150.
7. Place either the unattached reed switch magnet or the canopy
in place so that the reed switch is covered.
8. Turn the instrument on, set the speed to 166.2 rpm, select the
“rev” button to start the blade.
9. Turn on the lubricant pump and slowly pull the specimen arm
forward.
10. Sit the rod atop the saw blade and begin to section the bone at
the midshaft or slightly more distal (see Note 49).
11. Save the distal end for MMA embedding and coronal cross-
sectioning or discard it.
12. Adjust the micrometer on the Extech Labcut 150 with the
diamond wafering blade to the desired thickness for the first
wafer. Ensure the rod is advancing over the blade (see Note
50).
13. Section desired number of wafers.
14. Remove any plastic dust particles from the wafers by dipping
each in a sonicator filled with deionized water for 3–5 s. Dab
off any excess water.
15. Place a nickel size pool of Eukitt mounting media onto the
back of a labeled glass slide and mount the wafers in the pool.
16. Cover the glass slide with 2–3 layers of 1–2 ply paper towel cut
into slide size strips.
17. Place a blank glass slide on top of the paper towel strips and
fasten both slides together with 2–3 binder clips and allow the
slide to air dry overnight.
18. Remove the binder clips, blank slide, and paper towel strips
from the surface of the slide with mounted wafers to be sanded.
19. Using a table micrometer, zero out the slide thickness by
placing a portion of the glass slide that does not contain a
wafer between the clamps. Tighten the clamp on the slide
and clear the counter.
20. Open the micrometer clamp and move the slide so one of the
mounted wafers is measurable. Close the clamp down onto the
wafer and obtain a starting measurement of the wafer. Repeat
this step for all wafers.
21. Using 600 grit sandpaper, gently sand the surface of the wafers
until each measures 40–50 μm. Remeasure each wafer using
the micrometer throughout the sanding process in order to
determine when the desired thickness has been obtained.
22. Once all wafers have been sanded to the desired thickness, wipe
the sanding remnants from the surface of the slide with a damp
paper towel.
23. Add a thin layer of Eukitt mounting media to the surface of the
wafers and add a glass coverslip.
24. Add 2–3 small weights to the top of the coverslip to ensure any
air bubbles are removed and allow the slides to air dry
overnight.
3.11 MMA When processing tissues that are light sensitive due to fluoro-
Embedding chrome labeling, keep in the dark in between processing steps.
and Coronal
1. Add a thin layer of prefabricated mold embedding solution to
Sectioning of Femurs each prefabricated mold and allow the reagent to become
sticky/tacky.
2. Orient the tissue for coronal sectioning and fill remainder of
the mold with prefabricated mold embedding solution.
3. Add a paper label containing the specimen ID, and a paper clip
bent to a 45 –90 angle to each mold.
4. Cap the top of the mold with a silicone stopper and transfer
each mold to the bead bath at 25 C, ensuring the containers
are covered.
5. Remove molds from the bead bath daily and check until poly-
merization is complete (see Note 51).
6. Remove fully polymerized molds to grind and polish using the
MetaServ 250 Single Grinder/Polisher (see Note 52).
7. Apply a thin layer of the coating solution to a glass slide.
8. Place the embedded tissue mold in the microtome chuck and
face into the mold until the desired tissue is exposed.
9. Brush the exposed face of the block with a fine tip paint brush,
soaked in the cutting solution, and initiate the automatic
microtome to begin slicing a section from the block at 5 μm.
10. Cut a small rectangle from a Kimwipe, and soak in 70% ethanol.
11. Apply the Kimwipe to the surface of the mold and initiate
cutting on the microtome at a very slow speed.
12. As the Kimwipe and plastic section are slowly cut, grasp them
together with a pair of forceps. Continue to carefully pull the
section off the block as it is cut away by the microtome.
13. Transfer the section and wipe into a petri dish of 75% ethanol.
Carefully stretch the section if wrinkles are visible using two
pairs of forceps.
14. Once flat, place a piece of plastic film, just larger than the
section, on top of the section to keep it free of wrinkles.
15. Transfer the section and plastic film combination to the coated
slide.
16. Cover the slide with a thin sheet of plastic and lightly squeeze
out any excess ethanol using a gentle wiping motion with a
Kimwipe.
17. Cover the section with paper towel and roll flat with rubber
roller.
18. Each day sectioning is completed, stack the slides together with
a blank slide on the top and bottom of the stack and place in a
clamp. Transfer each stack to an oven set at 60 C, overnight.
19. Remove slides from the oven and inspect for damage.
20. Begin deplasticizing slides by soaking them in cellusolve for
1 h.
21. Place slides in acetone for 2–5 min.
22. Rinse slides in DI water for 2–5 min.
23. Continue to “Goldner’s Trichrome Staining 3.12” step 1 for
staining or step 10 to only coverslip for fluorochrome imaging.
3.12 Goldner’s 1. After deplasticizing samples, place slides in Wiegerts Hematox-

Trichrome Staining ylin for 20 min.
2. Rinse slides in running Tap Water for 10 min.
3. Place slides in working solution of Ponceau Acid Fuchsin for
10 min.
4. Place slides in 1% acetic acid for 30 s.
5. Place slides in filtered PTA-Orange G for 4 min.
6. Place slides in 1% acetic acid for 30 s.
7. Place slides in Light Green solution for 10 min.
8. Place slides in 1% acetic acid for 1 min.
9. Rinse slides in 2 changes of distilled water, for 1 min each.
10. Coverslip using the Ventana Symphony by using the proprie-
tary “Coverslip Only” program.
3.13 Paraffin 1. Fix the femurs in 10% NBF for 48 h, followed by storage in 70%
Embedding ethanol.
and Sectioning 2. Remove excess tissue from femurs and move to individual
of Femurs histocassette tissue cassettes.
3. Decalcify samples in 10% EDTA pH 7.4, stirring for
10–14 days at 4 C (see Note 53).
4. Dehydrate, clear, and infiltrate the samples in cassettes within a
tissue processor using 70% ethanol for 60 min, 80% ethanol for
60 min, 95% ethanol for 60 min 2, 100% ethanol for 60 min
3, xylene for 30 min 2, paraffin for 30 min, and paraffin for
45 min.
5. Remove the cassettes from the tissue processor and move them
to the embedding center.
6. Set the heated areas of the embedding center between 60 C
and 70 C.
7. Remove the tissue from the cassette and place it in an appropri-
ately sized heated mold. Orient all the samples the same way
depending on the type of cut desired (sagittal or coronal).
8. Hold the tissue specimen down with forceps while partially
filling the mold with molten paraffin. Secure the tissue by
quickly cooling the base of the mold (see Note 54).
9. Place the cassette bottom atop the mold and fill with paraffin.
10. Cool the blocks in a cooling area to set the paraffin until it has
solidified.
11. Remove the blocks from the mold by pulling upward on the
cassette.
12. Rough cut/face each block to expose the desired surface area
of the tissue(s). Precool paraffin blocks with the tissue side
down in a tray of ice water to facilitate sectioning.
13. Using a steel microtome knife or disposable blade, cut the
appropriate sections (see Note 55).
14. For histological sections, place the sections on the pre-labeled
glass slides with the IDs that correspond to each block being
sectioned. Mount sections by floating a paraffin ribbon into a
water bath set between 36 C and 42 C.
15. Dry paraffin sections in a lab oven set to 60 C for 10–20 min.
16. Remove the sections from the oven and allow the slides to cool
at room temperature.
17. Deparaffinize slides by soaking in fresh xylene for 3 min 3,
hand-dip slides into 100% ethanol 10, repeat in fresh 100%
ethanol, hand-dip slides in 95% ethanol 10, repeat in fresh
95% ethanol, and rinse in DI water.
18. Store slides in DI water until TRAP staining.
3.14 TRAP Staining 1. Use the Sigma-Aldrich product 387-A: Acid Phosphatase, Leu-
kocyte (TRAP) kit and protocol to stain deparaffinized
samples.
2. Coverslip the slides using an aqueous mounting media and
glass coverslips.
3.15 Imaging Slides 1. Image the trichrome stained samples at 20 magnification
using a brightfield microscope (Leica Aperio AT2—High Vol-
ume, Digital Whole Slide Scanning).
2. Image fluorochrome labeled samples at 20 magnification
(Zeiss Axio Imager.A2, Lumen Dynamics X-Cite Series
120 Q, QImaging QIClick Digital CCD Camera) using objec-
tive filter ET-DAPI/FITC/Texas Red (Chroma product #
69002). Use BIOQUANT OSTEO 19.2.6 software and follow
the BIOQUANT OSTEO 19.2.6 “Image Region and Image
Menu Introduction.”
3.16 Histo- The analysis steps are specific to BIOQUANT OSTEO 19.2.6
morphometry Manual.
1. To quantify trabecular bone parameters, open BIOQUANT
OSTEO 19.2.6 software and create a new dataset following
“The Data Region: Introduction.”
2. Measure trabecular parameters (Goldner’s Trichrome stained,
Fluorochrome labeled, TRAP stained slides) by following the
“Rodent Trabecular Bone Protocol.”
3. To quantify cortical bone parameters, start by creating new
dataset following “The Data Region: Introduction.”
4. Measure cortical bone parameters by following the “Basic Cor-
tical Bone Protocol.”
4 Notes
1. Male mice are more susceptible to developing DMM-induced

osteoarthritis [10] thus considered suitable materials for such
experiments, but the selection of mouse genders should
accommodate the goals of specific studies.
2. Mix equal volume of stock solutions A and B to make the
Modified Weigert’s Iron Hematoxylin working solution right
before use. The working solution has a shelf life of about
1 week when stored in dark.
3. Wrap in foil to prevent exposure as the solution is light
sensitive.
4. Stock solution is diluted to the 10 mg/kg recommended dose
for injecting mice.
5. Once prepared, the coating solution is stable at room tempera-

ture for 1 week.
6. Once prepared, the solution is stable at room temperature until
expiration date of the ethanol lot.
7. Sacrifice the mice in a CO2 chamber. Dissect both hind limbs
from the hip of the mice using surgical scissors and tweezers.
Remove about 1/3 of femur from the proximal end and 1/3 of
the tibia from the distal end using surgical scissors. Straighten
each knee and place in a labeled histological tissue cassette with
the patella facing up. This will help the following embedding
and sectioning in the frontal orientation.
8. Another common method for decalcification is incubation in
14% EDTA solution for 10 days. EDTA-decalcified tissues
generally have better result for immunohistochemistry than
formic-acid-decalcified ones. However, the Safranin O stain
shows better color intensity on the formic-acid-decalcified
sections.
9. Dehydrate the tissues in the following order: 70% ethanol for
1 h, 95% ethanol for 1 h, absolute ethanol for 1 h 2, absolute
ethanol for 1.5 h, and absolute ethanol for 2 h.
10. Deparaffinize sections, 2 changes of xylene, 3 min each, then
rehydrate in 2 changes of absolute alcohol, 2 min each. Incu-
bate in 95% ethanol for 2 min followed by 80% ethanol for
2 min. Wash sections briefly in distilled water.
11. According to the recommendation of Osteoarthritis Research
Society International (OARSI) [11], the knee joints should be
paraffin-embedded in frontal orientation and sectioned
through the entire joint at 80 mm interval. Each section should
be 5 mm.
12. The mouse OA knee damage categories followed by scores in
parentheses include normal (0), loss of Safranin O with no
structural changes (0.5), small fibrillations with no loss of
cartilage (1), vertical clefts down to the layer immediately
below the superficial layer and some loss of surface lamina
(2), vertical clefts/erosion to the calcified cartilage extending
to <25% of the articular surface (3), vertical clefts/erosion to
the calcified cartilage extending to 25–50% of the articular
surface (4), vertical clefts/erosion to the calcified cartilage
extending to 50–75% of the articular surface (5), vertical
clefts/erosion to the calcified cartilage extending >75% of
the articular surface (6). Use mean value of the scores from
the 4 quadrants of an individual knee joint for statistical
analyses.
13. To establish settings, double-click the indicator window to
open the X-ray energy settings. Set the voltage to 60 kV and
current to 167 μA. The “always maximize power” box should

be checked, which will automatically adjust the power to 10 W
and thus adjust voltage or current accordingly. Set the pixel
resolution to 2000 1200 and the X-ray camera image pixel
size to 6 μm. Apply a 0.5 mm aluminum filter.
14. Open “Preferences” in the “Options” tab. Un-check the two
flat-field correction boxes and click “OK.” Select “Acquisition
Modes” in the “Options” tab. Locate the window with the
arrow which corresponds to the filter, camera position, and
pixel size. Set the exposure to 360 and adjust the voltage and
current windows if needed. Check the option, “Acquire bright
+ dark for current mode.” Grab image to obtain a live preview
and right click to view the intensity profile. The average inten-
sity should be ~60%. Note: If the average intensity is not ~60%,
allow the X-ray to stabilize and adjust the exposure until it
reaches ~60%. Open “Preferences” in the “Options” tab.
Check the two flat-field correction boxes and click “OK.”
The average intensity should be 88% (2%).
15. Select “Set Oversized Scan” in the “Actions” tab. Click “Scout
Scan” and allow the scan to run to the bottom of the stage.
Enter the sample name in the “Prefix” window. Select the
“Top” button and click the top of the sample, followed by
the “Bottom” button and click the bottom of the sample.
Repeat this process for each sample in the scout scan preview
window.
16. Designate a file location by choosing “Browse” and select a
location. Set rotation step (deg.) to 0.4, averaging (frames) to
5, random movement to 10. Un-check the following: use
360-degree rotation, open chamber after scanning, camera
offset and partial width. Check “X-Ray off after scanning.”
17. Select “X/Y alignment with a reference scan” under the
“Actions” tab. Move the adjustable box to cover the sample.
Adjust the “Max. x-shift from start point (+)” and “Max.
y-shift from start point (+)” to 20. Change the “Matching
criteria and method” to “Least-Square.” Change the “Projec-
tion pairs to match” to “All.” Click “Match,” and after the
application is complete, click “Accept.”
18. In the “Start” tab, select the area in the sample to preview. For
a mouse femur, there will be multiple segments. Select “Pre-
view.” In the “Output” tab, change the graph to log view. To
toggle between graphs, double-click the graph. Click the tab
underneath the histogram and set the minimal value to
0. Adjust the maximal value, indicated by the red line, by
dragging the line and placing it just after the peaks from the
dataset end. Use this dynamic range for the rest of the samples
in the project.
19. In the “Output” tab, right click and drag over a nonporous
part of a sample. This will open an absorbance vs. distance
graph. This works best when using cortical bone. Record the
shape of the graph (cupping indicates beam-hardening, arching
indicates over-correction of beam-hardening, and flat indicates
little to no beam hardening). Repeat two or three times in
different proximal and distal regions of the sample. Click on
the “Settings” tab and perform the necessary corrections for
each graph type (increase the beam-hardening correction per-
centage for cupping, decrease the beam-hardening correction
percentage for arching, and no further correction needed for
flat graphs). Adjust beam-hardening settings until the graphs
become flat. The same beam-hardening settings are used for
the rest of the samples in a project. For a mouse femur, this
number typically ranges from 20% to 40%.
20. In the “Start” tab, select an area in the sample and click the
“Fine Tuning” button and select “Post-alignment.” Set num-
ber of trials to 5 and set parameter steps to 0.5 or 1. Click
“Start” and refer to the “Output” tab. Toggle through the
choices using the arrow buttons and choose an option that
minimizes changes in pores, edges, and dense particles. If the
fine tuning does not work in the range selected, manually
change the misalignment compensation in the “Settings” tab.
Repeat fine tuning for each part of the sample (there should be
3 or more segments for a mouse femur).
21. In the “Settings” tab, check “Ring Artifacts Reduction,” and
set to 10. In the “Start” tab, select an area in the sample, click
the “Fine Tuning” button, and select “Ring Artifacts Reduc-
tion.” Set number of trials to 5 and set parameter steps to
5. Click “Start” and refer to the “Output” tab. Toggle through
the choices using the arrow buttons and choose an option that
best reduces ring artifacts. This will only be done once per
sample.
22. This application should only be used if there is noise in the
background of the scan, which can be viewed in the “Output”
tab. To turn on smoothing, click the “Settings” tab and check
the “Smoothing” box. Adjust the levels until the noise in the
background is canceled. To see the effect of smoothing, choose
the “Start” tab, select an area in the sample and click “Pre-
view.” Use the same smoothing settings for the rest of the
samples in a project.
23. In the “Start” tab, select an area in the sample to view and click
“Preview.” In the “Output” tab, check “Use ROI,” and choose
a shape. Move and adjust the size of the ROI in the preview
window so that the sample is inside the ROI. Expand the ROI
in different segments of the femur to confirm the whole sample

is within the ROI (from proximal to distal epiphysis).
24. Choose the “Start” tab to see an overview of the scan. Move
the red lines at the top and bottom of the scan to include the
area of interest.
25. There are two ways to reconstruct and save the images. Option
1: Choose the “Start” tab and click the “Start” button. Option
2: Choose the “Start” tab and click the “Add to batch button.”
Continue to load more datasets to the batch by choosing the
“Add to batch button” after reconstructing each sample. When
all datasets are complete, click “Start Batch” in batch manager.
Folders with reconstructed files will appear in the scan folders.
26. Under the “Actions” tab, choose “Save” and “Select a VOI to
Save.” Establish settings in the “Options for saving a VOI”
window under “Saving options.” In the Resize X/Y option,
choose 1. In the View to save option, choose Transaxial (X-Y).
Create a new VOI folder within the reconstruction folder.
Click “OK” to save the VOI.
27. Select the “regions of interest preview” icon. Select the “inter-
polated region of interest” drop-down menu. Select “Round.”
Adjust the circle so that it is slightly smaller than the density
standard and place it in the center of the density standard.
28. Under “File,” choose “Save from ROI.” Create a new ROI
folder and un-check “Create folder named VOI.” Check “Save
ROI-file” and click “OK” to save the ROI file.
29. Select the “regions of interest preview” icon followed by the
“load regions of interest” icon. Open the “.roi” file.
30. Select the “binary selection preview” icon. Select the “From
dataset” tab in the “Histogram” window. Select the “Attenua-
tion coefficient” tab. Scroll to the bottom and record the mean
(total).
31. Scroll through the images to locate where the growth plate
separates. In the dataset window right click the slice chosen and
click “Selection Reference.” Double-click the black bar in the
dataset window to open the “Selection” window. Under the
“Analytic” tab set offset 0.25 mm and height to 2.5 mm.
32. To manually draw an ROI, choose the “regions of interest
preview” icon. Scroll to the lower boundary of the ROI and
draw a region that includes trabecular bone but excludes the
growth plate and cortical bone. Scroll toward the upper
boundary of the ROI and continue drawing skipping several
images in between. The ROIs drawn will interpolate automati-
cally. To save, choose “Save from ROI” under the “File” tab.
Create a new folder for the trabecular ROI and un-check
“Create Folder named VOI” and check “Save ROI-file.” Click

“OK” to save the ROI.
33. Save the ROI by choosing “Save from ROI” under the “File”
tab. Create a new folder for the trabecular ROI. Uncheck
“Create Folder named VOI” and check “Save ROI-file.”
Click “OK” to save the ROI. To determine the thresholding
scale, select the “binary selection preview” icon. Select the
“From dataset” tab in the histogram window. Choose the
“automatic thresholding” icon in the histogram window.
Record the upper and lower threshold values. For a project,
the thresholds for all samples should be recorded and then
averaged.
34. To build a task list to create a trabecular ROI and run analysis,
select the “custom processing preview” icon and select the
“Internal” tab to build a task list. Click + to add a function to
a task list. Do not run the task list yet. Add the functions in the
following order with the described settings: thresholding
(Global—enter the values determined in step 4), despeckle
(type: sweep, 3D space, remove: all except largest object, and
apply to: image), ROI shrink-wrap (mode: shrink-wrap, 2D
space, check “stretch over holes,” diameter: 32 pixels—adjust
as needed), bitwise operations (image ¼ image XOR region of
interest), morphological operations (type: opening, 3D space,
kernel: round, radius: 2, and apply to: image), morphological
operations (type: closing, 3D space, kernel: round, radius:
2, and apply to: image), morphological operations (type: open-
ing, 3D space, kernel: round, radius: 4, and apply to: image),
morphological operations (type: closing, 3D space, kernel:
round, radius: 4, and apply to: image), morphological opera-
tions (type: opening, 3D space, kernel: round, radius: 6, and
apply to: image), morphological operations (type: closing, 3D
space, kernel: round, radius: 6, and apply to: image), morpho-
logical operations (type: opening, 3D space, kernel: round,
radius: 8, and apply to: image), morphological operations
(type: closing, 3D space, kernel: round, radius: 8, and apply
to: image), morphological operations (type: opening, 3D
space, kernel: round, radius: 9, and apply to: image), morpho-
logical operations (type: closing, 3D space, kernel: round,
radius: 9, and apply to: image), morphological operations
(type: opening, 3D space, kernel: round, radius: 10, and
apply to: image), morphological operations (type: closing, 3D
space, kernel: round, radius: 10, and apply to: image), mor-
phological operations (type: erosion, 2D space, kernel: round,
radius: 3, apply to: image), bitwise operations (region of inter-
est ¼ COPY Image), reload (apply to: image), despeckle (type:
sweep, 2D space, remove: all except largest object, apply to:
region of interest), despeckle (type: remove pores, 2D space,
detected: by image borders, apply to: region of interest), save

bitmaps (apply to ROI, file format: BMP, check “custom sub-
folder” and name the folder, check “convert to monochrome”
1 bit), histogram (unit: bone mineral density, choose 3D space,
check “Inside VOI,” choose a file name and location), thresh-
olding (global thresholding, use the values previously recorded
for thresholding), despeckle (remove white speckles, choose
3D space, volume: less than 6 voxels, apply to image), des-
peckle (remove black speckles, choose 3D space, volume: less
than 6 voxels, apply to image), 3D analysis (check “basic
values,” “additional values,” “trabecular thickness” and “tra-
becular separation,” under the “save results as” region, select a
format to save the data output, and name the output of the
dataset), and 3D model (apply to: image inside ROI, type of
file: *.p3g, algorithm: Double-Time Cubes, check “Custom
filename” and name the file).
35. To apply the ROI to the open file, select the “regions of
interest preview” icon. Select the “load regions of interest”
icon and open the “.roi” file. To determine the thresholding
scale, select the “binary selection preview” icon. Choose the
“From dataset” tab in the histogram window and select the
“automatic thresholding” icon in the histogram window.
Record the upper and lower threshold values. For a project,
the thresholds for all samples should be recorded and then
averaged.
36. To add functions to a task list, click + on a highlighted func-
tion. Add the following functions to the task list in the follow-
ing order with the described settings: histogram (unit: bone
mineral density, choose 3D space, check “Inside VOI,” choose
a file name and location), thresholding (global thresholding,
use the values previously recorded for thresholding), despeckle
(remove white speckles, choose 3D space, volume: less than
6 voxels, apply to image), despeckle (remove black speckles,
choose 3D space, volume: less than 6 voxels, apply to image),
3D analysis (check “basic values,” “additional values,” “trabe-
cular thickness” and “trabecular separation,” under the “save
results as” region, select a format to save the data output, and
name the output of the dataset), and 3D model (apply to:
image inside ROI, type of file: *.p3g, algorithm: Double-
Time Cubes, check “Custom filename” and name the file).
37. Select the “BATMAN” icon. Choose the “Add” button and
load a trabecular ROI dataset (from step 3 if manual method
was used and step 4 if automated method was used). Under
the “Add” arrow button, choose “Load ROI” and select the “.
roi” file. Add all samples within a project and click “Start” to
run the task list on all samples.
38. To define a reference point, first select the “regions of interest

preview” icon. Define the length of the bone by using the
Z-position in the “Dataset” window and calculate 45% of the
femur’s length. Move the distance calculated from the distal
end of the femur. Select the slice in the “Dataset” window. To
make the position the selection reference, right-click and
choose “Selection Reference.” Double-click the black bar in
the dataset window to open the “Selection” window. Under
the “Analytic” tab set offset 0 mm and height to 0.6 mm.
39. At this location in the diaphysis, mostly cortical bone remains.
To create an ROI that contains the cortical bone, start by
choosing the “regions of interest preview” icon. Scroll to the
lower boundary of the ROI. Click the “interpolated region of
interest” dropdown and choose “Elliptic.” Include the entire
area of the bone within the ellipse. Scroll toward the upper
boundary of the ROI and continue moving the ellipse to cover
the entire tissue, skipping several images in between. The
ellipses will interpolate automatically.
40. To save the cortical region of interest (ROI) choose “Save from
ROI” under “File.” Create a new cortical ROI folder. Uncheck
“Create folder named VOI” and check “Save ROI-file.” Click
“OK” to save the ROI.
41. To apply an ROI, select the “regions of interest preview” icon.
Select the “load regions of interest” icon and open the “.roi”
file. To determine a thresholding scale, start by selecting the
“binary selection preview” icon. Select the “From dataset” tab
in the histogram window. Choose the “automatic threshold-
ing” icon in the histogram window. Record the upper and
lower threshold values. For a project, the thresholds for all
samples should be recorded and then averaged.
42. Choose the “custom processing preview” icon and select the
“Internal” tab to build a task list. Add the following functions
to the task list in the order with the described settings: thresh-
olding (global threshold, use values recorded from step 6),
despeckle (remove white speckles, choose 3D space, volume:
less than 6 voxels, apply to image), despeckle (remove black
speckles, choose 3D space, volume: less than 6 voxels, apply to
image), ROI shrink-wrap (mode: shrink-wrap, 2D space, check
“stretch over holes,” choose appropriate number of pixels for
phenotype for diameter, morphological operations and bitwise
operations (use a combination as needed to rid the sample of
any trabecular bone so that only cortical bone remains), des-
peckle (use in between morphological operations as necessary
to rid the region of trabecular bone, type: sweep, 2D or 3D
space depending on sample, remove: all except largest object,
apply to: image), 2D analysis (under the “add to report” region
choose “all results,” check “Append Summary results to file,”

and choose a file name and location), and 3D model (apply to:
image inside ROI, type of file: *.p3g, algorithm: double-time
cubes, check “custom filename” and name the file).
43. Add the following functions to the end of the task list in order
with the described settings: ROI shrink-wrap (mode: shrink-
wrap, 3D space to ensure that only the bone will be measured
when calculating BMD, check “stretch over holes,” diameter:
choose appropriate number of voxels for phenotype), reload
(apply to: image), and histogram (unit: bone mineral density,
3D space, check “Inside VOI” and “Append Summary results
to file,” and choose a file name and location).
44. To use Batch Manager, select the “BATMAN” icon. Choose
the “Add” button and load a cortical ROI dataset. Under the
“Add” arrow button, choose “Load ROI” and select the “.roi”
file. Load all samples within a project and click “Start” to run
the task list on all samples.
45. To minimize complications with IP injections, dosage can be
split between IP and subcutaneous injections.
46. In a tissue processor, in order, clear the samples for 2–4 h each,
in the following solutions: 70% ethanol, 95% ethanol 2, 100%
ethanol 3, and xylene 2.
47. Add a paper label containing the specimen ID to the end of the
tube. Fill the remainder of the tube with embedding solution
until the tube is approximately ¾ full.
48. Monitor the volume of embedding solution for evaporation
and add additional embedding solution to ensure the samples
remain submerged, when necessary. Polymerization is com-
plete when the embedding solution has turned to a hardened
plastic, the tube is smooth, and no longer tacky.
49. When the saw blade cuts completely through the rod, the saw
blade will automatically stop as it sits on the specimen arm.
50. Example: a 15 mm setting on the micrometer will yield a 35 μm
wafer.
51. Polymerization is complete when the embedding solution has
turned to a hardened plastic, the mold is smooth, and no
longer tacky.
52. Expose the surface of the polymerized block using the grinder
by holding the back of the block so that the face is in contact
with the abrasive paper. While the block is still wet, apply
pressure as needed and rotate the paper to polish away the
surface of the mold.
53. Replace with fresh EDTA after 7 days. Samples should be
flexible before processing.
54. To prevent cross-contamination, wipe the forceps with a clean

Kimwipe prior to placing them back into the forceps warmers.
55. For histological sections, 5 μm is the standard thickness.
Acknowledgements
Work in our laboratories has been or is supported by the Van Andel

Institute and the following grants from NIH: BOW (AR068668),
AGR (AR053237), and TY (AG061086).
References
1. Youlten SE, Baldock PA (2019) Using mouse 7. Bouxsein ML et al (2010) Guidelines for
genetics to understand human skeletal disease. assessment of bone microstructure in rodents
Bone 126:27–36 using micro-computed tomography. J Bone
2. Williams BO, Warman ML (2017) CRISPR/ Miner Res 25(7):1468–1486
CAS9 technologies. J Bone Miner Res 32 8. Parfitt AM et al (1987) Bone histomorphome-
(5):883–888 try: standardization of nomenclature, symbols,
3. Kamekura S et al (2005) Osteoarthritis devel- and units. Report of the ASBMR Histomor-
opment in novel experimental mouse models phometry nomenclature committee. J Bone
induced by knee joint instability. Osteoarthr Miner Res 2(6):595–610
Cartil 13(7):632–641 9. Kedlaya R et al (2013) Sclerostin inhibition
4. Glasson SS, Blanchet TJ, Morris EA (2007) reverses skeletal fragility in an Lrp5-deficient
The surgical destabilization of the medial mouse model of OPPG syndrome. Sci Transl
meniscus (DMM) model of osteoarthritis in Med 5(211):211ra158
the 129/SvEv mouse. Osteoarthr Cartil 15 10. Ma HL et al (2007) Osteoarthritis severity is
(9):1061–1069 sex dependent in a surgical mouse model.
5. Welch ID et al (2009) The retinoic acid bind- Osteoarthr Cartil 15(6):695–700
ing protein CRABP2 is increased in murine 11. Glasson SS et al (2010) The OARSI histopa-
models of degenerative joint disease. Arthritis thology initiative - recommendations for histo-
Res Ther 11(1):R14 logical assessments of osteoarthritis in the
6. Fang H et al (2018) Early changes of articular mouse. Osteoarthr Cartil 18(Suppl 3):
cartilage and subchondral bone in the DMM S17–S23
mouse model of osteoarthritis. Sci Rep 8
(1):2855
Chapter 12
Drill Hole Models to Investigate Bone Repair

Zhijun Li and Jill A. Helms
Abstract
Our understanding of the mechanisms underlying fracture healing is rapidly developing and is contributing
to new therapeutic strategies to enhance repair. To gain new insights, animal models must also evolve. From
initially imprecise, uncontrolled bone defects we now have precise injury models that still capture all of the
stages and phases of bone repair yet do so in a highly reproducible manner. The simple mono-cortical defect
model allows assessment of bone repair through a cartilage intermediate, e.g., endochondral ossification, as
well as direct bone repair, e.g., intramembranous healing. Cellular contributions of the periosteum can be
distinguished from contributions originating in the bone marrow. In this chapter, we focus on the
advantages of this bone repair model, as well as its limitations.
Key words Animal model, Bone regeneration, Fracture healing, Orthopedics
1 Introduction
Bone fractures are one of the most common traumatic injuries and
their etiologies typically involve falls and/or accidents [1]. Fractures
can also occur because of underlying disease processes, e.g., cancer
and osteoporosis that weaken the bone structure and increase the
likelihood of fracture. Fracture severity is typically classified accord-
ing to whether it is closed or open (compound), simple or com-
minuted, and infected or not [2, 3]. The more severe the fracture,
the longer is the healing time. Despite the known importance of
these variables, few of the conditions are replicated in animal
models.
Standardized animal models of bone repair typically simplify
fracture severity, in hopes of producing reproducible injuries. For
example, one commonly model using blunt trauma created by
ophthalmic forceps or three-point bending equipment that is deliv-
ered to the tibia or femur and which creates a closed compound
fracture [4, 5]. These fractures can then be allowed to heal with or
without stabilization. In stabilized scenarios, an intramedullary
“nail” (e.g., an insect pin) is typically inserted into the femur to
193
194 Zhijun Li and Jill A. Helms
align the bone ends. This method of stabilization is relatively easy,

but it also makes it impossible to ascertain how cells in the bone
marrow cavity contribute to the repair process [6, 7]. External, e.g.,
Ilizarov, fixators can be adapted to mouse models [8, 9] but tend to
be technique-sensitive and time-consuming.
To circumvent these issues, we and others have used a drill hole
to model the stages of fracture healing [10–12]. The method is
simple, highly reproducible, and requires only a few surgical tools.
Producing a mono-cortical “drill hole” in the tibia does not require
the placement of an intramedullary nail for stabilization; the stabi-
lization is provided by the second, intact cortex. Consequently, the
contribution of cells from the marrow cavity and surrounding soft
tissues to the healing process is not impinged upon by the presence
of internal or external devices [13, 14]. The holes can be of differ-
ent diameters but in our hands, a smaller diameter injury is prefera-
ble because the animal experiences minimal discomfort, and in the
tibia, the healing process is identical whether the injury is large or
small. The animal has several anatomical sites that are amenable for
drilling. In the appendicular skeleton, several sites have been used
including the proximal third of the tibia, the femoral neck, and the
metatarsus [15, 16]. The mandible, maxilla, and calvaria can also be
used [10, 17, 18].
In all appendicular locations, the surgery consists of producing
a complete osteotomy in one cortex, while leaving the far cortex
intact; the resulting gap between the bone ends constitutes the
fracture (see Fig. 1). The diameter of the osteotomy should be
approximately equal to the thickness of the cortex (i.e., in mice,
~0.4 mm, and in rats, ~0.8 mm) in order to provide sufficient
stabilization of the cut bone ends. If the diameter of the osteotomy
is increased, then the risk of spontaneous fracture of the far cortex is
also increased. When this happens, the fracture is considered unsta-
ble and healing ensues via a large, amorphous cartilage callus [19].
In rodents, the fracture healing process is complete within
~21 days. The initial inflammatory phase, characterized by the
accumulation of granulation tissues, takes place between post-
surgery days 1 and 3. The soft (cartilage) callus stage ranges from
post-surgery days 3 to 10; the hard callus phase is between post-
surgery days 10 and 14, and thereafter remodeling takes over until
the original bone contour is achieved.
One major advantage of this drill hole model is that while the
endosteum heals via intramembranous ossification (see Fig. 1), the
injured periosteum heals via endochondral ossification (see Fig. 2).
This spatial separation between endochondral and intramembra-
nous healing within the same injury site is especially useful when
trying to distinguish a disruption in hypertrophic cartilage forma-
tion from an overall disruption in bone repair [20].
Drill Hole Models to Investigate Bone Repair 195
Fig. 1 (a) General morphological view of fibular, (b) sagittal view, (c) coronal view of 2D/3D reconstruction of
intact fibular, red arrow points the surgical site. (d) Safranin O stained images show the monocortical defect in
Fig. 2 (a) Pentachrome stained images show the intact bone and healing site on post-injury (PID), (b) day
3, (e) 7, and (f) 28. (c) The PCNA expressing levels and (d) Collagen type II shows by immunostaining on PID3.
Abbreviations: gp growth plate, cb cortical bone, bm bone marrow
A second major advantage of the drill hole model is that the

defect can be filled with a biomaterial, a biologic, a scaffold, and/or
cells, and their contribution to the healing process can be readily
assessed. For example, we have placed collagen sponges soaked in
recombinant BMP2 into drill holes (see Fig. 3), thus eliciting in
rodents the same robust, endochondral bone repair from the peri-
osteum [21] that is reported in patients [22]. Moreover, the bone
that forms is heterotopic [23], also as reported in patient
populations [24].
Drill hole defects can also receive osteogenic cells, whose con-
tribution to the healing process can be monitored (see Fig. 3). For
example, osteogenic cells that co-expressed luciferase, the enzyme
that cleaves luciferin to produce photons of light, were placed into a
tibial defect, then in vivo live imaging was used to monitor the cells
as a function of time (see Fig. 3). Histology and immunostaining
with a luciferase antibody confirmed that the loss of signal was not
due to a loss of cells; rather, the diminishment in signal was because
the osteogenic cells secreted a mineralized matrix, the ability of the
photons of light to pass through the denser tissue resulted in a loss
of signal (see Fig. 3).
Fig. 1 (continued) the fibular. (e) Schematics shows gap between the bone ends constitutes the fracture.
Pentachrome stained images show healing site on post-injury (f) day 2 and (i) 21. Safranin O stained image
indicates heals via intramembranous ossification on post-injury (g) day 6; intramembranous healing show by
(h) TRAP stained image at the same time point. Abbreviations: gp growth plate, cb cortical bone, bm bone
marrow
Fig. 3 (a) Schematics and (b, c) Aniline blue stained images show the drill hole defects filled with a scaffold
with BMP2 and contribution to the healing process can be readily assessed. (d) Drill hole defects can receive
osteogenic cells that co-expressed luciferase, the enzyme that cleaves luciferin to produce photons of light,
were placed into a tibial defect, (e) then in vivo live imaging was used to monitor the cells as a function of time.
(f) Drill hole defects can be with growth factors(L-Wnt3a), pentachrome stained images show healing site with
(g) L-Wnt3a and with (h)L-PBS on PID 7. (i) The model can be adapted to reflect an osteonecrotic bone lesion,
(j) Aniline blue stained image shows the healing site on PID1. (k) DAPI staining show osteocytes at the cut bone
ends at the same time point. (l) The adapted drill holes model can be filled with a biomaterial, a biologic, a
scaffold, and/or cells, and their contribution to the healing process can be readily assessed (m, n). Abbrevia-
tions: gp growth plate, cb cortical bone, bm bone marrow
The model can be adapted to reflect an osteonecrotic bone

lesion (see Fig. 3). Although many animal models of osteonecrosis
exist [25], most are challenging to perform [26–28] and none fully
replicate the complex etiology of osteonecrosis in patients [29–
31]. Cryoablation can be performed immediately after creating
the drill hole, by placing the top of a metal drill bit in contact
with a piece of dry ice, then positioning the drill tip on the cut
bone ends [32]. The loss of DAPI staining in osteocytes at the cut
bone ends can be used to verify the extent of cryoablation
[32]. Using cryoablation, a simple drill hole injury can be con-
verted into a nonhealing osteonecrotic bone defect, which mimics
at least some of the features of osteonecrosis in patients.
The drill hole model has other advantages. Drilling sites are
often adjacent to anatomical landmarks such as on the tibial crest or
mid-shaft of the diaphysis of the femur, which allow for easy identi-
fication in tissue block during sectioning. Drill holes can also be
produced in the calvaria, either at the midline overlying the sagittal
sinus (see Fig. 4) or in the center of the frontal or parietal bones (see
Fig. 4). In the case of calvarial defects, care should be taken to avoid
any damage to the underlying dura mater (see Fig. 4). The absence
of a large marrow space means that even very small diameter, e.g.,
0.3 mm, calvarial defects are slow to heal. Larger calvarial defects
are often referred to as “critical size” defects, although the absolute
accuracy of that nomenclature has been disputed [33] .
There are limitations to drill hole models. First, the holes
created are small, and consequently surgeries typically require the
use of a microscope. Second, in the absence of nearby landmarks, it
may be difficult to identify the injury site once it has healed. The
healing time is also short, and at least in the appendicular skeleton,
the contributions of the marrow to the healing process are substan-
tial. As a consequence, the potential benefit of allogeneic cells or
biomaterials may not be appreciated because of the robust endoge-
nous healing response. In these cases, it may prove beneficial to
slow natural healing by cryoablation [32] or by genetic [20] or
biochemical [34] means.
2 Materials
2.1 Instrument As a survival surgery, sterile instruments must be used for each
Preparation animal. Therefore, all surgical instruments are placed in an envelope
and autoclaved prior to use. When performing multiple animal
surgeries at one time, instruments may be sterilized between ani-
mals using a glass bead sterilizer (30 s stay at 240–270 C). Take
care to ensure that instruments are no longer hot before touching
any tissues. Gloves should be changed between animals, and per-
sonal protective equipment (PPE) including a head cover, face
mask, and shoe covers should always be worn.
Fig. 4 (a) Skeletal staining show calvarial anatomy. (b) Drill holes can also be produced at the midline
overlying the sagittal sinus or in the center of the frontal or parietal bones. (c, d, e) 3D and 2D reconstruction of
calvarial defects model. (f) Schematics and (g) Aniline blue stained images show the case of calvarial defects.
(h) Calvarial defects can filled with a biomaterial, a biologic, a scaffold, and/or cells, (i) aniline blue stained
image and (k) BrdU immunoassayed image show their contribution to the healing process. Abbreviations:
P parietal bone, F frontal bone, IP intraparietal bone, N nasal bone, MN mandible
Table 1
Anesthetic agents
Agent name Dosage (mg/kg) Route

Ketamine hydrochloride 70–100 mg/kg Intraperitoneal (IP)
Xylazine 5–10 mg/kg Intraperitoneal (IP)
Isoflurane 2–5% gas with oxygen Inhalation (INH)
2.2 Anesthetic See Table 1.

Agents
2.3 Analgesic Agents See Table 2.
2.4 Anti-Infective See Table 3.

Agents
Table 2
Analgesic agents
Dosage
Agent name (mg/kg) Route Duration and frequency of administration
Buprenorphine 0.1 mg/kg Subcutaneous Every 6–12 h for 3 days post-op, then as needed.
(SQ)
Carprofen 5–10 mg/kg Subcutaneous Every 24 h for 3 days, then as needed
(SQ)
Buprenorphine 0.5–1.0 mg/ Subcutaneous Buprenorphine SR will be given before the start of
SR kg (SQ) surgery, and will last up to 72 h post-op.
Table 3
Anti-infective agents
Dosage
Agent name (mg/kg) Route Duration and frequency of administration
Cefazolin 25 mg/kg Subcutaneous The first dose intraoperatively and second dose 2 h
(SQ) postoperatively
Enrofloxacin 10 mg/kg Subcutaneous Only one dose intraoperatively, continue for a total of
(SQ) 2 days
Buprenorphine 0.5–1.0 mg/ Subcutaneous Buprenorphine SR will be given before the start of
SR kg (SQ) surgery, and will last up to 72 h post-op.
3 Methods
3.1 Animal 1. All animals are weighed prior to surgery and pre-surgical
Preparation weights are recorded on the surgical record.
2. Ophthalmic lubricant is placed in each eye to avoid corneal
drying during surgery.
3. Anesthetize the animal as per established regimens, such as
indicated in Table 1.
4. Remove hair with an electric clipper, then clean loose hairs with
ethanol pad.
5. At this stage, analgesics including Buprenorphine or Buprenor-
phine SR are given.
6. Next, the shaved skin is aseptically prepared and draped accord-
ing to APLAC Guidelines for Rodent Survival Surgery. Sterili-
zation of the area is performed using cotton-tipped applicators,
alternating between a Betadine solution and 70% ethanol for a
total of 3 times.
7. The animal is gently placed on top of a clean, absorbent surface
positioned on a validated heat source (e.g., circulating water
blanket, slide warmer, instant heat device) to prevent hypother-

mia during surgery.
8. Before making an incision, the depth of anesthesia is confirmed
via a toe-pinch; the withdrawal reflex should be abolished
before proceeding.
3.2 Anesthetic 1. Place the animal in the induction chamber.

3.2.1 Isoflurane 2. Adjust the oxygen flowmeter to 0.8–1.5 L/min.
Anesthesia 3. Adjust the isoflurane vaporizer to 3–5%.
4. Use the mask connected to the Bain circuit.
5. Adjust the flowmeter to 0.4–0.8 L/min.
6. Adjust the isoflurane vaporizer to 2–2.5%.
3.2.2 Injectable 1. Injectable anesthetic volume depends on sex, age, strain, body
Anesthesia weight, and the body condition of the animal.
2. Prepare the cocktail solution the day before surgery. Immedi-
ately before use, make sure that the solution is thoroughly
mixed.
3. In mice, injectable anesthetics are best administered via IP and
IV routes.
4. Duration of anesthesia is approximately 20–30 min.
3.3 Long Bone 1. After adequate anesthesia, an 8–10 mm skin incision is made in
Cortical Hole Drilling anterior surface of the tibia or on the lateral surface of the
femur.
2. A full-thickness flap is raised at the incision sites.
3. The overlying muscle is gently dissected off the tibia or femur.
4. The site is gently irrigated with sterile saline. Immediately
thereafter, a small hole is created using a dental drill running
at 1000–2000 rpm. The drill must be new and sharp; otherwise
too much pressure must be exerted to advance the drill and the
risk of touching the far cortex is increased.
5. The hole created in the anterior tibial plateau or in the
mid-shaft of femur should have an outer diameter of approxi-
mately 0.8–1.0 mm.
6. The drill hole should penetrate through a single tibial/femur
cortex only. No irrigation is required after creating the drill
hole. Bleeding usually stops spontaneously.
7. After creating the skeletal injury, soft tissues are closed with
absorbable suture.
8. The skin incision is closed with nonabsorbable sutures.
9. The nonabsorbable sutures are removed 7–10 days later.
10. Antibiotics are typically not required, although any signs of

prolonged inflammation, e.g., longer than 24 h, may require
the use of antibiotics; check with veterinary staff to ascertain
whether this is necessary (see Note 1).
3.4 Calvaria Cortical 1. Once adequate anesthesia has been verified, a 3 mm skin inci-
Bone Hole Drilling sion is made in the mid-line at the caudal portion of the skull
(see Note 2).
2. The periosteum is gently elevated and reflected.
3. A drill hole is created using of a micro-dissecting trephine and
sterile saline irrigation.
4. Take extreme care to leave the underlying dura mater
undisturbed.
5. The skin incision is closed with nonabsorbable suture.
6. The nonabsorbable sutures are removed 7–10 days later (see
Note 3).
3.5 Parameters 1. Animals are transferred into a clean cage which is partially over
Monitored After a 37 C heating pad. This allows animals, once ambulatory, to
Surgery choose whether to stay on the warm side or on the
non-heated side.
2. Separate awake animals from recovering animals.
3. Animals are allowed to ambulate freely but are closely moni-
tored for any evidence of pain as indicated by licking, biting,
crouching, vocalizations, teeth chattering, stiff walking, failure
to eat, self-imposed isolation/hiding, self-mutilation, rapid
breathing, open mouthed breathing, squinting, muscle rigidity,
lack of muscle tone, twitching/trembling, or abnormal
posture.
3.6 Early Euthanasia 1. The most likely complication of skeletal surgery is that the
Criteria injury site becomes infected. During the initial recovery period
(0–3 days post-surgery) closely evaluate each animal for signs of
inflammation/infection, such as redness and swelling, or signs
of exudate at the surgical site.
2. If signs of infection are found, consult with veterinarians about
the use of antibiotics. If signs of infection persist, the animal
should be euthanized.
3. The post-surgical weight of each mouse is monitored daily. If
any animal exhibits loss of weight greater than 10%, they are
typically removed from the study and euthanized.
4 Notes
1. A surgeon’s technique is a major influence on the outcome of

any animal surgery. Proper sterile technique and tissue
handling both decrease postoperative complications, increase
survival rates, and ensure reproducible results. Whenever there
is a doubt about an aspect of surgical care, consult a
veterinarian.
2. Handle all tissues gently, avoiding any unnecessary tissue dam-
age. Remember, healing duration depends on the overall health
of the surgical site.
3. The size of the hole is of utmost importance when developing
any bone drilling model on any sites. Consider the smallest
diameter hole possible, to lessen complicates and avoid unnec-
essary pain during the recovery process.
Acknowledgements
We thank previous Helms lab members including Drs. Philipp

Leucht, Benjamin Salmon, Bo Liu, Wei Jing, and Sylvain Mouraret
for their invaluable contributions over the years to our cumulative
understanding of bone healing.
References
1. Bell A, Templeman D, Weinlein JC (2016) 6. O’Neill KR, Stutz CM, Mignemi NA, Burns
Nonunion of the femur and tibia: an update. MC, Murry MR, Nyman JS et al (2012) Micro-
Orthop Clin North Am 47(2):365–375 computed tomography assessment of the pro-
2. Papakostidis C, Kanakaris NK, Pretel J, gression of fracture healing in mice. Bone 50
Faour O, Morell DJ, Giannoudis PV (2011) (6):1357–1367
Prevalence of complications of open tibial 7. Einhorn TA (1995) Enhancement of fracture-
shaft fractures stratified as per the Gustilo- healing. J Bone Joint Surg Am 77(6):940–956
Anderson classification. Injury 42 8. Tay BK, Le AX, Gould SE, Helms JA (1998)
(12):1408–1415 Histochemical and molecular analyses of dis-
3. Ktistakis I, Giannoudi M, Giannoudis PV traction osteogenesis in a mouse model. J
(2014) Infection rates after open tibial frac- Orthop Res 16(5):636–642
tures: are they decreasing? Injury 45 9. Choi P, Ogilvie C, Thompson Z, Miclau T,
(7):1025–1027 Helms JA (2004) Cellular and molecular char-
4. Cheng L, Ye F, Yang R, Lu X, Shi Y, Li L et al acterization of a murine non-union model. J
(2010) Osteoinduction of hydroxyapatite/ Orthop Res 22(5):1100–1107
beta-tricalcium phosphate bioceramics in mice 10. Leucht P, Jiang J, Cheng D, Liu B,
with a fractured fibula. Acta Biomater 6 Dhamdhere G, Fang MY et al (2013) Wnt3a
(4):1569–1574 reestablishes osteogenic capacity to bone grafts
5. Kayal RA, Siqueira M, Alblowi J, McLean J, from aged animals. J Bone Joint Surg Am 95
Krothapalli N, Faibish D et al (2010) (14):1278–1288
TNF-alpha mediates diabetes-enhanced chon- 11. Tanaka K, Tanaka S, Sakai A, Ninomiya T,
drocyte apoptosis during fracture healing and Arai Y, Nakamura T (2010) Deficiency of vita-
stimulates chondrocyte apoptosis through min A delays bone healing process in
FOXO1. J Bone Miner Res 25(7):1604–1615
association with reduced BMP2 expression 23. Minear S, Leucht P, Miller S, Helms JA (2010)
after drill-hole injury in mice. Bone 47 rBMP represses Wnt signaling and influences
(6):1006–1012 skeletal progenitor cell fate specification during
12. Behr B, Leucht P, Longaker MT, Quarto N bone repair. J Bone Miner Res 25
(2010) Fgf-9 is required for angiogenesis and (6):1196–1207
osteogenesis in long bone repair. Proc Natl 24. Lorenzo J, Horowitz M, Choi Y (2008)
Acad Sci U S A 107(26):11853–11858 Osteoimmunology: interactions of the bone
13. Jawad MU, Fritton KE, Ma T, Ren PG, Good- and immune system. Endocr Rev 29
man SB, Ke HZ et al (2013) Effects of scler- (4):403–440
ostin antibody on healing of a non-critical size 25. Fan M, Peng J, Wang A, Zhang L, Liu B, Ren Z
femoral bone defect. J Orthop Res 31 et al (2011) Emu model of full-range femoral
(1):155–163 head osteonecrosis induced focally by an alter-
14. Yang F, Wang J, Hou J, Guo H, Liu C (2013) nating freezing and heating insult. J Int Med
Bone regeneration using cell-mediated respon- Res 39(1):187–198
sive degradable PEG-based scaffolds incorpor- 26. Allen MR (2009) Bisphosphonates and osteo-
ating with rhBMP-2. Biomaterials 34 necrosis of the jaw: moving from the bedside to
(5):1514–1528 the bench. Cells Tissues Organs 189
15. Christou C, Oliver RA, Pelletier MH, Walsh (1–4):289–294
WR (2014) Ovine model for critical-size tibial 27. Allen MR, Kubek DJ, Burr DB, Ruggiero SL,
segmental defects. Comp Med 64(5):377–385 Chu TM (2011) Compromised osseous heal-
16. Berner A, Reichert JC, Woodruff MA, ing of dental extraction sites in zoledronic acid-
Saifzadeh S, Morris AJ, Epari DR et al (2013) treated dogs. Osteoporos Int 22(2):693–702
Autologous vs. allogenic mesenchymal progen- 28. Velez R, Soldado F, Hernandez A, Barber I,
itor cells for the reconstruction of critical sized Aguirre M (2011) A new preclinical femoral
segmental tibial bone defects in aged sheep. head osteonecrosis model in sheep. Arch
Acta Biomater 9(8):7874–7884 Orthop Trauma Surg 131(1):5–9
17. Yuan X, Pei X, Zhao Y, Li Z, Chen CH, Tulu 29. Zalavras CG, Lieberman JR (2014) Osteone-
US et al (2018) Biomechanics of immediate crosis of the femoral head: evaluation and treat-
postextraction implant osseointegration. J ment. J Am Acad Orthop Surg 22(7):455–464
Dent Res 97(9):987–994. https://doi.org/ 30. Hernigou P, Flouzat-Lachaniette CH,
10.1177/0022034518765757 Delambre J, Poignard A, Allain J, Chevallier
18. Chen T, Li J, Cordova LA, Liu B, Mouraret S, N et al (2015) Osteonecrosis repair with bone
Sun Q et al (2018) A WNT protein therapeutic marrow cell therapies: state of the clinical art.
improves the bone-forming capacity of auto- Bone 70:102–109
grafts from aged animals. Sci Rep 8(1):119 31. Hamadeh IS, Ngwa BA, Gong Y (2015) Drug
19. Colnot C, Romero DM, Huang S, Helms JA induced osteonecrosis of the jaw. Cancer Treat
(2005) Mechanisms of action of demineralized Rev 41(5):455–464
bone matrix in the repair of cortical bone 32. Salmon B, Liu B, Shen E, Chen T, Li J, Gillette
defects. Clin Orthop Relat Res 435:69–78 M et al (2017) WNT-activated bone grafts
20. Leucht P, Kim JB, Currey JA, Brunski J, Helms repair osteonecrotic lesions in aged animals.
JA (2007) FAK-mediated mechanotransduc- Sci Rep 7(1):14254
tion in skeletal regeneration. PLoS One 2(4): 33. Cooper GM, Mooney MP, Gosain AK, Camp-
e390 bell PG, Losee JE, Huard J (2010) Testing the
21. Minear S, Leucht P, Jiang J, Liu B, Zeng A, critical size in calvarial bone defects: revisiting
Fuerer C et al (2010) Wnt proteins promote the concept of a critical-size defect. Plast
bone regeneration. Sci Transl Med 2 Reconstr Surg 125(6):1685–1692
(29):29ra30 34. Leucht P, Lam K, Kim JB, Mackanos MA,
22. Neagu TP, Tiglis M, Cocolos I, Jecan CR Simanovskii DM, Longaker MT et al (2007)
(2016) The relationship between periosteum Accelerated bone repair after plasma laser cor-
and fracture healing. Romanian J Morphol ticotomies. Ann Surg 246(1):140–150
Embryol 57(4):1215–1220
Chapter 13
Generation and Experimental Outcomes of Closed Femoral

Fracture in Mice
Joseph L. Roberts, Christopher W. Kinter, and Hicham Drissi
Abstract
Fracture healing requires the integration of many cell types, growth factors, and cytokines that cannot be
adequately studied using in vitro and in silico models. This has prompted the development of highly
informative in vivo animal models to understand the complexities of fracture repair. Here, we describe a
modified procedure for mice, first developed for rats by Bonnarens and Einhorn, that does not require a
skin incision or suturing. This procedure involves boring a hole through the skin and articular surface of the
femoral condyle with a 25-gauge needle, fixation with a K-wire, and creation of a transverse mid-diaphyseal
fracture using a three-point bending fracture device. Fracture healing can be assessed using a variety of
techniques, including microcomputed tomography, torsion testing, histological and histomorphometric
analyses, and assessment of gene expression. There are many orthopedic trauma applications of this murine
femoral fracture model ranging from assessment of safety and efficacy of novel therapeutics to the influence
of specific genes on bone repair.
Key words Fracture, Murine, Bone repair, Healing, Closed fracture, micro-CT, Torsion testing
1 Introduction
The first report of the standard closed fracture in rats was described
by Bonnarens and Einhorn in 1984 [1]. It has since been adapted
to other model systems such as mice, which have become one of the
most widely accepted translational models to study fracture pheno-
types [2–4]. The availability of numerous transgenic mouse strains
coupled with the simplicity, speed, and reproducibility of this pro-
cedure have proven useful for studying both genetic and therapeu-
tic influences on healing [5]. Studies using this model have led to
significant advancements in our understanding of the mechanisms
by which fractures heal under both pathological and treatment
conditions.
After administration of anesthetics, this method involves the
creation of a hole in the femur medullary canal by perforating the
skin and articular surface of the femoral condyle with a 25-gauge
205
206 Joseph L. Roberts et al.
needle. The K-wire is then inserted into the medullary canal using a
retrograde approach and covered by the skin. Finally, the operated
limb is held perpendicular to the line of fracture, beneath a blunt
guillotine device, wherein a transverse fracture is created by three-
point bending. Administration of preoperative analgesics helps to
manage pain.
Fractured femurs are typically harvested at multiple timepoints
to evaluate the progression of healing. Fracture healing and quality
of newly formed bone can be quantitatively assessed using micro-
computed tomography (micro-CT) imaging and torsion testing,
respectively. Radiographic scoring is another useful qualitative mea-
sure of fracture healing. Typically, specimens are fixed in formalin as
needed and processed for histology. Histological and histomorpho-
metric assessments allow for the assessment of fracture callus matrix
composition through quantification of the distribution of cartilage,
bone, and fibrous tissue. Histological sections can be further uti-
lized for immunohistochemical analyses to detect cells and tissues
that express growth factors, cytokines, and other proteins of inter-
est during fracture healing. RT-qPCR from RNA isolated from
whole callus or from histological sections is also a powerful tool
to gain valuable insight into the relationship between different
experimental models and treatments and expression of genes dur-
ing fracture healing.
2 Materials
2.1 Closed 1. Personnel: It is optimal to have at least two researchers involved

Fracture Model in this surgery. The first researcher (assistant) is needed to
surgically prepare the mice, manipulate non-sterile objects,
and position the mouse. The second researcher (surgeon) per-
forms the surgeries under sterile conditions.
2. Anesthetics and analgesics: 2% isoflurane/oxygen mixture
delivered via nose-cone inhalation and buprenorphine SR
(1 mg/kg administered via subcutaneous injection).
3. Standard sterile surgical equipment: hemostat, chlorohexidine,
25-gauge needles, wire cutter, an appropriate length of 0.01500
diameter K-wire (average length per adult mouse femur:
15.0–15.5 mm; 316 LVM stainless steel, Small Parts), sterile
drapes. All surgical equipment should be opened and main-
tained in a sterile fashion.
4. Preoperative preparation: Hair clippers, chlorohexidine, 70%
isopropyl alcohol, gauze, Puralube Vet ointment, scale,
heating pad.
5. Blunt 3-point bending guillotine device.
6. Sterile gloves, masks, bouffant cap, sterile gown.
7. X-ray device.
Generation of Murine Femoral Fracture 207
2.2 Microcomputed 1. A micro-CT machine.

Tomography 2. Portable hard drive or storage device to extract and store data.
(Micro-CT)
3. Scanning medium such as 70% ethanol or 1 phosphate-
buffered saline.
4. Parafilm.
5. A testing protocol that indicates area of the scan/duration,
tube voltage (kVP), current x time product (mAs), and special
resolution (μm).
6. Foam or gauze.
7. Forceps.
2.3 Torsional Testing 1. Polymethylmethacrylate (PMMA) or dental cement (e.g.,

Stoelting Cat# 51458).
2. Forceps.
3. 1 cm2 Aluminum hollow cubes.
4. Cold saline solution.
5. Gauze.
6. Camera with video capabilities.
7. Caliper for measurements.
8. Personal protective equipment: gloves, laboratory coat, safety
glasses.
9. Waterproof/chemical-resistant laboratory marker.
10. Torsion mechanical testing apparatus/computer with a load
cell that is sensitive enough to detect sample differences in the
order of 0.5–5 Nmm with appropriate tools to manipulate the
machine.
2.4 Radiographic 1. Personnel: at least two trained individuals.

Scoring 2. Anteroposterior and lateral radiographs.
3. Scoring criteria and assessment sheets.
2.5 Histology 1. Paraffin embedding: 10% Neutral buffered formalin, an orbital

shaker, a chelator such as 14% EDTA solution (pH 7.1–7.2) for
decalcification of experimental specimens, 70%, 95%, and 100%
ethanol, xylene, paraffin wax, fume hood, razor blades, metal
base molds for embedding, embedding cassettes, chemical
resistant marker or pencil, and an embedding center to make
paraffin blocks.
2. Paraffin sectioning: Paraffin blocks to be sectioned, a micro-
tome capable of cutting sections as thin as 5 μm, a hot water
bath (set at 45 C), positively charged slides, and a rack for
drying slides.
3. Staining materials: Containers filled with 70%, 95%, and 100%

ethanol, xylene, distilled water, slide staining rack, oven capable
of reaching 56 C, cover slips, and appropriate mounting
solution.
4. Immunohistochemistry materials: 1 Tris–buffered saline,
0.1% Tween-20 (TBST), quenching solution (3% hydrogen
peroxide), blocking buffer (5% normal goat serum in TBST),
a water bath capable of reaching 55 C, microscope with cam-
era, an appropriate antigen retrieval solution (e.g., sodium
citrate), humidified chamber, hydrophobic pen, and appropri-
ate diluted primary and conjugated secondary antibodies (e.g.,
HRP-conjugated, fluorochrome-conjugated).
5. Histomorphometric quantification: Appropriate analysis soft-
ware and microscope with camera.
2.6 RNA Isolation 1. Cryogenic tissue homogenizer or a mortar and pestle.

and RT-qPCR 2. Liquid nitrogen.
3. Temperature-controlled centrifuge.
4. TRIzol.
5. Chloroform.
6. 100% Isopropanol.
7. 75% Ethanol.
8. RNase-free water.
9. Heating block capable of reaching 55–65 C.
10. Nuclease-free microcentrifuge tubes, pipettes, and filter pipette
tips (1000 μL, 200 μL, 10 μL).
11. Surgical scissors and forceps.
2.6.1 RNA Isolation from 1. Freshly cut histological sections.

Histological Sections 2. RNase-away.
3. RNase-free 10 μL tips.
4. Inverted light microscope with a 4 objective.
5. Xylene or other appropriate deparaffinization solution.
6. Kit capable of extracting RNA from formalin-fixed paraffin-
embedded tissue (e.g., Qiagen RNeasy FFPE or RecoverAll
total nucleic acid isolation kit for FFPE).
7. Nuclease-free microcentrifuge tubes.
2.6.2 Reverse 1. UV spectrophotometer or Nano-drop machine.

Transcription and RT-qPCR 2. cDNA synthesis kit.
3. Thermal cycler.
4. Real-time PCR machine.
5. Nuclease-free water.
6. Forward and reverse primers for genes of interest and house-
keeping genes.
7. PCR plates or strips compatible with machine.
8. Centrifuge with plate holder attachments.
9. PCR tubes.
3 Methods
3.1 Closed Femoral Surgeries should be performed in a clean surgical suite or procedure
Fracture Model room on a surface disinfected with 70% ethanol or Quatricide (see
Note 1).
1. Record body weight of mouse prior to fracture (see Notes
2 and 3).
2. Anesthetize the mouse by placing it in a container with air flow
consisting of 2% isoflurane/oxygen mixture.
3. Once unresponsive to stimuli (e.g., pedal reflex), remove the
anesthetized mouse and place on a heating pad, coat eyes with
Puralube vet ointment, insert nose into nose cone with contin-
uous flow of 2% isoflurane/oxygen mixture, and administer
Burprenorphine SR (1 mg/kg subcutaneously) (see Note 4).
4. Shave the entire surgical leg and adjacent skin well with clip-
pers, wipe shaven limb with gauze soaked in chlorohexidine,
then gauze soaked in 70% isopropyl alcohol.
5. Transfer the mouse to the surgical table covered in sterile
drapes, place head inside a nose cone to deliver continuous
2% isoflurane/oxygen mixture in a supine position.
6. Cut drapes to expose a hole only large enough to slip the
surgically prepared limb through.
7. The surgeon holds the operated limb with the middle finger
positioned underneath the knee and the index finger resting on
the quadriceps muscle. The thumb is held against the tibia for
balance.
8. Using their other hand, the surgeon perforates the skin and
articular surface of the femoral condyle with a 25-gauge needle,
rotating (without applying pressure) gently in a single direction
until the medullary cavity is punctured (see Notes 5 and 6) (see
Fig. 1a).
9. The surgeon picks up a K-wire of an appropriate length (aver-
age 15–15.5 mm) using a hemostat.
Fig. 1 Creation of mid-diaphyseal femoral fracture. (a) Perforate the skin and articular surface of the femur
using a 25G needle to puncture the medullary canal. (b) Using a hemostat, a K-wire is inserted into the
medullary canal until it reaches the proximal end of the femur. (c) Excess K-wire is cut leaving 1–2 mm
exposed. (d) The exposed K-wire is bent to form a 90 angle using a hemostat, then (e) covered under the skin.
(f) The surgeon positions the pinned femur perpendicular to the blunt guillotine and pressure is applied to
create a mid-diaphyseal fracture
10. The needle is removed from the marrow cavity and the K-wire
is inserted into the medullary canal using a retrograde approach
until it reaches the proximal end of the femur (see Note 5).
Correct depth of insertion is felt with resistance upon meeting
cortical bone of the proximal femur (see Fig. 1b).
11. Excess K-wire is cut with wire cutters, leaving approximately
3–4 mm of K-wire that is bent to form a 90 angle, then
covered under the skin without suturing (see Fig. 1c–e).
12. The surgical assistant holds the mouse in a prone position,
while the surgeon positions the pinned femur perpendicular
to the blunt guillotine (see Fig. 1f).
13. With their free hand, the surgical assistant raises the lower stage
which applies pressure via three-point bending until an audible
break is noted.
14. Take an X-ray to confirm location, type, and severity of the
fracture.
15. The animal is placed in a clean cage that has food pellets located
at the bottom in addition to the normal food and water.
16. Monitor mice daily for the first 3 days after fracture to ensure
that they bear weight on their operated limbs and are eating
and drinking.
17. Take radiographs at intervals (e.g., 7, 14, 21 days post-
fracture) to ensure that the K-wire is maintaining fixation of
the fracture and track healing.
3.2 Micro-CT Micro-CT is used to examine fracture healing at a later timepoints,

mainly 14 days post-fracture. Fractured femurs can be scanned at
earlier timepoints (i.e., day 7 and 10 post-fracture) to acquire 3D
images, but analyses cannot be accurately completed due to the
small amount of mineralized tissue within the callus at earlier
timepoints.
1. Fix dissected femurs in a sufficient volume of 10% neutral
buffered formalin for 1 week at 4 C prior to scanning.
2. Wash fixed bones in 1 PBS for 30 min at room temperature
on an orbital shaker. Dispose of PBS and replace with fresh
1 PBS.
3. Turn on the machine and verify the pre-programmed testing
conditions.
4. Carefully remove the K-wire pin from the medullary canal and
discard in a designated sharps container.
5. Wrap bones in foam or gauze and insert into a scanning tube
filled with appropriate scanning medium (e.g., 70% ethanol or
1 PBS). Cover top of tube with parafilm to prevent evapora-
tion during scanning.
6. Dock the attachment and close machine.
7. Repeat steps 4 and 5 if machine is equipped with a carousel
that allows batch scanning.
8. Acquire a scout-view, select the areas of interest and analysis
program. The machine computes the estimated completion
time of the scan (see Note 7).
9. Start the scan and wait for the indicated time until the next
sample(s) may be processed.
10. If machine is not equipped with a carousel, repeat steps 2–6
until all the samples are scanned.
11. Reconstruct scan for further analysis.
12. Images are then contoured (and a global threshold is applied if
applicable) to segment out existing cortical bone and analyzed
(see Fig. 2).
3.3 Torsion Testing Care must be taken to ensure the safety of both the individual
performing the test and those in the surrounding area when bio-
mechanical testing is performed. The ideal timepoint for torsion
Fig. 2 Microcomputed tomography (micro-CT) of fracture callus. (a) Manual contouring of the day 14 fracture
callus to exclude existing cortical bone. (b) Example of micro-CT reconstructions of a murine fracture callus at
day 14 post-fracture
testing is at the discretion of the researchers. Torsion testing is

typically most informative of quality of regenerated bone in the
later stages of fracture healing (21 days post-fracture).
1. At time of bone harvest, carefully remove muscle without dis-
turbing the callus leaving the K-wire inside the medullary
cavity. Wrap bone in saline-soaked gauze and store at 20 C
until testing.
2. When ready to perform testing, defrost the bone in a lukewarm
water bath until thawed.
3. Turn on the torsion mechanical testing apparatus with appro-
priate load cell attached.
4. Slowly and carefully pull the surgical K-wire from the fractured
femora. Rewrap the bones in saline-soaked gauze and place on
ice (see Fig. 3a).
5. Measure the total length, maximum outer diameter within the
diaphyseal region, and minimum outer diameter within the
diaphyseal region of the femur with a caliper and record.
6. Take pictures of all bones before testing and label as “before
failure” pictures.
7. Position the femoral condyles inside a 1 cm2 square mold. Pour
the PMMA within the mold to cement each extremity (see
Note 8), wrap bone in cold saline soaked gauze (see Note 9),
and allow it to solidify. Rotate the femur and place the femoral
head/greater trochanter inside another 1cm2 square mold. Fill
with PMMA and allow to solidify (see Note 10). Measure and
record gauge length for all specimens. Take pictures to confirm
sample alignment (see Fig. 3b–g).
8. Start and configure the software for the calibration. The testing
outputs/data analysis should be set (if applicable).
Fig. 3 Potting and torsion mechanical testing of fractured femora. (a) The K-wire is carefully removed from the
fractured limb using forceps. (b) 1 cm2 aluminum cubes are positioned and (c) PMMA powder is added to the
cube until full (d). (e) PMMA liquid is added using a pipette. (f) The femur is carefully positioned inside of
PMMA using forceps and allowed to solidify. (g) The opposite end of the femur is then potted. The fully potted
femur is placed in a torsion machine (h) and torsion is applied until failure (i). (j) Typical data obtained from
torsion testing
9. Place the first specimen in the testing apparatus and tighten one
end into place (see Fig. 3h). Remove the gauze around the
bone.
10. Being careful not to damage the specimen, slowly tighten the
second end until locked. After the position is set, zero the
torque and run the test while recording a video. Record inner
and outer maximum diaphyseal diameters for all samples after
testing. Take pictures to confirm site of failure (see Fig. 3i).
11. Repeat this procedure for all samples.
12. Turn off all machinery.
13. Extract the data and analyze the torque curves to verify testing
accuracy (see Fig. 3j).
3.4 Radiographic 1. Randomly shuffle or code radiographs of fractures to avoid any

Scoring experimental group associations (see Note 11).
2. The observers evaluate anteroposterior and lateral radio-
graphs in a blinded fashion using a validated radiographic
scoring system, subdivided into the following categories:
(a) periosteal and endosteal reaction, (b) callus opacity, and
(c) cortical remodeling and bridging (see Note 12) (see
Fig. 4).
Category 0 1 2 3 4
Periosteal
and
No reaction Mild reaction Moderate reaction Marked reaction
Endosteal
Reaction
Heterogeneous Heterogeneous
with minimal with moderate
Callus No evidence of Confluent with the
mineralization mineralization and
Opacity mineralization cortices, uniform
and cortices well partially confluent
demarcated with the cortices
Complete
Minimal cortical
Partial cortical cortical union
Cortical union (3 cortical
All cortical edges union (1-2 cortical (no cortical
Remodeling No apparent edges visible)
seen but ill- edges visible) with edges visible)
and remodeling without
defined visible medullary with well
Bridging reformation of the
canal demarcated
medullary canal
medullary canal
Fracture 1 Fracture 2 Fracture 3

Periosteal and Endosteal
3 3 3
Reaction
Callus Opacity 1 2 3
Cortical Remodeling and

1 2 4
Bridging
Fig. 4 Radiographic scoring criteria and representative scores of three independent fractured femora
(a) Periosteal and endosteal reaction is determined using a

four-point scoring system: 0 ¼ no reaction, 1 ¼ mild
reaction, 2 ¼ moderate reaction, and 3 ¼ marked
reaction.
(b) Callus opacity is determined using a four-point scoring
system: 0 ¼ no evidence of mineralization, 1 ¼ heteroge-
neous with minimal mineralization and cortices well
demarcated, 2 ¼ heterogeneous with moderate minerali-
zation and partially confluent with the cortices, and
3 ¼ confluent with the cortices, uniform.
(c) Cortical remodeling and bridging are determined using a
five-point scoring system: 0 ¼ no apparent remodeling,
1 ¼ all cortical edges seen but ill defined, 2 ¼ minimal
cortical union (three cortical edges visible) without refor-
mation of the medullary canal, 3 ¼ partial cortical union
(1–2 cortical edges visible) with visible medullary canal,

and 4 ¼ complete cortical union (no cortical edges visible)
with well-demarcated medullary canal.
3. Collect evaluation sheets, average scores for each reviewer, and
perform appropriate statistical analyses.
3.5 Histology The methodology outlined below describes general procedures for
histological fixation, sample preparation, sectioning, and histomor-
phometric quantification. Due to the wide variety of staining pro-
cedures, the experimenter is advised to consult relevant literature
specific to the staining procedures needed to visualize desired tissue
components (see Note 13).
3.5.1 Paraffin Embedding 1. After euthanasia, collect operative limbs and carefully remove
excessive soft tissue and muscle. Some soft tissue should be left
to maintain the structural integrity of the periosteum and the
fractured bone.
2. Fix bones in 10% neutral buffered formalin (20:1 fixative to
tissue ratio) at 4 C for 1 week.
3. Rinse fixed bones in PBS for 30 min on an orbital shaker.
4. Place bones in labeled embedding cassettes and submerge in a
beaker containing a sufficient volume of a decalcification agent
(e.g., 14% EDTA) (see Note 14). Place submerged bones on an
orbital shaker at room temperature. Change decalcification
solution every 2–3 days, for 2–3 weeks. Decalcification is com-
plete when bone is pliable and can be confirmed by X-ray.
5. Rinse decalcified bones twice by submerging in 1 PBS for
30 min with agitation.
6. To dehydrate bones, place cassettes in four separate changes of
70% ethanol and 95% ethanol for 20 min each. Followed by one
solution of 100% ethanol for 30 min, then overnight in 100%
ethanol with gentle shaking. The next day submerge the bones
in xylene for 10 min each for a total of 4 changes.
7. The bones are then placed in heated xylene:paraffin (1:1 ratio;
56 C) for 3 h, then allowed to solidify at room temperature
overnight. The following morning the xylene:paraffin contain-
ing the bones is melted, then the bones are placed in three
separate heated liquid paraffin (paraffin 1, 2, 3) solutions for
2 h each at 56 C.
8. Remove bones from cassettes and carefully retrieve the intra-
medullary pin from each fractured femur.
9. Turn-on embedding station to heat paraffin tank to 56 C and
to cool the cold plate.
10. Embed bones, with the femoral head pointing up, in hot
paraffin wax positioned in the center of a metallic mold on an
embedding station and allow the blocks to cool.
11. Remove any air bubbles that form in the center of the
wax mold.
12. Allow samples to cool on the embedding station for 30 min to
1 h before removing them from the metal molds and storing at
4 C or 20 C.
3.5.2 Paraffin Sectioning 1. Label all positively charged slides with specimen identification
(e.g., animal ID, genotype) and section number with a
chemical-resistant marker or pencil.
2. Turn on the hot water bath at least 30 min before sectioning.
3. Place a sharp microtome blade in the microtome.
4. Remove paraffin blocks from the refrigerator or 20 C
freezer and place on ice with paraffin side facing down toward
the ice.
5. Carefully position each block in the microtome and coarsely
section (15–20 μm) until the tissue is exposed.
6. Place the block on ice for 10–20 min.
7. Section the block at a setting of 5–7 μm until the mid-sagittal
plane is reached (see Note 15).
8. Carefully place sections on surface of the water in the water
bath to stretch the sections.
9. Allow paraffin sections to stand for 1–2 min in the water bath,
before carefully placing onto slides.
10. Repeat this procedure until the sample is fully sectioned or
until desired number of slides have been collected.
11. Place slides on a slide drying rack and allow slides to dry
overnight at room temperature.
3.5.3 Histological/IHC 1. Bake slides in an oven set at 56 C for 15–30 min or until
Staining paraffin is melted.
2. Rehydrate slides (generally 10 min in xylene, 5 min in xylene,
1 min in 100% ethanol, 1 min in 100% ethanol, 1 min in 95%
ethanol, 1 min in 95% ethanol, 5 min in deionized water).
3. Conduct histological staining/IHC procedure (differs depend-
ing on experiment: consult online literature) on rehydrated
slides (see Notes 16–18).
4. Dehydrate slides (generally 1 min in 95% ethanol, 1 min in
100% ethanol, 1 min in 100% ethanol, 1 min in xylene, 1 min in
xylene) (see Note 19).
5. Add appropriate mounting medium, carefully apply cover slip
and remove any bubbles by gently applying pressure to the
cover-slipped slide (see Note 20).
6. Allow slides to dry at room temperature overnight.
3.5.4 Histomorphometric 1. Visualize slides under a microscope using an appropriate

Quantification magnification.
2. Use software such as ImageJ, ImagePro, and Osteomeasure for
image analyses. Different analyses may require different soft-
ware (consult chosen software documentation for additional
details). Mineralized tissue, fibrous tissue, cartilage matrix can
be quantified, and normalized to the total callus area. The total
number of cells and cells expressing certain factors can be
counted and their total normalized to the area of analyses (see
Note 21).
3. Analyze data using appropriate statistical approaches.
3.6 Gene Expression Special care must be taken to ensure that solutions and work area
are not contaminated with RNases/DNases, which can compro-
mise the validity of the experiment. It is recommended to spray and
wipe all laboratory benches, pipettes, and pipette tip boxes with
RNase Away prior to starting the experiment.
3.6.1 RNA Isolation from 1. Isolate fracture calluses following euthanasia by removing all
Whole Fracture Callus surrounding soft tissue and muscle, removing the pin, and
surrounding bone from the intact fracture callus. Wrap the
callus in labeled aluminum foil and flash freeze in liquid nitro-
gen. Store bones at 80 C until ready for further processing.
2. Homogenize calluses into a fine powder using either a tissue
homogenizer or a mortar and pestle, maintained at 80 C.
3. Place homogenized samples in a nuclease-free microcentrifuge
tube and add an appropriate volume of TRIzol. Incubate for
5 min at room temperature.
4. Add 0.2 mL of chloroform per 1 mL TRIzol reagent to the
samples and shake tubes vigorously by hand for 15 s.
5. Incubate samples at room temperature for 3 min and subse-
quently centrifuge them at 12,000 g for 15 min at 4 C.
6. Carefully isolate the upper clear aqueous phase and pipette it
into a new tube, taking care not to disturb the white interphase
or lower pink aqueous phase.
7. Add 0.5 mL of 100% isopropanol per 1 mL TRIzol used, mix
by pipetting, and incubate at room temperature for 10 min.
Centrifuge the mixture at 12,000 g for 10 min at 4 C.
8. Remove the supernatant from the tube, leaving only the RNA
pellet, which is washed in 1 mL 75% ethanol per 1 mL
TRIzol used.
9. Briefly vortex the mixture and centrifuge at 7500 g for 5 min
at 4 C.
10. Carefully aspirate the ethanol wash using a pipette and air-dry
the pellet for 10–15 min. It is important to prevent complete
drying of the pellet.
11. Depending on the size of the RNA pellet, resuspend with
appropriate volume of RNase-free water (approximately
10–30 μL).
12. Incubate the resuspended pellet in a heating block set at 65 C
for 5 min (or 55 C for 10 min).
13. Proceed with reverse transcription or store isolated RNA at
80 C for downstream applications.
3.6.2 RNA Isolation from 1. Cut fresh sections (7–10 μm) from paraffin-embedded frac-
Histological Sections tured femurs.
2. Carefully place sections on surface of the water in the water
bath to stretch the sections.
3. Allow paraffin ribbon sections to stand for 1–2 min in the bath,
before carefully placing onto slides.
4. Repeat this procedure until a total of 5–7 sections have been
collected.
5. Allow sections to dry on a drying rack that has been cleaned
with RNase-Away.
6. Spray microscope stage with RNase-Away.
7. Visualize fracture callus under microscope and using a sterile
10 μL pipette tip carefully scrape the callus (see Fig. 5). The
scraped tissue will adhere to the pipette tip via electrostatic
forces.
8. Place scraped tissue into a nuclease-free microcentrifuge tube.
9. Repeat until all slides have been scraped.
Fig. 5 Histological sections of a day 14 fracture callus before and after scraping the callus (red outline)
10. Deparaffinize the collected scraped tissue using an appropriate

deparaffinization reagent (i.e., xylene) according to the RNA
isolation from FFPE kit manufacturer’s instructions.
11. Isolate RNA using an RNA isolation from FFPE kit according
to manufacturer’s instructions.
12. Proceed with reverse transcription or store isolated RNA at
80 C for downstream applications.
3.6.3 Reverse 1. Remove RNA from 80 C and thaw on ice.

Transcription 2. Determine RNA concentration and purity using a Nanodrop or
UV spectrometer (see Note 22). Repeat for each sample.
3. Reverse transcribe the same quantity of RNA for each sample
(typically 1 μg) according to the cDNA kit. cDNA obtained
from reverse transcription can be diluted and stored at 20 C.
3.6.4 Quantification 1. Prepare a master mix for each gene of interest and
of Gene Expression housekeeping gene, containing forward primer, reverse primer,
and SYBR green mix (if SYBR green is used). Vortex tubes to
mix, and centrifuge briefly (see Note 23).
2. First add an appropriate volume of master mix to each well of
the PCR plate. Then add the cDNA obtained from each callus
to the corresponding wells of the PCR plate. There should be
at least two, but ideally three replicates per cDNA and gene of
interest. It is also recommended to include a no-template
control (NTC) which only contains master mix and no cDNA.
3. Seal plate with plastic cover tightly and centrifuge briefly.
4. Place plate into the RT-qPCR machine. Add labels to the
software and run PCR using the optimized conditions for the
sample of interest (see Note 24).
5. Record CT values and enter them into an excel spreadsheet for
data analysis.
6. Perform data analysis for obtained CT values, relating expres-
sion of genes of interest to housekeeping genes (see Note 25).
4 Notes
1. All metallic equipment and the surgical tray should be cleaned

with enzymatic detergent prior to surgery. All metallic equip-
ment, gauze, and drapes should be wrapped in surgical drapes
and autoclaved prior to surgery. Needles should be opened in a
sterile fashion over the opened sterilized surgical tray.
2. It is important to monitor body weight during healing and to
determine the dose of analgesic to administer.
3. It is recommended to use mice that are older than 10 weeks of

age for fracture healing experiments.
4. Analgesics can also be administered postoperatively, but any
drug administration needs to be preapproved in the animal
protocol. Buprenorphine-HCL (Buprenex®) is an alternative
analgesic to buprenorphine SR but requires daily intramuscular
injections during the initial 3-day postoperative monitoring
period. NSAIDs should not be used as routine analgesics
because they can negatively influence fracture healing [6].
5. Before performing this procedure, it is suggested that the
surgeon practice inserting the K-wire.
6. An alternative to the method described here involves creation
of a medial parapatellar incision to expose the femoral condyles
prior to boring the hole in the medullary cavity. This would
require suturing and administration of antibiotics.
7. It is recommended to scan ex vivo samples with the following
parameters: 6 μm resolution, 70 kVp tube voltage, and tube
current 114 mA, with 200 ms integration time. Scan para-
meters can change depending on the particular experiment.
8. There are several alternatives to potting specimens with
PMMA. Common examples include low-temperature Field’s
metal and Bondo®. Fields metal is melted using a double
burner and poured into the metal molds, which then solidifies
in approximately 5 min. Bondo® (polyester resin sold in kits)
forms moldable putty when mixed with a hardener, which
subsequently sets and hardens into a solid geometry.
9. PMMA is hardened through an exothermic reaction wherein a
liquid methylmethycrylate monomer (should not exceed three
times the volume of the potted specimen region) is added to a
polymer powder. Bones should be kept in chilled saline-soaked
gauze when potting to preserve the biomechanical properties
of the bone.
10. The time for polymerization of PMMA is variable and depends
on the ratio of PMMA powder to liquid.
11. Researchers should ensure that group associations remain
coded throughout the evaluation process and that they do
not introduce any external bias.
12. It is best to have either orthopedic surgeons or experienced
orthopedic researchers who are familiar with the scoring scale
participate as observers for radiographic evaluation.
13. Processing of bones including decalcification can influence
staining and other downstream applications. When possible,
all bones belonging to a single experiment should be processed
at the same time to minimize variability within experiments.
14. Many decalcification agents are available, and the most appro-
priate one is dependent on the desired downstream applica-
tion. 10%–14% EDTA is a slower decalcifier but works well for
most histological stains used to assess fracture healing. Acid-
decalcifiers work quicker but should not be used if there is an
interest in completing enzymatic staining like TRAP.
15. Occasionally the embedded bones contain residual mineral. If
this is observed, the block should be surface decalcified by
removing the block from the microtome and placing it face
down in an appropriate decalcification solution (e.g., 14%
EDTA) for 30–60 min. This will only remove mineral near
the surface, so additional surface decalcification may be
required as the paraffin block is further sectioned.
16. Different staining procedures require different materials and
dyes. To determine if additional materials are needed and to
familiarize the researcher with each specific staining procedure,
it is recommended to consult online literature before begin-
ning an experiment. Tissue processing can also affect tissue
processing. The staining time for each stain must be optimized
for each experiment.
17. The tissue of interest will determine the type of histological
staining used. Hematoxylin and Eosin is useful for gross mor-
phology and identifying callus bone, while Safranin O/Fast
Green and Alcian Blue/Hematoxylin staining is useful for
identifying cartilage.
18. IHC primary antibody dilutions should be verified in the
researcher’s lab with positive and negative controls prior to
any IHC staining.
19. Dehydration of stained slides should not be completed for
certain stains such as TRAP because it can weaken staining
intensity. Slides with these stains should be coverslipped using
an aqueous mounting medium and sealed using fingernail
polish by painting around the edge of the coverslip.
20. If the applied pressure does not remove air bubbles from under
the coverslip, soak the slide in xylene until the coverslip falls off.
Add mounting medium to the exposed tissue and apply a new
coverslip.
21. There are different techniques for quantifying the number of
cells after IHC staining. Some studies set a baseline level of
staining intensity and count positive cells above that level.
Other studies count all cells with moderate staining and above.
22. Ideally RNA should have a 260/280 ratio of 1.8–2, with a
260/280 ratio of 2 indicating pure RNA. An abnormal
260/280 ratio suggests contamination with ethanol, phenol,
guanidine, or other reagents used to isolate RNA. Very low

concentrations of RNA (<10 ng/μL) may also give an abnor-
mal 260/280 ratio.
23. All primers should be validated prior to interpretation and
publication of data.
24. RT-qPCR conditions should be optimized and adhere to the
MIQUE guidelines [7].
25. There are multiple ways of quantifying gene expression from
CT values, including the ΔΔCT method and standard curve
method. The method used is at the discretion of the researcher.
Acknowledgements
This work was supported by NIH grant R01AG064464 to H.D.
References
1. Bonnarens F, Einhorn TA (1984) Production of 5. Marturano JE, Cleveland BC, Byrne MA,
a standard closed fracture in laboratory animal O’Connell SL, Wixted JJ, Billiar KL (2008) An
bone. J Orthop Res 2(1):97–101. https://doi. improved murine femur fracture device for bone
org/10.1002/jor.1100020115 healing studies. J Biomech 41(6):1222–1228.
2. Clifton KB, Paglia DN, Soung DY, Wolf JM, https://doi.org/10.1016/j.jbiomech.2008.01.
Moss IL, Drissi H (2014) Effects of Wnt5a hap- 029
loinsufficiency on bone repair. J Orthop Trauma 6. Cottrell J, O’Connor JP (2010) Effect of
28(8):e191–e197. https://doi.org/10.1097/ non-steroidal anti-inflammatory drugs on bone
bot.0000000000000041 healing. Pharmaceuticals (Basel) 3
3. Walia B, Lingenheld E, Duong L, Sanjay A, (5):1668–1693. https://doi.org/10.3390/
Drissi H (2018) A novel role for cathepsin K in ph3051668
periosteal osteoclast precursors during fracture 7. Bustin SA, Benes V, Garson JA, Hellemans J,
repair. Ann N Y Acad Sci 1415(1):57–68. Huggett J, Kubista M, Mueller R, Nolan T,
https://doi.org/10.1111/nyas.13629 Pfaffl MW, Shipley GL, Vandesompele J,
4. Hussein AI, Mancini C, Lybrand KE, Cooke Wittwer CT (2009) The MIQE guidelines: min-
ME, Matheny HE, Hogue BL, Tornetta P 3rd, imum information for publication of quantita-
Gerstenfeld LC (2018) Serum proteomic assess- tive real-time PCR experiments. Clin Chem 55
ment of the progression of fracture healing. J (4):611–622. https://doi.org/10.1373/
Orthop Res 36(4):1153–1163. https://doi. clinchem.2008.112797
org/10.1002/jor.23754
Chapter 14
Mouse Models of Osteoarthritis: Surgical Model

of Post-traumatic Osteoarthritis Induced by Destabilization
of the Medial Meniscus
Kirsty L. Culley, Purva Singh, Samantha Lessard, Mengying Wang,
Brennan Rourke, Mary B. Goldring, and Miguel Otero
Abstract
The surgical model of destabilization of the medial meniscus (DMM) has become a gold standard for
studying the onset and progression of post-traumatic osteoarthritis (OA). The DMM model mimics clinical
meniscal injury, a known predisposing factor for the development of human OA, and permits the study of
structural and biological changes over the course of the disease. In addition, when applied to genetically
modified or engineered mouse models, this surgical procedure permits dissection of the relative contribu-
tion of a given gene to OA initiation and/or progression. This chapter describes the requirements for the
surgical induction of OA in mouse models, and provides guidelines and tools for the subsequent histologi-
cal, immunohistochemical, and molecular analyses. Methods for the assessment of the contributions of
selected genes in genetically modified strains are also provided.
Key words Surgical model, Histology, Immunohistochemistry, RNA and DNA extraction
1 Introduction
Osteoarthritis (OA) is a “whole joint” disorder involving all joint

tissues, with progressive cartilage erosion as the major pathological
indicator leading to joint replacement surgery. Post-traumatic oste-
oarthritis (PTOA) represents a subset of OA in which the end stage
of the disease may be very similar to idiopathic OA, but the initial
causes, stages of development and progression, patient popula-
tions, and potential approaches to therapy are distinct. Biomechan-
ical instability of the joint is one of the known risk factors in the
pathogenesis of OA, and is the prevalent factor involved in the
development of PTOA, in particular. Over the years we have gained
understanding of the molecular mechanisms driving the
Kirsty L. Culley, Purva Singh, Mary B. Goldring, and Miguel Otero contributed equally to this work.
223
224 Kirsty L. Culley et al.
destruction of cartilage and other joint tissues in OA, based on

analyses of gene and protein expression in clinical material and in
cell culture models derived from human tissues. However, we
currently have no disease-modifying OA drug (DMOAD) partly
because observations in human joint tissues in situ can be made
only in retrieved surgical or postmortem specimens. Therefore,
following human patients longitudinally during the development
of OA disease is not possible. Furthermore, biomarkers in synovial
fluid, serum, and urine, as well as MRI, have yet to be fully char-
acterized for early diagnosis or therapeutic outcome in cohorts,
let alone in individual patients.
No OA animal model is entirely predictive of human responses,
and it still remains unclear how well any of the available models
resemble idiopathic OA in the aging human population. Spontane-
ous OA models such as the Str/ort mouse [1] are available and
widely used. However, the long-time course of disease develop-
ment over several months makes the spontaneous models less
attractive for examining the effects of therapies. A noninvasive tibial
loading model has recently been developed and represents an
attractive alternative for replicating OA in the mouse because
both cartilage and bone changes occur in a defined manner both
spatially and temporally [2, 3]. Mouse models with the same muta-
tions as in human chondrodysplasias develop OA with age due to
abnormal composition and structure of joint tissues, such as articu-
lar cartilage, and associated biomechanical instability. The molecu-
lar phenotypes resemble those reported in the profiling studies of
human cartilage [4, 5], in which both catabolic and anabolic gene
signatures have been identified.
Surgical models of PTOA, in which the ACL or other knee
ligaments are transected in different animal species, reflect many
aspects of PTOA in humans and have several advantages over
spontaneous models, including faster disease onset, decreased
genetic drift, and better reproducibility. Surgical instability models
in dog, guinea pig, rabbit, rat, sheep, and goat all are used widely to
replicate aspects of human OA disease. Due to the high cost of
maintaining larger species and the availability of genetically mod-
ified strains, surgical PTOA models in mice are preferred for pre-
liminary screening and for determining the in vivo influence of
knockout or transgenic overexpression of a given gene on the
initiation and progression of disease throughout a time course of
several weeks [6, 7]. Indeed, genetically modified mice and differ-
ent mouse strains are employed to uncover mechanisms associated
with risk factors such as biomechanical instability, injury, inflamma-
tion, obesity, and genetic mutations and permit gene profiling over
the time course of OA initiation and progression [8–12]. Although
aging does not inevitably lead to OA, age-related responses in joint
tissues of different mouse strains are accelerated in PTOA models
DMM Mouse Model of Post-traumatic Osteoarthritis 225
[11]. Common to all mouse models of OA are certain molecular

pathways, particularly those controlling expression and activation
of proteolytic enzymes, which determine initiation and progression
of cartilage damage. Together, the findings to date suggest that the
OA signature may be disease-specific and unrelated to aging. These
findings therefore lend credence to the possibility of identifying
gene signatures in at-risk populations, including those susceptible
to PTOA, prior to the onset of overt OA.
The surgical model of destabilization of the medial meniscus
(DMM) has become a gold standard in the field. DMM surgery was
used to demonstrate the importance of the key aggrecan- and
collagen-degrading enzymes in cartilage destruction in mice defi-
cient in Adamts5 [13, 14] and Mmp13 [15]. The DMM surgery
was also used in studies that demonstrated the contribution of
different kinases and transcription factors to cartilage degradation
[16–19]. Other studies relied in the DMM model to evaluate
OA-associated pain mechanisms [20] and to test pharmacologic
intervention to target OA pain [21]. More recently, the DMM was
combined with a nonsurgical tibial compression model to dissect
the impact of mechanical loading to the progression of PTOA
[22]. Reduced severity of OA and in some cases increased regener-
ation were demonstrated in PTOA mouse or rat models treated
with a syndecan-4 antibody [23], a Hedgehog signaling inhibitor
[24], recombinant human PTH(1–34) (teriparatide) [25], an
aggrecanase inhibitor [26], and kartogenin, a small molecule tar-
geting the chondrogenic program [27]. Glasson [6] highlighted
the importance of murine genetic background in PTOA models
while screening several mouse strains in the DMM model. The
129/SvEv mice are most susceptible to surgical OA induction,
whereas the least susceptible strain tested are DBA/1, which is
highly susceptible to autoimmune or inflammatory arthritis. The
C57BL/6 strain, used most frequently for generation of knockout
and transgenic mice, either alone or on a mixed background with
129/SvDv strains, has intermediate susceptibility, indicating its
utility in determining whether the chosen genetic modification is
expected to enhance or attenuate cartilage loss following challenge.
More recent studies using a population genetics approach to gen-
erate recombinant inbred mouse strains have also highlighted the
concept that strain-related differences in mice may reflect OA sus-
ceptibility differences in humans. These studies comparing strains
derived from the MRL/MpJ superhealer mouse with different
healing capacities [28, 29] showed inverse correlation between
cartilage healing and OA development in a PTOA model.
In this chapter, we provide detailed methodology and guide-
lines to conduct the DMM surgical model, with special emphasis on
the histological, immunohistochemical, and molecular analyses for
evaluation of the impact on the development and progression
of OA.
2 Materials
2.1 Conditional 1. Tamoxifen powder.

Deletion or Induction 2. Sunflower seed oil from Helianthus annuus.
of Transgene
3. Ethanol.
Expression
in Genetically 4. ½ cc Tuberculin syringe with 27-gauge needle ½-in needle.
Modified Mice 5. 2 cc Lactated Ringer Injection USP solution.
2.1.1 Tamoxifen 6. DietGel Recovery, Purified Dietary Supplement for Laboratory
Treatment for Deletion Rodents (Fisher and Son, NJ).
of Floxed Alleles
2.1.2 Doxycycline 1. Doxycycline hyclate powder.

Treatment for Control 2. Autoclaved water.
of Tetracycline-Regulated
Promoters
2.2 Anesthesia 1. 2 cc Lactated Ringer Injection USP solution.

Induction 2. 10 or 20 mL Luer-Lok tip syringe.
and Maintenance
3. 25-gauge Needle ⅝ .
4. Ketamine at 30 mg/kg of body weight (Ketaset, 100 mg/mL).
5. Xylazine at 5 mg/kg of body weight (Rompun, 20 mg/mL).
6. Acetylpromazine at 1 mg/kg body weight (Acepromazine,
10 mg/mL).
7. Sterile water, injectable.
8. ½ cc Tuberculin syringe with 27-gauge needle ½-in needle.
9. Buprenorphine at 0.5 mg/kg of body weight (Buprenex,
0.3 mg/mL).
10. ½ cc Tuberculin syringe with 27-gauge needle ½-in needle.
2.3 Preparation 1. Hazard Technology Golden A5 Electric Hair Clippers attached

of Surgical Site to a containment vacuum with HEPA filter (Oster Professional,
McMinnville, TN).
2. Surgical scrub brush/sponge.
3. Nail cleaner with 4% chlorhexidine gluconate
(BD Company, NJ).
4. 70% Isopropyl alcohol.
5. Gauze sponges.
2.4 Surgical All surgical tools must be sterilized prior to surgery using a PRI
Reagents MUS Sterilizer (LBR Scientific Inc., East Rutherford, NJ).
and Equipment
1. Zeiss surgical microscope, Super Lux 40.
2. Surgical blades; size 11 and size 15.

3. Q-tips.
4. Mini dissecting scissors (World Precision Instruments,
Sarasota, FL).
5. Student Vannas scissors (World Precision Instruments,
Sarasota, FL).
6. Dumont tweezers #5B, 45 degrees angled (World Precision
Instruments, Sarasota, FL).
7. Dumont tweezers #5, straight (World Precision Instruments,
Sarasota, FL).
8. Tyrell Hook (World Precision Instruments, Sarasota, FL).
9. Micro Castroviejo Needle Holder (World Precision Instru-
ments, Sarasota, FL).
10. 2 Feather Scalpel Handles #3 (Electron Microscopy Sciences,
Hatfield, PA).
11. Coated Vicryl Suture (8–0 (0.4 metric) 1200 (30 cm) TG140–8
6.5 mm 3/8c) (Ethicon Inc., Somerville, NJ).
12. VETClose Surgical Adhesive containing formulated cyanoac-
rylate (Henry Schein Animal Health, Dublin, Ohio).
13. 0.9% Sodium Chloride irrigation USP Solution.
14. Powdered Sterile Surgical Gloves.
15. Converters, Royal Silk Sterile Surgical Gown, Level 3.
16. Sterile Gauze Sponges.
2.5 Mouse Housing Housing of mice at 5 to 6 mice per cage in large cages, according to
NIH and IACUC guidelines, at least 2 weeks prior to surgery is
recommended to ensure that the mice are accustomed to one
another and will maintain a level of activity that will induce OA
postoperatively.
1. Thoren Ventilated Rack (Thoren Caging Systems Inc.,
Hazelton, PA).
2. Thoren weaning cages (model #2, polycarbonate, dimensions:
L12.125 W12.125 H5.625, Thoren Caging Systems Inc.,
Hazelton, PA).
3. Polished stainless steel wire for mice cages (Thoren Caging
Systems Inc., Hazelton, PA).
4. Nestpacks, Betachip, 100gr Bedding (Fisher and Son, NJ).
5. LabDiet 5053 irradiated, PicoLab Rodent Diet 20 (Fisher and
Son, NJ).
6. DietGel Recovery, Purified Dietary Supplement for Laboratory
Rodents (Fisher and Son, NJ).
7. Mouse Tunnels (BioServ Frenchtown, NJ).
2.6 Sample Fixation, 1. 10 Phosphate-buffered solution (PBS).

Decalcification, 2. 4% Paraformaldehyde (PFA): Heat 400 mL of 1 PBS to 60 C
and Processing and add 20 g of PFA while stirring. Add 500 μL of 2 N NaOH
2.6.1 Tissue Fixation
and continue to stir until the PFA goes into solution. Add
50 mL of 1 PBS, and add dH2O water to final volume of
500 mL; bring to pH 7.4 with hydrochloric acid. Filter the
solution through a 0.45 μM filter. Aliquot and freeze at 20 C
for long-term storage.
3. 2 N Sodium hydroxide: Dissolve 4 g of NaOH pellets in 40 mL
dH2O; make to a final volume of 50 mL.
4. Hydrochloric acid.
5. 0.45 μM Filter.
6. 15 mL Falcon tubes.
7. Rocker.
2.6.2 Tissue 1. 20% Sodium citrate dihydrate: Dissolve 100 g of sodium citrate
Decalcification dihydrate in 350 mL of 1 PBS; add dH2O to a final volume of
500 mL.
2. 45% Formic acid: To prepare 500 mL dilute 225 mL of 95%
formic acid in 275 mL of dH2O.
3. 2 N Sodium hydroxide: Dissolve 4 g of NaOH pellets in 40 mL
dH2O; make to a final volume of 50 mL
4. 10% EDTA: Add 100 g of EDTA to 850 mL dH2O while
stirring. Adjust the final volume to 1 L with dH2O and pH to
7.4 using sodium hydroxide pellets.
5. Ammonium oxalate monohydrate (AO): Weigh 10 g of AO and
add to 150–200 mL of dH2O while stirring. Heat gently to
dissolve all the ammonium sulfate and continue to add AO
until the solution becomes saturated. Precipitates of ammo-
nium sulfate crystals should form and will indicate that the
solution is fully saturated. Allow the solution to cool to room
temperature before use.
2.6.3 Tissue Processing 1. Tissue-Tek Biopsy Uni-Cassette (Sakura, Torrance, CA).

2. Specimen Foam Pads (Electron Microscopy Sciences,
Hatfield, PA).
3. Ethanol (EtOH).
4. Xylene.
5. Paraffin.
6. Spin Tissue Processor (Model STP120) (ThermoFisher Scien-
tific Microm, Rockford, IL).
7. Tissue Embedding Center (Model EC350-1) (ThermoFisher
Scientific Microm, Rockford, IL).
8. Base Molds (Electron Microscopy Sciences, Hatfield, PA).
2.6.4 Sectioning 1. HM 355S Automatic Microtome (ThermoFisher Scientific

Microm, Rockford, IL).
2. MX35 Ultra Blades (Richard-Allan Scientific, Kalamazoo, MI).
3. Diamond White Glass microscope slides (Globe Scientific Inc.,
Paramus, NJ).
4. 37 C Oven.
5. Slide storage file.
2.7 Histological 1. Xylene.

Staining 2. Ethanols (EtOH): 70%, 80%, 90%, and 100%.
3. 10 Phosphate-buffered solution (PBS).
4. Shandon instant hematoxylin solution: Prepare the day before
staining following manufacturer’s instructions. Briefly, mix
1 bottle of part A and 1 bottle of part B in 1 L of dH2O and
let stir overnight at room temperature. Filter the hematoxylin
solution through Fisherbrand Chromatography paper at every
use. After preparation, the hematoxylin solution can be stored
at room temperature and reused up to 1 month.
5. Scott’s buffer solution: Add 10 g of magnesium sulfate and 2 g of
sodium bicarbonate to 1 L of dH2O and let stir for 30 min
before using. Prepare freshly for every staining.
6. Fast Green 0.2% Solution.
7. 1% Acetic acid solution: Add 10 mL of glacial acetic acid to 1 L
of dH2O. Prepare freshly for every use.
8. Safranin O 0.5% Solution.
9. VectaMount (Vector Laboratories Inc., Burlingame, CA).
10. SuperSlip CoverGlass 20 50 1.
12. Chromatography paper.
2.8 Immuno- IHC and IF conditions (e.g., retrieval method, or primary and
histochemistry (IHC) secondary antibody concentration) may vary and require optimiza-
and Immuno- tion. The provided protocols are guidelines, and have been opti-
fluorescence (IF) mized for the specified antibodies utilizing the reagents indicated.
1. Xylene.
2. Ethanol (EtOH).
3. 10 Phosphate-buffered solution (PBS).
4. 2 mg/mL Hyaluronidase.
5. 3% H2O2.
6. Humid Chamber (BD Falcon BioDish XL Nontreated Surface,
BD Biosciences Discovery Labware, Bedford, MA).
7. Liquid Blocker, Super Pap Pen Mini (Ted Pella Inc.,

Redding, CA).
8. Blocking solution: 1.5% normal goat blocking serum in a 1
PBS solution containing 1.5% Tween.
9. Primary Antibody.
10. Secondary Antibody (Biotinylated Ab Vectastain ABC Elite
Kits, Vector Laboratories Inc., Burlingame, CA; Abcam, Cam-
bridge, MA; or Alexa Fluor 555 Molecular Probes Cell Signal-
ing, Billerica, MA).
11. Tween.
12. AB Reagent (Vectastain ABC Elite Kits, Vector Laboratories
Inc., Burlingame, CA).
13. DAB Chromogen (DAKO, CA) or Vector NovaRED (Vector
Laboratories Inc., Burlingame, CA).
14. 0.2% Fast Green solution.
15. VectaMount (Vector Laboratories Inc., Burlingame, CA).
16. ProLong Gold antifade reagent with DAPI (Molecular Probes
by Life Technologies Corp., Eugene, OR).
17. SuperSlip CoverGlass 20 50 1.
2.9 RNA and DNA 1. Sterile phosphate-buffered saline (PBS) (Cell culture stan-
Extraction for Gene dard); ice-cold for dissecting mouse knees.
Expression and DNA 2. Dissecting scissors.
Methylation Analyses
3. Straight forceps (Hampton Research, Aliso Viejo, CA).
of Mouse Articular
Cartilage
4. Curved tipped forceps (Hampton Research, Aliso Viejo, CA).
5. Two star quality micro scissors 300 blade.
2.9.1 Isolation
of Articular Cartilage 6. Stereo zoom microscope.
7. Microscope stand with focus mount.
8. Nova 2000 fiber optic illuminator and optic guides (Nikon,
Melville, NY).
9. 2 Feather Scalpel handle # 3.
10. Surgical Blades; size 11 and size 15.
11. RNAlater (Ambion Life technologies, Grand Island, NY).
12. DNase/RNase-free Eppendorf tubes.
13. Filter pipette tips (nuclease-free).
2.9.2 Total RNA Isolation 1. TRIzol® Reagent (Invitrogen, Grand Island, NY).
from Cartilage Using 2. QIAshredder (Qiagen Inc., Valencia, CA).
a Modified mirVana™
3. Chloroform:Isoamyl alcohol (IAA) 24:1.
Protocol
4. Phenol:chloroform:IAA, 125:24:1.
5. Molecular-grade ethanol.
6. MirVana miRNA Isolation Kit, without phenol (Ambion Life
Technologies, Grand Island, NY).
7. DNA-free Kit (Ambion Life Technologies, Grand Island, NY).
8. 3 M Sodium acetate pH 5.5, nuclease-free.
9. UltraPure Glycogen, nuclease-free.
10. Nuclease-free water.
11. Handheld homogenizer (VWR Pellet Mixer) (VWR Interna-
tional, Radnor, PA).
12. Nuclease-free pestles (Argos Technologies, Elgin, IL).
13. NanoDrop 2000 (ThermoFisher Scientific Inc., Rockford, IL).
14. Bioanalyzer (Model 2100) (Agilent Technologies, Santa
Clara, CA).
2.9.3 Total RNA isolation 1. TRIzol® Reagent (Invitrogen, Grand Island, NY).
from Cartilage Using 2. QIAshredder (Qiagen Inc., Valencia, CA).
a Modified Rneasy® Mini
3. Chloroform:Isoamyl alcohol (IAA) 24:1 (Sigma Aldrich,
Protocol
St. Louis, MO).
4. Molecular-grade ethanol (American Bioanalytical Inc.,
Natick, MA).
5. Rneasy mini RNA Isolation kit (Qiagen Inc., Valencia, CA).
6. RNase-Free DNase set (Qiagen Inc., Valencia, CA).
7. TissueLyser II (Qiagen Inc., Valencia, CA).
8. 2.0 mL Eppendorf safe-Lock tubes (Eppendorf,
Hauppauge, NY).
9. Stainless steel beads, 5 mm (Qiagen Inc., Valencia, CA).
10. TissueLyser single bead dispenser, 5 mm (Qiagen Inc.,
Valencia, CA).
Clara, CA).
2.9.4 DNA Isolation from 1. Molecular-grade ethanol (American Bioanalytical Inc.,

Cartilage Using a Modified Natick, MA).
Gentra® Puregene® DNA 2. Molecular-grade isopropanol (American Bioanalytical Inc.,
Isolation Protocol Natick, MA).
3. Nuclease-free water (Qiagen Inc., Valencia, CA).

4. Gentra Puregene DNA isolation kit (Qiagen Inc.,
Valencia, CA).
5. UltraPure Glycogen, nuclease-free (Invitrogen Life Technolo-
gies, Grand Island, NY).
Clara, CA).
3 Methods
3.1 Conditional Although conventional global and tissue-specific knockout models

Deletion or Induction have been used in combination with the DMM surgery to uncover
of Transgene the contributions of aggrecanases [13], collagenases [15], and
Expression transcription factors [16] to cartilage degradation, gene modifica-
in Genetically tions frequently present developmental alterations that add con-
Modified Mice founding factors and extra complexity. Thus, inducible models, in
which a gene can be knocked out in a specific tissue at a desired time
3.1.1 Tamoxifen point, have been developed. To determine if conditional ablation of
Treatment of Mice a gene in cartilage affects the onset and/or progression of PTOA,
for Conditional Gene mice containing the floxed alleles of the gene of interest are crossed
Ablation with mice harboring cartilage-specific inducible Cre recombinase
transgenes, such as the minimal Col2a1 promoter transgenes
(Col2a1-Cre-ER [30] or Col2a1-Cre-ERT2 [31] or the targeted
Agc1 (aggrecan)-CreERT2 knockin allele [17, 32, 33]). These
inducible Cre strains contain a transgene that expresses a modified
form of Cre recombinase, which is controlled by a mutated version
of the mouse estrogen receptor ligand-binding domain, which does
not bind natural estrogen at physiological concentrations, but
instead binds the estrogen-derivative tamoxifen. Therefore, admin-
istration of tamoxifen at a desired time point allows elucidation of
the role of a gene at different stages of embryonic development,
after birth, or in adult life. Described below is one method of
tamoxifen administration for Cre-recombinase-mediated deletion
of floxed genes in adult mice prior to DMM surgery.
3.1.2 Tamoxifen 1. Dissolve 3 mg of tamoxifen in 100 μL ethanol and vortex for

Administration by 5 min.
Intraperitoneal Injection 2. Add 900 μL of sunflower oil and incubate at 37 C to improve
solubility; it should be completely dissolved in approximately
1 h with periodic vortexing to speed up the process.
3. Aliquot the tamoxifen for storage. Store at 4 C for 2 weeks

(maximum) or at 20 C for months.
4. Administer medications intraperitoneally into the right lower
abdominal quadrant, with the animal’s anterior body tilted
down, via 0.5-cc tuberculin syringe with a 27-gauge ½-in
needle. To facilitate access to food and water, recovery gel
cups and food pellets are placed on the floor of the cage with
the mice.
5. One day after a tamoxifen injection, administer 2 cc of lactated
Ringer’s solution intraperitoneally to each mouse (see Note 1).
3.1.3 Doxycycline The tetracycline-controlled transcriptional activation Tet-Off sys-

Treatment tem [34] is widely used to achieve expression of transgenes con-
for Tetracycline-Inducible trolled by tissue-specific promoters. The Tet-Off system requires a
Transgene Expression responder construct (containing a tetracycline-responsive element,
TRE) that controls the expression of the transgene of interest and
an activator construct (with a Tet-controlled transcriptional activa-
tor, tTA). Binding of tTA to the TRE induces transcription of a
specific gene downstream of TRE. In the Tet-Off system, when
doxycycline is present it binds to tTA to prevent it from binding to
the TRE. However, when doxycycline is absent, tTA binds to the
TRE and activates tissue-specific transcription of the gene of inter-
est localized downstream of TRE. To study the effects of the
inducible overexpression of a given gene, use murine strains that
express tTA under the control of cartilage-specific promoters, such
as Col2a1 [35] or Comp [36]. The following steps describe work
with ComptTA-TRE transgenic mice, as described [16].
3.1.4 Doxycycline 1. Prepare doxycycline at a concentration of 1 mg/mL in drinking

Administration (See water.
Note 2) 2. Administer doxycycline orally, ad libitum, to female
pregnant mice.
3. Change water and doxycycline weekly.
4. Continue the treatment during the entire pregnancy period,
and until weaning at 1 month after birth.
5. At 1 month after birth, remove doxycycline from the drinking
water to induce transgene expression. Mice will be ready for
surgery at 12 weeks of age.
3.2 Surgical The following steps describe DMM surgeries (see Note 3) per-
Resection of Mouse formed in 12-week-old male mice, as described previously by
Knee Joints Glasson et al. [37]. Age- and sex-dependent changes in OA severity,
including associated gene expression changes, have been reported
and should be considered when designing experiments
[11, 38]. Briefly, unilateral joint instability is induced by microsur-
gical transection of the medial meniscotibial ligament (MMTL),
which anchors the medial meniscus to the tibial plateau (see Note
4). DMM surgery is completed in the right knee, leaving the left
knee as a non-operated control or a sham-operated control, in
which the meniscotibial ligament is localized but not transected
(see Note 5).
1. Administer to mice subcutaneously 2 cc of lactated Ringer
solution prior to anesthesia.
2. Administer medications by intraperitoneal injection into the
right lower abdominal quadrant, with the anterior body tilted
down, via a ½ cc tuberculin syringe with a 27-gauge ½-in
needle. This single dose provides 15–20 minof surgical anes-
thesia. If necessary, anesthesia may be prolonged by adminis-
tration of isoflurane gas via a nose cone.
3. Use electrical hair clippers to remove hair from the surgical site.
Prepare the site by scrubbing twice with a 4% chlorhexidine
surgical brush/sponge, and then wiping with 70% isopropyl
alcohol.
4. Perform arthrotomy to expose the femoral tibial joint. Make a
longitudinal incision of approximately 5 mm over the distal
patellar tendon to the proximal tibial plateau.
5. Incise the joint capsule immediately medial to the patellar
tendon with a size 15 blade, and gently lift the tendon with
45 degree Dumont tweezers to allow access of the Tyrell hook
by technician. Use the Tyrell hook to move the tendon to one
side to allow access to the joint compartment.
6. Optional (see Note 6): Perform blunt dissection of the fat pad
over the intercondylar area using 45 degree Dumont tweezer
to expose the medial meniscotibial ligament (MMTL). Control
mild hemorrhaging from the fat pad by applying pressure with
Q-tips.
7. Identify the MMTL running from the cranial horn of the
medial meniscus laterally onto the anterior tibial plateau.
Take care to identify and avoid the lateral meniscotibial liga-
ment (LMTL), which is posterior and has fibers running in a
similar direction.
8. Section the MMTL with a size 11 blade, with the blade
directed proximo-laterally to destabilize the medial meniscus.
Use the 45-degree Dumont tweezer to check if the MMTL is
fully transected and the medial meniscus destabilized.
9. Suture close the joint capsule and subcutaneous layer with
Coated Vicryl Suture (8–0) and close the skin by application
of VETClose Surgical Adhesive.
10. Immediately after anesthetic recovery administer an initial dose
of buprenorphine (Buprenex 0.3 mg/mL) at 0.05 mg/kg
subcutaneously. Place mice in cages that are maintained on
warming blankets and continue to monitor until the mice are

awake. Administer buprenorphine at 0.05 mg/kg subcutane-
ously every 8–12 h (or more frequently if needed) for a dura-
tion of 48 h postoperatively [39]. Analgesic support may be
extended for certain procedures as deemed by the IACUC
and/or veterinary input. Additional analgesia may be given at
the discretion of the veterinary staff. Routinely after short
procedures, mice are ambulatory and appetent. Animals that
are inappetent, nonambulatory, or manifest any other sign of
illness (porphyria, poor hair coat, weight loss, etc.) are exam-
ined and appropriate therapy is administered (e.g., antibiotics,
fluid therapy, additional analgesic, warming pad or euthanasia if
necessary) (see Note 7).
11. Place an unopened nest pack containing bedding (see Note 8)
and a mouse tunnel (to promote activity) (see Note 9) in the
cage and allow mice to move in their cages ad lib. For the first
48 postoperative hours, to facilitate access to food and water,
recovery gel cups and food pellets are placed on the floor of the
cage. Animals are assessed by veterinary staff on the second
postoperative day. Prophylactic or postoperative antibiotics are
not administered routinely for short procedures performed
under aseptic conditions.
12. Optional: Assess the effects of drugs or other reagents on the
course of OA development and progression post-surgery by
comparing the agent versus the vehicle in different groups of
DMM-operated animals (the number of animals required per
treatment group for histological analysis is determined by
power analysis). For a single-dosing regimen, administer to
separate knees the agent or vehicle (e.g., PBS) at 1 or
2 weeks post-surgery (pre-onset OA), or if one injection is
not sufficient, dosing at 1–3 times per week over a subsequent
4-week period may be attempted. Additionally drugs can be
delivered/administered systemically via methods such as intro-
duction into the diet or via osmotic pumps implanted
subcutaneously.
3.3 Histological The animals are sacrificed (according to IACUC guidelines) (see
Assessment of OA Note 10) at appropriate time points post-DMM surgery and the
Pathology knee joints collected for histological assessment to determine the
effects of time or genotype on experimental OA. The pathology is
assessed using a modified Mankin scoring system recommended by
OARSI [40] based on Safranin O-stained histological sections.
Meniscus, subchondral bone, synovial changes, and osteophyte
formation may also be examined in the same sections to evaluate
the overall condition of the joints. Guidelines and examples for
sample processing in order to complete histological evaluation are
provided in the following sections.
Fig. 1 Trimming mouse knee joints for embedding in paraffin. Once the leg has been dissected (a) it is
important to remove as much muscle (M) as possible (b and c) to prevent the knee from bending during
processing. The tibia (T) and femur (F) should be cut to approximately 0.5 in with the patellar tendon
(PT) located in the center of the specimen
3.3.1 Fixation 1. Immediately after sacrifice dissect knees for histological analysis
and remove skin and excess muscle with dissection scissors.
Trim the femur and tibia so that they are ~0.5-in in length
(see Fig. 1), then place the dissected knees in individual falcon
tubes or biopsy cassettes for tissue fixation in 4% PFA. It is
advisable that the volume of PFA (or other fixative) is at least
15–20 the volume of tissue.
2. Fix samples for the desired length of time depending on your
antibody/staining to be completed downstream (duration of
fixation is based on previous standardized histology or immu-
nohistochemistry (IHC)/immunofluorescence (IF) and can be
as short as 6 h or as long as overnight). During fixation, place
samples on a rocker or shaker.
3. Following fixation, wash the samples with dH2O for 1 h at
room temperature on a shaker, changing the dH2O every
15 min.
3.3.2 Decalcification There are several available methods for decalcification and, while
the appropriate method should be selected based on the
subsequent histological/immunohistochemical analyses, it should
remain consistent within each experimental model studied.
Depending on the method selected, the time required for decalcifi-
cation and subsequent retrieval method for immunostaining will
change. In this section we will detail two methods for decalcifica-
tion (EDTA and formic acid:sodium citrate), both of which are
suitable for reliable Safranin O/Fast green staining and scoring.
Other available methods include 5–10% nitric acid or 10% HCl.
Sodium Citrate/Formic Acid Decalcification

1. Mix 250 mL of 45% formic acid with 250 mL of 20% sodium
citrate dihydrate; add decalcification solution to sample. The
recommended volume of decalcification solution is at least
15–20 the volume of tissue.
2. Incubate samples at room temperature on a rocker for
2–3 days.
3. Assess samples to determine if they are decalcified after
2–3 days (see the decalcification test below using a solution of
saturated ammonium oxalate).
4. Change the 45% formic acid/20% sodium citrate dihydrate
fixation solution for fresh solution every 3 days. The decalcifi-
cation should be complete within approximately 5–7 days.
10% EDTA Decalcification

1. Add 10% EDTA, pH 7.4 to samples; the recommended volume
of decalcification solution is at least 15–20 the volume of
tissue.
2. Incubate samples on a rocker at 4 C for 7–14 days, changing
the solution two times per week.
Decalcification Test
1. Take 1 mL of the decalcification solution from the samples that
you are decalcifying and add 5 mL of saturated AO solution.
Mix well, incubate for 30 min at room temperature, and check
for a white precipitate by holding against a black background.
2. If a white precipitate forms, remove the decalcification solution
from the samples and add fresh decalcification solution. Incu-
bate the samples for 2–3 more days at room temperature on a
shaker. Decalcification is complete only if no white precipitate
is formed.
3. Once decalcification is complete, wash samples in dH2O for
6 h. Change the water every 30 min or place samples (in biopsy
cassettes) under a gentle flow of clean running water.
4. After washing, place the samples in 70% ethanol for temporary

storage at 4 C until ready for embedding (long-term storage at
4 C is not recommended) (see Note 11).
3.3.3 Processing 1. Place samples in Tissue-Tek Biopsy Uni-Cassettes. Use two

and Embedding specimen foam pads to stabilize the joint in the cassette if the
mouse knee is too small (see Fig. 2 for orientation of knee for
coronal sections and Fig. 3 for sagittal sections).
Fig. 2 Orientation of mouse knee joints for paraffin-embedded coronal sections.

(a) Coronal view of the trimmed knee with the patellar tendon (PT) is easily
visualized. (b) The specimen is placed in a metallic mold ready for paraffin
embedding. For coronal sectioning, orient the femur (F) facing upward and the
tibia (T) downward in relation to the mold, using forceps to ensure that the
patellar tendon is facing up and centered within the mold
Fig. 3 Orientation of mouse knee joints for paraffin-embedded sagittal sections.

(a) Sagittal view of the trimmed knee with the patellar tendon (PT) is easily
visualized. (b) The specimen is placed in a metallic mold ready for paraffin
embedding. For sagittal sectioning, orient the femur (F) and tibia (T) so they are
facing left in relation to the mold, using forceps to ensure that the patellar tendon
is facing to the left and the sample is centered in the mold
2. Process samples using a tissue processor with the following

guidelines: 70% ETOH (7 h or overnight), followed by 80%
EtOH (1 h), 90% EtOH (1 h), 100% EtOH (3 each for 1 h),
Xylene (2 each for 1.5 h), and paraffin (2 each for 2 h).
3. Embed samples in fresh paraffin in tissue molds using a tissue-
embedding station. Allow paraffin to set, remove samples from
molds, and store long-term at 4 C.
3.3.4 Sectioning 1. Cut serial coronal sections of 6 μm throughout the whole

embedded mouse knee joint using a microtome (see Note 12).
2. Mount three sections per Diamond White Glass microscope
slides.
3. Allow slides to dry overnight in a 37 C oven to remove any
water that may be trapped under sections.
4. Stain every fifth slide (every 90 μm) with Safranin O for OA
histological scoring, leaving intervening slides for immunohis-
tochemistry or immunofluorescence.
Optional: Prior to completing staining protocols, paraffin-
embedded sections can be placed on a slide heating block at
60 C overnight.
3.3.5 Histological 1. Deparaffinize in Xylene 2 for 8 min each (in the fume hood).
Staining 2. Rehydrate in an ethanol series: 100% EtOH (2, 5 min each);
95% EtOH (5 min); 85% EtOH (4 min); and 70% EtOH
(4 min).
3. Incubate slides in filtered hematoxylin for 30 s.
4. Rinse slides in tap water 3 times for 5 min each until water is
clear.
5. Incubate in Scott’s Buffer for 2 min.
6. Rinse slides in tap water 3 times for 5 min each.
7. Incubate in 0.2% Fast green for 4 min.
8. Rinse quickly in 1% acetic acid solution (dip 3 times).
9. Rinse quickly in dH2O.
10. Stain slides in 0.5% Safranin O for 5 min.
11. Rinse slides in 95% EtOH (dip 3 times).
12. Rinse slides in 100% EtOH (dip 3 times).
13. Incubate slides in 100% EtOH for approximately 2 min. If the
ethanol is still pink after 2 min, place the slides in a fresh
ethanol bath for another minute.
14. Incubate slides in Xylene for 3 min followed by fresh Xylene for
10 min (in the fume hood).
15. Mount sections with Vectamount medium and SuperSlip
CoverGlass.
Fig. 4 Representative photomicrograph of Safranin O-stained coronal section (a) and schematic (b) detailing
the anatomical structures of the knee. Three quadrants are visualized: medial femoral condyle (MFC), medial
tibial plateau (MTP), and lateral tibial plateau (LTP). The growth plate (GP) is located at the proximal end of tibia
(T) and fibula (F). Skeletal muscle (SM) can be seen surrounding the knee. The medial meniscus (M) can be
identified between the MTP and MFC
3.3.6 Histological A well-established histological scoring system, which utilizes Safra-

Scoring nin O-stained sections (see Fig. 4 for a representative section iden-
tifying joint structures), serves as the primary outcome measure to
determine the rate and extent of OA in the DMM model. Due to
the thin nature of mouse articular cartilage, a modified Mankin
histological grading scale recommended in the OARSI Histopa-
thology Atlas by Glasson et al. [40] is used (see Table 1).
1. Prepare 12–15 sections (approximately 90 μm apart) spanning
the knee joint for each experimental group. A minimum of
10 animals per group per time point (see Subheading 3.3.7
for further details).
2. Score the medial tibial plateau and medial femoral condyle of all
samples, since damage observed post-DMM surgery is located
primarily on the medial side of the joint (see Fig. 5 for grading
of four quadrants of the knee joint).
3. Repeat scoring by a minimum of two individuals.
4. Represent data as SUM Score (where the summed OA score of
a minimum total of 10 slides is graphically represented for each
Table 1
Mouse histological scoring system recommended by OARSI [40]
Grade Osteoarthritis damage

0 Normal
0.5 Loss of safranin O without structural changes
1 Small fibrillations without loss of cartilage
2 Fibrillation to the layer immediately below the superficial layer and some loss of surface lamina
3 Fibrillation/erosion to the calcified cartilage extending to <25% of the width of articular
cartilage
4 Fibrillation/erosion to the calcified cartilage extending to 25–50% of the width of articular
cartilage
5 Fibrillation/erosion to the calcified cartilage extending to 50–75% of the width of articular
cartilage
6 Fibrillation/erosion to the calcified cartilage extending to >75% of the width of articular
cartilage
Fig. 5 Schematic (a) and Safranin O-stained coronal section (b). Representation
of the four quadrants of the knee, which can be graded post-DMM surgery,
including the medial tibial plateau (MTP), medial femoral head (MFH), lateral
tibial plateau (LTP), and lateral femoral head (LFH). The majority of cartilage
damage will be observed on the medial side of the joint
mouse/knee) or MAX score (where the maximum score for

each mouse/knee is represented graphically) (see Fig. 6 for
representative sections of each OA grade).
3.3.7 Statistical Analysis 1. Perform power analysis for the number of mice required per
group. Using a 2-point difference as the definition of a statisti-
cal difference, group sample sizes of 7 in each group achieve
85.8% power to detect a 2-point difference in the modified
mouse score after adjusting for nonparametric Mann-Whitney
U test and setting significance to 0.017 (to adjust for multiple
comparisons). To account for attrition due to surgery or
anesthesia-related deaths, the total number per group is
adjusted to 10.
2. For histological scoring, Mann-Whitney U tests and Kruskal-
Wallis with Dunn’s post-analysis (Prism® Graph-Pad) are
needed as the nonparametric equivalents for the independent
samples t-test and one-way ANOVA (see Note 13).
3.3.8 Osteophyte Scoring The progressive erosion of the cartilage observed post-DMM sur-
gery is accompanied by osteophyte formation and development.
The DMM model is a valuable tool to assess the contribution of
certain genes to the formation and development of osteophytes
during OA. Indeed, Loeser and colleagues [41] described in detail
the relationship between osteophyte development in OA and genes
involved in morphogenesis, differentiation, and development. A
histological scoring system was developed to score both osteophyte
size and maturity, as described by Little et al. [15] (see Tables 2 and
3). This system complements the modified Mankin scoring system
recommended by OARSI [40] for grading cartilage degradation,
and follows the same scoring guidelines for statistical analysis.
Briefly, osteophyte scoring is performed on the same slides/sec-
tions used to assess cartilage degradation and obtain an OA score
for each knee. The osteophyte size is highly dependent on cartilage
destruction, with joints exhibiting a high OA score often also
having large mature osteophytes (see Fig. 7). In addition, osteo-
phytes are often localized close to areas of cartilage degradation,
and thus are predominately located on the medial side of the tibial
plateau. Initially, osteophytes are composed of cartilage (see Fig. 7a,
e), and then transition to a combination of both cartilage and bone
(b) before they progress to a boney appearance as the osteophyte
matures (see Fig. 7b, f).
3.4 Immuno- Localization of protein targets of interest, including MMP-13, or

histochemistry the presence of type II collagen cleavage epitopes (C1,2C) are
examined by immunohistochemistry (IHC) using commercially
available antibodies (e.g., from Abcam, or Ibex http://www.
ibex.ca) or in-house generated antibodies, if necessary (see
Fig. 6 Safranin O-stained coronal sections taken from the knee representing each OA grade of the modified
Mankin scoring system recommended by OARSI [40]. The grade presented represents the damage observed
on the medial tibial plateau. Graphs can be constructed based on scoring of multiple knee joints using Table 1
Table 2
Assessment of osteophyte size in mouse knee joints
Osteophyte size Features

0 None
1 Small (approx. Same thickness as the adjacent cartilage)
2 Medium (approx. 1–3 times the thickness of the adjacent cartilage)
3 Large (approx. >3 times the thickness of the adjacent cartilage)
Table 3
Assessment of osteophyte maturity in mouse knee joints
Osteophyte maturity Features

0 None
1 Predominantly cartilage
2 Mix of cartilage and bone
3 Predominantly bone
Note 14). Depending upon the selected antibody or the tissue

processing (fixation, decalcification method, etc.), the optimal anti-
gen retrieval method, the primary and secondary antibody concen-
tration, or the antibody blocking solution will vary, and therefore
each IHC protocol should be carefully optimized. Below is
provided an example of an IHC protocol using the C1,2C antibody
(IBEX #50–1035) on formalin-fixed, paraffin-embedded knee sec-
tions in mice post-DMM.
3.4.1 Immunoperoxidase 1. Deparaffinize sections in xylene, 2 times for 10 min each.

Staining 2. Rehydrate gradually through a series of graded ethanol con-
centrations: 100% (2 times, 3 min each); 95% (2 min); 85%
(1 min); 70% (1 min); 50% (1 min). Finally, wash in 1 PBS
(2 times, 5 min each).
3. Perform antigen retrieval treatment with 2 mg/mL hyaluroni-
dase for 30 min at 37 C in a humid chamber (see Note 15).
4. Wash in 1 PBS buffer (3 each for 5 min).
5. Incubate for 15 min at RT in 3% hydrogen peroxide in dH2O
to quench endogenous peroxidase activity.
6. Wash in 1 PBS buffer (2 each of 5 min).
7. Incubate in 1.5% normal goat blocking solution in a humid
chamber for at least 1 h at room temperature (see Note 16).
Fig. 7 Representative photomicrograph images of Safranin O-stained coronal sections taken from the knee
(wild type mouse at 8 weeks post-DMM surgery) with yellow outlines indicating osteophytes. The composition
of the osteophyte is dependent on its maturity: Early osteophytes are composed mainly of cartilage (a), and
then transition to a combination of both cartilage and bone (b) before they develop into mature osteophytes
that consist mainly of bone (c). The size and maturity of the osteophyte often correlates with the severity of OA
found in the joint post surgery, with small osteophytes consisting mainly of cartilage observed in joints with a
low histological score (d), and mature osteophytes consisting mainly of bone observed in highly damaged
joints (e). Osteophyte formation is modest in wild type mice subjected to DMM surgery, but may be enhanced
by certain gene modifications. Graphs can be constructed based on scoring of multiple knee joints using
Tables 2 and 3
8. Remove blocking solution from slides and incubate with pri-

mary antibody C1,2C at 1:200 dilution or the corresponding
isotype-matched negative control IgG (see Note 17) in the
normal goat blocking buffer with 1.5% Tween overnight at
4 C in a humid chamber.
9. Wash in 1 PBS buffer (2 each of 5 min). Optional: Add
0.5% Tween to the 1 PBS buffer from this step if the back-
ground is expected to be high.
10. Incubate with biotin-conjugated secondary antibody for 1 h at
room temperature in a humid chamber.
12. Incubate with avidin biotin enzyme reagent for 30 min at room
chamber in a humid chamber.
14. Incubate with DAB chromogen or Vector NovaRed until

desired stain intensity develops. Comparative slides (e.g., wild
type versus knockout and positive and negative controls)
should be monitored to determine the proper
development time.
15. Wash sections in dH2O for 2 min to stop the reaction.
16. The provided C1,2C IHC protocol does not include a coun-
terstaining step. If required, counterstaining methods include:
(a) Fast green 0.2% solution: stain for 2 min, followed by
quick washing in 95% EtOH.
(b) Hematoxylin (filtered): Stain for 30 s, followed by imme-
diate washing with several changes of dH2O, then with
95% EtOH.
17. Rinse quickly in 100% EtOH, and then perform a second wash
in 100% EtOH for 2 min.
18. Incubate in Xylene (2 for 5 min each).
19. Mount slides with 2 drops of Vectamount medium, cover with
a glass coverslip, and observe by light microscopy.
3.4.2 Immuno- The immunofluorescence (IF) protocol is commonly used when

fluorescence Staining there is a need to detect multiple cellular targets by simultaneous
Protocol labeling; a mix of primary antibodies is followed by a combination
of secondary antibodies conjugated to diverse fluorochromes emit-
ting light at different wavelengths. The following protocol is
intended as a general guide for immunofluorescence on paraffin-
embedded sections:
1. Follow steps 1–4 of the IHC protocol (see Subheading 3.4.1).
2. Incubate with normal blocking serum for 1 h at room temper-
ature in a humid chamber (see Note 16).
3. Incubate with optimized primary antibody concentration in a
humid chamber overnight at 4 C (see Note 17). If using a
primary antibody conjugated with a fluorochrome, omit step 5.
4. Wash in 1 PBS buffer with 0.5% Tween 20 (2 each of
5 min).
5. Incubate with fluorochrome-conjugated secondary antibody
for 2 h at room temperature in a humid chamber (make sure
from this step onward the samples are shielded from light).
Secondary antibodies are conjugated to a wide range of fluor-
ochromes to suit the users’ needs (e.g., IgG-FITC, IgG–TR,
IgG-CY3, IgG-Cy5, Alexa Fluor or DyLight, Chromeo, Sure-
Light, etc.). As cartilage autofluoresces in the green spectrum,
it is recommended to use fluorochromes that do not fall within
this spectrum.
6. Wash in 1 PBS buffer (2 each of 5 min) in a slide container

covered with aluminum foil.
7. Mount slides with an antifade mounting medium (e.g., Pro-
Long Gold antifade reagent with DAPI) and follow drying
instructions of the manufacturer.
8. Visualize staining with a fluorescence microscope.
3.5 Cartilage For nucleic acid isolation, cartilage is dissected from the femoral
Microdissection heads and tibial plateaus of the knee and homogenized in TRIzol
for RNA and DNA (Invitrogen). Below (see Subheadings 3.6 and 3.7) we describe
Extraction for Gene methods to isolate total RNA using mirVana™ miRNA isolation
Expression and DNA kits (Ambion), for work with mRNA and miRNA, or the RNeasy®
Methylation Analyses Mini (Qiagen) kits, following the manufacturer’s instructions with
additional modifications. Following these methods, an average of
50–80 ng of high-quality total RNA is obtained from the articular
cartilage isolated from one knee joint. This should serve as a guide
for the number of mice required to achieve a sufficient amount of
RNA for gene expression analyses.
For RNA isolation, it is critical to place each leg in RNAlater
(steps 1–5) as early as possible to obtain good RNA integrity; thus,
one person completes steps 1–7 and a second person completes
steps 8–12.
Person 1
1. Sacrifice mouse and immediately remove hind legs.
2. Use dissection scissors to quickly remove as much soft tissue
from the legs as possible.
3. Place the legs in separate Petri dishes and cover with ice-cold
1 PBS.
4. While working on one leg, keep the other leg in 1 PBS on ice,
complete steps 5–7 on the other leg.
5. Dissect remaining muscle and tendon with a scalpel and size
15 blade under the microscope. Remove dissected soft tissues
from the Petri dish, as these tissues may be sources of RNases.
6. Transfer cleaned leg to a fresh Petri dish and wash with ice-cold
1 PBS.
7. Transfer the washed leg into a fresh small Petri dish and cover
with RNAlater. Keep the leg on ice in RNAlater until ready to
complete steps 8–13.
Person 2: Keep legs in RNAlater at all times during the following
steps in the Petri dish placed under the dissection microscope, use a
scalpel with size 11 blade to separate the tibia and femur and expose
the articular surfaces.
8. Remove any remaining soft tissue or tendon surrounding the

articular surfaces with a scalpel or small dissection scissors. Be
careful not to damage the cartilage.
9. Place the bones in a fresh Petri dish and cover with RNA later,
as small amounts of soft tissue will make dissected cartilage
difficult to decipher.
10. While the leg is bathed in RNAlater use a scalpel with size
11 blade to carve the cartilage from the articular surfaces of
the tibia and femur.
11. Place the harvested cartilage in RNase-free Eppendorf tube
containing RNAlater (enough to completely cover the carti-
lage sample). Samples can be pooled in order to obtain a
higher RNA yield.
12. Either:
(a) Proceed immediately with the RNA isolation using the
mirVana kit and the modified protocol below; or.
(b) Place the cartilage in RNAlater overnight at 4 C.
Remove excess RNA later in the morning, snap freeze,
and store at 80 C until ready to isolate RNA.
RNAlater is not required for DNA isolation from microdis-
sected articular cartilage, which is performed as described above
using clean ice-cold 1 PBS in all steps (1–10). After microdissec-
tion (step 11), place articular cartilage in nuclease-free Eppendorf
tube and, step 12(a) proceed immediately with DNA isolation as
detailed in Subheading 3.8; or step 12(b) snap-freeze tissues in
liquid Nitrogen and store (liquid Nitrogen) until use.
High-quality DNA suitable for epigenomic analyses is obtained
using Gentra® Puregene® DNA isolation kits (Qiagen), following
the manufacturer’s instructions with additional modifications
(described in Subheading 3.8). Following this method, an average
of 200–450 ng of high-quality DNA is obtained from microdis-
sected articular cartilage from one knee joint. This should serve as a
guide for the number of mice required to achieve a sufficient
amount of DNA for analyses.
3.6 RNA Isolation The following procedure is performed according to the manufac-
from Cartilage Using turer’s instructions (Ambion) (see Note 18), except that additional
a Modified mirVana phenol:chloroform steps have been introduced to help improve the
miRNA Isolation Kit 260/280 values of the RNA isolated. To obtain good RNA integ-
(Ambion) Protocol rity (RIN) values, all reagents, pipet tips, and Eppendorf tubes must
be certified nuclease-free. In addition, isolation should be com-
pleted on a clean work space treated with RNAse Zap (Ambion).
1. Place the Eppendorf tube containing cartilage sample on ice
and add 500 μL of TRIzol to the tube. Homogenize the
cartilage sample in TRIzol using a 1.5-mL RNAse-free pestle
and handheld homogenizer, for approximately 10 min (or until

confident that the cartilage is fully homogenized). Each knee
can be homogenized separately, with the option of pooling
multiple samples onto the mirVana filter column at Step 18.
2. Add 200 μL TRIzol (to total volume of 700 μL) and vortex
to mix.
3. Transfer TRIzol with disrupted tissue to QIAshredder to fur-
ther homogenize and disrupt cells: centrifuge for 2 min at
16,000 g.
4. Transfer sample to a fresh nuclease-free Eppendorf tube, taking
care not to take or disrupt the pellet of matrix that will be
visible.
5. Add 300 μL TRIzol and vortex to mix (total volume: 1 mL).
6. Add 200 μL chloroform:IAA per 1 mL TRIzol and vortex for
15 s.
7. Incubate samples on ice for 5 min.
8. Centrifuge at 13,000 g for 10 min at 4 C.
9. Remove the aqueous phase, taking care not to disrupt the
interface, and transfer to new RNase-free tube (make note of
volume).
10. Add 1 volume of phenol:chloroform:IAA to the aqueous phase
and vortex for 15 s.
11. Incubate on ice for 10 min.
13. Remove aqueous phase and transfer to new RNase-free tube
(make note of volume).
14. Add 1 volume of chloroform:IAA to the aqueous phase, vortex
for 15 s to mix.
16. Centrifuge at 10,000 g, 10 min, 4 C.
17. Remove aqueous phase and transfer to new RNase-free tube
18. Add 1.25 aqueous volume of nuclease-free 100% ethanol
(room temperature) to the aqueous phase.
From this step onward, reagents from the mirVana miRNA
isolation kit will be used. These steps will isolate total RNA, includ-
ing miRNA, within the same fraction. However, the kit also pro-
vides the option for isolating miRNA and mRNA in different
fractions (refer to the user’s manual included with the kit for
further information).
1. For each sample, place a mirVana filter cartridge into a collec-

tion tube. Add the lysate/ethanol mix to the filter, 700 μL at a
time, and centrifuge (~15 s, 10,000 g, room temperature).
Discard flow-through. (For sample volumes larger than 700 μL
repeat in successive applications to the same filter).
2. Apply 700 μL of miRNA wash solution 1 (mirVana Kit) to
column and centrifuge (~5 to 10 s, 10,000 g, room temper-
ature). Discard flow-through.
3. Apply 500 μL of miRNA wash solution 2/3 (mirVana Kit) to
column and centrifuge (~5–10 s, 10,000 g, RT). Discard
flow-through.
4. Repeat step 19.
5. Spin column to remove residual fluid (1 min, 10,000 g, room
temperature).
6. Transfer filter cartridge to fresh collection tube.
7. Add 100–50 μL nuclease-free water (room temperature) to the
filter and leave for 1 min.
8. Centrifuge to elute (~30 s, 16,000 g) (see Note 19).
9. Apply the eluted 50 μL to the filter cartridge and spin again
(~30 s, 16,000 g) (see Note 20).
10. To a 50 μL-volume of RNA sample, add 0.1 volume (5 μL) of
DNase buffer and 1 μL DNase (all reagents provided with the
DNA-free Kit from Ambion).
11. Incubate at 37 C for 25 min.
12. Add 0.1 volume (5.5 μL) of inactivating reagent, mix well, and
incubate at room temperature for 2 min (mixing occasionally,
at least 3 times).
13. Centrifuge at 10,000 g, 1.5 min, 4 C.
14. Remove the supernatant and place in a nuclease-free 1.5-
mL tube.
15. Quantify RNA using a NanoDrop spectrophotometer.
16. RNA integrity value can be assessed at this point using a
Bioanalyzer (Agilent) through a service usually provided by
the institutional genomics core.
3.6.1 Ethanol This precipitation step can be completed to concentrate the RNA if
Precipitation the concentration obtained above is not enough for required
analyses.
1. On ice, add 94 μL of ice-cold water to each 50 μL of RNA
sample (volumes should now be about 144 μL).
2. Add 16 μL of cold 3 M NaOAc, pH 5.0, buffer to each sample
to make give a concentration of around 0.3 M. Mix by pipet-
ting up and down. Final volume will be 160 μL.
3. To each sample add 2 μL of 20 μg/μL UltraPure Glycogen.

Mix by pipetting up and down (see Note 21).
4. Add 2.5 parts of ice-cold 100% ethanol to each sample—
pipette up and down to mix.
5. Place the samples at 80 C for at least 1 h (can be left
overnight).
6. Centrifuge at maximum speed, 10 min at 4 C.
7. The RNA pellet should now be visible in the Eppendorf tube.
Decant ethanol.
8. Add 750 μL of ice-cold 70% ethanol to the pellet, gently
vortex, and centrifuge at 10,000 g for 5 min.
9. The RNA pellet should now be visible. Decant ethanol. Cen-
trifuge samples quickly (approximately 5 s) to collect any
remaining ethanol in the bottom of the tube and remove it
with a filtered, nuclease-free tip.
10. Resuspend RNA in 15 μL of sterile nuclease-free water.
11. Re-assess RNA concentration using a NanoDrop
spectrophotometer.
12. Assess RIN value using a Bioanalyzer (Agilent).
3.7 RNA Isolation This modified RNA isolation protocol combines the TRIzol® RNA
from Cartilage Using extraction method with the Qiagen Rneasy® mini kit protocol. The
a Modified RNeasy protocol modifications are intended to improve RNA purity
Mini RNA Isolation Kit (A260/280), while preserving RNA integrity (RIN > 7). All
(Qiagen) Protocol reagents, pipet tips, and Eppendorf tubes must be certified
nuclease-free. In addition, isolation should be completed on a
clean workspace treated with RNAse Zap (Ambion).
1. Label 2 mL microcentrifuge tubes compatible with TissueLy-
ser and dispense 1 stainless steel bead per tube.
2. Place microdissected cartilage (see Note 22) in the 2 mL micro-
centrifuge tube along with 700 μL of Trizol and homogenize
the cartilage at 300 Hz for 2 min using the TissueLyser II. If
needed, repeat the step again once. See Note 23 for subsequent
homogenization.
(a) Alternatively, if no TissueLyser is available, homogenize
the cartilage using 1.5 mL RNAse-free pestle and hand-
held homogenizer, for approximately 5–10 min on ice, as
described in Subheading 3.6.
3. Transfer the homogenized sample in 700 μL of TRIzol to a
QIAshredder to further homogenize and disrupt cells: centri-
fuge for 2 min at 16,000 g.
4. Transfer the flow-through to a fresh nuclease-free Eppendorf

tube, taking care not to disrupt or take the pellet of matrix that
will be visible.
5. Add 300 μL TRIzol and vortex to mix (total volume: 1 mL).
6. Add 200 μL chloroform:IAA per 1 mL TRIzol and vortex for
15 s.
7. Incubate samples on ice for 5 min.
9. Remove the aqueous phase, taking care not to disrupt the
interface, and transfer to a new nuclease-free tube (make note
of volume).
10. Add 1 volume of phenol:chloroform:IAA to the aqueous phase
and vortex for 15 s.
13. Remove aqueous phase and transfer to new nuclease-free tube
14. Add 1.25 aqueous volume of 100% molecular-grade ethanol
(room temperature) to the aqueous phase (for example, add
625 ul of 100% ethanol to 500 ul of aqueous phase).
15. Mix thoroughly by inverting the tube and let it stand at room
temperature for 2 min.
From this step onward, reagents from the RNAeasy RNA iso-
lation kit will be used and procedure will be followed as per man-
ufacture’s instruction (refer to the user’s manual included with the
kit for further information).
1. Add the lysate/ethanol mix to the Qiagen mini column filter,
700 μL at a time, and centrifuge (~15 s, 10,000 g, room
temperature). Discard flow-through. (For sample volumes
larger than 700 μL repeat in successive applications to the
same filter).
2. Wash the column with 350 μL of RW1 wash buffer, by cen-
trifuging for ~15 s at 10,000 g, room temperature.
3. DNase treatment: add 80 μL of DNase mixture directly on the
column (without touching the filter or the walls of the col-
umn), incubate at room temperature for 15 min.
(a) To make DNase mixture: add 70 μL of RDD buffer and
10 μL of DNase.
4. Add 350 μL of RW1 wash buffer to wash the column, and
centrifuge for ~15 s at 10,000 g, room temperature. Discard
the flow-through.
5. Add 500 μL of RPE buffer and centrifuge for ~15 s at

10,000 g, room temperature. Discard the flow-through.
6. Repeat washing with 500 μL RPE buffer, centrifuge for ~2 min
at 10,000 g, room temperature. Discard the collection tube.
7. Transfer the column to new collection tube and centrifuge for
1 min at 10,000 g, to remove any residual RPE buffer from
the column. Discard the collection tube.
8. Transfer the column to fresh nuclease-free 1.5 mL Eppendorf
tube. Add 30 μL of RNase-free water to the filter (without
touching the filter), let the column stand for 1 min at room
temperature. Centrifuge for 1.5 min at 10,000 g, room
temperature to elute the RNA.
(a) Optional: to concentrate and/or fully elute RNA, add
back the same 30 μL of eluted nuclease-free water to the
filter and repeat the elution step 7 (see Note 20).
9. Quantify RNA using a NanoDrop spectrophotometer.
10. RNA integrity value can be assessed at this point using a
Bioanalyzer.
3.8 DNA Isolation The Gentra® Puregene® DNA isolation protocol has been modified
from Cartilage Using to increase DNA yield and purity. All reagents, pipet tips, and
a Modified Gentra Eppendorf tubes must be certified nuclease-free, and all isolation
Puregene DNA steps should be completed in a clean workspace.
Isolation Kit (Qiagen) 1. Place freshly isolated or frozen knee cartilage in 1.5 mL
Protocol nuclease-free Eppendorf tube (see Note 22 for comment
regarding required amount of starting material).
2. Homogenize the cartilage using 1.5 mL pestle and handheld
motorized homogenizer.
3. Add 300 μL of cell lysis solution to the grounded cartilage, and
heat the sample at 65 C for 30 min.
4. Add 1.5 μL of proteinase K to the sample tube, shake, and
incubate at 55 C for 1 h. If cartilage is not completely dis-
solved, continue heating at 55 C for up to 3 hrs. Cartilage
must be completely digested before moving to the next steps.
5. Add 1.5 μL of RNase A solution, mix thoroughly by inverting
the tube, and incubate at 37 C for 45 min.
6. Quickly transfer the sample to ice and allow to cool for 1 min.
7. Add 100 μL of protein precipitation solution and vortex vigor-
ously for 20 s to mix, incubate on ice for 5 min.
8. Centrifuge sample tube for 3 min at 13000 g. A visible white-
colored precipitated protein pellet is formed at the bottom of
the tube.
9. If pellet is loose, repeat the incubation for 5 min on ice and

repeat step 8.
10. Transfer the supernatant to a fresh nuclease-free 1.5 mL
Eppendorf tube.
11. Add 300 μL of ice-cold molecular-grade isopropanol to the
supernatant.
12. Add 0.5 μL of glycogen (stock concentration: 20 μg/μL) to
the isopropanol and supernatant mixture, and invert 50 times.
13. Incubate the tube at 20 C for 1.5 h. Centrifuge for 10 min at
13000 g.
14. A transparent DNA pellet will be visible at this stage. Remove
the supernatant carefully and discard without disturbing the
DNA pellet.
15. Add 300 μL of 70% ethanol and invert tube several times to
wash DNA. Centrifuge at 13000 g for 1 min.
16. Carefully discard the supernatant without disturbing the DNA
pellet (whitish, at this stage).
17. Repeat steps 17 and 18 twice, to ensure a thorough washing of
the DNA pellet.
18. Carefully discard the supernatant and remove remaining excess
ethanol by completely (and carefully) draining the tube by
inverting on Kimwipes, taking care that the pellet remains in
the tube. Avoid overdrying the DNA pellet as DNA will be
difficult to resuspend.
19. Add 50 μL of DNA hybridization buffer to the tube with the
dry DNA pellet, incubate at 65 C for 1 h to dissolve DNA.
20. If needed, incubate at room temperature (15–25 C) overnight
with gentle shaking.
21. Quantify DNA using a NanoDrop spectrophotometer, and
perform required quality control steps to ensure that the qual-
ity is sufficient for downstream analyses.
4 Notes
1. To induce Cre-recombinase-mediated deletion of the floxed

gene only in adult chondrocytes, mice receive three intraperi-
toneal injections of either tamoxifen (knockout) or vehicle
alone (wild-type control) at 2-day intervals of 2.5 mg per
10 g of body weight of mouse. However, the concentration
and dose should be chosen carefully depending on the age of
the animal, and therefore the current literature should be
reviewed. The final injection is scheduled 1 week prior to
surgery, which is usually performed at 12 weeks of age,
allowing the mice time for complete gene ablation and recovery
from tamoxifen treatment before anesthesia. To observe the
effects of gene modification on OA development, comparisons
should be done in tamoxifen- versus vehicle-treated
littermates.
2. Different doxycycline concentrations have been used success-
fully, with no reported adverse effect [35, 36], but with differ-
ences in the time required for transgene activation to occur due
to the varying amount of time required for doxycycline to clear
from the mouse system. The concentrations achieved by the
doses administered are not sufficient to inhibit collagenase
activity. Adequate controls have to be used for comparison
after the DMM surgery [36].
3. All procedures must be approved by the Institutional Animal
Care and Use Committee (IACUC). Before live animals are
used, all personnel must obtain CLAS orientation and training,
including demonstration to veterinarians the surgical proce-
dure on cadaveric mice.
4. See Glasson et al. [37] for excellent photographs and sche-
matics to guide the surgery.
5. It is recommended that one surgeon perform all surgeries
involving each genetically modified mouse strain to reduce
variability. Comparisons between wild type and knockout or
transgenic strains are completed using littermates.
Non-operated or sham controls need to be checked to ensure
that there is no nonspecific effect of the genetic modification.
6. The fat pad may also be left in place and merely transected to
allow access and visualization of the MMTL.
7. Significant postoperative pain and debility are not anticipated.
Mice are monitored postoperatively by the Veterinary Staff, and
if there is evidence of pain, suffering, or illness, analgesia
and/or other treatment will be administered at their discretion.
Particular attention will be paid to ensure that animals are
ambulating normally after the procedure. For further informa-
tion on analgesic dugs, please see reviews [39, 42].
8. Since only male mice are used in the DMM model because of
the protection by estrogen in females [37], it is necessary to
establish the following procedures to avoid aggressive behavior.
Mice are housed together prior to surgery (completed as soon
as possible post weaning if males are not from the same litter),
and the same mice are housed together post-surgery in a cage
containing some of their original dirty bedding. Administra-
tion of analgesics is continued according to the protocol. Sur-
gical mice are monitored for the first few hours after surgery
and immediately the next morning. If a dominant male is
noticed, it is immediately separated from the group. The
remaining mice in the cage are monitored daily for fighting and
any new emerging dominant male is removed. A minimum of
3 mice must remain housed together post surgery to encourage
the level of activity required to promote OA initiation, devel-
opment, and progression.
9. Immediately after surgery, place one mouse tunnel per cage to
provide enrichment and promote the level of activity required
to promote OA initiation, development, and progression.
10. For histology, mice may be sacrificed by CO2 inhalation. For
RNA extraction, if a CO2 tank is not immediately available,
cervical dislocation may be required to avoid rigor mortis
before tissues can be dissected from the joints and placed in
extraction buffer.
11. Alternatively to test if the samples are decalcified, a needle can
be passed through the bone of the sample. If no resistance is
felt, the sample can be considered decalcified.
12. During sectioning check carefully every fifth slides to ensure
the knee is correctly orientated in the paraffin. The lateral
femoral condyle will usually appear in the first sections. Refer
to Figs. 4 and 5 to gain a good concept of the proper knee
orientation. If after 15 to 20 collected slides all four quadrants
are not identified, discard the knee.
13. Previous studies by Glasson et al. [37] found that the mean
maximum histological scores at 4 weeks postoperatively for the
unoperated, sham surgery, DMM, and ACLT groups, respec-
tively, were (S.E.M.) 1.1 0.1, 1.0 0.3, 3.7 1.5, and
4.3 0.4; and at 8 weeks the mean maximum scores were
1.2 0.3,1.2 0.2, 4.1 0.3, and 5.0 0.4.
14. The conditions for immunohistochemistry and immunofluo-
rescence (e.g., antigen retrieval method, primary and second-
ary antibody concentration) may vary and require
optimization. The provided protocols are guidelines and have
been optimized for applying the specified antibodies to mouse
knee joints utilizing the reagents indicated.
15. Hyaluronidase treatment is just one of many available antigen
retrieval methods. In general, the retrieval method depends
upon the antibody selected and the tissue processing, and
therefore requires optimization for the antibody used for
detection of specific antigen epitopes in a given tissue. It is
advisable to perform optimization steps comparing conditions
without antigen retrieval with one or two antigen retrieval
methods. Retrieval methods include:
(a) Heat retrieval in a sodium citrate buffer pH 6.0 (20 min at
95 C).
(b) Treatment with hyaluronidase (2 mg/mL 30 min at
37 C).
(c) Treatment with pepsin (5 mg/mL in 0.02% HCL, for

45 min at 37 C).
(d) Treatment with 0.05% saponin solution for 30 min RT.
16. Blocking solutions are used to reduce background and dimin-
ish nonspecific staining. Many blocking methods exist and the
correct method should be used for your chosen antibody. If
normal serum is used for blocking, the correct serum should be
chosen to avoid interaction with primary and secondary anti-
bodies, and the tissue being stained. Ideally the serum chosen
should be derived from the same species in which the second-
ary antibody is raised, or from an unrelated species. Increasing
the incubation time with blocking serum can further reduce
background staining. In addition, adding non-ionic detergents
(such as Tween) can reduce nonspecific hydrophobic interac-
tions and help permeabilize the tissue to reach intracellular
epitopes.
17. Optimal antibody concentrations should be determined empir-
ically, by titration in the blocking buffer. Always incubate slides
with positive (tissue known to express your protein of interest
control or commercially available positive controls) and nega-
tive controls (isotype non-immune immunoglobulin control at
the same concentration as the primary antibody, or a tissue that
does not express the antigen).
18. The mirVana miRNA isolation kit (Ambion) can be purchased
with or without phenol. This protocol is optimized using the
mirVana kit without phenol, with the addition of TRIzol
(Invitrogen).
19. Some downstream sequencing protocols require a minimum
amount of RNA at a certain concentration (ng) per μL. The
50 μL elution volume is a smaller volume than that recom-
mended by the manufacturer, but can result in more concen-
trated RNA, which can sometimes prevent the need to
complete ethanol precipitation in order to concentrate the
sample.
20. Re-apply the eluted sample back onto the filter to increase both
yield and concentration of RNA.
21. Glycogen can interfere with some downstream sequencing
methods and should be investigated before use as a carrier.
22. Pooled microdissected cartilage from 3 knee joints yielded
optimal results (high purity and integrity, and required concen-
tration) for RNA-Seq (and RTqPCR validation) and Reduced
Representation of Bisulfite Sequencing analyses in cartilage
samples retrieved at different time points after DMM surgeries.
23. If some fragments or cartilage remain intact after the Tissue-
Lyser step, use 1.5-mL RNAse-free pestle and handheld
homogenizer as described in Subheading 3.6.
Acknowledgements
Research related to this topic was supported by National Institutes

of Health grants R21-AG049980, R01-AG022021, and
RC4-AR060546. Kirsty L. Culley, Purva Singh, Mary
B. Goldring, and Miguel Otero contributed equally to this work.
References
1. Poulet B, Ulici V, Stone TC, Pead M, and down-regulated in experimental osteoar-
Gburcik V, Constantinou E, Palmer DB, thritis. J Biol Chem 286(43):37758–37767.
Beier F, Timmons JA, Pitsillides AA (2012) https://doi.org/10.1074/jbc.M111.248039
Time-series transcriptional profiling yields 9. Yasuhara R, Ohta Y, Yuasa T, Kondo N,
new perspectives on susceptibility to murine Hoang T, Addya S, Fortina P, Pacifici M,
osteoarthritis. Arthritis Rheum 64 Iwamoto M, Enomoto-Iwamoto M (2011)
(10):3256–3266. https://doi.org/10.1002/ Roles of beta-catenin signaling in phenotypic
art.34572 expression and proliferation of articular carti-
2. Poulet B, Hamilton RW, Shefelbine S, Pitsil- lage superficial zone cells. Lab Investig 91
lides AA (2011) Characterizing a novel and (12):1739–1752. https://doi.org/10.1038/
adjustable noninvasive murine joint loading labinvest.2011.144
model. Arthritis Rheum 63(1):137–147. 10. Lodewyckx L, Cailotto F, Thysen S, Luyten FP,
https://doi.org/10.1002/art.27765 Lories RJ (2012) Tight regulation of wingless-
3. Ko FC, Dragomir C, Plumb DA, Goldring SR, type signaling in the articular cartilage - sub-
Wright TM, Goldring MB, van der Meulen chondral bone biomechanical unit: transcrip-
MC (2013) In vivo cyclic compression causes tomics in Frzb-knockout mice. Arthritis Res
cartilage degeneration and subchondral bone Ther 14(1):R16. https://doi.org/10.1186/
changes in mouse tibiae. Arthritis Rheum 65 ar3695
(6):1569–1578. https://doi.org/10.1002/ 11. Loeser RF, Olex AL, McNulty MA, Carlson
art.37906 CS, Callahan MF, Ferguson CM, Chou J,
4. Sato T, Konomi K, Yamasaki S, Aratani S, Leng X, Fetrow JS (2012) Microarray analysis
Tsuchimochi K, Yokouchi M, Masuko-Hongo- reveals age-related differences in gene expres-
K, Yagishita N, Nakamura H, Komiya S, sion during the development of osteoarthritis
Beppu M, Aoki H, Nishioka K, Nakajima T in mice. Arthritis Rheum 64(3):705–717.
(2006) Comparative analysis of gene expres- https://doi.org/10.1002/art.33388
sion profiles in intact and damaged regions of 12. Nuka S, Zhou W, Henry SP, Gendron CM,
human osteoarthritic cartilage. Arthritis Schultz JB, Shinomura T, Johnson J, Wang Y,
Rheum 54(3):808–817 Keene DR, Ramirez-Solis R, Behringer RR,
5. Aigner T, Fundel K, Saas J, Gebhard PM, Young MF, Hook M (2010) Phenotypic char-
Haag J, Weiss T, Zien A, Obermayr F, acterization of epiphycan-deficient and epiphy-
Zimmer R, Bartnik E (2006) Large-scale gene can/biglycan double-deficient mice.
expression profiling reveals major pathogenetic Osteoarthritis Cartilage 18(1):88–96.
pathways of cartilage degeneration in osteoar- https://doi.org/10.1016/j.joca.2009.11.006
thritis. Arthritis Rheum 54(11):3533–3544 13. Glasson SS, Askew R, Sheppard B, Carito B,
6. Glasson SS (2007) In vivo osteoarthritis target Blanchet T, Ma HL, Flannery CR, Peluso D,
validation utilizing genetically-modified mice. Kanki K, Yang Z, Majumdar MK, Morris EA
Curr Drug Targets 8(2):367–376 (2005) Deletion of active ADAMTS5 prevents
7. Little CB, Fosang AJ (2010) Is cartilage matrix cartilage degradation in a murine model of
breakdown an appropriate therapeutic target in osteoarthritis. Nature 434(7033):644–648
osteoarthritis--insights from studies of aggre- 14. Stanton H, Rogerson FM, East CJ, Golub SB,
can and collagen proteolysis? Curr Drug Tar- Lawlor KE, Meeker CT, Little CB, Last K,
gets 11(5):561–575. https://doi.org/10. Farmer PJ, Campbell IK, Fourie AM, Fosang
2174/138945010791011956 AJ (2005) ADAMTS5 is the major aggrecanase
8. Bernardo BC, Belluoccio D, Rowley L, Little in mouse cartilage in vivo and in vitro. Nature
CB, Hansen U, Bateman JF (2011) Cartilage 434(7033):648–652
intermediate layer protein 2 (CILP-2) is 15. Little CB, Barai A, Burkhardt D, Smith SM,
expressed in articular and meniscal cartilage Fosang AJ, Werb Z, Shah M, Thompson EW
(2009) Matrix metalloproteinase 13-deficient 23. Echtermeyer F, Bertrand J, Dreier R,

mice are resistant to osteoarthritic cartilage Meinecke I, Neugebauer K, Fuerst M, Lee YJ,
erosion but not chondrocyte hypertrophy or Song YW, Herzog C, Theilmeier G, Pap T
osteophyte development. Arthritis Rheum 60 (2009) Syndecan-4 regulates ADAMTS-5 acti-
(12):3723–3733 vation and cartilage breakdown in osteoarthri-
16. Wondimu EB, Culley KL, Quinn J, Chang J, tis. Nat Med 15(9):1072–1076. http://www.
Dragomir CL, Plumb DA, Goldring MB, nature.com/nm/journal/v15/n9/suppinfo/
Otero M (2018) Elf3 contributes to cartilage nm.1998_S1.html
degradation in vivo in a surgical model of post- 24. Lin AC, Seeto BL, Bartoszko JM, Khoury MA,
traumatic osteoarthritis. Sci Rep 8(1):6438. Whetstone H, Ho L, Hsu C, Ali AS, Alman BA
https://doi.org/10.1038/s41598-018- (2009) Modulating hedgehog signaling can
24695-3 attenuate the severity of osteoarthritis. Nat
17. Culley KL, Lessard SG, Green JD, Quinn J, Med 15(12):1421–1425
Chang J, Khilnani T, Wondimu EB, Dragomir 25. Sampson ER, Hilton MJ, Tian Y, Chen D,
CL, Marcu KB, Goldring MB, Otero M (2019) Schwarz EM, Mooney RA, Bukata SV, O’Keefe
Inducible knockout of CHUK/IKKalpha in RJ, Awad H, Puzas JE, Rosier RN, Zuscik MJ
adult chondrocytes reduces progression of car- (2011) Teriparatide as a chondroregenerative
tilage degradation in a surgical model of osteo- therapy for injury-induced osteoarthritis. Sci
arthritis. Sci Rep 9(1):8905. https://doi.org/ Transl Med 3(101):101ra193. https://doi.
10.1038/s41598-019-45334-5 org/10.1126/scitranslmed.3002214
18. Ismail HM, Miotla-Zarebska J, Troeberg L, 26. Chockalingam PS, Sun W, Rivera-Bermudez
Tang X, Stott B, Yamamoto K, Nagase H, MA, Zeng W, Dufield DR, Larsson S, Lohman-
Fosang AJ, Vincent TL, Saklatvala J (2016) der LS, Flannery CR, Glasson SS, Georgiadis
Brief report: JNK-2 controls aggrecan degra- KE, Morris EA (2011) Elevated aggrecanase
dation in murine articular cartilage and the activity in a rat model of joint injury is attenu-
development of experimental osteoarthritis. ated by an aggrecanase specific inhibitor. Oste-
Arthritis Rheumatol 68(5):1165–1171. oarthritis Cartilage 19(3):315–323. https://
https://doi.org/10.1002/art.39547 doi.org/10.1016/j.joca.2010.12.004
19. Rowe MA, Harper LR, McNulty MA, Lau AG, 27. Johnson K, Zhu S, Tremblay MS, Payette JN,
Carlson CS, Leng L, Bucala RJ, Miller RA, Wang J, Bouchez LC, Meeusen S, Althage A,
Loeser RF (2017) Reduced osteoarthritis Cho CY, Wu X, Schultz PG (2012) A stem cell-
severity in aged mice with deletion of macro- based approach to cartilage repair. Science 336
phage migration inhibitory factor. Arthritis (6082):717–721. https://doi.org/10.1126/
Rheumatol 69(2):352–361. https://doi.org/ science.1215157
10.1002/art.39844 28. Rai MF, Hashimoto S, Johnson EE, Janiszak
20. Miotla Zarebska J, Chanalaris A, Driscoll C, KL, Fitzgerald J, Heber-Katz E, Cheverud JM,
Burleigh A, Miller RE, Malfait AM, Stott B, Sandell LJ (2012) Heritability of articular car-
Vincent TL (2017) CCL2 and CCR2 regulate tilage regeneration and its association with
pain-related behaviour and early gene expres- ear-wound healing. Arthritis Rheum 64:2300.
sion in post-traumatic murine osteoarthritis https://doi.org/10.1002/art.34396
but contribute little to chondropathy. Osteoar- 29. Hashimoto S, Rai MF, Janiszak KL, Cheverud
thritis Cartilage 25(3):406–412. https://doi. JM, Sandell LJ (2012) Cartilage and bone
org/10.1016/j.joca.2016.10.008 changes during development of post-traumatic
21. Hwang SM, Feigenson M, Begun DL, Shull osteoarthritis in selected LGXSM recombinant
LC, Culley KL, Otero M, Goldring MB, Ta inbred mice. Osteoarthritis Cartilage 20
LE, Kakar S, Bradley EW, Westendorf JJ (6):562–571. https://doi.org/10.1016/j.
(2018) Phlpp inhibitors block pain and carti- joca.2012.01.022
lage degradation associated with osteoarthritis. 30. Nakamura E, Nguyen MT, Mackem S (2006)
J Orthop Res 36(5):1487–1497. https://doi. Kinetics of tamoxifen-regulated Cre activity in
org/10.1002/jor.23781 mice using a cartilage-specific CreER(T) to
22. Holyoak DT, Chlebek C, Kim MJ, Wright TM, assay temporal activity windows along the
Otero M, van der Meulen MCH (2019) proximodistal limb skeleton. Dev Dyn 235
Low-level cyclic tibial compression attenuates (9):2603–2612. https://doi.org/10.1002/
early osteoarthritis progression after joint dvdy.20892
injury in mice. Osteoarthritis Cartilage 27 31. Dao DY, Jonason JH, Zhang Y, Hsu W,
(10):1526–1536. https://doi.org/10.1016/j. Chen D, Hilton MJ, O’Keefe RJ (2012)
joca.2019.06.005
Cartilage-specific beta-catenin signaling regu- (3):1338–1346. https://doi.org/10.1016/j.

lates chondrocyte maturation, generation of ajpath.2011.05.023
ossification centers, and perichondrial bone 37. Glasson SS, Blanchet TJ, Morris EA (2007)
formation during skeletal development. J The surgical destabilization of the medial
Bone Min Res 27(8):1680–1694. https://doi. meniscus (DMM) model of osteoarthritis in
org/10.1002/jbmr.1639 the 129/SvEv mouse. Osteoarthr Cartil 15
32. Henry SP, Jang CW, Deng JM, Zhang Z, Beh- (9):1061–1069
ringer RR, de Crombrugghe B (2009) Gener- 38. Ma HL, Blanchet TJ, Peluso D, Hopkins B,
ation of aggrecan-CreERT2 knockin mice for Morris EA, Glasson SS (2007) Osteoarthritis
inducible Cre activity in adult cartilage. Genesis severity is sex dependent in a surgical mouse
47(12):805–814. https://doi.org/10.1002/ model. Osteoarthritis Cartilage 15
dvg.20564 (6):695–700. https://doi.org/10.1016/j.
33. Henry SP, Liang S, Akdemir KC, de Crom- joca.2006.11.005
brugghe B (2012) The postnatal role of Sox9 39. Flecknell PA (1996) Laboratory animal anes-
in cartilage. J Bone Min Res 27 thesia, 2nd edn. Academic Press, London
(12):2511–2525. https://doi.org/10.1002/ 40. Glasson SS, Chambers MG, Van Den Berg WB,
jbmr.1696 Little CB (2010) The OARSI histopathology
34. Gossen M, Bujard H (1992) Tight control of initiative - recommendations for histological
gene expression in mammalian cells by assessments of osteoarthritis in the mouse.
tetracycline-responsive promoters. Proc Natl Osteoarthritis Cartilage 18(Suppl 3):
Acad Sci U S A 89(12):5547–5551 S17–S23. https://doi.org/10.1016/j.joca.
35. Grover J, Roughley PJ (2006) Generation of a 2010.05.025
transgenic mouse in which Cre recombinase is 41. Loeser RF, Olex AL, McNulty MA, Carlson
expressed under control of the type II collagen CS, Callahan M, Ferguson C, Fetrow JS
promoter and doxycycline administration. (2013) Disease progression and phasic changes
Matrix Biol 25(3):158–165. https://doi.org/ in gene expression in a mouse model of osteo-
10.1016/j.matbio.2005.11.003 arthritis. PLoS One 8(1):e54633. https://doi.
36. Xu L, Polur I, Servais JM, Hsieh S, Lee PL, org/10.1371/journal.pone.0054633
Goldring MB, Li Y (2011) Intact pericellular 42. Jenkins WL (1987) Pharmacologic aspects of
matrix of articular cartilage is required for analgesic drugs in animals: an overview. J Am
unactivated discoidin domain receptor 2 in Vet Med Assoc 191(10):1231–1240
the mouse model. Am J Pathol 179
Chapter 15
Immunostaining of Skeletal Tissues

Crystal Idleburg, Madelyn R. Lorenz, Elizabeth N. DeLassus,
Erica L. Scheller, and Deborah J. Veis
Abstract
Immunostaining is the process of identifying proteins in tissue sections by incubating the sample with
antibodies specific to the protein of interest, then visualizing the bound antibody using a chromogen
(immunohistochemistry or IHC) or fluorescence (immunofluorescence or IF). Unlike in situ hybridization,
which identifies gene transcripts in cells, immunostaining identifies the products themselves and provides
information about their localization within cells (nuclear, cytoplasmic, or membrane) or extracellular
matrix. This can be particularly important in the context of bone and cartilage because they contain
many cell types as well as matrix components, each with distinct protein expression patterns. As the number
of antibodies continues to grow, this technique has become vital for research laboratories studying the
skeleton. Here, we describe a detailed protocol for antibody-based in situ analysis of bone and associated
tissues, addressing specific issues associated with staining of hard and matrix-rich tissues.
Key words Immunostaining, Immunohistochemistry, Immunofluorescence, Bone, Cartilage, Decal-

cification, Fixation, Antibodies, Antigen retrieval
1 Introduction
In both clinical and research studies, histology-based methods are

critical for describing phenotypes in patients and in experimental
organisms. There are 3–4 main steps for immunostaining:
1. Incubation with antigen-specific primary antibody.
2. Incubation with an enzyme-, biotin-, or fluorophore-
conjugated secondary antibody.
3. Detection of secondary antibody via an enzymatic reaction that
produces a colored precipitate (IHC).
4. Imaging using standard light (IHC) and/or confocal fluores-
cence microscopy (IF).
Crystal Idleburg and Madelyn R. Lorenz contributed equally to this work.
261
262 Crystal Idleburg et al.
However, the tissue collection and processing steps that come

prior to immunostaining are crucial and can greatly affect the image
quality. This is particularly true for bone and cartilage where it is
necessary to decalcify tissue while maintaining matrix components
such as proteoglycans. Therefore careful attention must be paid to
each step from tissue harvest and fixation to decalcification and
antigen retrieval [1]. Mistakes and overprocessing at any of these
steps can damage antigenic epitopes, tissue morphology, or adhe-
sion of tissue to slides, making it difficult to assess morphology and
obtain good staining.
1.1 Fixation Fixation is the process of treating tissue with solutions that preserve
gross morphology as well as molecular structures within the tissue
and should be started as soon as possible after harvest [2, 3]. Pene-
tration of fixative is determined by the size and nature of the tissue
of interest. Soft tissues and small pieces of tissue will fix faster than
larger or harder tissues. The standard fixative for paraffin embed-
ding is 10% neutral buffered formalin (NBF), while the most typical
for frozen sections is 4% paraformaldehyde (PFA). However, when
applied properly, either fixative can be used for each embedding
method. As with most fixatives, these solutions preserve tissue by
cross-linking the proteins. Therefore, tissues fixed in 10% NBF and
4% PFA usually require antigen retrieval before incubation with
primary antibody. Because of the cross-linking action, it is impor-
tant to avoid over-fixation as this can lead to excessive cross-linking,
which may mask the antigens, or dehydration, which may produce
an undulation artifact during sectioning [2]. Fixation in neutral
buffered zinc formalin, which prevents excessive cross-linking,
should also be considered for frozen sections, as it often eliminates
the need for antigen retrieval steps [4]. Although unfixed frozen
sections are useful in some situations such as reporter cell lines or
mice that express fluorescent proteins, fixation is almost always
required when performing antibody-based detection, and results
are generally better when tissue is fixed up front, rather than
dipping sections into fixative.
To ensure proper preservation when working with bone or
cartilage, it is necessary to clean away any unwanted soft tissue
such as skin and muscle. This allows for fixative penetration in a
timely manner and avoids under- or over-fixation. It also makes it
easier to orient the bone during embedding. Additionally, unde-
sired autofluorescence of blood cells can be removed by perfusing
the animal with phosphate-buffered saline (PBS) and 10% NBF or
4% PFA to rinse and fix the vasculature, followed by a post-fix of
immersion in 10% NBF or 4% PFA [5].
1.2 Decalcification Before embedding bone in paraffin or OCT, it is essential to soften

the tissue by lowering the calcium content (i.e., decalcification).
The duration of decalcification and degree of calcium ion removal
Immunostaining of Skeletal Tissues 263
are influenced by the solution used. Most commercial solutions are

acids, either mineral or organic, and soften bones quickly, but they
can easily damage the tissue and are generally not compatible with
immunostaining. Another less common method is the use of an
ion-exchange resin to exchange ammonium ions for calcium ions.
While this method yields the best morphology, especially in bone
marrow specimens, it is also the most expensive option [2]. The
most useful decalcifying method for immunostaining is treatment
with 14% ethylene diamine tetraacetic acid (EDTA) [2, 3, 6]. This
gentler chelating agent may decalcify hard tissues more slowly but is
less likely to damage tissue or affect antigenicity. Even with EDTA,
it is important to monitor and optimize the decalcification dura-
tion. Failure to do so can lead to poor morphology and weak
antibody-based staining. Lastly, if decalcification is not an option
for your desired antigen or preparation, specialized tape transfer
methods can be used to capture calcified sections of frozen or resin-
embedded tissues [7].
1.3 Antigen Retrieval Due to the cross-linking action of most fixatives, it is often neces-
sary to unmask antigens before staining [8]. The choice of retrieval
method will vary according to the antigens and antibodies used.
There are several methods of antigen retrieval but they fall into two
main categories, enzyme digestion and heat treatment. Each
retrieval method presents its own challenges and needs optimiza-
tion for different specimen types. Enzyme digestion requires preci-
sion in pH and duration of treatment because different tissues will
digest at different rates. The challenge in heat retrieval is in treating
the tissue long enough to ensure antigen retrieval without causing
it to lift off from the slide, which is a common problem when
working with cartilage and bone. For frozen sections, if retrieval
is necessary, enzymatic methods should be of primary consideration
due to the fragility of hydrated, aqueous-treated tissues. Antigenic-
ity in frozen sections can also be improved by use of a buffer
solution containing detergent (such as Tween 0.05–0.1%).
1.4 Data Analysis To accurately interpret staining, it is important to know the stan-
dard morphology and staining patterns in the tissue of interest.
Textbooks on histology, pathology, and developmental biology can
be a good resource for identifying the cells and structures. To
interpret the staining itself, the first priority is in determining
whether the signal (whether a chromogen or fluorescence) is spe-
cific or represents nonspecific background. Having both negative
and positive controls is crucial in making this determination. Nega-
tive control slides can be generated in two ways: no primary anti-
body or isotype- and species-matched immunoglobulin or serum
(which contains immunoglobulin) instead of primary antibody step
[8, 9]. While the no primary negative controls are usually accept-
able, the isotype control is the gold standard because it is possible to
see interactions between nonimmune immunoglobulins from one

species and target tissues from another. When using fluorophores,
the “no primary” control indicates native tissue autofluorescence.
Another good option for negative control comparisons is the use of
knockout tissue or cell lines with all the same reagents as test slides,
including the primary antibody. Ideally, control slides should have
no staining at all. However, some background is often unavoidable
and usually related to specific structures/cells, and thus the nega-
tive control slides must be directly compared to the test slides in
similar areas to demonstrate specificity.
Other important factors to consider are which tissues, cells, or
organelles are stained and whether the staining pattern makes
physiological sense based on known molecular pathways. If
knockout/wild-type pairs of the exact site and experimental condi-
tions are available, staining in the wild-type samples can be inter-
preted as positive. When such pairs are not in hand, positive control
slides from other sites or conditions with previously established
antigen expression are very useful here, although different tissue
types may have quite different staining patterns as well as nonspe-
cific background. Complementary techniques such as in situ RNA
hybridization, which identifies gene expression in specific cells, laser
capture microdissection with RNA analysis, or tissue fractionation
with protein or RNA analysis can also be used to verify findings.
2 Materials
2.1 IHC on Paraffin 1. Phosphate-buffered saline (PBS).

Sections 2. Citrate buffer, pH 6: Make 0.1 M stock solutions of citric acid
and trisodium citrate. To 450 mL dd water add 9 mL of citric
acid stock solution and 41 mL of sodium citrate stock solution.
The pH of this final solution should be about 6.0 0.1.
3. 10% Neutral buffered formalin (NBF).
4. 4% Paraformaldehyde (PFA) can be purchased as a 16% stock.
Making your own from powder is hazardous, and respiratory
precautions must be taken.
5. Peroxidase block (3% H2O2 in methanol): 25 mL 30% H2O2 to
a final of 250 mL in 100% methanol chilled at 20 C.
6. Positively charged slides such as Fisherbrand Superfrost Plus
(Thermo).
7. Vectastain ABC Kit (Vector Labs).
8. DAB Chromogen (Biocare).
9. 14% Free acid EDTA, pH 7.2–7.4 (EDTA Decalcification
Buffer): Mix 140 g EDTA free acid with 700 mL distilled
water. While stirring, slowly add 30 mL ammonia hydroxide
at 30-min intervals (for a total of 90 mL ammonia). Check the

pH. EDTA will not dissolve until pH is close to 7.2 If not up to
7.2, add the remaining 10 mL ammonia hydroxide dropwise to
get to pH 7.2, while constantly stirring. Add ~300 mL distilled
water for final volume of 1 L. It is critical that this solution is
made properly. If you overshoot the pH, do not attempt to
correct with HCl—just start over, as the excess chloride ions
will prevent proper EDTA chelation of ionized calcium [2].
10. Methanol.
11. Graded ethanols (30%, 50%, 70%): It is least expensive to dilute
from a purchased 70% stock, but 95% can also be used. These
concentrations do not have to be very exact. Adding about
50 or 70 mL of 70% ethanol and diluting up to 100 mL in
ddH2O is sufficient.
12. Xylene.
13. Coplin jars.
14. Humidity chamber: Any container with a lid can be lined with
damp paper towels to make one, and several companies
sell them.
15. Mounting medium, xylene compatible such as Richard-Allan
Scientific™ Mounting Medium (Thermo).
16. Proteinase K in 10 mM Tris–HCl, pH 7.4–8.0. Make a 10 mg/
mL Proteinase K (Roche) stock solution in ddH2O. 1 M Tris–
HCl pH 8.0 is commercially available. Dilute 1 M Tris–HCl
stock solution at 1:100 dilution (10 μL/mL) to make a 10 mM
Tris–HCl diluent solution. To 1 mL of this diluent solution,
add 1 μL of the Proteinase K stock solution to get a final
concentration of 10 μg/μL of Proteinase K.
17. 10% Blocking serum in PBS (see Note 1).
18. Primary antibody.
19. Secondary antibody (enzyme or biotin-conjugated).
20. Coverslips.
21. Mounting media, such as Permount (Fisher) or Cytoseal
(Richard-Allan).
2.2 IF on Frozen 1. Tris-NaCl-Tween Buffer (TNT): Combine 50 mL 1 M Tris–

Sections HCL pH 7.4, 25 mL 3 M NaCl, and 250 μL Tween-20. Bring
the solution to 500 mL total using dH2O and mix. Filter the
solution at .2 μm for best results. PBS with 0.1% Tween
20 (PBST) may also be used in place of TNT.
2. 10% Neutral buffered formalin (NBF), buffered aqueous zinc
formalin (ZF), or 4% paraformaldehyde (PFA). PFA can be
purchased as a 16% stock. Making your own PFA from powder
is hazardous, and respiratory precautions must be taken.
3. 10% Blocking serum in TNT (see Note 1).

4. Primary antibody.
5. Fluorophore-conjugated secondary antibody.
6. DAPI counterstain.
7. Positively charged slides such as Fisherbrand Superfrost Plus
(Thermo).
8. Coverslips.
9. PAP pen, such as SuperPAP with 4 mm tip (Biotium).
10. Aspirator vacuum pump.
11. Disposable freeze-safe cassettes for embedding.
12. 14% Free acid EDTA, pH 7.2 (EDTA Decalcification Buffer):
See Subheading 2.1 for more details.
13. Humidity chamber.
14. Glycerol-based aqueous mounting medium, such as
Fluoromount-G Mounting Medium (Affymetrix).
15. Clear nail polish.
3 Methods
3.1 IHC on Paraffin 1. If perfusing for rodent studies, an approved dose of anesthetic
Sections should be administered as determined by your local animal
safety committee. When the animal is fully sedated, it can be
3.1.1 Tissue Preparation
perfused with an intracardial injection of 1 M PBS, followed by
10% NBF or 4% PFA (typically 10 mL to 25 mL each). Con-
stant flow rate can be applied with a peristaltic syringe pump or
other method to improve the quality of the perfusion
fixation [5].
2. Immediately after dissection, fix bones in 10% neutral buffered
formalin or 4% paraformaldehyde for 24–72 h at room temper-
ature or 4 C. Fixative volume should be 15–20 times tissue
volume. To ensure complete penetration, tissue should be
agitated on a shaker during fixation (see Note 2).
3. Rinse tissue in PBS or distilled deionized water (ddH2O)
6 times, 15 min each.
4. Decalcify in 14% free acid EDTA, pH 7.2–7.4, with rocking,
changing solution daily (on weekdays only is OK). The number
of days required for decalcification of mouse bones is as follows
(see Note 3):
(a) Embryo > E17.5: 1–2 days.
(b) Postnatal days (P) 1–4: 3 days.
(c) P5-P10: 4–5 days.
(d) P10-P21: 7–10 days.

(e) Adults: 10–14 days.
5. Rinse tissue in PBS or ddH2O 6 times, 15 min each, to stop the
decalcification process (see Note 4).
6. Dehydrate the tissue through a series of ethanol solutions, with
rocking.
(a) 30% ethanol for 15 min.
(b) 50% ethanol for 15 min.
(c) 70% ethanol for 15 min.
7. Place tissue in the tissue processor for dehydration, clearing,
and infiltration before embedding. Typically, a 4-h processing
program works well for most machines with mouse long bones
(see Note 5).
8. When processing is complete, the bones are embedded in
paraffin. It is important to determine the plane of interest
before this point in order to orient the tissue properly.
9. Tissue sections are then cut at 5 μm using a microtome, floated
onto a 45 C water bath and placed on color frost slides (see
Note 6).
10. Slides are dried at room temperature overnight before storage
(see Note 7).
3.1.2 Antibody-Based 1. Heat slides in a 55 C oven for 1 h (see Note 8).

Staining 2. Incubate slides in Xylene 3 5 min.
3. Incubate slides in 100% ethanol 3 3 min.
8. Rinse in ddH2O 3 5 min.
9. DO NOT LET YOUR SLIDES DRY OUT AFTER THIS
POINT. Incubate slides in Peroxidase block for 10 min (see
Note 9).
10. Rinse slides once in ddH2O (2–3 dips) and then wash in PBS
3 5 min.
11. Start the antigen retrieval process by incubating slides in citrate
buffer in a covered Coplin jar at 55 C overnight (see Note 10).
12. Rinse slides in PBS 3 5 min.
13. Block endogenous biotin with avidin/biotin block according
to manufacturer instructions. We find that this is important in
skeletal tissues.
14. Incubate in 10% serum diluted in PBS for 60 min at room

temperature in humidified chamber (see Note 1).
15. Drain off serum solution. Do not rinse slides or let the slides
dry completely.
16. Incubate in primary antibody diluted in PBS with 1.5% serum
(see Note 11) overnight at 4 C or for 1 h at room temperature
in humidified chamber (see Note 12).
17. Rinse slides 3 5 min in PBS.
18. Incubate sections in biotinylated secondary antibody for
30 min, following data sheet from manufacturer for dilutions
(see Note 13).
19. Rinse slides 3 5 min in PBS.
20. Prepare ABC solution as per manufacturer’s instructions and
incubate with slides for 30 min.
21. Prepare DAB substrate. Also prepare an extra Coplin jar with
water to stop the substrate reaction (see Note 14).
22. For developing the slides you will need a light microscope. Lay
out all your reagents—the substrate solution, several Coplin
jars with water (see Note 15), and slides should all be easy to
reach. Some reactions take as little as 30 s before developing
background so there is little margin for error at this point.
23. Place your positive (+ve) and negative ( ve) control slides on
the microscope stage and add a drop of substrate solution to
each. The ideal time interval will give you the most intense
staining in your +ve control without giving any staining in the
ve. The maximum staining time should be 5 min or less. Place
the slides in the extra Coplin jars of ddH2O to stop the reac-
tion. Do not return them to jars with undeveloped slides since
this will start the substrate reaction prematurely.
24. Develop the rest of the slides one at a time at the optimal time
determined in step 23 Once a slide has developed put it in the
extra Coplin Jar with water.
25. Rinse slides well in ddH2O.
26. Counter stain the slides in Harris hematoxylin for 30 s to
1 min.
27. Wash in running tap distilled water for 10 min.
28. Dehydrate and clear through 2 changes of 95% ethanol,
3 changes of 100% ethanol, and then xylene.
29. Add coverslip with mounting medium. In a fume hood, place a
drop or thin line of mounting medium on the edge of the
coverslip on the benchtop and touch it with the edge of an
inverted slide at a 45o angle, and gently bring it down onto
coverslip. Avoid bubbles under the coverslip (see Note 16).
Wipe the excess xylene and mounting medium with gauze or

Kimwipe prior to viewing under microscope. Allow slides to
dry and xylene to evaporate in a fume hood.
3.2 IF on Frozen 1. Perfusion, fixation, decalcification, and washing are the same as
Sections Subheading 3.1.1, steps 1–5 above.
3.2.1 Tissue Preparation 2. Incubate tissue in 30% sucrose solution at 4 C for 3–5 days.
This helps to prevent formation of damaging ice crystals during
freezing/embedding. Tissue should sink to the bottom of the
container when ready for embedding.
3. Embed sucrose-permeated tissue in OCT Compound (Tissue-
Tek), paying close attention to orientation (see Note 17). Store
frozen blocks at –80 C until sectioning.
4. Tissue sections are then cut at the desired thickness (generally
from 10–100 μm) using a cryostat and placed on color frost
slides (see Note 18).
5. Store frozen slides at –80 C for long-term storage.
3.2.2 Antibody-Based NOTE: All steps should be performed in a humidified chamber.

Staining
1. Remove slides from freezer and allow to thaw in a humidifying
chamber for 10 min.
2. Draw a PAP pen barrier around tissue prior to rinsing for
maximum adherence of hydrophobic barrier to slide (see
Note 19).
3. Pipette enough PBS to create surface tension over the tissue
inside the hydrophobic barrier (generally 200 μL or more)
onto the slide to wash OCT from the tissue for 5 min. Then
aspirate the liquid.
4. Pipette PBST buffer to rinse for 5 min. Then aspirate the
liquid.
5. Incubate in 10% serum diluted in TNT for 60 min at room
temperature in humidified chamber (see Note 1). Then aspirate
the liquid.
6. Incubate in primary antibody diluted in TNT (see Notes 11
and 12). Close lid and wrap in plastic wrap, taking care not to
disturb the slides. Incubate slides at 4 C. For incubation time
suggestions, see Note 20.
7. After incubation, aspirate the antibody and rinse with PBST
3 10 min.
8. Incubate in secondary antibody diluted in TNT (see Notes 11
and 12). Close lid and wrap in plastic wrap, taking care not to
disturb the slides. Incubate slides at 4 C. For incubation time
and temperature suggestions, see Note 21.
9. After incubation, aspirate the antibody and rinse with PBST
3 10 min.
10. If the mounting media used contains DAPI, skip to step 12. If
not, add DAPI counterstain cocktail to slide and incubate for
5–7 min. Aspirate the liquid immediately to avoid overstaining.
For DAPI concentration suggestions, see Note 22.
11. Aspirate the liquid immediately to avoid overstaining. Rinse
with PBST 3 5 min.
12. Place 2–3 drops or a thin line of glycerol-based aqueous
mounting media along the edge of a coverslip. Carefully place
the slide (tissue down) on top, beginning on the edge with the
mounting media. This helps to prevent bubbles.
13. Seal slide edges with clear nail polish. This prevents the slides
from drying out and media from escaping, as aqueous media
does not fully set.
14. Keep slides at 4 C. Room temperature slides may dry out,
mold, or lose fluorescence more quickly (see Note 23).
4 Notes
1. Blocking and primary antibody incubation serum should be

from the same species as the secondary antibody (i.e., if the
secondary antibody is goat anti-rabbit then goat serum should
be used to block slides).
2. For example, 1–2 mouse femurs and/or tibiae should be
placed in a 15 mL tube with at least 10 mL fixative. Most
tissues will be properly fixed in 24 h, but large bones, such as
from rabbits, might require longer fixation. However, antige-
nicity can be reduced with longer fixation so optimization of fix
time may be needed.
3. The first time you do this, include a test bone of the same type
as you will analyze in the decalcification and use this one to
bend and compress. A fully decalcified bone should bend easily
and not break.
4. Total rinse time should be about 2 h.
5. First step in processor should be 70% ethanol. Whole bone
specimens from larger animals may require 6–8 h processing
times.
6. Slides can be checked using light, darkfield, or phase micros-
copy at this point for proper plane of section.
7. Do not skip this step and go straight to heating slides at 55 C.
The tissue is likely to fall of the slides during staining if you
do this.
8. Alternatively, slides may be baked overnight at 55 C.
9. Start with 10 min and extend if background is high on negative
controls. The 30% hydrogen peroxide can also be diluted in
PBS instead of cold methanol. In addition, there are commer-

cial peroxidase block alternatives such as Biocare 1, which may
be preferable because they have shorter incubation times.
10. Slides can also be heated to 95 C in citrate buffer for 10 min,
followed by cooling in hot buffer for 15 min. However, over-
night citrate buffer treatment is preferable to high heat because
it preserves tissue morphology better. If the high heat method
is used, make sure the buffer does not come to a full boil as this
will cause the tissue to fall off the slide. Other alternative
retrieval methods include enzymatic digestion at 37 C with
proteinase K (10 μg/mL in 10 mM Tris–HCl pH 7.4–8.0 for
20 min) or hyaluronidase (1% in PBS for 30 min). Avoid using
reagents generated in donkeys (serum or antibodies) when
using hyaluronidase since it increases background staining. In
addition, some antigens do not require retrieval. This is a step
that must be optimized for each antigen.
11. Data from the antibody manufacturer are good sources for
determining the primary and secondary antibody concentra-
tions to use. However, users may have to run serial dilution
experiments to determine the optimal concentration for spe-
cific tissue/antibody combinations.
12. Depending on the size of your tissue, you will need 50–200 μL
of antibody solution. Thick frozen sections may require up to
300 μL. Cut a piece of parafilm the same size as the slide and
float this on top to retain the antibody over the tissue, taking
care not to have any bubbles. Do not use a glass or plastic
coverslip. You do not need to use the parafilm for shorter
incubation, but make sure the tissue is covered with the solu-
tion. Do not allow the tissue dry or you will get very high
background.
13. Alternatively, incubate in secondary antibody conjugated to
horseradish peroxidase, diluted with 1.5% serum in PBS for
30 min at room temperature. If you do this, then skip step 20
and go straight to step 21.
14. We usually use DAB solution from Biocare, although other
chromagens are available. DAB generates a brown precipitate
and is very carcinogenic. Take care to use gloves and follow
manufacturer’s instructions.
15. Deionized water is sufficient in most cases, but milliQ or
ddH2O is fine to use if available.
16. You can use gentle pressure to push small bubbles to the edge
of the coverslip. If the bubbles are very large, you probably did
not use enough mounting medium. Put the whole slide with
the coverslip back into xylene to float off the coverslip and start
over. If you try to pry off the coverslip, you are likely to damage
the tissue.
17. Embedding frozen tissue is best accomplished using a cryo-

plate for even freezing of the sample.
18. Section thickness depends on structure or antigen of interest.
Thicker sections can be used for larger structures, such as
vasculature or nerves. Thinner sections may be used for molec-
ular or cellular antigens of interest. Of note, stronger deter-
gents may be needed to facilitate antibody penetration into
thick sections. Section thickness is also limited by decreasing
optical transparency and subsequent image quality.
19. The PAP pen may need to be vortexed prior to use. Take care
to only press the pen tip gently on the slide, as the liquid may
rush and pool. If greater adherence to the slide is needed, the
PAP pen outline may be re-enforced by spreading with a cotton
tip applicator. This barrier allows for use of smaller aliquots of
antibody and also prevents drying out of tissue over long
periods of time.
20. Thin sections (10–20 μm) may be incubated in primary anti-
body overnight. Thick sections (20 μm+) should be incubated
for 48-h to allow maximum penetration and equilibration.
21. Secondary antibody incubation times may vary. Thin sections
(10–20 μm) may be incubated at room temperature for 1–3 h.
Thick sections (20 μm+) should be incubated for 24 h at 4 C.
22. DAPI concentration may range from 0.1 μg/mL to 10 μg/
mL. For best results, troubleshoot concentrations when a new
batch is received for your desired application. For example,
high concentrations will result in densely stained, more
uniform nuclei, while low concentrations will reveal differences
between chromocenters and chromatin compaction if imaging
of nuclear architecture is needed.
23. The quality of the fluorescence will begin to degrade after the
first few months of storage. Use of mounting media containing
DAPI may require imaging within the first few days of staining.
For best results, specimens should be imaged as soon as possi-
ble to ensure the most accurate representation of the antigens.
References
1. Bord S (2003) Protein localization in 4. L’Hoste RJ Jr, Tourres MA (1995) Using Zinc
wax-embedded and frozen sections of bone Formalin as a Routine Fixative in the Histology
using immunohistochemistry. Methods Mol Laboratory. Lab Med 26(3):210–214
Med 80:237–247 5. Gage GJ, Kipke DR, Shain W (2012) Whole
2. Freida CL, Hladik C. Histotechnology.; A self animal perfusion fixation for rodents. J Vis Exp
instructional text , 2009 65
3. Sheehan DC, Hrapchak BB (1980) Theory and 6. Hao Z, Kalscheur VL, Muir P (2002) Decalcifi-
Practice of Histotechnology. 2nd ed. Bancroft cation of bone for histochemistry and immuno-
JD, Stevens J, editors. St. Louis, MO, The CV histochemistry procedures. J Histotechnol 25
Mosby Company (1):33–37
7. Serowoky MA, Patel DD, Hsieh JW, Mariani FV 9. Watkins S (2003) Immunohistochemistry. In:
(2018) The use of commercially available adhe- Ausubel FM, Brent R, Kingston RE, Moore
sive tapes to preserve cartilage and bone tissue DD, Seidman JG, Smith JA et al (eds) Current
integrity during cryosectioning. BioTechniques protocols in molecular biology. John Wiley &
65(4):191–196 Sons Inc., Hoboken, NJ. p. Supplement 7 Unit
8. Battifora H (1999) Quality assurance issues in 14.6
immunohistochemistry. J Histotechnol 22
(3):169–175
Chaptert 16
Mapping Regional Cortical Bone Responses to Local

Changes in Loading and Systemic Stimuli
Sara H. Windahl, Peter J. Delisser, and Gabriel L. Galea
Abstract
Quantification of cortical bone mass and architecture using μCT is commonplace in osteoporosis and
osteoarthritis research. Different groups often report substantially divergent mouse cortical bone responses
to nominally comparable interventions. In the case of studies assessing bones’ responses to externally
applied loading, these differences are commonly associated with methodological differences in the loading
regime. This chapter describes a widely published, standardized method of in vivo mouse tibia axial loading
to produce lamellar bone formation. Despite uniform application of axial loading, changes in bone mass are
highly site-specific within individual bones. For example, the mouse proximal tibia rapidly accrues new bone
following axial loading, but this osteogenic response tapers to produce undetectable differences distally.
Consequently, the bone sites selected for comparisons substantially influence the magnitude of differences
observed. Application of the freely available Site Specificity software allows site-specific responses to be
identified by rapidly quantifying cortical bone mass at each 1% site along the bone’s length. This high-
content screening tool has been informatively applied to study the local effects of changes in loading as well
as systemic interventions including hormonal treatment and aging. Automated multisite analyses of cortical
mass is increasingly identifying site-specific effects of “systemic” interventions such as global gene deletions.
Biological mechanisms underlying this apparent regionalization of cortical responses are largely unknown
but may start to be elucidated by increasingly widespread application of Site Specificity methods.
Key words Mechanical loading, Micro-computed tomography, Site specificity
1 Introduction
Mechanical loading is the primary functional determinant of bone

mass and architecture. Increases in loading-engendered mechanical
strain locally trigger new bone formation, whereas reduced loading
during disuse favors resorption [1]. The resulting mechano-
adaptive feedback loop, called the Mechanostat, locally adapts
bone mass to its load-bearing function [2]. Mechanisms underlying
the Mechanostat have largely been elucidated through the
Sara H. Windahl and Gabriel L. Galea contributed equally to this work.
275
276 Sara H. Windahl et al.
application of in vivo loading models to the bones of various

vertebrates [3]. Axial loading of the mouse tibia is now the most
commonly applied in vivo loading model. Within the loaded bone,
osteocytes and osteoblasts rapidly respond through wide-ranging
transcriptomic changes [4], increased osteoblast and osteocyte
metabolic activity [1, 5], and predictable downregulation of the
anti-osteogenic Wnt antagonist sclerostin [6, 7]. However, the
resulting loading-induced changes in cortical bone mass reported
by different groups vary markedly. Some of these differences can be
attributed to methodological changes such as insertion of rest
periods, different waveforms, or studying mice of different ages [3].
Here we describe a widely published, standardized method of
in vivo mouse tibia axial loading to produce lamellar bone forma-
tion [8–11]. Prior to performing in vivo loading it is essential to
measure the strength of cadaveric bones in order to calculate peak
loads which need to be applied to generate comparable strains
between experimental groups. In wild-type mice, the change in
cortical bone mass following loading is linearly related to peak
strain when habitual loading is eliminated [12, 13]. The methods
required for this strain-gauging procedure have been extensively
described elsewhere [14, 15]. The standard in vivo loading proto-
col described here is intended to generate lamellar bone formation
in 16-week-old C57BL/6 mice of both genders.
An additional confounding variable is the site of the bone in
which functional adaptation is assessed. While seemingly paradoxi-
cal, applying mechanical loading axially through the mouse tibia is
now known to cause bone gain in a site-specific manner
[16, 17]. The proximal tibia shows a much greater increase in
cortical bone mass than the distal tibia following 2 weeks of
tri-weekly axial loading [13]. Although the biological mechanisms
underlying this site-specificity are unknown, engineering-inspired
hypotheses such as the effect of greater proximal curvature have
been proposed to explain it [18]. What is even more perplexing is
that “systemic” interventions such as deletion of osteogenic or anti-
osteogenic genes also regionally alter cortical bone mass. One of
the first studies to describe genetically determined site-specificity
investigated the effect of ERα deletion in osteoclasts on femoral
bone mass [19]. Measurement of femoral bone mineral density in
20 equal segments along the length of the femur using dual X-ray
absorptiometry identified genotype-related differences proximally,
but not distally. The time and resources required to scan an entire
mouse bone initially limited the analysis of multiple sites along each
bone using micro-computed tomography (μCT). As μCT scanning
became faster and more accessible, various groups reported changes
in bone mass in response to loading as well as systemic interven-
tions at multiple sites along the length of the bone [17, 20, 21]. We
further extended this multisite analysis approach, initially to pro-
vide a global picture of compartmentalized cortical bone
Mapping Regional Cortical Bone Responses to Local Changes in Loading. . . 277
phenotypes caused by Prkcα deletion [22]. This global gene dele-

tion produced very regional in-growth of trabecular-like bone into
the endosteal cavity at specific sites of long bones.
Problems faced early-on in applying site-specific analyses of
bone mass included requirement for extensive computing power
to analyze each μCT slice through an adult mouse bone (typically
~2000 slices at 10 μm voxel sizes), needing to exclude the fibula
from studies focused on the tibia, and difficulties in aligning each
bone at a meaningful landmark. To address these we developed a
simple Site Specificity workflow, which was validated and reported
as a sensitive, rapid screening tool [23]. Site Specificity follows a
very simple workflow with open source MATLAB-executed code.
Similar workflows have also been described by others [24]. Applica-
tion of Site Specificity allows quantitative discrimination between
regional and global changes in cortical mass in response to changes
in loading, aging, genetic mutations, and hormonal treatment
[10, 23, 25].
In this chapter we detail the methods involved in axially loading
the mouse tibia, imaging the tibia using μCT analysis, and screening
for site-specific versus global changes in cortical bone mass.
2 Materials
2.1 Strain Gauging 1. Strain gauge, 120 Ω: Micro Measurements (Vishay Precision
and In Vivo Axial Tibial group, NC, USA), catalog number EA-06-031DE-120.
Loading 2. PCT-3 M gauge installation tape, Micro Measurements.
3. M-Flux, Micro Measurements.
4. Single conductor wire e.g.134 AWP, Micro Measurements.
5. M-Bond 200 adhesive and Catalyst, Micro Measurements.
6. Tin lead soldering wirer 60/40 0.8 mm.
7. Soldering station.
8. Multimeter, e.g., Digital multimeter, auto-range, VWR (Lut-
terworth, UK), catalog number 620–1920.
9. Dissection microscope with extra light.
10. Glass plate.
11. Cotton swabs.
12. Scalpels and forceps.
13. Isopropanol or 70% ethanol.
14. Polyurethane spray (can be purchased from local hardware
stores).
2.2 Ex Vivo μCT 1. There are several μCT instruments. Methods and settings
Imaging described in Subheading 3.2 Ex vivo μCT imaging below is
optimized for the use of an 1172 model, Bruker microCT
(Aartselaar, Belgium).
2. BMD calibration phantoms, 0.25 and 0.75 g/cm3, Bruker
microCT (Aartselaar, Belgium).
2.3 Site Specificity 1. MATLAB® 2012 or later.

and Statistical 2. FIJI (NIH, ImageJ) with BoneJ installed.
Analysis
3. IBM SPSS Statistics.
3 Methods
3.1 Strain Gauging Pre-gauging is performed to ensure that matched osteogenic

and In Vivo Axial Tibial strains are engendered by the loads applied to each experimental
Loading group. Pre-gauging should be performed before the loading exper-
iment on a subset of animals from the same group to be used in the
loading experiment. When multiple groups are to be compared, the
load:strain relationship should be calculated for each group (see
Note 1). Protocols related to anesthetizing and working with live
mice are beyond the scope of this chapter.
3.1.1 Preparation 1. Place the gauge on the glass board (shining side up, Fig. 1).
of the GAUGES 2. Secure one of the long edges of the gauge to the board with
adhesive masking tape.
3. Cut three sides of the gauge as shown in Fig. 1.
4. Lightly scratch the terminals with the blade to improve solder
adherence.
5. Apply a drop of M-Flux to each terminal.
6. Apply solder to each terminal.
Fig. 1 Sites for trimming the strain gauge. The gauge is attached to a glass plate
at the bottom of the image. The gauge is then cut with a scalpel in the indicated
order 1–4
Fig. 2 Photomicrograph of strain gauge with the attached wires before the final
cut. Solder points (arrows) with wire attached. The side cuts have been made
and the top cut is all that remains to be made. The top edge of the gauge is
immobilized by masking tape. The scale bar indicates 0.5 mm
7. Prepare two lengths of wires to be soldered onto the strain

gauge:
(a) Trim wire lengths to connect the gauge from the mouse’s
leg to the strain gauge unit.
(b) Remove the insulation from the end of the wires.
(c) Dip the ends in M-Flux.
(d) Apply solder to the ends.
8. Solder the wires to the terminals of the gauge (see Fig. 2).
9. Test the soldered gauge with a multimeter to ensure the resis-
tance remains within the tolerance of the gauge (120 0.2 Ω is
perfect; 118–121 Ω should work, below 120 is better than
above) and then tape the wires together.
10. Cut the gauge, loosening it from the tape (line 4, Fig. 1).
11. Spray the gauge with polyurethane.
12. Test the gauge again with the multimeter.
3.1.2 Attach the Strain 1. Euthanize the mouse using an ethically approved method
Gauge to the Tibia of a which the experimenter is competent and confident
Recently Euthanized Mouse performing.
to Pre-gauge (See Fig. 3) 2. Locally dissect the area of the right tibia where the strain gauge
should be applied.
3. Wipe the bone with a cotton swab dipped in EtOH or isopro-
panol to degrease the bone and facilitate attachment of the
gauge.
3.1.3 Attachment 1. Place a drop of catalyst and a drop of glue on a piece of foil.
of the Gauge to the Mouse 2. Dip the gauge in catalyst.
Tibia
Fig. 3 Attachment site of the gauge. The gauge is attached so that the center of
the gauge is located approximately at 37% of the bone’s length proximally. The
arrows indicate where the top of the gauge will be located
3. Immediately dip the gauge in glue and touch a clean bit of the
foil surface to remove excess glue. Too much glue will disable
the function of the gauge.
4. Hold the gauge steady on the bone with forceps to attach it to
the bone. See Fig. 3 for the correct site. The center of the
gauge should be where the tibia is the widest. The upper part of
the gauge should be positioned where there is an indention at
the medial side.
5. Test the gauge with a multimeter.
6. Secure with polyurethane (on gauge and bone if stored cold in
PBS before use).
3.1.4 Pre-gauging Place the Limb to Be Pre-gauged in Your Loading Device

Protocol and Subject it to Multiple, Graded Loads, Recording the Strain
Gauge Resistance at Each Load
A typical pre-gauging waveform is described below:
1. Ramp Load 0.5 N/s to 0.5 N (pre-load)

2. Dwell Load 7s
3. Ramp Load 460 N/s to 6 N (peak compressive load)
4. Dwell Load 5s
5. Ramp Load 460 N/s to 0.5 N
(continued)
6. Dwell Load 2s
7. Repeat 3–6 two times, recording the resistance value achieved each
time (in total three recordings per load).
8. Repeat 3–7 with increasing peak loads in step 3.
9. Ramp Displace 1.9 mm/s to 0.8 mm (unload)
Test each load at least three times on each bone (see table
above). The resulting load versus gauge resistance plot can then
be used to generate load:strain graphs by using the gauge factor to
convert resistance values into strain values.
3.1.5 Axially Load Once you know what loads need to be applied experimentally, place
the Tibia an anesthetized mouse in the loading device so that the knee is fixed
of Anesthetized Mice in the upper cup, and the foot is fixed in the lower cup. Make sure
that the foot and knee are aligned so that the joint is loaded axially
(see Note 2). The sequence below describes a standard rest-inserted
loading program applying 40 cycles at 11 N compressive load.
1. Ramp Load 0.5 N/s to 0.5 N (pre-load)

2. Dwell Load 2s
3. Ramp Load 393 N/s to 11 N (peak load, Note 3)
4. Dwell Load 0.05 s
5. Ramp Load 393 N/s to 0.5 N
6. Dwell Load 10 s (Rest period)
7. Repeat 3–6 39 repeats (repetitions)
8. Ramp Displace 1.9 mm/s to 1.9 mm (Unload)
3.2 Ex Vivo μCT 1. Euthanize the mouse using an ethically approved method
Imaging which the experimenter is competent and confident
performing.
3.2.1 Dissect the Left
(Control) and Right 2. Using sharp-tipped scissors, make a skin incision over the
(Exogenously Loaded) Tibia femur. The skin can then be manually pulled distally over the
calcaneus (see Note 3).
3. Proximally, flex the knee (stifle) and insert the sharp tip of a
10A scalpel laterally to medially through the popliteus to sever
muscle and tendon attachments. Similarly, cut through the
straight patellar ligament and collateral ligaments on either
side of the knee. This should destabilize the joint, allowing
you to insert the scalpel into the joint to cut the collateral
ligaments and disarticulate the femur from the tibia (see
Note 4).
4. Distally, flex the ankle (hock) and cut through the intertarsal
joints. It is not necessary to disarticulate the talus from the tibia
and doing so can occasionally damage the medial malleolus and
artifactually shorten the tibia, affecting bone measurement.
3.2.2 Fix and Dehydrate 1. Fix the bone and muscle in chilled 4% PFA for 2 days.
the Tibia 2. Dehydrate the tibia and muscle in 70% ethanol (see Note 5).
3.2.3 μCT Scan Each Methods and settings described here may need to be adapted to the
Bone specific μCT scanner available following the manufacturers’
instructions.
1. Prepare the bone by rolling the bone in non-PVC cling film
and place in the container (i.e., a straw). Attach the straw to the
μCT stage stand before screwing it into the scanner. If the bone
can be scanned dry, make sure the bone has dried out for
~10 min before the scan. Evaporation during the scan can
alter the density values. If the bone needs to remain moist,
then it can be scanned in 70% ethanol within a cut-down 1 mL
syringe using syringe stoppers above and below the bone to
keep the bone stable and avoid evaporation.
2. Select the parameters that best suit your bone. For mouse
bones, we use a 0.5 mm aluminum filter to reduce beam
hardening. Beam hardening is what happens when the bone
preferentially absorbs low energy photons and results in a dark
looking surface. The filter will decrease the low energy output
and improve the image. The whole tibiae and surrounding
muscles are typically imaged with a voxel size of 4.8 μm
(110 μm3). The applied X-ray voltage is 49–50 kV, current of
200 mA, with 0.5 mm aluminum filtration. Scans are obtained
over 180 degrees with a 0.6-degree rotation step. The images
can be reconstructed and binarized with global thresholding
(values: 1.000–1.160) using the NRecon Bruker software.
3. If the entire bone does not fit on one scanning run, an “over-
size scan” may be needed to scan larger bones, which can then
be merged digitally in post-processing to generate a single
bone series.
4. Once scanning is complete, replace bone in 70% ethanol solu-
tion and proceed with processing for additional tests (see
Note 6).
3.2.4 Reconstruction Reconstruct the RAW 2D Radiographic Images Generated from

the CT Scanner Following the Manufacturer’s Instructions
This should provide reconstructed *.bmp files representing
cross-sections through the bone, which can be used in Site Speci-
ficity analysis (see Note 7).
3.3 Site Specificity 1. Place the reconstructed cross-section μCT images for each
and Statistical bone in a single folder, henceforth the “Inpath.” Ensure
Analysis there are no blank (“spacer”) slices in the folder as bone length
is calculated from the number of slices in this folder.
2. Create a local copy of the SiteSpecificityV2.m script. This can
be downloaded from https://www.researchgate.net/publica
tion/295858230_SiteSpecificityV2 and is available with the
original manuscript [23].
3. Open MATLAB®, direct it to the directory containing SiteSpe-
cificityV2.m and open the script for editing.
4. Modify the Inpath line (line 13) to the folder containing the
bone to be analyzed (see Note 4). For example, change this line
to read:inpath ¼ ‘C:\Site specificity\Example\’
5. Save the script file and run it. The script will create an “Out-
put” folder within the Inpath. It will count the number of
images of the folder and identify the images corresponding to
each 1% site along the bone’s length (see Note 8). The
corresponding 100 images each undergo the following
processing (see Fig. 2a, b):
(a) Locally adaptive binarization to identify the bone and
exclude soft tissue.
(b) Identification of the largest continuous object (the tibia)
and exclusion of smaller objects (the fibula and trabecular
bone).
(c) Quantification of the binarized area. This is exported as
“Bone Area” and corresponds to Cortical Area (Ct.Ar) in
conventional μCT analysis. The output report will display
this as pixels2 and should be converted to mm2 using the
voxel size known from the μCT reconstruction settings.
(d) Quantification of empty space within the binarized area
(see Note 9). This is reported as “Marrow Area” (i.e., Ma.
Ar) in pixels2.
(e) The perimeter of the binarized area is also provided. If
required, Bone Area and Marrow Area can be summer to
calculate Total Tissue Area (Tt.Ar).
6. The Output folder will now contain 100 images saved as .gif
files corresponding to each 1% site along the bone’s length, .txt
result files with the cross-sectional measurements at each site,
and a single .csv file amalgamating the results (see Note 10).
The .gif files retain the same pixel dimensions as the original
image but are of much smaller size such that they can be easily
processed in additional software if needed.
7. To quantify parameters beyond those directly calculated by Site

Specificity, drag and drop the Output folder into Fiji. This will
open the subsampled cross-section images as a stack. Addi-
tional parameters, such as moments of inertia and cross-
sectional thickness, can now be calculated using the BoneJ
plugin [26].
8. Iteratively repeat the process until all bones have been ana-
lyzed. The resulting data can now be graphed as in Fig. 4c to
confirm successful analysis. If using Site Specificity as a screen-
ing tool with a small number of bones, you can stop at this
point. The following steps describe statistical comparison
between treatment groups using the mixed model procedure
in SPSS.
Fig. 4 Execution and application of the Site Specificity workflow. (a) Schematic representation of Site
Specificity analysis applied to the tibia of a 3-week-old mouse. The region of the bone analyzed by Site
Specificity are shown in the binarized reconstruction (right). Parts of the fibula (red highlight) not connected to
the tibia are identified and excluded. Images corresponding to each 1% site along the bone’s length are
identified. Proximal (including the growth plate, green) and distal sites are excluded from the analysis. (b)
Cross-section through the proximal tibia and fibula (red) at the 20% site from the proximal end. The binarized
tibia images are saved and can be exported into other analysis software such as BoneJ. Cortical area (Ct.Ar),
Marrow area (Ma.Ar), and external perimeter are provided as part of the in-built workflow. (c) Quantification of
Ct.Ar and Ma.Ar in the tibiae of five 3-week-old mice showing the expected pattern of cortical bone differences
in typical inter-mouse variability
(a) Mixed models are used to test the effect of random effects
(e.g., cage number), fixed effects (e.g., loading), and fixed
covariates (e.g., % site along the bone’s length). Strengths
and limitations of mixed models are described elsewhere
(e.g., [27]).
(b) Set up a datasheet with the following headings: Animal
ID, Site (% bone length), and Treatment (e.g., loading,
drug, genotype). Enter all data in a single list under each
of these headings. Set the appropriate variable types in the
SPSS “Variable view” tab.
(c) In SPSS, select the linear mixed models option. Specify
the Animal ID as the “subject,” Site as the “repeated
measure” (see Note 11), and click continue. Set up your
model as needed on the next screen (see the SPSS help
function on your version of the software if you are not
familiar with how to do this).
(d) An important outcome of Site Specificity analysis is quan-
tification of whether the response to treatment is signifi-
cantly different between different sites. This is shown by a
significant Site by Treatment interaction (indicates the
response to treatment is significantly different between
different sites).
(e) To request a direct comparison between different sites
with a Bonferroni post hoc comparison, click “Paste” to
obtain the model syntax and add the following text at the
end of the syntax before the full stop:
/EMMEANS ¼ TABLES(Treatment*Site) COMPARE
(Treatment) ADJ(Bonferroni)
(f) Post hoc comparisons are generated in a table under the
subheading “Pairwise Comparisons” (see Notes 12 and
13).
(g) The same syntax can now be rerun for other bone para-
meters in the same experiment.
4 Notes
1. Be careful when applying a treatment/genetic modification

which alters bone structure before starting a loading experi-
ment. If the same load is applied to bones with different mass
and architecture, different levels of osteogenic strain will be
engendered.
2. If the tibia is not placed directly vertical between the upper and
lower cups, the knee will be forced forward or back, which is
likely to rupture the cruciate ligaments. If this happens in an
anesthetized mouse, cull the mouse before it wakes up.
3. Most materials testing devices will avoid exceeding the speci-

fied load and, at rapid loading rates with short dwell times,
often fail to achieve the desired peak load. You may need to set
a slightly higher programmed peak load in order to achieve the
desired force (e.g., set a peak load of 11.2 N to achieve a load
of 11 N).
4. It is usually easy to evert and unroll mouse skin off the tibia up
to the calcaneus, but harder to fully remove it. There is no need
to pull it further. If the tibia is to be scanned and analyzed,
ensure the femur is grasped instead to avoid fracturing the
fibula. The fibula is itself a loading-responsive bone which can
be analyzed separately.
5. Disarticulating the tibia and femur can be tricky in some mice.
It is essential to obtain the entire tibia to direct the analysis at
each percent of the bone’s length. If the experimenter finds this
difficult at first, cut through the distal femur instead. Handling
tissues with wet gloves on wet tissue paper sometimes helps and
a dissection microscope can be used if needed.
6. The same bone can be used for histology and histomorphome-
try end points as well as μCT scanning. For histology, the bone
should be fixed (e.g., in paraformaldehyde) prior to dehydra-
tion. It is often helpful to trim away some of the overlying
muscles to aid fixative penetration but be careful not to break
the fibula.
7. SiteSpecifictyV2.m is intended to analyze .bmp image files. If a
different file type is used, this can be changed in line 26 (“d ¼
dir([inpath, ‘*.bmp’]);”).
8. The level of subsampling is set to 100 images, with each image
corresponding to a 1% site along the bone’s length by default.
This subsampling can be changed in line
31 (“num_samples ¼ 100;”).
9. Large pores and blood vessels which breach the cortex in a
single slice effectively cause the marrow area to become contin-
uous with the background, producing a small or zero Ma.Ar
measurement which should be excluded from subsequent ana-
lyses. Straight trans-cortical breaches in a single slice are rare in
the cortex of young adult mice and the Ct.Ar measurement
primarily used to screen bone mass is unaffected. Workarounds
are therefore not directly implemented in Site Specificity, but if
these breaches are problematic (e.g., genetic model with very
low bone mass), the subsampled binarized images saved in the
Output folder can easily be reanalyzed in conventional μCT
analysis software (e.g., CtAN, Bruker) implementing a shrink-
wrap function. The small number and size of subsampled
images means these can be batch-processed very quickly.
10. If the extremes of the bone have very faint signal which is of
soft-tissue equivalence, such that they binarize to produce a
black image, that slice will be excluded from the analysis. The
total number of output images and values will therefore be less
than 100. Additionally, the proximal and distal extremes of the
bone contain trabecular bone which cannot be analyzed using
Site Specificity. We therefore exclude the proximal and distal
10–15% of the tibia from our analyses. This may not be neces-
sary when analyzing less trabecular bone such as the
mouse ulna.
11. The structure of the model used will vary depending on the
endpoint. When testing whether an intervention enhances or
diminishes the response to loading, it may be more appropriate
to calculate the percentage difference between loaded and
control limbs for each mouse and use this loading response as
the dependent variable. This is especially true if the interven-
tion itself changes baseline bone mass (see Strain Gauging
methods above).
12. Displaying regions of significant differences between treatment
groups across multiple sites can be challenging. We and others
have adopted various methods to show this in previous
publications.
13. Calculating the statistical power to identify differences as sta-
tistically significant using a complex mixed model approach is
challenging. It is possible to simulate changes in bone mass by
resizing the Output images in Fiji by a known percentage. This
provides an indication of the proportion of sites at which that
percentage change would be detected as statistically significant.
In our initial validation of this analysis method [23], a 10%
change in Ct.Ar was detected as significant at 98.8% of sites
analyzed.
References
1. Hillam RA, Skerry TM (1995) Inhibition of 4. Galea GL, Meakin LB, Harris MA, Delisser PJ,
bone resorption and stimulation of formation Lanyon LE, Harris SE, Price JS (2017) Old age
by mechanical loading of the modeling rat ulna and the associated impairment of bones’ adap-
in vivo. J Bone Miner Res 10:683–689 tation to loading are associated with transcrip-
2. Skerry TM (2006) One mechanostat or many? tomic changes in cellular metabolism, cell-
Modifications of the site-specific response of matrix interactions and the cell cycle. Gene
bone to mechanical loading by nature and nur- 599:36–52
ture. J Musculoskelet Neuronal Interact 5. Skerry TM, Bitensky L, Chayen J, Lanyon LE
6:122–127 (1989) Early strain-related changes in enzyme
3. Meakin LB, Price JS, Lanyon LE (2014b) The activity in osteocytes following bone loading
Contribution of Experimental in vivo Models in vivo. J Bone Miner Res 4:783–788
to Understanding the Mechanisms of Adapta- 6. Moustafa A, Sugiyama T, Prasad J, Zaman G,
tion to Mechanical Loading in Bone. Front Gross TS, Lanyon LE, Price JS (2012)
Endocrinol (Lausanne) 5:154 Mechanical loading-related changes in
osteocyte sclerostin expression in mice are relationship between bone formation and
more closely associated with the subsequent mechanical stimulus within cortical bone:
osteogenic response than the peak strains Combining 3D fluorochrome mapping and
engendered. Osteoporos Int 23:1225–1234 poroelastic finite element modelling. Bone
7. Robling AG, Niziolek PJ, Baldridge LA, Con- Rep 8:72–80
don KW, Allen MR, Alam I, Mantila SM, 17. Sugiyama T, Saxon LK, Zaman G, Moustafa A,
Gluhak-Heinrich J, Bellido TM, Harris SE, Sunters A, Price JS, Lanyon LE (2008)
Turner CH (2008) Mechanical stimulation of Mechanical loading enhances the anabolic
bone in vivo reduces osteocyte expression of effects of intermittent parathyroid hormone
Sost/sclerostin. J Biol Chem 283:5866–5875 (1-34) on trabecular and cortical bone in
8. Bergstrom I, Isaksson H, Koskela A, mice. Bone 43:238–248
Tuukkanen J, Ohlsson C, Andersson G, Wind- 18. Dodge T, Wanis M, Ayoub R, Zhao L, Watts
ahl SH (2018) Prednisolone treatment reduces NB, Bhattacharya A, Akkus O, Robling A,
the osteogenic effects of loading in mice. Bone Yokota H (2012) Mechanical loading, damp-
112:10–18 ing, and load-driven bone formation in mouse
9. Lionikaite V, Henning P, Drevinge C, Shah FA, tibiae. Bone 51:810–818
Palmquist A, Wikstrom P, Windahl SH, Lerner 19. Nakamura T, Imai Y, Matsumoto T, Sato S,
UH (2019) Vitamin A decreases the anabolic Takeuchi K, Igarashi K, Harada Y, Azuma Y,
bone response to mechanical loading by sup- Krust A, Yamamoto Y, Nishina H, Takeda S,
pressing bone formation. FASEB J Takayanagi H, Metzger D, Kanno J,
33:5237–5247 Takaoka K, Martin TJ, Chambon P, Kato S
10. Meakin LB, Todd H, Delisser PJ, Galea GL, (2007) Estrogen prevents bone loss via estro-
Moustafa A, Lanyon LE, Windahl SH, Price JS gen receptor alpha and induction of Fas ligand
(2017) Parathyroid hormone’s enhancement in osteoclasts. Cell 130:811–823
of bones’ osteogenic response to loading is 20. de Souza RL, Pitsillides AA, Lanyon LE, Skerry
affected by ageing in a dose- and time- TM, Chenu C (2005) Sympathetic nervous
dependent manner. Bone 98:59–67 system does not mediate the load-induced cor-
11. Svensson J, Windahl SH, Saxon L, Sjogren K, tical new bone formation. J Bone Miner Res
Koskela A, Tuukkanen J, Ohlsson C (2016) 20:2159–2168
Liver-derived IGF-I regulates cortical bone 21. Javaheri B, Carriero A, Wood M, de Souza R,
mass but is dispensable for the osteogenic Lee PD, Shefelbine S, Pitsillides AA (2018)
response to mechanical loading in female Transient peak-strain matching partially
mice. Am J Physiol Endocrinol Metab 311: recovers the age-impaired mechanoadaptive
E138–E144 cortical bone response. Sci Rep 8:6636
12. Meakin LB, Galea GL, Sugiyama T, Lanyon 22. Galea GL, Meakin LB, Williams CM, Hulin-
LE, Price JS (2014a) Age-related impairment Curtis SL, Lanyon LE, Poole AW, Price JS
of bones’ adaptive response to loading in mice (2014) Protein kinase Calpha (PKCalpha) reg-
is associated with sex-related deficiencies in ulates bone architecture and osteoblast activity.
osteoblasts but no change in osteocytes. J J Biol Chem 289:25509–25522
Bone Miner Res 29:1859–1871 23. Galea GL, Hannuna S, Meakin LB, Delisser PJ,
13. Sugiyama T, Meakin LB, Browne WJ, Galea Lanyon LE, Price JS (2015) Quantification of
GL, Price JS, Lanyon LE (2012) Bones’ adap- alterations in cortical bone geometry using site
tive response to mechanical loading is essen- specificity software in mouse models of aging
tially linear between the low strains associated and the responses to ovariectomy and altered
with disuse and the high strains associated with loading. Front Endocrinol (Lausanne) 6:52
the lamellar/woven bone transition. J Bone 24. Javaheri B, Herbert E, Hopkinson M,
Miner Res 27:1784–1793 Al-Jazzar A, Pitsillides AA (2019b) Sost hap-
14. Javaheri B, Bravenboer N, Bakker AD, van der loinsufficiency provokes peracute lethal cardiac
Veen A, de Souza RL, Saxon L, Pitsillides AA tamponade without rescuing the osteopenia in
(2019a) In vivo models of mechanical loading. a mouse model of excess glucocorticoids. Am J
Methods Mol Biol 1914:369–390 Pathol 189:753–761
15. Melville KM, Robling AG, van der Meulen MC 25. Orriss IR, Lanham S, Savery D, Greene NDE,
(2015) In vivo axial loading of the mouse tibia. Stanier P, Oreffo R, Copp AJ, Galea GL (2018)
Methods Mol Biol 1226:99–115 Spina bifida-predisposing heterozygous muta-
16. Carrieroa A, Pereirab AF, Wilson AJ, tions in planar cell polarity genes and Zic2
Castagno S, Javaheri B, Pitsillides AA, reduce bone mass in young mice. Sci Rep
Marenzana M, Shefelbine SJ (2018) Spatial 8:3325
26. Doube M, Klosowski MM, Arganda-Carreras I, 27. Yang J, Zaitlen NA, Goddard ME, Visscher
Cordelieres FP, Dougherty RP, Jackson JS, PM, Price AL (2014) Advantages and pitfalls
Schmid B, Hutchinson JR, Shefelbine SJ in the application of mixed-model association
(2010) BoneJ: free and extensible bone image methods. Nat Genet 46:100–106
analysis in ImageJ. Bone 47:1076–1079
Chapter 17
Pain and Activity Measurements

David H. H. Molstad and Elizabeth W. Bradley
Abstract
Musculoskeletal pain contributes significantly to chronic pain experienced by adults and to health care use.
This chapter details several methods to evaluate pain and physical activity in mice that can be applied to
preclinical orthopedic models. These methods include the von Frey filament assay that measures mechanical
allodynia, open-field activity assays for evaluation of ambulation, and incapacitance measurements to
determine static weight bearing.
Key words Functional assays, Mechanical allodynia, von Frey, Incapacitance, Evoked pain, Open-field
activity assays, Static weight bearing
1 Introduction
One of the main reasons patients visit physicians for a musculoskel-

etal problem is not the physical damage of tissue degeneration, but
the associated pain [1]. Moreover, musculoskeletal pain is poorly
controlled and contributes significantly to opioid use [2, 3]. Muscu-
loskeletal pain can be either nociceptive (e.g., fracture, torn liga-
ment) or neuropathic (e.g., osteoarthritis, low back pain).
Nociceptive pain arises when there is potential for tissue damage
or as a result of tissue damage; this type of pain is acute (e.g., pain
lasting under 6 months) and has a protective function [4]. In
contrast, neuropathic pain, otherwise known as chronic pain, is
initiated by the central nervous system [4]. Whereas acute pain
serves a protective function, chronic pain does not and instead
can lead to disability [5].
Functional measurements of pain and locomotion in preclinical
models in the research setting may therefore be helpful to assess
analgesic benefits of potential therapeutics. In addition, inclusion of
pain assessment in a research design may also shed light to mecha-
nistic action. Measurements of allodynia and hyperalgesia in pre-
clinical models include evoked mechanical-sensitivity, static and
dynamic weight bearing, gait and spontaneous pain measurements
291
292 David H. H. Molstad and Elizabeth W. Bradley
[6]. In this chapter we will detail methods to perform von Frey

filament assays, open-field activity measurements, and static weight
bearing (i.e., incapacitance) determinations. The methods
described below are for use in mouse models, but could be adapted
for use in other rodents.
In combination, these functional assessments of pain are useful
tools for evaluation of pain in murine models of musculoskeletal
conditions. These musculoskeletal models include pain associated
with cartilage degradation due to aging, as well as chemically
induced or surgically induced models of joint destruction [7–
9]. Bone pain arising as a result of induced fracture, bone defects,
or tumor metastasis to bone can also be evaluated using these
functional assessments [10, 11].
The von Frey filament assay quantifies evoked mechanical allo-
dynia or hyperalgesia in murine models (see Fig. 1). The theory
behind the von Frey assay is that in a pain state otherwise innocuous
stimuli, such as a light touch, elicit a pain response [12]. In this
method, animals are placed in a suspended enclosure over a mesh
grid and monofilament fibers with defined bending forces are
applied to the plantar surface of the hind limb. If the force applied
from the filament is nocifensive, the animal withdraws its paw.
Underlying pain influences the threshold force needed to elicit a
nocifensive response and causes a shift in the pain response curve
(see Fig. 1); thus, the von Frey assay measures mechanical allodynia
by determining the force required to produce nocifensive
responses.
Fig. 1 Pain sensitization enhances responses to mechanical stimuli.

Centralization of the pain response during chronic pain shifts the pain
response curve from baseline pain (gray curve) to sensitized pain (blue curve).
Allodynia (black shaded region) results when stimuli that are normally not
perceived as painful become nocifensive. In contrast, hyperalgesia (red
shaded region) characterizes an increased response to painful stimuli
Pain and Activity Measurements 293
The testing period consists of two phases. The first is testing

before experimental intervention (e.g., injury) to establish von Frey
percent withdrawal responses for each animal. The second phase is
testing following intervention. The von Frey monofilament test can
be repeated on each animal over the duration of the study to
determine how interventions affect pain associated with musculo-
skeletal injury. For example, one would expect paw withdrawal
percentage to increase following destabilization of the medial
meniscus surgery [8]. In a fracture-healing model, paw withdrawal
percentage would decrease along with bone healing. There are a
number of factors that can influence the reproducibility of data
obtained from the von Frey filament assays [13].
The advantages of the von Frey filament assays are that it is an
elicited response and that handling-induced stress is eliminated
because the test subject is unrestrained within the apparatus. The
limitations of the approach are the long acclimation periods and
labor-intensive testing. It is also subject to reader bias, and subject
responses may vary depending on the reader (e.g., male or female
reader). There are three different methods for performing the von
Frey filament assay: the up-down method, the ascending method,
and the percent response method [5]. In this chapter, we will
describe the percent response method, as this approach has a
lower chance of a learned response and is less labor intensive. In
this method, animals are first acclimated to the testing enclosure
and test operator. This is followed by baseline testing (e.g., before
injury) and testing following experimentation. Monofilaments of
increasing strength are then tested for a defined number of times
and the percent withdrawal for each filament is determined.
Open-field activity arenas are used to track the movement of
research animals in three-dimensional space. These arenas monitor
distance traveled, active ambulating time, resting time, rearing time
and rearing events, average speed and stereotypic movement time
(e.g., small movements such as grooming). These measurements
are made in discrete increments, which can be summed to reflect
total activity over a period of time (e.g., 10 increments of 2 min for
a total of 20 min). A two-dimensional grid of infrared photocell
pairs is used to define the x-y axes within the arena. A second set of
infrared photocells placed 2.5 cm above the arena floor monitor
movement in the z-axis due to rearing. Simultaneous interruption
of the infrared beams places the animal within this three-
dimensional grid of the arena. A single arena, or set of arenas run
in parallel, can be used to track changes in activity over time during
an experiment. Time of day will impact animal activity, so it is best
to use a consistent time of day for all experimental recordings.
Similarly, the environment the activity arenas are placed in can
also impact animal activity. The arena should be free of other
animals and sudden and/or loud noises.
The incapacitance assay measures static weight bearing of each

hind limb of an animal. The animal is within an animal holder at an
incline such that most of the animal’s weight is placed on the hind
limbs. A force plate below each hind limb then measures spontane-
ous postural changes reflected by the relative amount of weight
placed on each limb. It can therefore be used to determine how an
injury to one limb affects how much weight the animal will place on
the affected and contralateral limb [14]. The advantage of this
method is the spontaneous measurement of postural changes.
The disadvantage of this method is that it relies on the animal
voluntarily assuming the correct position within the enclosure.
This method can also only be used in models of unilateral injury.
2 Materials
2.1 von Frey 1. von Frey monofilament fibers, Semmes-Weinstein,

Filament Assays 0.008–300 g.
2. Ugo Basile Grid Platform for Plantar Stimulation (Stoelting,
Wood Dale, IL, Product #57816).
3. Multiple Configuration Animal Enclosure (Stoelting, Wood
Dale, IL, Product #57823 or Bioseb, Pinellas Park, FL,
BIO-PVF).
2.2 Open-Field 1. Activity Animal Meter: Opto-Verimex 5 Auto-Track, Colum-

Activity Measurements bus Instruments, Columbus, OH, or an equivalent system.
2. PC Station with appropriate hardware.
2.3 Static Weight 1. Incapacitance meter (Bioseb, Pinellas Park, FL, BIO-SWB-
Bearing Incapacitance TOUCH-M) or equivalent instrument.
Assays (a) Animal holder/restrainer (rat or mouse).
(b) Control unit.
(c) Platform with sensors.
(d) Footswitch to start/stop experiment hands free.
2. PC Station with appropriate hardware.
3. Lab tape.
3 Method
3.1 von Frey Two weeks prior to baseline readings:

Filament Assays
1. Introduce animals to experimental chambers several times each
3.1.1 Acclimation Period week for 1–2 h.
2. On the day of the assay, allow the mice to become acclimated to
their surroundings for about 15–30 min, or until the mice are
no longer exploring.
3.1.2 On Day of Testing 1. Assemble the animal enclosure/grid platform (see Fig. 2).
Ensure that each enclosure is fully accessible through the
underlying mesh of the grid platform. Note: Opaque material
should be used to separate mice so that they cannot see each
other during acclimation or testing periods (see Fig. 2).
2. Place animals in enclosure chambers. Let animals acclimate to
enclosure until they are no longer actively exploring. If animals
are properly acclimated to the enclosure chambers, this should
take under 15 min (see Note 1).
Fig. 2 Equipment for the von Frey Monofilament assays. (a) Assembly of the von Frey equipment includes
multiple animal enclosures placed above a suspended grid platform. Mice are placed within chambers so that
the plantar surface can be accessed through the mesh grid. An opaque wall separates chambers so that
animals cannot see each other during acclimation and evaluation periods. (b) A mouse placed within the
animal enclosure chamber demonstrating access of the hind limb plantar surface through the mesh grid. (c)
Monofilaments ranging in fiber bowing weight used to perform the von Frey assay. (d) The von Frey
monofilament is applied until the fiber bows to achieve a defined force
3. Testing should not be performed while the animal is rearing,

grooming, or actively exploring the animal enclosure cham-
ber (see Note 1).
4. Start with a low monofilament fiber bowing weight (e.g.,
0.16 g for mice).
5. Apply the fiber perpendicularly to the center of the plantar
surface of the hind paw of the animal’s affected limb (e.g.,
right or left) five times. Record the numbers of paw withdra-
wals out of five tests.
6. Repeat the monofilament stimulation for the affected limb of
each animal within the animal enclosure.
7. Test the contralateral paw of each animal (e.g., five monofila-
ment stimuli).
8. Retest each hind limb (affected followed by contralateral) two
times for a total of 15 potential paw withdrawal responses
per paw.
9. Repeat steps 5–8 using increasing fiber bowing weights (e.g.,
0.4 g and 0.6 g for mice).
10. Determine the average percent paw withdrawal by dividing the
number for withdrawals for each hind limb of each animal by
15 (total number of tests), then converting to percentage.
Report values for a fiber bowing weight that elicits 50% paw
withdrawal in the control baseline condition.
3.2 Open-Field 1. Transport mice to the area where activity arenas are located.
Activity Measurements Allow mice 15 min in cages to recover from transport before
starting analyses.
2. Open the Opto-track software. Open a new experiment, select
the location to save the file, and name the file.
3. Select arenas to be used in each run by checking the appropriate
checkbox within the software. Adjust the total time for the
analysis to 20 min or desired alternate time.
4. Verify that the software is localizing objects in the three-
dimensional space by tracing along the arena surface. Place
one animal in each arena. Allow the animal to acclimate within
the arena for 2 min (see Note 2).
5. Hit the “Start All” button within the Opto-track software.
Ensure the software is tracking the animal accurately. Allow
time to elapse.
6. When the experiment is complete, return animals to cages.
Select “Analyze Experiment,” and then export all bin files.
This will save a .csv file to the location you selected that can
be viewed with Excel software.
7. Open the .csv file in Excel. To obtain the total distance and
time measurements, sum all measured intervals. The speed of
the animal can be obtained through the average of all speed
values obtained during the experiment.
8. Determine changes between experimental groups and or
over time.
3.3 Static Weight 1. Transport mice to the area where the incapacitance meter is
Bearing Incapacitance located. Allow mice 15 min in cages to recover from transport
Assays before starting analyses.
2. Place the animal within the animal enclosure (see Fig. 3). Tap-
ing down the animal’s tail to the base of the enclosure will help
the animal assume correct placement.
3. Ensure that the animal has both upper limbs placed equally on
the inclined portion of the animal enclosure and that both hind
feet are located on the force plate (see Fig. 3 and Note 3).
4. Use the footswitch to take a 5-s measurement.
5. Repeat for a total of three consecutive 5-s measurements. Use
the average of these three measurements.
6. Determine the percent weight bearing on the affected limb
using the following equation:
% weight on ipsilateral limb ¼ (weight placed on ipsilateral
limb/summed weight placed on both limbs) 100%
7. Determine changes between experimental groups and/or
over time.
Fig. 3 Incapacitance measurement equipment. The incapacitance assay measures static weight bearing on
each hind limb of a mouse. (a) Side view of a mouse placed in the incapacitance chamber. (b) Forward view of
a mouse within the incapacitance chamber. (c) Rear view of a mouse placed in an incapacitance chamber.
Note the foot on the left is not contacting the force plate within the chamber. Lab tape is also used to secure
the animal’s tail to prevent the animal from turning around within the chamber
4 Notes
1. During the acclimation period of the von Frey assay, the reader
should be present. Acclimation of the experimental animals to
the enclosures will ensure that habituation occurs on test days.
Fiber bowing weights can be adjusted if a response is either not
elicited or is saturated (e.g., paw withdrawal is either 0% or
100%). During testing, be careful not to brush or scratch the
plantar surface with the monofilament. Contact with the outer
margins of the plantar surface should also be avoided. A posi-
tive response is evident if the animal actively withdraws its paw
from the stimulus.
2. The activity assay software should display the correct location
(lower left hand is 0,0 and the upper right is 30,30 within the
Cartesian plane, see (Fig. 4) and register that the z-plane is
broken. If software is not registering location correctly, ensure
that the arena walls or other items do not break infrared beams.
3. For static weight bearing assays, the animal must assume the
correct positioning within the enclosure before an accurate
reading can be taken. This includes the animal facing forwards
with both forelimbs placed on the incline and both hind feet
placed on the force plates. Likewise, it is difficult to determine if
an animal will not place its foot on the force plate due to pain or
due to non-compliance.
Fig. 4 Open-field activity assays. (a) An open-field activity assay arena. (b) Each activity arena tracks the
three-dimensional movement of a mouse within the arena. Cartesian coordinates for each corner of the arena
are shown and can be used to verify correct tracking of an animal by the infrared grid
Acknowledgements
This work was made possible by training and research grants from
the National Institutes of Health (AR065397 and AR072634) and
the University of Minnesota Stem Cell Institute and Board of
Regents. These contents are solely the responsibility of the authors
and do not necessarily represent the official views of the NIH.
References
1. Hawker GA (2017) The assessment of muscu- 8. Hwang SM, Feigenson M, Begun DL, Shull
loskeletal pain. Clin Exp Rheumatol 35(Suppl LC, Culley KL, Otero M, Goldring MB, Ta
107):8–12 LE, Kakar S, Bradley EW, Westendorf JJ
2. Carnide N, Hogg-Johnson S, Cote P, Irvin E, (2017) Phlpp inhibitors block pain and carti-
Van Eerd D, Koehoorn M, Furlan AD (2017) lage degradation associated with osteoarthritis.
Early prescription opioid use for musculoskele- J Orthop Res 36(5):1487–1497
tal disorders and work outcomes: a systematic 9. Xu M, Bradley EW, Weivoda MM, Hwang SM,
review of the literature. Clin J Pain Pirtskhalava T, Decklever T, Curran GL,
33:647–658 Ogrodnik M, Jurk D, Johnson KO, Lowe V,
3. Moshfegh J, George SZ, Sun E (2018) Risk Tchkonia T, Westendorf JJ, Kirkland JL (2017)
and risk factors for chronic opioid use among Transplanted senescent cells induce an
opioid-naive patients with newly diagnosed osteoarthritis-like condition in mice. J Geron-
musculoskeletal pain in the neck, shoulder, tol Ser A Biol Sci Med Sci 72:780–785
knee, or low back. Ann Intern Med 170 10. Majuta LA, Guedon JG, Mitchell SAT, Kus-
(7):504–505 kowski MA, Mantyh PW (2017) Mice with
4. Nicholson B (2006) Differential diagnosis: cancer-induced bone pain show a marked
nociceptive and neuropathic pain. Am J decline in day/night activity. Pain Rep 2:e614
Manag Care 12:S256–S262 11. Minville V, Laffosse JM, Fourcade O, Girolami
5. Deuis JR, Dvorakova LS, Vetter I (2017) JP, Tack I (2008) Mouse model of fracture
Methods used to evaluate pain behaviors in pain. Anesthesiology 108:467–472
rodents. Front Mol Neurosci 10:284 12. Lolignier S, Eijkelkamp N, Wood JN (2015)
6. Miller RE, Malfait AM (2017) Osteoarthritis Mechanical allodynia. Pflugers Archiv
pain: what are we learning from animal models? 467:133–139
Best Pract Res Clin Rheumatol 31:676–687 13. Mogil JS (2017) Laboratory environmental
7. Bradley EW, Carpio LR, McGee-Lawrence factors and pain behavior: the relevance of
ME, Castillejo Becerra C, Amanatullah DF, unknown unknowns to reproducibility and
Ta LE, Otero M, Goldring MB, Kakar S, Wes- translation. Lab Anim 46:136–141
tendorf JJ (2016) Phlpp1 facilitates post- 14. Pomonis JD, Boulet JM, Gottshall SL,
traumatic osteoarthritis and is induced by Phillips S, Sellers R, Bunton T, Walker K
inflammation and promoter demethylation in (2005) Development and pharmacological
human osteoarthritis. Osteoarthr Cartil 24 characterization of a rat model of osteoarthritis
(6):1021–1028 pain. Pain 114:339–346
INDEX
A E
Annulus fibrosus (AF)...............................................41–51 Endosteal mesenchymal progenitors ..........30, 33, 35–38
Antibodies .............................21, 91–93, 95, 96, 99, 100, Enzymatic digestions .............29–38, 42, 90, 98, 99, 271
196, 208, 221, 225, 229, 230, 236, 242, 244,
245, 256, 257, 261–266, 268, 270–272 F
Antigen retrieval................ 208, 244, 256, 262, 263, 267 Fixation ............................... 50, 168, 211, 214, 228–229,
236, 237, 244, 262, 266, 268, 270
B
Fluorescence activated cell sorting (FACS) ............89–91,
Bone marrow...........................15, 18, 19, 25, 29–38, 55, 93–97, 99, 100
57, 72, 194, 263 Fluorescence recovery ....... 112, 114–117, 121, 123, 134
Bone repair ........................................................... 193–203 Fluorescence recovery after photobleaching
Bones ...................................3–11, 15–26, 29, 30, 32, 33, (FRAP)...................................................... 109–137
35, 37, 38, 55, 57, 89, 90, 96, 97, 99, 154, 167, Fractures ............................ 193, 194, 205–222, 291, 292
174–178, 182, 185, 186, 188–190, 193, 194,
196, 198, 201–203, 206, 210, 212–214, 217, G
220, 221, 224, 234, 242, 248, 256, 262, 263,
Gene expression regulation ....................... 4, 50, 65, 110,
266, 267, 270, 275–287, 292, 293 142, 143, 172, 217–219, 222, 226, 230–233,
247, 264
C
Cartilage .................................... 21, 41, 53–69, 112, 142, H
154, 171, 183, 194, 206, 217, 221, 223–225,
Healing .............................193, 194, 196, 198, 203, 205,
230–232, 240, 242, 245, 247–254, 257, 262, 206, 210–212, 219–221, 225, 293
263, 292 Histology ........................37, 38, 50, 166, 172, 196, 206,
Cell isolation.......................................... 45–47, 51, 90, 97 207, 214–217, 236, 256, 263, 286
Chondrocytes ................................ 36, 41, 42, 48, 53–69,
Histomorphometry .............................165, 171, 182, 286
96, 111, 118, 119, 142, 254 Hydroxymethylcytosine ....................................... 101–107
Chondrogenesis ..................................15, 18, 20, 21, 102
Closed fracture ..................................................... 205, 206 I
Co-culture ................................53, 54, 58–61, 63, 65, 68
Collagenase................................ 4–11, 31, 37, 43, 46, 47, Immunofluorescence (IF).......................... 114, 116, 135,
50, 56, 57, 91, 232, 255 229, 230, 236, 239, 245, 247, 256, 261, 265,
Colony forming unit-fibroblast (CFU-F).............. 30, 33, 268–270
37, 38 Immunohistochemistry (IHC).................. 183, 208, 216,
Cryopreservation............................................................. 72 221, 229, 230, 236, 239, 242, 244–246, 256,
Cultures .............................. 5, 6, 8, 9, 11, 16, 19–21, 23, 261, 264, 266–268
25, 26, 29–38, 42–44, 46, 48–50, 54, 58–61, 63, Immunostaining........................... 21, 196, 237, 261–263
67, 68, 72, 74–78, 80–82, 86, 87, 109–137, 145, Incapacitance ............................................... 292, 294, 297
151, 168, 169, 177, 178, 224, 230 Induced pluripotent stem cells (iPSCs) ..................71–73,
78, 80–85
D Intervertebral disc (IVD) .........................................41–51
Isolation ........................ 3–11, 18, 29–38, 42, 43, 51, 57,
Decalcification .............................21, 183, 207, 214, 220, 89, 90, 96, 97, 202, 208–209, 217–219,
221, 228–229, 237, 244, 262–264, 266–268, 270 230–232, 247–254, 257
https://doi.org/10.1007/978-1-0716-0989-7, © Springer Science+Business Media, LLC, part of Springer Nature 2021
301
OSTEOPOROSIS AND OSTEOARTHRITIS
302 Index
M P
Matrix deposition......................................................53, 54 Pain ..........172, 202, 203, 206, 225, 255, 291–293, 298
Mechanical allodynia..................................................... 292 Protein dynamics.................................110, 121, 123, 132
Mechanical loading ..................................... 225, 275, 276
Mesenchymal stem cells .......................29, 54, 72, 74, 78, R
80–82, 96, 98, 111, 119 Regenerative medicine ................................................ v, 54
Mesenchymal stromal cells (MSCs) ................. 15–27, 29, Reprogramming ..............................71–73, 78, 80–82, 85
53–69, 78, 80
RNA and DNA extraction ..................230–232, 247, 248
Micro-computed tomography (micro-CT/μCT) ......206,
207, 210, 276 S
Mouse embryonic fibroblasts (MEFs) ...... 71, 73, 75–78,
80–82, 86 Single cell sequencing ......................................... v, 89–100
Site specificity ...................................... 276–278, 282–287
N Static weight bearing .................................. 292, 294, 298
Surgical models .................................................... 223–257
Next-generation sequencing ............................... 102–104
Nucleus pulposus (NP).............................................41–51 T
O Tissue engineering ................................................. 54, 154
Torsional testing............................................................ 207
Open field activity assays............................................... 298
Transcription factor activity.......................................... 110
Osteoarthritis ............................101, 102, 142, 165, 166, Trophic effects...........................................................54, 55
172, 173, 182, 183, 223–258, 291
Osteoblasts ................................4, 5, 7, 8, 10, 11, 30, 36, V
89, 96, 157, 276
Osteocytes ...........................................3–12, 97, 198, 276 Von Frey .......................................................292–296, 298

2020 Osteoporosis and Osteoarthritis.-humana

Uploaded by

Copyright:

Available Formats

You might also like

2020 Osteoporosis and Osteoarthritis.-humana

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

2020 Osteoporosis and Osteoarthritis.-humana

Uploaded by

Copyright:

Available Formats

Methods in

Molecular Biology 2221

Andre J. van Wijnen

For further volumes:

Andre J. van Wijnen and Marina S. Ganshina

ISSN 1064-3745 ISSN 1940-6029 (electronic)

© Springer Science+Business Media, LLC, part of Springer Nature 2021

Musculoskeletal regenerative medicine and orthopedic research examine molecular mechan-

Rochester, MN, USA Andre J. van Wijnen

PART I CELLULAR AND MOLECULAR BIOLOGY OF OSTEOARTHRITIS

1 Isolation of Murine and Human Osteocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

PART II IN VIVO MODELS OF SKELETAL TISSUE INJURY, DEGENERATION,

11 Generation and Characterization of Mouse Models

NIDHI BHUTANI • Department of Orthopaedic Surgery, Stanford University, Stanford, CA,

CRYSTAL IDLEBURG • Musculoskeletal Research Center, Histology and Morphometry Core,

AMBER RATH STERN • Engineering Systems Inc., Charlotte, NC, USA

Cellular and Molecular Biology of Osteoarthritis

Isolation of Murine and Human Osteocytes

Key words Osteocyte, Isolation, Age, Culture, Collagenase, Mice

bones of 16-day-old chick embryos [4]. Osteocyte morphology

hypermineralized bone, they produced a very low yield rate mainly

This technique facilitates the isolation of osteocytes from skeletally

2.1 Isolation 1. Collagenase solution: Dissolve 300 active units/mL of collage-

2. EDTA solution: Prepare the 5 mM ethylenediaminetetraacetic

2.2 Isolation 1. Collagenase solution: Dissolve 2 mg/mL (approx.

2. EDTA/BSA solution: Dissolve tetra sodium EDTA in calcium

The osteocyte isolation protocol from mouse bone takes approxi-

4. Wash the bones in sequential dishes/wells of a six-well plate

25. Collagenase Treatment 5: Incubate the bone pieces in warmed

9. Add 40 mL EDTA solution (5 mM in HBSS with 0.1% BSA,

1. The collagenase solution must be prepared fresh the morning

9. These will be a mix of osteoblastic and osteocytic cells.

Expansion and Chondrogenic Differentiation of Human Bone

Key words Tissue engineering, Regenerative medicine, Cartilage, Chondrogenesis, Mesenchymal

In this manuscript we describe the materials and methods we use to

previous book chapter (https://doi.org/10.1007/978-3-319-

All the experiments regarding the manipulation of cells of human

Materials Company Catalogue

2.2 Chondrogenic 1. Complete chondrogenic medium (CCM; see Note 4):

Materials Company Catalogue

(c) 40 μg/mL L-Proline.

The most commonly used method for the isolation of BM-MSC

3.2 Chondrogenic BM-MSC are trypsinized as indicated in Subheading 3.1. However,

3.3.1 Histological After chondrogenic differentiation we generally fixate the pellet

If not differentially indicated, all the components need to be stored

thawed and 995 μL of CEM (without ascorbic acid-2-

6 months. Once thawed, the stock solution can be stared

(h) Ascorbic-acid preparation: dissolve the powder in

9. Unless specifically requested by the research question, we gen-

This work is part of Medical Delta RegMed4 program.

A Novel Enzymatic Digestion Approach for Isolation

Almost a half century ago, Alexander Friedenstein and colleagues

tissue regeneration. Therefore, they have been intensively investi-

2. Tissue culture incubator with temperature and gas composi-

2.3 Reagents 1. 70% Ethanol.

9. Chondrogenic medium: High glucose DMEM, 0.1 μM dexa-

3.1 Harvest 1. Euthanize the animal by CO2 inhalation.

endosteum remain attached to the bone after the flushing and

13. Perform a cell count by diluting a cell aliquot 1:10 in 3% acetic